Effect of modified graphene oxide on Cu and phosphorus in eutrophic river sediments

With the acceleration of industrialization, a large number of heavy metals are discharged into lakes through atmospheric deposition, surface runoff, and waste water, and most heavy metals enter the sediments. If the water environment changes, heavy metals in the sediments would be released again, which would be a serious threat to the ecological environment (Yan et al. 2016). According to another survey, the sediments of lakes and rivers in China have been seriously polluted by heavy metals (Jiang et al. 2018). Therefore, it is an urgent problem to find stable and efficient adsorbents to immobilize heavy metals in the sediments.

In recent years, graphene oxide (GO) has exhibited good water dispersibility and biocompatibility due to its effective oxygen-containing functional groups, such as hydroxyl, carboxyl and epoxy groups on the surface. These properties could be effectively used for the adsorption of heavy metal ions and dyes (Ahmad et al. 2019). However, the small particle size and easy aggregation features of GO caused an unsatisfactory immobilizing effect (Ramesha et al. 2011). Therefore, researchers prefer to immobilize heavy metals by modifying GO. At present, the modification processes can be divided into four methods, which are physical, chemical, magnetic modification and impregnation (Zhao et al. 2019). Among them, chemical modification and magnetic coating are the most widely used methods. Magnetic coating could enhance the specific surface area of GO, and increase the efficiency of modified materials to immobilize heavy metals (Shraban et al. 2020). The composite of Fe₃O₄ and graphene oxide shows excellent adsorption capacity for Pb (Hu et al. 2017). GO with Fe₃O₄ loaded on the surface can improve the electron transfer between metal oxides and organic materials, and adsorb heavy metals more efficiently than GO (El-Shafai et al. 2020). Chemical modification can positively change the number and type of functional groups on the surface of GO. Amine functionalization introduces ammonia ions into GO surface, which can enhance GO's electronegativity and allow it to adsorb heavy metal ions more efficiently (Wang et al. 2018). In addition, chitosan-modified GO has higher adsorption capability for heavy metals, a wider pH range, and stronger adaptability to the environment than GO (Cao et al. 2019). From this point of view, both magnetic and chemical modification can improve the efficiency of GO to immobilize heavy metals. Researchers have paid more attention to the treatment of heavy metals and organic pollutants in water (Liu et al. 2019b). However, due to the lack of research on the stable solidification of heavy metals in the sediments, researchers have been insufficiently aware of the effects of modified materials on immobilizing heavy metals in the sediments.

Generally, released excessive phosphorus into water body would cause eutrophication and reduce the concentration of dissolved oxygen (DO) in water (Jin et al. 2019). As a typical grassland lake in the cold and arid regions, Ulansuhai nur has arid regional climatic conditions, where rainwater dilution and the purification capability of the environment cannot offset the damage of external pollutants to the ecosystem. Heavy metals contamination and eutrophication in the sediments have increased in Ulansuhai nur year by year. Nitrogen and phosphorus in overlying water affect the immobility of heavy metals in the sediments to some extent (Rinklebe et al. 2016b). Phosphates can immobilize heavy metals in the sediments. However, the effect of immobilization depends on the available phosphorus content in the soil or sediments, as soluble phosphorus performed considerably better than insoluble phosphorus in the process of restoration (Su et al. 2015). Nutrients (such as N, P, and S) changed the distribution of microbial communities in the sediments of the Bohai Sea in China, which in turn adversely influenced the immobility of heavy metals in the sediments (Lu et al. 2019). A study has also shown that heavy metals (such as As, Ba, Cd, Cu, Pb, and Sr) in the sediments may be released under aerobic conditions (Rinklebe et al. 2016b), which is inconsistent with previous research. When the pH and oxidation–reduction potential (ORP) change in the sediments, heavy metals may reenter the overlying water and cause secondary pollution (Rinklebe et al. 2016a). Therefore, it is of great significance to study the influence of modified materials on the immobilization of heavy metals by simulating the changes of water environment.

In this paper, by changing the type and number of oxygen-containing functional groups on the surface of GO and optimizing its specific surface area, two modified materials were prepared to immobilize Cu in the Ulansuhai nur sediments and reduce the ecological risk of Cu. In the present study, we compared the effect of G-F with that of G-N-C on immobilizing Cu in the sediments; we designed the experiments based on the actual situation of severe eutrophication in Ulansuhai nur in recent years, studied the immobilization effect of the modified materials on copper under different environmental conditions, and initially analyzed the immobilization mechanism. This study may help to understand the influence of water environment changes on heavy metals immobilization in the sediments.

GO (0.05 g) was added to 50 mL of deionized water and sonicated for 30 min to form a GO suspension. Subsequently, 0.5 g of FeSO₄^.7H₂O was accurately weighed and added into 2.5 mL of deionized water. 2 mol L⁻¹ NaOH was prepared, and added into the FeSO₄ solution. The mixed solution was then added into the GO suspension and ultrasonically shaken for 30 min to allow the reaction to proceed fully. The obtained mixture was heated to reflux at reflux temperature for 3 h, then filtered by a quantitative filter paper. The filtered matter was dispersed in an acetone solution, and finally dried at 180 °C overnight.

GO (0.3 g) was dispersed in 150 mL of ethanol and sonicated for 30 min at room temperature. Subsequently, 2.3 mL of 3-aminopropyl triethoxysilane (APTES) was added drop-wise into the GO suspension in N₂ atmosphere. Then the mixture was stirred and maintained at reflux temperature for 24 h. The filtered materials were washed five times with ethanol, and finally dried to obtain aminated graphene oxide (G-N).

0.5 g chitosan (CS) was dissolved in 10 mL 1% (v/v) acetic acid solution, and 1 g G-N was slowly added into the CS solution. The mixture was magnetically stirred for 10 min, then slowly heated to 50 °C, and stirred for 3 h. The final mixture was washed several times with acetone and filtered by a quantitative filter paper; the filtered materials were dried in vacuo to obtain G-N-C.

Scanning electron microscopy (SEM) images of GO, G-F and G-N-C were obtained using field emission environment SEM (FEI QUANTA 650 FEG). Fourier transform infrared spectroscopy (FTIR) (Thermo Nicolet-870 spectrophotometer) was used to record the FTIR spectra of GO, G-F and G-N-C, which ranged from 400 cm⁻¹ to 4,000 cm⁻¹. X-ray diffraction (XRD) patterns were retrieved using an X-ray diffractometer (RigakuD/max-2550 VK/PC diffractometer equipped with Cu Ka radiation at 40 kV and 40 mA) to verify the crystal structures of GO, G-F and G-N-C.

In this experiment, sediment sampling points and determination of the total amount of Cu were placed in an externally specified index (ESI). Cu was repaired by GO, G-F and G-N-C in situ. A total of 16 G-point sediment samples (10.000 ± 0.005 g) were accurately weighed. Four experimental design groups and a control group (10 g air-dried precipitate only) were investigated. The test groups included GO group [10 g air-dried precipitate, 5, 10, and 15% (w/w) GO], G-F group [10 g air-dried precipitate, 5, 10 and 15% (w/w) G-F], and G-N-C group [10 g air-dried precipitate, 5, 10 and 15% (w/w) G-N-C]. Each sample was placed in a 150 mL vial with 100 mL of deionized water. Three replicates were carried out for each experimental group. The samples were then incubated at room temperature for 35 days. The BCR extraction was done at the 1st, 3rd, 7th, 15th, 25th and 35th day. The concentrations of Cu in the overlying and interstitial water were measured at the 35th day.

A total of 16 G-point precipitation samples (10.000 ± 0.005 g) were accurately weighed and placed in a 250 mL beaker, respectively. Subsequently, they were divided into four groups: one 200 mL deionized water group, and three sodium phosphate solution (0.025, 0.5 and 1 mgL⁻¹) groups. The deionized water group was taken as a control group. An amount of 15% (w/w) of GO, G-F and G-N-C were respectively added into the other three groups. Then, the samples were kept in a hermetic space for 35 days. The experimental designs were repeated three times. Concentration of phosphorus, available phosphorus, pH, ORP and DO values in the overlying water were measured at the 1st, 3rd, 7th, 15th, 25th and 35th day. Finally, the BCR extraction was performed.

Figure 1(a) presents the typical wrinkled layered structure of GO. Figure 1(b) shows that Fe₃O₄ loaded on GO surface significantly increased the specific surface area of GO, which improved the adsorption sites for Cu. Figure 1(c) displays the grooves on the surface of G-N-C based on the different stiffnesses between chitosan and GO. Comparing with GO (in Figure 1(a)), the mass density of G-N-C (in Figure 1(c)) significantly increased, which indicated that GO and CS were coupled.

GO, CS, G-N, G-N-C and G-F composites were described by FTIR analysis. As shown in Figure 2(b), the –OH, C = O, –NH₂ and C–OH vibrations respectively occurred at 3,390, 1,720, 1,620 and 1,050 cm⁻¹ in GO spectrum (Yadav et al. 2014). In Figure 2(a), –OH at 3,390 cm⁻¹ moved to a high frequency, which indicated an enhancement of the hydrogen bonding between Fe₃O₄ and GO. –CH₂ vibration occurred at 2,880 cm⁻¹. The tensile vibrations of C–C, C–O and C = C were evident at 1,000–1,500 cm⁻¹ (Yan et al. 2016), and the C = O bond at 1,720 cm⁻¹ moved to 1,600 cm⁻¹, forming –COO⁻ (Song et al. 2017). G-F vibration at 604 cm⁻¹ was attributed to the Fe–O group (Yan et al. 2016), which demonstrated that Fe₃O₄ was loaded on the GO surface. In Figure 2(c), the C = O peak at 1,720 cm⁻¹ disappeared, and the –NH₂ at 1,620 cm⁻¹ moved to 1,560 cm⁻¹. At the same time, the peak of C–OH at 1,070 cm⁻¹ further widened (Sun et al. 2019). As shown in Figure 2(d), –OH vibration at 3,420 cm⁻¹, –NH₂ peak at 1,620 cm⁻¹, and C–O vibration at 1,410 cm⁻¹ were evident in the FTIR spectrum of CS. As shown in Figure 2(e), the broad peak at 3,430 cm⁻¹ was ascribed to the presence of –OH and –NH₂. The characteristic peak of –CH bond appeared at 2,930 cm⁻¹, while the C–OH bond was evident at 1,100 cm⁻¹. The change at 1,470–1,270 cm⁻¹ described the stretching of the C–N bond and the bending of the NH₂ and -NH⁻ groups (Zhuang et al. 2009). Hence, the new material G-N-C was successfully made.

Structural analyses of GO, CS powder, G-F and G-N-C composites were conducted via XRD (2θ = 10°–90°). As shown in Figure 3(a), 26.42° was the characteristic diffraction peak of GO, which corresponded to the 002 plane and exhibited narrow and strong features. For G-F composites in Figure 3(c), new diffraction crystalline peaks appeared at 2θ = 18.31°, 30.121°, 35.479°, 53.495°, 57.026° and 62.622° respectively, corresponding to 111, 220, 311, 422, 511 and 440 (Zhang et al. 2019b). These results indicated that Fe₃O₄ had been successfully loaded on the GO surface. In Figure 3(b), the characteristic peak of CS appeared at 2θ = 25.726° due to its amorphous structure (Samuel et al. 2019). As shown in Figure 3(d), the characteristic peaks of CS and GO respectively occurred at 2θ = 25.726° and 2θ = 26.42°. Therefore, certain GO and CS with amorphous structures were presented. Moreover, physical graft exchange between CS and GO might occur (Song et al. 2017).

The total amount of Cu in the sediments was 55.6 mg kg⁻¹. Heavy metals in the sediments could be divided into four different forms (acid soluble form, reducible form, oxidizable form, residual form), and the aforementioned four forms of heavy metals were interconvertible in the sediments. Figure 4 shows the composition changes of Cu in the sediments of GO, G-F and G-N-C. GO, G-F and G-N-C had different mass fractions.

Figure 4(a) clearly shows that the migration and conversion capacity of Cu significantly decreased when stable materials were added at the 35th day. Comparing with the result in the blank test (without GO, G-F and G-N-C), the acid extractable state respectively decreased by 7.76%, 13.36% and 19.28% after adding 5, 10 and 15% of GO. With an increased amount of GO, the residual state also slightly changed. Besides, comparing with the result in the control group, when 5, 10 and 15% of G-F were respectively added, the acid extractable state respectively decreased by 22.42, 29.35 and 32.36%, while the residual state respectively increased by 19.77, 26.38 and 26.49%. In addition, after adding 5, 10 and 15% of G-N-C, the acid extractable state respectively decreased by 21.42, 27.37 and 27.5%, while the residual state respectively increased by 19.02, 19.63 and 19.99%. The experimental results clearly showed that the acid extractable portion of Cu exhibited different reduced levels when the mass fractions of the three stable materials (i.e. GO, G-F and G-N-C) increased. Moreover, the residual state increased with an increase in mass fraction of the dosing stabilizer. The above mentioned three stable materials had the best ability for immobilizing Cu at 15% quality score.

Figure 4(a) exhibits that the order, in ability, for immobilizing Cu was G-F > G-N-C > GO when the added amounts of GO, G-F and G-N-C were approximately 15 wt%. The FTIR spectrum shows a large number of carboxyl and hydroxyl groups in G-F. The SEM indicates that Fe₃O₄ was embedded on the GO surface, which increased the specific surface area of GO and provided abundant adsorption sites for immobilizing Cu. Both FTIR and SEM demonstrate that CS-grafted GO increased the type of high-quality functional groups on the GO surface, while it did not expand the specific surface area of the GO. Therefore, the efficiency of G-F for immobilizing Cu was excellent due to an increase in specific surface area of G-F. Moreover, the oxygen-containing functional groups present on the surface of G-F accelerate the ion exchange between Cu and stable materials. Hence, the bioavailability of Cu was reduced and the Cu was immobilized in the deposit.

Figure 4(b) and 4(c) illustrate the concentration of Cu in overlying and interstitial water. In the control group (without GO, G-F and G-N-C), the concentration of Cu in overlying and interstitial water were respectively 0.033 mg L⁻¹ and 0.034 mg L⁻¹. The concentration of Cu in the overlying and interstitial water significantly decreased with an increase in the mass fractions of GO, G-F and G-N-C. When the stable materials with mass fractions of 5, 10 and 15% were added, the Cu content in overlying and interstitial water may be arranged in descending order as time increases: G-F> G-N-C> GO. Comparing with the result in the control group (without GO, G-F and G-N-C), the concentration of Cu in overlying and interstitial water respectively reduced by 44.6–53.5% and 3–42.7%. In the G-N-C experimental group, the concentration of Cu in overlying and interstitial water respectively reduced by 38.5–65% and 14.4–45.4%. In the G-F experimental group, the concentration of Cu in overlying and interstitial water respectively decreased by 61.5–66.3% and 22.4–47.1%. The exchange in soluble contaminants between the sediments and overlying water could be achieved through the interstitial water, and biological toxicity was closely related to the interstitial water. G-F excelled in reducing the concentration of Cu in overlying and interstitial water. Comparing with GO and G-N-C, G-F was more effective in reducing the risk of Cu being released back into the overlying water.

Previous studies indicated that phosphate compounds could effectively immobilize heavy metals such as Pb, Cd, Cu and Zn. However, phosphates in overlying water changed the DO of water, and the Ph and ORP of the sediments, which directly increased the ecological risk of heavy metals (Banks et al. 2012). In this study, three stable materials (GO, G-F and G-N-C) were added to investigate their effect on immobilizing Cu under different concentrations of phosphorus-containing overlying water.

The results in the control group in Figure 5 show that the stability of Cu continuously decreases as the concentration of phosphates in the overlying water increases (0.25, 0.5, and 1 mg L⁻¹). In the experimental group, when deionized water was added for aging for 35 days, the phosphate concentration in the overlying water was lower than the limit. The residual state of the GO group increased by 6.77%, while the acid extractable state decreased by 12.28%. The residual state of the G-F group increased by 10.36%, and the acid extractable state decreased by 13.37%. The residual state of the G-N-C group increased by 7.12%, while the acid extractable state decreased by 12.77%. When the concentration of phosphates in the overlying water was 0.25 mg L⁻¹, the residual state in the GO group increased by 8.68%, while the acid extractable state decreased by 16.56%. The residual state in the G-F group increased by 16.76%, while the acid extractable state decreased by 20%. The residual state in the G-N-C group increased by 15.89%, while the acid extractable state decreased by 18.89%. When the concentration of phosphates in the overlying water was 0.5 mg L⁻¹, and the aging time was 35 days, the residual state of the GO group increased by 10.25%, while the acid extractable state decreased by 12.3%. The residual state of the G-F group increased by 13.76%, while the acid extractable state decreased by 13.02%. The residual state in the G-N-C group increased by 12.5%, while the acid extractable state decreased by 12.89%. When the concentration of phosphates in the overlying water was 1 mg L⁻¹, and the aging time was 35 days, the residual state of the GO group increased by 12.5%, while the acid extractable state decreased by 12.3%. The residual state of the G-F group increased by 15.78%, while the acid extractable state decreased by 15%. In the G-N-C group, the residual state increased by 15.01%, while the acid extractable state decreased by 13.89%. Hence, prolonging the aging time led to a high concentration of phosphates in the overlying water and high bioavailability of Cu. The researchers also found a similar pattern in a previous study, and concentration of phosphates was closely related to the mobility of heavy metals (Liu et al. 2019a). After adding identical amount of GO, G-F and G-N-C, with an increase in the concentration of phosphates in the overlying water, the efficiencies for immobilizing Cu by the three materials increased at the 35th day. Moreover, the efficiency for immobilizing Cu by G-F was higher than those of GO and G-N-C.

Anion-induced adsorption and changing the negative charge of the adsorbent surface were the main mechanisms for immobilizing cations (such as Pb, Cu, Cd, Cr, etc) (Sanderson et al. 2016). Studies had shown that the presence of strengthened the adsorption of Pb on metal oxides, which might be related to co-adsorption. There were two reasons for co-adsorption, the first being related to the formation of surface ternary complexes, and the other being electrostatic attraction between positively charged Pb²⁺ and negatively charged ⁠. (Zhao et al. 2016). As shown in Figures 5 and 6, the increase in available phosphorus after adding GO and G-N-C was more notable than that under G-F; however, the effect for immobilizing Cu by GO or G-N-C was worse than that by G-F. Maybe the competitive adsorption for Cu between phosphates and stabilizer in the overlying water occurred after adding GO and G-N-C. Cu and phosphates formed unstable metal phosphates, which weakened the immobilization effect for Cu. When iron oxide was used to remove Pb and in water at the same time, iron oxide would first form complexes with Pb, and then these complexes would act as adsorption materials to further adsorb in water (Shi et al. 2020). Similar to Pb, phosphates had a good adsorption capacity for Cu (Zeng et al. 2017). Phosphate could form phosphate complexes with Cu, these complexes were insoluble in water, but soluble in ammonia, dilute acid and ammonium salt solution. In recent years, due to the inflow of external pollutants, the pH value of Ulansuhai nur had significantly changed, which would increase the possibility of copper phosphates in the sediments being re-released in the overlying water, causing harm to the ecological environment. After adding the stable materials, Cu could complex with the iron oxide on the surface of the G-F, thereby forming a more efficient adsorption material to adsorb ⁠. G-F shows good performance in the aspect of immobilizing Cu and adsorbing phosphates.

The experiment also measured the DO concentration of the overlying water, and the pH and ORP values of the sediments. At the same concentration of phosphates in the overlying water after 35 days of aging, Figure 7(b) shows that the pH value of the experimental group increased compared with the control group. In the experimental group, as the concentration of phosphates in the overlying water increased, the pH value in the original sediments was lower than that in the sediments after adding deionized water. At a concentration of phosphates in the overlying water of 0.25 mg L⁻¹, after adding GO, G-F and G-N-C stabilizers, comparing with the deionized water group, the decreased level in pH of the sediments was smallest when G-F was added. The sediments' ORP value in Figure S2 shows an opposite trend. This may be because the phosphates were unstable in the aqueous solution. In addition, under the weak alkaline conditions, would be hydrolyzed and release H⁺, which lowered the pH of the sediments. Increasing concentration of phosphates in the overlying water will cause a decrease in the pH value of the sediments. Generally, increasing pH value is beneficial for the adsorption and co-precipitation of Cu and metallic iron ore. In addition, G-F will reduce the possibility of forming unstable metal phosphates with Cu. Therefore, the effect of G-F on immobilizing Cu was better than that of GO and G-N-C.

The DO concentration in the overlying water confirmed the superior immobilizing capability of G-F for Cu. When the concentration of phosphates in the overlying water was identical after 35 days, Figure 7(a) shows that the dissolved oxygen gradually increased compared with the control group. In the low oxygen state, phosphates were commonly adhered to the surface of the deposit (Zhang et al. 2019a), and Cu formed an unstable heavy metal phosphate with phosphates, which increased the ecological risk of Cu. In the experimental group, with an increase of phosphate concentration, after adding GO, G-F and G-N-C stabilizers, the reduction of levels of dissolved oxygen in the G-F group was higher than that in the GO and G-N-C groups. G-F reduced the possibility of phosphate adhering to the surface of the sediment; furthermore, the efficiency of immobilizing Cu by G-F was higher than that by GO and G-N-C.

In this study, the effects of GO, G-F and G-N-C on immobilizing Cu in the sediments were compared. The results are discussed as follows. (1) Through the characterization of G-F and G-N-C, the results showed that the specific surface area of GO increased with the amount of Fe₃O₄ loaded, by grafting chitosan to GO, the number of surface functional groups increased and the structure became tight. (2) Increasing the mass fraction of the modified materials from 5% to 15%, the effect of immobilizing Cu in the sediments strengthened. (3) The efficiency of stable materials on immobilizing Cu from high to low is: G-F> G-N-C> GO. (4) Lower concentration of phosphates in the overlying water helped stabilize materials to reduce the ecological risk of Cu in the sediments. At the same concentration of phosphates in the overlying water, G-F had the best ability to immobilize Cu in the sediments, and it could be inferred that G-F could play a positive role in the treatment of phosphorus-rich water. G-F is a high-efficient and environmentally friendly material that can be used for contaminated site remediation. It is of great importance that the effect of eutrophication on the stability of heavy metals should also be taken into consideration.

This study was financially supported by the National Natural Science Foundation of China (No. 21107041); the Young and Middle-Aged Academic Backbone Program of the Inner Mongolia University of Technology; the Natural Science Foundation of Inner Mongolia (No. 2020MS05028); the Department of Education of Inner Mongolia Autonomous Region (No. NJZY17095).

All relevant data are included in the paper or its Supplementary Information.