Adsorptive removal of Cr(VI) and Cu(II) ions from aqueous solutions by a natural moss

Contamination of heavy metals (HMs) from industrial activities to nature induces severe threats to living organisms and the environment. The accumulation of HMs in different body organs generates unwanted side effects because of their non-biodegradable nature and long biological half-lives (Alalwan et al. 2020; Pushkar et al. 2021; Ismail et al. 2022). Chromium, cadmium, copper, nickel, mercury, and lead are among the most common HMs that cause toxic effects after exceeding a certain concentration level in living organisms (Tumolo et al. 2020). The toxicity of chromium and its effects on the ecosystem are of great importance and require attention because of its worldwide and direct usage in plating, steel industry, and leather tanning, and utilization as an additive in the production of pigments, catalysts, and corrosion inhibitors (Mohan et al. 2011; Sharma et al. 2021). Chromium has two stable forms, Cr(III) and Cr(VI), among a wide range of its oxidation states (Zhang et al. 2014; Giri et al. 2021). Although Cr(III), an essential nutrient, is necessary for living metabolisms, it is dangerous in case of overdose (Behera et al. 2020). On the other hand, Cr(VI) exhibits higher toxicity than Cr(III) compounds, and the toxicity potential between these oxidation states is complex (Bost et al. 2016; Zhang et al. 2021). The cellular uptake of Cr(VI) is more effective than Cr(III), therefore Cr(VI) penetrates the cell smoothly, while Cr(III) cannot (Sharma et al. 2021). An important detail about the biological systems is the generation of free radicals during the reduction of chromium, which can form intracellular targeting complexes in the body (Pushkar et al. 2021). Chromium is reported to enter the cell and alter the DNA through a mechanism of reducing to other transition states, Cr(V) and Cr(IV) intermediately, and Cr(III) finally as an end by-product during the generation of free radicals through enzymatic and non-enzymatic activities (Pushkar et al. 2021; Sharma et al. 2021). Finally, chromium exposure results in a spectrum of damages, including red blood cell damage (Zhang et al. 2014), oxidative DNA damage, chromosomal aberrations, DNA strand breaks, and Cr-DNA adducts and a number of ailments due to normal hematological parameters’ alterations (Nickens et al. 2010).

Copper is commonly used in petroleum refinery, mining, plating, plumbing, automotive, machinery, and architecture owing to its corrosion resistance, high electrical conductivity, malleability, and strength, and such industrial applications cause large amounts of industrial effluents contaminated by copper (Lee et al. 2021). Besides industrial activities, copper and its compounds widely used in some fields such as valves, coatings, alloys, and interior plumbing, also drive attention since it directly contaminates drinking water in variable concentrations (Bost et al. 2016; Lee et al. 2021). Another reason for Cu(II) contamination is the usage of CuSO₄·5H₂O as an additive to surface waters to prevent algae formation (Zong et al. 2022). Copper is a nutrient for organisms, also it is a pollutant for the environment. Copper is crucial as a catalytic cofactor for organisms for some processes occurring biologically. On the other hand, the overloading of copper causes oxidative damage and death of the cells due to the Fenton-type redox reactions (Bost et al. 2016).

Removal of hazardous HMs during discharge control of the aquatic environment is one of the challenges of the 21st century, and treatment becomes a complex problem to obtain HMs contaminated wastewater with concentrations of less than 0.5 g L⁻¹ (Volesky 2001). Electrodialysis, conventional ion-exchange, reverse osmosis, cementation, and precipitation have one or more of being high-cost, technically complicated, high energy and chemical requiring, secondary polluting demerits for HMs uptake from large amounts of industrial wastewater (Ye et al. 2012). Also, complexation, adsorption, diffusion, chelation, or precipitation are the possible contributing mechanisms depending on the specifications of a biomass during the biosorption of HMs (Volesky 1990). Adsorption is an inexpensive and technically basic method for the removal of HM ions from wastewater (Cao et al. 2013; Sobhanardakani et al. 2016). The most important step in the adsorption process is the selection of adsorbent with high adsorption capacity, abundance, and low cost. Activated carbon (AC) and biochar are considered good options as adsorbents in adsorption studies due to their porous structure with a large surface area which leads to a high adsorption capacity by providing active sites for the adsorption of pollutant species (Nworie et al. 2022; Jurgelane & Locs 2023). However, AC is quite expensive and its selectivity and reusability are limited. Similarly, biochar is a relatively expensive material that requires significant energy to produce. Additionally, since biochar forms agglomerates in aqueous solution, it is difficult to separate it from the solution by centrifugation or filtration. Due to these disadvantages, both AC and biochar are less preferred in the adsorption studies. Materials available locally in nature, agricultural wastes, industrial products, and rock minerals with high-metal uptake capacities have been proposed as low-cost adsorbents in recent studies (Ye et al. 2012). Biological materials, which are also used to obtain coagulants in the treatment of wastewater, provide potentially available and more operative sorbents in HM retention from wastewater (Gong et al. 2005; Peña-Guzmán & Ortiz-Gutierrez 2022).

Removal of Cr(VI) ions via adsorption onto several adsorbents, such as mango and jackfruit seed kernel (Giri et al. 2021), Pinus brutia ten (Ozdes et al. 2014), hoop pine leaves (Dilshad et al. 2021), and spent Ulva prolifera macroalga (Vinayagam et al. 2022), has been experimentally studied. Cu(II) is a highly demanded HM in industrial applications and anthropogenic activities which require high-cost wastewater treatment methods. Bio-modified geopolymer composite (Mama et al. 2022), sludge of industrial palm oil mill (Lee et al. 2021), immobilized bacteria (Duran et al. 2009), Rosa canina seeds (Serencam et al. 2014), and magnetic composites of attapulgite/chitosan blended with bacterial cellulose nanofibrils (Chen et al. 2020) have been used for the removal of Cu(II) ions.

Studies have been noticed in which different moss species are utilized in the adsorption of HM ions. Low et al. (1997) have evaluated the adsorption of Cr(III) ions from synthetic solution and tannery waste using Calymperes delessertii Besch. Lee et al. (1995) have used natural and copper-coated moss (Calymperes delessertii Besch) for the removal of Cr(III) and Cr(VI) from aqueous solutions. According to our literature review, Racomitrium ericoides (Brid.) Brid., a type of moss has not previously been used as an adsorbent for the removal of organic or inorganic pollutants, or as a biomonitor in determination of accumulation levels of the pollutants. There is only one study in which R. ericoides (Brid.) Brid. was utilized as a bioindicator for determining the concentrations of ¹³⁷Cs in the High arctic (Saniewski et al. 2020). In the present study, R. ericoides (Brid.) Brid. (REB) was utilized as an adsorbent for the first time in removal of Cr(VI) and Cu(II) ions from aqueous media. Target ions were successfully removed from model solutions and a new and cost-effective method was developed as an alternative to the high-cost techniques to eliminate Cr(VI) and Cu(II) from industrial wastewater. The utilization of a three-layer artificial neural network (ANN) model was assessed to speculate the uptake efficiency of HMs. In the model, some independent experimental parameters, such as solution pH, equilibrium time, REB quantity, and initial HM concentration were employed as input parameters to train the neural network. Simultaneously, the percentage of adsorption, a dependent variable, was regarded as the output layer of the neural network. On the other hand, significant experimental factors that directly affect the adsorption efficiency of HMs were investigated in detail through the batch system and the most suitable conditions were determined. Reusability studies were carried out to evaluate the stability of the adsorbent, and desorption conditions were also investigated to determine the regeneration possibilities of the adsorbent. The kinetic and isotherm models were evaluated according to the experimental data to detail the process in terms of mechanism.

All chemicals used at different stages of the study including Cu(NO₃)₂·3H₂O, K₂Cr₂O₇, NaCl, CaCl₂, NaHCO₃, Na₂CO₃, NaOH, HNO₃, and HCl were of analytical purity and obtained from Fluka (Buch, Switzerland) or Merck (Darmstadt, Germany). To prepare the stock solutions of Cr(VI) and Cu(II) ions at 5,000 mg L⁻¹, the calculated amounts of Cu(NO₃)₂·3H₂O and K₂Cr₂O₇ were weighed with Sartorius BP 1106 model analytical balance. After dissolving with distilled water, the final volume was made up to 250 mL. Different concentrations of metal solutions used in the experiments and the calibration solutions used in the metal analysis by Flame Atomic Absorption Spectrometer (FAAS) were obtained by diluting the stock metal solutions. Diluted solutions of NaOH and HNO₃ were employed to adjust the initial pH of the metal solutions. To clean the glassware used throughout the studies, tap water and distilled water were used, respectively, after being kept in 5% HNO₃ solution for a certain period of time.

R. ericoides (Brid.) Brid. (REB) picked up from the countryside of Trabzon, a city in the Blacksea region of Türkiye, was washed with tap water and distilled water to clear any environmental dust, and then laid on a filter paper for 3 days in a dark and moisture-free environment before being dried in an oven at 105 °C for 3 h. Dry moss was ground in a blender. The particles smaller than 180 μm were stored in a glass bottle until the experimental studies and used as an adsorbent directly without being treated by any supporting chemicals or modifiers. A Fourier Transform Infrared Spectrometer (Perkin Elmer 1,600 Series) was utilized to explicate the functional groups of REB while its morphological structure was clearly studied by Scanning Electron Microscope (ZIESS Evo Ls 10). A QuantaChrome Autosorb iQ2 Automated Gas Sorption Analyzer was used to obtain the N₂ adsorption/desorption isotherms. The surface area of REB was calculated using Brauner–Emmer–Teller (BET) methods. Apart from these techniques, point-of-zero-charge (pH_pzc) determination was performed for the characterization of the adsorbent.

The ANN model was employed using the Neural Network Toolbox in MATLAB R2017b mathematical software. The architecture of the model consisted of three basic layers: the input layer, hidden layer, and output layer. The input parameters selected for this research were solution pH, contact time, REB quantity, and initial HM concentration, while the adsorption percentage was characterized as the output parameter. A neural network model employing three layers and utilizing backpropagation (BP) was trained. The hidden layer was characterized by a tangent sigmoid transfer function (tansig), while the output layer featured a linear transfer function (purelin). The training of the network was conducted using the Levenberg–Marquardt backpropagation (LMB) algorithm. The performance of the network was assessed using the correlation coefficient (R²). The optimal architecture of the ANN model was determined as 4-6-1.

The specific surface area of REB calculated by the BET method was found to be 5.15 m² g⁻¹. The pH value at which the adsorbent's surface becomes electrically neutral is the term ‘point-of-zero charge’ (pH_pzc) in the literature (Moharami & Jalali 2013). The pH_pzc value of REB was determined to be 3.4. Functional groups on the REB surface were electrically charged as negative at higher pH values (pH > pH_pzc) and charged as positive at lower pH values (pH < pH_pzc).

Cr(VI) ions are usually available in aqueous solutions in the forms of dichromates (⁠⁠), chromates (⁠⁠) and acid chromates (⁠⁠), and some other oxyanions (Ozdes et al. 2014). Acid chromate anions are dominantly more than other forms at lower pH values. In contrast, chromate ions are more prevalent than others at higher pH levels. As the initial pH is increased from 1.0 to 2.0, q_e values increased from 8.2 to 10.3 mg g⁻¹, and then with an increasing initial pH up to 4.0, adsorbed amount of Cr(VI) decreased to 2.5 mg g⁻¹ (Figure 3). At higher values of initial pH (pH > pH_pzc), functional groups of the REB surface charged negatively and repulsion forces between the REB surface and Cr(VI)-related anions caused a decrease in adsorption. Similar to these findings, Yaashikaa et al. (2019) have also deduced that the increase in initial pH caused a decrease in the adsorption efficiency of Cr(VI) onto mixed adsorbent, and hence the optimum pH value for Cr(VI) adsorption was reported as 2.0. As a result, subsequent experiments were planned by optimizing the initial pH of Cr(VI) and Cu(II) solutions to 2.0 and 5.0, respectively.

The C value and the rate constants (k_id,1 and k_id,2) were obtained by evaluating the IPD model. According to data in Supplementary material, Table S1, the rate constants of intraparticle diffusion (k_id,2) were smaller than the rate constants of film diffusion (k_id,1), which represent that intraparticle diffusion was the rate-limiting step in Cu(II) and Cr(VI) adsorption in this study. C parameters in the behavior of not cutting across the origin gave the additional information that not the intraparticle diffusion by itself, but both the surface film and intraparticle diffusion possibly limited the reaction rate in the explanation of the complex mechanism of Cu(II) and Cr(VI) pollutants' adsorption onto REB (Maliyekkal et al. 2008).

Isotherm models were evaluated to elucidate the adsorption mechanism in the uptake process of the relevant ions. Herein, possible mechanisms of Cu(II) and Cr(VI) adsorption onto REB at the equilibrium are defined by Langmuir, Freundlich, and Dubinin–Radushkevich (D–R) isotherm models.

The Langmuir model relies on monolayer adsorption which consists of specific homogeneous sites and prevents the interactions between adsorbed species (Langmuir 1918), while the characteristic of the Freundlich model is multilayer adsorption owing to heterogeneous surface built by interactions between species adsorbed by differently energized surface sites (Freundlich 1906). Extra knowledge about the adsorption type is gained from the D–R isotherm model (Dubinin & Radushkevich 1947).

Equilibrium data were used to calculate the isotherm constants (Table 1) through the linear graphics of C_e/q_e vs. C_e (Langmuir), lnq_e vs. lnC_e (Freundlich), and lnq_e vs. ε² (D–R), as demonstrated in Supplementary material, Figure S1 for Cu(II) and in Supplementary material, Figure S2 for Cr(VI) ions. The correlation coefficients (R²) higher than 0.9 for the Langmuir and Freundlich models were both acceptable for expressing the mechanisms of Cr(VI) and Cu(II) adsorption, indicating an external surface that consisted of some active sites distributed homogeneously and some other active sites distributed heterogeneously on REB. Bayisa et al. (2021) also reported that the same models fitted well for Cr(VI) adsorption onto Catha edulis by considering the R² values close to each other. The maximum adsorption capacity of REB was obtained as 41.2 mg g⁻¹ for Cr(VI) and 22.7 mg g⁻¹ for Cu(II) adsorption, by using the equation as described in the Langmuir model. Maximum Cr(VI) and Cu(II) adsorption capacities of REB and various adsorbents (Sharma & Forster 1993; Duran et al. 2009; Mohan et al. 2011; Serencam et al. 2014; González et al. 2016; Zhu et al. 2018; Yaashikaa et al. 2019; Chen et al. 2020; Yahya et al. 2020; Lee et al. 2021; Ozdemir et al. 2021; Vinayagam et al. 2022) were tabulated for a comparison (Table 2).

Between the lowest (50 mg L⁻¹) and the highest (1,000 mg L⁻¹) initial HM concentrations, R_L values ranged from 0.727 to 0.118 for Cr(VI) and 0.794 to 0.161 for Cu(II) ions, indicating that Cr(VI) and Cu(II) removal by REB is favorable over the method developed in this study. The n values acquired as 2.0 for Cr(VI) and 2.6 for Cu(II) from the Freundlich model also emphasized the favorability of Cr(VI) and Cu(II) retention by REB. Calculated values of E by using the D–R model were 8.51 kJ mol⁻¹ for Cr(VI) and 9.90 kJ mol⁻¹ for Cu(II) retention, both of which correspond to ion-exchange-based adsorption range (8–16 kJ mol⁻¹). Lee et al. (2021) also reported that Cu(II) sequestration by the sludge of industrial palm oil mill occurred via ion-exchange mechanism with a mean adsorption energy E of 9.3 kJ mol⁻¹.

The desorption of Cu(II) ions was assessed by using HCl and NaOH as the eluting agents with the concentrations differing between 0.5 and 2.0 mol L⁻¹, and the Cr(VI) desorption was evaluated by using NaOH as an eluting agent in a concentration range of 0.25–2.00 mol L⁻¹ to enable the recovery of the HM ions. HM ion concentrations were 100 mg L⁻¹ with initial pH values of 2.0 for Cr(VI) and 5.0 for Cu(II), and REB dosage was 5.0 g L⁻¹ for all samples in the stage of adsorption. Suspensions were subjected to an adsorption process for 360 min and then centrifuged. Pollutant-loaded wet REB samples were washed with distilled water and dried in an ambient atmosphere for 24 h. HM ion loaded dry adsorbent was treated by each of the eluting agents in a volume of 10 mL for 360 min. The suspensions were centrifuged to separate the adsorbent from the solution before FAAS analysis of the remaining ions. The desorption percentages of Cu(II) and Cr(VI) ions are given in Table 3.

The results of desorption studies showed that the regeneration efficiency of Cu(II) ions nearly did not change by increasing the concentration of NaOH. However, the desorption percentage of Cu(II) has increased from 66.6 to 76.8% by increasing HCl concentration from 0.5 to 1.0 mol L⁻¹. On the other hand, increasing the concentration of NaOH from 0.25 to 2.00 mol L⁻¹ resulted in an increase in the desorption percentages for Cr(VI) from 15.5 to 21.1% (Table 3). Clearly, Cr(VI) and Cu(II) ions could not succeed to be desorbed from REB. This result conceivably commented as the adsorption of these ions onto the adsorptive sites of REB may have occurred chemically and mostly irreversibly because of the strong bonding interactions between the surface functional groups of moss and Cu(II) or Cr(VI) ions.

To further reduce the cost of the developed process and to have an idea about the stability of the adsorbent, the effectiveness of REB in adsorbing HMs was tested without regeneration. The experiments were carried out using an initial Cu(II) and Cr(VI) concentration of 50 mg L⁻¹ with 5.0 g L⁻¹ of REB suspension. After the experiments were continued for 360 min, the adsorbent was separated and then treated with another 50 mg L⁻¹ of HM solution. The process was repeated for five times. It was concluded that although the adsorption capacity of REB decreased to a certain extent with each subsequent loading, the adsorbent could be used at least five times without regeneration (Supplementary material, Figure S3).

In order to apply the ANN model for the uptake efficiency of HMs onto REB, the dataset, comprising 25 experimental points, was randomly divided into three subgroups: training (70%), validation (15%), and testing (15%). A three-layer BP neural network model was utilized with tansig in the hidden layer and purelin in the output layer. Different neuron numbers were assessed, and the optimal performance was obtained with six neurons in the hidden layer. Subsequently, several training studies were carried out to obtain the best weights, errors, and correlations. The ANN regression plots depicted the relationship between the network output and the adsorption efficiency for Cu(II) and Cr(VI) are given in Supplementary material, Figures S4 and S5, respectively. These plots show the training, validation, testing, and overall prediction sets. The regression analysis demonstrated a close match between the network output values and the experimental adsorption efficiency. Remarkably high overall prediction set R² values of 0.994 for Cu(II) and 0.990 for Cr(VI) were obtained. These findings underscore the suitability of the ANN model for accurately predicting HM adsorption efficiency.

In this study, REB was used for the first time as a low-cost, efficient adsorbent for the removal of HM ions from aqueous media. Removal of Cr(VI) and Cu(II) ions was studied through batch experiments, and various parameters affecting the adsorption efficiency were investigated to figure out the best conditions of the relevant process. REB was identified to have a higher capacity of adsorption than many other adsorbents used for the removal of Cr(VI) and Cu(II) in the literature. Optimum initial pH values were observed as 2.0 and 5.0 for Cr(VI) and Cu(II) adsorption, respectively. Adsorption efficiencies reached their maximum values at 360 min of contact time at which both systems reached the equilibrium. The PSO kinetic model was dominant in the adsorption, and E values computed by evaluating the isotherm models pointed out that the occurring process was based on ion-exchange mechanism. Cr(VI) and Cu(II) contamination requires great attention because of the wide usage of these metals in industrial applications which result in large amounts of polluted wastewater.

In general, the present study suggests that the use of readily available natural moss for the adsorption of HM ions can be an economical and sustainable alternative to reduce environmental pollution and clean water resources. Compared to chemical adsorbents or other methods, the use of moss is more environmentally friendly. Studies in the literature show that natural moss can be used successfully for the adsorption of pollutant species. In the current study, the effectiveness of natural moss in the adsorption of HMs was assessed by providing detailed analyses on topics such as optimum experimental conditions, adsorption kinetics and isotherms, thus introducing this natural material into the literature. One limitation of the study is the low adsorption capacity of natural REB. Therefore, in our future studies, we will increase the adsorption capacity by modifying REB in different methods. Additionally, the usability of both natural and modified REB in the adsorption of organic pollutant species will be tested.

The financial support from the Unit of the Scientific Research Projects of Karadeniz Technical University is gratefully acknowledged.

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

The authors declare there is no conflict.