Graphene oxide/cellulose membranes in adsorption of divalent metal ions

Abstract

In this paper, graphene oxide/cellulose membranes were prepared in order to perform effective adsorption of heavy metal ions: cobalt, nickel, copper, zinc, cadmium and lead. Two types of membranes were fabricated, i.e. pressed and non-pressed membranes. The experiment showed that the pressed membranes are highly durable at different pH values, even in basic solutions, and they can be applied in separation/removal of heavy metal ions during vigorous shaking in the aqueous solution. The non-pressed membranes were proved to be less stable, however, they can be successfully applied in the filtration process at the high flow-rates. The results of the batch experiments and the measurements by the inductively coupled plasma atomic emission spectroscopy (ICP-OES) indicated that the maximum adsorption can be achieved at pH 4–8. Adsorption isotherms and kinetic studies indicated that the sorption of the metal ions on the membranes occurs in a monolayer coverage, hence it is controlled by the chemical adsorption involving the strong surface complexation of metal ions with the oxygen-containing groups on the surface of graphene oxide. The maximum adsorption capacity values of Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and Pb(II) on the graphene oxide/cellulose membranes at the pH of 4.5 are 15.5, 14.3, 26.6, 16.7, 26.8, 107.9 mg g⁻¹, respectively. The competitive adsorption experiments showed the affinities of prepared membranes for the metal ions in the order of Pb > Cu > Cd > Zn ≥ Ni ≥ Co. The affinity order agrees with the first stability constant of the associated metal hydroxide and acetate. The adsorption properties of the graphene oxide/cellulose membranes, their reusability (more than 10 cycles) and durability in the aqueous solutions open the path to removal of heavy metals from water solution. The membranes can be also used in the field of analytical chemistry for the preconcentration and/or separation of trace and ultratrace metal ions.

---

Introduction

Over the last few years, graphene and its derivative graphene oxide (GO) have become two of the most intensively studied materials. This results from their exceptional electrical, electrochemical, optical, mechanical and chemical properties. Graphene has a single- or a few-layer thickness of sp²-hybridized carbon atoms arranged in a honeycomb pattern. GO, prepared by the strong oxidation of graphite has large quantities of oxygen atoms present on its surface in the forms of epoxy, hydroxyl, and carboxyl groups. Both graphene and GO have large surface area, which results in high adsorption capacity. However, the adsorptive properties of graphene and GO are completely different.¹ Graphene is a hydrophobic adsorbent with the strong affinity for carbon-based ring structures due to the strong van der Waals interactions. In contrast to graphene, GO is a highly hydrophilic adsorbent. Functional groups containing oxygen atoms have a lone electron pair and they can efficiently bind a metal ion to form a metal complex. Recently published papers have shown that GO has impressive adsorption properties toward metal ions, e.g.: 842–1119 mg g⁻¹ Pb(II),^2,3 68 mg g⁻¹ Co(II),⁴ 106–345 mg g⁻¹ Cd(II),^3,4 175 mg g⁻¹ Eu(III),⁵ 98–299 mg g⁻¹ U(VI),^6,7 118–294 mg g⁻¹ Cu(II).^3,8 The suitable treatment of the GO surface can also improve the adsorption capacity or enhance the selectivity toward metal ions.^9–13 Although GO has impressive adsorption capacities toward metal ions, its practical use in the metal removal can be seriously hampered.

It is basically due to the fact that GO forms very stable suspensions and very small sizes of GO particles can cause high pressure in the filtration systems. The difficulties associated with very small sizes of GO particles can be overcome by covalent binding of the GO nanosheets to the support^14,15 or by magnetic nanocomposites preparation^16–18 that can be recovered from the solutions by an external magnetic field. Another solution is to use the membrane technology. Although, dozens of papers devoted to the fabrication of GO-based membranes and their application in various technological areas, e.g. gas^19,20 or molecular separation,^21,22 fuel cells,²³ biosensors,²⁴ have been published, the literature data of adsorption of heavy metal ions from the aqueous solutions are still limited.

The application of the GO membranes for the purposes of metal ions separation/removal from the aqueous solutions requires highly controlled pH since the adsorption process depends highly on the protonation and deprotonation of the binding sites of the functional groups present on the GO surface and on the metal species present in the solution. Unfortunately, GO nanosheets can form very unstable membranes at desired pH due to the negative charges generated on the GO surface, being a result of the electrostatic repulsion between GO nanosheets. As a consequence, the GO membranes become nondurable in the aqueous solutions. Yeh et al. demonstrated that the multivalent cationic metal, i.e. Al³⁺, can effectively cross-link the GO nanosheets and strengthen the membrane to be obtained.²⁵ However, the authors observed a strong relation between the property of the GO membranes and the pH values. The membranes are highly stable in acidic solutions due to Al³⁺ cross-linking. In neutral and basic solutions, the membranes can still disintegrate in water because of the inability of neutral Al(OH)₃ and the anionic Al(OH)₄⁻ to cross-link the negatively charged GO nanosheets. Much more stable GO membranes can be fabricated by the cross-linking using ethylenediamine (EDA),²⁶ dicarboxylic acids, diols or polyols.²⁷ Nellore et al. developed CNT-bridged GO membrane using bis amine poly(ethylene glycol) as cross-linking agent.²⁸ The experiments show that CNT-bridged GO membranes modified with glutathione (γ-L-glutamyl-L-cysteinyl-glycine) can be used for the efficient removal of As(III), As(V), and Pb(II) ions from water samples. Stable GO membranes can be also prepared using poly(vinyl alcohol) (PVA). Then, the oxygen-containing functional groups on the GO surface can form strong hydrogen bonding with hydroxy groups on the PVA chains.²⁹ Such membranes were used for the removal of Cu(II), Cd(II) and Ni(II) from the aqueous solutions.³⁰ Stable GO impregnated mixed matrix membranes were prepared using polysulfone for the adsorption of heavy metal ions, Cr(VI), Cu(II), Pb(II) and Cd(II).³¹ The literature shows that GO-based membranes can be also stabilized by the hydrogen bonding with cellulose as matrix.³² However, the adsorptive properties of GO/cellulose composite membranes toward metal ions and their stability in aqueous solutions have not been studied.

In this paper, the GO/cellulose membranes were prepared for effective separation/removal of divalent metal ions. Two types of GO/cellulose membranes were fabricated, i.e. pressed and non-pressed membranes. The paper shows that the pressed GO/cellulose membranes are highly durable in a broad pH range. Therefore, they can be applied for the adsorption of heavy metal ions during vigorous shaking in the aqueous solutions. The non-pressed membranes are less stable, however, they can be successfully used in the filtration process at high flow-rates.

Experimental section

Reagents and solutions

Metal stock solutions were purchased from Sigma-Aldrich; nitric acid (65%, p.a.), sulfuric acid (98%, p.a.), ethanol (p.a.), ammonium hydroxide solution (25%, p.a.), potassium manganate (p.a.), urea (p.a.), and sodium nitrate (p.a.) were from POCh (Gliwice, Poland); sodium hydroxide (p.a) was purchased from Chempur; microcrystalline cellulose (ashless quality) was from Macherey-Nagel (Düren, Germany); 0.45 μm nitrocellulose membrane filters were purchased from Merck Millipore. Standard solutions were diluted with high purity water obtained from Milli-Q system (Millipore, Molsheim, France).

Instruments

The microstructural observations were conducted on a JEOL-7600F scanning electron microscope (SEM) equipped with the Oxford X-ray energy-dispersive spectrometer (EDS). The chemical composition of GO was analyzed by X-ray photoelectron spectroscopy (XPS). The measurements were performed using a PHI 5700/660 Physical Electronic spectrometer with a monochromated Al Kα radiation. The spectra were analyzed with a hemispherical mirror assuring energy resolution of 0.3 eV. Three hours after placing the samples in situ at 10⁻¹⁰ hPa vacuum, their surface was clean enough for the measurements. The binding energy in the range −2 to 1400 eV and the core-level characteristic peak for C 1s were measured. The background was subtracted using the Tougaard approximation. Powder diffraction data (XRD) were collected on X'Pert PRO X-ray diffractometer with PIXcel ultrafast line detector and Soller slits for Cu Kα radiation. The measurements were done in Bragg–Brentano geometry. The Raman spectra were taken at room temperature using RenishawInVia Raman spectrometer equipped with confocal DM 2500 Leica optical microscope, a thermoelectrically cooled Ren Cam CCD as a detector, and a diode laser operating at 830 nm. ICP-OES, an optical emission spectrometer, ICP Model Spectroblue (Spectro Analytical Instruments GmbH, Germany, http://www.spectro.com) was used for the determination of metal ions. The spectrometer was used with the following parameters: polychromator: focal length 750 mm, holographic master grating, wavelength range 165–770 nm; detector: thermally stabilized optical system, parallel readout architecture; generator: frequency 27.12 MHz, power 0.7–1.7 kW, air cooled; nebulizer Cross Flow type; exhaust system requirements: torch box: 200–300 m³ per h, generator: 250–300 m³ per h. The wavelengths were Pb: 283.305 nm, Zn: 213.856 nm, Ni: 231.604 nm, Co: 230.786 nm, Cu: 327.396 nm, Cd: 214.438 nm.

The batch adsorption experiments were performed using orbital platform shaker Unimax 1010 (Heidolph, Germany), while experiments under controlled flow-rate with the use of multi-channel peristaltic pump PD 5201 (Heidolph, Germany).

Synthesis of GO

GO was synthesized by Hummers' method:³³ 3.0 g of graphite flakes and 1.5 g of NaNO₃ were added to 70 mL of concentrated H₂SO₄. The mixture was cooled to 0 °C, and 9 g of KMnO₄ was added slowly in portions to keep the reaction temperature below 20 °C. Then, the mixture was warmed to 35 °C and stirred for 12 h. The reaction mixture was cooled to room temperature and poured onto 400 mL of ice with 3 mL of 30% H₂O₂. Next, the solid product was separated by centrifugation at 5000 rpm and washed 20 times with water and 30 times with 5% HCl. Each time, the solid was redispersed by ultrasonication and collected by centrifugation. Next, the solid was washed with deionized water. The centrifugation and ultrasonication with a new portion of deionized water were recycled ca. 20 times until the solution was neutral. The obtained GO was dried at 100 °C.

Preparation of GO/cellulose membranes

100 mg of GO was dispersed in 70 mL of water by a 30 minute sonication. Then, 100 mg of urea, 150 mg of sodium hydroxide and 100 mg of microcrystalline cellulose were added to the suspension of GO. The mixture was sonicated for 15 min. After dissolution of cellulose, 1 mL of 30% H₂SO₄ was added, and the mixture was sonicated for 15 min. The suspension was put into 100 mL flask and filled with water up to the mark. Directly before use the mixture was dispersed for 30 min in an ultrasonic bath in order to homogenize the suspension. To prepare GO/cellulose membrane (containing 2 mg of GO) 2 mL of the mixture was passed through a Millipore membrane (25 mm diameter, 0.45 μm pore size) using a filtration assembly (25 mm, Sigma-Aldrich). The layer of GO/cellulose collected onto the membrane was washed with redistilled water. The prepared membranes were dried for 24 h in room temperature. Finally, the membranes were pressed under 130 kN for 1 min.

Batch adsorption experiment

The batch adsorption experiments (the effect of pH, the effect of contact time and adsorption isotherms) were carried out using 25 mL of single-metal aqueous solutions with the desired pH and metal concentration. The pH values of the solutions were adjusted with nitric acid or ammonia solutions. Then, the solution with GO/cellulose membrane were shaken (250 rpm) within 15–180 min (kinetic study) or shaken for 180 min to achieve adsorption equilibrium (in the case of adsorption isotherms). The amount of metal ions adsorbed on the membrane calculated from the difference between the initial concentration C₀ (mg L⁻¹) and equilibrium concentration C_e (mg L⁻¹), was determined by the ICP-OES: q_e = (C₀ − C_e)V/m_GO, where V is the volume of the suspension, and m_GO is the mass of GO. The initial concentrations C₀ in kinetic study, pH and flow-rate effects were 0.25 mg L⁻¹ of each metal ions. In the case of adsorption isotherms, C₀ was as follows: 0.5–20 mg L⁻¹ Cu(II), 1–12 mg L⁻¹ Zn(II), 1–14 mg L⁻¹ Co(II), 0.5–17 mg L⁻¹ Ni(II), 1–20 mg L⁻¹ Cd(II), 4–40 mg L⁻¹ Pb(II).

Results and discussion

Characterization of GO and GO/cellulose membranes

The GO used for the synthesis of membranes were characterized by the XPS, XRD and Raman spectroscopy. The high-resolution C 1s spectra are presented in Fig. 1a. The C 1s spectrum of graphite shows the main peak at 284.4 eV due to the graphitic carbon, whereas spectrum of GO reveals additionally other four peaks at 285.8, 286.9, 288.1, and 289.5 eV assigned to C–OH, C–O–C, C=O, and O–C=O.^2,34 Fig. 1b presents the XRD patterns, which are characteristic of graphite and GO.^35,36 The intense 002 peak in graphite at 2θ = 26.5° corresponds to coherently scattering hexagonal carbon layers with d₀₀₂ spacing of 3.4 Å. In GO the interlayer distance growths up to 6.7 Å due to the intercalation of the functional groups into the carbon sheets and manifests as a broad and less intense peak at 13.14°. From the XRD pattern of GO, it could be inferred that the original graphite powders had been completely oxidized. Additionally, the thickness of coherently scattering layers in GO is reduced to ∼5 nm compared to ∼60 nm graphite sheets. Fig. 1c presents the Raman spectra characteristic of graphite and GO.³⁷ The main features of the spectra are two prominent bands which correspond to the so-called G and D modes. The G band at 1582 cm⁻¹ and 1600 cm⁻¹ in graphite and GO, respectively, is related to the E_2g vibrational mode of ordered in-plane sp² carbons. The remaining bands, D and D′, are related to the disorder of the edge carbons.³⁸ They appear at 1312 cm⁻¹ (D), 1613 cm⁻¹ (D′ shoulder band) in graphite and at 1322 cm⁻¹ (D) in GO. In GO the D′ is merged with the G band. The broadening of all bands in GO as well as the increase in the relative intensity of the D band compared to that of the G band confirm higher disorder of edge carbons in GO. The disorder arises from the small size of the GO sheets, as it is indicated by the XRD; the presence of the oxygen functional groups as well as the number of areas of sp² carbons with alternating patterns of single-double carbon bonds generate blue shift of the G band (from 1582 to 1600 cm⁻¹).³⁹ Fig. 2 shows photograph and SEM images of synthesized GO/cellulose membrane. As can be seen, GO on the surface of membranes has highly wrinkled structure. This structure in concert with its softness and flexibility of GO nanosheets allow achieving excellent contact with analyzed solution in flow conditions.^15,40

Fig. 1 High-resolution XPS spectra of C 1s (a), XRD patterns (b) and Raman spectra (c) obtained for graphite and GO.

Fig. 2 Photograph (a) and SEM images (b–d) of the GO/cellulose membranes.

The durability of prepared GO/cellulose membranes was checked during vigorous shaking (250 rpm) in aqueous solution at different pH values within 180 min. As can be seen in Fig. 3 the pressed GO/cellulose membranes are highly durable at different pH, even at high pH values. Another situation can be observed for non-pressed membranes. In this case, a thin layer of GO/cellulose was peeled off from the surface of nitrocellulose during vigorous shaking. Although, the non-pressed GO/cellulose membranes cannot be applied, when vigorous shaking is applied, they can be successfully used in filtration process (Subsection Kinetic study). In our research we also checked the durability of pressed and non-pressed GO thin layer (without cellulose) collected onto the nitrocellulose membranes. As can be seen the GO membranes are not durable. Moreover, their stability strongly depends on pH values. This is in accordance with our previous studies on the stability of GO dispersion in water.³ In acidic solution GO has a tendency to agglomerate and precipitate. At pH above pH_pzc (point of zero charge; pH_pzc of GO is 3.8–3.9),² the surface charge of GO is negative. Then, the negative charges generated on the GO surface result in electrostatic repulsion between GO nanosheets. As a consequence, GO membranes become nondurable.

Fig. 3 Stability of pressed and non-pressed GO and the GO/cellulose membranes in aqueous solutions at different pH during vigorous shaking.

Adsorption of metal ions on GO/cellulose membranes

Effect of pH. The adsorption of metal ions depends on the protonation and deprotonation of binding sites of the chelating molecules onto the surface of GO/cellulose membrane, as well as on the metal species present in the solution. Therefore, the acidity of the analyzed sample plays an important role in adsorption of metal ions. The adsorption of metal ions on GO/cellulose membranes was investigated at pH ranging from 1 to 8. As can be seen from Fig. 4, the adsorption of Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and Pb(II) increases quickly at pH 2–4 and reaches the maximum value, and then remains constant at pH 4–8. At low pH, the protonation of the oxygen functional groups of GO reduces the ability in chelating with metal cations which leads to fewer adsorption capacities. Moreover, at pH below 4 (pH_pzc of GO), the positive charges generated on the GO surface result in the electrostatic repulsion. At pH > 4, the positive charge density on the surface sites of the GO decreases which leads to the decrease in the repulsion between the membrane surface and metal cations that are predominant at pH 1–6.¹¹ Of course, the mechanism of chelation is also possible, besides the electrostatic interactions. The metal ions can be chelated by the neighboring carboxyl and hydroxyl groups remaining on the GO surface. Therefore, some adsorption is also observed at pH < 4 despite the low dissociation of the functional groups and competition between H⁺ and metal ions. The adsorption percentage at pH 3 equals 51, 57, 76, 39, 63 and 88% for, Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and Pb(II), respectively. The highest adsorption percentage of Cu(II) and Pb(II) at pH 3 can be explained by the strong affinity of these ions to the GO. Taking into account that the high adsorption of metal ions from aqueous solution (pH from 4 to 8) and in order to prevent precipitation of metal hydroxides, the next adsorption experiments were performed at pH 4.5. Under such conditions, the adsorption of metal ions is dominated by the strong surface complexation of metal ions with the oxygen-containing groups on the surface of GO. It should be noticed that divalent metal ions Me²⁺ can form ammine complexes Me(NH₃)²⁺, Me(NH₃)₂²⁺, Me(NH₃)₃²⁺, etc. if ammonia is present in solution (e.g. due to the pH adjustment using ammonia solutions).⁴¹ However, the next experiments (kinetics, adsorption isotherms, competitive adsorption) were performed at pH 4.5 using only nitric acid to adjust pH.

Fig. 4 Effect of pH; conditions: C₀ = 0.25 mg L⁻¹, T = 25 °C, time = 180 min, (m V⁻¹) = 80 mg L⁻¹.

Kinetic study. The contact time between the metal ions and the membrane is crucial to obtain high adsorption percentage. Because of the high stability of pressed membranes in aqueous solutions (Subsection Characterization of GO and GO/cellulose membranes) the kinetic study was performed using bath adsorption experiment under vigorous shaking. Fig. 5a shows that the adsorption of metal ions increases remarkably at the beginning of the experiment and then reaches the equilibrium state. The percentage adsorption reached 90% after 130 min for Co(II), Ni(II), Cu(II), Zn(II), Cd(II), and after 105 min for Pb(II). The adsorption at 95% requires 180 min for Co(II) and Ni(II), 160 min for Cu(II) and Zn(II), 150 min for Cd(II), and 130 min for Pb(II). It is worth noting here that the time required to reach the equilibrium is longer than in the case of GO dispersed in aqueous solutions. Then percentage adsorption reached 90% after 4–15 min for Cu(II), Zn(II), Cd(II) and Pb(II), respectively.³ It is due to such features of GO nanosheets as large surface area, wrinkled structure and excellent dispersibility in water allow achieving very good contact with solution. As a consequence the adsorption of metal ions and the equilibrium state is achieved very quickly. The kinetics of an adsorption on GO/cellulose membranes was studied using the pseudo-second order rate adsorption kinetic model:⁴²
where q_e and q_t (mg g⁻¹) are the capacities of metal ions adsorbed at the equilibrium and time t (min), respectively, k₂ is the pseudo-second-order rate constant (g mg⁻¹ min⁻¹). The calculated kinetic parameters for the adsorption of metal ions on the GO/cellulose membranes at initial concentrations of 0.25 mg L⁻¹ of metal ions are presented in Table 1. As can be seen, the experimental data for the adsorption of metal ions are very well fitted by the this kinetics model (r = 0.994–0.999). Moreover, the experimental q_e values (3.125 mg g⁻¹) are close to q_e values calculated from the pseudo-second-order kinetic model. Data may indicate that the adsorption of metal ions on the GO/cellulose membranes is controlled by the chemisorption, which involves the surface complexation with the oxygen-containing groups on the surface of GO. Therefore, the adsorption capacity is proportional to the number of active sites occupied on the membrane. The adsorption of metal ions was also studied under flow condition. Fig. 5b shows that the adsorption higher than 90% can be obtained: for Co(II) and Ni(II) at flow-rate of 12.5 mL min⁻¹, for Cu(II), Zn(II) and Cd(II) at flow-rate of 25 mL min⁻¹, and for Pb(II) at flow-rate up to 38 mL min⁻¹. The adsorption of Pb(II) reaches the maximum value of 100% and remains constants at a wide range of flow-rate, i.e. 0.8–25 mL min⁻¹.

Fig. 5 Kinetics (a) and flow-rate effect (b); conditions: C₀ = 0.25 mg L⁻¹, pH = 4.5, T = 25 °C, (m V⁻¹) = 80 mg L⁻¹.

Table 1 Kinetic parameters of pseudo-second-order equation for adsorption of Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and Pb(II) on the GO/cellulose membranes

Metal ion	qe (mg g^−1)	k2	r
Co(II)	3.47 ± 0.094	0.0094 ± 0.00089	0.9981
Ni(II)	3.57 ± 0.17	0.0076 ± 0.0011	0.9941
Cu(II)	3.86 ± 0.080	0.0053 ± 0.00024	0.9989
Zn(II)	3.51 ± 0.14	0.0097 ± 0.0015	0.9959
Cd(II)	3.88 ± 0.13	0.0055 ± 0.00043	0.9970
Pb(II)	4.02 ± 0.18	0.0055 ± 0.00056	0.9952

---

Adsorption isotherms. The adsorption of metal ions on the GO/cellulose membranes was simulated using Langmuir^43,44 and Freundlich⁴⁵ isotherm models:

q_e = K_FC_e^1/n (1)

where q_max is the maximum amount of metal ions adsorbed per unit weight of GO at the high equilibrium concentration of metal ions (mg g⁻¹), K_L is the constant related to the free energy of adsorption (L mg⁻¹), K_F (mg¹⁻ⁿ Lⁿ g⁻¹) and n are Freundlich constants related to the adsorption capacity and the adsorption intensity, respectively. The Langmuir and Freundlich adsorption isotherms are presented in Fig. 6. Isotherm parameters obtained by fitting the adsorption equilibrium data to the isotherm models are included in Table 2. It can be noticed that the adsorption isotherms are better fitted by the Langmuir model than by the Freundlich model, which suggests a chemical adsorption process, i.e. the metal ions can be chelated by the neighboring oxygen-containing groups on the surface of GO. Fig. 7 shows the SEM image of the GO/cellulose membranes after adsorption of Pb(II) ions. It can be noticed that distribution of Pb on the GO nanosheets remaining onto the surface of GO/cellulose membrane is rather homogeneous. The maximum adsorption capacity q_max values of Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and Pb(II) on GO/cellulose membranes are 15.5, 14.3, 26.6, 16.7, 26.8, 107.9 mg g⁻¹, respectively. These values are lower than that for the pure GO dispersed in the aqueous solution. It can result from the lower surface area and from the fact that cellulose can cover the GO nanosheets in some degree and as a consequence all active sites of GO are not available for adsorption of metal ions.

Fig. 6 The Langmuir and Freundlich adsorption isotherms; conditions: pH = 4.5, T = 25 °C, time = 180 min, (m V⁻¹) = 80 mg L⁻¹.

Table 2 Parameters of Langmuir and Freundlich models for adsorption of Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and Pb(II) on the GO/cellulose membranes

Metal ion	Langmuir			Freundlich
	qmax	KL	R^2	KF	n	R^2
Co(II)	15.5 ± 0.22	1.72 ± 0.10	0.9441	10.0 ± 0.10	5.9 ± 0.39	0.9351
Ni(II)	14.3 ± 0.40	5.87 ± 0.35	0.9466	9.9 ± 0.10	5.9 ± 0.34	0.9426
Cu(II)	26.6 ± 0.39	6.06 ± 0.47	0.9956	20.8 ± 0.23	9.3 ± 1.42	0.7491
Zn(II)	16.7 ± 0.15	2.89 ± 0.13	0.9670	12.5 ± 0.14	8.3 ± 1.1	0.7877
Cd(II)	26.8 ± 0.22	3.51 ± 0.17	0.9959	20.4 ± 0.14	8.6 ± 0.79	0.8265
Pb(II)	107.9 ± 1.94	2.13 ± 0.09	0.9743	68.1 ± 0.53	5.9 ± 0.53	0.8829

---

Fig. 7 SEM image of GO nanosheet with adsorbed Pb(II) ions.

The affinity of metal ions to the GO/cellulose membranes. The difference in the affinity of metal ions to the GO/cellulose membranes was studied by the competitive adsorption using fifteen binary mixtures: Pb/Cd, Pb/Ni, Pb/Cu, Pb/Zn, Pb/Co, Ni/Co, Ni/Zn, Ni/Cd, Ni/Cu, Zn/Cu, Zn/Cd, Zn/Co, Cd/Co, Cd/Cu, Cu/Co. The binary mixtures contained the metal ion in initial amount related to the maximum capacity of GO/cellulose membranes for a single metal at pH 4.5. The experimental data for binary mixtures presented in Fig. 8 show that the affinities of the GO/cellulose membranes for these metal ions follow the order of Pb > Cu > Cd > Zn ≥ Ni ≥ Co. The affinity order agrees with the first stability constant of the associated metal hydroxide (Me²⁺ + OH⁻ ↔ Me(OH)⁺; log K₁ = 7.82, 7.0, 4.17, 4.4, 4.97 for Pb(OH)⁺, Cu(OH)⁺, Cd(OH)⁺, Zn(OH)⁺, Ni(OH)⁺, respectively),⁴⁶ and the first stability constant of the associated metal acetate (Me²⁺ + Ac⁻ ↔ Me(Ac)⁺; log K₁ = 2.52, 2.16, 1.5, 1.5, 1.12, 1.5 for Pb(Ac)⁺, Cu(Ac)⁺, Cd(Ac)⁺, Zn(Ac)⁺, Ni(Ac)⁺, Co(Ac)⁺, respectively).⁴⁶ The correlation between the affinity of metal ions to the GO/cellulose membranes and the first stability constant of the associated metal hydroxide or acetate can be explained by the mechanism of metal adsorption based on the complexation of metal ions with surface functional groups (e.g. –OH and –COOH). The same affinity order was obtained for the density functional theory studies,⁴⁷ and experimental data for suspension of GO.³ It is also worth noting here that the similar results were also observed for the adsorption of divalent metal ions onto the inorganic oxides and organic surfaces.⁴⁸ In the case of inorganic oxides, a strong correlation between the pH at which 50% of the metal had been removed from solution (pH₅₀) and the first hydrolysis constant was observed. For organic surfaces, pH₅₀ agreed very well with the metal–carboxyl ligand stability constants. The same results were obtained for the metal ion adsorption onto the biological substrates, which suggests the complexation of metal ions with surface functional groups.^49,50

Fig. 8 The affinity of metal ions to the GO/cellulose membranes; conditions: pH = 4.5, T = 25 °C, time = 180 min, (m V⁻¹) = 80 mg L⁻¹.

Effect of ionic strength. Fig. 9 shows that the maximum adsorption capacity of the GO/cellulose membranes decreases with the ionic strength. This effect can result from the following aspects:^30,51,52 (i) the number of active sites on the adsorbent surface can be reduced due to the adsorption of the ionic species, (ii) the ionic strength influences on the activity coefficient of the metal ions and thus limits the metal ion transfer from the solution to solid surfaces,⁴ and (iii) the ionic strength can influence particle aggregation by emerging the electrostatic interactions.⁵³

Fig. 9 Effect of the ionic strength; conditions: pH = 4.5, T = 25 °C, time = 180 min, (m V⁻¹) = 80 mg L⁻¹.

Desorption and reuse studies. For the potential practical application, it is important to study the possibility of desorbing the metal ions adsorbed on the GO/cellulose membranes for the recycling purposes. The stability and potential reusability of the GO/cellulose membranes was assessed by monitoring the change in recovery of Pb(II) through several adsorption/elution cycles. For each cycle, the aqueous solution of pH 4.5 containing Pb(II) ions was passed through the membrane, then Pb(II) ions were eluted with 0.1 mol L⁻¹ HNO₃, and finally the membrane was conditioned with acetate buffer of pH 4.5. The results presented in Fig. 10 show that the GO/cellulose membrane can be reused at least ten times without any loss in its adsorption capacity.

Fig. 10 Reuse of the GO/cellulose membranes; conditions: C_Pb(II) = 0.25 mg L⁻¹, pH = 4.5, T = 25 °C, (m V⁻¹) = 80 mg L⁻¹, eluent: 0.1 mol L⁻¹ HNO₃.

Conclusions

Two type of GO/cellulose membranes (pressed and non-pressed membranes) for highly effective adsorption of metal ions were fabricated. The research shows that the pressed membranes are highly durable at different pH values. Despite the negative charges generated on the GO surface and electrostatic repulsion between GO nanosheets, the membranes are also highly stable in basic solutions. As a consequence, the pressed membranes can be applied in separation/removal of heavy metal ions during vigorous shaking in the aqueous solution. The non-pressed membranes are less stable, however, they can be successfully used in filtration process at high flow-rates.

Adsorption isotherms and kinetic studies suggest that sorption of metal ions on membranes is monolayer coverage and adsorption is controlled by the chemical adsorption involving the strong surface complexation of metal ions with the oxygen-containing groups on the surface of GO. The competitive adsorption experiments showed the affinities of prepared membranes for the metal ions follow the order of Pb > Cu > Cd > Zn ≥ Ni ≥ Co. The affinity order agrees with the first stability constant of the associated metal hydroxide and acetate. The maximum adsorption capacity values of Co(II), Ni(II), Cu(II), Zn(II), Cd(II) and Pb(II) on the GO/cellulose membranes at pH 4.5 are 15.5, 14.3, 26.6, 16.7, 26.8, 107.9 mg g⁻¹, respectively. Reviewing the recently published papers (see Table 3) these values can be considered as satisfactory. Moreover, GO/cellulose can be regenerated several times, more than 10 cycles. The adsorption properties of the GO/cellulose membranes, their reusability (more than 10 cycles) and durability in aqueous solutions open the path for the removal of heavy metals from water solution. The membranes can be also used in analytical chemistry for the preconcentration and/or separation of trace and ultratrace metal ions.

Table 3 Recently published papers on application of membranes in adsorption of metal ions^a

Membrane	Metal ion	Maximum adsorption capacity, mg g^−1	pH	Regeneration, number of cycles	Ref.
a AAO – anodic aluminum oxide, PSF – polysulfone, PVA – poly(vinyl alcohol), PAA – polyacrylic acid, PVT – polyvinyltetrazole, PVDF – polyvinylidene fluoride, A-HNTs – halloysite nanotubes functionalized with 3-aminopropyltriethoxy-silane, PVT – polyvinyltetrazole, PAN – polyacrylonitrile, PEO – poly ethylene oxide, PAMAM – polyamidoamine.
Chitosan	Cu(II)	87.5	5	5	54
NH2-MCM-41	Cu(II)	3.7	6	3	55
PVT-co-PAN	Cu(II)	44.3	5	—	56
PSF/Fe2O3	Pb(II)	13.2	7	—	57
Pb-imprinted PVA/PAA	Pb(II)	207	5	6	58
AAO-polyrhodanine	Pb(II)	481	6	5	59
A-HNTs@PVDF	Cu(II), Cd(II)	0.50, 0.45	5.5	3	60
Nafion/PVA	Cu(II), Co(II)	42.5, 24.7	5.9	5	61
PAMAM	Cu(II), Cd(II), Pb(II)	37.4, 28.7, 54.6	4.0	5	62
GO/PSF	Cu(II), Cd(II), Pb(II)	75, 68, 79	6.5–6.7	3	31
GO/PVA	Cu(II), Ni(II), Cd(II)	72.6, 62.3, 83.8	5.7	6	30
PEO/chitosan	Cu(II), Ni(II), Cd(II), Pb(II)	164, 175, 144, 135	5–6	5	63
GO/cellulose	Co(II), Ni(II), Cu(II), Zn(II), Cd(II), Pb(II)	15.5, 14.3, 26.6, 16.7, 26.8, 107.9	4.5	10	This work

---

Acknowledgements

The project was supported by the National Science Centre (Poland) by the Grant No. 2015/17/B/ST4/03870.

References

1. R. Sitko, B. Zawisza and E. Malicka, TrAC, Trends Anal. Chem., 2013, 51, 33–43.
2. G. Zhao, X. Ren, X. Gao, X. Tan, J. Li, C. Chen, Y. Huang and X. Wang, Dalton Trans., 2011, 10945–10952.
3. R. Sitko, E. Turek, B. Zawisza, E. Malicka, E. Talik, J. Heimann, A. Gagor, B. Feist and R. Wrzalik, Dalton Trans., 2013, 42, 5682–5689.
4. G. Zhao, J. Li, X. Ren, C. Chen and X. Wang, Environ. Sci. Technol., 2011, 45, 10454–10462.
5. Y. Sun, Q. Wang, C. Chen, X. Tan and X. Wang, Environ. Sci. Technol., 2012, 46, 6020–6027.
6. G. Zhao, T. Wen, X. Yang, S. Yang, J. Liao, J. Hu, D. Shao and X. Wang, Dalton Trans., 2012, 41, 6182–6188.
7. Z. Li, F. Chen, L. Yuan, Y. Liu, Y. Zhao, Z. Chai and W. Shi, Chem. Eng. J., 2012, 210, 539–546.
8. W. Wu, Y. Yang, H. Zhou, T. Ye, Z. Huang, R. Liu and Y. Kuang, Water, Air, Soil Pollut., 2013, 224, 1372–1379.
9. C. J. Madadrang, H. Y. Kim, G. Gao, N. Wang, J. Zhu, H. Feng, M. Gorring, M. L. Kasner and S. Hou, ACS Appl. Mater. Interfaces, 2012, 4, 1186–1193.
10. R. Sitko, P. Janik, B. Feist, E. Talik and A. Gagor, ACS Appl. Mater. Interfaces, 2014, 6, 20144–20153.
11. R. Sitko, P. Janik, B. Zawisza, E. Talik, E. Margui and I. Queralt, Anal. Chem., 2015, 87, 3535–3542.
12. G. Liu, S. Gui, H. Zhou, F. Zeng, Y. Zhou and H. Ye, Dalton Trans., 2014, 6977–6980.
13. I. E. M. Carpio, J. D. Mangadlao, H. N. Nguyen, R. C. Advincula and D. F. Rodrigues, Carbon, 2014, 77, 289–301.
14. Q. Liu, J. Shi, J. Sun, T. Wang, L. Zeng and G. Jiang, Angew. Chem., Int. Ed., 2011, 50, 5913–5917.
15. R. Sitko, B. Zawisza, E. Talik, P. Janik, G. Osoba, B. Feist and E. Malicka, Anal. Chim. Acta, 2014, 834, 22–29.
16. V. Chandra, J. Park, Y. Chun, J. W. Lee, I.-C. Hwang and K. S. Kim, ACS Nano, 2010, 4, 3979–3986.
17. J. Li, S. Zhang, C. Chen, G. Zhao, X. Yang, J. Li and X. Wang, ACS Appl. Mater. Interfaces, 2012, 4, 4991–5000.
18. M. Liu, C. Chen, J. Hu, X. Wu and X. Wang, J. Phys. Chem. C, 2011, 115, 25234–25240.
19. H. W. Kim, H. W. Yoon, S.-M. Yoon, B. M. Yoo, B. K. Ahn, Y. H. Cho, H. J. Shin, H. Yang, U. Paik, S. Kwon, J.-Y. Choi and H. B. Park, Science, 2013, 342, 91–95.
20. H. Li, Z. Song, X. Zhang, Y. Huang, S. Li, Y. Mao, H. J. Ploehn, Y. Bao and M. Yu, Science, 2013, 342, 95–98.
21. R. K. Joshi, P. Carbone, F. C. Wang, V. G. Kravets, Y. Su, I. V. Grigorieva, H. A. Wu, A. K. Geim and R. R. Nair, Science, 2014, 343, 752–754.
22. B. Mi, Science, 2014, 343, 740–742.
23. H. Zarrin, D. Higgins, Y. Jun, Z. Chen and M. Fowler, J. Phys. Chem. C, 2011, 115, 20774–20781.
24. X. Liu, L. Xie and H. Li, J. Electroanal. Chem., 2012, 682, 158–163.
25. C.-N. Yeh, K. Raidongia, J. Shao, Q.-H. Yang and J. Huang, Nat. Chem., 2015, 7, 166–170.
26. Y. Zhang, S. Zhang and T.-S. Chung, Environ. Sci. Technol., 2015, 49, 10235–10242.
27. Z. Jia and Y. Wang, J. Mater. Chem. A, 2015, 3, 4405–4412.
28. B. P. V. Nellore, R. Kanchanapally, F. Pedraza, S. S. Sinha, A. Pramanik, A. T. Hamme, Z. Arslan, D. Sardar and P. C. Ray, ACS Appl. Mater. Interfaces, 2015, 7, 19210–19218.
29. X. Zhao, Q. Zhang, Y. Hao, Y. Li, Y. Fang and D. Chen, Macromolecules, 2010, 43, 9411–9416.
30. P. Tan, J. Sun, Y. Hu, Z. Fang, Q. Bi, Y. Chen and J. Cheng, J. Hazard. Mater., 2015, 297, 251–260.
31. R. Mukherjee, P. Bhunia and S. De, Chem. Eng. J., 2016, 292, 284–297.
32. D. Han, L. Yan, W. Chen, W. Li and P. R. Bangal, Carbohydr. Polym., 2011, 83, 966–972.
33. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806–4814.
34. S. Stankovich, D. A. Dikin, R. D. Piner, K. A. Kohlhaas, A. Kleinhammes, Y. Jia, Y. Wu, S. T. Nguyen and R. S. Ruoff, Carbon, 2007, 45, 1558–1565.
35. K. Krishnamoorthy, M. Veerapandian, K. Yun and S.-J. Kim, Carbon, 2013, 53, 38–49.
36. S. Abdolhosseinzadeh, H. Asgharzadeh and H. S. Kim, Sci. Rep., 2015, 5, 10160.
37. S. Reich and C. Thomsen, Philos. Trans. R. Soc., A, 2004, 362, 2271–2288.
38. L. G. Cançado, A. Jorio, E. H. M. Ferreira, F. Stavale, C. A. Achete, R. B. Capaz, M. V. O. Moutinho, A. Lombardo, T. S. Kulmala and A. C. Ferrari, Nano Lett., 2011, 11, 3190–3196.
39. K. N. Kudin, B. Ozbas, H. C. Schniepp, R. K. Prud'homme, I. A. Aksay and R. Car, Nano Lett., 2008, 8, 36–41.
40. M. Hirata, T. Gotou, S. Horiuchi, M. Fujiwara and M. Ohba, Carbon, 2004, 42, 2929–2937.
41. D. W. Fuerstenau and K. Osseo-Asare, J. Colloid Interface Sci., 1987, 118, 524–542.
42. Y. S. Ho and G. McKay, Process Biochem., 1999, 34, 451–465.
43. I. Langmuir, J. Am. Chem. Soc., 1916, 38, 2221–2295.
44. I. Langmuir, J. Am. Chem. Soc., 1918, 40, 1361–1403.
45. H. M. F. Freundlich, Z. Phys. Chem., 1906, 57, 385–470.
46. J. A. Dean, Lange's handbook of chemistry, McGraw-Hill, New York, 1999.
47. S. Yang, C. Chen, Y. Chen, J. Li, D. Wang, X. Wang and W. Hu, ChemPlusChem, 2015, 80, 480–484.
48. D. S. Eldridge, R. J. Crawford and I. H. Harding, J. Colloid Interface Sci., 2015, 454, 20–26.
49. J. B. Fein, C. J. Daughney, N. Yee and T. A. Davis, Geochim. Cosmochim. Acta, 1997, 61, 3319–3328.
50. J. B. Fein, A. M. Martin and P. G. Wightman, Geochim. Cosmochim. Acta, 2001, 65, 4267–4273.
51. C. Chen and X. Wang, Ind. Eng. Chem. Res., 2006, 45, 9144–9149.
52. S. Yang, J. Hu, C. Chen, D. Shao and X. Wang, Environ. Sci. Technol., 2011, 45, 3621–3627.
53. W. Stumm, Chemistry of the Solid-Water Interface: Processes at the Mineral-Water and Particle-Water Interface in Natural Systems, John Wiley & Son Inc., 1992.
54. X. Wang, Y. Li, H. Li and C. Yang, Carbohydr. Polym., 2016, 146, 274–281.
55. Y. Bao, X. Yan, W. Du, X. Xie, Z. Pan, J. Zhou and L. Li, Chem. Eng. J., 2015, 281, 460–467.
56. M. Kumar, R. Shevate, R. Hilke and K.-V. Peinemann, Chem. Eng. J., 2016, 301, 306–314.
57. N. Abdullah, R. J. Gohari, N. Yusof, A. F. Ismail, J. Juhana, W. J. Lau and T. Matsuura, Chem. Eng. J., 2016, 289, 28–37.
58. J. He, A. Liu and J. P. Chen, J. Colloid Interface Sci., 2015, 439, 162–169.
59. J. Song, H. Oh, H. Kong and J. Jang, J. Hazard. Mater., 2011, 187, 311–317.
60. G. Zeng, Y. He, Y. Zhan, L. Zhang, Y. Pan, C. Zhang and Z. Yu, J. Hazard. Mater., 2016, 317, 60–72.
61. D. K. Sharma, F. Li and Y.-n. Wu, Colloids Surf., A, 2014, 457, 236–243.
62. Q. Zhang, N. Wang, L. Zhao, T. Xu and Y. Cheng, ACS Appl. Mater. Interfaces, 2013, 5, 1907–1912.
63. M. Aliabadi, M. Irani, J. Ismaeili, H. Piri and M. J. Parnian, Chem. Eng. J., 2013, 220, 237–243.

---