Biosorption of Cu²⁺ by Eichhornia crassipes: Physicochemical characterization, biosorption modeling and mechanism

2.1 Chemical reagents and preparation of solutions

All the chemical reagents are from Merck and BDH grade. Stock solutions of copper (1000 and 500 mg L⁻¹) were prepared from Cu(NO₃)₂·3H₂O in ultrapure water. A stock NaOH (0.3 M) solution was freshly prepared in deionized water. Stock solutions of 0.3 M and 0.1 M HNO₃ were freshly prepared and standardized by NaOH standard solution (Komy et al., 2006). Periodic standardization for NaOH was made by using standard solution of 0.1 M oxalic acid. All solutions were kept in a refrigerator at 5 °C until measurements were undertaken.

A Milli-Q water purification system was used. All glassware were cleansed for 1 week in 1:1 HNO₃, 1 week in 1:1 HCl and 1 week in ultrapure water to prevent any contaminations (Komy, 1993).

2.2 Sampling and preparation of biomass

The living biomass sample was collected from large floating masses in the River Nile at Sohag city (∼500 km from Cairo), Egypt. The leaves and stems parts of the plant were used in the present study to obtain the dead biomass. For removal of any sand and other trapped debris, the parts were washed with Nile, tap and ultrapure waters. The biomass was then freeze-dried, ground and sieved through 0.25 mm mesh sieve. The produced powder was washed in 0.001 M NaNO₃ and three times in 0.01 M NaNO₃ (Plette et al., 1995) and finally dried.

2.3 Physicochemical characteristics of natural sample

Biomass C, H, N and P analysis were made with a LECO CHN-600 analyzer model CHN-600. Jenway UV/Vis spectrophotometer model 6405 for quantification of total protein and carbohydrate in the biomass by using methods described by Lowery et al. (1951) and Hedge and Hofreiter (1962), respectively.

Surface area and pore size of the biomass were determined by standard BET assumption using N₂-adsorption isotherm using a Nova surface analyzer model 2000 with a method described by Gregg and Sing (1982).

For qualitative/quantitative description of the amino acid content in the biomass, Eppendorf amino acid analyzer model IC-3000 was used. Shimadzu Infrared (FT-IR) spectrophotometer model 470 was used to investigate the functional groups on the biomass.

SEM photographs was used to describe the surface morphology of E. crassipes biomass before and after biosorption of Cu²⁺ while EDAX analysis were used as a primary fast test to identify the capability of biomass to accumulate Cu²⁺ on its surface. Therefore, EDAX analysis was performed for 30 mg of the dry biomass before and after the biosorption experiment (in a solution containing 3.7 mg L⁻¹ of Cu²⁺ and maintained for 150 min of shaking).

2.4 Influence of pH and time on Cu²⁺ biosorption by E. crassipes

For estimation of the influence of pH on Cu²⁺ absorption, 300 mg of dried biomass was mixed with 148 μL of 500 mg L⁻¹ Cu²⁺ (final concentration 3.7 mg L⁻¹) in a set of 10 flasks at different pH’s (2.5–6.0) and completed to a total volume of 20 mL with NaNO₃ (0.1 M). The flasks were then agitated for 3 h at 25 °C to reach equilibrium and centrifuged at 10,000 rpm for 20 min. The resulting supernatants were analyzed for the residual Cu²⁺ by atomic absorption spectrometry using Buck scientific AAS model 210 VGP (USA). The Cu²⁺ uptake at each pH was calculated using the following equation:

For estimation of the effect of equilibrium time on Cu²⁺ absorption, 300 mg of dried biomass was mixed with 148 μL of 500 mg L⁻¹ Cu²⁺ (final concentration 3.7 mg L⁻¹) in a set of six flasks at pH 4.5 and completed to a total volume of 20 mL with NaNO₃ (0.1 M). Afterwards, the flasks were agitated at different time intervals (25–240 min) at 25 °C to attain the sorption equilibrium, centrifuged at 10,000 rpm and finally analyzed for the residual Cu²⁺ AAS as before. Again, the Cu²⁺ uptake at each pH was calculated using Eq. (1) and the overall experiment was repeated three times for precision.

2.5 Isothermal studies

For carrying out the isothermal study, 300 mg dried biomass was mixed with 20 mL NaNO₃ (0.1 N) in 10 flasks at fixed pH 3.5. Then, Cu²⁺ was added increasingly to obtain a range of Cu²⁺ concentration from 2 to 25 mg L⁻¹. The resulting mixtures were stirred for 3 h at 25 °C, centrifuged at 10,000 rpm and analyzed for the residual Cu²⁺ by AAS as before. This experiment was repeated at pH 4.5 and 5.5 for comparison.

Two Langmuir linearization models by Pardo et al. (2003) and Norton et al. (2004) were used to calculate the biosorption parameters at each pH.

Eqs. (2) and (3) have given their authors’ name for simplicity in the further discussion.

2.6 Proton binding and Cu²⁺binding constants

For determining the values of proton bindings (types, amounts and binding constants) of biomass, conductometric and potentiometric titrations were performed using a Jenway conductivity meter model 4320 and a Orion digital pH/mV meter model 701A, respectively. In the conductometric (potentiometric) titration, a suspension of 200 (500 mg) dried biomass in 200 mL NaNO₃ (0.1 N) was titrated against HNO₃ and NaOH solutions. For a good description of the acid–base properties, constant ionic strength (0.1 M NaNO₃) was maintained during the titration (Benedetti et al., 1995).

It should be mentioned that the equivalent point in the conductometric titration is represented by an intersection of two straight lines (James and Parks, 1982). The amount of proton binding (N_T, mol g⁻¹) was calculated by summing up the two equivalent points in the NaOH and HNO₃ titrations. While in the potentiometric titration, the total amount of proton binding (N_T, mol g⁻¹) was calculated from the difference between the total proton amounts in the presence and absence of biomass.

Non-Ideal Competitive Absorption model (NICA), a theoretical model, was used for quantitative description of the protonation behavior of E. crassipes surface:

For estimation of Cu²⁺ binding constants (stability constant of Cu-biomass system), 300 mg dry weight of biomass was mixed with 148 μL of 500 mg L⁻¹ Cu²⁺ (final concentration 3.7 mg L⁻¹) at different pH’s (2–6) and completed to a total volume of 20 mL with 0.1 M NaNO₃. The suspensions were then shaken for 3 h at 25 °C, centrifuged at 10,000 rpm to exclude the biomass and analyzed for the residual Cu²⁺ in the supernatant using AAS technique.

For quantitative description of Cu²⁺ binding constants (K_Mi), a theoretical model (NICA) was used:

3.1 Physicochemical characteristics of biomass

Table 1 shows the results of elemental analysis, total protein, carbohydrates, surface area and pore size for E. crassipes biomass.

The biomass shows a significant content of total protein and carbohydrates as well as the elemental analysis is illustrated in Table 1. This reflected that the biomass tissue has abundant function groups (—COOH, —NH₂, —NH—, —OH, C⚌O, and ${\text{PO}}_4^{-3}$ ) that giving a primary anticipation for the biomass capability to react with the Cu²⁺ through chelation with those sites (Komy et al., 2006).

E. crassipes has a relative high surface area and pore size (4.16 m² g⁻¹ and 35.93 Å) what present an evidence for the great physical contact between the biomass surface and the Cu²⁺ ions in solutions and suggest the entrapment of the Cu²⁺ ions inside those large pores. The former results are comparable to those (4.56 m² g⁻¹ and 1.17 Å) of Pseudomonas biomass in a previous study (Komy et al., 2006).

Fig. 1 describes the scores (%) of amino acids in the biomass. Proline, Glutamic, Aspartic and Leucine acids represent the major (74%) group of amino acids in the biomass while, Histidine, Isoleucine and Methionine are minor (4.12%). The percent of Proline and Isoleucine in the present study was found to be slightly higher than those obtained for the same biomass studied by Ghabbour et al. (2004). This could be attributed to the difference in the biomass habitat where temperature, grazing and nutrients influence the chemical composition of the growing plant from region to another (Weaver, 1946). The above result ensures the abundance of certain chelating centers (—COOH and —NH₂) in the biomass which are capable of capturing Cu²⁺ ions from aqueous solutions.

Close

Figure 1

Percentages of amino acids in E. crassipes biomass (g individual amino acid/g total amino acids).

Export to PPT

Figs. 2a and b describe the FT-IR spectra before and after Cu²⁺ (3.7 mg L⁻¹) biosorption, respectively. Fig. 2a displays a number of absorption peaks indicating the complex nature of the examined biomass as follows; 3421.2 cm⁻¹ (bonded, —OH and —NH), 2926.4 cm⁻¹ and 600.9 cm⁻¹ (C—H), 1252 cm⁻¹ (P⚌O phosphonate or phosphoramide), 1061.9 cm⁻¹ (C—O of aliphatic alcohol) and 1635.8 cm⁻¹ (strong, asymmetric stretching of R—COO⁻, aromatic C⚌C, C⚌O in R—CHO, C⚌O of quinones or in conjugation with alkenes) (Yang et al., 2010). Fig. 2b shows a remarkable shift in FT-IR bands for some ionizable functional groups after the biosorption of Cu²⁺. Generally, all of the absorption bands (stretching) after biosorption of Cu²⁺ were shifted from their original positions. There are three obvious shifts (>15 cm⁻¹) in the absorption bands of O—H bending from 1384.1 to 1402.4 cm⁻¹, P⚌O stretching from 1252 to 1274 cm⁻¹ and in the absorption band of O—C stretching from 1061.9 to 1078.3 cm⁻¹. This result is an indication of the biosorption process. Similar results of –OH and –C⚌O groups were obtained in a previous study on the same biomass collected from China (Zhou et al., 2011). Moreover, a previous study on Cr⁶⁺ biosorption by E. crassipes, collected from India, revealed a shift in the absorption band of —OH group after biosorption process (Mohanty et al., 2006). This suggests that —OH group is a main constituent in E. crassipes samples around the world and acts as a major chelating center in the plant material.

Close

Figure 2

Infrared spectrum of dry E. crassipes biomass (a: before biosorption and b: after biosorption of 3.7 mg L⁻¹ Cu²⁺ by 300 mg dry biomass).

Export to PPT

SEM photographs in Fig. 3a and b are describing the surface morphology of E. crassipes biomass before and after biosorption of Cu²⁺, respectively. Fig. 3a indicates that the untreated biomass is characterized with a hollow structure. A similar SEM result was observed in Brazil by Schneider et al. (1995). Fig. 3b shows that Cu²⁺ ions possess sphere and needle shapes with domination of the latter and uniformly distributed over the surface of biomass. Moreover, the uptake of Cu²⁺ in the pores was appeared to be higher than the in rest of biomass surface (Fig. 3b, intrinsic image).

Close

Figure 3

Scanning electron microscope images of E. crassipes biomass (a: before biosorption and b: after biosorption of 3.7 mg L⁻¹ Cu²⁺ by 300 mg dry biomass).

Export to PPT

According to EDAX analysis result, Table 1, shows that the biomass contains a natural amount of Cu representing 0.7% of its dry weight and it should be absorbed during the plant’s life (Omanayi et al., 2011). This quite higher value of naturally adsorbed Cu might corresponds to the high contamination of Nile water with Cu²⁺. After biosorption process, Cu²⁺ amount in the biomass surface increased to 1.54% that is mainly attributed to the biosorption process (Iqbal et al., 2009). These results match with the FT-IR results, which exhibit significant shifts in the bands of —OH, —P⚌O and C⚌O groups due to Cu²⁺ complexation.

Accordingly, the above results of biomass characteristics suggest the heterogeneity of biomass surface and its variable chemical content with the presence of mainly three (—OH, COO⁻, and —P⚌O) acidic sites on the biomass responsible for the Cu²⁺ biosorption.

3.2 Influence of pH and time on Cu²⁺ biosorption by E. crassipes

3.2.1

3.2.1 Influence of pH

Fig. 4a illustrates the influence of pH (2.5–6.0) on the absorption of Cu²⁺ onto surface of E. crassipes. It was observed that with increasing the pH from 2.5 to 3.25, a rapid increase in the Cu²⁺ uptake by the biomass took place. This is probably due to additional active sites are introduced to copper biosorption within this pH range. On increasing the pH from 3.25 to 5.5, a slight decrease in the Cu²⁺ uptake was observed. At pH > 5.5, the absorption of Cu²⁺ decreases rapidly due to the formation of hydroxylated complexes of Cu²⁺. This finding suggests that the amount of Cu²⁺ uptake is dependent on the pH value. As a result, the optimum pH range for Cu²⁺ biosorption is 3.25–5.5. Likely, Schneider et al., (1995) stated that optimum pH range of Cu²⁺ uptake by E. crassipes biomass was between 5.0 and 6.6. This variation might be attributed to the difference in the chemical composition of the two biomasses. Thus, the pH dependence of Cu²⁺ uptake is related to solution chemistry and functional groups of the biomass. In this respect, Sheng et al., (2004) indicated that all Cu²⁺ species in all biosystems exist in the ionic form at pH < 4.0. Subsequently, the increase in Cu²⁺ uptake by the biomass in the present study at pH 2.5–3.25 cannot be explained on the base of change in metal speciation, as hydrogen ions will compete with Cu²⁺ for the active sites of cell wall of biomass.

Close

Figure 4

Influence of (a: pH and b: contact time) on Cu²⁺ biosorption in a system composed from 300 mg biomass, 3.7 mg L⁻¹ Cu²⁺ and 20 mL of 0.1 M NaNO₃.

Export to PPT

3.3 Biosorption isothermal study for Cu²⁺ – E. crassipes system

Maximum absorption capacity (q_max) and Langmuir constants (b) and (K_f) were evaluated at pH 3.5, 4.5 and 5.5 by applying Pardo-model and Norton-model using Eqs. (2) and (3), respectively. Results of their values are shown in Table 2.

Generally, good fitting between the experimental and theoretical data was obtained on applying the two models. This indicates that the two models are adequate to describe the behavior of Cu²⁺ biosorption by E. crassipes biomass where the values of R² (correlation coefficient) exceed 94% in the two models (Pardo et al., 2003). R² coefficient gives the amount of variance explained by the model, so it can be used to evaluate the goodness of fitting at different pH values (Pardo et al., 2003). Moreover, a good Fitting by Pardo-model was noticed at the three pH’s with the best one observed at at pH 4.5 with highest R² values (Table 2).

Fig. 5a–c shows the linear fitting at pH (3.5, 4.5 and 5.5) against the concentration of Cu²⁺ using Norton-model. Similarly, the value of R² in Table 2 indicated that the best fit was at pH 4.5 in Norton-model. This result matches well with that obtained by Pardo-model at pH 4.5 suggesting that at such pH the maximum absorption was fulfilled. Thus, based on Norton-model, the linear fittings for copper-biomass system, at the three pH’s, can be represented as follows:

Close

Figure 5

Linearized biosorption isotherms for Cu²⁺-E. crassipes system by using Norton-model at (a: pH 3.5, b: pH 4.5 and c: pH 5.5).

Export to PPT

The values of q_max, b and K_f parameters were used to evaluate the biosorption isotherm in this work. These data are compared with E. crassipes biomass collected from Brazil (Schneider et al., 1995) and India (Dave et al., 2009). It is found that K_f value in the present study (0.39) is higher 1.62 times than that of Brazilian sample (0.24), whereas it is lower 16 times than that of Indian sample (6.05) (Table 2). Similarly, the b value in the present study is lower 0.55 times than of India-sample (0.029) (Dave et al., 2009). Likewise, q_max in the present study obtained by Norton-model (24.89) is very close to that obtained by the Brazilian sample (23.1) (Schneider et al., 1995), while q_max obtained by Pardo-model (27.7) is 1.2 times lower than that obtained from the Indian sample (33.4) (Dave et al., 2009). The inconsistency in the values of the q_max, b and K_f, calculated for Cu²⁺ biosorption by the same biomass species from different locations, may be attributed to the variation in their habitat which alter the chemical composition and physical characteristics of the biomass and hence the ability and mechanism of biosorption (Omanayi et al., 2011).

3.4 Acid–base properties and proton bindings on E. crassipes

A recent study by Mane et al. (2011) on the same biomass showed that the higher acidic content of the plant is directly connected to its metal binding functions. Accordingly, the importance of estimating the types, amounts and binding constants of acidic sites on the biomass has been aroused to discuss the biosorption mechanism of Cu²⁺ by E. crassipes.

Figs. 6a–c represent the acid–base titrations of E. crassipes biomass by increasing the pH from 2 to 11.5 using potentiometric and conductometric titrations. The dotted curve in Fig. 6a shows the potentiometric titration for the biomass suspension system; a relation between X_Hexp, the equilibrium concentrations of total protons in mol g⁻¹ (Y-axis), vs. pH values (X-axis). Noticeably, the inflection in the dotted curve (X_Hexp) is broad and poorly defined to describe the protonation properties as it is highly affected by the amount of protons on the biomass.

Close

Figure 6

(a) Potentiometric, (b and c) conductometric titration curves for proton bindings and (d) NLLSR fitting for Cu²⁺ biosorption by E. crassipes biomass at different pH’s. Analogous to a(d) curves: the dotted X_Hexp(X_Mexp) and solid X_Htheor(X_Mtheor) lines represent the experimental and theoretical data, respectively.

Export to PPT

This result ensures the diversity of binding sites on the tested biomass and leads us to use a theoretical model (NICA), which is reported by Plette et al. (1996). The results of biomass characterization awakened the existence of mainly three active sites (—OH, —COO⁻ and $\text{—}{\text{PO}}_4^{-3}$ ) describing the biosorption of Cu²⁺ by the biomass. Therefore, the theoretical (NICA) fitting will be performed on the assumption of the presence of three acidic sites. More description on the theoretical model was described elsewhere (Komy et al., 2006; Seki and Suzuki, 2002; Komy, 2004). Thus, Eq. (4) can be rewritten as follows with considering the three sites:

To evaluate these constants, a NICA model was applied using the experimental X_Hexp in Eq. (9). Practically, the experimental X_Hexp (dotted curve) was fitted with the theoretical X_Htheor (solid line) in Fig. 6a using Microsoft excel 2003 (solver add-in). As could be observed from Fig. 6a, there is a great fitting between the theoretical and experimental values at each pH (R² = 0.997). The results of the six parameters (N₁, N₂, N₃, K_H1, K_H2 and K_H3) are listed in Table 3. The total number of acidic sites (N_T = N₁ + N₂ + N₃) was found to be N_T = 1.65 × 10⁻² mol g⁻¹.

To confirm the last result (N_T), the total amount of acidic sites was re-measured by conductometric titration for the same biomass system, as shown in Fig. 6b and c. The total amount of acidic sites conductometrically was N_T = 1.69 × 10⁻² mol g⁻¹ (calculated as mentioned in Section 2.5). There is high agreement between the two N_T values with a relative error of 2.39%.

No reported data on the acidic sites of E. crassipes was found, so a comparison between the present data of E. crassipes with that of Pseudomonas aeruginosa (Komy et al., 2006) and Cumin (Komy, 2004) species was made. Values of N₂ = 1.16 × 10⁻² mol g⁻¹ and pK_H2 = 1.94 in the present study are analogous to P. aeruginosa (N₂ = 1.21 × 10⁻² mol g⁻¹ and pK_H2 = 1.92) (Komy et al., 2006) but quite similar to Cumin (N_1,2 = 7.87 × 10⁻³ mol g⁻¹ and pK_H1,2 = 1.88) (Komy, 2004). Accordingly, this acidic site is mainly found in all of E. crassipes, P. aeruginosa and Cumin. The values of pK_H are identical to Glutamic (αpK_COO− = 2.2) and Aspartic (αpK_COO− = 2.1) acids (Solomons, 1994) which are amino acids characterizing all the three biomasses.

N₁ and N₃ in the present study are of lower values compared to those of P. aeruginosa (N₁ = 2.16 × 10⁻² and N₃ = 6.87 × 10⁻³ mol g⁻¹), while pK_H1 and pK_H3 are so close to the pK_H1 = 1.66 and pK_H3 = 2.16 of P. aeruginosa biomass (Komy et al., 2006). In addition, Proline and $\text{—}{\text{PO}}_4^{3-}$ have pK_H of 2.0 (Solomons, 1994) and 2.15 (Harvey, 1956), respectively, indicating that the other acidic sites (1 and 3) correspond to the Proline and phosphate centers which are found in both E. crassipes and P. aeruginosa (Komy et al., 2006).

The difference between N_i values in the present study and the other two biomasses may be related to: (i) variation in percentage of Proline, Glutamic, Apartic and Leucine acids from one species to another, (ii) the protein series on E. crassipes may partially hydrolyzed to poly peptides (i.e. more free sites are produced), (iii) the presence of an additional ( ${\text{PO}}_4^{3-}$ ) site in E. crassipes and (iv) the low surface area (4.18 m² g⁻¹) of biomass. Finally, results in (Table 3) show that the values of pK_Hi of E. crassipes biomass are close to each other (1.8, 1.9 and 2.0) but lower than the standard values of pK_COOH corresponding to the major amino acids (Proline = 2.0, Glutamic = 2.2, Aspartic = 2.1 and Leucine = 2.3) (Solomons, 1994) and pKa ( ${\text{PO}}_4^{3-}=2.1$ ) (Harvey, 1956) in E. crassipes biomass.

3.5 Binding constants of Cu²⁺

A Non-Linear Least Squares Regression (NLLSR) coupled with NICA model were used to evaluate Cu²⁺ binding constants (K_M1, K_M2 and K_M3) with the three acidic sites. More description and details on the NICA model were described elsewhere (Komy et al., 2006; Seki and Suzuki, 2002; Komy, 2004). Shortly, the NLLSR method with NICA model was applied with the following two assumptions: (i) the biomass contains three main acidic sites (—OH, —COOH and ${\text{PO}}_4^{3-}$ ) responsible for the biosorption of Cu²⁺, (ii) the biosorption process corresponds to monodentate binding sites. So, Eq. (5), can be rewritten with considering the values of (N₁, N₂, N₃, K_H2, K_H2 and K_H3), as follows:

Fig. 6d illustrates the non-linear fitting between the X_Mexp (dotted line) and X_Mtheor (solid line) vs. the pH. A distinctive fitting between the theoretical and experimental lines was obtained (Y-residual = 6.88 × 10⁻¹⁵ and R² = 0.997), Fig. 6d.

The results in Table 3 indicate that pK_M1, pK_M2 and pK_M3 values are almost identical suggesting —OH, —COOH and ${\text{PO}}_4^{3-}$ sites to be the major acidic sites to bind with Cu²⁺. In addition, the NLLSR method with NICA can be applied successfully for determining the Cu²⁺ binding constants.

Taking into consideration the complexity of the chemical composition of the E. crassipes, several mechanisms (ion exchange, complexation, coordination and microprecipitation) can occur at the same time, depending on the aqueous environment (Sheng et al., 2004). In view of that, the interaction between Cu²⁺ and the E. crassipes biomass can be suggested as follows: