Application Of Response Surface Methodology And Artificial Neural Network In Modelling The Adsorption Of Chromium (VI) And Lead (II) Using Corn Cob As Biosorbent

ABSTRACT

This study aimed at optimizing the adsorption of chromium and lead with corn cob using Response Surface Methodology (RSM) and Artificial Neural Network (ANN). The treated and untreated corn cobs were characterized using Fourier Transform Infra-Red (FTIR), Scanning Electronic Microscope (SEM) and X-Ray Diffraction (XRD). The experiment was designed with Box-Behnken experimental design of RSM in Design Expert version 10, with four factors which generated 29 experimental runs in order to optimize the removal of lead (II) and chromium (VI) ions from waste water. The effect of the process parameters; pH (2 to 10), contact time (10 to 70 minutes), adsorbent dosage (0.1 to 1.0g) and initial concentration (20 to 100 mg/L) on the removal efficiency of the chromium and lead were evaluated using RSM. The 29 experimental data was divided into subsets, training (70%), validation (15%) and test (15%) for the Artificial Neural Network modelling in MATLAB 2019 environment and was quantified using Mean Square Error (MSE) and correlation coefficient (R₂). The experimental result showed maximum lead removal efficiency of 0.415 at concentration of 20 mg/L, pH of 6, dosage of 0.55 g at 10 minutes and 0.935 at concentration of 100 mg/L, pH of 6, dosage of 1 g at 10 minutes were obtained for the treated and untreated corn cob respectively. The maximum chromium removal efficiency of 0.695 at concentration of 60 mg/L, pH of 6, dosage of 0.55 g at 40 minutes and 0.875 at concentration of 100 mg/L, pH of 6, dosage of 1 g at 40 minutes were obtained for the untreated and treated corn cob respectively. The multiple regression for the removal of lead (Pb) using untreated and treated corn cob resulted to 2FI and quadratic model with R-squared values of 0.0.734702 and 0.72372 respectively, while the removal of chromium using untreated and treated corn cob gave quadratic models with R-squared values of 0.840365 and 0.812564 respectively. The Cr (VI) and Pb (II) linearity and high values of correlation coefficients (R²) of 0.9944 and 0.9901 for Langmuir model, 0.9604 and 0.9753 for Freundlich isotherm and 0.8593 and 0.8034 for Temkin isotherm respectively. The Levenberg–Marquardt algorithm ANN models for predicting the removal efficiency of lead and chromium using the treated corn cob generated R² value of 0. 999999 and 0.999999, compared to the RSM model with R² value of 0.734702 and 0.840365 while the untreated corn cob has ANN R² value of 0.999999 and 0.999999, compared to the RSM model with R² value of 0.72372 and 0.812564. The comparison of the ANN and RSM models showed ANN to be a better predictor than RSM.

 

 TABLE OF CONTENTS

Cover page

Title page                                                                                                                                i

Declaration                                                                                                                              ii

Certification                                                                                                                            iii

Dedication                                                                                                                              iv

Acknowledgements                                                                                                                v

Table of contents                                                                                                                    vi

List of tables                                                                                                                           viii

List of figures                                                                                                                                     ix

Abstract                                                                                                                                  xi

 

CHAPTER 1

INTRODUCTION

1.1       Background of the study                                                                                            1

1.2       Statement of problem                                                                                                 5

1.3       Objectives of the study                                                                                               6

1.4       Justification of the study                                                                                            6

1.5       Scope of the study and its limitations                                                                                    7

 

CHAPTER 2

LITERATURE REVIEW

2.1       Heavy metals                                                                                                              8

2.1.1    Chromium (VI)                                                                                                           8

2.1.2    Lead                                                                                                                            9

2.2       Adsorption                                                                                                                  10

2.2.1    Mechanism of Adsorption                                                                                          12

2.2.2    Types of Adsorption                                                                                                   13

2.2.2.1 Physical adsorption                                                                                                     14

2.2.2.2 Chemical adsorption                                                                                                   14

2.2.3    Factors affecting adsorption                                                                                       15

2.2.3.1 Nature of adsorbent                                                                                                    15

2.2.3.2 Nature of solute (adsorbate)                                                                                       16

2.2.3.3 Nature of solvent                                                                                                        17

2.2.3.4 Influence of temperature                                                                                            17

2.2.3.5 Influence of pH                                                                                                          18

2.2.3.6 Contact time or residence time                                                                                   18

2.3       Agricultural adsorbents                                                                                               19

2.3.1    Corn Cob                                                                                                                    21

2.4       Response Surface Methodology                                                                                 23

2.5       Artificial Neural Network                                                                                           27

2.5.1    Learning Process                                                                                                         28

2.5.2    Generalization                                                                                                             29

2.5.3    Selecting the number of Hidden Layers                                                                     30

2.5.4    Pre-Process and Post-Process of the Training Patterns                                               31

2.6       Gap in the literature                                                                                                    33

 

 

CHAPTER THREE

MATERIALS AND METHOD

3.1       Materials/equipment                                                                                                    34

3.2       Preparation of adsorbent                                                                                             35

3.2.1    Carbonization                                                                                                              35

3.2.2    Impregnation                                                                                                               35

3.3       Preparation of chromium stock solution/working solution                                         36

3.4       Batch adsorption experiment                                                                                      36

3.4.1    Effect of pH                                                                                                               38

3.4.2    Effect of initial concentration                                                                                     38

3.4.3    Effect of adsorbent dosage                                                                                        39

3.5       Design of experiment                                                                                                  39

3.6       Adsorption isotherm model                                                                                        40

3.7       Adsorption kinetics                                                                                                     42

3.8       ANN model                                                                                                                43

 

CHAPTER 4

RESULT AND DISCUSSION

4.1       Characterization of the adsorbent                                                                               45

4.1.1    Fourier Transform Infra-Red                                                                                      45

4.1.2    X-ray Diffraction                                                                                                        48

4.1.3    Scanning Electron Microscopy (SEM)                                                                       49   

4.2       Response surface methodology model for removal efficiency                                   50

4.2.1    ANOVA for % Lead Removal Using Untreated Corn adsorbent                             52

4.2.1.1 Effect of process parameters on the removal of lead using untreated corncob          53

4.2.2    ANOVA for% Lead Removal Using Treated Corn Cob                                           56

4.2.2.1 Effect of process parameters on the removal of lead using treated corncob              57

4.2.3    ANOVA for % Chromium Removal Using Untreated Corn Cob                             60

4.2.3.1 Effect of process parameters on the %removal of chromium using untreated corncob61

4.2.4    ANOVA for % Chromium Removal Using Treated Corn Cob                                 64

4.2.4.1 Effect of process parameters on the %removal of chromium using treated corncob65

4.3       Artificial Neural Network (ANN) model                                                                   68

4.4       Comparison of Response Surface Methodology and Artificial Neural Network       72

4.5       Batch adsorption result                                                                                               73

4.5.1    Effect of contact time on removal efficiency                                                                         73

4.5.2    Effect of temperature on removal efficiency                                                              74

4.5.3    Effect of initial concentration on removal efficiency                                                 75

4.5.4    Effect of pH on removal efficiency                                                                            76

4.5.5    Effect of adsorbent dosage on removal efficiency                                                     77

4.6       Equilibrium adsorption isotherms                                                                               78

4.7       Adsorption Kinetics                                                                                                    81

 

CHAPTER FIVE

CONCLUSION AND RECOMMENDATION

5.1       Conclusion                                                                                                                  84

5.2       Recommendations                                                                                                      84

5.3       Contribution to knowledge                                                                                         85

 

REFERENCES

 

LIST OF TABLES

Table                                                  Title                                                                       Page

3.1:      List of equipment                                                                                                        34

3.2:      Experimental ranges of the factors for Box Behnken Design                                                39

3.3       Adsorption Isotherm models                                                                                      41

4.1:      FTIR result for the untreated and treated corn cob                                                    46

4.2:      Experimental design for removal efficiency                                                               51

4.3:      Analysis of Variance for the removal of lead using untreated corn cob                    53

4.4:      Analysis of Variance for the removal of lead using treated corn cob                                    57

4.5:      Analysis of variance for the removal of chromium using untreated corn cob            61

4.6:      Analysis of Variance for the removal of Chromium using treated corn cob              65

4.7:      ANN comparison of 11 backpropagation (BP) algorithms for removal efficiency

using untreated and treated corn cob                                                                                     69

 

4.8:      Comparison of the correlation coefficients of RSM and ANN models                     72

 

 

 

 

 

 

 

 

 

LIST OF FIGURES

Figure                                                 Title                                                                Page

2.1:      Adsorption mechanism                                                                                               12

2.2:      Pathways of adsorption process                                                                                 13

2.3:      Training platform of ANN                                                                                          28

2.4:      Result of ANN training platform                                                                              30

2.5:      ANN platform showing hidden neurons                                                                    31

3.1:      Scheme of adsorption of pollutants by the batch process                                          38

3.2:      Artificial Neural Network for the adsorption process                                                            44

4.1:      FTIR result for untreated corn cob                                                                             47

4.2:      FTIR of treated corn cob                                                                                            47

4.3:      XRD for untreated corn cob                                                                                       48

4.4:      XRD for treated corn cob                                                                                           49

4.5:a)   SEM for untreated corn cob b) SEM for Treated corn cob                                        50

4.6:      Effect of concentration and pH on % removal of lead using untreated corncob       54

4.7:      Effect of concentration and dosage on % removal of lead using untreated corncob 54

4.8:      Effect of concentration and time on % removal of lead using untreated corncob     54

4.9:      Effect of dosage and pH on % removal of lead using untreated corncob                 55

4.10:    Effect of pH and time on % removal of lead using untreated corncob                      55

4.11:    Effect of dosage and time on % removal of lead using untreated corncob               55

4.12:    Effect of concentration and pH on % removal of lead using treated corncob           58

4.13:    Effect of concentration and dosage on % removal of lead using treated corncob     58

4.14:    Effect of concentration and time on % removal of lead using treated corncob         58

4.15:    Effect of dosage and pH on % removal of lead using treated corncob                     59

4.16:    Effect of pH and time on % removal of lead using treated corncob                          59

4.17:    Effect of dosage and time on % removal of lead using treated corncob                   59

4.18:    Effect of concentration and pH on %removal of chromium using untreated corncob62

4.19:    Effect of concentration and dosage on %removal of chromium using untreated

corncob                                                                                                                       62

4.20:    Effect of concentration and time on %removal of chromium using untreated corncob62

4.21:    Effect of pH and dosage on %removal of chromium using untreated corncob         63

4.22:    Effect of pH and time on %removal of chromium using untreated corncob             63

4.23:    Effect of time and dosage on %removal of chromium using untreated corncob       63

4.24:    Effect of concentration and pH on %removal of chromium using treated corncob   66

4.25:    Effect of concentration and dosage on %removal of chromium using treated corncob66

4.26:    Effect of concentration and time on %removal of chromium using treated corncob 66

4.27:    Effect of pH and dosage on %removal of chromium using treated corncob             67

4.28:    Effect of pH and time on %removal of chromium using treated corncob                 67

4.29:    Effect of dosage and time on %removal of chromium using treated corncob           67

4.30:    Training, validation and test mean squared error using Levenberg–Marquardt

            backpropagation for lead removal using (a) untreated (b) treated corn cob               70

 

4.31:    Training, validation and test mean squared error using Levenberg–Marquardt

backpropagation for chromium removal using (a) untreated (b) treated corn cob      70

 

4.32:    ANN regression plots for the removal of Lead using (a) untreated corn cob

(b) treated corn cob                                                                                                     71

 

4.33:    ANN regression plots for the removal of Chromium using (a) untreated corn cob

(b) treated corn cob                                                                                                     71

 

4.34:    Effect of contact time on removal efficiency                                                                         73

4.35:    Effect of temperature on removal efficiency                                                              74

4.36:    Effect of initial concentration on removal efficiency                                                 75

4.37:    Effect of pH on removal efficiency                                                                            76

4.38:    Effect of adsorbent dosage                                                                                        77

4.39:    Langmuir adsorption isotherm of Pb (II)                                                                    79

4.40:    Langmuir adsorption isotherm of Cr (VI)                                                                  80

4.41:    Freundlich adsorption isotherm                                                                                  80

4.42:    Temkins adsorption isotherm                                                                                      81

4.43:    The plots of pseudo-first-order kinetic model for Cr(VI) and Pb(II)

adsorption onto corncob                                                                                             82

4.44:    The plots of pseudo-second-order kinetic model for Cr(VI) and Pb(II)

adsorption onto corncob                                                                                             83