0 Introduction

Nowadays, water pollution is a global issue. According to the latest United Nations world water development report, there are more than 2 billion people who have limited access to safe drinking water^[1]. Studies revealed that the river and air at Kuantan, Malaysia were polluted by bauxite, which consists of high amount of aluminium oxide and ferric oxide, due to mining activities^[2]. Study reported by Hafiza et al.^[3] showed that heavy metal concentration level in groundwater at Rosob village, Ampar Tenang and Machang exceeded the permissible limit set by Malaysian drinking water quality standard. Well water at Rosob village contains 409.05 ug/L of Mn²⁺. This is higher than 100 ug/L, which is the permissible limit set by Malaysian Ministry of Health. Recent study by Yunus et al.^[4] showed that the coastal sediment in Peninsular Malaysia is contaminated by low, moderate and chronic level of heavy metals comprised of Pb, Hg, Cd, Cu, and Zn.

Among all the pollutants in water, heavy metal such as uranium exhibits high toxicity even at very low concentration level^[5]. Heavy metal is non-degradable and persistent in the environment. Due to their high solubility in water, they can easily penetrate living organisms, and eventually accumulate in human body after entering the food chain. Trivial amount of heavy metals consumption such as copper is essential to know in human and animal bodies^[6]. However, at higher dose of ingestion, they become threats, thus causing serious health disorders such as stunted growth, cancer, organ damage, damage of nervous system, and fatality^[7].

Heavy metal polluted water is mainly generated from industrial activities such as mining, electroplating, wood processing, and painting. With the availability of advance technologies, heavy metals can be removed by chemical precipitation and membrane filtration^[6]. However, chemical precipitation consumes large amount of chemical such as lime and oxidants which require physicochemical monitoring of effluent especially pH^[7]. Additionally, toxic sludge is also produced^[7-8]. Meanwhile, high pressure is required for both reverse osmosis and nanofiltration filtration^[9-10]. Bisht et al.^[11]reviewed several methods for heavy metals removal and concluded that adsorption is the most popular method due to its low cost, availability, and eco-friendliness.

The use of agricultural wastes for heavy metals gained a lot of attention from researchers lately due to its biodegradable properties, availability in large quantity, and high mechanical strength^[12]. Numerous efforts have been done to process the waste to achieve high heavy metals adsorption, as shown in Table 1. Generally, the samples were firstly washed and dried in the oven before treatment and dried again before being pulverized for adsorption use. The findings showed that both acids and alkali treated samples have positive impacts to heavy metal removal^[13-20]. However, limited research works have been done to provide alternative treatment process for agricultural waste. Additionally, the potential of other agricultural waste such as banana and dragon fruit peels for Cu²⁺ removal has not been discovered in a comprehensive manner.

Thus, the objective of this study is to propose a new approach to treat banana peel by using ZnCl₂ prior to alkali and acid treatment for Cu²⁺ removal and explore the potential of dragon fruit peels as an adsorbent for Cu²⁺ removal. Seven adsorption isotherms including Linear, Freundlich, Redlich-Peterson, Sips, Toth, Langmuir, and Volmer were studied to describe the adsorption mechanism.

1 Materials and Methods 1.1 Materials

The skins of banana and dragon fruit were firstly peeled, cleaned, and cut into small pieces before being dried in the oven for 24 h at 105 ℃. After the drying process, the samples were stored at room temperature. All the samples were pulverized before being used. The images of the dried samples are shown in Fig. 1. Chemicals such as CuCl₂, NaOH, and HCl were used without further purification.

1.2 Chemical Activation of Sample Adsorbents

The dried sample was chemically activated by using ZnCl₂^[21]. 20 g of dried sample was soaked in 10 g/mL of Zn solution for approximately 2 h. The sample was filtrated and dried in oven for 24 h at 80 ℃. After the drying process, the sample was treated with 0.1 M HCl and 0.1 M NaOH subsequently to remove the residual organic compounds. Then, the sample was washed with hot water until pH 7 was achieved. Lastly, the sample was dried in oven at 80 ℃ before being used for the Cu²⁺ adsorption experiment. Functional groups of the dried fruit peels were characterized by using Fourier Transform Infrared Spectroscopy (FTIR).

1.3 Batch Adsorption Studies

The adsorption experiment was conducted at room temperature by dispersing 0.5 g to 1.5 g of dried and pulverized banana (B), dragon fruit (D), chemically activated banana (Act.-B), and dragon fruit (Act.-D) to 100 mL of 10×10^-6 Cu²⁺ solutions. The pH of the solution was initially adjusted to pH 6. The experiment was conducted for 2 h. The sample was filtered using syringe filter (Thermo fisher, 0.22 micron). The concentration of Cu²⁺ before (C_i) and after (C_f) was measured using Atomic Absorption Spectrometer (Agilent Technologies, Model: 240 AA). The percentage of removed Cu²⁺ was calculated by using the following equation:

$ {\rm{ Percentage\;of\;removal }} = \frac{{\left( {{C_{\rm{i}}} - {C_{\rm{f}}}} \right)}}{{{C_{\rm{i}}}}} \times 100\% $

(1)

Isotherm study was conducted by repeating the adsorption experiment using 1 g of adsorbent in 10×10^-6 - 40×10^-6 of Cu²⁺ solution. In this study, only Act.-B was selected due to the best adsorptive performance. The equilibrium relations for the best adsorbents were identified by fitting the data to isotherms, as shown in Table 2.

The equilibrium adsorption was calculated by using Eq. (2) below, where c_e is the concentration of Cu²⁺ at equilibrium point, V is the volume of the sample, and m is the mass of the adsorbent used.

$ {q_{\rm{e}}} = \frac{{\left( {{c_{\rm{i}}} - {c_{\rm{e}}}} \right)V}}{m} $

(2)

The following statistical parameters (Eqs. (3)-(6)) were calculated to verify the best isotherm to present the adsorption study.

$ {R^2} = \frac{{\sum {{{\left( {{q_{{\rm{mean }}}} - {q_{{\rm{cal}}}}} \right)}^2}} }}{{\sum {{{\left( {{q_{{\rm{cal }}}} - {q_{{\rm{mean }}}}} \right)}^2}} + \sum {{{\left( {{q_{{\rm{cal }}}} - {q_{{\rm{exp }}}}} \right)}^2}} }} $

(3)

$ {\chi ^2} = \frac{{\sum {{{\left( {{q_{\exp }} - {q_{{\rm{cal}}}}} \right)}^2}} }}{{q_{{\rm{cal}}}^2}} $

(4)

$ {\rm{SSE}} = \sum {{{\left( {{q_{\exp }} - {q_{{\rm{cal}}}}} \right)}^2}} $

(5)

$ {\rm{RMSE}} = \sqrt {\frac{1}{{{N_{\exp }}}}\sum {{{\left( {{q_{\exp }} - {q_{{\rm{cal}}}}} \right)}^2}} } $

(6)

q_exp is the experimental adsorption capacity, q_cal and q_mean are the calculated and average value of experimental adsorption capacity respectively, N_exp is the number of data points.

2 Results and Discussion 2.1 FTIR

Figs. 2(a) and (b) show the FTIR spectra for both B and Act.-B. The broad band within the range of 3000 cm^-1 to 3600 cm^-1 and the peak at 3855 cm^-1 indicate the existence of hydroxyl group (-OH), which is available in cellulose of fruit peels^[22]. The peak around 2920 cm^-1 is due to the stretching of alkane group (C-H)^[23]. Peak at 1730 cm^-1 in Fig. 2(a) is due to the presence of carboxylic acid or ester group (C=O)^[23]. Meanwhile, the bands between 1450 cm^-1and1700 cm^-1 region are usually assigned to the alkene double bond (C=C). Lastly, the peak in the range of 1050 cm^-1 to 1450 cm^-1 is caused by the presence of C-O bond in carboxylic acid or alcohol group^[22]. All the characteristic bands observed in this study is consistent with the FTIR spectra reported by other researchers^[23-24]. It is notable that ZnCl₂ activation treatment contributed different impact on different fruit peel. ZnCl₂ caused the formation of new C=C bonds at 1506.83 cm^-1 in Act.-B, while C-H and C=O bonding disappear in Act.-D as shown in Table 3. This finding is similar with the finding reported by Xie et al.^[25] where the intensity for C-H and C=O bands was reduced after the activation treatment on the orange peel. This showed the crystallinity of cellulose was affected and destructed in the activation treatment.

Generally, intensities of the peaks would increase and slightly shift after the heavy metal adsorption process, especially the O-H, C-O bonding due to the sharing of electron valence^[26]. Similar finding was also reported by Lahieb Faisal^[27] where the dried banana peels used to adsorb copper showed significant shifts in the bonding between 1300 cm^-1 and 1716 cm^-1. Thus, it is reasonable to deduce that the copper adsorption would contribute similar impact to the adsorbents in this study.

2.2 Adsorption Performance

Fig. 3 shows the adsorptive performance of 4 types of adsorbents on Cu²⁺ removal. Generally, when the mass of adsorbent increased, the Cu²⁺ removal rate increased. This is due to higher number of available active site for binding purpose when the mass of adsorbent increased^[28]. This can be clearly seen from the adsorption trend of B. The Cu²⁺ removal increased from 4.92 % to 10.53 % and 12.67 % when the mass of B increased from 0.5 g to 1.0 g and 1.5 g. Similar trend is observed when D and Act.-D were used as the absorbent. However, it is notable that the degree of removal rate was reduced when the mass of D and Act.-D was increased from 1.0 g to 1.5 g. Additional 2% and 4% of Cu²⁺ removal was achieved when the mass of D and Act.-D was increased by 50%. It is reasonable to deduce that the active sites of the adsorbent were aggregated when the adsorbent was provided in excess^[16]. This reduces the number of active binding sites for adsorption purpose.

This effect is very significant when Act.-B was adopted as the adsorbent. Approximately 3% increment in Cu²⁺ removal rate was found when the mass was increased from 0.5 g to 1.0 g. This suggested that the optimum dosage to remove 10×10^-6 of Cu²⁺ is 0.5 g of Act.-B, where the final concentration is 0.636×10^-6. This meets the discharge limit of Standard B stated in The Environment Quality (Industrial Effluents) Regulation 2009, which is 1 mg/L^[29]. However, in order to meet the discharge limit of Standard A (0.2 mg/L), 1.5 g of Act.-B can be used to treat the Cu²⁺ to 0.128 mg/L.

Fig. 3 also shows that Act.-B and Act.-D performed better in Cu²⁺ removal compared with untreated B and D. When dried B and D were immersed in ZnCl₂ solution, the solution intercalated into the carbon matrix in both B and D. The reaction between ZnCl₂ and carbon atoms occurred between the extended interlayers of carbon. Pores were generated within the sample when it was heated in oven^[30]. The pores increase the surface area for Cu²⁺ adsorption, thus explaining the higher Cu²⁺ removal rate in Act.-B and Act.-D compared with B and D. This is further supported by the finding from Wirasnita et al.^[31] where the surface of ZnCl₂ activated empty fruit bunch (EFB) for adsorption purpose was 86.62 m²/g, while 9.09 m²/g was reported for untreated EFB. Similar finding was reported by Spagnoli et al.^[32] where equilibrium adsorption capacity was 350 mg methylene blue/g for ZnCl₂ treated cashew nut shell while approximately 70 mg methylene blue/g was reported for untreated cashew nut shell.

It is notable that Act.-B performed at least 30% better than Act.-D in terms of Cu²⁺ removal. This is due to the presence of hydroxyl and carboxyl groups in Act.-B, as shown in Fig. 2 and Table 3. Furthermore, only hydroxyl groups are present in Act.-D. The positive impact of oxygen containing functional groups for heavy metals removal is revealed by Yang et al.^[33], and it has been proved by Li et al.^[34] in Pb²⁺ removal, Tian et al.^[35] in Cr⁶⁺, and Tan et al.^[36] in Cu²⁺ and Pb²⁺.

2.3 Adsorption Isotherm

Adsorption isotherm study was conducted on the best performed adsorbent Act.-B with a maximum q_e value of 1872.8 mg/g. The adsorptive capacity of the adsorbents in this study is comparable with the data reported by Hossain et al.^[16], Prabu et al.^[17] and Feizi and Jalai^[19], as shown in Table 1. Fig. 4 shows the predicted and experimental Cu²⁺ adsorption using Act.-B. Freundlich, Redlich-Peterson, Sips, and Toth models fitted the experimental data well with R² more than 0.99, as shown in Table 4. Among these models, Freundlich model showed the lowest SSE, χ², and RMSE values. This indicates heterogeneous adsorption due to the diversity of adsorption sites^[23] and multilayer adsorption^[37]. Redlich-Peterson and Sips models are the hybrid Freundlich and Langmuir models^[38]. These models could be reduced to either Langmuir or Freundlich under specific condition. The condition includes low initial absorbate concentration of feed solution, which reduces Sips model to Freundlich model. This explains the high R² obtained from Redlich-Peterson and Sips models in this case.

According to Wang and Hwang^[39], layer thickness, layer coverage, effective coverage, deactivated fraction, isosteric heat and equilibrium constants of adsorption are the essential properties of a multilayer adsorption process. These properties are interrelated and it is strongly governed by the surface characteristic of adsorbents, especially the pore size distribution^[40]. Attempts such as steam pyrolysis and surface modification could be done to improve the properties of Act.-B in the future work.

3 Conclusions

A new approach to treat banana and dragon fruit peels by using ZnCl₂ prior to alkali and acid treatment was conducted in this study. Maximum adsorption capacity at equilibrium was reported as 3.4 mg Cu²⁺/g Act-B and 3.0 mg Cu²⁺/g Act-D, respectively. This result is slightly better compared with the data reported in the literature. It is notable that Act-B performed at least 30% better than Act-D in terms of Cu²⁺ removal. This is due to the presence of hydroxyl and carboxyl groups in Act-B. Adsorption isotherm study showed that the experimental data for Act-B fitted well with Freundlich model, compared with the remaining 6 isotherms. This suggested multilayer adsorption, steam pyrolysis, and surface modification of adsorbent could further enhance the adsorption rate. Lastly, the finding of this study also showed that Act-B can be used to treat Cu²⁺ polluted water up to Malaysia water quality standard.