*3.1.4 Thermogravimetric analysis (TGA)*

Thermogravimetric makes it possible to follow, as a function of temperature, the weight loss evolution of each sample, mainly caused by dehydration or/and by the decomposition of the organic matter it contains. The thermal stability of BEC-ED comparing to BEC is studied basing on the TGA thermograms shown in **Figure 6**. The thermal behavior of BEC shows two stages of thermal decomposition, the first one is observed between 35 and 250°C attributed to the solvents and adsorbed water vaporization [79]. The strong weight loss (70%) corresponds to the degradation of the grafted entities and the cellulose backbone is observed in the temperature range of 250–600°C. For BEC-ED, no thermal event was observed below 250°C, indicating the absence of traces of solvents. The thermal decomposition of BEC-ED is noticed from 250°C up to 450°C with a mass loss of 90%. In addition, a low degradation of the thermal stability of BEC-ED compared to BEC has been noted,

*New Ethylenediamine Crosslinked 2D-Cellulose Adsorbent for Nanoencapsulation Removal… DOI: http://dx.doi.org/10.5772/intechopen.98709*

**Figure 6.** *TGA thermogram profiles of BEC and BEC-ED.*


**Table 1.**

*Coded and actual variables and their levels.*

and this is possibly due to the decrease of hydrogen interaction density and their replacement by hydrophobic interactions, and the supramolecular separation of the polymer chains caused by grafted ethylenediamine (ED).

#### **3.2 Response surface methodology (RSM) modeling procedure**

In the current work, is focused on *3-level Box–Behnken design* (BBD) in response surface methodology for seeking the optimal conditions for the removal efficiency of Cu(II) and Pb(II) onto BEC-ED. The three variables affecting the current process are pH at 4.5, 6.0, and 7.5, contact time at 5, 17.5, and 30 min, and adsorbent amount at 10, 20 and 30 mg. The complete design consisted of three different levels (1, 0, and + 1) and 3-variable (pH—X1, Adsorbent amount —X2, and contact time —X3). The layout of the factorial design is shown in **Table 1**. A total of 17 experiments were used in this study to evaluate the effects of the three input variables on Pb(II) and Cu(II) removal efficiency. The full picture of experiments with their responses (Pb(II) and Cu(II) removals) are tabulated in **Tables 2** and **3**, respectively.

The analysis of variance (ANOVA) was applied to the experimental runs, and then the results of the Box–Behnken design table are calculated and fitted by a suitable polynomial equation. According to the model's evaluation in **Tables 4** and **5**, which focuses on maximum R2 , predicted R<sup>2</sup> , and adjusted R2 , the quadratic polynomial (Eq. (1)) model was chosen and well-fitted for all three independent parameters and responses (Cu(II) and Pb(II) removal efficiency).


#### **Table 2.**

*The BBD matrix design with three independent factors and the corresponding experimental results.*


#### **Table 3.**

*ANOVA analyses of the quadratic model and determination coefficients for Pb(II) adsorption efficiency.*

*New Ethylenediamine Crosslinked 2D-Cellulose Adsorbent for Nanoencapsulation Removal… DOI: http://dx.doi.org/10.5772/intechopen.98709*


#### **Table 4.**

*Model summary statistics Pb(II).*


#### **Table 5.** *Model summary statistics Cu(II).*

Therefore, the predictive polynomial quadratic response model can be described as the following equation (Eq. (1)) [80]:

$$\mathbf{Y} = \mathfrak{g}\_0 + \sum\_{\mathbf{i}=1}^{\mathrm{n}} \mathfrak{f}\_{\mathbf{i}} \mathbf{X}\_{\mathbf{i}} + \sum\_{\mathbf{i}=1}^{\mathrm{n}} \mathfrak{f}\_{\mathbf{i}\mathbf{i}} \mathbf{X}\_{\mathbf{i}}^2 + \sum\_{\mathbf{i}=1}^{\mathrm{n}} \sum\_{\mathbf{i}>1}^{\mathrm{n}} \mathfrak{f}\_{\mathbf{j}} \mathbf{X}\_{\mathbf{i}} \mathbf{X}\_{\mathbf{j}} \tag{1}$$

Where Y is the predicted response, β0 and βi are the intercept coefficient, and the linear coefficient respectively, βii and βij are the quadratic and the interaction coefficients, respectively, while Xi and Xj represent the coded values of the independent variables.

An ANOVA analysis for Cu (II) and Pb (II) removals was performed, and the results are presented in **Tables 3** and **6**, respectively. According to ANOVA analysis, the results obtained showed that the F and P-values less than 1000 and 0.0500, respectively. This confirmed that the model terms are significant. While **Lack of Fit F-value** in the ANOVA tables introduces an insignificant error with regard to the pure error. The response for Cu(II) and Pb(II) removal efficiency was determined with real factors by the following expressions (Eqs. (2) and (3)):

$$Pb(II)\;Removal = -480.655 + 154.152 \ast pH + 6.49305 \ast Amount + 3.56334 \ast Time$$
  $+0.0508333 \ast pH \ast Amount + -0.0221333 \ast pH \ast Time + 0.00434 \ast Amount \ast Time$   $+-12.9344 \ast pH^2 + -0.14465 \ast Asborbent \; amount^2 + -0.074704 \ast Time^2$ 

$$\begin{aligned} \text{Cu} (\text{II}) &\text{Rnoval} = -462.971 + 146.053 \ast pH + 7.18408 \ast Amount + 3.02441 \ast Time \\ &+ 0.00816667 \ast pH \ast Amount + -0.1164 \ast pH \ast Time + 0.011 \ast Amount \ast Time + \\ &- 11.9436 \ast pH^2 + -0.157755 \ast Amount^2 + -0.0532832 \ast Time^2 \end{aligned}$$

(3)

Statistical diagnostics test is an excellent and effective tool for confirming the model presented. These diagnostic plots are given in **Figure 7**. By classifying the


#### **Figure 7.**

*Diagnostic plots for adsorption of PbII): Probability plot for the studentized residuals (a), comparison between actual and predicted values (b), plot of the externally studentized residuals vs. experimental run number (c), diagnostic plots for adsorption of Cu(II): Probability plot for the studentized residuals (d), comparison between actual and predicted values (e), plot of the externally studentized residuals vs. experimental run number (f).*

proportion of normal probability in terms of residuals, it can be observed that the datum-points are approximately straight-line (**Figure 7a** and **d**). Into the other diagnostic plots, the actual responses were compared to their residuals based on

#### *New Ethylenediamine Crosslinked 2D-Cellulose Adsorbent for Nanoencapsulation Removal… DOI: http://dx.doi.org/10.5772/intechopen.98709*

predicted responses, suggesting that the quadratic model was required to predict removal efficiency in the experimental parameters (**Figure 7b** and **d**). In addition, as shown in the plot (**Figure 7c** and **f**) the data showed a good homogeneity. In **Figure 7**, the dispersion of the residuals is dispersed randomly about 5, confirming that the results are coherent with the model. In the other diagnostic plots, actual responses were compared to their residues based on predicted responses, implying that the quadratic model was necessary to predict removal efficiency in experimental parameters.

By employing the RSM method, the evaluated models (Eqs. (2) and (3)) are used to design the 3-D graphs and find the optimal conditions for Pb (II) and Cu(II) removal efficiency. It can be seen from **Figures 8** and **9** that the retention of Pb(II) and Cu(II) ions onto BEC-ED increases with increases of the pH solution. The removal efficiency reached a maximum of around 6. When the pH is higher than 6 or lower than 5 the adsorption decreased rapidly. This could be explained by that in the acidic environment, the active groups responsible for the adsorption process exist mainly in the NH3+ form, and they prevent the retention of Pb (II) and Cu(I) ions on the amino groups of BEC. When the pH increases from 2 to 5, the active sites of the chelator become in the form of free NH2 amines, which facilitate chelation on Pb(II) [60]. In addition, at high pH, the formation of lead and copper hydroxides (**Figure 10**) limits their adsorption on the BEC-ED surface, and as shown in **Figures 8** and **9**, at high pHs, the removal efficiency of Pb(II) and Cu(II) ions is significantly diminished. Contact time was also examined and as given in **Figures 8** and **9**. The results of the retention of Pb(II) and Cu(II) onto BEC-ED revealed that the maximum adsorption equilibrium can be achieved rapidly around 16 min.

In conclusion, based on the desirability optimization with three factors, the best removal efficiency was 99.52% and 97.5% for Pb(II) and Cu(II), respectively and was obtained at pH: 5.94, adsorbent amount: 22.2 mg, and contact time: 21.53 min (**Figure 11**).

#### **Figure 8.**

*(a, b and c) 3D response surface plot and, (d, e and f) contour plot for the effect of factors on the Pb(II) removal efficiency.*

**Figure 9.**

*(a, b, and c) 3D response surface plot and, (d, e and f) contour plot for the effect of factors on the Cu(II) removal efficiency.*

**Figure 10.** *Speciation diagram of lead (a) and copper (b) as a function of pH in ultrapure water, determined by the hydra/medusa program [81].*

#### **3.3 Equilibrium isotherms**

Adsorption isotherms for lead and copper were made by carrying out batch adsorption studies. Lead and copper adsorption was studied onto BEC-ED in a large concentration range (from 15 to 250 mg/L), to better model the retention mechanisms. The adsorption experiments were performed at room temperature by using a mass of adsorbent 22.2 mg with 50 mL of the aqueous solution, at pH 5.94 and contact time 21.53 min. The quantity of the lead and copper ions adsorbed onto the BEC-ED at equilibrium, qe (mg/g), and the adsorption percentage was calculated by the following Equations [82]:

$$q\_{\epsilon} = \frac{(C\_0 - C\_{\epsilon})}{m} \times V \tag{4}$$

*New Ethylenediamine Crosslinked 2D-Cellulose Adsorbent for Nanoencapsulation Removal… DOI: http://dx.doi.org/10.5772/intechopen.98709*

**Figure 11.**

*Desirability approach function optimization for Pb(II) and Cu(II) in terms of removal efficiency (%) and desirability.*

$$\text{\%Adsorption} = \left(\mathbf{1} - \frac{\mathbf{C}\_{\varepsilon}}{\mathbf{C}\_{0}}\right) \times \mathbf{100} \tag{5}$$

Where, C0 and Ce are the metal ion initial concentration and concentration at equilibrium (mg/L), respectively. V is the volume of solution (L) and m is the adsorbent amount (g).

The adsorption isotherms of Cu(II) and Pb(II) on BEC-ED were modeled using the Freundlich (Eq. (6)) [83] and Langmuir (Eq. (7)) [84] models equations:

$$q\_{\epsilon} = K\_F C\_{\epsilon}^{1/n} \tag{6}$$

Where, n and KF are Freundlich constants represent the heterogeneity index, and the adsorption coefficient, respectively.

$$q\_{\epsilon} = \frac{q\_{m}K\_{L}C\_{\epsilon}}{1 + K\_{L}q\_{m}}\tag{7}$$

Where, KL (L/mg) and qm (mg/g) are the Langmuir constant and the maximum adsorption capacity, respectively.

The equilibrium isotherm obtained for lead and copper adsorption on BEC-ED is shown in **Figure 12**. Lead adsorption was greater than that of copper 43.85 mg/g. This could be explained by the higher reactivity of lead than copper, which can have

**Figure 12.** *Equilibrium isotherms for Pb and Cu adsorption on BEC-ED (solid lines and dash-dotted represent Langmuir and Freundlich fitting, respectively).*

stronger interactions with the lone pairs of electrons of the nitrogen atoms of amino groups than that of the Cu, as previously demonstrated elsewhere [63, 85]. Therefore, the Pb(II) ions can rapidly form a stable complex with -NH2 groups on the surface of the BEC- ED.

The adsorption isotherm was modeled both by Freundlich (Eq. (6)) and Langmuir (Eq. (7)) models. The results revealed that the equilibrium isotherms data (**Figure 12** and **Table 7**) correlated better with the Langmuir model with a maximum adsorption capacity estimated at 50.76 mg/g and 39.68 mg/g for Pb(II) and Cu (II), respectively. This implies that the BEC-ED surface is homogeneous, which indicates that lead and copper ions adsorption follows monolayer adsorption.

To better understand the retention mechanisms of copper and lead adsorption on BEC-ED, the effect of lead and copper ions adsorption on the morphology of BEC-ED was monitored by SEM analysis. **Figure 13** shows the possible interactions between BEC-ED and the Pb(II) and Cu (II) ions, as well as the proposed mechanism. SEM pictures showed distortion in the morphology of BEC-ED under the adsorption forces of Pb(II) and Cu(II) ions. However, this effect is probably caused by the interactions between the metal ions and the donor sites of the grafted groups (ED), where the internal compression of the laminated structure has caused a very remarkable separation of the polymeric layers. Indeed, the approximate calculation


**Table 7.**

*Freundlich and Langmuir constants for lead and copper adsorption on BEC-ED.*

*New Ethylenediamine Crosslinked 2D-Cellulose Adsorbent for Nanoencapsulation Removal… DOI: http://dx.doi.org/10.5772/intechopen.98709*

**Figure 13.** *Impact of the adsorption of lead and copper ions on the morphology of BEC-ED.*

carried out to compare the inter-chain distances showed a decrease in the latter. The results obtained showed the capability of BEC-ED for adsorbing Cu(II) via metal interactions with NH2 groups of ethylenediamine [56, 60], resulting in the decrease in the inter-layer distance from 11.50 Å to 07.34 Å for copper and 08.27 Å for lead.

For the purpose to assess the potential retention of lead and copper ions retention provided by BEC-ED compared to other adsorbents, the results achieved through this study were compared with the adsorption abilities of some conventional natural and synthetic cellulose in the literature (**Table 8**). It has been found that the lead and copper retention capacity of BEC-ED is among the higher results. Therefore, considering the retention capabilities of other adsorbents, accessibility, environment friendly biomaterial, and low cost, it may be concluded that the BEC-ED adsorbent demonstrated its ability to efficiently eliminate lead and copper ions in simple media.


#### **Table 8.**

*Comparison of lead and copper adsorption capacity of BEC-ED with conventional natural and synthetic cellulose adsorbents.*
