Introduction

Nowadays, environmental conservation is of increasing social and economic concerns worldwide [1]. As an impartible member of the environment, water bodies have been a global issue from both quantity (as a result of the rapid industrialization, unsustainable production patterns, population growth, etc.) and quality (because of heavy and diverse contamination) standpoints [2]. In fact, water contamination due to the discharge of harmful wastes like petroleum products, fertilizers, pesticides, pharmaceuticals, dyes, and heavy metals from various sources into the environment has become an important challenging issue of society today [3]. From the contaminants of particular concern, heavy metals (HMs) have received much special attention because of their toxicological and physiological effects on the environment, human beings, animals, and plants even in small quantities [4].

Heavy metal contamination is mainly originating from many industries (such as metallurgical, mining, chemical, tannery, battery, and nuclear), of which they may directly or indirectly enter the food chain. The most adverse aspect of HMs is that they cannot be readily excreted from the human body following metabolism in the liver and easily accumulate in organs such as the kidneys and brain, thus slowly altering and damaging regular body function [5, 6].

Some toxic HMs of particular anxiety in the treatment of industrial wastewaters are mercury, lead, chromium, and copper. These metals impose very extensive effects on humans. For instance, mercury ions can obliterate the central nervous system, impair pulmonary function, and induce breathlessness and chest disconformity. The mercury ions also cause human chromosome breakage and genetic defects, thus conducing to an abnormality of chromosome distribution [7]. Exposure to lead could have toxicological impacts on the immune, renal, hematopoietic, cardiovascular, reproductive, and pulmonary systems [8]. Furthermore, copper is toxic at elevated concentrations and is persistent and bioaccumulative in the environment [9]. It is, therefore, imperative to eliminate these toxic pollutants from the effluents efficiently and selectively prior to discharging into the environment.

During recent decades, a variety of removal methods including chemical precipitation, membrane filtration, liquid–liquid extraction, ion exchange, biosorption, and adsorption have been developed. Among these methods, adsorption has the advances of convenient operation, cost-effectiveness, high efficacy, and lack of secondary contamination for the elimination of HMs at low concentrations [10]. Over the years, a great number of adsorbents have been developed to remove HM ions from aqueous media, like activated carbon, agricultural and industrial wastes, natural clay minerals, ion-exchange resins, polymers, and so forth [18,19,20,21,22]. Nonetheless, conventional adsorbents often have a relatively weak affinity to HMs, especially mercury ion, resulting in low adsorption capacity [11]. Thus, the development of emerging and effectual adsorbents for their removal declares a great deal of priority in wastewater treatment.

GO is a two-dimensional (2D) carbon material that represents exceptional properties like excellent adsorption behavior. However, pure GO is found to be non-magnetic with limited adsorbing abilities. One of the effective ways to solve the problem is to couple GO with other nanoparticles like metal, metal oxide, and sulfides [12]. Over the past few years, magnetic ferrites of the type MFe2O4 (M = Fe, Co, Ni, Cu, Mn, etc.) with cubic spinel structure have been anchored on functionalized GO to enhance its activity in degradation of organic pollutants, adsorption of various contaminants, high-density information storage, solar cells, magnetic drug delivery, and anti-microbial applications [13,14,15]. Among ferrites, NiFe2O4 exhibits noticeable characteristics of moderate saturation magnetization, favorable visible-light response, excellent chemical, and photochemical stability as well as mechanical hardness [14, 16, 17]. The adsorbability of NiFe2O4/GO can be further increased by functionalizing it by an eco-friendly and biologically active tripeptide, namely glutathione. Glutathione (GSH) is an important antioxidant in almost all living beings including plants, animals, fungi, and even some bacteria and archaea [18,19,20]. Since GSH consists of L-glutamate, l-cysteine, and glycine amino acids, it could capture the HM ions via its –SH, two free –COOH, –NH2, and two –CONH– groups. Therefore, it can be employed to improve the adsorption performance of GO [20]. According to our literature survey, until now, little work has been done on the preparation of NiFe2O4/GO functionalized with GSH (GSH-NiFe2O4/GO) composite, and their applications for removal of heavy metal ions from water environments.

This work, therefore, aimed to develop a novel magnetic GO nanocomposite functionalized with GSH as a natural, cheap, harmless, and eco-friendly biologic reagent for the removal of HMs of Hg(II), Cu(II), and, Pb(II) from aqueous solutions. The as-prepared nanocomposite was characterized by different techniques such as field emission-scanning electron microscopy (FE-SEM), energy-dispersive X-ray spectroscopy (EDS), Fourier transform-infrared spectroscopy (FT–IR), X-ray diffraction (XRD) patterns, vibrating sample magnetometry (VSM), and Brunner–Emmett–Teller (BET) surface area analysis. The adsorption process was performed by considering the main effective parameters of pH, time, and initial concentration. The adsorption kinetics, isotherms, thermodynamics, and recycling studies were also conducted to facilitate scaling-up the results. The spiked groundwater samples were applied to study the efficacy of the prepared adsorbent for the exclusion of Pb(II), Hg(II), and, Cu(II) ions and to evaluate the influences of coexisting ions on the removal process.

Materials and methods

Chemicals

All chemicals and reagents used in this study were of analytical grade or higher, including Hg(NO3)2·H2O, Cu(NO3)2·3H2O, Pb(NO3)2, NaNO3, graphite powder < 20 μm, KMnO4, NaNO3, H2O2 [(30% (w/w) in H2O)], H2SO4 (98%), FeCl3·6H2O, NiCl2·6H2O, glutathione, and NH3 (32%) solution. The solutions were prepared with deionized water (DW).

Preparation of GSH-NiFe2O4/GO nanocomposite

GO was successfully prepared using the modified method of Hummer [21]. Briefly, 2.5 g NaNO3 was dissolved into 115-mL concentrated H2SO4 in an ice bath beaker, stirred for 15 min. The mixture was exposed to 2.5 g graphite powder and stirred for 20 min. Then, 15 g KMnO4 was added to the suspension and stirred for 70 min. The beaker was then muffled and kept into an ultrasonic bath at 35–40 °C and stirred further for 15 h. Afterward, 250-mL DW and 25-mL H2O2 were gradually added into the mixture at 95 °C and stirred for 15 min, respectively. The brown solution was centrifuged for 10 min at 12,000 rpm and rinsed alternatively with DW and mixture of HCl (5%, v/v) and water until pH reached to about 7. The product was freeze-dried for 48 h for future consumption.

The GO was magnified by the chemical precipitation of Ni(II) and Fe(III) on its surface. For this, 1 g of as-prepared GO was dispersed ultrasonically in 250-mL DW for 60 min. Then, 6.5 g FeCl3·6H2O and 2.5 g NiCl2·6H2O were added into the suspension, stirred in the ultrasonic for 30 min at 80 °C. The mixture was transferred into a round-bottom flask, exposed to the heater in a paraffin bath at 80 °C, and stirred magnetically. 10-mL NH3 (32%) was added drop-wise into the solution for 15 min. While the system was full of N2 gas and the decanter was closed, the solution was refluxed at 80 °C. After stirring vigorously overnight, NiFe2O4/GO was separated magnetically, washed with DW several times until the filtrate pH reached to neutral, and dried at 60 °C in a vacuum oven for 24 h.

For anchoring of GSH on the surface of the magnetic GO nanocomposite, 0.5 g GSH was magnetically dissolved into 100-mL DW. 50 mL of the solution was then mixed with 0.5 g NiFe2O4/GO nanocomposite, covered with an aluminum foil, and was shaken for 2 h. GSH-NiFe2O4/GO composite was then separated using a magnetic field, washed several times with ethanol and DW water and thoroughly dried under vacuum for further consumption.

Instrumental techniques

The morphology and particle size of GSH-NiFe2O4/GO composite along with the quantification of the contributed elements were examined by field emission scanning electron microscopy (FE–SEM, Sigma-Aldrich, Zeiss) on gold-coated equipped with energy-dispersive X-Ray analysis spectroscopy (EDX, Sigma-Aldrich, Zeiss). Identification of (in)organic, polymeric links, and functional groups was conducted using Fourier infrared transformation spectroscopy (FT–IR, Nicolet Thermo Nexus 870) in the range of wavenumbers 400–4000 cm−1. The purity and crystal structure of the nanocomposite was examined by an X-ray diffractometer (XRD) (Xʼ pert PRO, Panalytica) with Cu Kα source (λ = 1.54056 Å) at 100 mA and 40 kV. The magnetic features of the composite were implemented by a vibrating magnetometer/alternating gradient force magnetometer (VSM, Kavir magnetic Co., Iran). Brunner–Emmett–Teller (BET) surface area was determined by TriStar II Plus (Micromeritics) instrument. The pH of the samples was adjusted by using either NaOH or HCl 0.1 mol L−1 solutions and determined by a pH meter (PerkinElmer, USA). The HMs concentrations were measured by mercury analyzer (MA, Milestone, DMA 80, Italy) and the flame atomic absorption instrument (nov AA 400p, Germany) with flow rate of 5 mL min−1 at the wavelength of 217 nm and 324.7 nm for lead and copper, respectively.

Batch experiments

Batch adsorption experiments were conducted in 100-mL beakers containing 25 mL of a single ion solution with a certain initial concentration and 10 mg of GSH-NiFe2O4/GO which were shaken on a shaker under room temperature at 250 rpm for specific moments. Then, the solid–liquid phases were separated by a magnet, followed by filtration. An atomic absorption spectrophotometer was used for determining the concentration of metal ions in the solution. The impacts of various parameters, such as solution pH (3–7), contact time (5–120 min), and initial ion concentration (5–120 mg L−1), were studied on the adsorption of metal ions. Each set of experiments was done in triplicates and the average values reported with the deviation within ± 5%.

The adsorption capacity at time \(t\;(q_{t} ,{\text{mg g}}^{ - 1} )\) and the removal efficiency (%) of metal ion were calculated according to Eqs. (1) and (2), respectively.

$$q_{t} = \frac{{(C_{0} - C_{t} ) V }}{m}$$
(1)
$${\text{Removal}}\;{\text{efficiency}}\; ( {{\% )}} = \frac{{(C_{0} - C_{t} ) }}{{C_{0} }} \times 100$$
(2)

where C0 and \(C_{t} \; ( {\text{mg L}}^{ - 1} )\) are the concentration of the target metal ion in solution before and after adsorption, respectively. V(L) is the volume of the metal ion solution and m(g) is the mass of GSH-NiFe2O4/GO applied.

To investigate the effect of coexisting ions on the adsorption process, different types of cations comprising Fe(II), Mn(II), Zn(II), and Cd(II) were chosen and investigated at the same concentration of target metal ion (20 mg L−1) in binary solutions and contact time of 60 min with adsorbent. Then, the supernatants were separated for analysis of the residual metal ions.

The propriety of GSH-NiFe2O4/GO for practical environmental remediation was evaluated by collecting groundwater samples from Kukeneh village in Blukat Rural District, Rahmatabad and Blukat District, Rudbar County, Gilan Province, Iran. The water samples were used without pretreatment prior to the adsorption, except the filtration through a cellulose membrane filter with the pore size of 0.45 μm. Experiments were done by spiking groundwater samples at 5 mg L−1 concentration of the tested metal ion with adjusting the desired pH. 0.1 g of GSH-NiFe2O4/GO was dispersed to 10 mL of spiked sample and then shaken for 30 min. After the adsorption process, removal efficiency (%) of metal ion was measured.

The regeneration experiments were accomplished on GSH-NiFe2O4/GO using six adsorption–desorption cycles. For each adsorption cycle, 30 mg GSH-NiFe2O4/GO was added to 25 mL of metal ion with an initial concentration of 10 mg L−1 and shaken for 15 min. For desorption cycle, metal-loaded GSH-NiFe2O4/GO was collected after filtration, dispersed into 10 mL of eluent HNO3 (0.1 mol L−1), and shaken for 20 min at a speed of 250 rpm to be desorbed.

Results and discussion

Material characterization

The FE–SEM micrograph of (a) GO, (b) NiFe2O4/GO, and (c) GSH-NiFe2O4/GO is illustrated in the left panel of Fig. 1. It is clear from Fig. 1a that GO had layered, thin, slightly crinkled, and relatively wrinkled structures. Compared with GO, NiFe2O4/GO material demonstrated a much coarser surface, which disclosed that a number of nickel ferrite particles with the average size of 40–90 nm anchored to the surface of GO nanosheets (Fig. 1b). After modification with GSH, no conspicuous morphological change was observed (Fig. 1c).

Fig. 1
figure 1

High magnification FE-SEM images (panels on the left-hand side) and EDX spectra (panels on the right-hand side) with the insets of elemental analysis of GO (a, d), NiFe2O4/GO (b, e), and GSH-NiFe2O4/GO (c, f)

Full size image

The elemental distribution spectrum of samples is also presented in the right panel of Fig. 1. As shown in Fig. 1d, e, the decorating GO with NiFe2O4 particles led to the appearance of nickel and iron characteristic peaks along with other elements, such as carbon and oxygen. The functionalization with GSH is associated with the presence of sulfur and nitrogen peaks, which can be attributed to thiol, amide, and amine groups of GSH (Fig. 1f).

The FT-IR spectra of the synthesized GO, NiFe2O4/GO, and GSH-NiFe2O4/GO are shown in Fig. 2a–c, respectively. The adsorption peaks at 3443, 1724, 1634, 1406, 1260, and 1036 cm−1 in the spectrum of GO show the presence of O–H, C=O in carboxyl groups, aromatic C=C, C–OH, epoxy C–O, and alkoxy C–O vibrations [22, 23]. The spectrum of NiFe2O4/GO displays two strong peaks at 445 and 477 cm−1 corresponding to the stretching vibrations of metal–oxygen at tetrahedral and octahedral sites, respectively, confirming that NiFe2O4 was successfully anchored onto GO nanosheets [11, 24]. After functionalization with GSH, five new characteristic peaks appeared (Fig. 2c): the 2518 cm−1 peak associated with S–H stretching, the 1331 cm−1 peak attributed to C=N stretching, the 1060 cm−1 corresponded to the N–H bending, the 922 cm−1 assigned to the S–H bending, and N–H wagging was at 826 cm−1 [23, 25]. What’s more, for better comparing the spectrum of GSH-NiFe2O4/GO was compared with the spectrum of pristine glutathione (Fig. 2d). The appearance of glutathione adsorption peaks in the spectrum of GSH-NiFe2O4/GO proved the effectiveness of the functionalization process.

Fig. 2
figure 2

FT-IR spectra of (a) GO, (b) NiFe2O4/GO, (c) GSH-NiFe2O4/GO, and (d) GSH

Full size image

In order to explore the crystalline structure of as-prepared materials, the XRD technique was performed. Figure 3a, b exhibits typical XRD patterns of GO and GSH-NiFe2O4/GO, respectively. In Fig. 3a, an intense diffraction peak at 2θ value of about 12.3° with 7.1 nm d-spacing is observed, which is corresponding to the (001) crystalline plane of GO. In Fig. 3b, seven characteristic peaks of NiFe2O4 at 2θ values of 18.3°, 30.1°, 35.6°, 45.9° 54.3°, 57.1°, and 62.6° which can be attributed to the (111), (220), (311), (400), (422), (511), and (440) reflections of spinel-type NiFe2O4 (JCPDS 54-0964), respectively, indicating the existence of NiFe2O4 particles in the as-prepared GSH-NiFe2O4/GO composite [16, 17]. Moreover, the diffraction peaks of GO (002) are still present in the XRD pattern of GSH-NiFe2O4/GO, illustrating that the main structure of GO remained quite invariable when it was coated with NiFe2O4 and functionalized by GSH. The average crystallite size of NiFe2O4 particles was approximately calculated as 39.3 nm, according to the Debye–Scherrer equation [26].

Fig. 3
figure 3

XRD patterns of a GO and b GSH-NiFe2O4/GO

Full size image

Magnetization measurements were performed at room temperature for evaluating the magnetic property of GSH-NiFe2O4/GO nanocomposite. The magnetization curve shown in Fig. 4a implies its super paramagnetic behavior. The saturation magnetization of 8.9 emu g−1 is observed at ± 10 kOe. This value is far less than that reported for bare NiFe2O4, 36.1 emu g−1 [14]. This difference could be attributed to the less mass fraction of magnetic component in the synthesized composite. The electron exchange between the NiFe2O4 and GO may also cause quenching of the magnetic momentum in GSH-NiFe2O4/GO [27]. However, the magnetic property of GSH-NiFe2O4/GO was still sufficient to meet the requirements for magnetic of separation.

Fig. 4
figure 4

a Magnetic hysteresis loops of GSH-NiFe2O4/GO (the inset shows the photograph of GSH-NiFe2O4/GO attracted to the magnet) and b N2 adsorption–desorption isotherm of GSH-NiFe2O4/GO and the relevant pore size distribution (inset)

Full size image

To investigate the surface area and the pore structure of GSH-NiFe2O4/GO composite, an N2 gas adsorption–desorption isotherm and BJH (Bareet–Joyner–Halenda) pore size distribution curve were performed, as shown in Fig. 4b. The isotherm curve represents the type IV N2 adsorption–desorption as well as H3 hysteresis loop. The BET surface area and desorption pore volume of GSH-NiFe2O4/GO were 75.8110 m2 g−1 and 0.0645 cm3 g−1, 3.8062 nm, respectively, which was far higher than those of GO of 3.5386 m2 g−1 and 0.0171 cm3 g−1. The mean pore width for composite was 3.8062 nm, calculated by BJH analysis.

Adsorption exploration

Effect of solution pH value

To evaluate the effect of pH, the adsorption of HM ions (with the initial concentration of 10 mg L−1) was explored at the pH range of 3.0–7.0, and the results are shown in Fig. 5a. In case of Pb(II), the pH values ≥ 6 have not been studied because of the precipitation of Pb(II) ion, forming hydroxide in the solution. As shown in Fig. 5a, the pH of the aqueous media plays a vital role in the adsorption of HM ions by using GSH-NiFe2O4/GO. The removal efficiency of all four HM ions is ≥ 94%, indicating that GSH-NiFe2O4/GO is highly efficient to eliminate HMs from aqueous media. Figure 5a also shows a distinct increase tendency in removal efficiencies of Pb(II), Hg(II), and Cu(II) with the gradual increase in pH to 5, 6, and 6. The high values of pH (> 6) lead to a reduction in the removal of Hg(II) and Cu(II) ions.

Fig. 5
figure 5

a The effect of solution pH value on the removal efficiency of metal ions using GSH-NiFe2O4/GO and b point of zero charge (pHpzc) of GSH-NiFe2O4/GO and (conditions: m = 10 mg, V = 25 mL, C0 = 10 mg L−1, and t = 120 min)

Full size image

The observed dependence of the process on pH may be attributed to change in the solubility of metal ions as well as the adsorbent surface charge with pH. The functionalized GSH contains thiol (–SH), carboxyl (–COOH), amino (–NH2), and amide (–CONH–) groups that may affect the surface charge of GSH-NiFe2O4/GO. The pH of zero-point charge (pHpzc) for GSH-NiFe2O4/GO is found to be about 4.8 (Fig. 5b). At pH < pHpzc, the adsorbent surface is positively charged, both in the case of the protonation of active groups present on the coated GSH and magnetic Fe3O4 particles, which could lead to little removal efficiency owing to the electrostatic repulsive forces between cations [Pb(II), Hg(II), or Cu(II)] and the protonated adsorbent. Additionally, in acidic solution, the excessive H+ ions may compete seriously with the positive-charged metal ions for the adsorption sites, which diminish the removal ability. On the other hand, the enhancement of cation removal with the increase in pH above pHpzc can be explained based on the deprotonation of adsorbent (resulting in the negatively charged of the prepared adsorbent surface) and the reduction of H+ concentrations in the aqueous medium. Hence, the strong electrostatic attraction between adsorbent and cations along with less competitive adsorption of H+ makes the removal quantity of cations increase remarkably [20]. However, with the increment in pH beyond 6, the removal efficiencies of Cu(II) and Hg(II) started to decrease. It seems that at the alkaline pH values, hydroxide ions (OH−) tend to combine with the aforementioned ions in competition with the functional groups on the nanocomposite, leading to the formation of metal hydroxide species [as M(OH)+ and M(OH)2]. Consequently, no adsorption experiment implemented at solution pH > 7.

Effect of contact time and adsorption kinetics

The relationship between time and adsorption capacity can be investigated by the kinetic models, which in turn, can be used to calculate and delineate the adsorption rate. For this reason, the equilibrium time is one of the most effective factors on adsorption capacity [28]. Figure 6 displays the effect of contact time at the range of 5–120 min on the removal efficiency of HM ions with the initial concentration of 10 mg L−1. As shown in Fig. 6, rapid adsorption of HMs by GSH-NiFe2O4/GO occurs up to 60 min of reaction time, and adsorption slows down thereafter, lastly reaches equilibrium in approximately 90 min which 99%, 97%, and 94% of Hg(II), Cu(II), and Pb(II) ions were removed, respectively. Such efficient and rapid performance suggests that the applied nanocomposite can be considered as a potential material in the removal of the studied HMs, which in turn, is attributed to its abundant unoccupied active adsorption sites accessible on the surface of GSH-NiFe2O4/GO, and the absence of internal diffusion resistance with the first 90 min of contact time. Beyond this time, insignificant changes occurred in the amount of ion removal owing to the saturation of active binding sites. Hence, 90 min was selected as the equilibrium time for the study of further adsorption properties. The experimental equilibrium adsorption capacity (\(q_{{e , {\text{exp}}}} ,\;{\text{mg g}}^{ - 1}\)) of GSH-NiFe2O4/GO was found to be 24.76, 24.25, and 23.60 for Hg(II), Cu(II), and Pb(II), respectively.

Fig. 6
figure 6

Effect of contact time on the removal efficiency of metal ions using GSH-NiFe2O4/GO (conditions: m = 10 mg, pH 6 for both Hg(II) and Cu(II), pH 5 for Pb(II), V = 25 mL, and C0 = 10 mg L−1)

Full size image

To investigate the adsorption mechanism and potential the rate-limiting steps, the adsorption kinetic data of Hg(II), Cu(II), and Pb(II) were evaluated by testing nonlinear (pseudo-first-order, pseudo-second-order, and simplified Elovich) and linear (intraparticle diffusion) kinetic models at various time intervals (5–120 min). Linear and nonlinear regression analyses were carried out using Excel (2013) and Mathematica (9) software, respectively.

The pseudo-first-order (PFO), pseudo-second-order (PSO), and Elovich kinetic models can be nonlinearly represented by Eqs. (3), (4), and (5), respectively.

$$q_{t} = q_{e} (1 - {\text{e}}^{{ - k_{1} t}} )$$
(3)
$$q_{t} = \frac{{k_{2} q_{e}^{2} t}}{{1 + k_{2} q_{e} t}}$$
(4)
$$q_{t} = \frac{1}{\beta }{\text{ ln(}}\alpha \beta t )$$
(5)

where qe (mg g−1) and qt (mg g−1) are the quantity of HM ion adsorbed per unit mass of the adsorbent at equilibrium and time t (min), respectively; k1 (min−1) and k2 (g mg−1 min−1) are the adsorption rate constant of PFO and PSO kinetic models, respectively; \(\alpha \;({\text{mg g}}^{ - 1} { \hbox{min} }^{ - 1} )\) is the initial adsorption rate constant and \(\beta \;({\text{g mg}}^{ - 1} )\) is the desorption constant. The initial adsorption rate, h (mg g −1 min−1) at t → 0, is determined from k2 and qe as Eq. (5).

$$h = k_{2} q_{e}^{2}$$
(6)

The nonlinear fitting results of the three above-mentioned kinetic models are shown in Fig. S1a–c for Hg(II), Cu(II), and Pb(II), respectively. All relevant kinetic parameters as well as correlation coefficient values (R2) are also given in Table 1. On account of the best goodness-of-fit (R2 > 0.999), it seems that compared to PFO kinetic model, the PSO and Elovich models can describe better the adsorption of selected three HM ions studied by GSH-NiFe2O4/GO. In addition, fitted equilibrium adsorption capacities obtained from PSO model (qe,cal) agree more closely with the experimental values (qe,exp) given in the last row of Table 1. Kinetic studies implied that the adsorption may be occurred by the chemical process involving valence forces through sharing or exchange of electrons between GSH-NiFe2O4/GO adsorbent and metal ions [20].

Table 1 Kinetic fitting parameters of PFO, PSO, and simplified Elovich models for the adsorption of metal ion onto GSH-NiFe2O4/GO
Full size table

The calculated values of k2 and h from PSO kinetic model were much greater for Hg(II) than those obtained for Pb(II) or Cu(II), suggesting that the uptake of Hg(II) was more rapid than uptake of Pb(II) or Cu(II). Consequently, the binding sites on GSH-NiFe2O4/GO had a higher affinity for Hg(II) ions.

The linear mathematical equation [Eq. (7)] of intra-particle (Weber and Morris) diffusion model was utilized in order to study the effect of mass transfer resistance on the binding site of pollutants to the surface of the adsorbent.

$$q_{t} = k_{i} t^{1/2} + C_{i}$$
(7)

where ki (g mg−1 min−1/2) is the intra-particle diffusion rate and Ci (mg g −1) is the boundary-layer thickness at stage i.

The linear plot of qt versus \(t^{1/2}\) for adsorption of Hg(II), Cu(II), and Pb(II) is exhibited in Fig. S2a–c, in the respective order, and the values of ki, Ci, and R2 are also listed in Table 2. The plots in Fig. S2 indicate three different stages, implying that the uptake of HM ions onto GSH-NiFe2O4/GO takes place via three stages so that the magnitude of the rate constants are as follows: k1 > k2 > k3 (Table 2). The first linear stage indicates the external diffusion, wherein metal ions transport from the solution to the external surface of adsorbent and interacts facilely with its active sites. The second linear stage denotes the intra-particle diffusion, wherein the adsorbed metal ion penetrates into internal pores of adsorbent. The third stage refers to the gradual saturation adsorption process and final equilibrium step [29, 30]. The substantial resistance to mass transfer takes place in the second stage. The resultant plots in Fig. S2 also exhibits notable deviations in linearity from the origin and with nonzero intercepts, which signify that adsorption of metal ions onto GSH-NiFe2O4/GO involves some chemisorption and is not entirely governed by intra-particle diffusion.

Table 2 Kinetic fitting parameters of intra-particle diffusion model for the adsorption of metal ion onto GSH-NiFe2O4/GO
Full size table

Effect of initial HM ion concentration and adsorption isotherms

To investigate the influence of the initial metal ion concentration, a series of eight experiments for each ion studied, with different concentrations (5–120 mg L−1) were carried out using 10 mg of GSH-NiFe2O4/GO in 25 mL solution, given pH for each ion, and 90 min of contact time. The obtained data are shown in Fig. 7. It can be seen that an increase in the initial concentration of metal ion from 5 to 120 mg L−1 leads to a decrease in the removal efficiency from 100%, 99%, and 97% to 77%, 71%, and 66% for Hg(II), Cu(II), and Pb(II), respectively. This can be due to less probability of getting the free active sites of GSH-NiFe2O4/GO by the metal ions and, therefore, the saturation of the adsorbent at higher concentrations, which declines the amount of metal uptake.

Fig. 7
figure 7

Effect of the initial metal ion concentration on the removal efficiency of metal ions using GSH-NiFe2O4/GO (conditions: m = 10 mg, pH 6 for both Hg(II) and Cu(II), pH 5 for Pb(II), V = 25 mL, and t = 90 min)

Full size image

The interaction between adsorbents and adsorbates at equilibrium can be found through the investigation of isotherm models. In this work, Langmuir [Eq. (8)], Freundlich [Eq. (9)], Dubinin–Radushkevich (D–R) [Eq. (10)], and Temkin [Eq. (11)] equilibrium isotherms were applied to determine the adsorption mechanism of HMs by GSH-NiFe2O4/GO. The models can be nonlinearly expressed by the following equations.

$$q_{e} = \frac{{q_{m} b C_{e} }}{{1 + b C_{e} }}$$
(8)
$$q_{e} = K_{F} C_{e}^{1/n}$$
(9)
$$q_{e} = q_{t} e^{{ - \beta \varepsilon^{2} }}$$
(10)
$$q_{e} = B_{T} {\text{ ln(}}A_{T} C_{e} )$$
(11)

where \(q_{m} \; ( {\text{mg g}}^{ - 1} )\) and \(b\;({\text{L mg}}^{ - 1} )\) are the maximum adsorption capacity and Langmuir equilibrium constant, respectively; \(K_{F} \left( {{\text{mg}}^{{1 - \frac{1}{n}}} {\text{L}}^{{\frac{1}{n}}} {\text{g}}^{ - 1} } \right)\) and n are Freundlich isotherm constant and Freundlich dimensionless exponent related to the adsorption intensity, respectively;\(\beta \;({\text{mol}}^{2} {\text{J}}^{ - 2} )\) and \(q_{t} \; ( {\text{mg g}}^{ - 1} )\) are D–R isotherm constant, and the theoretical adsorption capacity, respectively; ε is Polanyi potential connected with the equilibrium concentration of adsorbate as \(\varepsilon = RT { \ln }\left[ {1 + 1/C_{e} } \right]\) where R (J mol−1 K−1) and T (K), respectively, are the gas constant and the temperature; \(A_{T} \; ( {\text{L mg}}^{ - 1} )\) and \(B_{T}\) are Temkin constant connected with equilibrium binding energy and a constant related to the heat of adsorption, respectively; the value of \(\beta\) is used to compute the mean free energy of adsorption \(E\;({\text{kJ mol}}^{{{-}1}} )\), as E = 1/(2β)1/2.

The nonlinear fitting results of the isotherm models are shown in Fig. 8a–c for Hg(II), Cu(II), and Pb(II), in the respective order. All pertinent isotherm parameters together with correlation coefficient values (R2) are also listed in Table 3. The values of R2 of Freundlich isotherm are higher than 0.999 for tested three metal ions. This clearly indicates that the adsorption equilibrium data are better fitted to Freundlich isotherm than Langmuir, D–R, or Temkin isotherms. Hence, the adsorption process of metal ions is governed by multilayer formation and the binding sites on the surface of GSH-NiFe2O4/GO reveal the heterogeneous nature. The constant values of 1/n are all less than one. This outcome confirms that selected metal ions interact effectively with the adsorbent; accordingly, the adsorption occurs beneficially under applied conditions [31].

Fig. 8
figure 8

Nonlinear isotherm models of Langmuir, Freundlich, Dubinin–Radushkevich, and Temkin for a Hg(II), b Cu(II), and c Pb(II) adsorption onto GSH-NiFe2O4/GO. The green circles are the experimental data. (Color figure online)

Full size image
Table 3 Isothermal fitting parameters for the adsorption of metal ion onto GSH-NiFe2O4/GO
Full size table

The value of the mean free energy of adsorption, E, lies between 8.0 and 16.0 kJ mol−1, implying that the adsorption may be dominated by the chemical adsorption process [32].

Based on the Langmuir equation, the maximum capacity (qm) was found to be 272.94, 266.22, and 264.16 for Hg(II), Cu(II), and Pb(II) ions, respectively (Table 3). It is noted that Hg(II) ion adsorption onto GSH-NiFe2O4/GO is greater than the other two metal ions, which could be due to the thiol functional group in addition to the amide, amine, and carboxyl groups contained in glutathione which is more favorable to bind mercury ions.

The maximum adsorption capacities of GSH-NiFe2O4/GO are higher than or comparable to the adsorption capacity of earlier adsorbents, as shown in Table 4. Therefore, it can be concluded that GSH-NiFe2O4/GO adsorbent has a superior potential for the removal of HM ions from aqueous media.

Table 4 Comparison of the adsorption performance of GSH-NiFe2O4/GO with various adsorbents for Hg(II), Cu(II), and Pb(II) ions
Full size table

Effect of interfering ions

Complex waters usually contain a variety of ions, which may compete with the target ions for binding sites of the applied adsorbent. For this reason, investigation of such coexisting ions’ effect is of great practical importance, especially for adsorption-based studies, and it fundamentally determines the scaling-up of the results [33]. Therefore, the adsorption of HMs by GSH-NiFe2O4/GO was investigated in the presence of some coexisting ions including Mn(II), Zn(II), Fe(II), and Cd(II), all at the concentration of 20 mg L−1. The experimental results, shown in Fig. 9, indicated that the studied ions effect negligibly on Hg(II) removal, whereas Pb(II) adsorption is obviously decreased in the presence of Mn(II) and Hg(II) ions. Moreover, Cu(II) removal is affected considerably by the presence of Mn(II), Cd(II), and Hg(II) ions. The coexisting Fe(II) and Zn(II) ions show a minor effect on Pb(II) and Cu(II) uptake. The competition of HM ions for adsorption related probably with the same valence of heavy metals.

Fig. 9
figure 9

Effect of the competitive cations on Pb(II), Hg(II), and Cu(II) removal (conditions: m = 10 mg, pH 5 for Pb(II), pH 6 for both Hg(II) and Cu(II), V = 25 mL, and C0 = 20 mg L−1, and t = 60 min)

Full size image

Removal of metal ions from the real water sample

Practical use of an adsorbent to real samples is the ultimate objective of adsorption investigations for wastewater treatment. Therefore, the capability of GSH-NiFe2O4/GO for the removal of selected metal ion from environmental groundwater samples was tested. Since the concentration of metal ions was found to be below the detection limit, the water sample was spiked with 5 mg L−1 of a single metal ion and then treated with 0.1 g of GSH-NiFe2O4/GO. The tests were done in triplicates under the optimum conditions and the analytical results were reported in Table 5. The removal efficiencies of 99.5% for Hg(II), 98.9% for Cu(II), and 97.5% for Pb(II) ions were achieved, implying that the adsorbent has a superior performance in removing all three metal ions from groundwater samples without important matrix impact.

Table 5 Determination of metal ions in the groundwater samples
Full size table

Reusability of GSH-NiFe2O4/GO

In order to assess the reusability of GSH-NiFe2O4/GO, it was subjected to several loadings with the target HM ions and subsequent washing with 0.1 HNO3 mol L−1 solution under stirring to regenerate the adsorbent. The results are presented in Fig. 10. It was observed that the removal efficiency remained in the range of 77.5–81% after six consecutive cycles of adsorption–desorption, which shows that GSH-NiFe2O4/GO could be applied as a reusable adsorbent for adsorption of studied ions in real-scales applications.

Fig. 10
figure 10

Reusability study of GSH-NiFe2O4/GO

Full size image

Conclusions

In this study, a novel magnetic composite material, GSH-NiFe2O4/GO, was successfully prepared via a quite convenient method. It represented a superior adsorption performance in capturing three heavy metal ions: Hg(II), Cu(II), and Pb(II). The ion removal efficiency was dependent on solution pH, operating time, and initial concentration of metal ions. The optimal pH value was 5 and 6 for Pb(II) and both Hg(II) and Cu(II), respectively, with more than 94% of removal efficiency in individual aqueous solutions. The adsorption attained equilibrium within approximately 90 min. The experimental data nicely followed together by both PSO and Elovich kinetic models as well as Freundlich isotherm model. The calculated maximum adsorption capacity of GSH-NiFe2O4/GO was derived to 272.94, 266.22, and 264.16 mg g−1 for Hg(II), Cu(II), and Pb(II), respectively, which was higher than or comparable to the adsorption capacity of many reported adsorbents in the literatures. In a binary solution, the removal efficiency of Hg(II) was still considerable. However, the removal percentage of Pb(II) and Cu(II) was relatively declined in the presence of Hg(II), Mn(II), and Cd(II) by competition. GSH-NiFe2O4/GO has also confirmed to be applicable in removing metal ions from groundwater samples. It was believed that the abundant functional groups consisting of carboxylic acid, amine, amide, and thiol groups mainly participated in the chemical adsorption of process owing to surface complexation in the binding sites. These findings could promote the development of innovative materials as the adsorbent in water pollution remediation.