EXPERIMENTAL INVESTIGATION of an INTEGRATED SOLAR DRIVEN WASTEWATER TREATMENT SYSTEM for TRIGENERATION APPLICATIONS by MURAT

EXPERIMENTAL INVESTIGATION OF AN INTEGRATED SOLAR DRIVEN WASTEWATER TREATMENT SYSTEM FOR TRIGENERATION APPLICATIONS

By MURAT EMRE DEMIR

A Thesis Submitted in Partial Fulfillment of the Requirement for the Degree of Doctor of Philosophy in Mechanical Engineering

University of Ontario Institute of Technology Faculty of Engineering and Applied Science

Oshawa, Ontario, Canada August 2018

ABSTRACT Hydrogen is a promising alternative as an energy carrier (and carbon-free fuel) to meet the global power demand. Hydrogen production systems can efficiently harvest energy from renewable sources. Even though photo-electrochemical systems offer an attractive potential for both hydrogen production and wastewater treatment systems, their application in industrial scales is not satisfactory. This thesis study proposes a trigeneration system for electricity, hydrogen and clean water production. TiO2 photocatalyst is used in a photoelectrochemical (PEC) reactor, which cleans the water via Fenton-like process and produces hydrogen. A novel solar thermoelectric generator (TEG) unit drives the reactor. For the hot side, the phase change material (PCM) which is heated by the concentrator feeds TEG, and the wastewater stream provides the cold surface. The PEC system is investigated experimentally while the design and calculations of the TEG-PCM subsystem are analyzed theoretically in this study. Moreover, in this study, the synergistic effects of advanced oxidation reactions (AOPs) in a combination of TiO2 photocatalysis are comparatively investigated for hydrogen production and wastewater treatment applications. The synergistic effects of Fenton, Fenton-like, photocatalysis (TiO2/UV) and UV photolysis (H2O2/UV) are investigated individually and in a combination of each other. The effects of various parameters, including pH, type of the electrode and electrolyte and the UV light, on the performance of the combined system are also investigated experimentally. A numerical modeling study is performed to predict the temperature, heat transfer rate and generated power distributions on the thermoelectric generator which is included in the integrated system. The thermodynamic properties of the matter flows at each stage and state point are obtained for the integrated system. The overall energy and exergy efficiencies of the integrated system are 5.2% and 5.5% respectively. The total cost rate of the system is determined to be 0.28 $/h from exergoeconomic analysis results. After 40 minutes of the operation time of the PEC reactor under 2.3V of practical cell potential and 60°C of cell temperature, the reactor produces 22.6 mg of hydrogen and removes 87% of COD. The proposed TEG unit configuration generates 551.2 W of electricity with 7.46% heat to electricity efficiency for the case when the temperature of wastewater is 25°C, and the molten salt is at phase changing temperature of 323°C.

Keywords: Efficiency, wastewater treatment, energy, exergy, hydrogen production, photoelectrochemistry, solar energy. i

ACKNOWLEDGMENTS

First and foremost, I would like to express my most profound gratitude and appreciation to my supervisor Professor Dr. Ibrahim Dincer for giving me this unique opportunity to work in this area, his continuous support and accurate advice throughout my research study. He has provided me with exceptional guidance and encouragement to overcome the challenges that I have faced during my Ph.D. research. I am sincerely honored by having Professor Dr. Dincer as supervisor and inspirational mentor.

In addition, I would also like to thank Professor Dr. Bestami Ozkaya for his feedbacks and motivation at critical times. I am so thankful to my colleagues at CERL, Burak Yuzer and Ghassan Chehade who helped me a lot to build my setup and conduct my experiments. In addition, I would like to thank Aaditya Geed who helped me during the literature review and preliminary studies. I am genuinely grateful to all my colleagues and friends in Dr. Dincer`s research group who give me the support and motivation during this journey, particularly, Dr. Yusuf Bicer, Dr. Muhammad Ezzat, Dr. Abdullah Al-Zahrani, Dr. Farrukh Khalid, Maan Al-Zareer, Azzam Abu-Rayash, Osamah Siddiqui, and Huseyin Karasu.

The support provided by the Turkish Academy of Sciences (TÜBA) and the Natural Sciences and Engineering Research Council of Canada is gratefully acknowledged.

Special thanks and love go to my dear wife Busra, and my daughter Zeynep Sare. Their existence has always been my primary motivation during my stay in Canada. I am sincerely thankful for their sacrifice and blessings.

Last but not the least; I would like to express my most profound love and thanks to my parents Hulya and Ahmet Demir for their endless support, determination, inspiration and reassurance throughout my life. Their prayers and good wishes have always been with me. I will always be grateful to them. I would also express my thanks to my elder brother Ali Fatih for being such a strong role model for me.

TABLE OF CONTENTS ABSTRACT ...... i ACKNOWLEDGMENTS ...... ii TABLE OF CONTENTS ...... iii LIST OF FIGURES ...... vi LIST OF TABLES ...... xii NOMENCLATURE ...... xiv CHAPTER 1: INTRODUCTION ...... 1 1.1 Hydrogen Production ...... 3 1.2 Hydrogen from Renewable Resources ...... 4 1.2.1 Hydrogen from Biomass ...... 5 1.2.2 Wind Energy Based Hydrogen Production ...... 6 1.2.3 Geothermal Energy Based Hydrogen Production ...... 6 1.2.4 Solar Energy Based Hydrogen Production ...... 7 1.3 Motivation ...... 10 1.4 Objectives ...... 11 CHAPTER 2: BACKGROUND AND LITERATURE REVIEW ...... 13 2.1 Hydrogen Production Methods from Wastewater ...... 13 2.1.1 Dark Fermentation ...... 14 2.1.2 Photo Fermentation...... 15 2.1.3 Microbial Electrolysis Cells ...... 18 2.1.4 Electrochemical Disinfection ...... 20 2.1.5 Electrodialysis ...... 22 2.1.6 Electrocoagulation ...... 23 2.2 Advanced Oxidation Processes for Wastewater Treatment ...... 25 2.2.1 Electrochemical Oxidation ...... 26 2.2.2 Anodic Oxidation ...... 27 2.2.3 Electro-Fenton ...... 27 2.2.4 Photocatalysis ...... 28 2.3 Thermoelectric Generators ...... 29 2.4 Literature Review ...... 30 2.5 Main Gaps in the Literature ...... 34 CHAPTER 3: SYSTEM DEVELOPMENT AND ANALYSIS ...... 36

iii

3.1 System Description ...... 36 3.2 Alternative integrated system layouts ...... 39 CHAPTER 4: EXPERIMENTAL APPARATUS AND PROCEDURE ...... 42

4.1. TiO2 Sol-gel Dip Coating ...... 42 4.2.PEC setup ...... 44 4.2.1 Single Compartment Cell PEC Reactor...... 45 4.2.2 Small Scale PEC Reactor ...... 47 4.2.3 PEC Reactor with Membrane Electrode Assembly ...... 50 4.3 Analytical Methods ...... 59 4.3.1 COD analysis ...... 59 4.3.2 Iron Measurements and Decolorization Assessment ...... 60 4.4 Experimental Uncertainty Analysis ...... 61 CHAPTER 5: ANALYSIS AND MODELING ...... 63 5.1 Thermodynamic Analyses ...... 63 5.1.1 Solar Concentrator ...... 65 5.1.2 PCM Unit ...... 66 5.1.3 TEG Unit ...... 67 5.1.4 PEC Reactor ...... 69 5.1.5 Overall Energy and Exergy Efficiencies ...... 74 5.1.6 Sustainability index ...... 75 5.2 Exergoeconomic analysis ...... 75 5.2.1 Scale-up analysis ...... 79 5.3 Optimization Study ...... 82 CHAPTER 6: RESULTS AND DISCUSSION ...... 85

6.1 TiO2 Nanoparticle Deposition via Sol-Gel Dip Coating Technique and Photoelectrode Characterization Study Results ...... 86 6.1.1 Small Scale PEC Reactor electrodes ...... 94 6.2 Single Compartment Cell PEC Reactor ...... 97 6.3 Synergistic Effects of AOPs for Hydrogen Production and Wastewater Treatment Processes ...... 102 6.4 Thermoelectric Generator ...... 115 6.5 Molten Salt storage...... 119 6.6 PEC Reactor and Integrated System Results...... 121 6.7 Exergoeconomic assessment results...... 131 iv

6.7.1 Scale-up analysis results ...... 136 6.8 The Optimization Study Results ...... 141 CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS ...... 149 7.1 Conclusions ...... 149 7.2 Recommendations ...... 152 REFERENCES ...... 154

LIST OF FIGURES Fig. 1.1 Global renewable internal freshwater resources and energy usage per capita (data from Ref. [1,2]) ...... 1 Fig. 1.2 Share of renewables on worldwide electricity generation (data from Ref. [4]) ...... 2

Fig. 1.3 Global H2 production capacity as of 2016 (data from Ref. [7]) ...... 4 Fig. 1.4 Paths of generation of hydrogen from renewable energy sources (modified from Ref. [14]) ...... 5 Fig. 1.5 A simple layout of a PV hydrogen production system...... 9 Fig. 2.1 Classification of hydrogen production methods from wastewater...... 13 Fig. 2.2. Comparison of the solar based bio-hydrogen production processes (modified from ref. [51])...... 17 Fig. 2.3 Diagram of a conventional two-chamber microbial electrolysis cell (modified from ref.[65]) ...... 19 Fig. 2.4 A single-chamber microbial electrolysis cell (modified from [72]) ...... 20 Fig. 2.5 A wastewater electrolysis cell (modified from ref. [79]) ...... 21 Fig. 2.6 A simple layout of the electrodialysis cell. (modified from ref.[86]) ...... 22 Fig. 2.7 Classification of advanced oxidation processes (modified from ref. [102]) ...... 26 Fig. 2.8 A simple layout of a thermoelectric generator ...... 29 Fig. 3.1 Layout of the system integrated the system...... 36 Fig. 3.2 Block diagram of the system integrated the system...... 38 Fig. 3.3 Illustration of the alternative integrated system layout 1...... 39 Fig. 3.4 Illustration of the alternative integrated system layout 2...... 40 Fig. 3.5 Illustration of the alternative integrated system layout 3...... 40 Fig. 3.6 Illustration of the alternative integrated system layout 4...... 41 Fig. 4.1 Block diagram describing the procedure implemented for preparing sol-gel applied by dip coating on steel anode ...... 42 Fig. 4.2 Custom made dip coating mechanism with adjustable withdrawal speed...... 43 Fig. 4.3 Temperature controlled furnace used for the annealing process...... 44 Fig. 4.4 Experimental setup of the single compartment cell PEC reactor ...... 46 Fig. 4.5 Experimental setup with three interconnected columns used in small-scale tests...... 47 Fig. 4.6 Structure of potassium hydrogen phthalate (KHP)...... 49 Fig. 4.7 Structure of Reactive Black 5 (RB5) dye...... 50 vi

Fig. 4.8 The PEC reactor for simultaneous hydrogen production and wastewater treatment...... 51 Fig. 4.9 The assembly of the PEC reactor...... 52 Fig. 4.10 Anode compartment of the PEC reactor ...... 54 Fig. 4.11 Hydrogen production measurement by manometer with video tracking...... 55 Fig. 4.12 Gamry Instruments Reference 30k booster and Reference 3000 [132,133]...... 55 Fig. 4.13 Pyranometer used in the study [150]...... 56 Fig. 4.14 Thermocouple wire used in the study [151]...... 57 Fig. 4.15 Data logger used in the study [153] ...... 58 Fig. 4.16 Block digester [154] (a) and COD and a multiparameter bench photometer [136] used in this study ...... 61 Fig. 5.1 PCM-TEG module used in this study...... 69 Fig. 5.2 Hydroxyl radical formation mechanisms in the reactors used in this study...... 72 Fig. 5.3 The layout of the scale-up system...... 79 Fig. 5.4 Genetic algorithm cycle (modified from [182]) ...... 83 Fig. 6.1 The sequence of the results sections ...... 85

Fig. 6.2 TiO2 coated stainless steel plate before (a) and after (b) the annealing process...... 86

Fig. 6.3 Cyclic voltammetry graphs of the cell with TiO2 coated stainless steel working electrode and graphite counter electrode...... 87

Fig. 6.4 Partial cyclic voltammetry graphs of the cell with the TiO2 coated stainless steel working electrode and the copper counter electrode ...... 88

Fig. 6.5 Partial cyclic voltammetry graphs of the cell with the TiO2 coated stainless steel working electrode and the nickel counter electrode...... 88

Fig. 6.6 Cyclic voltammetry graphs of the cell with TiO2 coated stainless steel working electrode and various counter electrodes (vs. Ag/AgCl)...... 89

Fig. 6.7 Current vs. time graph with an applied potential of -1.4V (vs Ag/AgCl) of the TiO2 coated photoanode and graphite cathode under light and no light conditions...... 90 Fig. 6.8 Cyclic voltammetry results (-0.3 to 0.9 V vs. Ag/AgCl reference electrode) of coated steel plate and graphite rod counter electrode...... 91 Fig. 6.9 Linear voltammetry sweep results with an initial potential of 0.15V (vs. Ag/AgCl) under the chopped light...... 91 Fig. 6.10 Current vs. time graph with the potentiostatic applied potential of 1.4 V (vs. Ag/AgCl) of TiO2 coated photoanode and graphite cathode...... 92 vii

Fig. 6.11 The production yield under the potentiostatic applied potential of 1.4 V (vs. Ag/AgCl) of TiO2 coated photoanode and graphite cathode under light and no light conditions (V vs. Ag/AgCl)...... 93

Fig. 6.12 Open circuit potential of TiO2 coated stainless steel with Na2SO4 (0.14 M) electrolyte solution under chopped UV light condition...... 95 Fig. 6.13 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V.

(TiO2 coated stainless steel used as the anode under UV light and dark conditions with 0.14 M

Na2SO4 as the electrolyte solution)...... 95 Fig. 6.14 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V. Graphite used as the anode under UV light and dark conditions with 0.14 M Na2SO4 as an electrolyte solution...... 96 Fig. 6.15 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V.

Stainless steel used as the anode under UV light and dark conditions with 0.14 M Na2SO4 as an electrolyte solution...... 97 Fig. 6.16 (a) Current vs. time graph and (b) hydrogen production rate from different types of synthetic wastewater (RB5, sucrose) and tap water by using photo-assisted electro-Fenton process (PAEF)...... 98 Fig. 6.17 (a) Current vs. time graph and (b) hydrogen production rate from different types of synthetic wastewater (RB5, sucrose) and tap water by using photo-assisted electro-Fenton process (PAEF) ...... 98 Fig. 6.18 The RB5 degradation of treated and untreated samples for a solution (25 l) containing 6×10-5 M RB5 (55 ppm of initial COD) at the beginning of the electrolysis ...... 99 Fig. 6.19 The RB5 solution before and after the treatment ...... 100 Fig. 6.20 The comparison of the optical density of the RB5 solution before and after the electrolysis...... 100 Fig. 6.21 The electroproduction of iron ions on an SS rod anode in KHP solution and under UV light...... 102 Fig. 6.22 The RB5 degradation for Experiment 21 and 22. Spectra recorded for the RB5 solution, anolyte and catholyte at the end of electrolysis and 1 day after...... 104 Fig. 6.23 Current vs. time graph of the RB5 solution electrolysis with a different pH level of dye solution under UV light condition with the TiO2 coated stainless steel photoanode and graphite cathode...... 106 viii

Fig. 6.24 KHP degradation and hydrogen production in PEC reactor under various experimental conditions...... 107 Fig. 6.25 RB5 degradation and hydrogen production in PEC reactor under various experimental conditions...... 107 Fig. 6.26 Power usage(a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the KHP solution as the electrolyte. The steel anode and the graphite cathode is used with (E1) and without UV light(E2) conditions ...... 109 Fig. 6.27 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the KHP solution as the electrolyte. The graphite anode and the graphite cathode is used with (E3) and without UV light(E4) conditions ...... 110 Fig. 6.28 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. (Steel anode and graphite cathode) ...... 110 Fig. 6.29 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. The graphite anode and graphite cathode is used with (E9) and without UV light(E10) conditions ...... 111 Fig. 6.30 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. The TiO2 coated steel photoanode and the graphite cathode is used with (E12) and without UV light(E11) conditions ...... 112 Fig. 6.31. Optical absorbance values of RB5 solution for various experiments after one day from the electrolysis...... 114 Fig. 6.32 Variation of thermoelectric performance with temperature ...... 115 Fig. 6.33 Change of heat to electricity efficiency and generated electricity of the TEG with the inlet temperature of the cold stream...... 116 Fig. 6.34 Change of exit stream temperature and heat transfer rate of the TEG with the inlet temperature of the cold stream...... 117 Fig. 6.35 The temperature distribution of the TEG module...... 118 Fig. 6.36 Change of heat to electricity efficiency of the TEG with the inlet temperature of the cold stream...... 119 Fig. 6.37 The variation of heat losses through the molten salt tank with the changing insulation thickness ...... 120

Fig. 6.38 Current-time curve of the reactor measured by the potentiostat device under two different operating temperature ...... 122 Fig. 6.39 The variation of the reactor efficiencies with operating temperature...... 123 Fig. 6.40 The variation of the hydrogen production rate with operating temperature ...... 123 Fig. 6.41 Current-time curve of the PEC reactor measured by the potentiostat device under two different operating temperature ...... 124 Fig. 6.42 KHP degradation and hydrogen production in PEC reactor under different temperatures (0.14 M KHP anolyte and catholyte under 40 minutes operation of the reactor)...... 125 Fig. 6.43 KHP degradation and hydrogen production in the PEC reactor under different temperatures ( 0.14 M KHP anolyte and 0.1 M H2SO4 catholyte under 40 minutes operation of the reactor)...... 126 Fig. 6.44 Color change of textile wastewater (RB5 solution) in PEC reactor during 5 hours of operation...... 127 Fig. 6.45 Optical absorbance of textile wastewater (RB5 solution) in PEC reactor during 5 hours of operation...... 128 Fig. 6.46 Optical absorbance of RB5 solution (at 590 nm) in PEC reactor during 52 minutes of operation...... 128 Fig. 6.47 Reaction rate of the optical absorbance removal of RB5 solution...... 129 Fig. 6.48 Exergy destructions and losses of the system components...... 131 Fig. 6.49 The cost breakdown of the proposed system...... 132 Fig. 6.50 Exergy destruction rates of the subsystems...... 133 Fig. 6.51 The share of exergy destruction cost rate in each component of the integrated system...... 133 Fig. 6.52 The effects of the interest rate on the system cost rates and exergoeconomic factor. . 134 Fig. 6.53 The effects of the annual operation time on the system cost rates and exergoeconomic factor...... 135 Fig. 6.54 The effects of the electricity cost on the system cost rates and exergoeconomic factor...... 136 Fig. 6.55 The price range of the water desalination system with various technologies (modified from [199])...... 139 Fig. 6.56. The sensitivity analysis for hydrogen cost rate ($/kg) with various parameters...... 140 Fig. 6.57. Cost of hydrogen the after system improvements...... 141 x

Fig. 6.58 The variation of the objective function and system outputs through the iterations of the optimization study...... 147

LIST OF TABLES Table 1.1 Selected solar hydrogen production approaches. (modified from [17,26]) ...... 8

Table 2.1 H2 production rates of anaerobic reactors treating agricultural waste (adapted from ref. [35])...... 15 Table 2.2 Advanced oxidation processes (adapted from [100]) ...... 25 Table 4.1 Features of the membrane used in the PEC reactor [146] ...... 53 Table 4.2 Accuracies related to measurement devices...... 62 Table 5.1 Design parameters of the solar field ...... 65 Table 5.2 Properties of the selected molten salt [124]...... 66 Table 5.3 Properties of the thermoelectric materials [123] ...... 69 Table 5.4 Material costs of the integrated system...... 77 Table 5.5 The financial parameters of the proposed system...... 78 Table 5.6 The specifications and parameters of the scale-up system...... 80 Table 5.7 The cost parameters...... 81 Table 6.1 Comparison of photocurrent and applied voltage for different substrate materials ..... 94 Table 6.2 Measured data of the single cell reactor with the uncertainty range ...... 101 Table 6.3. Experimental conditions of the comparative study...... 103

Table 6.4 Effect of sacrificed iron and H2O2 concentration ratio on COD removal efficiency . 109 Table 6.5 Absorbance of the substrates at specific wavelengths, which were, referred in German wastewater discharge standards...... 113 Table 6.6 Measured data of the comparative study with the uncertainty ranges...... 114 Table 6.7 Design parameters of the PEM electrolyser [20] ...... 121 Table 6.8 The exergoeconomic results of the components in the integrated system ...... 134 Table 6.9. The costs of PEC plant...... 137 Table 6.10 Fixed operating costs of the plant...... 137 Table 6.11 The direct capital costs of the components in 1000 kg/day PEC concentrated light hydrogen production plant...... 138 Table 6.12 Ground and surface waters extraction volumetic costs in different [199,200]...... 140 Table 6.13 Multi-objective optimization results for the total cost rate and exergy destruction rate of the solar PCM subsystem including the sensitivities ...... 142 Table 6.14 Multi-objective optimization results for the total cost rate and induced thermoelectricity, sustainability index of the TEG subsystem including the sensitivities ...... 143 xii

xiii

NOMENCLATURE A Area (m2), pre-exponential factor C Concentration (mol/L)

Ċ D Cost rate of exergy destruction ($/h) cp Specific heat at constant pressure (kJ/K) DNI Direct normal irradiance (W/m2)

DP Depletion factor E Actual PEM electric potential (V)

Eact Activation overpotential (V)

Econc Concentration overpotential (V)

Eohm Ohmic overpotential (V)

Er Reversible potential (V) Eẋ Exergy rate (kW) ex Specific exergy (kJ/kg) f Exergoeconomic factor F Faraday constant (9.6485 × 104 C/mol) h Specific enthalpy (kJ/kg), convective heat transfer coefficient W/(m2 K) hfg Enthalpy of vaporization (kJ/kg) HHV Higher heating value (kJ)

ΔHm Enthalpy of fusion (kJ/kg) I Current (A), Irradiance (W/m2) j Current density (A/m2) 2 j0 Exchange current density (A/m ) 2 jL Limiting current density (A/m ) k Thermal conductivity (W/m K), turbulent kinetic energy (m2/s2), reaction rate LHV Lower heating value (kJ) ṁ Mass flow rate (kg/s) M Molality (mol/L), Molar mass (g/mol) N Number of moles Ṅ Molar flow rate, mol/s NA Avogadro number

xiv

P Pressure (kPa) R Universal gas constant (8.314 kJ/kmol-K) q Electrical charge (A s) Q̇ Heat rate (kW) s Specific entropy (kJ/kg K) S Seebeck coefficient (V/K) t Time (s) T Temperature (K) V Voltage Ẇ Work rate (kW) yi Molar fraction of species Ż Cost of owning and operating the system ($/h) ZT̅ Modified dimensionless figure of merit

Greek letters

η Energy efficiency ψ Exergy efficiency Δ Change λ Wavelength (nm) ϕ Quantum efficiency σ Electrical conductivity (S/m), Stefan-Boltzmann constant (5.76 10−8 W/m2K4) δ Membrane water content ε Surface emissivity, the rate of kinetic energy dissipation (m2/s3) ρ Density (kg/m3) ħ Planck’s constant (6.62606957 × 10-34 m2kg/s)

Subscripts and superscripts a Anode, air ad Adsorbed c Cathode, concentrator, cold side ch Chemical d Destruction

xv dl Destruction and losses e Electrical en Energy ex Exergy h Hot side HEX Heat Exchanger i State number in Inlet of a component loss Heat losses m Membrane, melting n n-type thermoelectric material out Outlet of a component ov Overall p p-type thermoelectric material ph photo r receiver 0 Ambient condition 1, 2, … i State points, Experiment number

Acronyms

AOP Advanced Oxidation COD Chemical oxygen demand CPV Concentrated Photovoltaics CST Concentrated solar thermal energy CERL Clean Energy Research Laboratory DF Dark fermentation DSSC Dye sensitized solar cell E Experiment EC Electrochemical ED Electrodialysis EES Engineering Equation Solver GA Genetic Algorithm xvi

GHG Greenhouse gas HDPE High density polyethylene HEX Heat exchanger kgoe kilograms of oil equivalent MACRS Modified Accelerated Cost Recovery System MEC Microbial electrolysis cell MFC Microbial fuel cell MPC Microbial photoelectrochemical cell N, n Total number OCP Open circuit potential PAEF Photo-assisted electro-Fenton process PCM Phase Change Material PF Photo fermentation PEC Photoelectrochemical PEM Proton Exchange Membrane PtB Platinum black PV Photovoltaics RCS Reactive chorine species SI Sustainability index SS Stainless steel SCCM Standard cubic centimeters per minute TEG Thermoelectric generator TEM Thermoelectric material TES Thermal energy storage UOIT University of Ontario Institute of Technology US Ultrasound UV Ultraviolet UV/V Ultraviolet-visible spectral region WW Wastewater VFA Volatile fatty acid

xvii

CHAPTER 1: INTRODUCTION

In today’s world, there is a significant interest in sustainable development to meet the demand of increasing global needs such as energy and clean water due to population growth and higher living standards. The high demands on energy and fresh water, bring about renewable fresh water shortage and higher energy consumption rate. Fig. 1.1 exhibits the global renewable internal freshwater resources and energy usage per capita. Renewable internal freshwater resources such as internal river flows and groundwater from rainfall diminished due to increasing water demand, population growth, industrial development and the expansion of irrigated agriculture over the past century. Renewable internal freshwater resources per capita range from 13391m3 to 5926m3 in the past 52 years [1]. Similarly, 1,336 kilograms of oil equivalent (kgoe) energy was consumed by per capita worldwide in 1971, and it reached 1,929 kgoe by 2014. According to the trends, these numbers are expected to go up in the upcoming years [2].

16000 2500

14000 2000 12000

10000 1500

8000

per capita) per 3 Renewable internal 1000 6000 freshwater resources

per capita (1000 m (1000 4000 Energy use 500

2000 ) capita per capita(kgoe per use Energy Renewable internal freshwater resources resources freshwater internal Renewable 0 0 1960 1970 1980 1990 2000 2010 Year Fig. 1.1 Global renewable internal freshwater resources and energy usage per capita (data from Ref. [1,2])

Energy is one of the critical factors for technological developments, society, world politics, and the environment. All of our daily works are dependent on abundant and continuous energy supply. According to the World Bank data, 78 % of the total energy consumption comes from fossil fuels. The dependence on fossil fuels is not sustainable. The finite existence of the conventional energy sources brings potential energy crisis with it. Moreover, carbon-based fuels are very detrimental to the environment. 87% of the carbon emissions caused by humans is due to the usage of fossil fuels [3]. Hence, it is of importance to focus on seeking alternative clean energy sources to meet the rising energy demand in an environmentally friendly manner. As clean energy sources, renewable energy can be considered sustainable alternatives due to their significant advantages over fossil fuels.

Fig. 1.2 Share of renewables on worldwide electricity generation (data from Ref. [4])

Currently, around 22.8% of the global electricity generation comes from renewables and in particular hydropower as illustrated in Fig. 1.2. However, in many cases, it is impossible to cover the electrical power demand by using only one renewable energy resource due to its intermittent nature. Additionally, renewables cannot be dispatched to meet the demand of a power system, which is also considered as one of the main drawbacks of renewable energy. Thus, energy storage mediums with hybrid systems ought to be established to overcome these issues. Hydrogen is a promising alternative to address the

2 aforementioned issues and can be used as an energy carrier. Some of the main advantages of hydrogen include the following [5]:  High energy conversion efficiency  No emissions when used  Abundant sources in nature  Availability of various storage alternatives  Availability of transportation over long distances  Conversion into various fuel options  High heating value, compared with that of many conventional fuels (HHV: 141.80 MJ/kg and LHV:119.96 MJ/kg)  Easily producible from water using renewable sources with smallest environmental impact.

1.1 Hydrogen Production Hydrogen is the most abundant element in the universe. While it is not readily available in the environment, it can be produced via different reactions. Various hydrogen production methods are developed and can be categorized into six main group including electrochemical, thermochemical, photochemical, radiochemical, biochemical and hybrid (combination of two or more methods) [6]. The increase in hydrogen production capacities in the world can be seen in Fig. 1.3. The last decade has witnessed a growing hydrogen production capacity. The global production rate has increased by 12.3% in ten years and reached over 400 Mm3 per day [7]. As expected, higher hydrogen production requires more energy. Amongst the existing hydrogen manufacturing methods, fossil fuel-based techniques are currently the most common for hydrogen production. Over 95% of the global hydrogen production is presently met by various fossil fuel sources [8]. Steadily increasing the usage of carbon-based fuels leads to severe concern regarding CO2 concentration in the atmosphere. Furthermore, oil prices rise dramatically as a result [9]. Combustion of fossil fuels for hydrogen production conflicts with the concept of sustainable energy. Thus, an alternative energy source that is highly available in nature with manageable collateral effects (particularly environmental) should be used for hydrogen production. In this context, renewables seem the most appealing option among the other energy sources.

450 400 350

300 /day) 3 250 200

150 (Million m (Million 100

Hydrogen production capacity capacity production Hydrogen 50 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Fig. 1.3 Global H2 production capacity as of 2016 (data from Ref. [7])

1.2 Hydrogen from Renewable Resources Presently, a high majority of the hydrogen production is attained by the utilization of hydrocarbons via thermochemical processes which is also known as reforming [9–12].

However, the finite existence of the fossil fuels and the CO2 emission related environmental concerns have raised the need for renewable energy sources. Hence, shifting the focus of future energy scenarios towards renewables is not only justified but also required. Fig. 1.4 shows various pathways for hydrogen production from renewables. All the presented renewable energy forms can be used to generate electrical power for hydrogen production. Since the photonic energy can only be found through solar radiation, solar energy is used in photochemical hydrogen production applications. Specific microorganisms, manipulate biochemical energy to construct their essential materials. Biochemical energy is mostly stored in organic matters like carbohydrates, glucose, sucrose, proteins, and cellulose. The photosynthetic mechanisms allow plants and microorganisms to produce hydrogen [13]. Hence, both solar energy and biomass together can be used in a combination. Besides, biomass can also be utilized without the presence of light with dark fermentation. Energy from the Earth can also be used (so-called:

4 geothermal) for hydrogen production via thermal energy form (thermochemical water splitting) or electrical form (electrolysis).

Renewable Energy

Renewable Ocean Tides and Solar Biomass Geothermal Wind Hydro energy Thermal Waves

Photonic Biochemical Thermal Electrical Energy Energy Energy Energy

Hydrogen

Fig. 1.4 Paths of generation of hydrogen from renewable energy sources (modified from Ref. [14])

1.2.1 Hydrogen from Biomass Biomass is a feedstock, which is defined as any renewable, organic matter that can be utilized directly as a fuel, or converted to another form of fuel or energy product. Biomass is considered as one of the significant sustainable energy sources in the world, and it is expected to replace fossil fuel in the future. It is going to play a vital role in the future of global energy scenario. Thus, it has attracted considerable interest in the last years [8,15,16]. Biomass currently contributes 12% of total global energy production [17]. Hydrogen can be manufactured from various kind of biomass such as corn starch, sugar cane juice, forestry, industrial and municipal waste, and agricultural and industrial crops residues [18]. Numerous types of processes can be performed to extract hydrogen from biomass such as thermochemical conversion, dark fermentation, anaerobic digestion, photo-fermentation and sequential dark and photo-fermentation processes [6]. Currently, biomass based bio-hydrogen production is not available for large-scale applications. Even though biomass is considered as an indigenous feedstock, its application fields are limited

5 due to its drawbacks such as lower hydrogen production rates, seasonal availability, operational and transportation difficulties [19]. The chemical reactions of biomass-based hydrogen production methods are quite similar to the carbon-based fuel methods. Gasification and pyrolysis are recognized as the most favorable medium-term technologies for the commercialization of biomass-to-hydrogen technology [20]. Biomass carries the potential to diminish greenhouse gas (GHG) emissions as it absorbs back CO2 through photosynthesis [21].

1.2.2 Wind Energy Based Hydrogen Production Electricity generation via wind power has the least adverse environmental impact compared with the other energy resources [22]. In 2012, over 280 GW electric power was produced by the wind energy, accounting for approximately 2% of the overall global electricity demand [8]. Especially for distributed systems, the generated electricity by the wind turbine could be used in water electrolysis for zero-emission hydrogen production [17]. However, higher investment costs and lack of dispatchability are still considered as the main drawbacks of the wind to hydrogen technology.

1.2.3 Geothermal Energy Based Hydrogen Production Geothermal energy appears to be a promising sustainable and renewable energy option for hydrogen production, owing to its relatively low cost and abundance. Hydrogen can be produced via geothermal energy based pathways such as thermal dissociation of water or water electrolysis. Geothermal energy can be utilized for producing the required electricity for electrolysis or heat energy can be used in thermochemical processes. Thermochemical cycles were initially proposed in the late 20th century as high potential and an alternative method for hydrogen manufacturing. Thermochemical water splitting for hydrogen production is a series of chemical processes that achieve the decomposition of water into oxygen and hydrogen by using high-temperature heat (500°C to 2,000°C). The thermochemical process can also be done in a hybrid form by combining heat with electricity. Thermochemical cycles usually comprise of chain reactions, where water is thermally decomposed, and all other chemicals are recycled. Only water is consumed as it decomposes into the H2 and O2 [23].

1.2.4 Solar Energy Based Hydrogen Production Due to the high availability of solar power and its wide range of applicability for hydrogen production, considerable interest has emerged for this source. Hence, much effort has been invested in the research and development of solar-based systems. The amount of solar irradiance reaching the surface of the world in around half an hour corresponds to the same amount of energy as the annual global energy demand [24]. Solar energy can be utilized in a variety of hydrogen production applications including photo- fermentation, artificial photosynthesis, photovoltaic-electrolysis and photo-electrolysis. In addition, the solar thermal form can also be used in thermochemical water splitting processes. In Table 1.1, the primary solar hydrogen production methods are tabulated and briefly explained. As can be seen in this table, hydrogen can be extracted via solar-based processes with the help of various energy forms, including:  Photonic energy  Solar thermal energy  Electrical energy  Biochemical energy Photovoltaic-based solar hydrogen production pathways are mainly simple water electrolysis, which is driven by PV electricity (as seen in Fig. 1.5). In solar thermal energy, solar energy is used for hydrogen production in solar thermal form. The solar thermal energy can be utilized in either low-temperature applications or concentrated solar power (high temperature) applications for hydrogen manufacturing. Thermolysis is a method for concerted solar energy-based water decomposition, which consists of a single-step process at elevated temperature (to 2,227-2,727 oC). For instance, at 2,727 oC and 100 kPa, dissociation ratio of the water into the essentials is 64% [25]. Since the thermolysis process is reversible, one of the main challenges in the application is to prevent the recomposition of water from hydrogen and oxygen.

Table 1.1 Selected solar hydrogen production approaches. (modified from [17,26]) Hydrogen Operation Material Brief description of End generation conditions resources Process Products technique Hydrogen Low PV driven water PV- Electrolysis Water Oxygen Temperature electrolysis.

Photoelectrolysis of Hydrogen Low PEC Water water with a hybrid Oxygen Temperature cell. Decomposition of Hydrogen Dark Low Biomass+ wastewater or Carbon fermentation Temperature wastewater organic matter by dioxide microorganisms Decomposition of wastewater or Hydrogen Low Wastewater+ organic matter by Photo-biological Carbon Temperature biomass microorganisms and dioxide sunlight into hydrogen. Thermolysis by The breakdown of concentrated High water molecules by Hydrogen Water solar thermal Temperature the action of heat at Oxygen energy (CST) higher temperatures Combination of heat Thermochemical High Water + Metal sources with Hydrogen process by CST Temperature oxides chemical reactions to Oxygen split water Reaction of coal with Hydrogen Gasification by High O and steam under Carbon Water+ Coal 2 CST Temperature higher temperatures monoxide and pressures Thermal breakdown Natural gas, oil Fuel cracking by High of natural gas, oil, Hydrogen or other CST Temperature and other Carbon hydrocarbons hydrocarbons Hydrogen Water, Natural Steam reforming of Steam reforming High Carbon gas, oil or other fuels under higher by CST Temperature dioxide hydrocarbons temperatures

High- Solar thermal Temperature High electricity driven Hydrogen Water Electrolysis by Temperature high-temperature Oxygen CST water electrolysis

Another critical point for the process is to construct the thermolysis reactor endurable at higher temperature range. The selected materials should be able to stand the temperature range of 1500 K with substantial temperature fluctuations. Besides, the chosen material should be durable to chemically active medium for longer a lifetime [27,28]. Apart from reaching and maintaining greater temperature range required for the process, other obstacles are summarized as follows:  Recombination of gaseous products may cause the formation of an explosive mixture. To prevent that situation, feasible systems for separating hydrogen from the other thermolysis products have to be developed.  The lack of suitable materials capable of enduring a chemically active environment.  The lack of suitable materials can stand the desired temperatures.

Fig. 1.5 A simple layout of a PV hydrogen production system.

The recent research interests have been shifted on thermochemical cycles, while scientists are tackling these issues [29]. Thermochemical water splitting is an alternative to electrolytic decomposition of water. The process requires multiple chemical reactions and lowers the required temperature to drive conventional one-step high-temperature water thermolysis. Currently, one of the significant challenges of the process is attaining higher efficiencies at lower temperature levels [30]. On the other hand, thermochemical cycles

9 yield promising results to be considered as a potential method for hydrogen manufacturing [20]. The processes like photovoltaics, biophotolysis, artificial photosynthesis are considered as the low-temperature methods where, coal gasification, thermolysis, fuel cracking, and thermochemical cycles are regarded as high-temperature processes [26]. Light-dependent biological hydrogen manufacturing processes follow the same principles as algal and plant photosynthesis. Photo-fermentation, direct and indirect biophotolysis are the main photobiological hydrogen production methods. Photosynthesis of algae and plants drives the hydrogen production by splitting the water into O2 and a reducing agent, which is strong enough to reduce protons or carbon dioxide to produce H2 or carbohydrates respectively [26].

1.3 Motivation According to the trends in water usage and climate change scenario, almost 50% of the global population will suffer from water scarcity by 2030 [31]. Water scarcity has already become a critical global issue. Several countries have been focusing their interest on wastewater utilization for various purposes. Furthermore, various wastewater treatment methods have been introduced in the last century. All those processes require some sort of energy. Driving the water treatment process with fossil fuels is not eco-friendly nor sustainable. Hence, it is of importance to investigate, develop, and commercialize renewable-based, eco-friendly and highly efficient water purification systems for a more sustainable future. Solar irradiance is the most abundant and available energy source on earth. However, it is not stable and has the intermittency problem, which limits its applications. Consequently, eﬃcient and low-cost solutions for solar energy storage in a compact medium are essential to enhance the performance of solar systems. Solar energy based combined systems have several advantages with multiple outputs at higher efficiencies [32]. Hydrogen appears to be a potential energy carrier (and carbon-free fuel) for meeting the global power demand. Hydrogen production systems can effectively harvest energy from renewable sources [13]. Even though photoelectrochemical systems offer attractive potential for both hydrogen production and wastewater treatment systems, their application in the industrial

10 scale is not satisfactory. Therefore, the underlying motivation of this study is the potential of combining photoelectrochemical hydrogen production methods with wastewater treatment systems by storing the solar energy in both thermal and chemical forms.

1.4 Objectives The primary purpose of this thesis study is to propose and develop a novel solar-based trigeneration system for wastewater treatment, photoelectrochemical hydrogen production and thermoelectric power generation. The suggested integrated system maximizes the utilization of solar energy by storing it in thermal and chemical forms (PCM and hydrogen). In more detail, the specific objectives of this thesis are as follows:  To design and analyze a solar-driven photoelectrochemical system for wastewater treatment, hydrogen production, and electricity generation. The proposed system should be integrated with a latent heat storage embedded TEG subsystem to operate continuously. In a photoelectrochemical reactor, cell voltage and current are simultaneously produced upon absorption of solar light by photoanode. Furthermore, the photoanode is used to assist the wastewater treatment process by generating hydroxyl radicals via a photocatalytic process. AOPs are investigated for wastewater treatment processes in the anode section.  To conduct a comprehensive thermodynamic modeling for the proposed integrated system and complete the performance assessment, energy and exergy analyses of the system thermodynamically. a) All the balance equations are carried out (mass, energy, entropy, and exergy) for all the system elements. b) Exergy destruction rate and energy losses for each component are investigated to obtain exergy efficiencies for individual system elements. c) A numerical study is performed to model the heat transfer on the TEG module and assess the heat to electricity conversion performance. d) To perform an electrochemical modeling for water electrolysis in order to investigate the relationship between the production yields and the temperature.  To build a lab-scale experimental setup and conduct required experiments considering the appropriate materials research for each system components. The following tasks are taken into account in the designing process.

a) To select suitable methods for system monitoring (to measure hydrogen production rate, electricity usage and degraded organics in the wastewater) b) To select the main variables such as temperature, irradiation, electrical current,

voltage, COD value and H2 production to investigate production yields and system performance assessment. c) To define the working conditions of the experimental setup by determining the input parameters to be varied (temperature, wastewater, irradiation, applied voltage), control parameters to be monitored (temperature, applied current, pH,),

and output parameters to be measured (H2 production, COD value, current generation).  To perform several experimental studies on the system based on different parameters and environmental conditions to examine the influence of primary variables on production and system efficiency: a) PEC process: Investigate the impact of irradiance level, operating temperature, wastewater type and applied voltage on the hydrogen production rate, COD removal efficiency, energy efficiency and exergy efficiencies. b) Advanced Oxidation process: Investigate the synergistic effects of AOPs on hydrogen production and wastewater treatment applications. The AOPs that take place in the anode section are investigated individually and in combinations of the other AOPs.  To perform an exergoeconomic analysis to evaluate the viability of the proposed system and relate exergy with the cost and environmental impact. Cost determination of each sub-system, cost of exergy destruction and estimation of the purchase cost of each component are all determined in this analysis.  To perform a multi-objective optimization on the system parameters such as operating temperature, ambient temperature, irradiance, operation time and financial conditions such as interest rate with respect to (a) exergy analysis: to maximize the exergy efficiency (ii) exergoeconomic analysis: to reduce the associated cost, and (iii) water analysis: to maximize the COD degradation.

CHAPTER 2: BACKGROUND AND LITERATURE REVIEW

In this chapter, hydrogen production from wastewater methods including biological, electrochemical, photochemical and hybrid processes (in a combination of two or more processes) are explained in detail. Then, advanced oxidation processes for water treatment applications and reaction mechanisms are briefly presented. The related studies in the literature are shown and the main gaps are discussed in the final part of the chapter.

2.1 Hydrogen Production Methods from Wastewater Organic matter based hydrogen production methods are considered one of the most promising options for sustainable energy production. These processes are favored as they degrade organic matters while producing hydrogen from wastewater.

Photo Fermantation Photo-biological Bio-photolysis

Biological Dark Fermantation

Microbial Electrolysis Bio-electrochemical Hydrogen Cell Production Methods from Wastewater Electrochemical (EC) Electrochemical disinfection Electro - Fenton

Photoelectrocatalytic Photoelectrochemical disinfection (PEC)

Photocatalytic Photochemical disinfection

Fig. 2.1 Classification of hydrogen production methods from wastewater.

Hydrogen production combined with wastewater treatment methods can be categorized as presented in Fig. 2.1. As exemplified in the figure, these methods consist of biological, electrochemical, solar assisted or hybrid (combination of two or more methods) processes. In this section, all these methods are explained individually in detail.

2.1.1 Dark Fermentation Dark Fermentation is an anaerobic breakdown of organic substances, and it is recognized as the most straightforward method for attaining biological hydrogen. Fermentation occurs in the absence of light at room conditions. It yields in higher H2 production rates as compared to photosynthetic approaches [33]. Dark Fermentation can be performed using many types of organic substances including glucose, carbohydrates, starch, cellulose, by-products of food and agricultural industries. Microorganisms involved in the production of bio-hydrogen are strict anaerobes. Studies have been conducted to evaluate the potential use of solid substrate conversions such as agricultural waste and food waste. In addition to bio-hydrogen production, specific volatile fatty acids and alcohols can be produced by adjusting the operating conditions based on the metabolic pathways driven by the bacteria [33]. There are many factors such as temperature, substrate type, pre-treatment, bioreactor conditions that affect the production of bio-hydrogen. Various studies evaluated the bio-hydrogen production through different food wastes at several temperatures. One hypothesis that has come up is that bio-hydrogen production may be inversely correlated with the cellulose and lignin content of the waste [34]. Bio-hydrogen harvests from several crop substrates in the current literature are tabulated in Table 2.1. In order to make all the results comparable among each other, all the listed studies in Table 2.1 have been selected using the same operating temperature of 35°C and the same type of operation mode (batch type experiments). To determine the effect of the pre-treatment process on the overall performances, more comprehensive studies should be conducted [35]. The limitations of this process at industrial scale, stem from poor production and conversion of fermentative biological processes [34,36]. Optimization of operating conditions of the reactors remains a significant factor regarding the improvement of the biological H2 production. One of the most critical parameters is the pH, which regulates the anaerobic fermentation process. The level of pH does not only affect the bio-hydrogen

14 yields, but can also modify the by-products of the process. For optimal H2 manufacturing, a pH of 5-6 for food wastes and a neutral pH for crop residues are recommended [34]. The partial pressure of H2 is another limiting parameter during the fermentation of organic matter [34]. Another critical parameter is the temperature that affects both biological H2 production efficiency and bacterial metabolisms in mixed cultures [34].

Table 2.1 H2 production rates of anaerobic reactors treating agricultural waste (adapted from ref. [35]) Maximum achieved Type of the H2 production rate Pretreatment process Reference substrate -1 (ml H2 g vs) Corn straw 68 1.5 MPa for 10 minutes [37] Corn stover 49 220 ℃ for 3 minutes [38] Grass silage 6 No pretreatment of the [39] feedstock Rice bran 61 Not determined [40] Wheat bran 43 Not determined [40] Rice 96 No pretreatment of the [41] feedstock Carrot 71 No pretreatment of the [41] feedstock Cabbage 62 No pretreatment of the [41] feedstock Chicken skin 10 No pretreatment of the [41] feedstock Egg 7 No pretreatment of the [41] feedstock Lean meat 8 No pretreatment of the [41] feedstock Food waste 60 Not determined [42] Food waste 77 No pretreatment of the [43] feedstock

2.1.2 Photo Fermentation Photo-fermentation is similar to dark fermentation; however, the main difference is that photo-fermentation is conducted in the presence of light. Photo-fermentation uses non- oxygenic photosynthetic bacteria that utilizes sunlight and biomass to produce hydrogen. The purple, non-sulfur and green sulfur bacteria are some of the well-known examples of photosynthetic bacteria [44].

Photosynthetic purple bacteria is considered as the most promising photobiological hydrogen yield method due to its high applicability under various conditions. Moreover, the process can utilize either industrial by-products or wastes from dark fermentation as organic acids, through nitrogenase enzymes [45–49]. Organic acids are utilized as electron donors in this process and transported to the nitrogenase, which has no nitrogen, and transfers the electron to one proton [50]. Strict control of environmental conditions is required to produce H2 efficiently via photo- fermentation. The optimal pH and temperature ranges are between 6.8-7.5 and 31-36 °C respectively [51]. Designing photo-fermentation reactors (photo-bioreactors) are challenging due to the light distribution, which constitutes the most critical parameter affecting H2 production rate. Thus, photo-reactors usually have a high surface area to enhance the light efficiency in the photo-fermentation process. A brief comparison of the solar-based bio-hydrogen production processes is presented in Fig. 2.2. Solar biophotolysis of water appears to be the cleanest method for biohydrogen yield. However, poor hydrogen production rate, oxygen inhabitation and strict light obligation of the process are considered as the main drawback of this approach [51– 53]. There have been numerous studies on photo-fermentation divided based on the process used; some of them are subsequently explained in this section. Continuous Photo-Fermentation: Most studies on continuous photo-fermentation process use a combination of photo-fermentation and dark fermentation. Theoretically, yields of up to 80% are reported depending on the substrate used during fermentation. The experimental periods are long and varied from 25 hours [54] to 120 hours[55]. Currently, sequencing dark/photo-fermentation processes are used to increase bio- hydrogen production yield (as seen in Eq.2.1 and Eq.2.2). Both hydrogen and organic acids are simultaneously formed during the dark fermentation process. These low molecular weight organic acids cannot be consumed while this process is taking place. The fatty acids produced in the first stage will be removed when photo-fermentation is used as a second step process, and accordingly, the production yield of hydrogen will be increased. The two most important advantages of using this combined system are the increase of the theoretical hydrogen yield to 12 molH2 / mol substrate and provide the high-performance wastewater treatment:

First step: Dark fermentation C6H12O6 + 2H2O → 2CH3COOH + 2CO2 + 4H2 (2.1)

Second step: Photo-fermentation 2CH3COOH + 4H2O + light → 4CO2 + 8H2 (2.2) Batch Fermentation: Especially in the past decade, batch photo-fermentative hydrogen gas production from different substrates, has been studied extensively. The majority of the studies in the open literature use pure carbon sources in sterile and nutrient- rich fermentation media for batch experiments. [51]. Glucose was one of the nutrients investigated with R. sphaeroides as bio-catalysts by Fang et al. [56]. Almost all the glucose was turned into hydrogen, volatile fatty acid (VFA), and CO2, though the hydrogen production was measured at 6.6% of theoretical. Special living organisms

PHOTO- Light requirement Fermantation

Strict operation

conditions Bio Hydrogen Hydrogen Bio

– Relatively slow Production

Solar Sensitve to O Biophotolysis 2 formation

Light requirement

Fig. 2.2. Comparison of the solar based bio-hydrogen production processes (modified from ref. [51])

Researchers have studied various operating conditions to obtain better hydrogen yields at higher light efficiency. Basak and Das [57], used an annular photobioreactor to enhance H2 production through enhanced light penetration. Eroglu et al.[58] used a temperature controlled flat plate solar bioreactor. The highest yield of hydrogen was achieved by using R. sphaeroides-RV strains, and the optimum sugar concentration was found to be 5g/L [59].

Fed-batch photo-fermentation: In Fed-batch applications, one or more substrates are supplied to the reactor during the cultivation process and wherein the products remain in the reactor until the end of the experiment [60]. Fed-batch offers an advantage over the two other types, but it has not been investigated as much as other methods. Yields of up to 80% theoretical have been reported using the fed-batch process by the researchers [56]. Semi-continuous photo-fermentation: That kind of processes can be operated at different times in a day with various time intervals. Wu et al. [60], built an experimental setup and investigate the influence of the parameters such as the initial feeding time of the reactor, the time interval between feeding times and the feeding volume ratio. They obtained the optimum operating conditions for their photo-fermentation reactor as being feeding volume ratio of 40%, the initial feeding time of 3 days, and interval feeding time of one day. In order to better understand the photofermentation process and estimate the possible outputs of scale-up applications, building a mathematical model is essential. In this regard, Akbari and Vaziri [61], introduced a new modified mathematical model to predict the specific growth rate of bacteria, substrate consumption and H2 production rates by the photo-fermentation process. Numerous combinations of the proposed models were applied, and experimental data validated forecasts of the models.

2.1.3 Microbial Electrolysis Cells Microbial Electrolysis Cells (MECs) involve the production of hydrogen gas using biomass and wastewater and other renewable resources [62,63]. A microbial electrolysis cell is a hydrogen production technology, which is related to microbial fuel cells (MFC). While MFC generates electricity via the microbial decomposition of organic matters, microbial electrolysis cell partially reverses the process for hydrogen manufacturing from organic substances. The process is driven by electricity [64]. MECs use electrochemically active bacteria to oxidize organic matter and generate

CO2, electrons, and protons. To produce hydrogen, the reactor needs to be fed by an external electrical potential between 0.2V to 0.8V under pH = 7, T = 30 °C and P = 1atm. These conditions allow for the bacteria to transfer electrons where they travel to the cathode and combine with the free protons in the solution to form hydrogen [65].

There are various reactor designs used to achieve sustainable hydrogen production levels. These reactors mainly classified according to the number of the chamber in the reactor.

Two-chambered MECs: A Two Chambered MEC typically consists of an organic matter anode, a cathode and a membrane used to prevent microbial consumption of hydrogen (as seen in Fig 2.3). Cation change membranes are used in MEC. The generated ions during the fermentation cause a concentration polarization on the membrane surface and depending on the current density on the surface; the proton diffusion rate decreases with the reactor performance. Many variations have been tested by researchers on the two chamber MEC concept to improve the efficiency of the reactor [66–68]. The results in the literature show that optimum operating temperature is around 30oC. Also, lower pH levels in the cathode of the chamber enhance the hydrogen production. In order to decrease the supplied electric current to the system, an acidic medium could be prepared for the cathode chamber [68]. + -

Organic matter

2H++2e-

+ + - H CO2+H +e H2

Anode Membrane Cathode Fig. 2.3 Diagram of a conventional two-chamber microbial electrolysis cell (modified from ref.[65]) Single Chamber MEC: Single Chamber MECs work on the principle of hydrogen being insoluble in water, high hydrogen production rates and slow microbial conversion of hydrogen to methane. The

19 largest source of hydrogen losses in this type of reactor design is due to hydrogen losses to methanogens [64,65,69,70]. The recent studies on single chamber MECs show that membrane free MEC designs can improve both high hydrogen recoveries and production rates. This is due to reduced internal resistance in the system, which means the applied voltage is used more to drive the reaction than to overcome the resistance [65]. Also, to reduce the potential losses related to membrane and improve the system efficiency, a new membrane free microbial electrolysis cell was proposed using large graphite brush anodes under short electrode distance. Since single chamber membrane-free microbial electrolysis cell eliminates the membrane, they are considered as compact and economical alternatives. However, one of the leading concerns about membrane free MECs is the microbial hydrogen losses to methanogens. Methanogens compete with electrochemically active microorganisms for both product (hydrogen) and electron donor substrate [65,71].

Ag/AgCl reference Gas collection electrode tube

Cathode Brush anode connection connection

Fig. 2.4 A single-chamber microbial electrolysis cell (modified from [72])

2.1.4 Electrochemical Disinfection Electrochemical disinfection is considered as a feasible alternative to conventional decentralized wastewater treatment systems. Electrochemical (EC) disinfection process

20 has been reported to be significantly effective in killing a broad spectrum of microbial pathogens in various water mediums [73–79]. Organic species are degraded biologically during the electrochemical disinfection process by several kinds of oxidants, which are manufactured by the electrolysis process. When chloride is available inherently as it is presented in seawater and sewage water or it is artificially added, reactive chlorine species (RCS) such as free chlorine and chlorine radical species are produced electrochemically and act as the main sterilizers.

Anode vicinity Bulk phase Cathode vicinity

O2 H2 Cl- Cl-

RCS RCS

- H+ OH pH pH Anode Cathode

Fig. 2.5 A wastewater electrolysis cell (modified from ref. [79])

Human urine can also be used as an electrolyte as it contains a high concentration of chloride [80], though presence of chlorine as mentioned earlier is also a cause for concern as excess levels can be toxic, as well as some of the byproducts formed when excess voltage is applied are incredibly toxic to both plants and humans [81]. Reactive oxygen species that include hydrogen peroxide and ozone generated during the electrolysis of water can improve the disinfection efficiency [82]. During the electrochemical disinfection process, water is forced to circulate in a disinfector, which consists of electrically charged electrodes. Fig. 2.5 depicts the scheme of the toilet water electrolysis cell, which is located at the California Institute of Technology, United States. The electrochemical disinfection system works inline, which is treating 20 L of wastewater each time in a batch mode. The supernatant of the sedimentation tank supplied the

21 wastewater for the EC reactor, which also acts as a simple anaerobic digester. The treated sewage water is sent back to the sterilized water storage tank for later use [79]. Electrochemical disinfection appears to be a promising approach to inactivate bacteria and viruses efficiently from the wastewater. On the other hand, it should be noted that the treated sewage wastewater is safe for only non-potable reuse. These reactors could be utilized as toilet flushing or agricultural irrigation [79].

2.1.5 Electrodialysis The electrodialysis process is a process, which is not directly related to hydrogen production as long as it is not combined with another process. It can be used before or after the PEC reactor to desalinate the electrolyte. Electrodialysis (ED) is a tertiary water treatment, which can be used to treat soluble, organic and inorganic substances [83]. Electrodialysis works with the utilization of a semi-permeable membrane that is under the influence of an electric current [84,85]. ED is an ion separation technique consists of an electric field and semipermeable ion-selective membranes as presented in Fig 2.6.

Treated water

EW EW

Anode

Feed water

EW Cation-exchange membrane

Anion-exchange membrane

Fig. 2.6 A simple layout of the electrodialysis cell. (modified from ref.[86])

Electrical current leads the migration of anions towards the anode (which is positively charged electrode) and cathode (which is negatively charged electrode). Subsequently, while the cations and anions free dilute stream is accumulated in the center, the concentrated (brine) streams are passed to the side cells. Even though the electrodialysis cell is a tool for freshwater production, it only can desalinate the water. In this approach inactivation of biological organisms is not possible [55]. Electrodialysis is usually operated in a continuous or batch mode, with two electrodes where the voltage is applied. To obtain the optimum degree of mineralization the arrangement of the membrane can be set as series or parallel [87,88]. The factors that affect the electrodialysis treatment include the following: [83]  pH  Temperature  Applied current  Properties of pollutants  The selective ion exchange membrane  Wastewater flow rate  Fouling  Scaling of wastewater  The number of stages and the configuration of the treatment. The cost of water treatment through electrodialysis varies widely from $15 to $400 per million liters of treated water and offers a reduction in the total dissolved solids [83]. Electrodialysis can be used to deionize water before sending it to a photoelectrochemical reactor for hydrogen production and organic matter degradation. The process can be performed in steps or combined in a single reactor. Electrolysis can also be used to produce hydrogen indirectly as described by Tufa et al. [89] which can provide a maximum yield of 44 cm3h-1cm-2 for the integrated system. The proposed electrodialysis setup in the study is coupled with an alkaline polymer electrolyte (APE) water electrolysis cell.

2.1.6 Electrocoagulation The electrocoagulation (EC) process can be defined as the process where the introduction of the coagulants and removal of the suspended solids, metals, colloidal material including

23 other dissolved solids from wastewater or water [90]. One of the additional advantages of the EC process is that the process can remove harmful microorganisms. The usual operational method of the EC operation is that a direct current is applied, and the surficial electrode is sacrificed, which dissolves in the solution. The desolation of the sacrificial electrode results in increasing the metal concentration in the solution, which later on precipitates as hydroxides and oxides. Some of the fundamentals on the EC process requires the understanding of the coagulation, which is a traditional physicochemical treatment via phase separation for the decontamination of the wastewater in order to reduce its effect on the environment when it is released [90,91]. A background history review of the EC method dates back to as early as 2000 BCE, EC was used in water purification and potabilization [90,92]. The formation and the aggregation of a colloidal system are what the process is based on, and in order to further enhance the coagulation coagulating agents can be used. In coagulation water treatment process, adding agents such as Al3+ and Fe3+ salts enhance and favors the formation of pollutant aggregates [90,93]. In general, the following are the most critical steps that take place during the EC treatment [90,94–96]  Electronic reactions  Destabilization of the pollutants  Formation of aggregates  Removal of coagulated pollutants by sedimentation  Electrochemical and chemical reactions promoting the catholic reduction Considering the advantages of the EC approach and comparing it with the physicochemical treatment of coagulation. Here is a list of the common advantages that has been reported by several authors [90,97–99]:  More effective  pH control is not necessary  The highly-pure electrogenerated coagulant improves the pollutants removal  The operating costs are much lower However, the advantages come with the disadvantages, which can be listed as follows:  The possible anode passivation

 The consumption of sacrificial anodes

2.2 Advanced Oxidation Processes for Wastewater Treatment

Ozone (O3) has been utilized as an oxidant for wastewater purification, chemical reagent and chemical substance in industrial applications for over a century. O3 is considered as one of the most potent oxidants and sterilizing substance within the conventional oxidants.

Table 2.2 Advanced oxidation processes (adapted from [100])

Type of Advanced oxidation process Process inlets

2+ H2O2/Fe Fenton

3+ H2O2/Fe Fenton like

TiO2/hv/O2 Photocatalysis

O3/UV

O3/H2O2

O3/US

H2O2/US

3+ H2O2/Fe - Oxalate

2+ 3+ H2O2/Fe (Fe )/ UV Photo-assisted Fenton

+2 Mn /Oxalic acid/ O3

Theoretically, O3 capable of oxidizing organic compounds to CO2, H2O and inorganic materials to their maximum stable oxidation states. However, in a real application, O3 is relatively selective in its oxidation reactions, and these reactions are slow for water treatment applications [101]. Therefore, O3 is introduced with H2O2 and UV to enhance the treatment level. The previous studies show that utilization of ozone/ultraviolet radiation and ozone/hydrogen peroxide has better results in terms of oxidation than the ozone itself. The processes which employ hydrogen peroxide, ozone and/or ultraviolet radiation to inactivate the organic substances in the wastewater is named Advanced Oxidation Processes (AOPs) [101].

AOPs can be categorized either as homogeneous or heterogeneous processes. Moreover, homogeneous processes also can be further classified under processes, which do not use energy, and processes that use energy (as seen in Fig. 2.7 and Table 2.2) [102]. In this section, due to their applicability with electrolyzers, homogeneous and electrical energy used AOP applications are described mainly following a study in [102].

Advanced Oxidation Processes

Homogeneous Heterogeneous Processes Processes

Without Catalytic Using Energy Energy Ozonation

O3 In Alkaline Ultraviolet Ultrasound Electrical Photocatalytic Medium Radiation (Us) Energy Energy Ozonation

Electro Heterogeneous O /H o O /UV O /US chemical 3 2 2 3 3 Photocatalysis Oxidation

Anodic H O /Catalyst H o /UV H o /US 2 2 2 2 2 2 Oxidation

Electro- O /H o /UV 3 2 2 Fenton

Photo-Fenton 2+ Fe /H2O2/UV

Fig. 2.7 Classification of advanced oxidation processes (modified from ref. [102])

2.2.1 Electrochemical Oxidation

As briefly stated in the previous sections, in electrochemical oxidation, migration of oxygen from the water/solvent to the compound that desired to be oxidized occurs under

26 either direct or indirect anodic oxidation processes. Utilization of the electrical power as a vector for decontamination is recognized as the primary characteristic of this method. The substance to be degraded reacts with oxidants that electrochemically generated in situ [102]. The electrochemical oxidation occurs in the following steps [103]: Cathode reactions + − O2 + 2H + 2e → H2O2 (2.1) Anode reactions + • − H2O → H + (OH )ads + e (2.2) • + − (OH )ads → (O)ads or O2 + H + e (2.3)

(O)ads + O2 → O3 (2.4)

2.2.2 Anodic Oxidation In this type of advanced oxidation process, the biological effluents are degraded by hydroxyl radicals, which are entirely produced in an anode from the process of water oxidation [102]. The anodic oxidation depends on the slow reaction of organic contaminants with adsorbed hydroxyl radicals at the surface of a high oxygen overvoltage [104]. Platinum, lead dioxide (it could be doped or undoped), doped tin dioxide, iridium(IV) oxide and, boron- doped diamond are the most common type of electrodes in these applications [104].

2.2.3 Electro-Fenton

In the electro-Fenton process, H2O2 is manufactured electrochemically by means of the cathodic reduction of dissolved O2 on a carbon electrode, as explained in the following reaction [105]: + − O2(g) + 2H + e → H2O2 (2.5) Fe+2 is usually added externally to the treated water to improve the oxidizing power of hydrogen peroxide [104]: 2+ 2+ • Fe + H2O2 → Fe(OH) + OH (2.6) Here, OH• is a non-selective and very active oxidant agent, which reacts with organics yielding dehydrogenated or hydroxylated derivatives, until their conversion into carbon dioxide and inorganic ions are attained [104]. One of the main benefits of this method is that reaction Eq. (2.6) is propagated from Fe2+ regeneration, which is mainly by the

27 reduction of Fe3+ ions manufactured in it with hydrogen peroxide. Nevertheless, Fe2+ ions can also be quickly destroyed by OH• leading to Fe3+ and hydroxide ion [104]: Fe2+ + OH• → Fe+3 + OH− (2.7) Usually, the electro-Fenton treatments are considered expensive, and in these applications, salt is added to enhance the electrical conductivity of the medium [102]. Therefore, to improve the performance of the process by regeneration of Fe2+ from supplementary photo- reduction of Fe3+ ions and decomposition of Fe3+ complexes, photoelectron Fenton process is developed. This method includes the simultaneous irradiation of the water with ultraviolet light to accelerate the oxidation process [104,106]: Fe(OH)2+ + hv → Fe2+ + OH• (2.8) One of the primary drawbacks in this application is strict pH requirement and sludge reforming which may lead to disposal problems [103].

2.2.4 Photocatalysis

Employing a semiconductor metal (for example TiO2) as a photocatalyst and O2 as an oxidizing agent is the basis of the photocatalytic process [106]. So far, several types of photocatalysts have been tested, and titanium dioxide appears to be the best option regarding the features of it as higher stability, decent performance and reasonable cost [100,107]. The photocatalytic process starts with when the photons reach to the photocatalytic surface and are absorbed due to electron-hole pairs [100]. − + TiO2 + hv → e + h (2.9) The vast amount of the reducing power of formed electrons in the reaction above allows them to reduce some metals and dissolved O2 in superoxide radical ion form. The remaining holes can oxidize adsorbed water or hydroxide to reactive hydroxide radicals as follows [103]: + • + TiO2(h ) + H2Oad → TiO2 + HOad + H (2.10) + − • TiO2(h ) + HOad → TiO2 + HOad (2.11) Also, part of the adsorbed compound can be oxidized directly via electron migration as follows [103]:

+ •+ TiO2(h ) + RXad → TiO2 + ROad (2.12) However, it results in a decline in the quantum yield due to the recombination of the quantity of electron-hole pairs. Despite the many research towards the photocatalytic method, the photocatalytic process has not commercialized sufficiently [100].

2.3 Thermoelectric Generators Thermoelectric generators (TEG) are semiconductor-based devices, which can convert heat to electricity by Seebeck effect and vice versa. As they have a straightforward structure without any mechanical part, they are quite suitable for small-scale applications. Furthermore, they have a low maintenance cost and no operational expenses. Thermoelectric generators are also environmentally friendly devices, as they do not emit any of greenhouse gasses. R

QL Heat Sink

P-type N-type

Heat Source

Fig. 2.8 A simple layout of a thermoelectric generator

All of those aforementioned specifications of TEGs attracts the attention of researchers. Despite all the advantages TEGs have, TEG technology could not reach the sufficient level. The main reason for that situation is their relatively low efficiencies. Additionally, as thermoelectricity is still a developing technology; thermoelectric generator costs are higher compared with the conventional power generation systems. Despite the drawbacks, they can excite the industry‘s attention due to their advantages stated above [108,109]. Thermoelectric generators are more convenient for small-scale and decentralized applications where the installation of conventional heat engines are not feasible. For

29 instance, the exhaust gasses of vehicles are considered as unavoidable losses. Due to the space limitation of an automobile, it is quite hard to recover the waste heat of the engine with typical heat engines [108]. A thermoelectric unit can be used such cases. The simple sketch of a TEG is depicted in Fig. 2.8. Usually, a thermoelectric generator consists of two dissimilar thermoelectric materials as being negatively doped (n- type) and a positively doped (p-type). N-type and p-type thermoelectric materials are connected from their terminals. A thermoelectric could also be built from circuit consists of one type of thermoelectric material (n or p-type) and one electrically conductive material. However, such devices would not be practical as the thermoelectric potential is induced only from one side. The temperature difference between hot and cold sides of the thermoelectric materials creates an electrical potential which results in the electrical current for the system. The performance of a thermoelectric generator can be assessed by the figure of merit ZT which is the function of thermal conductivity, Seebeck coefficient, electrical conductivity and temperature of the thermoelectric materials [108,110]. A detailed explanation of the figure of merit and heat to electricity efficiency of the thermoelectric generator is given in the analysis section.

2.4 Literature Review In this section, a literature review is presented about novel hydrogen production systems from wastewater, advanced oxidation systems, and photoelectrochemical systems. Recent advances in the solar-based trigeneration systems are also covered. Yang et al.[111] developed and analyzed a stable anaerobic fermentation process for a dairy product, cheese whey. They run reactors at two different operational modes of batch and continuous. Both systems were including mixed bacterial groups under mesophilic conditions. Experiments of the batch mode reactor were conducted to investigate the degradation rate and assess the performance of the H2 production from cheese whey wastewater. Various feed to organisms ratios were tested. Experiments of the continuous type reactor were performed to investigate the effect of operational parameters of hydraulic retention time, pH and organics loading rate on bio-hydrogen production. They have reported that pH range between 4 -5 was the optimum conditions for the cheese whey wastewater fermentation. Also, it was observed that batch and continuous reactions yield different types of products.

Lin et al. [112] operated an integrated dark fermentation and methane reforming system for bio-hydrogen production in an eco-friendly way. Beverage wastewater was selected as the substrate of the dark fermentation, and its effluent was fed to an anaerobic methane reformer reactor. In the experiments, a steel-silk catalyst which was consisting of a plasma-assisted methane-reformer was used. One of the benefits of this process was carbon monoxide, which is considered the most toxic gaseous for PEM fuel cells, not produced directly or indirectly during the reactions. Zhi et al. [113] developed and investigated a technical-economic model, based on the fermentative production of H2 to predict the possible improvement in a citric wastewater plant. Also, the proposed model was able to evaluate the size, annual cost, overall efficiency and capital and electricity costs under various formations. In a stand- alone scenario, the possible power generation from H2 was not enough for driving the system. Therefore a PV array is installed to meet insufficient load. The simulations indicated that the combination of the bio-hydrogen and solar was more costly than the anaerobic-anoxic-oxic process. Eroglu et al. [58] conducted an experimental study in which Rhodobacter sphaeroides O.U.001 were used for bio-hydrogen production from olive mill wastewater which was the sole substrate sources. The system was able to produce H2 under both anaerobic and light conditions. Rhodobacter can easily shift to the fermentative mode and generate organic acids for the cases when the illumination was less than the required level. When the lighting was adequate, the microorganisms switch their metabolism into the photo-fermentation mode, and organic acids were turned into H2 effectively. Mishra et al. [114] proposed a dark and photo-fermentation combined system to maximize the H2 production from palm oil mill wastewater. The first stage of the process was the degradation of the effluent by an anaerobic bacteria, Clostridium butyricum LS2 and the second stage process was performed by a photoheterotrophic bacteria, Rhodopseudomonas palustris. The substrate of the reactor was diluted with the wastewater for better illumination in the second chamber. Results showed that combined dark and photo-fermentation layout improved the system`s performance significantly. The integrated system witnessed almost four times increase in the H2 production and two times

31 growth in the COD based organic removal efficiency compared to the dark fermentation stand-alone system. Ajayi et al. [115] suggested a system for solar assisted microbial electrolysis cell. Dye-sensitized solar cell (DSSC) was used to supply essential electrical need by microbial electrolysis cell. The MEC consisted of a photo-electrode (dye-sensitized photoactive material coated conducting glass) and a counter electrode (Pt-coated conducting glass). The results showed that unlike the traditional MEC where external sources supplied the electrical energy, hybrid MEC– Dye-sensitized solar cell could harvest the extra energy for hydrogen yield directly from sunlight. The maximum cathodic recovery efficiency was obtained 86% in the batch experiments. Baeza et al. [116] constructed a laboratory scale MEC plant (130 L) based on a cassette configuration to produce hydrogen from urban wastewater. The experiments yielded reasonable results. Over 4 L hydrogen production per day with a gas purity of 95% was recorded. Cathodic gas recovery was obtained as 82%, and electrical energy recovery found to be 121%. In this method, it was possible to reach 75% of organic matter removal efficiency. Salgado et al. [117] conducted experimental studies about solar driven photocatalytic H2 production methods. Au/TiO2 utilized as the photo-catalyst in the setup. Various kind of effluent has been tested including domestic wastewater, industrial wastewater, and alcohol. Moreover, the effect of ionic strength on the production was investigated. Decontamination of the mentioned effluents has been performed under direct solar irradiation. Au/TiO2 as a photo-catalyst showed satisfactory results for H2 production. Results revealed that, while the rate of dissolved organics had a positive effect on hydrogen production, ionic strength negatively affected H2 yield. Formic acid as a waste had the highest hydrogen yield amongst the other options. He et al. [118] introduced a novel light-driven- microbial photo-electrochemical cell (MPC) in order to overcome the conventional MEC`s issues associated with high cost of cathode catalyst or external electrical power requirement which made them impractical. The MPC system contained a titanium dioxide photo-cathode and a microbial anode. Basically, the system was an integration of microbial anode with a photoactive cathode. In the anode side (anode), active bacteria produced electrons electrochemically, and then

32 electrons migrated to the cathode side. In the photoactive cathode side, TiO2 absorbed the photon to generated electrons and holes at TiO2`s conduction and valence band respectively. The study showed the promising potential of MPC systems for future applications. Huang et al. [79] designed and analyzed a system for concurrent electrochemical disinfection of domestic wastewater and hydrogen production. In their work, the treated domestic wastewater was planned to be reused for agricultural irrigation or toilet flush. Various types of microorganisms have been tested to assess the disinfection efficiency individually. The results showed that electrochemical disinfection combined wastewater electrolysis cell was able to inactivate both bacteria and viruses successfully without the need of electrolyte support. Photoactive BiOx/ TiO2 anode was used in this study to supply reactive chlorine species (RCS) which was the primary disinfectant, to the system. Kim et al. [119] introduced a novel three-chamber microbial system coupled with an electrodialysis chamber for simultaneous desalination and organic degradation. Exoelectrogenic microorganisms oxidized organic materials and migrated electrons to the anode side, while H2 was produced at the cathode side. This process assisted by the electrical potential. Salinity level decreased in wastewater via electrodialysis process. The reactor consisted of an anode, a cation exchange membrane, concentrated waste chamber, anion exchange membrane and a cathode respectively. Two membrane formation of the reactor allows the separation of ions.

Park et al. [120] introduced a new multi-layered BiOx–TiO2 electrodes for the oxidation of chemical pollutants coupled with the H2 production to achieve synergistic enhancement. The photoactive electrodes were consists of a mixed-metal oxide group consisting of one layer of TaOx–IrOx, BiOx–SnO2 as the second layer, and BiOx–TiO2 as the third layer deposited in a series on both sides of Ti foil. The main findings of the study were adequate. The oxidation of the organic matter was improved by 70% under the light. Solis et al. [9] came out with a novel idea by combining hydrogen production and electro-Fenton oxidation process in a PEM electrolyzer. While Fenton-type process decolorized the textile effluent in the anode side, hydrogen was also simultaneously produced on the cathode side. A PV array drove the developed system. Therefore the system simultaneously provides wastewater treatment, electricity production, and

33 hydrogen production. An electro- Fenton-like process drove the decolorization process. In +2 +3, order to conduct Fenton-type processes, Fe / Fe and H2O2 should be supplied to the waste. The suggested system provided these ions by sacrificing the stainless steel electrochemically. Hence, as long as hydrogen peroxide introduced to the system, the Fenton-like process was able to take place. Since Fenton reaction is more effective at acidic mediums, H2SO4 was selected as the electrolyte in their study. Tokumura et al. [121] developed a system that combines the photo-Fenton process with hydrogen production. The main outputs of the system were decolorized water, hydrogen, and electricity as in Solis et al.`s [9] study. However, in this reactor membrane was not presented. Unlike the Solis et al.`s study, the experimental setup of this research consisted of three adjacent reactors which were not directly connected to each other. While the first reactor produces energy and generates ferrous ions for the photo-Fenton process, the second and third reactors utilized these ions with the presence of hydrogen peroxide for the color removal process. The treatment works by acidic corrosion of iron electrode and hydrogen peroxide addition under ultraviolet illumination. Ultraviolet light leads the reduction of the complex intermediate of Fenton reaction, Fe(OH)2+ to Fe2+ and hydroxyl radical. To predict the outputs of the actual system a dynamic simulation model of the experimental layout was also built. They observed that color removal efficiency increased with both iron ion concentration and ultraviolet light intensity until H2O2 concentration reached a threshold value. The developed model match with the experimental values satisfactory.

2.5 Main Gaps in the Literature Various researchers have investigated individual methods of the hydrogen production and wastewater treatment applications to demonstrate each approaches positive and negative characteristics, technical barriers that need to be overcome before the approaches commercialize and to determine the technical requirements of the methods. In both photo and dark fermentation methods, oxygen has a substantial inhibitory effect. Moreover, the light conversion efficiency of the photofermentation is relatively low and dark fermentation products contain carbon dioxide which requires a separation process [44,122]. On the other hand, the electrochemical water splitting processes can supply high hydrogen purity. However, these reactions require relatively high electrical potential, which increases the

34 production costs of hydrogen. One way to reduce the energy need in the electrolyzer is replacing the anodic reaction of the O2 generation with an alternative reaction that requires less energy. It could be achieved by anodic oxidation of iron via electrogeneration. Fenton reaction in a PEM electrolyzer was tested only once in the literature before [9]. Promising results were obtained for decolorization of water. Several advantageous of Fenton- reactions have been observed as complete degradation of organic matters, quick breakdown depending on the hydrogen peroxide concentration, production of no harmful byproducts. However, studies on simultaneously Fenton-reaction and hydrogen production systems are limited in the literature. Photocatalysts also have been used in oxidation processes. Nevertheless, the utilization of ultraviolet radiation and photocatalyst is not sufficient only by itself. They should be considered in an integrated cycle with another advance oxidation process. Photocatalysts assisted photoelectron-Fenton process in a PEM electrolyzer has not been studied for both hydrogen production and water treatment before. More studies should be focused on photo-Fenton processes. Moreover, the synergistic effects of advanced oxidation reactions in a combination of TiO2 photocatalysis for hydrogen production and wastewater treatment applications are not studied in the literature before. The synergetic effect of advanced oxidation processes (AOP) such as Fenton, Fenton-like, photocatalysis

(TiO2/UV) and UV photolysis (H2O2/UV) should be investigated individually and in a combination of each other. Influence of operating conditions including pH level, type of the electrode and electrolyte and the UV light, on the performance of the combined system should also be studied to design more effective systems.

CHAPTER 3: SYSTEM DEVELOPMENT AND ANALYSIS

In this chapter, an integrated trigeneration system for producing hydrogen, treating wastewater and generating thermoelectricity is proposed and the system is explained in detailed. Moreover, four more alternative layouts are proposed and discussed comparatively.

3.1 System Description The proposed integrated trigeneration system is depicted in Fig. 3.1. The integrated system consists of a wastewater tank, a latent heat storage subsystem, a thermoelectric generator subsystem, a heat exchanger for wastewater preheating and TEG cooling, a photoelectron chemical reactor for hydrogen production and wastewater treatment, a reflective mirror, and pumps for working fluid circulation.

Solar Wastewater, Treated water irradiance Electricity

Solar light

Hydrogen tank

PEC reactor for H2 and wastewater H2O treatment H+

Mirror H2O

Heat Membrane H2O Exchanger

TEG Solar irradiance

TE G PCM PCM

Wastewater tank

Fig. 3.1 Layout of the system integrated the system.

Initially, wastewater is pumped from the storage to the heat exchanger where it is preheated before the electrolysis and oxidation process. That part of the system, which includes TEG, heat exchanger and molten salt reactor, increases overall the system performance in five-way:  First of all, elevating the temperature of wastewater rises the hydrogen production rate, Fenton process efficiency, and electrolysis efficiency.  Secondly, the wastewater rejects its heat to the TEG and increases the temperature difference between the top and bottom legs of the TEG, which results in a rise of the TEG electricity generation, and heat to electricity efficiency. TEGs are the devices, which make a direct conversion of the temperature difference to an electrical potential. For a thermoelectric generator, it is obligatory to have a temperature difference between the legs of the generator.  Thirdly, For the hot side, TEG is fed by the phase change material (PCM) which is heated by the concentrator, and the wastewater stream provides the cold surface. Utilization of PCM precludes to exceed the working temperature range of the TEG. Because focusing the concentrator directly to the top of the TEG can lead excessive temperature on the surface, which may cause a malfunction. Instead, by using PCM, operating temperature can be adjusted and maintained by only changing the mass or type of the PCM.  Forth, PCM supplies the thermal storage for the system and make possible of the trigeneration even after sunset.  Lastly, PCM top of the TEG also acts as a voltage regulator. Because during the day solar, irradiance varies, hence the temperature of the hot surface of the TEG changes without the presence of PCM. However, as the phase change of a substance occurs under constant temperature, PCM avoids the temperature fluctuations and therefore voltage variations generated by the TEG. After wastewater leaves the heat exchanger, it enters to the photoelectrochemical reactor where the water treatment and hydrogen production occurs. The reactor uses TiO2 as the photocatalyst. For better hydrogen production and oxidation, solar illumination should reach to the all photocatalyst surfaces. A quartz glass assembly with high optical transparency is used on the photoanode side of the reactor. A comprehensive parametric

37 study is conducted to see the effects of design parameters such as (wastewater type, pH. level, temperature, solar light) on the overall system energy and exergy efficiency. Fig. 3.2 shows the block diagram of the system. The study consists of experimental and theoretical parts. The photoelectrochemical (PEC) reactor is experimentally investigated with a lab- scale system while the TEG-PCM subsystem is theoretically analyzed. Energy and exergy analyses are conducted to assess the performance of the trigeneration system, and a comparative assessment is executed to investigate the effects of various system parameters on the system outputs. A numerical modeling is performed by the COMSOL Multiphysics software package [123] to obtain the electricity generation, temperature distribution and heat transfer characteristics of the TEG unit. All the other components of the integrated system are modeled and analyzed in the Engineering Equation Solver (EES) [124]. Hydrogen production, wastewater treatment, and electricity generation capacities of the system are then calculated. Wastewater treatment process is quantified by measuring the decrease in chemical oxygen demand (COD) level in the wastewater.

Theoretical part of the study Experimental part of the study

Fig. 3.2 Block diagram of the system integrated the system.

3.2 Alternative integrated system layouts In this section, four different alternative integrated systems are proposed for trigeneration. Alternative system 1 is presented in Fig. 3.3 omits the latent heat storage system and instead the system uses a voltage regulator and battery packs. Induced thermoelectricity varies during the day with the fluctuating solar irradiance. Hence, to regulate the output, a voltage regulator is used in this layout. Also, this layout uses a battery pack to store excess electricity to use for the periods when the solar light is insufficient to drive PEC reactor. Even though this system eliminates the thermal energy storage medium, it brings two more component. Furthermore, since the concentrated solar power is directly focusing on the TEG, it constrains the solar dish area. In order not to overheat the TEG device during the peak hours of solar intensity, solar dish area is needed to be adjusted accordingly.

Fig. 3.3 Illustration of the alternative integrated system layout 1.

Alternative integrated system layout 2 is presented in Fig. 3.4. Unlike the first two systems, this system is utilizing concentrating photovoltaics (CPV). Due to the concentrated light, CPV heats up considerably which limits its efficiency. The undesirable heat is discharged to wastewater in this layout. Similar to the other layouts, heated up wastewater increases Fenton process and electrolysis efficiency. The third alternative integrated system layout is depicted in Fig. 3.5. This layout can be considered as the combination of the first two alternative systems. A TEG unit is embedded into a CPV unit to utilize the potential of high-temperature of the CPV. The unit

TEG uses the CPV as the heat source, while the wastewater stream is working as the heat sink. The temperature difference between the two sides of the TEG induces the electricity. However, the generated thermoelectricity is not going to be as high as the proposed system in this study due to the lower surface temperature.

Fig. 3.4 Illustration of the alternative integrated system layout 2.

Fig. 3.5 Illustration of the alternative integrated system layout 3.

On the other hand, it should be noted that electricity production by the TEG, is not the primary source of this system. CPV is also providing the required electricity for the PEC reactor. Fig. 3.6 shows a system without a solar dish, thermal energy storage, and wastewater heater. A PV pack is producing electricity via photonic energy. Since the solar irradiance is not concentrated, less electricity output is expected in this system compared to the systems with CPV. Furthermore, since the wastewater heater is not used in this layout, less hydrogen production, less conversion efficiency and less COD removal efficiency is expected as a result. All four layouts can be used for trigeneration. However, those systems require extra electrical components such as a voltage regulator and battery which will bring additional maintenance cost with it. Hence, the main system configuration more convenient for decentralized applications.

Fig. 3.6 Illustration of the alternative integrated system layout 4.

CHAPTER 4: EXPERIMENTAL APPARATUS AND PROCEDURE

In this section, a brief explanation of the experimental set-up is given with the materials and methods used in the experiments. In this study, the experimental procedure consists of three main groups. The first group of experiments is TiO2 nanoparticle deposition and characterization of the coated and uncoated electrode samples with and without the presence of UV-light. The second part of the experimental system is the PEC reactor both for wastewater treatment and hydrogen production. Finally, the third part of the experimental system is the experimental analysis of open and closed reflux COD measurements to quantify the degradation of organics in the solutions. All experiments are conducted in batch type, in this thesis study.

4.1. TiO2 Sol-gel Dip Coating

In the present thesis study, a low-cost sol-gel method is preferred to deposit TiO2 nanoparticles on the anode for photoelectrochemical hydrogen production. The block diagram of the preparation methods for sol-gel applied by dip coating on steel anode can be seen in Fig. 4.1.

500 ml distilled or pure water

25 ml TTIP

5 ml Acetic Acid

3, 5 ml Nitric Acid

Heat 80 °C for 30 min

Mix for 2h

TiO NP sol-gel 2 Fig. 4.1 Block diagram describing the procedure implemented for preparing sol-gel applied by dip coating on steel anode

In this method, firstly 500 of deionized water is poured into a beaker. Then 25 ml Titanium(IV) isopropoxide [125], 5 ml of acetic acid [126] and 3.5 ml of nitric acid [127] are mixed into the solution. A magnetic stirrer with a resistant heater is used to heat the solution from ambient temperature to 80°C for 30 minutes. The prepared “sol” is continuously stirred for 2 hours at a constant temperature to form a homogenous solution at pH 1. In this study, a custom-made dip coating machine with adjustable withdrawal speed is used. The picture of the dip coating setup is presented in Fig. 4.2.

Fig. 4.2 Custom made dip coating mechanism with adjustable withdrawal speed

A series of pre-samples consisting of stainless-steel washer are coated using the custom-made dip-coater machine at room temperature and pressure to achieve better dipping frequency and annealing temperature. Finally, stainless-steel plate anode with an active area of 180 cm2 is coated at with a withdrawal speed of 2.5 mm/s and then dried for 1 hour at room temperature. The coated anode is annealed under temperature controlled furnace (as seen in Fig. 4.3) with an initial temperature of 50C to 500C by increments of

15 C/min and kept at the constant temperature for 30 min. After all, the photo-activity and hydrogen production performance are tested under solar simulator light [128] in 0.05 M sodium hydroxide (NaOH) electrolyte. The further information on coating process and the results were published in International Hydrogen Energy Journal and for further information, it can be examined [129].

Fig. 4.3 Temperature controlled furnace used for the annealing process.

4.2.PEC setup Photo-electrochemical system for simultaneous wastewater treatment and hydrogen production consists of mainly a PEC reactor, light source, power source, flowmeter, liquid column and wastewater tanks. The photo-electrochemical reactor experiments are tested under a solar simulator (artificial light). In order to understand the influence of design parameters, anode and cathode reactions and synergetic effects of advanced oxidation reactions on hydrogen production and wastewater treatment processes, several reactor formations are tested. The PEC reactor is investigated with various layouts such as a single compartment cell, three interconnected column, and proton exchange membrane electrolyzer. Moreover, different electrode pairs such as in the combination of a graphite

44 rod, TiO2 nanoparticle coated graphite rod, stainless steel rod, TiO2 coated stainless steel rod, nickel mesh, copper rod, large-scale TiO2 coated stainless steel plate and large stainless steel plate are tested through the experiments. COD removal and hydrogen production rates are monitored for alternative waste options such as Reactive Black 5 solution (textile waste), sucrose solution (food industry), potassium hydrogen phthalate (KHP) solution and tap water.

4.2.1 Single Compartment Cell PEC Reactor

For the first step of the PEC design, the experiments are conducted in the rectangle batch reactor with a volume of 30 L. TiO2 deposited stainless steel anode with an active area of 180 cm2, and a graphite rod cathode are used as electrode pair in the reactor. Sol-gel dip coating procedure is applied to the deposit titanium dioxide nanoparticles on the surface of photoanode for simultaneous hydrogen production and wastewater treatment. The photoactivity of the photoanode is tested under solar simulator light [128] in 0.05 M sodium hydroxide (NaOH) electrolyte. Further information on the single cell reactor experiments can be found in the reference [130]. Electrolyte preparation of Single Compartment Cell PEC Reactor

Reactive Black 5 dye [131] (RB5, CAS # 17095-24-8) and table sugar (sucrose, C12H22O11) are dissolved in water to simulate the textile and food industry wastewater respectively.

0.025 M Na2SO4 is added to the solution to ensure the electrical conductivity of the solution is sufficient while the synthetic wastewaters are getting prepared. The pH value of the synthetic wastewaters and tap water are adjusted to 3 by adding concentrated H2SO4. 200 mg/L hydrogen peroxide (50% H2O2, CAS # 7722-84-1) is added to prepared wastewater to form Fenton reagent [130]. Experimental setup of Single Compartment Cell PEC Reactor

All the experiments are conducted in the rectangle batch reactor. TiO2 nanoparticle deposited photo-anode, and graphite cathode is selected as the electrode pair. The solar simulator (irradiance of 600 W/m2) illuminates the reactor to increase hydrogen production efficiency and produce Fe2+/Fe3+ couples from photo-anode. Electrolytes which are RB5, sugar and tap water are treated by using the photo-assisted electro-Fenton process (PAEF) to generate hydrogen gas. The potentiostat with a booster [132,133] is used to monitor, record and provide electric charge required for the hydrogen production and anode iron

45 oxidation. The photoelectrochemical reactor is used for the in-situ production of hydrogen and Fenton reagent is presented in Fig. 4.4. The PEC reactor is located under the solar simulator to illuminate the area of the anode with a surface area of 180 cm2. The working and working sense wires of the potentiostat form working electrode, which are connected to the graphite rod (cathode), while the counter and counter sense form the counter electrode which are connected to the TiO2 deposited stainless steel (photoanode). The cathode is placed into a glass chamber to collect the yielded hydrogen gas. After the experiment, a test kit is used [134] to detect the iron ion concentration in the wastewater to obtain the sacrificial rate of the anode [130].

Fig. 4.4 Experimental setup of the single compartment cell PEC reactor

4.2.2 Small Scale PEC Reactor A modular, small-scale PEC reactor is developed and built to comparatively investigate the synergistic effects of advanced oxidation reactions in a combination of TiO2 photocatalysis for hydrogen production and wastewater treatment applications. An experimental study is conducted with a photoelectrochemical reactor consisting of three interconnected columns and a graphite cathode, a photoanode and a UV- light source for simultaneous hydrogen production and wastewater treatment applications. The synergetic effect of advanced oxidation processes (AOP) such as Fenton, Fenton-like, photocatalysis (TiO2/UV) and UV photolysis (H2O2/UV) investigated individually and in a combination of each other. Similar to the previous setup, Fenton type reagent in the reactor is formed by anodic sacrificial of stainless steel electrode with the presence of H2O2. The impact of various parameters including pH level, type of the electrode, electrolyte and the UV light on the performance of the combined system are also investigated experimentally.

Graphite Photo Cathode Anode

UV light source

Potentiostat

Fig. 4.5 Experimental setup with three interconnected columns used in small-scale tests.

The batch experiments are conducted in a reactor with three interconnected column as depicted in Fig. 4.5 to investigate anode and cathode reactions separately. A single column type reactor is also used during the small-scale experiments.

Four different anodes are used in the experiments such as stainless steel, graphite and titanium dioxide nano deposited stainless steel and graphite rods. Sol-gel dip coating method is used for TiO2 coating on stainless steel. Experiments are executed under two different wastewaters. Reactive Black 5 dye [131] (RB5, CAS # 17095-24-8) is dissolved in water in order to simulate the textile wastewater. Decolorization and COD degradation is monitored in these experiments. Potassium hydrogen phthalate (KHP) solution [135] is selected as a colorless wastewater option. The photoactivity of the photoanode is tested under UV light. Induced photocurrents from the photoanode are measured by the potentiostat [132] in open circuit potential mode, and the photoelectrochemical process is assisted electrically by the potentiostat. Experiments are conducted under potentiostatic and galvanostatic mode. A UV- spectrophotometer is used to monitor the decolorization of the samples by scanning the optical absorbance values within the range of 340 nm to 900 nm. A Multiparameter Bench Photometer [136] is used to measure the COD values before and after the experiments. The closed reflux method is used for COD measurements [137]. To ensure the oxidation in the reactor is going under whether Fenton or Fenton-like processes, Fe2+ and Fe3+ ion sacrificial rates are also obtained by a spectrophotometric method. Further information on the small-scale reactor experiments can be found in the reference [138]. Electrode preparation of Small Scale Reactor In the present study, four different anode option is tested such as graphite rod, stainless steel rod and semiconductor coated stainless steel and graphite rods. TiO2 is selected as the photocatalyst due to the high corrosion resistant nature and the ability to form hydroxyl radicals in the water when the photons hit the coated surface. [100,107,139]. A low-cost sol-gel method which is explained in section 4.1 is utilized to deposit TiO2 nanoparticles on the stainless steel anode for photoelectrochemical hydrogen production and wastewater treatment [138]. Electrolyte preparation of Small Scale Reactor Reactive Black 5 [131] (RB5, CAS # 17095-24-8) and potassium hydrogen phthalate (KHP) solutions [135] are selected as wastewaters respectively. Molecular structures of the RB5 and KHP are depicted in Fig. 4.6 and Fig. 4.7 respectively. During the preparation of the textile waste (RB5 solution), 50 mg of RB5 dye is dissolved in 500 ml of deionized

48 water and pH is raised to 11.5 with the addition of NaOH. The prepared 100 mg/L RB5 dye solution is heated up to 80 ℃ and stirred for 4 hours to complete hydrolysis of the RB5 dye [140]. Later on, 20 g/L Na2SO4 is added to the solution to ensure the sufficient electrical conductivity. During the first set of experiments pH level is kept as 10.5, and the second set of experiments it is dropped to pH=3 with the addition of H2SO4 to observe the effect of pH on the reaction [138].

Fig. 4.6 Structure of potassium hydrogen phthalate (KHP).

The KHP solution is prepared by mainly following the methods of the Standard Methods Committee [135], the concentration of the KHP is set twice as the standard COD solution. 850 mg of KHP is dissolved in deionized water and diluted to 1 liter. Potassium hydrogen phthalate theoretically has a COD of 1.176 mg O2/mg [141] therefore this solution has a theoretical chemical oxygen demand of 1mg O2/mL [135]. Similar to the RB5 solution, pH of the KHP solution is adjusted as pH=3 with the addition of H2SO4 and electrical conductivity is increased by dissolving 20 g/L Na2SO4. For the cases where hydrogen peroxide is used, the concentration of H2O2 is set as 200 mg/L and 400 mg/L [140].

Fig. 4.7 Structure of Reactive Black 5 (RB5) dye.

Analytical Methods Decolorization of the textile wastewater is monitored before, at the end of the experiment and after 24 hours with a UV-spectrophotometer. The photometer scans the optical absorbance of the samples from 340 nm to 900 nm wavelength. Moreover, the optical absorbance at 436, 525 and 620 nm wavelengths are measured according to the German wastewater discharge standards [142]. COD is measured by the colorimetric closed reflux method [137]. Mercury (II) sulfate, sulphuric acid, potassium dichromate used as the digestion solution, while sulphuric acid and silver sulfate is used as the sulphuric acid reagent [143]. Ferrous and ferric ion concentrations are measured for the experiments with coated and uncoated stainless steel anodes. Determining the type and concentration of iron ions are crucial since it directly affects the AOP`s characteristics. Ferrous ions with hydrogen peroxide compose the Fenton reagent, and Ferric ion with hydrogen peroxide takes place in Fenton-like reaction. A calibration curve is plotted by using iron ion reagents [144] and optical absorbance values of the samples at 510 nm. Iron sulfate and the sulphuric acid solution are used in the benchmark experiments. After plotting the calibration curve, iron sacrificial rates are determined for the stainless steel electrodes.

4.2.3 PEC Reactor with Membrane Electrode Assembly The photo-electrochemical system for simultaneous wastewater treatment and hydrogen production system consist of mainly a PEC reactor, light source, power source, flowmeter, liquid column (manometer) and a quartz glass case. The Photo-electrochemical reactor experiments are tested under solar simulator (artificial light) illumination. The experimental configuration is presented in Fig. 4.8.

The assembly of the PEC system is depicted in Fig. 4.9. The photoelectrochemical reactor consists of Nafion membrane (PtB- IrRuOx), stainless steel cathode and TiO2 coated photoanode and a stainless steel sacrificial anode. The reactor casing material is selected as high-density polyethylene (HDPE) due to its strong moisture resistance. In addition, HDPE is also resistant to most of the chemicals, flame retardant and machinable [20].

Fig. 4.8 The PEC reactor for simultaneous hydrogen production and wastewater treatment.

Unlike the conventional PEM electrolyzer configurations, this system has two parallel anode plate as shown in Fig. 4.10. Since it is observed in the first set experiments that TiO2 deposition constrains the iron sacrificial, a stainless steel anode is added to the reactor. TiO2 deposited photoanode is located after the quartz glass to ensure the sufficient amount of UV light reaches the photoactive surface. A sacrificial anode is installed close to the proton exchange membrane as similar to the conventional PEM electrolyzer designs to provide electrical conductivity. The Nafion membrane consists of two dissimilar electro-

51 catalysts coated surfaces. Iridium ruthenium oxide and platinum black are coated on anode and cathode sides respectively. The electro-catalysts of Nafion membrane has the same 2 density on both sides as 3 mg/cm . The specifications of the membrane are tabulated in Table 4.1. The produced hydrogen is measured by three different methods. The first method is connecting a flowmeter to the outlet of the cathode. Omega FMA-1600A flowmeter [145] is used in the experiments, which can measure the gas streams in a range of 0-100 SSCM. The second method is attaching a liquid column to the outlet of the cathode. By measuring the displacement level of the water column, hydrogen production in a time interval can be calculated.

Fig. 4.9 The assembly of the PEC reactor.

Finally, the third method is based on Faraday`s law in which the quantity of hydrogen production by the electrochemical process is associated with the electrical charge consumed by the electrolyzer. Each method has its own benefits and drawbacks. Faraday`s law can give the most accurate results as long as the current efficiency of the cell is obtained successfully. Measurements by the flowmeter have some challenges. Since the hydrogen is the lightest element on the periodic table, there would be possible losses due to hydrogen leaks through the fittings of the reactor. Also, it is also observed in the preliminary tests that, hydrogen can accumulate in local areas, which causes a pulse flow instead of

52 continuous flow. Hence, in this method, it is difficult to record instantaneous hydrogen production. However, the recordings can be logged for a sufficiently long time interval, and an average production rate can be calculated from it.

Table 4.1 Features of the membrane used in the PEC reactor [146] Properties Value Thickness 127 micrometers (5 mil) Basic Weight (g/m²) 250 43 (6.2) in MD, 32 (4.6) in TD - Method: Tensile Strength - max. (MPa) ASTM D 882 Non-Std Modulus (MPa) 249 (36) - Method: ASTM D 882 225 (MD), 310 (TD) - Method: ASTM D Elongation to Break (%) 882 Tear Resistance - Initial (g/mm) 6000 (MD, TD) - Method: ASTM D 1004 >100 (MD), >150 (TD) - Method: ASTM Tear Resistance - Propagating (g/mm) D 1922 Specific Gravity (23 °C, 50% RH) 1.98 Available Acid Capacity (meq/g) 0.90 min Total Acid Capacity (meq/g) 0.95 to 1.01 Conductivity (S/cm) 0.10 min

Hydrogen measurement via liquid column is based on the principle of the manometer. As the hydrogen is produced by the cell, its partial pressure applies force on the water column and displaces the water level. By measuring the water displacement, it is possible to obtain volume and partial pressure of the H2, and under the ideal gas assumption, the accumulated mass in the column can also be calculated. The volumetric collection method also provides hydrogen production values in a time interval. In order to get more accurate data in time, the change of the water column`s level can be recorded by a camera. Then, video recordings can be processed by a video analysis and modeling tool called Tracker [147]. The software can track displacements of the objects automatically by time. Also, it detects the location of the selected object by time and plots the displacement-time curves. As long as the calibrations of the measurements are done with an object of known length (such as a ruler), it also can provide accurate results. The interface of the software is presented in Fig. 4.11 In this method due to the partial pressure of H2 some part of it dissolves in the water. The dissolved gas is

53 proportional to the partial pressure, which is also known as Henry's Law. The dissolved amount of gas also should be taken into account in this method.

Fig. 4.10 Anode compartment of the PEC reactor.

Another challenge on the measurement methods is obtaining the Ferrous (Fe2+) and Ferric (Fe3+) ions precisely. Since these ions are not added to the cell externally but produced electrochemically by sacrificial of the anode, it is impossible to estimate the concentration of them without doing measurements. Iron ions could be measured in many different ways, however; every extra step for measurement of hydrogen yield brings more uncertainty with it.

Fig. 4.11 Hydrogen production measurement by manometer with video tracking.

A potentiostat is an electrical device that controls the electrical potential difference between an electrode pair of working and reference. The device can supply and measure the electricity concurrently. Most of the cases, controlled and measured parameters in a potentiostat are the cell potential and the cell current respectively.

Potentiostat Booster

Fig. 4.12 Gamry Instruments Reference 30k booster and Reference 3000 [132,133].

In this study Gamry 3000 high-performance Potentiostat/Galvanostat/ZRA [132] is used which has a maximum electrical potential and current of ±32 Volts and ±3 A

55 respectively (as seen in Fig 4.12). In order to provide additional supply to the potentiostat, Gamry Reference 30K Booster [133] is connected to the potentiostat which raises the current limit up to ± 30 Amps with compliance limits of +20 and -2.5 V [148]. Characterizations of photoanodes are performed with the potentiostat in various modes including linear voltammetry, cyclic voltammetry, potentiostatic, galvanostatic and open circuit potential. The potentiostat is also used for supply and record electrical input to the reactor. Pyranometer Solar irradiance at the surface of the Earth is mainly defined as the overall radiation across a wavelength range of 280 to 4000 nanometer, which is also classified as shortwave radiation. Total solar radiation, diffuse and direct beam are recognized as global shortwave irradiance which is expressed in watts per square meter [149]. The solar simulator illuminates the reactor in laboratory tests. The TSS-208 Trisol Solar Simulator from OAI Instruments [128] is used in the experiments. Pyranometers are the devices for measuring global shortwave irradiance. The pyranometer currently available in the CERL is a silicon cell based device as shown in Fig. 4.13 It is sensitive approximately 80% of the total shortwave irradiance which corresponds to a spectrum between 350-1100 nanometer wavelength. Nevertheless, the silicone cell could be calibrated to predict total shortwave radiation in the full solar spectrum as well [149]. In the experimental study, a pyranometer [150] is used to measure the solar irradiance level reached the surface of the reactor.

Irradiance Data cable sensor

Fig. 4.13 Pyranometer used in the study [150].

Thermocouples A thermocouple is a sensor device, which consists of two dissimilar metal leg as depicted in Fig. 4.14 and utilized for temperature measurement purposes. The thermoelectric legs are joint together from their ends creating a junction. Thermocouples mainly work with the same principle with TEGs. They create an electric potential by the temperature difference between legs, which means these devices do not directly measure the temperature but temperature difference. Therefore, each thermocouple set should be calibrated to a reference temperature. The most popular method is the cold junction for the calibration. Most of the modern data logger devices have their own calibration unit.

Positive and negative wires of the thermocouple PFA insulation

Fig. 4.14 Thermocouple wire used in the study [151].

There are several types of thermocouples, which are classified regarding their features such as operating temperature range, resistance to vibration, durability, resistance to chemicals, and application compatibility. The most common types of thermocouples are namely: Type J, K, T, and E, which are recognized as “Base Metal” thermocouples. For high-temperature applications, R, S and B type of thermocouples are preferable [152]. In this study, Omega TT-K-24, K type 24 AWG of diameter thermocouple wire [151] is used. Néoflon PFA insulation material is used to support endurance of the wire under relatively high temperatures. Thermocouples are used to measure the electrolyte temperatures in this study.

Data Logger A data logger is an electronic instrument that records the data measured by sensors (temperature, pressure, electric current, resistance and potential) over time. The data logger used in the experiments is presented in Fig. 4.15 In this study, OMEGA DAQPRO-5300 Portable Handheld Data Logger [153] is used to record and plot the temperature data by time.

Sensor input/Alarm output

Display screen PC USB connection socket

DC Power input

Fig. 4.15 Data logger used in the study [153]

The data logger has its internal memory which can store 512,000 sample, and the device has its own display screen for visualizing graphs and tables of the collected data. Therefore, it is not required for the device to be connected to the computer during the experiments. The data logger is also a highly responsive instrument, which can be able to do up to 4000 sampling in a second, which is necessary for the heat transfer calculations. Thermocouple legs are connected to positive and negative channels of the data logger from their ends. The data logger measures the electric potential between the ends of the thermocouple wires due to the temperature difference. Mainly, a data logger measures the

58 electrical voltage and calculates the surface temperature based on the type of the thermocouple wire and reference temperature.

4.3 Analytical Methods In this section analytical methods to determine chemical oxygen demand, decolorization and electro-generated iron ions are introduced. COD is measured by open and colorimetric closed reflux methods [137]. While mercury (II) sulfate, sulphuric acid, potassium dichromate used as the digestion solution; sulphuric acid and silver sulfate is used as the sulphuric acid reagent [143]. Iron concentration and decolorization assessments are performed based on spectrophotometric methods.

4.3.1 COD analysis

In the suggested three generation system the main outputs are electricity, hydrogen, and clean water. In the sections 4.1 and 4.2 measuring of electricity output and hydrogen production is described. In this section, quantifying the degradation of organic matters in the wastewater is explained. The chemical oxygen demand value shows the amount of required oxygen for the oxidation of all the organics in a solution in a chemical way. There are several ways of COD measurement. In this study, titrimetric open reflux and colorimetric closed reflux methods are explained. Standard Methods Committee`s experimental procedure [137] is followed in the experiments. The open reflux method is based on the principle of oxidation of the organics by hot chromic and sulfuric acids. High temperature and acidic conditions accelerate the oxidation. Potassium dichromate (K2Cr2O7) is the main reagent in this process. K2Cr2O7 is able to oxidize several types of organic matters almost entirely to CO2 and H2O. The wastewater sample, which is taken from the PEC reactor, is refluxed in the strong acidic solution with a known excess amount of K2Cr2O7. After digestion period, the remaining unreduced potassium dichromate is titrated with (NH4)2Fe(SO4)2·6H2O to determine the amount of potassium dichromate consumed, and the oxidizable substance is determined in terms of oxygen equivalent [135]. The experimental apparatus of this method consists of a magnetic stirrer with a heating plate, a glass pipet, a glass buret and 500 mL Erlenmeyer flasks. Moreover, the reagents

59 are a standard potassium dichromate solution, ferroin indicator solution and sulfuric acid reagent. The open reflux method is used only for the first set of experiments. The rest of the tests are measured by the closed reflux colorimetric method since it requires fewer chemical reagents compare to the open reflux method. The method is based on the principle of oxidation of the COD material the dichromate ion oxidizes, which leads a change in the state of chromium from the hexavalent (VI) to the trivalent (III) which are both colored 2- and hence absorbs the visible light. Dichromate ion (Cr2O7 ) has a poor optical transmittance in the spectral region of 400-nm, on the other hand, chromic ion`s (Cr3+) optical transmittance is much greater. Moreover, the chromic has a high absorbance in the 600-nm region of the spectrum, whereas the dichromate has almost full (=%100) transmittance. In 9M sulfuric acid solution, these chromium species have different molar extinction fractions. The test solutions used in this method includes a sample digestion solution and sulfuric acid. This mixture is heated in a block digester [154] (as seen in Fig. 4.16) preheated to 150°C and reflux for 2 hours in a protective shield. For the determination 3+ of COD values, increase in Cr ion is measured by a spectrophotometer equipped COD measurement device [136]. COD values are read before and after the experiments.

4.3.2 Iron Measurements and Decolorization Assessment

Decolorization of the textile waste is monitored before, right after and in 24 hour periods with a UV-spectrophotometer. The photometer scans the optical absorbance of the samples from 340 to 900 nm wavelength. Moreover, the optical absorbance at 436, 525 and 620 nm wavelengths are read according to German wastewater discharge standards [142]. The ferrous and ferric ion concentrations are measured for the experiments with coated and uncoated stainless steel anodes. Determining the type and concentration of iron ions are crucial since it directly affects the AOP`s characteristics. While ferrous ions with hydrogen peroxide compose the Fenton reagent, Ferric ion with hydrogen peroxide forms in the Fenton-like reagent. A calibration curve is plotted by using iron ion reagents [144] and optical absorbance values of the samples at 510 nm. Iron sulfate and the sulphuric acid solutions are used in benchmark experiments. After plotting the calibration curve, iron sacrificial rates are determined for the stainless steel electrodes.

(a) (b) Fig. 4.16 Block digester [154] (a) and COD and a multiparameter bench photometer [136] used in this study

4.4 Experimental Uncertainty Analysis Uncertainty analysis is one of the essentials of an experimental study as it shows the error band of the work. In an experimental study, there are two main sources of inaccuracies. The first source of error is a statistical fluctuation in the measured data due to experimental conditions (random error) and the second type of errors is a systematic inaccuracy which is usually coming from the measuring instruments (systematic errors). Random (precision) errors always show up in an experimental study, and they presumably have different values for the same repeated tests. The random errors cannot be controlled or avoided, but they can be approximated by comparing multiple measurements and their standard deviations. Conducting more experiments under the same conditions helps the better estimation of the precision errors. On the other hand, systematic (bias) uncertainties are consistent and repeatable errors associated with faulty equipment or flawed experimental design. Falsely calibrated equipment may lead to this error. In overall, the uncertainty of an experiment is a function of bias and precision errors, and it can be defined as follows:

2 2 0.5 Ui = (Bi + Pi ) (4.1)

61 where U, B and P represent the uncertainty of the variable, bias (systematic) and precision (random) errors respectively. For instance, an experiment where the independent parameters are electric potential (V), temperature (i) and volume flow rate (v) uncertainty can be calculated as follows:

∂i 2 ∂i 2 ∂i 2 B2 = ( ) B2 + ( ) B2 + ( ) B2 (4.2) i ∂V V ∂T T ∂v v After obtaining each parameter`s uncertainty by using the Eq. 4.1, overall uncertainty can be calculated from Eq. 4.2. Uncertainty propagation tool of Equation Solver (EES) is used to calculate the uncertainty of the experimental data. For further information about the method used by the software, NIST Technical Note 1297 [157] can be investigated. Table 4.2 shows the devices used in this study with their accuracy ranges.

Table 4.2 Accuracies related to measurement devices. Measurement Measurement Device Accuracy Parameter Range Electric ±0.2% of Potentiostat (Booster) [133] potential ± 32 Volts scale ±0.2% difference of reading Electric ± 1 mV ±0.3% of Potentiostat [132] potential ±11 V reading difference ±0.2% of Potentiostat (Booster) [133] Electric current ±30 A scale ±0.2% of reading ± 10 pA ±0.3% of Potentiostat [132] Electric current ±3 A range COD and Mulitparameter 0-1500 mg/l ± 15 mg/l ± 4% of mg/l COD Bench Photometer [136] COD reading Data Logger with K type Temperature -250 to 1200°C ±0.5% Thermocouple [153] Solar Simulator [128] Irradiance 800-1100 W/m2 ± 20.28 W/m2

Pyranometer [150] Irradiance 0-2200 W/m2 ± 5 %

Flowmeter [145] ±(0.8% of reading Volume flow 0-100 SCCM + rate 0.2% FS) pH meter [155] pH 0-14 pH 0.001 pH Spectrophotometer [156] Spectrum 340-1000 nm ±0.01 A at 0.3 A ±0.05 A at 1.0 A

CHAPTER 5: ANALYSIS AND MODELING

In this chapter, a detailed explanation of the methodology used in this study is presented. The thermodynamic analysis is conducted for both experimental and theoretical part of the study. All the thermodynamic quantities such as energy and exergy efficiencies and exergy destructions are calculated for all subsystems. A numerical model is built in the COMSOL Multiphysics software to obtain power generation, heat transfer, and temperature distribution through the thermoelectric generator. Another mathematical model is also developed for the PEM electrolyzer. An exergoeconomic study is executed to evaluate the economic aspects of the system. Furthermore, a multi-objective optimization method is employed to find the optimum operating conditions and corresponding total cost rates, sustainability index and production rates. Optimization, thermodynamic and exergoeconomic analyses are calculated in EES while the COMSOL Multiphysics software is used for the TEG calculations.

5.1 Thermodynamic Analyses The thermodynamic analysis of the proposed system is examined under five subunits including a solar concentrator, a latent heat storage, a thermoelectric generator unit embedded into a heat exchanger, and a PEC reactor. In this section, thermodynamic calculations of the sub-systems are introduced. First, the amount of solar irradiation reaching to the solar concentrator is calculated. Then the transferred heat from the solar receiver to the molten salt medium is determined. A numerical study is carried out to determine the produced thermoelectric power by the TEG. Also, heat transfer, power density, and temperature distributions on the TEG are modeled in this section. General balance equations (mass, energy, entropy, exergy) for the developed system is performed for all the system components. Mass balance equation can be shown for any elements of the system as follows: dm ∑ ṁ − ∑ ṁ = cv (5.1) in in out out dt

dm Here, cv represents the change of the mass in the control volume by time and ṁ dt demonstrates the mass flow rate of the working fluid. Energy balance equation for energy balance (EBE) for any system element can be defined as follows: ̇ ̇ Q − W + ∑ ṁ inhin − ∑ ṁ out hout = 0 (5.2) where Ẇ and Q̇ show the work rate and heat transfer rate which is passing through boundaries of the system and h indicates the specific enthalpy of the stream. General energy balance equation for any system element can be shown as follows [158]:

̇ Qi ̇ ∑ ∑ ̇ (5.3) Ex − ExWi + ṁ inexin − ṁ out exout − Exdi = 0 Here, ̇ and ̇ Qi indicate the exergy rate by heat and work respectively. ̇ is the rate ExWi Ex Exdi of exergy destruction by the system elements. For any components, specific exergies can be determined as follows: exi = hi − h0 − T0(si − s0) + exch (5.4) th exi shows the specific exergy of the i state point. Specific entorpy and enthalpy at the reference temperature T0 are shown as s0 and h0 respectively. The ambient conditions are set as the reference conditions. exch indicates the specific chemical exergy of the stream. The specific chemical exergy is determined only for the cases if any chemical reaction takes place [159]. In this study, the only chemical reaction occurs in the PEC reactor. Specific chemical exergy can be obtained as follows [158] exch = ∑ y푖 exch,i + R T0 ∑ y푖 ln 푦푖 (5.5)

In the calculation of chemical exergy, y푖 shows the molar fraction of the substance ‘i’, in the corresponding mixture. In this study, the thermodynamic model is built under the following assumptions [160]:  No chemical reactions take place except the PEC reactor reactions.  The kinetic and potential energy changes are neglected.  The heat losses through the PEC reactor are neglected.  The pressure drop and heat losses in through the pipelines are neglected. The COMSOL Multiphysics software package is utilized to model and simulate the non-isothermal and turbulent fluid flow in the thermoelectric generator where the thermoelectrical energy conversion occurs. The exit stream results (temperature, pressure,

64 specific enthalpy and generated electricity) of the thermoelectric generator is sent to the integrated system model, which is built in EES. Then, overall system performance is evaluated by using the Engineering Equation Solver. The overall energy and exergy efficiencies of the overall system are calculated, and the capacity of hydrogen production, water treatment, and electricity generation is determined. Moreover, a parametric study is executed to investigate the impact of the design parameters on the overall system.

5.1.1 Solar Concentrator

During the daytime, when there is an adequate amount of solar light, the transferred heat from the concentrator to the receiver can be defined as follows [161]:

Q̇ c = ηc × Ac × I (5.6) 2 where I and Ac and,ηc represents the solar direct normal irradiance level (W/m ), reflective 2 surface area(m ) and the efficiency of the concentrator respectively. The term of ηc takes into account all kind of deficiencies including shadow losses, dirt, and heat losses due to radiation. Moreover, in this study, heat transfer through the receiver surface is considered. Since the temperature at the focal point reaches elevated temperatures, radiative heat transfer losses are not neglected. Also, convection due to surrounding air flow is considered as heat losses. Since the heat losses due to conduction are minor losses compared to the convection and conduction losses, they are assumed to be zero.

Table 5.1 Design parameters of the solar field Parameters Value PCM tank base area 0.74 m2 2 Direct normal irradiance level (Iave) 7.43 kWh/m day Number of receivers 4 Concentrator efficiency 80% Wind speed 5 m/s Thermal conductivity of the insulation 0.045 W/m K material Insulation material thickness 5 cm

The heat losses in the receiver can be expressed as follows: ̇ 4 4 Qloss = ha A푟 (T푟 − T0 ) + σ ϵ A푟 (Tr − T0 ) (5.7)

65 where ε, Ar, and Tr stand for the emissivity, area and the temperature of the receiver cavity respectively. σ and ha show Stefan-Boltzmann constant and convective heat transfer coefficient of the ambient air respectively. ha can be estimated by the following empirical formula as [162]:

2 ha = 10.45 − va + 10 √va (W⁄m K) (5.8)

Here, va is the ambient wind speed in m/s. After defining all the losses, the heat transfer rate from the concentrator to the receiver cavity can be calculated as follows [163]:

Q̇ r = Q̇ c − Q̇ loss (5.9)

5.1.2 PCM Unit

In this study, a molten salt mixture consists of 58%LiCl-42%RbCl is selected based on its melting temperature and specific heat. When the solar light is available, the molten salt tank is charged by the concentrated irradiance. The properties of the molten salt used in this study is listed in Table 5.2. The heat rate transfer during the charging process of the PCM can be given as follows:

Q̇ r = ṁ cp,molte salt (T2 − Tm) + ṁ m ΔHm = ṁ (h2 − h1 ) (5.10)

Here, ṁ m shows the melting rate of the molten salt. ΔHm and Tm are the enthalpy of fusion and melting temperature of the molten salt respectively. Subscripts 2 and 1 indicate the outlet and inlet stream of the receiver respectively. In this study, the molten salt is selected only based on the thermophysical properties. In order to assess the molten salt`s viability on practical applications, the chemical properties also should be taken into account for the lifetime of the system and safety.

Table 5.2 Properties of the selected molten salt [124]. Compound Melting Density Specific Thermal Temperature, 훒 heat cp conductivity T (°C) 3 (kJ/kg (W/m K) m (kg/m ) K) 58%LiCl- 311 2500 0.89 0.39 42%RbCl (liquid) (liquid) (liquid)

5.1.3 TEG Unit The thermoelectric effect is a phenomenon which provides direct and reversible conversion between temperature and electrical potential. This phenomenon is known as “Seebeck Effect.” Usually, a thermoelectric material consists of two dissimilar semiconductor-based material as being negatively doped (n-type) and positively doped (p-typed) materials. Even though constructing a TEG with two similar type thermoelectric legs is possible, it would not be efficient. In recent years these devices are gaining popularity due to their favorable features such as having no moving part, emitting no greenhouse gasses, having straightforward structure, having low operating and maintenance cost. On the other hand, thermoelectric technology is still developing, and the efficiency of these devices has not reached the desired levels. They are not economically viable in all cases compared to the conventional power generation options. These devices are favored for the cases where a compact design is the required or the heat losses are unavoidable, and there are no other conventional ways to recover it. As it is explained earlier, each thermoelectric generator consists of two thermoelectric legs. In this study n-type and p-type Bismuth Telluride thermoelectric materials (TEM) are selected as the thermoelectric pairs. The performance of a TEG is assessed based on the modified dimensionless figure of merit ZT̅ which is a function of the thermoelectric properties of the TEMs and temperature. ZT̅ can be defined as follows [110,160]:

2 (Sp−Sn) T̅ ZT̅ = 2 (5.11) 1 1 k k 2 [( n)2+( p) ] σn σp

Here, subscripts n and p shows the type of the thermoelectric material. σ, k and S stand for electrical conductivity (S/m), thermal conductivity (W/m K) and Seebeck Coefficient (V/K) respectively. The maximum heat to electricity conversion efficiency of a TEG is directly related to the dimensionless figure of merit and the hot and cold side temperatures of the TEG, and it can be given as follows [164]:

Ẇ Th−Tc √ 1+ZT̅ −1 η = = Tc (5.12) Q̇ Th √ 1+ZT̅ − Th

Here, Q̇ and Ẇ show the heat transfer rate and thermoelectric power. Th and Tc stand for cold and hot side temperatures of the TEG. Also as it can be stated from that Eq. 5.11 and Eq. 5.12 that, a suitable thermoelectric device should have high electrical conductivity and Seebeck coefficient while it has low thermal conductivity. Moreover, in order to have a better thermoelectric yield, the system should operate under high temperatures with substantial temperature differences. In this study, the TEG unit is placed between the heat exchanger and the molten salt reactor as it presented in Figure 5.1. The temperature profile of the TEG along with the heat transfer characteristics should be determined to calculate the generated thermoelectricity throughout the TEG. Hence, a numerical model is built and analyze in the COMSOL Multiphysics software to calculate heat transfer, power density, and temperature distributions. [160]. A non- isothermal turbulent flow model is built in the software by applying k-ϵ turbulence model. Under steady-state conditions, the governing equations such as continuity and momentum can be given for the wastewater flow inside the TEG as follows [123,160]:

2 2 ρ(퐮 . ∇)퐮 = ∇ . [ (−p퐈 + (μ + μ )(∇퐮 + (∇퐮)T) − (μ + μ ) (∇. 퐮) − 퐈 ρk퐈)] + 퐅` (5.13) T 3 T 3 ∇. (ρ퐮) = 0 (5.14)

μT ρ(퐮 . ∇)k = ∇ . [(μ + ) ∇k] + Pk − ρϵ (5.15) σk 2 μT ϵ ϵ ρ(퐮 . ∇)ϵ = ∇ . [(μ + ) ∇ϵ] + Cc1 Pk − Cc2ρ (5.16) σϵ k k 2 2 P = μ [∇퐮: (∇퐮 + (∇퐮)T) (∇. 퐮)2] − ρk∇. 퐮 (5.17) k T 3 3 where k2 μ = ρC , C = 1.44, C = 1.92, C = 0.09, T μ ϵ c1 c2 μ

σk = 1, σϵ = 1.3 Here, I, F, P and μ show the identity tensor, body forces, pressure and the dynamic viscosity of air respectively. k and ϵ indicate the turbulent kinetic energy and the rate of dissipation of kinetic energy [160]. The structure and dimensions of the TEG module are presented in Fig. 5.1. A rectangular cross-sectional channel is placed at the top of the TEG unit, and the bottom

68 part of the TEG unit is connected to the PCM tanks surface to create and maintain the temperature difference between the sides of the TEG. In the numerical model, it is assumed that the hot side temperature of the thermoelectric leg is equal to the melting temperature of the selected PCM. The thermoelectric properties of n and p-type thermoelectric material are assumed the same (except Seebeck coefficients, which is negative for n-type). The properties are taken from the COMSOL Multiphysics library (As seen in Table 5.3). Further information on the TEG-PCM setup analysis in this study can be found in the reference [165]. Table 5.3 Properties of the thermoelectric materials [123] Thermoelectric k (W/m σ (S/m) μS (V/K) Material K) n & p type Bi2Te3 1.9 111,110 ± 192 (negative value for n, positive value for p type)

Water inlet TEG cold side TEG modules TEG hot side Water exit

T= constant = PCM`s Melting Temp 60 mm Copper walls 30 mm Fig. 5.1 PCM-TEG module used in this study.

5.1.4 PEC Reactor

In this study, a PEC reactor for simultaneous wastewater treatment and hydrogen production is used. The thermoelectric generator feeds the electrical load of the PEC electrolyzer. The TiO2 nanoparticles coated anode, and stainless steel cathode are suggested for the reactor. A brief description of the electro-Fenton reactions take place in the PEC reactor is presented in this section

The absorption of photons possessing energy equal to or higher than that semiconductor (3.2 eV for Titanium dioxide) can be shown as follows [166]: − + TiO2 + hν → TiO2(e ) + TiO2(h ) (5.18) The considerable reducing power of the formed electrons led a reduction in some metals and dissolved O2 in superoxide radical ion form while the remaining holes are able to oxidize adsorbed water or hydroxide to reactive hydroxide radicals as follows [100]: + • + TiO2(h ) + H2Oad → TiO2 + HOad + H (5.19) + − • TiO2(h ) + HOad → TiO2 + HOad (5.20) These reactions are vital for photo-oxidative degradation processes due to the high concentration of adsorbed water or hydroxide on the particle surface. Furthermore, some portion of the adsorbed compound can be oxidized directly via electron migration as follows [100]: + •+ TiO2(h ) + RXad → TiO2 + ROad (5.21)

In this study, in order to observe the effect of UV/TiO2 system on the wastewater degradation and hydrogen production, two other experimental studies are also conducted as being one without UV light, the other one without TiO2 coating. Reactions of an anode and cathode of a PEM electrolyzer for water splitting can be given as follows: + − H2O → O2 + 2H + 2e (5.22) + − 2H + 2e → H2 (5.23) This reaction requires 1.23V theoretically in acid mediums. However, the true cell voltage is higher such as 1.8-2.0 V. Hydrogen production is directly linked to the features of the

O2 evolution at the anode side. One of the methods to reduce the energy requirement in this process is replacing the anodic reaction of the oxygen generation with oxidation of iron ions electrochemically [9]. Fe − 2e− → Fe2+ (5.24) 2+ 3+ Electric potential at anode leads oxidation of Fe to Fe [9]. Fe2+ − e− → Fe3+ (5.25) The theoretical energy demand of the reaction Eq. 5.24 is ΔG0 = 74.2 kJ/mol which is 162.9 kJ/mol less than the oxygen evaluation reaction in water electrolysis.

The Fenton type reactions take place with the presence of H2O2 and iron ions in an acidic medium. Hydrogen peroxide reacts with the iron catalyst and generates hydroxyl radicals as follows [167]: 2+ 3+ • − Feaq + H2O2 → Feaq + HO + OH (5.26) 2+ 3+ where the addition (or generation from Feaq ) of Feaq increases the radical chain mechanism. 3+ 2 + Feaq + H2O2 → Fe − OOH + H (5.27) 2+ 3+ • − Feaq + H2O2 → Feaq + HO + OH (5.28) 2+ 3+ • − Feaq + H2O2 → Feaq + HO + OH (5.29) 2+ • + 3+ Feaq + HO2 + H → Feaq + H2O2 (5.30) 3+ • 2+ + Feaq + HO2 → Feaq + O2 + H (5.31) The chemical reaction rate of Eq. 5.26 is much higher than the corresponding reaction Eq. 5.27. It can be stated that Fenton reaction (Eq. 5.26) which has an excess amount of hydrogen peroxide is rate-limiting only for the initial time interval (in the order of seconds or minutes). After this short-term, the oxidation rate is controlled by the rate of reaction in Eq. 5.27 [167–170]: Any process which accelerates the reduction of ferric ions to ferrous can be utilized as a catalyst for the reaction chain above. In electro-Fenton applications, this is achieved electrochemically. One way to accelerate this reaction is irradiation by UV/V light as follows [171]: hν 3+ 2+ • + Feaq + H2O → Feaq + HO + H (5.32) The UV light can also be used in the process of hydrogen peroxide decomposition for hydroxyl radical production by the following photochemical reactions [172]: hν • H2O2 → 2 HO (5.33) • • H2O2 + HO → H2O + HO2 (5.34) • − • − HO + HO2 → HO2 + OH (5.35) − • H2O2 + HO2 → HO + H2O + O2 (5.36)

Fig. 5.2 shows the hydroxyl radical formation pathways in the PEC reactor. It is expected for the reactor to have all the AOP reactions shown in the Fig. 2.7. The produced

71 hydrogen is collected in the cathode column, and the produced hydrogen is measured with the help of volumetric methods and potentiostat together. The quantity of yield gas via an electrochemical process is directly associated with the electrical charge consumed by the cell, which is also defined by Faraday’s law.

1 N = ( ) q (5.37) n F N×M ṁ = (5.38) t Here, F, N, n, and q represent the Faraday constant (96485 A s/mol), electron migration per hydrogen production, hydrogen production (mol) and electrical charge (A s) respectively. t M ṁ and show the overall operation time of the cell (s), the molar mass of hydrogen (g/mol) and hydrogen production rate (g/s), respectively. The electrical charge can be obtained by determining the electrical potential of the cell. The required electrical potential supplied to the reactor, E can be obtained as follows:

E = Er + Eact + Eohm + Econc (5.39)

Here, E is the practical (actual) cell potential, Er is the ideal cell potential, Eact is the activation polarization, Econc is the concentration polarization, and Eohm is the ohmic polarization.

Fig. 5.2 Hydroxyl radical formation mechanisms in the reactors used in this study.

The losses associated with sluggish electron kinetics are calculated in the activation overpotential losses. Activation overpotential losses occur at both anode and cathode sides of the reactor, which can be given as:

Eact = Eact,a + Eact,c (5.40) R T j Eact,a = ln( ) (5.41) αan F j0,a

R T j Eact,c = ln ( ) (5.42) αcn F j0,c 2 2 where j shows the current density (A/cm ), j0 is the exchange current density (A/cm ) and R is the universal gas constant (J/mol K),.

Eohm can be identified as the resistance to the flow of protons in the membrane, which can be calculated as

Eohm = j Rohm (5.43)

Rohm = δm/σm (5.44) where δm and σm show the membrane water content and the conductivity of the membrane (Ω−1 cm−1) respectively. In an electrolysis cell, the required electrical power to drive the reactions is given as follows:

We = E x j x Acell x n (5.45)

Here, Acell is surface the area of an electrolyzer cell and n represents the total amount of cells in the electrolyzer stack. The performance evaluation of PEM electrolyzers can be made via several criteria. Among them, energy efficiency is one of the most commonly used assessment methods. One of the definitions of energy efficiency for the electrolyzer is the ratio of total energy consumption by the electrolyzer to the potential amount of heat released by the combustion of produced hydrogen. For a PEM electrolyzer driven by only an external electric supply, energy efficiency can be defined as follows:

HVH2 mH2 ηPEM = (5.46) We Another efficiency relation can be defined for the PEM electrolyzer based on the ideal and actual electric overpotentials to see the impact of the irreversibilities on the reactions:

E η = r (5.47) PEM,2 E The quantum efficiency of the PEC reactor can be defined as the ratio between the mole flow rate of the electrons migrated due to produce hydrogen and the mole flow rate of the photons [173,174]: 2ṅ ϕ = H2 (5.48) ṅ ph

Here ϕ is the quantum efficiency, ṅ H2 indicates the molar flow rate of the produced hydrogen and ṅ ph shows the molar flow rate of the photons which can be calculated as follows [173,174]::

λ2 λ Iλ ṅ ph ∫ dλ (5.49) λ1 ħ c NA

where λ is the wavelength, Iλ is the spectral irradiance, ħ is Planck’s constant, c is the speed of light, and NA is Avogadro’s number. The photonic energy and exergy content in the spectral range can be obtained as follows [174] :

λ2 Ė ph = ∫ Iλ dλ (5.50) λ1

λ2 To Ė xph = ∫ Iλ (1 − )dλ (5.51) λ1 Tλ Here,

5.33016 ×10−3 T = (5.52) λ λ

5.1.5 Overall Energy and Exergy Efficiencies The overall system performance is assessed by energy and exergy efficiencies. Thermodynamic properties as temperature, pressure, exergy destructions, specific enthalpy, specific entropy are obtained for all the system elements. The thermoelectric generator unit is modeled on COMSOL Multiphysics software program, and all the remaining elements of the suggested system is modeled in the Engineering Equation Solver (EES) software package. The overall system exergy efficiency can be defined by taking into account all the inputs and useful commodities as

∑Ė xdl ψ = 1 − ̇ (5.53) ̇ Qsolar Ex

where, ∑Eẋ dl shows the overall exergy destructions plus losses throughout the system. Since the all the exergy input is coming from the sun, it is used as the only system ̇ ̇ Qsolar input. Ex is taking into account both the irradiance reaching the PEC reactor surface and the concentrator. Since the fluid flow through the TEG unit is much higher than the other streams in the system, any fluctuations in the exit stream effects the system results considerably. Therefore, two different energy efficiency definitions are used in this study. In the first definition, the water is omitted from the efficiency definition to evaluate the effect of power and hydrogen production on the system performance more clearly. In the second definition, exit stream of the TEG unit is considered as a useful output since it can be used in space heating applications. The same approach is applied to the exergy efficiency. ̇ ṁ H2HVH2+Wnet,e +ṁ H2O(Δh) ηov = (5.54) Q̇ solar ̇ ṁ H2HVH2+Wnet,e ηov,2 = (5.55) Q̇ solar

5.1.6 Sustainability index Sustainability index is a tool to understand the performance assessment of a system, which shows how sustainable a system is in actual practice. Sustainability index can be considered as an indicator for the effects of potential improvements on the system for more effective utilization of the available feedstock. The definition of the sustainability index (SI) is given with the following formula [24]: 1 SI = (5.56) 1−ηex Here;

−1 Eẋ d SI = DP 푎푛푑 DP = (5.57) Eẋ 푖푛

DP represents the depletion factor, which is the ratio between the exergy destruction rate of the system and the exergy inlet rate to the system.

5.2 Exergoeconomic analysis In this chapter, a cost analysis method, which is also taking into account exergy destructions throughout the system, is introduced. Exergoeconomic analysis of the integrated system is carried out by combining numerical and experimental findings and

75 costs in the study. First, the costs of the selected equipment are listed as given in Table 5.4. Then, exergoeconomic analysis of the integrated system is performed. The costs are expressed in U.S dollars. The assembly of the PEC system mainly consists of a Nafion membrane (PtB- IrRuOx), stainless steel cathode, TiO2 coated photoanode and a stainless steel sacrificial anode. The reactor casing material is selected as high-density polyethylene (HDPE) due to its strong moisture resistance. HDPE is also resistant to most of the chemicals, flame retardant and machinable [20]. A thermoelectric generator, molten salt and heat exchanger are equipment to build the TEG and Solar-PCM sub-systems of the integrated configuration. In total $7,453 is needed to build the reactor for lab-scale testing. As it can be seen in Table 5.4, the highest expense share is coming from the TEG and Nafion membrane which corresponds to 40.2% and 26.8% of the overall capital cost respectively. The selected catalysts are responsible for the high price of the Nafion membrane. The electrode is designed considering to allow the produced gas discharge without accumulating inside the reactor. Such a design requires channels and detailed machining. Also, the acrylic with a machined hole is used to place a quartz glass in the center of the reactor window. Therefore machining cost of the overall setup is obtained as $500 for the system. The exergy cost for the streams in any cost rate balance can be expressed as follows [175]:

Ċ = c Ė x (5.58) where c stands for cost per unit of exergy ($/kWh) and Ė x stands for exergy rate (Ẇ ).

The exergy cost rates could be related to matter, electricity, and heat. It can be defined respectively as [17]:

Ċ m = cm Ė xm (5.59)

Ċ e = ce Ė xe (5.60)

T Ċ = c Ė = c Q̇ (1 − 0) (5.61) m h h T

The general cost balance equation can be defined as follows [20]

∑Ċ in + Ẇ cin + Ż = ∑Ċ out + Ẇ cout (5.62)

Table 5.4 Material costs of the integrated system. Material Quantity Unit Price Total rice Nafion membrane (with PtB and IrRuO2 coating) 1 $2,000 $2,000 PEC reactor case, High Density Polyethylene (HDPE) 1 $250 $250 SS Electrodes (Sacrificial Anode and Cathode) 2 $200 $400 SS Electrode plate (photoanode plate without coating) 1 $30 $30 Clear Acrylic glass for PEC reactor window 2 $60 $120 Quartz glass for PEC reactor window 1 $150 $150 Bolts, nuts, and washers 40 $1 $40 Gasket sheet 1 $20 $20 Sealants (silicone, resin, adhesives etc.) 1 $50 $50 Rubber piping 1 $20 $20 Machining (for reactor window, case and electrodes) 1 $500 $500 Sol-gel chemicals 3 $50 $150 Solar Concentrator 1 $123 $123 Thermoelectric Generator 60 $50 $3000 Heat exchanger 1 $200 $100 Molten salt 1 $300 $100 Support structure 1 $100 $100 Overall $7,453

The capital recovery factor (CRF) is another parameter for the cost analysis, which represents the ratio between a fixed annuity and present value of receiving that annuity for a certain length of time. CRF can be described as follows:

i (1+i)n CRF = − 1 (5.63) (1+i)n where i and n stand for the interest rate, and the number of annuities received respectively.

The component related cost (Ż ) takes into account life cycle stages and operation time of the components and can be defined as follows:

Z CRF φ Ż = i (5.64) i 3600 N

Here, N is the annual operating hours of the component, CRF is capital recovery factor, 휑 is for maintenance factor and Ż is the capital cost of component i. The total cost includes the capital cost (CC) and the operation and maintenance cost (OM). The relation between OM and CC can be defined as follows:

OM = CC OMratio (5.65)

The total capital cost of the construction can be defined as

TCC = CRF (CC + OM) (5.66)

The cost of destruction is calculated via exergy destruction rate of the component as

Ċ d = ci Ė xd (5.67)

Here, in order to take into account all kind of waste, Ċ d term also includes exergy losses. The total cost rate includes the component related cost and cost of losses.

Ċ t = Ċ d + Ż (5.68)

The total cost rate represents the annual expenses of the component, and for a better system, it should be minimized. Exergoeconomic factor is another parameter, which shows the contribution of the destruction in the total cost, and it can be expressed as

Ż f = (5.69) Ż +Ċ d

Table 5.5 The financial parameters of the proposed system. Financial parameter Value Interest rate 2% Lifetime of the components 10 years Electricity cost 0.10 $/kWh Thermal energy cost 0.04 $/kWh OMratio 45.5 Annual operational time of the system 3000 hours

The exergoeconomic parameters such as Ż and Ċ d are indicators of the importance of the component in the system optimization, whereas the parameter f gives a relative measure of the cost-effectiveness of the corresponding component [20]. The calculated results in the exergoeconomic analysis of the integrated system are presented in results and discussion section. The financial parameters, which are taken into account in this study, are listed in Table 5.5.

5.2.1 Scale-up analysis

An economic analysis is performed for a commercial scale hydrogen production and wastewater treatment plant which produces 1000 kg of hydrogen, treats 222 m3 wastewater daily. Since the mass production of the commodities cost less, the scale-up analysis can provide more realistic results to illustrate the viability of the proposed system.

Solar irradiance

TEG-PCM Power Plant

Electricity Electricity Hydrogen COD free wastewater Concentrated Light PEC Hydrogen Salt Electrodialysis Hydrogen storage production and Plant Wastewater treatment

Desalinated water

Wastewater Water storage

Fig. 5.3 The layout of the scale-up system.

The cost analysis is executed with the model developed by the U.S. Department of Energy, “Hydrogen & Fuel Cells Program, Future Central Hydrogen Production from

Photoelectrochemical Process” [176]. The related model analyzes the technical and financial aspects of central and forecourt hydrogen production technologies. The model uses a scenario for a photoelectrochemical solar concentrator system with heliostats/concentrators to focus the solar irradiance at a 10:1 intensity ratio onto multi- junction PV/PEC cell receivers which are immersed in an electrolyte reservoir and pressurized to 300 psi. The PV cells are in electrical contact with an electrolyte reservoir and treat wastewater on the anode side and produce hydrogen gas on the cathode side.

Table 5.6 The specifications and parameters of the scale-up system. Operating Capacity Factor (%) 95.0 Plant Design Capacity (kg of H2/day) 1,053 Daily Output (kg/day) 1,000 Yearly Output (kg/year) 365,128 Daily wastewater treatment (m3/day) 222

A standard discounted cash flow rate of return methodology is implemented which calculates the minimum levelized cost of the produced H2. The calculated cost of hydrogen also includes a specified after-tax internal rate of return from the production technology. Capital costs, operating and maintenance cost, capacity factors are also other main financial parameters affects the cost rate. The layout of the integrated system is presented in Fig. 5.3. The integrated commercial system is considered for a textile factory. The operating specifications and financial parameters of the system are tabulated in Table 5.6 and Table 5.7. The costs are expressed in U.S dollars. A post-treatment unit with a bipolar membrane assembly stack is built to desalinate the wastewater and produce acid and base from it. The acid and base outputs of the bipolar membrane electrodialysis are considered to be used in the Fenton type process. Since the textile wastewater consists high concentration salt and it is conductive enough, no external salt is added to waste before sending it to the PEC reactor. Hydrogen peroxide consumption rate is kept constant as the experimental study. Based on the experimental findings in this study, the required hydrogen peroxide to treat 222 m3 of wastewater is calculated as 88.7 kg H2O2. The bulk price of the H2O2 solution with a purity of 50% is considered through the calculations.

Table 5.7 The cost parameters. Specified Reference year 2010 Predicted start-up year 2020 Basis year 2005 Construction Period in years 2 1st Year Capital Spent in Construction 30% 2nd Year Capital Spent in Construction 70% Start-up Time in years 0.25 Specified Plant life in years 40 Specified study period in years 40 Length of the Depreciation Schedule in years 20 Depreciation Method MACRS Equity Financing 100% During the Start-up % of Fixed Operation Costs 100% Revenues 75% Variable Operating Costs 75% Specified portion of depreciable capital investment 10% Decommissioning costs Specified portion of capital investment 10% Salvage value Specified Inflation rate 1.9% Internal rate of return (after applying tax) 10.0% Taxes State 7.0% Federal 36.0% Total 40.48% Working Capital 15% Specified annual change in operation costs

The TEG unit is considered to provide the required electricity. The cost analysis part is not covered the costs related to TEG, electrodialysis, and storage units. Reference year is selected 2010 and the start-up year of the system is assumed as 2020. The prices of the system components are extracted from the model library. Some expenses are modified based on the scenario. The system is considered to work with an operating capacity factor 2 of 95%. The plant is thought to be constructed on a 0.1 km area. The specified plant`s lifetime is 40 years, and the financial analysis covers all the lifespan of the system. The length of the depreciation of the system is 20 years where it is calculated via the Modified

Accelerated Cost Recovery System (MACRS).

5.3 Optimization Study Optimization of an energy system consists of modifying the system layout and designing the parameters based on one or multiple design objectives. If the system performance is not only linked with a single parameter, a multiple parameter optimization will provide more accurate results to optimize the objectives. Maximizing the efficiency whereas minimizing the cost and environmental impact can be considered as an example of multiple objective optimization problems [177]. Hence as the first step of the optimization study, it is crucial to identify objectives, system boundaries, and environmental conditions. As the second step, optimization criteria should be identified such as energy, economic or environmental criteria which could also be the basis of assessment. Decision variable should be chosen as the following step. These variables have to be independent and address the related characteristics of the system in a suitable manner. Furthermore, their impact on the system should be disguisable from the parameters with less significance. Then as the next step, a mathematical model should be built which defines the effect of selected independent variables on the performance of the system [17]. There are various optimization methods presented which are suitable to solve multi-objective optimization problems. Genetic Algorithms (GAs) are considered one of the most promising methods [178]. In this thesis study, a genetic algorithm is used to execute the optimization analysis. The genetic algorithm is an evolutionary algorithm method which is inspired by the biological evolution [179–181]. In this thesis study, the genetic algorithm is executed due to the features of it such as - No requirement of an initial condition - Adaptability of working with multiple variables - Ability to find global optimum point - Ability to use populations - Utilization of objective function formation The GA method is considered as the one of the most robust technique as it can find the global optimum even though there are local optimum points.

Population Off Springs Decoded String New Generation Parents Genetic Population Operators

Manipulation Mates Reproduction

Genetic Operators

Fig. 5.4 Genetic algorithm cycle (modified from [182])

However, this process may take a longer time compared to the other optimization methods. GA tries to mimic the concept of evaluation. In the genetic algorithm, the solution vector (x) is defined as a chromosome, and groups of chromosomes form a population. Then, one of those population is randomly selected considering the data range and constraints. The chromosomes in the selected population are evaluated based on their fitness to the objective function. Then, a new population is formed in a stochastic manner from the selected chromosomes from the existing population. This process is called “breeding.”. The assets of the chromosomes which are successfully carried to the next population are categorized by encoded values of its independent variables. The method algorithm is presented in Fig. 5.4. The genetic algorithm based multi-objective optimization study is executed by using Engineering Equation Solver software. Three objective functions are considered here

83 for optimization: sustainability index of the overall system (to be maximized), COD removal efficiency (to be maximized) and overall cost rate of the system (to be minimized). The objective functions stated above can be defined as follows: 1 SI = (5.70) 1−ηex

Ċ Total = ∑(Ż i + Ċ d,i) = Ċ d,PCM−TEG + Ċ d,PEC + 푍̇PCM−TEG + 푍̇PEC (5.71)

CODwastewater−CODtreatedwater ηCOD = (5.72) CODwastewater Here, SI shows sustainability index which indicates how sustainable the proposed system is in real applications. SI is a function of the exergy efficiency of the system which is directly linked with the exergy inlet of system and exergy destruction and losses of the system (as seen in Eq. 5.55). Hence, the sustainability index is also an indicator of a system`s ability to utilize the feedstock to produce useful commodities. Ċ Total is the total cost rate of the system which represents the annual expenses of the system including the costs due to irreversibilities. ηCOD represents the chemical oxygen demand removal efficiency. The chemical oxygen demand is used to quantify the amount of oxygen, which is needed for the oxidation of all organic substances in water chemically and it is desired to be minimized in water treatment applications. The constraints that are considered in the optimization study are defined as:

 The reference environment temperature: 0 °C < Tref < 50 °C  The PEC reactor operating temperature: 10 °C – 80 °C.  The TEG heat – electricity efficiency: 0.01 – 0.15.  The average daily solar irradiance: 3 kWh/m2 8 kWh/m2  The Interest rate: 1% - 10%  The Annual operating time: 2000h – 4000h The constraints are set based on the limits of the operating conditions. The lower limit of the reference temperature is set according to the freezing point of the water at 1 atm pressure. The higher limits of the PEC reactors` operating temperature is set as the Nafion membrane`s maximum working temperature and daily solar irradiance is selected based on the solar energy potential of solar energy rich and poor regions.

CHAPTER 6: RESULTS AND DISCUSSION

In this chapter, the observed results are discussed and comparatively analyzed considering data from experimental, theoretical and numerical methods. The analysis sequence of the system presented under nine category is depicted in Fig. 6.1. First, TiO2 deposition on the stainless steel electrode and its characterization results are presented, and then small-scale PEC reactor is tested for, and synergistic effects of advanced oxidation processes on hydrogen production and synthetic wastewater treatment are presented. Thereafter, generated thermoelectricity and molten salt sub-system results are shown based on the numerical and analytical calculations. Then, photoelectrochemical reactor results are shown under various conditions. Based on all the sub-system results integrated system results are calculated. Optimization and exergoeconomic results are then presented to assess the system`s feasibility in different aspects.

TiO2 Nanoparticle Deposition via Sol–Gel Dip Coating Technique and Photoelectrode Characterization Study Results

Single Compartment Cell PEC Reactor

Synergistic effects of AOPs for hydrogen production and wastewater treatment processes

Thermoelectric generator

Molten Salt Storage

PEC Reactor with Membrane Electrode Assembly Results and Discoussion Integrated System Results

Exergoeconomic Analysis Results

Optimization Study Results

Fig. 6.1 The sequence of the results sections

6.1 TiO2 Nanoparticle Deposition via Sol-Gel Dip Coating Technique and Photoelectrode Characterization Study Results The procedure of sol-gel dip coating is explained in the experimental apparatus and procedure chapter. The TiO2 nanoparticle coated anode plate is tested for photoelectrochemical characterization in a solution of 0.05 M of NaOH under the light, no light, and chopped light condition. In order to compare the results with different counter electrodes, various types of metals are used. The surface area of the electrode in the study is 180 cm2. The solar simulator is utilized for imitating solar illumination to characterize the photoactivity of the formed semiconductor films. The picture of the large electrode before and after the annealing process is depicted in Fig. 6.2. A rainbow colored pattern is observed on the coated surface after the annealing process, which is indicating the formation of titanium oxide sol-gel, Ti(OH)4. The coated layer shows a similar physical characteristic to the previous applications presented in the literature [129].

(a) (b) Fig. 6.2 TiO2 coated stainless steel plate before (a) and after (b) the annealing process.

Initially, the metal plate is tested under a measured irradiance of 600 W/m2 in the solar simulator. An electrolyte solution of 0.05 M of NaOH is prepared. The working electrode is TiO2 coated photoanode and the counter electrode is a graphite rod. The solar

86 simulator light is shut off to obtain the no-light conditions, and the tests are repeated in sequence to avoid any external effect. The cyclic voltammetry (CV) is a method for potentiodynamic electrochemical measurement where the working electrode (TiO2 coated SS) ramped linearly with respect to time, and after the electric potential reaches its maximum set value, the working electrode is ramped back similarly. Fig. 6.3 illustrates the cyclic voltammetry sweep with an applied potential scan limit between -0.4V to 0.8V. The figure shows the current potential responses of the TiO2 coated metal plate to a cyclic potentiodynamic scan with a rate of 10 mV/s. Maximum 5mA of shift is observed in the cyclic voltammetry analysis. Essentially, the current is improved by the effective transference of the majority holes from

TiO2 towards the external circuit to combine with the majority electrons [129].

30 Graphite with no light Graphite with light

Current (mA)Current 0

-10

-20 -400 -200 0 200 400 600 800 mV (vs Ag/AgCl)

Fig. 6.3 Cyclic voltammetry graphs of the cell with TiO2 coated stainless steel working electrode and graphite counter electrode.

The same testing is repeated with the electrode pair of TiO2 and Cu. The working electrode is TiO2 photoanode, and the counter electrode is the copper rod in Fig. 6.4. The solar simulator light is shut off to observe the performance of no-light conditions, and the tests are repeated in a sequence to minimize any external effect. Since the copper electrode

87 has a higher electrical conductivity than graphite electrode, more currents are observed at the same corresponding voltage. However, The copper electrode shows weaker photoresponse than the graphite as a counter electrode [129].

7.00 6.90 6.80 6.70 6.60 6.50 6.40 6.30 Copper with no light Copper with light Current (mA)Current 6.20 6.10 6.00 350 355 360 365 370 375 380 mV (vs Ag/AgCl)

Fig. 6.4 Partial cyclic voltammetry graphs of the cell with the TiO2 coated stainless steel working electrode and the copper counter electrode.

8.00 7.80 7.60 Nickel with no light Nickel with light 7.40 7.20 7.00 6.80

Current (mA)Current 6.60 6.40 6.20 6.00 350 355 360 365 370 375 380 mV (vs Ag/AgCl)

Fig. 6.5 Partial cyclic voltammetry graphs of the cell with the TiO2 coated stainless steel working electrode and the nickel counter electrode.

The Ni and Cu electrodes witness the similar photoelectrochemical trend, which results in similar currents as depicted in Fig. 6.5. The comparison of all three counter electrodes is shown in Fig. 6.6. Under the same operating conditions, graphite induces more photocurrent when it is paired with TiO2 coated photoanode. At its maximum response, the electrode pair with graphite shows more than three times better photoactivity than the other alternatives [129]. Fig. 6.7 illustrates the potentiostatic scan of the graphite rode (counter electrode) with the TiO2 nanoparticle coated SS anode (working electrode) under the applied voltage of -1.4 V (vs. Ag/AgCl). It can be stated that the electrode pair of TiO2 coated SS and graphite has better anodic and cathodic response under the light and dark conditions, which corresponds to average current densities of 0.21 mA/cm2 and 0.26 mA/cm2 respectively. The photocurrent density is calculated as 0.05 mA/cm2 where Momeni and Ghayeb [183] 2 recorded a maximum of 0.039 mA/cm with the undoped TiO2 substrate.

30 Copper Nickel

20 Graphite

Current (mA)Current 0

-10 (d)

-20 -400 -200 0 200 400 600 800 1000 mV (vs Ag/AgCl)

Fig. 6.6 Cyclic voltammetry graphs of the cell with TiO2 coated stainless steel working electrode and various counter electrodes (vs. Ag/AgCl).

0 with light

-20 with no light

-40

-60

-80 Current (mA)Current -100

-120

-140 0 50 100 150 200 250 300 Time (s)

Fig. 6.7 Current vs. time graph with an applied potential of -1.4V (vs Ag/AgCl) of the TiO2 coated photoanode and graphite cathode under light and no light conditions.

Fig. 6.8 shows the cyclic voltammetry scan between -1.6 V to 1.4 V (vs. Ag/AgCl) of TiO2 coated SS/graphite electrode pair in 0.05 molar sodium hydroxide electrolyte. Under the light and dark conditions around 10 mA of current difference is observed which can be attributed to photoactivity of the TiO2. Linear sweep voltammetry (LSV) is a voltammetric method where the electric current at a working electrode (in this case TiO2 coated SS anode) is measured while the potential between the working electrode and a reference electrode (Ag/AgCl) is swept linearly in time. The oxidation or reduction of species in the electrolyte is recorded as their peak or trough in the current signal at the electric potential where the species starts to be oxidized or reduced. An LSV scan is conducted from 0.15 V to 1.4 V (vs. Ag/AgCl) at a constant scan rate of 10 mV/s (as seen in Fig. 6.9).

100

with no light 50 with light

-50 Current (mA)Current -100

-150 -2.00 -1.50 -1.00 -0.50 0.00 0.50 1.00 1.50 Voltage (V)

Fig. 6.8 Cyclic voltammetry results (-0.3 to 0.9 V vs. Ag/AgCl reference electrode) of coated steel plate and graphite rod counter electrode.

0.45 0.40 0.35 0.30

0.25 Photocurrent changes of 20mA due to light chopping 0.42 0.20 0.40

0.38 Current (A) Current 0.15 0.36

Current(A) 0.34 0.32 0.10 0.30 1.30 1.32 1.34 1.36 1.38 1.40 0.05 Voltage (V) 0.00 1.00 1.10 1.20 1.30 1.40 1.50 Voltage (V)

Fig. 6.9 Linear voltammetry sweep results with an initial potential of 0.15V (vs. Ag/AgCl) under the chopped light.

In order to see the current jumps due to photoresponse, the solar simulator light is shut off and turn on with a time interval of 10 seconds. A maximum of 10 mA of current difference is obtained between under the light and no light conditions. A potentiostatic scan is a method for electrochemical measurement where the power supply imposes a constant potential on the working electrode (TiO2 coated SS) for a certain time interval. Fig. 6.10 illustrates the photocurrent change over time under the applied potentiostatic voltage of 1.4 V (vs. Ag/AgCl). The hydrogen production is obtained by using a water displacement column. Initially, the produced hydrogen is collected from the cathode side, circulated through a closed vessel and it flows into the water column. Fig 6.11 presents the data for hydrogen production. After 20 minutes of operation of the reactor, the produced hydrogen is measured. 17 ml and 14 ml of hydrogen are produced in 20 minutes under the light and no light conditions respectively (in 0.05 M NaOH electrolyte solution under 1.4 V vs. Ag/AgCl).

130 with light 128 with no light 126

124

122

Current (mA)Current 120

118

116

114 0 200 400 600 800 1000 1200 Time (s) Fig. 6.10 Current vs. time graph with the potentiostatic applied potential of 1.4 V (vs. Ag/AgCl) of TiO2 coated photoanode and graphite cathode.

18 16 14 12 10

8 production (ml)production

2 6 H 4 2 0 with no light with light Fig. 6.11 The production yield under the potentiostatic applied potential of 1.4 V (vs. Ag/AgCl) of TiO2 coated photoanode and graphite cathode under light and no light conditions (V vs. Ag/AgCl).

The photocurrent densities calculated in this study for the undoped TiO2 coated SS is compared with the presented data in the literature. As it can be seen in Table 6.1, the obtained values are in within the range of results of the previously published studies. Under the dark condition, both energy and exergy efficiencies of the reactor are obtained as 72.9% and 71.7% respectively. These efficiencies are obtained for illuminated conditions as 1.3% and 3.4% respectively. The drastic drop in calculated efficiencies can be attributed to the high energy input of the solar light with a relatively low photon to current efficiency. Therefore, another definition of the energy efficiency is defined to make the results of light and dark conditions comparable. In this definition, the solar light input is not considered as the energy input and only supplied electrical energy is taken into account. Consequently, energy and exergy efficiencies are found to be 88.5% and 87.1% respectively. As a result of this chapter, the PEC reactor witnesses 21% and 15% enhancement of hydrogen production and efficiencies under the artificial solar light.

Table 6.1 Comparison of photocurrent and applied voltage for different substrate materials Authors Year Photocurrent Substrate material Voltage density applied -2 Mascaretti et al., 2017 2017 40 μA cm Ar/O2-TiO2 photoanode 0.75V (vs [184] Ag/AgCl ) Y. Zhao, Hoivik, and 2016 1.59 mA cm-2 nanopore/nanotube TNTs NA Wang, 2016 [185] electrode -2 Fan et al., 2016 [186] 2015 0.12 mA/cm TiO2/RGO/Cu2O 0.8V (vs Ag/AgCl) -2 Tripathi and Chawla, 2015 18 mA/ cm TiO2-CuO 0.8 (V vs 2016 [187] SCE) -2 Momeni and Ghayeb, 2015 0.36 mA/cm Cr-TiO2NT 1.0 (V vs. 2015 [183] Ag/AgCl) -2 Fa et al., 2014 [188] 2014 80.2 μA/cm TiO2–rGO 1.0 (V vs SCE) -2 Arakelyan et al., 2008 2007 3 mA/cm Thin film n- TiO2 0.85 (V vs [189] SCE)

6.1.1 Small Scale PEC Reactor electrodes

The TiO2 nanoparticle deposited stainless steel rod is tested for photoelectrochemical characterization in a KHP solution under UV light. A graphite rod is deployed as the counter electrode in a full cell experiment. The experiments are conducted under UV light, no UV light and chopped UV light conditions to observe the effect of induced photocurrents. All the experiments are repeated in a sequence to avoid any external effect. All the electrodes used in this section has a surface area of fewer than 500 mm2. The open-circuit potential (OCP) is an electrochemical measurement which shows the voltage difference between two electrodes of an electrolyzer when disconnected from any circuit. Unlike the other methods presented in this study, there is no external load is applied to the reactor in the OCP measurements. Fig. 6.12 presents an open circuit potential scan of the photoanode and graphite counter electrode under chopped light illumination.

The figure proves the photoresponses of the TiO2 coated stainless steel rod to an open circuit potential (OCP) scan. The semiconductor layer of the photoanode is activated under UV light where the electron-hole pair is generated which increases electric potential [138].

-100 UV light on -105 ↓ -110

-115

-120

-125

-130 Voltage Voltage (mV) -135 ↑ -140 UV light off

-145

-150 0 50 100 150 200 Time (s)

Fig. 6.12 Open circuit potential of TiO2 coated stainless steel with Na2SO4 (0.14 M) electrolyte solution under chopped UV light condition.

) Dark UV-Light 2 5

Current Density (A/m Density Current 1

0 0 5 10 15 20 Time (min) Fig. 6.13 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V. (TiO2 coated stainless steel used as the anode under UV light and dark conditions with 0.14 M Na2SO4 as the electrolyte solution).

A jump of around 25 mV reveals the photoactivity of the coated TiO2 photoanode where the photocurrent shows a positive potential bias, indicating an n-type signal of the coated metal, which implies that there is a sufficient overpotential for the reduction of wastewater on irradiated TiO2. As it is presented in Fig.6.13, 6.14, and 6.15, photoactivity of the coated sample and uncoated samples are comparatively investigated under potentitostatic mode. 0.14 M Na2SO4 is used as the electrolyte in these experiments. It can be seen in Fig. 6.13 that under 1.4-volt cell potential, the electric current witnesses highest increase in percentage. Around 38% of an increase of electric current is observed with the reactor formation where the TiO2 coated photoanode, and the graphite cathode is used under the UV light. The test is repeated under similar conditions with different electrode pairs. In Fig. 6.14 graphite is used as the anode whereas, in Fig. 6.13, stainless steel is selected as the anode electrode. Similar to the results in Fig. 6.3, the graphite electrode shows some photoresponse. However, no significant impact of UV light is observed on the stainless steel electrode [138].

12 ) 2 Dark UV-Light 10

4 Current Density (A/m Density Current 2

0 0 5 10 15 20 Time (min)

Fig. 6.14 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V. Graphite used as the anode under UV light and dark conditions with 0.14 M Na2SO4 as an electrolyte solution.

Dark UV-light

) 2 10

4 Current Density (A/m Density Current 2

0 0 5 10 15 20 Time (min)

Fig. 6.15 Current density graphs of experiments conducted under potentiostatic mode with 1.4 V. Stainless steel used as the anode under UV light and dark conditions with 0.14 M Na2SO4 as an electrolyte solution.

6.2 Single Compartment Cell PEC Reactor The experiments are conducted in the rectangle batch reactor with a volume of 30 L as it is explained in the experimental apparatus and procedure section. A TiO2 deposited stainless steel anode with an active area of 180 cm2, and a graphite rod cathode are used as electrode the pair in the reactor. Fig. 6.16 shows current vs. time graph and hydrogen production rate for different types of synthetic wastewater (RB5, sucrose) and tap water by using photo-assisted electro-Fenton process (PAEF) with an average input power of 2 W. While the highest current flow occurs under tap water by around 230 mA, the lowest electric current is measured in the experiment uses sucrose solution as the electrolyte. Fig. 6.17 indicates the total hydrogen production during the electrolysis process of 40 minutes. The results show that sucrose has the highest hydrogen production rate of 2.1 ml/min while tap water has the lowest hydrogen production rate of 0.7 ml/min [130].

250 Sugar 245 Water 240 RB5 235 230 225 220 Current (mA)Current 215 210 205 200 0 10 20 30 40 Time (min) Fig. 6.16 (a) Current vs. time graph and (b) hydrogen production rate from different types of synthetic wastewater (RB5, sucrose) and tap water by using photo-assisted electro-Fenton process (PAEF).

2.5

2.0

1.5

1.0

0.5 Hydrogen production rate (ml/min)rateproduction Hydrogen

0.0 RB5 Sugar Water

Fig. 6.17 (a) Current vs. time graph and (b) hydrogen production rate from different types of synthetic wastewater (RB5, sucrose) and tap water by using photo-assisted electro-Fenton process (PAEF)

Fig. 6.18 shows the decolorization of the RB5 solution for treated and untreated samples. The solution (20 l) is containing 6x10-5 M RB5 (55 ppm of initial COD) at the beginning of the process. It is observed that the absorption bands decrease with time and the reaction goes on with the formed iron ions and hydrogen peroxide [130].

1.5 Treated sample (17 h)

Treated sample (27 h)

1 Untreated sample (27 h)

Treated sample (47 h) Absrorbance

0.5

0 340 440 540 640 740 840 Wavelength (nm)

Fig. 6.18 The RB5 degradation of treated and untreated samples for a solution (25 l) containing 6×10-5 M RB5 (55 ppm of initial COD) at the beginning of the electrolysis

After the first day, the color of sample turns completly transparent, and no noticeable change in absorbance is observed. Fig. 6.19 shows the untreated and treated sample pictures. The textile waste becomes colorless after the photoelectrochemical process. Fig 6.20 depicts the comparison of color degradation of the RB5 solution just after the experiment. It is observed that between the wavelengths (λ) of 340 nm to 740 nm optical absorbance of the sample considerably decreases. In the scope of this study, also the COD reduction is measured. 33% of COD reduction is achieved in the first 17h for the RB5 solution. Table 6.2 shows the main measured data in this section and their uncertainty

99 ranges. In order to understand the effect of Fenton type reactions on the treatment process, iron ion concentrations are also measured via colorimetric methods immediately after the experiment. The iron concentration in the substrate is obtained as 0.5 ppm at the end of the experiments.

Fig. 6.19 The RB5 solution before and after the treatment

Initial 4 Post experiment

2 Absrorbance

0 340 440 540 640 740 840 Wavelength (nm)

Fig. 6.20 The comparison of the optical density of the RB5 solution before and after the electrolysis.

100

Table 6.2 Measured data of the single cell reactor with the uncertainty range Uncertainty Calculation Identifier Value range ± Unit Exergy efficiency 4.4 ± 0.628 % Energy efficiency 2.1 ± 0.336 % H2 production H2 production (RB5) 5.54 ± 0.277 mg H2 production H2 Production (Sugar) 5.1 ± 0.255 mg H2 production H2 Production (Water) 5.84 ± 0.292 mg Absorbance 17 h after exp. at 436 nm 0.094 ± 0.0009 cm-1 Absorbance 17 h after exp. at 525 nm 0.062 ± 0.0006 cm-1 Absorbance 17 h after exp. at 620 nm 0.039 ± 0.0004 cm-1 Absorbance 27 h after exp. at 436 nm 0.068 ± 0.0007 cm-1 Absorbance 27 h after exp. at 525 nm 0.044 ± 0.0004 cm-1 Absorbance 27 h after exp. at 620 nm 0.036 ± 0.0004 cm-1 27 h after (untreated Absorbance sample) at 436 nm 0.559 ± 0.0056 cm-1 27 h after (untreated Absorbance sample) at 525 nm 0.770 ± 0.0077 cm-1 27 h after (untreated Absorbance sample) at 620 nm 1.647 ± 0.0165 cm-1 Absorbance 47 h after exp. at 436 nm 0.063 ± 0.0006 cm-1 Absorbance 47 h after exp. at 525 nm 0.016 ± 0.0002 cm-1 Absorbance 47 h after exp. at 620 nm 0.002 ± 0.00002 cm-1

Energy and exergy efficiencies are obtained for the PEC systems using RB5 solution, sucrose solution, and tap water as substrates. The average energy and exergy efficiencies are found to be 2.1% and 4.4% respectively. Similar to the results in section 6.1.1, as the solar energy input to the system is dominating the energy inlet (artificial solar light and electricity inputs are the inlets), the efficiencies are limited by the conversion efficiency of the titanium dioxide nanoparticle coated surface, and hence the efficiencies are dropped accordingly. However, it should be noted that the system induces photocurrents under irradiance and it leads an increase in the hydrogen production rate. So that the system produces more hydrogen and hydroxyl radicals under solar irradiance. As the solar energy is the most abundant and accessible form of energy in the world, it can be used in such an application even with the low conversion efficiencies. An alternative efficiency definition is also defined to see improvement in the electrolyzer performance under light condition. Since the solar irradiance dominates energy and exergy the input sides of the system (with a low photon to hydrogen conversion), it is hard to compare the system performance under illumination and dark conditions. Hence the solar terms are omitted

101 from the efficiency definitions. The energy efficiency of the reactor is obtained 73% and 88% with light and dark conditions respectively. Moreover, the exergy efficiencies are found to be 72% and 87% with light and dark conditions respectively. Faraday efficiency of the textile wastewater (RB5 solution) is calculated as 98% in this study. For further experiments with a single cell reactor in this study (all configurations without a membrane), this faraday efficiency is used in the calculations. To enhance this process some modifications are done in the final form of the reactor such as using two dissimilar anode plates, quartz glass assembly, a membrane electrode assembly, closer electrode positioning which can be named as one of the main improvements in the reactor [130].

6.3 Synergistic Effects of AOPs for Hydrogen Production and Wastewater Treatment Processes A small-scale modular PEC reactor is developed and built to comparatively investigate the synergistic effects of advanced oxidation reactions in a combination of TiO2 photocatalysis for hydrogen production and wastewater treatment applications as explained in

18 16 500 mA 14 12 10 8 6 4

2 Total Iron ion concentration (ppm)concentrationion Iron Total 0 0 10 20 30 40 50 Time (min) Fig. 6.21 The electroproduction of iron ions on an SS rod anode in KHP solution and under UV light.

102 experimental setup and procedure section. After characterization of the photoanode is evaluated as in Section 6.1, the sacrificial iron rate of a stainless steel electrode is measured by a spectrophotometric method to ensure the iron concentration is sufficient for Fenton like processes. Fig. 6.21 indicates that total iron concentration of the steel anode almost shows a linear profile by time. At the end of 40 minutes of operation of the reactor with a 200 ml of solution, iron concentration is increased from 0 ppm to 16.5 ppm [138].

Table 6.3. Experimental conditions of the comparative study.

103

It is observed that ferrous ion concentration is generated less than 0.5 ppm which also means that ferric ion is the dominant species in the solution. Based on the iron ion concentration it can be stated that the iron sacrificial rate is sufficient for Fenton type reagent as long as hydrogen peroxide is added to the solution [190,191]. Another test is performed with the TiO2 nanoparticle deposited stainless steel anode. At the end of 40 minutes of electrolysis, less than 0.5 ppm iron sacrificial is observed from the coated photoanode which also means that Fenton type processes have less contribution to the experiments when the TiO2 coated anode is used [138]. In this section, various combinations of experimental conditions are tested to investigate the synergetic effect of advanced oxidation processes in a wastewater electrolyzer for hydrogen production and wastewater oxidation. The conditions of each experiment are tabulated in Table 6.3. For the experiments with UV light, only the anode part is exposed to illumination.

3.00 Anolyte (with UV) Catholyte (with UV) 2.50 Initial solution Anolyte (dark) Catholyte (dark)

) 2.00 Mixed electrolyte (with UV) (24h) 1 - Mixed electrolyte (dark) (24h)

1.50

Absorbance (cm Absorbance 1.00

0.50

0.00 400 500 600 700 800 900 Wavelength (nm) Fig. 6.22 The RB5 degradation for Experiment 21 and 22. Spectra recorded for the RB5 solution, anolyte and catholyte at the end of electrolysis and 1 day after.

104

All the experiments are operated for 40 minutes except Experiments 13-18 which are conducted for 20 minutes. In all experiments, reactions were terminated after 24 hours from the experiment by adding 1 M NaOH until pH level reaches 10. Fig. 6.22 illustrates the decolorization results of experiments 20 and 21 for the anode and cathode sections separately. The effect of UV and H2O2 photolysis on decolorization in the electrolyzer also can be seen in Fig. 6.22. As expected, the anolyte of Experiment 20 witnesses the highest RB5 degradation and followed by the catholyte of the same reactor. It should be noted that there is a mass transfer due to diffusion of electrolytes as there is no membrane in between the anode and cathode sides [138]. It can also be seen in Fig. 6.22 that the least color removal is observed in the catholyte of Experiment 21, which is operated under dark conditions. It can also be seen in the same figure that more RB5 degradation occurs in the anolyte compared to catholyte under same conditions. For experiment 21, both electrodes are kept dark, and both sides have the same amount of H2O2 concentration. However, the anolyte is witnessed more decolorization. This situation can be explained by the charge of the dye solution. Karcher et al. [192] were studied removal of reactive dyes from textile wastewater by using anion exchange resins. In that study, RB5 was successfully removed from the textile wastewater by anion exchange resins which also shows RB5 is an anionic based dye. This is almost the same in the present reactor since the anode side of the reactor removes RB5 with a higher efficiency while using graphite anode and graphite cathode. Positively charged graphite rod adsorbs the RB5 dye through the surface, which rises RB5 removal efficiency [138]. In Fig. 6.23 potentiostatic electrolysis of the RB5 solution with different pH levels under UV light condition is presented. Initially, RB5 based textile waste is introduced to the reactor without adjusting its pH (10.5). Then its pH is dropped to 3 for another testing. It can be stated from the graph that by dropping the pH of the electrolyte from 10.5 to 3 it is possible to have an increase in the current around 6% which results in more hydrogen production. Moreover, as the Fenton reaction is efficient in acidic mediums, it increases the COD removal efficiency naturally with the presence of H2O2 and iron concentration.

105

pH 3 20

pH 10.5

19 (mA)

18 Current

16 0 10 20 30 40 50 60 Time (min) Fig. 6.23 Current vs. time graph of the RB5 solution electrolysis with a different pH level of dye solution under UV light condition with the TiO2 coated stainless steel photoanode and graphite cathode.

The organic matter degradation and hydrogen production performances of experiments with the KHP solution are shown in Fig. 6.24. All the presented COD removal efficiencies are representing the COD removal of the substrate after one day from the electrolysis. The maximum degradation efficiency is observed in Experiment 5 by 98%, whereas minimum degradation efficiency is obtained for Experiment 4 (Without UV/

Graphite anode/H2O2) by 48%. In Experiment 5, advanced oxidation processes as photocatalysis (UV/TiO2) and photolysis (UV/H2O2) are responsible for the water treatment. Moreover, with the small amount iron concentration sacrificial from the coated photoanode, electro-Fenton type processes also contribute to hydroxyl radical forming, respectively.

106

100 10 90 KHP solution 9 80 8 70 7 60 6 50 5 40 4 30 3

COD removal efficiency (%) efficiency removal COD 20 2 Hydrogen production (mg/Wh) production Hydrogen 10 1 0 0 E1 E2 E3 E4 E5 E6 COD removal efficiency (%) Hydrogen production (mg/Wh)

Fig. 6.24 KHP degradation and hydrogen production in PEC reactor under various experimental conditions. 100 10 RB5 solution 90 9

80 8

70 7

60 6

50 5

40 4

30 3

20 2 (mg/Wh) production Hydrogen COD removal efficiency (%) efficiency removal COD 10 1

0 0 E7 E8 E9 E10 E11 E12 COD removal efficiency (%) Hydrogen production (mg/Wh)

Fig. 6.25 RB5 degradation and hydrogen production in PEC reactor under various experimental conditions.

107

The COD removal efficiency of E1 experimental run (as seen in Fig. 6.24) is observed as 94% with the electrolysis detention time of 40 minutes, which is one of the highest recorded value considering all experimental runs given in Table 6.3. Further, highest hydrogen production rate is obtained at E1-run of photo-electro Fenton type process, when compared with the other experimental runs. Fenton/Fenton like reagent forms by the iron oxidation (with the presence of H2O2) in the anode side of the reactor which requires 162.9 kJ mol-1 less energy than oxygen evolution [9]. Hence less energy is consumed to yield hydrogen than conventional water electrolysis. Consequently, a higher amount of hydrogen is produced during the electrolysis [138]. Fig. 6.25 reveals the COD removal and hydrogen production results of the experiments with the RB5 solution. Similar to the experiments with the KHP solution,

Experiment 12, where the TiO2 coated photoanode is used with H2O2 under UV light has the highest COD removal efficiency by 77%, and Experiment 15 where graphite anode is used with H2O2 has the least COD degradation by 26%. Moreover, similar to the results with the KHP solution, the maximum hydrogen is produced in Experiment 8. It is observed that experiments with the KHP solution have more organic matter degradation relative to the experiments with the RB5 solution. It can be explained in two ways. Firstly, the KHP solution is colorless and has more optical transmittance compared to RB5. Therefore, UV light cannot penetrate the RB5 as it can do through the KHP solution. Hence, it affects the degradation rate positively for the KHP solution. Secondly, as it is depicted in Fig.4.6 and Fig. 4.7, RB5 has a more complex structure than KHP, and it requires more hydroxyl radicals for the treatment. The effect of sacrificed iron of the anode and H2O2 concentration on COD removal is summarized in Table 6.4 The uncoated stainless steel sacrifices 17 mg/L total iron in 40 minutes of reaction time [138]. 3+, Under no UV light condition, 200 mg/L H2O2 reacts with 17 mg/L Fe and 79% COD was removed from the KHP solution. When UV light applied to the process, COD 3+ removal efficiency increases to 94% due to the reaction of UV with Fe and H2O2. The 2+ 3+ Fe -Fe /H2O2 ratio of the process which is conducted with steel anode under UV light condition is determined as 0.053. This ratio is very close to the optimum Fe/H2O2 ratio (0.055) of Fenton oxidation for RB5 removal obtained by Meric et al.[140].

108

Table 6.4 Effect of sacrificed iron and H2O2 concentration ratio on COD removal efficiency 2+ COD H2O2 Sacrificed Fe - 2+ 3+ 3+ Anode UV light removal, reacted, Fe -Fe , Fe /H2O2 % mg/L mg/L ratio Uncoated Stainless Steel Yes 94 323 17 0.053 Uncoated Stainless Steel No 79 200 17 0.085 Graphite Yes 55 173 - - Graphite No 48 99 - -

Electric power usage and overall energy consumption during the PEC reactor operation are also investigated comparatively under various conditions. Fig. 6.26 illustrates the power usage and overall energy consumption data to drive PEC reactor with and without light conditions. The reactor operates under the galvanostatic mode with steel- graphite electrode pairs and utilizes KHP as the electrolyte. During the 40 minutes of operation, the reactor uses 2.87 W electric power. The presence of the UV light drops the power usage by 3.6% and power consumption becomes 2.77W.

3.1 8.0 7.0 2.9 6.0 with UV-light 5.0 dark 2.7 4.0 with UV-light

dark 3.0 Power Power (W) 2.5 Energy (kJ) Average Power Consumption: 2.0 with UV-light: 2.77 W 1.0 dark: 2.87 W 2.3 0.0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min)

(a) (b) Fig. 6.26 Power usage(a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the KHP solution as the electrolyte. The steel anode and the graphite cathode is used with (E1) and without UV light(E2) conditions

109

3.2 8.0

3.1 7.0

3.0 6.0 with UV-light 2.9 dark 5.0 2.8 with UV-light 4.0 2.7 dark

Power Power (W) 3.0 2.6 Energy (kJ) 2.0 2.5 Average Power Consumption: with UV-light: 2.94 W 2.4 dark: 3.05 W 1.0 2.3 0.0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min) (a) (b) Fig. 6.27 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the KHP solution as the electrolyte. The graphite anode and the graphite cathode is used with (E3) and without UV light(E4) conditions

3.0 7.0

2.9 with UV-light 6.0 with UV-light dark dark 2.8 5.0

2.7 4.0

2.6 3.0

Energy Energy (kJ) Power Power (W) 2.5 2.0 Average Power Consumption: with UV-light: 2.61 W 2.4 dark: 2.76 W 1.0

2.3 0.0 0 10 20 30 40 0 20 40 Time (min) Time (min) Fig. 6.28 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. (Steel anode and graphite cathode)

110

3.0 8.0

with UV-light 7.0 with UV-light 2.9 dark 6.0 dark

2.8 5.0

4.0 2.7

Energy Energy (kJ) 3.0 Power Power (W)

Average Power Consumption: 2.0 2.6 with UV-light: 2.81 W dark: 2.84 W 1.0

2.5 0.0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min)

Fig. 6.29 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. The graphite anode and graphite cathode is used with (E9) and without UV light(E10) conditions

Fig. 6.27 shows power usage and overall electrical energy consumption for the electrode pairs of graphite anode and a graphite cathode. The reactor is run under the galvanostatic mode with light and without light conditions. At the end of 40 minutes of operation, the reactor uses 3.05 W electric power on average. Similar to the Fig. 6.26 with the presence of UV the electric power consumption declines by 3.6% and becomes 2.94 W Fig. 6.28 shows the results for the reactor formation with a stainless steel anode and a graphite cathode. The RB5 solution is used as the electrolyte in these set of experiment and the electric power supplied in the galvanostatic mode for 40 minutes. The experimental run under the dark condition has an electric power requirement of 2.76 W and consumes 6,624 J energy in 40 minutes of operation. Under the UV light, these numbers drop to 2.61 W and 6,264 J respectively. The power consumption diminishes by 5.7% with the presence of the UV light. Fig. 6.29 illustrates the results for the reactor formation with the graphite anode and graphite cathode. The reactor operates under the galvanostatic mode with and without light conditions. The synthetic textile waste is used in the experimental runs. Under the dark conditions, the reactor uses 2.84 W electric power and consumes 6,816 J of energy. When the UV lamp operates there is slightly change is observed. The power usage of the reactor

111 drops to 2.81 W which corresponds to 1.1% change in the consumption. When the Fig. 6.28 and Fig. 6.29 are compared, the big difference in power usages between the cases can be seen. This situation can be attributed to the impact of Fenton and Fenton type processes on the system. As it is explained in the previous chapters, iron oxidation requires less power than oxygen evaluation which results in less power demand by the anode reactions [138].

2.9 7.0

with UV-light 6.0 with UV-light dark 2.8 dark 5.0

4.0 2.7

3.0

Energy Energy (kJ) Power Power (W)

2.6 2.0 Average Power Consumption: with UV-light: 2.62 W 1.0 dark: 2.73 W 2.5 0.0 0 10 20 30 40 0 10 20 30 40 Time (min) Time (min) Fig. 6.30 Power (a) and overall energy consumption graph (b) of the PEC reactor under the galvanostatic mode with the RB5 solution as the electrolyte. The TiO2 coated steel photoanode and the graphite cathode is used with (E12) and without UV light(E11) conditions

Fig. 6.30 shows results for the TiO2 coated photoanode and the graphite cathode electrode pairs. The RB5 solution is used as the electrolyte and the system is operated under the galvanostatic mode similar to the other cases. During the 40 minutes of operation, the reactor uses 2.76 W electric power and with the presence of UV electric power consumption drops by 4.1% and becomes 2.62 W. It can be seen in Fig. 6.28 and Fig. 6.30 that the PEC reactor formation with uncoated stainless steel is consuming less energy than the reactor with photoanode. As the

TiO2 avoids the electro-generation of iron ions, Fenton-type processes occur in an insufficient level in the corresponding reactor. Therefore, the reactor with a stainless steel anode requires less/almost same power as the one with photoanode. Absorbance values of the treated and untreated RB5 solutions can be seen in Fig.

6.31. The highest color removal (84%) obtained with reactor formation which uses TiO2

112 coated stainless steel as the photoanode under UV light condition (Table 6.5). Even though 2+ 3+ a small amount of photo-electro produced Fe /Fe from the TiO2 coated stainless steel lead to Fenton process, the formed Fenton reagent contributes the decolorization process [138].

Table 6.5 Absorbance of the substrates at specific wavelengths, which were, referred in German wastewater discharge standards. Absorbance (cm-1)

Substrate 436 (nm) 525 (nm) 620 (nm)

RB5 Influent 1.002 1.55 0.77

E8 (after 1 day) 0.33 0.36 0.14

E9 (after 1 day) 0.56 0.45 0.19

E10 (after 1 day) 0.53 0.5 0.2

E11 (after 1 day) 0.52 0.46 0.17

E12 (after 1 day) 0.20 0.27 0.08

Moreover, UV light excites TiO2 nanoparticles coated stainless steel to produce OH● radicals which enhances the color removal efficiency. As it is presented in Fig. 6.31, E8, which is conducted with the stainless steel as the anode and under the UV light condition has the second best color removal efficiency by 75%. The color removal efficiency difference between E8 and E9 can be explained by the absence of TiO2 nanoparticles. TiO2 nanoparticles in E8 have around 10% positive effect on color removal efficiency for the RB5 solution. Also, the experiments E9 and E10, which are conducted with the graphite rod as the anode and under the light and no light conditions have almost the same color removal efficiencies. In these experiments, dye absorption on the graphite rod’s surface is observed which assists the color removal mechanism. The main measured data of the comparative study is tabulated with their uncertainty ranges in Table 6.6. As it is presented in this section, Cases with a TiO2 coated electrode under UV light illumination gives better results compared to other combinations. Around 10 mg/l of COD removal is calculated for the water analysis tests and 0.07 W electric power uncertainty is calculated for the potentiostat measurements [138].

113

Table 6.6 Measured data of the comparative study with the uncertainty ranges. Electric power COD removal COD removal H2 (mg/Wh) Experiment usage (W) (mg/l) efficiency (%) E1 6.594 ± 0.1458 2.768 ± 0.07187 99.37 ± 12.06 94.04 ± 0.9623 E2 6.366 ± 0.1407 2.867 ± 0.07444 83.07 ± 12.31 78.61 ± 3.452 E3 6.200 ± 0.1371 2.944 ± 0.07644 58.56 ± 13.18 55.42 ± 7.196 E4 5.978 ± 0.1322 3.053 ± 0.07927 50.36 ± 13.59 47.66 ± 8.448 E5 6.066 ± 0.1342 3.009 ± 0.07812 103.5 ± 12.04 97.95 ± 0.3315 E6 6.088 ± 0.1347 2.998 ± 0.07784 52.24 ± 13.49 49.44 ± 8.161 E7 6.618 ± 0.1463 2.758 ± 0.07161 53.45 ± 9.077 70.04 ± 4.835 E8 7.001 ± 0.1548 2.607 ± 0.06769 58.18 ± 8.937 76.24 ± 3.835 E9 6.500 ± 0.1438 2.808 ± 0.07290 20.11 ± 10.8 26.35 ± 11.89 E10 6.417 ± 0.1419 2.842 ± 0.07379 38.31 ± 9.714 50.2 ± 8.038 E11 6.683 ± 0.1478 2.731 ± 0.07091 58.98 ± 8.917 77.29 ± 3.666 E12 6.956 ± 0.1539 2.624 ± 0.06813 45.45 ± 9.379 59.56 ± 6.527 RB5 solution initial COD: KHP solution initial COD: 105.7 ± 12.04 76.31 ± 8.695

2.5 RB5 Influent E8 E9 E10

2 E11 E12

)

1 - 1.5

1 Absorbance (cm Absorbance

0.5

0 340 440 540 640 740 840 Wavelength (nm) Fig. 6.31. Optical absorbance values of RB5 solution for various experiments after one day from the electrolysis.

114

6.4 Thermoelectric Generator In the proposed system, the PEC reactor is driven by the battery stack consists of thermoelectric generators. A preliminary study is conducted to evaluate the feasibility of the TEG devices. Also in the previous studies, TEGs have been investigated as the part of Ph.D. study [108,160,193,194].

4 20.0 o Tc=50 C 18.0 3.5 V I Efficiency 16.0 3 14.0 2.5 12.0

2 10.0

8.0 Efficiency (%) Efficiency

Electric potential (V) (V) potential Electric 1.5 Electrical current (A) current Electrical 6.0 1 4.0 0.5 2.0

0 0.0 100 200 300 400 500 Hot Side Temperature (oC)

Fig. 6.32 Variation of thermoelectric performance with temperature

TEC solid-state power generators company`s Calcium/Manganese – Bismuth Telluride-based high-temperature thermoelectric generator is selected for voltage output results corresponding to temperature differences [195]. For the operating temperatures below 300 oC, thermoelectric properties are obtained by applying linear regression of the data in the spec sheet of the TEG. Sahin et al.`s [196] temperature and figure of merit

115 relation is used in the preliminary study. Heat to electricity conversion efficiency is obtained by using Eq. (5.12).

700 8.00

7.90 600 7.80

500 7.70 7.60 400 7.50

7.40 300 Generated Electricity 7.30 (W)

Generated Electricity (W) Electricity Generated 200 7.20 Heat to Electricity Efficiency (%) (%) Efficiency Electricity to Heat 7.10 100 7.00

0 6.90 275 285 295 305 315 Inlet Stream Temperature (K)

Fig. 6.33 Change of heat to electricity efficiency and generated electricity of the TEG with the inlet temperature of the cold stream.

Fig. 6.32 presents the performance of the thermoelectric generator between the range of 100oC-500oC hot side temperature and constant (50oC) cold side temperature. The voltage and current values are obtained from the data sheet of a BiTe2 based TEG unit [195]. Electric power output is positively affected by the increasing temperature. It can be explained in three ways. First, the thermoelectric properties as Seebeck coefficient, electrical conductivity, and thermal conductivity are functions of the temperature, which also directly affect the thermoelectric efficiency. Secondly, The increase in the temperature of the hot side causes in a decrease in

Tc/Th ratio which has a positive effect on the thermoelectric efficiency. This situation results in a rise in the thermoelectric efficiency. Lastly, the increase in the hot side

116 temperature leads greater heat flow between the top and bottom side of the thermoelectric generator. Higher heat transfer rate results in higher power output by the TEG as well.

355 7600

350 Exit Stream Temperature (K) 7400

345 Heat Transfer Rate (W) 7200 340

335 7000

330 6800

325

6600 (W) Rate Transfer Heat Exit Stream Temperature (K) Temperature Stream Exit 320 6400 315

310 6200 275 285 295 305 315 Inlet Stream Temperature (K)

Fig. 6.34 Change of exit stream temperature and heat transfer rate of the TEG with the inlet temperature of the cold stream.

The numerical analysis executed by COMSOL Multiphysics software presents the results for temperature distributions, heat transfer characteristics, thermoelectricity generation and efficiency values are calculated. The variation of the heat to electricity efficiency of the thermoelectric generator with the cold stream temperature is presented in Fig. 6.33. The hot side temperature of the TEG is set constant as the melting temperature of the PCM while the temperature of the cold stream is varying from 5°C to 45°C. The increase of the temperature of the cold stream (wastewater flow) has an adverse effect on both electricity generation and efficiency. Since Tc/Th ratio decreases with the increasing cold stream temperature, the figure of merit and the efficiency values drop. Moreover, the

117 diminishing temperature difference between hot and cold sides have also a negative impact on the thermoelectric power output.[108,160,165,193,194]. Fig. 6.34 illustrates the heat transfer characteristics of the TEG with the increasing wastewater temperature. The exit stream temperature of the TEG increases from 314 K to 348 K with the inlet temperature growth from 278 K to 318 K. As it explained earlier, due to decreasing temperature differences between hot surface and the inlet wastewater stream, less heat transfer rate is observed on TEG. Increasing the temperature of the inlet stream from 25°C to 45 °C decreases the heat transfer rate from 7.5 kW to 6.3 kW.

Temperature (K)

Fig. 6.35 The temperature distribution of the TEG module.

Results of the COMSOL modeling are presented in Fig. 6.35 and Fig 6.36 which show how the temperature and power production varies in the TEG unit respectively. Temperature and power density distributions through the thermoelectric generator are shown in Fig. 6.35 and Fig. 6.36 respectively. The highest temperature is obtained closer to the exit region while the highest power density is obtained closer to the inlet region. The generated power density has the maximum value at the inlet of the TEG by around 1500

118

W/m2 and diminishes to 900 W/m2 at the exit of the TEG. In order to have a homogeneous temperature distribution and power densities, the mass flow rate of the inlet stream can be increased. However, it will lead to a lower exit temperature. In this study, the cold-stream is desired to have a temperature of around 60°C. Therefore, the mass flow rate is adjusted accordingly (0.05 kg/s) [165].

Fig. 6.36 Change of heat to electricity efficiency of the TEG with the inlet temperature of the cold stream.

6.5 Molten Salt storage In the suggested system, molten salt supplies heat for the thermoelectric generator, PEC reactor and store the heat for later usage. A preliminary study is performed for a cylindrical tank with a 10 cm of diameter and height. A molten salt in a storage medium in a cylindrical tank mentioned corresponds to 534 grams. It is obtained that, to heat up that amount of molten from 25oC to 350 oC with the concentrator under 700 W/m2, 5 minutes and 16 seconds heating is required. It is obtained that after all the molten salt changes its phase from solid to liquid 31.94 °C temperature drop occurs in a one-hour storage period with 5 cm fiberglass insulation material. Later, the molten salts cool down to melting temperature

119 and around 8 hours all the molten salt change its phase from solid to liquid. To enhance the thermal management, the fiberglass-based insulation material is replaced with the vacuum insulated products. In the proposed system, the molten salt should maintain its temperature at melting temperature to have a stable power output. Fig. 6.37 represents the heat losses through a rectangular molten salt tank, which has a lateral area of 8.34 m2. The insulation material thickness is varied to obtain the most suitable value for the system.

2000

1800

1600

1400

1200

1000 (W)

loss 800

. Q 600

400

200

0 0 2 4 6 8 10 Insulation thickness (cm)

Fig. 6.37 The variation of heat losses through the molten salt tank with the changing insulation thickness

As it can be seen in Fig. 6.37 the effect of insulation material on the heat losses are significant for the first 5 cm. Then, there is only a slight enhancement of the insulation is obtained. Therefore, 5 cm of insulation material is used in this study. It should also be noted that more insulation thickness increases the heat transfer surface, which results in more heat losses after the optimum point.

120

6.6 PEC Reactor and Integrated System Results As the first step of the study, a PEM electrolyser is investigated theoretically and experimentally. The PEM electrolyzer without the presence of semiconductor deposition on anode surface is used in the first experiments. Design parameters of the reactor are tabulated in Table 6.7. The reactor is run under different temperatures to see the influence of temperature on the hydrogen production and efficiencies. Since the proposed system includes a heating process before the PEC reactor, this first step results study has vital importance in the thesis work. Gamry 3000 high-performance Potentiostat/Galvanostat/ZRA [132] is used as the power supply and electric current and potential measurement device. The first experiment is run under potentiostatic conditions, which supplies constant electric voltage to the system. A maximum current value is set before the experimental study to avoid any potential problems due to over electric current passage through the electrolyzer. Table 6.7 Design parameters of the PEM electrolyser [20] Cathode exchange current density 3.20×10-7 A/m2 Anode exchange current density 1.70×10-9 A/m2 Hydrogen pressure 101.325 kPa Water pressure 101.325 kPa Oxygen pressure 101.325 kPa Membrane conductivity 0.102 S/cm Membrane thickness 0.0127 cm PEM cell active area 0.026 m2 Ambient temperature 25 oC

Each experiments take 15 minutes for each condition. Two different operating temperatures are tested as being 24oC and 44oC. 3 V of electric potential is supplied to both cases. Fig. 6.38 presents the time and electric curve of the PEM electrolyzer under different temperatures. As it stated in Chapter 5, the actual electric potential of the PEM electrolyzer increases with the temperature, which increases hydrogen production. Also, a mathematical model has been developed of a reactor for hydrogen production in the scope of this study. Results have been published in International Journal of Hydrogen Energy [160]. Considerable amount of increase in hydrogen production rate is observed with the increasing temperature from 24oC to 44 oC.

121

Fig. 6.38 Current-time curve of the reactor measured by the potentiostat device under two different operating temperature

The variation of efficiency in the operating temperature range between 8.89oC and o 80 C is given in Fig. 6.39. The first definition of the reactor efficiency (ηPEM) increases from 71.1% to 84.6% and the second definition of the reactor efficiency (ηPEM,2) arises from 59.0% and reaches the maximum value by 64% with increasing temperature. Here,

ηPEM represents the ratio between supplied energy to the PEM electrolyzer and the energy released from the consumption of the produced H2, ηPEM,2 shows the proportion of the reversible electric potential (Er) to the theoretical electric potential. The experimental and theoretical results are proving the benefits of the integrated system. Heating up the electrolyte with the PCM-TEG unit almost doubles the production rate within the operating range. Increase in production rate can be considered as one of the performance indicators for a PEM electrolyzer. Exergy efficiency is another parameter should be taken into in the designing process. As it can be seen in Fig. 6.40, exergy efficiency of the PEM electrolyzer is increasing around 10% with the increasing temperature [160].

122

85 ηPEM ηPEM,2

Efficiency Efficiency (%) 65

50 10 20 30 40 50 60 70 80 Operating Temperature (oC)

Fig. 6.39 The variation of the reactor efficiencies with operating temperature.

75 1.6 ψPEM Hydrogen Production 1.55 70

1.5

65 1.45

1.4

60 Exergy Exergy Efficiency (%) 1.35

55 Hydrogen Hydrogen Production (mg) 1.3

50 1.25 10 20 30 40 50 60 70 80 Operating Temperature (oC) Fig. 6.40 The variation of the hydrogen production rate with operating temperature

123

The reason behind the improvement of the performance of the PEM electrolyser with temperature can be explained by the decrease in the Gibbs free energy for decomposing the water molecules. Less requirement for the Gibbs free energy, in turn, reduces the electrical potential needed to break the water molecule bond to produce the hydrogen based on the Nernst equation. The second experimental set is conducted with the wastewater as the anolyte under the solar simulator light. 600 W/m2 of the solar simulator irradiance is applied throughout the experiments. The two anode structure is used in the experiments. The first anode is photoanode (TiO2 nanoparticle deposited), and the other one is the sacrificial anode (stainless steel). The experiments take 40 minutes and performance analysis is evaluated based on hydrogen production and COD removal efficiencies. The KHP solution as explained in the section is used in this test. 1.5 V of electric potential is applied under the potentiostatic mode with two different temperatures. Similar to the first set of experiments, 40°C and 24°C are selected as the operating temperatures. Fig. 6.41 shows the current flow through the PEC reactor corresponding to the applied voltage. Similar to the PEM electrolyzer experiments and the theoretical model, the PEC reactor is also showing a positive trend with the increasing temperature. As it is given in Fig. 6.41 cell current is rising from 300 mA to approximately 400 mA with the increasing temperature from 24 °C to 40°C.

Fig. 6.41 Current-time curve of the PEC reactor measured by the potentiostat device under two different operating temperature

124

Fig. 6.42 shows the KHP degradation and hydrogen production in the PEC reactor after 40 minutes of operation. Similar to the water electrolysis with temperature increase the required Gibbs free energy for water splitting decreases. Hence, the electrical required potential to break the water molecule bond decreases based on Nernst equation. While 6.60 mg of hydrogen is produced at the end of the photoelectrochemical reaction at 24°C, 9.05 mg of hydrogen is produced at 40°C. Moreover, the temperature has also a positive effect on the COD removal process. The COD removal efficiency of the reactor rises from 44% to 87% with the increasing temperature from 24 °C to 40 °C.

100 10.00

90 9.00

80 8.00

70 7.00

60 6.00

50 5.00

40 4.00 CODremoval efficiency (%)

30 3.00 Hydrogen Hydrogen production (mg) 20 2.00

10 1.00

0 0.00 24°C 40°C COD removal efficiency (%) Hydrogen production (mg)

Fig. 6.42 KHP degradation and hydrogen production in PEC reactor under different temperatures (0.14 M KHP anolyte and catholyte under 40 minutes operation of the reactor).

Temperature is one of the vital factor affecting catalytic oxidation reaction rates in the Fenton process. Results in this study and the previous studies in the literature show that COD removal efficiency, affected by temperature and increases with the elevating temperature [197]. The impact of the temperature on COD degradation can be attributed to the kinetic parameters of Fenton treatment which was calculated by Sun et al. [198].

E ln k = ln A − ( a ) (6.1) RT

125 where k is the specific reaction rate which controls process, A is the pre-exponential (or frequency) factor, T is the temperature of the substrate (K), Ea is the activation energy (kJ/mol), and R is the ideal gas constant (0.0083 kJ/ mol K). The obtained results of experiments and Eq. 6.1 with a clear indication that the reaction rate of COD degradation is significantly affected by reaction temperature and increased with the raising of temperature. Since the primary goal of the PEC reactor is treating wastewater in the anode compartment while producing hydrogen in the cathode side, another set of tests are conducted with under wastewater catholyte and water (with sulphuric acid) conditions. Fig. 6.43 represents the COD reduction and hydrogen production in the PEC reactor with 0.14 M KHP solution anolyte and 0.1 M sulphuric acid solution catholyte under different temperatures and 40 minutes run. Unlike the previous experiment, the system is operated under relatively higher electric potential. Parallel to the results in Fig 6.41 hydrogen production and COD removal efficiencies are increasing with the temperature. While 17.6 mg of hydrogen is produced at the end of the photoelectrochemical reaction at 20°C, 22.6 mg of hydrogen is produced at 60oC. Furthermore, COD degradation efficiency is enhanced with the temperature. The COD removal efficiency of the system rises from 44% to 87% with the increasing temperature.

100 90 20 80 70 15 60 50 10 40 30

20 5

COD removal efficiency (%) efficiency removal COD Hydrogen production (mg) production Hydrogen 10 0 0 20°C 60°C COD removal efficiency (%) Hydrogen production (mg)

Fig. 6.43 KHP degradation and hydrogen production in the PEC reactor under different temperatures ( 0.14 M KHP anolyte and 0.1 M H2SO4 catholyte under 40 minutes operation of the reactor).

126

Furthermore, the performance of the PEC reactor is also tested in decolorization aspect. The textile wastewater with 0.0014 M Reactive Black 5 dye is treated in the PEC reactor for 5 hours, and at the end of each one hour period 10 ml of sample treated water is taken and pictured (as seen in Fig. 6.44). The highest color change is observed in the first 3 hours then it decolorization process slows down. Fig. 6.44 shows the change in absorbance values during the 5 hours operation of the PEC reactor. Since the excess iron ions generation occurs in the reactor, which does not react with hydrogen peroxide, a yellowish colored solution is formed (yellow for Fe3+ and light green for Fe2+) and the absorbance values are obtained higher than expected. Hence, to see the impact of Fenton type reactions on the wastewater, iron ions are precipitated in higher pH levels with NaOH to form Fe (OH)2. The optical transmittance of the treated sample is increased considerably during the operation time of 5 hours. It should be noted that the decolorization process continues even after the operation time of the PEC reactor and wastewater becomes colorless as long as sufficient H2O2 is added to the reactor.

Fig. 6.44 Color change of textile wastewater (RB5 solution) in PEC reactor during 5 hours of operation.

127

3.5 Influent 3 3 h

) 5h 1 - 2.5 3h high pH 2 5h high pH

1.5

1 Absorbance (cm 0.5

0 400 450 500 550 600 650 700 Wavelength (nm) Fig. 6.45 Optical absorbance of textile wastewater (RB5 solution) in PEC reactor during 5 hours of operation.

Another experimental study is conducted to calculate the reaction rate coefficient of the process. The optical absorbance of the textile wastewater is measured periodically at the peak absorbance wavelength of RB5. The wastewater sample with an initial absorbance of 1.47 cm-1 (at 590 nm) is analyzed. The reactor runs for 52 minutes during the experiment under 2.3 V electric potential. The change in the optical absorbance at 590 nm with time is illustrated in Fig. 6.46.

1.6

1.4

1.2

)

1 - 1

0.8

0.6 Absrobance Absrobance (cm 0.4

0.2

0 0 10 20 30 40 50 60 Time (s) Fig. 6.46 Optical absorbance of RB5 solution (at 590 nm) in PEC reactor during 52 minutes of operation.

128

Moreover, the integration method is implemented to obtain the reaction rate of the process. Based on the literature and Fig. 6.46, the reaction rate of decolorization is expected to be irreversible and first order. The reaction rate (k) of an irreversible first-order process can be calculated as follows:

d[A] r = = −k [A] (6.2) A dt C (t) d[A] t ∫ a = −k ∫ dt (6.3) Ca(0) [A] 0 ln Ca(t) − ln Ca(0) = −k t (6.4) −k t Ca(t) − Ca(0) = e (6.5) Here, r represents the rate of reaction (mol/L), A shows the concentration of the reactant

(mol/L). Ca(t) is the concentration of the reactant at time t. The concentration of RB5 is responsible for the black color to the textile waste. By Beer’s law, the absorbance of the RB5 solution is directly proportional to the concentration of the Reactive Black 5 in solution. Hence plotting the absorbance as a function of time is principally the same as plotting the concentration as a function of time.

1.6

1.4 y = 0.0272x R² = 0.9924 1.2

0.8 ln ln (A(t)/A(0))

- 0.6

0.4

0.2

0 0 10 20 30 40 50 60 Time (s) Fig. 6.47 Reaction rate of the optical absorbance removal of RB5 solution.

129

As it can be seen in Fig. 6.47, the plot of the –ln A(t)/A(0) with respect to time is linear. A(t) represents the optical absorbance of the textile wastewater at time t. The linearity shows the reaction takes place as the first order. The coefficient of determination R2 (over 0.99) of the fitted curve also validates the results. The slope of the plotting “0.0272” shows the reaction constant for the optical absorbance change. The reaction constant of RB5 is also can be defined as “ c 0.0272”. Here c stands for a constant number, as the optical absorbance and RB5 concentration are directly proportional. After obtaining experimental results from the PEC reactor with membrane electrode assembly, all the results are combined for integrated system calculations. The thermodynamic calculations of the system components are performed by both EES and COMSOL Multiphysics software. The overall energy and exergy efficiencies of the system overall system for the cases where the heat of the treated water is utilized for space heating applications are calculated as 93.7% and 9.3% respectively. Fig. 6.48 shows the exergy destruction rates of the components used in this study. Due to low heat to electricity conversion and high-temperature differences on the system boundaries, the TEG unit has the most top exergy destruction. Similarly, since the concentrator witnesses losses at higher temperatures, it has the second largest destruction in the system. As the most of the heat is utilized for heating and solar heat used as heat in the final product, the energy efficiency is obtained considerably high. On the other hand, the quality of the energy is drastically drop in the heat exchanger part. PCM at 595K is used for heating water from about 25°C to 60°C which results in a massive entropy generation and exergy destruction for that part. Hence energy efficiency is found high where the exergy efficiency is obtained relatively low. In order to better understand the system performance another efficiency definition is used which excludes the space heating load on the overall system

(ηov and ψov). These energy and exergy efficiencies are found to be 5.5% and 5.2%.

The Bi2Te3 based TEG is operating with a heat to electricity efficiency of 7.46%. Based on the experimental data, the PEC reactor produces 14 mg of hydrogen per hour under 1.5 V electric potential and 600 W of solar irradiance. 40 minutes of electrolysis removes 87.2% of COD within a day. It is obtained that by increasing the water`s temperature from 24°C to 40°C it is possible to enhance electrolyzer efficiency by 4.7%, respectively.

130

10000

1000 rate (kW)rate

andlosses 100

10 Exergy destructionExergy

1 PCM storage Concentrator Receiver TEG PEC

Fig. 6.48 Exergy destructions and losses of the system components.

6.7 Exergoeconomic assessment results In this chapter, a cost analysis method that is also taking into account exergy destructions throughout the system results are presented. Exergoeconomic analysis of the integrated system is carried out by combining numerical and experimental findings and cost in the study. First, the costs of the selected pieces of equipment are listed, and the cost breakdown of the system is obtained as depicted in Fig. 6.49. The main expenses to build such a system is coming from the PtB and IrRuOx catalysts coated Nafion membrane and BiTe2 based TEG unit. These two components are corresponding to 67% of the total setup expenses. The exergoeconomic assessment is based on the stream`s exergy rates and exergy destruction rates of the related sub-systems. Hence, the exergy destruction rates of the subsystems are presented in Fig. 6.50. The highest wasted exergy is obtained for the TEG unit. Relatively lower heat to electricity efficiencies of the component leads greater exergy destruction rates.

131

PEC reactor case Machining SS Electrodes 3% 7% 5% Molten salt Bolts, nuts and SS Electrode plate 4% HEX washers (photoanode) 3% 1% <1%

Quartz glass Sol-gel chemicals Gasket sheet 2% 2% <1% Concentrator Membrane Acrylic glass 2% Rubber piping 27% 2% Other Sealants <1% 3% 1%

Support structure TEG 1% 40%

Fig. 6.49 The cost breakdown of the proposed system.

Note that the temperature difference between the heat source and heat sink in the TEG unit is mainly responsible for the exergy destruction rate. Consequently, the exergy destruction rates are quite higher than the other subsystems. 2.8 kW exergy destruction rate is obtained at the TEG unit. Also concentrator witnesses a high exergy destruction which can be attributed to the duty of the concentrator. As the only light to heat interaction occurs at high temperatures in the concentrator, exergy destruction at the subsystem is obtained high. The cost rates and the corresponding costs of exergy destructions for each subsystem are listed in Table 6.8. The PEC reactor has the highest investment cost due to the expensive materials (membrane and quartz glass) in the structure and high cost of electricity input. The TEG setup follows the PEC reactor with an investment cost rate of 0.1214 $/h. The least investment cost is observed in the Solar-PCM subsystem by less than 0.02 $/h. Since the TEG unit uses most of the heating potential of the Solar-PCM unit and

132 converts heat to electricity with an efficiency less than 10%, cost of exergy destruction for this element is obtained highest among the system (as seen in Fig.6.51).

10000

1000

100

Exergy destruction Exergy rate (W) 10

1 Concentrator + PCM TEG PEC

Fig. 6.50 Exergy destruction rates of the subsystems.

PEC TEG 32% 42%

Concentrator + PCM 26%

Fig. 6.51 The share of exergy destruction cost rate in each component of the integrated system.

133

Table 6.8 The exergoeconomic results of the components in the integrated system Sub- Capital Cost Rate of Exergy Exergoeconomic Annual systems Cost ($) Destruction 퐂̇퐝 ($/h) Factor-f (%) Investment Cost Rate 퐙̇ ($/h) TEG $3,200 0.03124 0.7953 0.1214 Solar-PCM $523 0.01983 0.5 0.01983 PEC $3,730 0.02376 0.8562 0.1415 Overall $7,453 0.07483 0.7907 0.2827

A parametric study is conducted to investigate the financial parameters such as interest rate or cost of electricity on the system components cost rates. Even though these parameters are independent of the material itself, they are directly related with the financial conditions of the region where the system is constructed. As shown in Fig. 6.52, the interest rate has an adverse effect on the system cost rates. While the total cost rate of the integrated system is 0.3664 $/h at 2% interest rate, it increases to 0.8759 $/h at 15% interest. The annual cost rate and the cost of exergy destruction of the system also have similar trends. Annual investment cost rate witnesses an increase from 0.28 $/h to 0.54 $/h with the increasing interest rate.

1 0.9 0.8

0.7 )

- 0.6 0.5

0.4 factor f ( ffactor 0.3 0.2 0.1

Cost rate ($/h) and Exergoeconomic Exergoeconomic and ($/h) rate Cost 0 1 3 5 7 9 11 13 15 Interest rate (%) Ct,ov Cd,ov Z,ov f,ov Fig. 6.52 The effects of the interest rate on the system cost rates and exergoeconomic factor.

134

Another parameter which has vital importance on the cost flow is the operation time of the system. The operation time depends on the parameters such as the energy source (solar light for this case), market demand and maintenance periods. The base value of the operation time is selected as 3000 h annually for the proposed system. As the system in this study has a thermal energy storage unit, it is able to operate continuously even after sunset. The Fig. 6.53 depicts the effects of annual operation time on total cost rates. All the cost rates drop drastically with the increasing operation time. More operation time shortens the capital recovery period. Therefore, it can be stated that energy storage is not only improving the system by increasing the system`s annual capacity, it also helps the system to quickly recover the capital investment The cost of heat and electricity is considered equal in this exergoeconomic assessment. Since the system is producing more electricity than it consumes, the increase in the cost of electricity should have a positive impact on the cost.

1.8 ) - 1.6

1.4

1.2

0.8

0.6

0.4

0.2

Cost rate ($/h) and Exergoeconomic factor ( factor and Exergoeconomic($/h) rateCost 1000 1500 2000 2500 3000 Anuual operation time (h) Ct,ov Cd,ov Z,ov f,ov

Fig. 6.53 The effects of the annual operation time on the system cost rates and exergoeconomic factor.

135

The effect of electricity/heating on the overall system can be seen in Fig. 6.54. Since the system heats up the excess amount of wastewater in the TEG unit, its heating potential is taken into account in the base case. The total cost rate of the overall system declines by 26% with the increasing electricity prices from 0.01 $/kWh to 0.15 $/kWh. Moreover, with the increasing cost of electricity and heat, the exergoeconomic factor of the system increases. The value of f witnesses an increase from 0.63 at 0.01 $/kWh till it reaches a maximum value of 0.86 at an electricity and heating cost of 0.15 $/kWh. It should be noted that the improvement of the cost rate includes the potential heating cost of the waste heat of the thermoelectric generator. For better cost management, the corresponding waste heat can be utilized in applications such as space heating, drying, etc.

1 0.9 0.8

0.7 ) - 0.6 0.5

factor f ( ffactor 0.4 0.3 0.2 0.1 Cost rate ($/h) and Exergoeconomic Exergoeconomic and ($/h) rate Cost 0 0.01 0.03 0.05 0.07 0.09 0.11 0.13 0.15 Electricity cost ($/kWh) Ct,ov Cd,ov Z,ov f,ov

Fig. 6.54 The effects of the electricity cost on the system cost rates and exergoeconomic factor.

6.7.1 Scale-up analysis results

The cost analysis of the scale-up system which produces 1000 kg of hydrogen and treats 222 m3 of wastewater per day. The main costs of the plant are tabulated in Table 6.9. The model takes into account all of the costs related to capital, decommissioning, salvage and

136 operation and maintenance. The costs are expressed in U.S dollars. The capital cost of the plant for hydrogen production and wastewater treatment is calculated as $7,714,878. Only around 2% of the capital cost is related with non-depreciable costs of the system.

Table 6.9. The costs of PEC plant Plant Costs* 2010$ Capital costs Depreciable portion 7,541,620 Non-Depreciable portion 173,258 Total investment 7,714,878 Cost of decommissioning 754,162 Plant salvage value 771,488 Annual fixed operating costs in $/year 280,028 * Inflated to start-up ** Not including tax credits and incentives Table 6.10 Fixed operating costs of the plant. Fixed Operating Costs 2010$ Plant staff presented as # of FTEs employed by the plant 0.36318 Burdened labor cost, including overhead in $/man-hr 53.56 Labor cost in $/year 40,459 G&A in $/year 8,092 Permits, fees and licensing costs in $/year 1,176.42 Insurance and property taxes in $/year 127,825 Production repairs and maintenance in $/year 24,929.20 Hydrogen peroxide cost in $/year 29,502.90 Summation 231,985

The model also takes into account indirect depreciable expenses such as site preparation, engineering and design processes, up-front permitting, and project contingency costs. Table 6.10 shows the fixed operating of the plant. The total yearly operating cost is projected $231,985. The property tax and insurance are mainly contributing to these expenditures by $127,825 per year. The labor and burdened labor cost are also responsible for the significant portion of the operational cost. In this regard, the direct capital cost of the system elements including the PEC cell body, concentrating and containment system, electrodes and condensers and sensors and monitoring systems are given in Table 6.11.

137

Table 6.11 The direct capital costs of the components in 1000 kg/day PEC concentrated light hydrogen production plant. Installed Major equipment Costs ($) Concentrating plexiglass and containment system 3,778,491 PEC Electrode 576,447 Pump for make-up water 250 Manifold piping 19,033 Collection Piping 4,716 Column Collection 2,225 Final Collection 507 Condenser 8,350 Manifold Piping (diameter) 19,033 Collection Piping (diameter) 4,716 Column Collection Piping (diameter) 2,225 Final Collection Piping (diameter) 507 34,831 PLC 3,529 The building of the control room 20,619 The control room wiring panel 3,529 Computer and visualization accessories 1,765 Monitoring Software 5,057 Controller for the water level 83,143 Sensors Hydrogen area 160,934 Pressure 7,306 Hydrogen flow meter 6,470 Wiring For instrument 477 Power applications 239 Conduit 7,135 Piping Installation 9,347 Reactor feed installation 75 Gas processing Subassembly install 2,505 Reactor Foundation & Erection 352,926 Control System Install 90,061 5,206,448

The concentrating Plexiglas and containment system cost is estimated $3,778,491 which forms the highest expense rate throughout the system components. This value is followed by the PEC electrode ($576,447) and reactor foundation and erection ($352,926) costs respectively. The required PEC electrode area to produce 1,053 kg of hydrogen and

138

233 m3 of treated water at the maximum capacity is obtained as 22,667 m2. The lowest costs are obtained as the costs of reactor feed installation, pump for make-up water, collection piping, wiring related instrument and power.

) 3

m 11.3

/ $

( 10 10.0

8 6.3 6 4.0 4.4 4 3.4 2 1.3 0.4 Cost of desalinated water water of desalinated Cost 0 Solar Collector Fuel based plants Wind RO PV RO Plants

Fig. 6.55 The price range of the water desalination system with various technologies (modified from [199]).

As it is depicted in Fig. 6.55 and Table 6.12, the cost of water directly extraction from ground and surface varies $0.40 to $0.75 per m3 in different regions. Moreover, for the desalination process, it can cost from $0.44 up to $11.29 per m3. Unlike the conventional hydrogen production plants, the proposed system also provided treated water as a by-product and sends the treated wastewater to the electrodialysis plant for desalination. The yearly benefit from wastewater is estimated $81,030 by considering the 3 revenue per cubic meter of treated wastewater as 1.0 $/m .

139

Table 6.12 Ground and surface waters extraction volumetic costs in different [199,200]. 3 Region Process Cost ($/m ) Surface – ground water The United States and production without 0.40–0.75 Europe distribution Surface – ground water Western Australia production with 0.45–0.61 distribution

Operating Capacity Factor (100%, 95%, 90%) 5.48 5.75 6.05 Plant Design Capacity (kg of H2/day) (1,106, 1,053, 1,000) 5.49 6.05 Total Capital Investment 5.58 5.92 ($6,072K, $6,391K, $6,711K) Costs of Costs of After-tax Real IRR H2 for 5% H for 5% 5.60 5.90 2 (10%, 10%, 11%) decrease increase in the in the Total Fixed Operating Cost parameter ($0,220K, $0,232K, $0,244K) 5.72 5.78 parameter Utilities Consumption (% of baseline) (95%, 100%, 105%) 5.75 5.75 Feedstock Consumption (% of baseline) (95%, 100%, 105%) 5.75 5.75

$5.0 $5.2 $5.4 $5.6 $5.8 $6.0 $6.2

Fig. 6.56. The sensitivity analysis for hydrogen cost rate ($/kg) with various parameters.

Fig. 6.56 shows the sensitivity of the 1 kg of hydrogen based on various parameters. As it can be seen in the figure, by the increasing plant capacity factor from 95% to 100%, it is possible to decrease the cost from 5.75 kg/$ to 5.48 kg/$. In contrast, for the case when the system operates with a capacity factor of 90% the unit price increases by 30 cents. Another important parameter affecting the cost is after-tax real interest rate. Changing this rate from around 10% to 11% cause a $0.3 increase in the cost. The waterfall diagram in Fig. 6.57 shows the cost of 1 kg of hydrogen after completing potential improvements through the system. By improving 5% the given steps in Fig. 6.57, it is possible to drop the unit price to $5.29.

140

$5.80

↓0.27 $5.70

$5.60

$5.50 ↓0.16 $5.40 $5.75

$5.30 ↓0.03 ↓0.00 $5.20 $5.29 $5.10

$5.00 Baseline Operating Total Capital Total Fixed Utilities Feedstock Adjusted Capacity Investment Operating Consumption Consumption Factor = -5% Cost = -5% = -5% = +5% = -5% Fig. 6.57. Cost of hydrogen the after system improvements.

6.8 The Optimization Study Results The optimization of an energy system consists of modifications to the system layout and designing parameters based on one or multiple designing goals. As the system performance is not only linked with a single parameter, a multiple parameter optimization will provide more accurate results to optimize the objectives. Maximizing the efficiency while minimizing the cost and environmental impact can be considered as an example of multiple objective optimization problems [177]. In this study Genetic Algorithm is selected to perform multi-objective optimization. The genetic algorithm based multi-objective optimization study is executed by using Engineering Equation Solver software. Three objective functions are considered here for optimization: sustainability index of the overall system (to be maximized), COD removal efficiency (to be maximized) and overall cost rate of the system(to be minimized). Optimization toolbox of Engineering Equation Solver is used to execute Generic Algorithm (GA) on the objective functions. The GA is performed for 64 generations with

141

32 individuals. Higher values of the maximum mutation lead the algorithm to deep search for the optimum point from distant locations whereas smaller values more focus on around current optimum. Most of the other parameters in GA is set as default and restricted from any kind of adjustment in EES software. [20].

Table 6.13 Multi-objective optimization results for the total cost rate and exergy destruction rate of the solar PCM subsystem including the sensitivities

Multi-Objective Best Multi-Objective Best Decision Exergy destruction ±20 Optimum Exergy destruction and parameter and Best Total Cost % Best Total Cost Rate Rate ̇ Interest rate, I Ct,PCM =0.0285 $/h (%) 1 1 Ė xd,PCM = 863 W ̇ Annual operation Ct,PCM =0.0285 $/h time, t (h) 4000 3200 Ė xd,PCM = 863 W Ċ t,PCM =0.0282 $/h Daily average ̇ Exd,PCM = 863 W ̇ solar irradiance, Ct,PCM =0.0282.2 $/h 2 Iav (kWh/m ) 3 3.6 Ė xd,PCM = 940 W ̇ Reference Ct,PCM =0.0282.2 $/h temperatire,T0 (K) 323.2 278.2 Ė xd,PCM = 1101 W

Initially, all the subsystems are optimized based on their performance criteria. While the exergy destruction rate is the main concern in the PCM-Solar system due to high temperatures, hydrogen production rate, COD removal efficiencies are more vital for the PEC reactor. Cost is considered one of the main parameters for all systems to be minimized. Weighting factors are used to combine multiple objective functions into a single objective function. 0.5 weighting factors are assigned for the functions to be maximized whereas - 0.5 is used for the functions, which are requested to be minimized, such as exergy destruction, energy losses and cost. The multi-objective optimization results for the Solar-PCM subsystem are tabulated in Table 6.13. The system is optimized based on the results giving the optimum cost rate and exergy destruction and losses rate. Furthermore, the sensitivity of the decision parameters within a range of 20% is presented in the same table. For the Solar-PCM subsystem, optimum results are obtained at the minimum interest rate, maximum operation time, and minimum solar irradiance and maximum ambient temperature.

142

Table 6.14 Multi-objective optimization results for the total cost rate and induced thermoelectricity, sustainability index of the TEG subsystem including the sensitivities Multi-Objective Multi-Objective Best Best Sustainability Sustainability Index, Index, Best Total Best Total Cost Rate, Cost Rate, and Decision parameter Optimum and Thermoelectricity ±20% Thermoelectricity Ċ t,TEG = 0.1118 $/h Ċ d,TEG =0.02450 $/h SITEG = 1.173 Interest rate, I (%) 1 0.012 Ẇ TEG = 627.5 W Ċ t,TEG = 0.1531 $/h Ċ d,TEG =0.04516 $/h Annual operation Ċ t,TEG = 0.1099 $/h SITEG = 1.173 ̇ time, t (h) 4000 Ċ d,TEG =0.02357 $/h 3200 WTEG = 627.5 W SITEG = 1.173 Ċ t,TEG = 0.1224 $/h ̇ WTEG = 627.5 W Ċ d,TEG =0.03612 Daily average solar $/h irradiance, Iav SITEG = 1.173 2 (kWh/m ) 8 6.4 Ẇ TEG = 502 W Ċ t,TEG = 0.161 $/h Ċ d,TEG =0.05314 $/h Reference SITEG = 1.177 temperatire,T0 (K) 278.2 323.2 Ẇ TEG = 547.6 W

Since less interest rate and more operation time have a positive impact on the CRF, the optimum financial results are obtained in the point closer to the lower bound of the interest rate and upper bound of the operation time. Moreover, less solar irradiance means, less entropy generation in the concentrator, which results in lower rates of exergy destructions. In the second case, the multi-object optimization study is executed for the TEG subsystem. An optimum global point is obtained which considers the best possible values for total cost rate, cost of exergy destruction rate, induced thermoelectricity, and sustainability index. The findings of the second case are tabulated in Table 6.14.

143

Table 6.15 Multi-objective optimization results for the total cost rate, sustainability index, COD removal rate and hydrogen production rate of the PEC subsystem including the sensitivities Multi-Objective Best Sustainability Index, Best Total Multi-Objective Best Cost Rate, Best Sustainability Index, Best Hydrogen Total Cost Rate, Best Production Rate, Hydrogen Production Best COD Decision Rate, Best COD Removal Removal parameter Optimum Efficiency ±20% Efficiency Ċ t,PEC = 0.06491 $ /h SIPEC = 1.516

ṁ H2 = 11.06 10−3 mg⁄s Interest rate, I CODremoval (%) 1 0.012 = 100%

Ċ t,PEC = 0.1133 $ /h SIPEC = 1.516

ṁ H2 Annual = 11.06 10−3 mg⁄s operation time, t CODremoval (h) 4000 3200 = 100%

Ċ t,PEC = 0.06706 $ /h Ċ = 0.06277 $/h t,PEC SI = 1.558 SI = 1.516 PEC PEC ṁ ṁ = 11.06 10−3 mg⁄s H2 H2 = 10.72 10−3 mg⁄s CODremoval = 100% Photoanode 0.024 CODremoval Surface (m2) 0.03 m2 m2 = 94.38 %

Ċ t,PEC = 0.06706 $ /h SIPEC = 1.551

ṁ H2 Daily average = 10.72 10−3 mg⁄s solar irradiance, CODremoval 2 Iav (kWh/m ) 8 6.4 = 94.38 %

Ċ t,PEC = 0.09959 $ /h SIPEC = 1.558

ṁ H2 PEC operating = 8.12 10−3 mg⁄s temperatire,TPE CODremoval C (K) 353.2 298.2 = 100%

144

Unlike the PCM-Solar subsystem, TEG unit requires lower ambient temperature (which is also the temperature of the wastewater stream used as a coolant in the TEG unit) for a better thermoelectric yield. Table 6.16 Multi-objective optimization results of the integrated system including the sensitivities Multi-Objective Best Multi-Objective Best Sustainability Index, Sustainability Index, Best Total Cost Rate, Best Total Cost Rate, Decision ±20 Optimum Best Hydrogen Best Hydrogen parameter % Production Rate, Production Rate, Best Best COD Removal COD Removal Efficiency Efficiency Ċ t,ov = 0.213 $/h Ċ d,ov = 0.009754 $/h SIov = 1.054 −3 ⁄ ṁ H2 = 11.1 10 mg s Interest rate, I CODremoval = 100 % (%) 1 1.2 Ẇ TEG = 547.6 W Ċ t,ov = 0.3092 $/h Ċ d,ov = 0.05785 $/h SIov = 1.054 −3 ⁄ ṁ H2 = 11.1 10 mg s Annual operation CODremoval = 100 % Ċ = 0.2086 $/h time, t (h) 4000 t,ov 3200 Ẇ TEG = 547.6 W Ċ = 0.007591 $/h d,ov Ċ t,ov = 0.2179 $/h SI = 1.054 ov Ċ d,ov = 0.01189 $/h ṁ H 2 SIov = 1.054 = 11.1 10−3 mg⁄s −3 ṁ H = 10.7 10 mg⁄s COD = 100 % 2 Photoanode removal 0.024 COD = 100 % ̇ removal Surface (m2) 0.03 m2 WTEG = 547.6 W m2 Ẇ = 547.6 W TEG Ċ t,ov = 0.2239 $/h Ċ d,ov = 0.02284 $/h SIov = 1.053 −3 ⁄ Daily average ṁ H2 = 10.7 10 mg s solar irradiance, CODremoval = 94.49 % 2 Iav (kWh/m ) 8 6.4 Ẇ TEG = 438.1 W Ċ t,ov = 0.2188 $/h Ċ d,ov = 0.01775 $/h SIPEC = 1.058 −3 ⁄ ṁ H2 = 9.91 10 mg s Reference CODremoval = 97.03 % temperatire,T0 (K) 323.2 298.2 Ẇ TEG = 592 W

145

The temperature difference between hot and cold sides enhances the efficiency and capacity of the TEG unit.627.5 W is induced in the global optimum point of the TEG unit.

Sustainability index (SITEG)and the total cost rate Ċ t,TEG of TEG unit is calculated as 1.173 and 0.1099 $/h respectively. For the third case, a multi-objective optimization analysis is performed for the PEC reactor by considering the best total cost rate, hydrogen production rate, COD removal efficiency, and sustainability index. Table 6.15 presents the results of the case. Even though higher photo anode surfaces leads more dissipated heat which results in exergy losses, the positive impact of it on both hydrogen production rates and COD removal efficiencies. Hence, higher values of the photoanode surfaces are favored for better hydrogen production rates. Also, solar irradiance shows a positive impact to the COD removal efficiency results. As it is explained in section 5.3, UV-light enhances the COD removal rate for the Fenton process. 100% removal rate is obtained for the case where the operating temperature is at 80°C and the average solar irradiance is at 8 kWh/m2. Similar to the other results, lower interest rate and greater operational time diminish the total cost rate of the system. At the global optimum point, sustainability index (SITEG)and the total cost rate Ċ t,TEG of photoelectron chemical reactor is calculated as 1.516 and 0.06277 $/h respectively. The rest of the analysis results can be seen in Table 6.15 For the last case, a global optimum point is obtained for the integrated system which takes into account, best conditions for system`s sustainability index, hydrogen production rate, total cost rate, induced thermoelectric power, and COD removal rate. Table 6.16 presents the results of the optimization study. Similar to the previous results, a lower interest rate and higher operating times are favored due to better capital recovery factor. Moreover, at higher temperatures it is observed that system has better hydrogen production rates. It can be attributed to the decrease in Gibbs free energy for decomposing water molecules.

146

650 1.060

600 1.058 550

500 1.056

450

(W) 1.054

TEG 400

350 1.052

Sustainability Index (-) Index Sustainability

300 1.050 250

1.048 200 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 (17434) (17434) Iterations Iterations (a) (b)

1.2e-5 0.056

1.1e-5 0.054

1.0e-5 0.052

9.0e-6

0.050

8.0e-6 efficiency Exergy

0.048

Hydrogen production rate (kg/s) Hydrogen 7.0e-6

0.046 6.0e-6 0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400 (17434) (17434) Iterations Iterations (c) (d)

1.0

100 0.9

0.8 90

0.7 80

0.6

70 0.5

Total Cost Rate, $/h)( 0.4 efficiency (%) COD_removal 60

0.3 50 0.2 0 50 100 150 200 250 300 350 400 (17434) 0 50 100 150 200 250 300 350 400 Iterations (17434) Iterations (e) (f)

2.0 a) Thermoelectric generator work 1.8

1.6 b) Sustainability index

1.4 c) Hydrogen production rate 1.2 d) Exergy efficiency 1.0 e) Total cost rate 0.8

0.6 f) COD removal efficiency

Value of the functionobjective 0.4 g) Defined objective function of the system 0.2 0 50 100 150 200 250 300 350 400 Iterations (g) Fig. 6.58 The variation of the objective function and system outputs through the iterations of the optimization study.

147

Less requirement for the Gibbs free energy, in turn, reduces the electrical potential needed to break the water molecule bond to produce the hydrogen based on the Nernst equation. On the other hand, higher ambient temperatures heat up the wastewater stream which is used as a coolant of the TEG to induce thermoelectricity. As the TEG unit cannot be cooled as sufficiently with the high-temperature water, the thermoelectric efficiency and generated thermoelectricity declines with the increasing temperature. For this scenario at the global optimum point, total cost rate, sustainability index and hydrogen production rate of the system are obtained as 0.2086 $/kWh, 1.054 and 11.1 10-3 mg/s respectively. In this case, all the weighting factors are considered equal. For the alternative scenarios where the hydrogen production rate is favored compared to cost, weighting factors can be adjusted accordingly. Fig. 6.58 shows the variation of the objective function and the system through the development of an optimization study. The solution is converged through the study at the 17434 iterations, however, to save space and provide a readable figure only 400 iterations are shown in the graph with the initial 300 iterations and the last 100 iterations from a total of 17,434. Fig. 6.58 (a) and 6.58 (b) shows similar behavior in the achieving the optimum, which indicate a close relationship between the production of power from the thermoelectric generator and the sustainability index. Similar behavior can also be seen in the exergy efficiency (Fig. 6.58 (d)), this trend is expected as the sustainability index is a function of exergy efficiency and highly dependent on it. Fig. 6.586 (c) shows that the development and the evolving of the hydrogen production rate are achieved at the end of the study, which is similar to the minimization of the system total cost rate in Fig. 6.58. (e).

148

CHAPTER 7: CONCLUSIONS AND RECOMMENDATIONS

In this chapter, the main findings derived from this thesis are summarized and briefly explained. Based on the gained experience and obtained results, further recommendations are also proposed for future studies in this research area.

7.1 Conclusions The aim of this Ph.D. thesis study is to develop and analyze a solar energy driven system for hydrogen, clean water and electricity production. The primary objectives of this Ph.D. study are set after a comprehensive literature review on wastewater treatment and bio- hydrogen manufacturing technologies. Photo electro-Fenton reaction appears to be a desirable option amongst the existing hydrogen production and wastewater treatment systems. The main two advantages of this reaction are the high oxidation capacity and lower energy requirement. Photocatalytic materials could enhance this process. A photoelectrochemical process is proposed, which treats the wastewater via Fenton-based oxidation and produced hydrogen. The reactor is driven by a solar – thermoelectric battery pack. The proposed experimental setup for the suggested system is introduced, and all the elements are described individually. Experiments are conducted in the Clean Energy Research Lab (CERL) under several operating parameters such as wastewater type, wastewater temperature, applied current and voltage, under light and no light conditions. All results are recorded and plotted using a data acquisition system. The chemical oxygen demand is used to quantify the amount of oxygen needed for the chemical oxidation of all organic substances in water. Optimization of the study is also carried out for the overall system and subsystems. There are three main contributions from this thesis study. First of all, photocatalysts assisted photoelectron-Fenton process in a PEM electrolyzer is used for the first time in the literature for both hydrogen production and water treatment. Secondly, the synergetic effects of advanced oxidation processes (AOP) such as Fenton, Fenton-like, photocatalysis

(TiO2/UV) and UV photolysis (H2O2/UV) are investigated individually and in a combination of each other for the first time in the literature. The influence of operating conditions including pH level, type of the electrode and electrolyte and the UV light, on

149 the performance of the combined system are studied to design more effective systems. Lastly, a novel TEG-PCM system is designed and theoretically analyzed, providing stable thermoelectricity output, eliminating the requirement of a voltage regulator and acting as an electrolyte heater, which enhances the hydrogen production and wastewater treatment efficiencies. The main findings obtained from the experimental part of this thesis study can be summarized as follows:

• The energy and exergy efficiencies of the single cell PEC reactor under 1.4 V of cell potential and dark conditions are obtained as 72.85% and 71.69% respectively. Under 600 W/m2 of artificial solar light illumination, both energy and exergy efficiencies of the respective reactor are obtained as 88.46% and 87.05% respectively.

• The maximum hydrogen production rate of the PEC reactor is obtained as 33.9 mg/h at 2.3 V of cell potential and 60°C operating temperature.

• The COD removal efficiency of the PEC reactor rises from 44% to 87% with the increasing temperature from 20°C to 60°C.

• The hydrogen production rate of the PEC reactor rises from 26.7 mg/h to 33.9 mg/h with the increasing temperature from 20°C to 60°C.

• Under 600W/m2 of the solar simulator irradiance, the maximum induced photocurrent

from the TiO2 coated stainless steel anode is observed as 10 mA.

• For the single compartment reactor, the optical absorbance of the treated samples decreases sharply in the first 17 hours, and the decolorization is completed in 48 hours.

• In the small scale-experiments, the highest COD degradation rate is obtained as 97.9%

under the experimental condition, which combines UV/TiO2, UV/H2O2, and photoelectro Fenton type processes.

• In the small scale experiments, the highest H2 production rates from the electrolysis of KHP (6.59 mg/Wh) and RB5 (7.00 mg/Wh) are obtained under the experimental conditions which have the highest rate of photo-electro Fenton type processes.

150

• A maximum 48.5% improvement in the COD removal efficiency is observed for the

experimental condition where KHP solution is used as the electrolyte and TiO2 coated stainless steel is used as the photoanode.

• The graphite rod under the positive charge adsorbs RB5-based textile wastewater, which assists the decolorization process.

• The TiO2 coating on the stainless steel anode restricts the electrogeneration of the iron ions which limits Fenton-type processes in the reactor. Less than 0.5 ppm of iron sacrificial is observed in the electrolyzer after 40 minutes of operation under 500 mA of electric current.

• The KHP solution has 20.7% of higher COD removal in the electrolyzer compared to

RB5-based textile wastewater (under UV light, 0.5 A electric current with TiO2 coated photoanode and graphite cathode).

The following main findings are obtained from the thesis study focussed on the present conceptually developed integrated system:

• The overall energy and exergy efficiencies of the integrated system under 2.3 V of cell potential, 25°C of ambient temperature, and 1 atm of ambient pressure are calculated as 5.5% and 5.2% respectively.

• The thermoelectric generators working around 250oC of hot side and 50°C of cold side temperature could generate 1.8 V of electric potential, which is sufficient for running the PEC reactor.

• With the increasing temperature of the cold stream (wastewater flow) from 5°C to 45°C, the induced thermoelectricity drops from 642 W to 474 W, and hence the energy efficiency of the TEG decreases from 7.87% to 6.98%, respectively.

• The capital cost of the lab scale integrated system is mainly dominated by the cost of TEG pack, and the PtB-IrRuOx coated Nafion membrane which corresponds to 40% and 27% of the total cost respectively.

151

• For the case where the system annual operation time is taken to be 1000 hours for the lab scale system, the total calculated cost rate is about 1.609 $/h. For an annual operation, the time increases up to 3000 hours where the total cost rate may decrease down to 0.3977 $/h.

• In the scale-up analyses, the concentrated light based PEC hydrogen production plant

in which the design capacity of H2 production is considered as 1,053 kg/day, the cost of hydrogen is calculated to be 5,75 $/kg, respectively.

• Under the optimum conditions, the total cost rate (Ċ t,ov), the sustainability index

(SIov), the exergy efficiency (ηex), the hydrogen production rate (ṁ H2), the COD

removal efficiency (ηCOD), and the generated thermoelectric generator power (Ẇ TEG) of the system are obtained as 0.2086 $/h, 1.054, 5.1%, 11.1 × 10-3 mg/s, 100% and 547.6 W, respectively.

• The proposed system is easily able to treat 180 liters of wastewater and produce 15 kWh of electricity on a daily basis. Note that these numbers correspond to daily wastewater production and electricity consumption per person.

7.2 Recommendations This section focuses on providing recommendations for further studies that might follow the presented work in this thesis. The work presented in this thesis investigates the photoelectrochemical hydrogen production and wastewater degradation in a single reactor under various conditions. In order to develop the study to a wider perspective and expand the usage opportunities for the proposed systems, the recommendations are listed below:

• A scaled-up integrated system, which is capable of treating farmhouse, house, or factory wastes, should be tested experimentally. This system should employ the hybrid PEC reactor integrated with PCM and TEG. • A comprehensive life cycle assessment study for the proposed system is necessary to confirm that overall life cycle emission of the proposed system is minimal compared to its alternatives

• Alternative photocatalyst coating methods, such as electrodeposition, spin coating,

152

vapor deposition and spray pyrolysis should be tested in order to determine if sol-gel dip coating is the most suitable method on the photoanode.

• Doping materials should be investigated to enhance the photoabsorption, hydroxyl radical generation, and overall solar-to-hydrogen efficiency of the experimental PEC system.

• In-situ production of hydrogen peroxide should be implemented to the PEC reactor to reduce the cost and eliminate the possible risks of transportation and handling processes of the concentrated hydrogen peroxide.

• The excess iron ions which are electro-generated for Fenton type processes in the anolyte should be recovered.

• Various types of wastewater from various industrial sectors should be examined to verify the performance of the system for all types of wastes.

153

REFERENCES

[1] Worldbank, Renewable internal freshwater resources per capita n.d. http://data.worldbank.org/indicator/ER.H2O.INTR.PC?end=2014&start=1966 (accessed May 9, 2017). [2] Worldbank, Energy use per capita n.d. http://data.worldbank.org/indicator/EG.USE.PCAP.KG.OE (accessed May 9, 2017). [3] World Bank. Fossil fuel energy consumption. Data, Energy Foss Fuels Consum 2016. http://data.worldbank.org/indicator/EG.USE.COMM.FO.ZS. [4] REN21. Renewables 2015-Global Status Report. vol. 4. 2015. doi:10.1016/0267- 3649(88)90030-1. [5] Acar C, Dincer I, Naterer GF. Review of photocatalytic water-splitting methods for sustainable hydrogen production. Int J Energy Res 2016;40:1449–73. doi:10.1002/er.3549. [6] Dincer I, Zamfirescu C. Sustainable hydrogen production options and the role of IAHE. Int J Hydrogen Energy 2012;37:16266–86. doi:10.1016/j.ijhydene.2012.02.133. [7] Hydrogen Analysis Resource Center n.d. http://hydrogen.pnl.gov/tools/lower-and- higher-heating-values-fuels (accessed December 14, 2016). [8] Hosseini SE, Wahid MA. Hydrogen production from renewable and sustainable energy resources: Promising green energy carrier for clean development. Renew Sustain Energy Rev 2016;57:850–66. doi:10.1016/j.rser.2015.12.112. [9] Navarro-Solís I, Villalba-Almendra L, Alvarez-Gallegos A. H2 production by PEM electrolysis, assisted by textile effluent treatment and a solar photovoltaic cell. Int J Hydrogen Energy 2010;35:10833–41. doi:10.1016/j.ijhydene.2010.07.086. [10] Ewan BCR, Allen RWK. A figure of merit assessment of the routes to hydrogen. Int J Hydrogen Energy 2005;30:809–19. doi:10.1016/j.ijhydene.2005.02.003. [11] Stoll RE, von Linde F. Hydrogen-what are the costs? Hydrocarb Process 2000:42– 6. [12] Tugnoli A, Landucci G, Cozzani V. Sustainability assessment of hydrogen production by steam reforming. Int J Hydrogen Energy 2008;33:4345–57. doi:10.1016/j.ijhydene.2008.06.011. [13] Dincer I, Zamfirescu C. Sustainable Hydrogen Production. vol. 305. 1st ed. Amsterdam: Elsevier; 2016. [14] Dincer I. Green methods for hydrogen production. Int J Hydrogen Energy 2012;37:1954–71. doi:10.1016/j.ijhydene.2011.03.173. [15] Hosseini SE, Wahid MA. Development of biogas combustion in combined heat and power generation. Renew Sustain Energy Rev 2014;40:868–75. doi:10.1016/j.rser.2014.07.204. [16] Caputo AC, Palumbo M, Pelagagge PM, Scacchia F. Economics of biomass energy utilization in combustion and gasification plants: effects of logistic variables. Biomass and Bioenergy 2005;28:35–51. doi:10.1016/j.biombioe.2004.04.009. [17] Acar C. Experminental Investigation and Analyses of Continuous Type Hybrid Photoelectrochemical Hydrogen. University of Ontario Institute of Technology, 2016.

154

[18] Biomass Feedstocks, Department of Energy n.d. https://energy.gov/eere/bioenergy/biomass-feedstocks (accessed June 7, 2017). [19] Acar C, Ghosh S, Dincer I, Zamfirescu C. Evaluation of a new continuous type hybrid photo-electrochemical system. Int J Hydrogen Energy 2015;40:11112–24. doi:10.1016/j.ijhydene.2014.12.036. [20] Bicer Y. Investigation of Novel Ammonia Production Options Using Photoelectrochemical Hydrogen. University of Ontario Institute of Technology, 2017. [21] Clark JH, Deswarte FEI. Introduction to chemicals from biomass. 1st ed. Cornwall: John Wiley & Sons, Ltd.; 2015. [22] Denholm P, Kulcinski GL, Holloway T. Emissions and energy efficiency assessment of baseload wind energy systems. Environ Sci Technol 2005;39:1903– 11. doi:10.1021/es049946p. [23] Balta MT, Dincer I, Hepbasli A. Geothermal-based hydrogen production using thermochemical and hybrid cycles: A review and analysis. Int J Energy Res 2009;34:757–75. doi:10.1002/er.1589. [24] Dincer I, Zamfirescu C. Sustainable Energy Systems and Applications. vol. 6. 2011. doi:10.1007/978-0-387-95861-3. [25] Steinfeld A. Solar thermochemical production of hydrogen –– a review 2005;78:603–15. doi:10.1016/j.solener.2003.12.012. [26] Joshi AS, Dincer I, Reddy B V. Solar hydrogen production: A comparative performance assessment. Int J Hydrogen Energy 2011;36:11246–57. doi:10.1016/j.ijhydene.2010.11.122. [27] Dincer I, Joshi AS. Solar based hydrogen production systems. 2013. doi:10.1007/978-1-4614-7431-9. [28] Baykara SZ. Experimental solar water thermolysis 2004;29:1459–69. doi:10.1016/j.ijhydene.2004.02.011. [29] Division S, Faculty ES. Renewable hydrogen production . The role of Solar Thermal Water Gabriel Vicens 1990. [30] Naterer GF, Suppiah S, Stolberg L, Lewis M, Ferrandon M, Wang Z, et al. Clean hydrogen production with the Cu e Cl cycle e Progress of international consortium , I : Experimental unit operations 2011;6. doi:10.1016/j.ijhydene.2011.08.012. [31] International Decade for Action “Water for Life” 2005-2015. Focus Areas: Water scarcity n.d. [32] Dincer I, Acar C. A review on clean energy solutions for better sustainability. Int J Energy Res 2015;39:585–606. doi:10.1002/er.3329. [33] Das D. Hydrogen production by biological processes: a survey of literature. Int J Hydrogen Energy 2001;26:13–28. doi:10.1016/S0360-3199(00)00058-6. [34] Cardoso V, Romão BB, Silva FTM, Santos JG, Batista FRX, Ferreira JS. Hydrogen Production by Dark Fermentation. Chem Eng Trans 2014;38:481–6. doi:10.3303/CET1438081. [35] Guo XM, Trably E, Latrille E, Carrre H, Steyer JP. Hydrogen production from agricultural waste by dark fermentation: A review. Int J Hydrogen Energy 2010;35:10660–73. doi:10.1016/j.ijhydene.2010.03.008. [36] Levin DB, Pitt L, Love M. Biohydrogen production: Prospects and limitations to practical application. Int J Hydrogen Energy 2004;29:173–85. doi:10.1016/S0360-

155

3199(03)00094-6. [37] Li D, Chen H. Biological hydrogen production from steam-exploded straw by simultaneous saccharification and fermentation. Int J Hydrogen Energy 2007;32:1742–8. doi:10.1016/j.ijhydene.2006.12.011. [38] Datar R, Huang J, Maness P-C, Mohagheghi A, Czernik S, Chornet E. Hydrogen production from the fermentation of corn stover biomass pretreated with a steam- explosion process. Int J Hydrogen Energy 2007;32:932–9. doi:10.1016/j.ijhydene.2006.09.027. [39] Karlsson A, Vallin L, Ejlertsson J. Effects of temperature, hydraulic retention time and hydrogen extraction rate on hydrogen production from the fermentation of food industry residues and manure. Int J Hydrogen Energy 2008;33:953–62. doi:10.1016/j.ijhydene.2007.10.055. [40] Noike T, Mizuno O. Hydrogen fermentation of organic municipal wastes. Water Sci Technol 2000;42. [41] Okamoto M, Miyahara T, Mizuno O, Noike T. Biological hydrogen potential of materials characteristic of the organic fraction of municipal solid wastes. Water Sci Technol 2000;41. [42] Kim S-H, Han S-K, Shin H-S. Feasibility of biohydrogen production by anaerobic co-digestion of food waste and sewage sludge. Int J Hydrogen Energy 2004;29:1607–16. doi:10.1016/j.ijhydene.2004.02.018. [43] Lay J-J, Fan K-S, Hwang J-I, Chang J-I, Hsu P-C. Factors Affecting Hydrogen Production from Food Wastes by Clostridium Rich Composts. J Environ Eng 2005;131:595–602. doi:10.1061/(ASCE)0733-9372(2005)131:4(595). [44] Akroum-Amrouche D, Abdi N, Lounici H, Mameri N. Biohydrogen production by dark and photo-fermentation processes. Proc 2013 Int Renew Sustain Energy Conf IRSEC 2013 2013:499–503. doi:10.1109/IRSEC.2013.6529679. [45] Gandía LM, Arzamendi G, Diéguez PM. Renewable Hydrogen Energy: An Overview. 2013. doi:10.1016/B978-0-444-56352-1.00001-5. [46] Yetis M, Gündüz U, Eroglu I, Yücel M, Türker L. Photoproduction of hydrogen from sugar refinery wastewater by Rhodobacter sphaeroides O.U. 001. Int J Hydrogen Energy 2000;25:1035–41. doi:10.1016/S0360-3199(00)00027-6. [47] Das D. Hydrogen production by biological processes: a survey of literature. Int J Hydrogen Energy 2001;26:13–28. doi:10.1016/S0360-3199(00)00058-6. [48] Holladay JD, Hu J, King DL, Wang Y. An overview of hydrogen production technologies. Catal Today 2009;139:244–60. doi:10.1016/j.cattod.2008.08.039. [49] Ni M, Leung DYC, Leung MKH, Sumathy K. An overview of hydrogen production from biomass. Fuel Process Technol 2006;87:461–72. doi:10.1016/j.fuproc.2005.11.003. [50] Kırtay E. Recent advances in production of hydrogen from biomass. Energy Convers Manag 2011;52:1778–89. doi:10.1016/j.enconman.2010.11.010. [51] Argun H, Kargi F. Bio-hydrogen production by different operational modes of dark and photo-fermentation: An overview. Int J Hydrogen Energy 2011;36:7443–59. doi:10.1016/j.ijhydene.2011.03.116. [52] Kapdan IK, Kargi F. Bio-hydrogen production from waste materials. Enzyme Microb Technol 2006;38:569–82. doi:10.1016/j.enzmictec.2005.09.015. [53] Yu J, Takahashi P. Biophotolysis-based Hydrogen Production by Cyanobacteria and

156

Green Microalgae. Commun Curr Res Educ Top Trends Appl Microbiol 2007;1:79–89. [54] Fascetti E, D’Addario E, Todini O, Robertiello A. Photosynthetic hydrogen evolution with volatile organic acids derived from the fermentation of source selected municipal solid wastes. Int J Hydrogen Energy 1998;23:753–60. doi:10.1016/S0360-3199(97)00123-7. [55] Zhang Y, Van der Bruggen B, Pinoy L, Meesschaert B. Separation of nutrient ions and organic compounds from salts in RO concentrates by standard and monovalent selective ion-exchange membranes used in electrodialysis. J Memb Sci 2009;332:104–12. doi:10.1016/j.memsci.2009.01.030. [56] Fang H, Zhu H, Zhang T. Phototrophic hydrogen production from glucose by pure and co-cultures of Clostridium butyricum and Rhodobacter sphaeroides. Int J Hydrogen Energy 2006;31:2223–30. doi:10.1016/j.ijhydene.2006.03.005. [57] Basak N, Das D. Photofermentative hydrogen production using purple non-sulfur bacteria Rhodobacter sphaeroides O.U.001 in an annular photobioreactor: A case study. Biomass and Bioenergy 2009;33:911–9. doi:10.1016/j.biombioe.2009.02.007. [58] Eroǧlu I, Tabanoǧlu A, Gündüz U, Eroǧlu E, Yücel M. Hydrogen production by Rhodobacter sphaeroides O.U.001 in a flat plate solar bioreactor. Int J Hydrogen Energy 2008;33:531–41. doi:10.1016/j.ijhydene.2007.09.025. [59] Kapdan IK, Kargi F, Oztekin R, Argun H. Bio-hydrogen production from acid hydrolyzed wheat starch by photo-fermentation using different Rhodobacter sp. Int J Hydrogen Energy 2009;34:2201–7. doi:10.1016/j.ijhydene.2009.01.017. [60] Wu YN, Wen HQ, Zhu JN, Ding J, Ren NQ, Liu BF. Best mode for photo- fermentation hydrogen production: The semi-continuous operation. Int J Hydrogen Energy 2016;41:16048–54. doi:10.1016/j.ijhydene.2016.05.182. [61] Akbari L, Mahmoodzadeh Vaziri B. Comprehensive modeling of photo- fermentation process for prediction of hydrogen production. Int J Hydrogen Energy 2017;42:14449–57. doi:10.1016/j.ijhydene.2017.04.119. [62] Logan BE, Call D, Cheng S, Hamelers HVM, Sleutels THJA, Jeremiasse AW, et al. Microbial Electrolysis Cells for High Yield Hydrogen Gas Production from Organic Matter. Environ Sci Technol 2008;42:8630–40. doi:10.1021/es801553z. [63] Meda US. Bio-Hydrogen Production in Microbial Electrolysis Cell Using Waste Water From Sugar Industry. ResearchGate 2015;4:452–8. [64] Guo K, Tang X, Du Z, Li H. Hydrogen production from acetate in a cathode-on-top single-chamber microbial electrolysis cell with a mipor cathode. Biochem Eng J 2010;51:48–52. doi:10.1016/j.bej.2010.05.001. [65] Kadier A, Simayi Y, Abdeshahian P, Azman NF, Chandrasekhar K, Kalil MS. A comprehensive review of microbial electrolysis cells (MEC) reactor designs and configurations for sustainable hydrogen gas production. Alexandria Eng J 2014;55:427–43. doi:10.1016/j.aej.2015.10.008. [66] Liu H, Grot S, Logan BE. Electrochemically Assisted Microbial Production of Hydrogen from Acetate 2005. doi:10.1021/ES050244P. [67] Cheng S, Logan BE. Sustainable and efficient biohydrogen production via electrohydrogenesis. Proc Natl Acad Sci U S A 2007;104:18871–3. doi:10.1073/pnas.0706379104.

157

[68] Kyazze G, Popov A, Dinsdale R, Esteves S, Hawkes F, Premier G, et al. Influence of catholyte pH and temperature on hydrogen production from acetate using a two chamber concentric tubular microbial electrolysis cell. Int J Hydrogen Energy 2010;35:7716–22. doi:10.1016/j.ijhydene.2010.05.036. [69] Call D, Logan BE. Hydrogen Production in a Single Chamber Microbial Electrolysis Cell Lacking a Membrane. Environ Sci Technol 2008;42:3401–6. doi:10.1021/es8001822. [70] Hu H, Fan Y, Liu H. Hydrogen production using single-chamber membrane-free microbial electrolysis cells. Water Res 2008;42:4172–8. doi:10.1016/j.watres.2008.06.015. [71] Zhang Y, Angelidaki I. Microbial electrolysis cells turning to be versatile technology: Recent advances and future challenges. Water Res 2014;56:11–25. doi:10.1016/j.watres.2014.02.031. [72] Call D, Logan BE. Hydrogen production in a single chamber microbial electrolysis cell lacking a membrane. Environ Sci Technol 2008;42:3401–6. doi:10.1021/es8001822. [73] Stoner GE, Cahen GL, Sachyani M, Gileadi E. The mechanism of low frequency a.c. electrochemical disinfection. Bioelectrochemistry Bioenerg 1982;9:229–43. doi:10.1016/0302-4598(82)80013-5. [74] Patermarakis G, Fountoukidis E. Disinfection of water by electrochemical treatment. Water Res 1990;24:1491–6. doi:10.1016/0043-1354(90)90083-I. [75] Grahl T, Märkl H. Killing of microorganisms by pulsed electric fields. Appl Microbiol Biotechnol 1996;45:148–57. [76] Cano A, Cañizares P, Barrera-Díaz C, Sáez C, Rodrigo MA. Use of conductive- diamond electrochemical-oxidation for the disinfection of several actual treated wastewaters. Chem Eng J 2012;211–212:463–9. doi:10.1016/j.cej.2012.09.071. [77] Li XY, Ding F, Lo PSY, Sin SHP. Electrochemical Disinfection of Saline Wastewater Effluent. J Environ Eng 2002;128:697–704. doi:10.1061/(ASCE)0733- 9372(2002)128:8(697). [78] Jeon B-H, Dempsey BA, Royer RA, Burgos WD. Low-Temperature Oxygen Trap for Maintaining Strict Anoxic Conditions. J Environ Eng 2004;130:1407–10. doi:10.1061/(ASCE)0733-9372(2004)130. [79] Huang X, Qu Y, Cid CA, Finke C, Hoffmann MR, Lim K, et al. Electrochemical disinfection of toilet wastewater using wastewater electrolysis cell. Water Res 2016;92:164–72. doi:10.1016/j.watres.2016.01.040. [80] Kim J, Choi WJK, Choi J, Hoffmann MR, Park H. Electrolysis of urea and urine for solar hydrogen. Catal Today 2013;199:2–7. doi:10.1016/j.cattod.2012.02.009. [81] Bergmann MEH, Rollin J. Product and by-product formation in laboratory studies on disinfection electrolysis of water using boron-doped diamond anodes. Catal Today 2007;124:198–203. doi:10.1016/j.cattod.2007.03.038. [82] Bergmann H, Koparal AT, Koparal AS, Ehrig F. The influence of products and by- products obtained by drinking water electrolysis on microorganisms. Microchem J 2008;89:98–107. doi:10.1016/j.microc.2007.12.007. [83] Gupta VK, Ali I, Saleh TA, Nayak A, Agarwal S. Chemical treatment technologies for waste-water recycling—an overview. RSC Adv 2012;2:6380. doi:10.1039/c2ra20340e.

158

[84] Gottberg V, Antonia JM, Siwak LR. Electrodialysis reversal process. Int Desalin Water Reuse Q 1998;7:33–7. [85] Tongwen X. Electrodialysis processes with bipolar membranes (EDBM) in environmental protection - A review. Resour Conserv Recycl 2002;37:1–22. doi:10.1016/S0921-3449(02)00032-0. [86] Singh AK, Shahi VK. Electro-Membrane Processes. Encycl. Membr. Sci. Technol., Hoboken, NJ, USA: John Wiley & Sons, Inc.; 2013. doi:10.1002/9781118522318.emst084. [87] Jakobsen M, Fritt-Rasmussen J, Nielsen S, Ottosen LM. Electrodialytic removal of cadmium from wastewater sludge. J Hazard Mater 2004;106:127–32. doi:10.1016/j.jhazmat.2003.10.005. [88] Ribeiro AB, Mateus EP, Ottosen LM, Bech-Nielsen G. Electrodialytic Removal of Cu, Cr, and As from Chromated Copper Arsenate-Treated Timber Waste. Environ Sci Technol 2000;34:784–8. doi:10.1021/es990442e. [89] Ashu R, Rugiero E, Chanda D, Hnàt J, Baak W Van, Veerman J, et al. Salinity gradient power-reverse electrodialysis and alkaline polymer electrolyte water electrolysis for hydrogen production 2016;514:155–64. doi:10.1016/j.memsci.2016.04.067. [90] Garcia-Segura S, Eiband MMSG, de Melo JV, Martínez-Huitle CA. Electrocoagulation and advanced electrocoagulation processes: A general review about the fundamentals, emerging applications and its association with other technologies. J Electroanal Chem 2017;801:267–99. doi:10.1016/j.jelechem.2017.07.047. [91] Hakizimana JN, Gourich B, Chafi M, Stiriba Y, Vial C, Drogui P, et al. Electrocoagulation process in water treatment: A review of electrocoagulation modeling approaches. Desalination 2017;404:1–21. doi:10.1016/J.DESAL.2016.10.011. [92] Bratby J. Coagulants, Coagulation and Flocculation in Water and Wastewater Treatment. 2nd ed. London: IWA Publishing; 2006. [93] Lee CS, Robinson J, Chong MF. A review on application of flocculants in wastewater treatment. Process Saf Environ Prot 2014;92:489–508. doi:10.1016/J.PSEP.2014.04.010. [94] Mollah MYA, Morkovsky P, Gomes JAG, Kesmez M, Parga J, Cocke DL. Fundamentals, present and future perspectives of electrocoagulation. J Hazard Mater 2004;114:199–210. doi:10.1016/j.jhazmat.2004.08.009. [95] Al Aji B, Yavuz Y, Koparal AS. Electrocoagulation of heavy metals containing model wastewater using monopolar iron electrodes. Sep Purif Technol 2012;86:248–54. doi:10.1016/J.SEPPUR.2011.11.011. [96] Can OT, Bayramoglu M, Kobya M. Decolorization of Reactive Dye Solutions by Electrocoagulation Using Aluminum Electrodes. Ind Eng Chem Res 2003;42:3391– 6. doi:10.1021/ie020951g. [97] Brillas E, Martínez-Huitle CA. Decontamination of wastewaters containing synthetic organic dyes by electrochemical methods. An updated review. Appl Catal B Environ 2015;166–167:603–43. doi:10.1016/J.APCATB.2014.11.016. [98] Lin SH, Peng CF. Treatment of textile wastewater by electrochemical method. Water Res 1994;28:277–82. doi:10.1016/0043-1354(94)90264-X.

159

[99] Meunier N, Drogui P, Montané C, Hausler R, Mercier G, Blais J-F. Comparison between electrocoagulation and chemical precipitation for metals removal from acidic soil leachate. J Hazard Mater 2006;137:581–90. doi:10.1016/J.JHAZMAT.2006.02.050. [100] Andreozzi R, Caprio V, Insola A, Marotta R. Advanced oxidation processes (AOP) for water purification and recovery. Catal Today 1999;53:51–9. doi:10.1016/S0920- 5861(99)00102-9. [101] Glaze WH, Kang J-W, Chapin DH. The Chemistry of Water Treatment Processes Involving Ozone, Hydrogen Peroxide and Ultraviolet Radiation. Ozone Sci Eng 1987;9:335–52. doi:10.1080/01919518708552148. [102] Poyatos JM, Muñio MM, Almecija MC, Torres JC, Hontoria E, Osorio F. Advanced oxidation processes for wastewater treatment: State of the art. Water Air Soil Pollut 2010;205:187–204. doi:10.1007/s11270-009-0065-1. [103] Amadelli R, De Battisti A, Girenko D., Kovalyov S., Velichenko A. Electrochemical oxidation of trans-3,4-dihydroxycinnamic acid at PbO2 electrodes: direct electrolysis and ozone mediated reactions compared. Electrochim Acta 2000;46:341–7. doi:10.1016/S0013-4686(00)00590-9. [104] Boye B, Dieng MM, Brillas E. Anodic oxidation, electro-Fenton and photoelectro- Fenton treatments of 2,4,5-trichlorophenoxyacetic acid. J Electroanal Chem 2003;557:135–46. doi:10.1016/S0022-0728(03)00366-8. [105] Peralta-Hernández JM, Meas-Vong Y, Rodríguez FJ, Chapman TW, Maldonado MI, Godínez LA. Comparison of hydrogen peroxide-based processes for treating dye-containing wastewater: Decolorization and destruction of Orange II azo dye in dilute solution. Dye Pigment 2008;76:656–62. doi:10.1016/j.dyepig.2007.01.001. [106] Pignatello JJ. Dark and photoassisted iron(3+)-catalyzed degradation of chlorophenoxy herbicides by hydrogen peroxide. Environ Sci Technol 1992;26:944–51. doi:10.1021/es00029a012. [107] Rajeshwar K. Photoelectrochemistry and the environment. J Appl Electrochem 1995;25. doi:10.1007/BF00242533. [108] Demir ME, Dincer I. Development and heat transfer analysis of a new heat recovery system with thermoelectric generator. Int J Heat Mass Transf 2017;108:2002–10. doi:10.1016/j.ijheatmasstransfer.2016.12.102. [109] Islam S, Dincer I, Yilbas BS. Energetic and exergetic performance analyses of a solar energy-based integrated system for multigeneration including thermoelectric generators. Energy 2015;93:1246–58. doi:10.1016/j.energy.2015.09.111. [110] Kong LB, Li T, Hng HH, Boey F, Zhang T, Li S. Waste Energy Harvesting. vol. 24. 2014. doi:10.1007/978-3-642-54634-1. [111] Yang P, Zhang R, McGarvey J, Benemann J. Biohydrogen production from cheese processing wastewater by anaerobic fermentation using mixed microbial communities. Int J Hydrogen Energy 2007;32:4761–71. doi:10.1016/j.ijhydene.2007.07.038. [112] Lin CY, Leu HJ, Lee KH. Hydrogen production from beverage wastewater via dark fermentation and room-temperature methane reforming. Int J Hydrogen Energy 2016;41:21736–46. doi:10.1016/j.ijhydene.2016.07.028. [113] Zhi X, Yang H, Berthold S, Doetsch C, Shen J. Potential improvement to a citric wastewater treatment plant using bio-hydrogen and a hybrid energy system. J Power

160

Sources 2010;195:6945–53. doi:10.1016/j.jpowsour.2010.04.046. [114] Mishra P, Thakur S, Singh L, Ab Wahid Z, Sakinah M. Enhanced hydrogen production from palm oil mill effluent using two stage sequential dark and photo fermentation. Int J Hydrogen Energy 2016;41:18431–40. doi:10.1016/j.ijhydene.2016.07.138. [115] Ajayi FF, Kim KY, Chae KJ, Choi M-J, Kim SY, Chang IS, et al. Study of hydrogen production in light assisted microbial electrolysis cell operated with dye sensitized solar cell. Int J Hydrogen Energy 2009;34:9297–304. doi:10.1016/j.ijhydene.2009.09.096. [116] Baeza JA, Martínez-Miró À, Guerrero J, Ruiz Y, Guisasola A. Bioelectrochemical hydrogen production from urban wastewater on a pilot scale. J Power Sources 2017;356:500–9. doi:10.1016/j.jpowsour.2017.02.087. [117] Arzate Salgado SY, Ramírez Zamora RM, Zanella R, Peral J, Malato S, Maldonado MI. Photocatalytic hydrogen production in a solar pilot plant using a Au/TiO 2 photo catalyst. Int J Hydrogen Energy 2016;41:11933–40. doi:10.1016/j.ijhydene.2016.05.039. [118] He YR, Yan FF, Yu HQ, Yuan SJ, Tong ZH, Sheng GP. Hydrogen production in a light-driven photoelectrochemical cell. Appl Energy 2014;113:164–8. doi:10.1016/j.apenergy.2013.07.020. [119] Kim Y, Logan BE. Simultaneous removal of organic matter and salt ions from saline wastewater in bioelectrochemical systems. Desalination 2013;308:115–21. doi:10.1016/j.desal.2012.07.031. [120] Park H, Bak A, Ahn YY, Choi J, Hoffmannn MR. Photoelectrochemical performance of multi-layered BiO x-TiO 2/Ti electrodes for degradation of phenol and production of molecular hydrogen in water. J Hazard Mater 2012;211–212:47–54. doi:10.1016/j.jhazmat.2011.05.009. [121] Tokumura M, Morito R, Kawase Y. Photo-Fenton process for simultaneous colored wastewater treatment and electricity and hydrogen production. Chem Eng J 2013;221:81–9. doi:10.1016/j.cej.2013.01.075. [122] Das D, Veziroglu TN. Advances in biological hydrogen production processes. Int J Hydrogen Energy 2008;33:6046–57. doi:10.1016/j.ijhydene.2008.07.098. [123] COMSOL Inc. COMSOL Multiphysics® Modeling Software Version 5.2 n.d. [124] EES: Engineering Equation Solver 2016. [125] Titanium(IV) isopropoxide CAS No. 546-68-9 | Sigma-Aldrich n.d. https://www.sigmaaldrich.com/catalog/search?term=546-68-9&interface=CAS No.&N=0+&mode=partialmax&lang=en®ion=CA&focus=product (accessed December 6, 2017). [126] Acetic acid CAS No. 64-19-7 | Sigma-Aldrich n.d. https://www.sigmaaldrich.com/catalog/search?term=64-19-7&interface=CAS No.&N=0+&mode=partialmax&lang=en®ion=CA&focus=product (accessed December 6, 2017). [127] Nitric acid CAS No. 7697-37-2 | Sigma-Aldrich n.d. https://www.sigmaaldrich.com/catalog/search?term=7697-37-2&interface=CAS No.&N=0+&mode=partialmax&lang=en®ion=CA&focus=product (accessed December 6, 2017). [128] Class AAA Standard Solar Simulator n.d. http://www.oainet.com/oai-

161

solarsimulator-pp.html (accessed December 6, 2017). [129] Chehade G, Demir ME, Dincer I, Yuzer B, Selcuk H. Experimental investigation and analysis of a new photoelectrochemical reactor for hydrogen production. Int J Hydrogen Energy 2018;43:12049–58. doi:10.1016/j.ijhydene.2018.04.110. [130] Demir ME, Chehade G, Dincer I, Yuzer B, Selcuk H. Design and analysis of a new photoelectrochemical cell for hydrogen production and wastewater treatment (Under Review). Electrochim Acta 2018. [131] Reactive Black 5 Dye content ≥50 % | Sigma-Aldrich n.d. https://www.sigmaaldrich.com/catalog/product/sial/306452?lang=en®ion=CA (accessed December 25, 2017). [132] Gamry Reference 3000 Potentiostat/Galvanostat/ZRA n.d. https://www.gamry.com/potentiostats/reference-3000/ (accessed May 18, 2017). [133] Reference 30k Booster for Reference 3000 Potentiostat n.d. https://www.gamry.com/potentiostats/reference-30k-booster/ (accessed May 14, 2018). [134] OMEGA Engineering Canada | Water Testing Kits n.d. https://www.omega.ca/pptst_eng/WT_WATERTEST_KITS.html#order (accessed December 29, 2017). [135] Standard Methods Committee. 5220 Chemical Oxygen Demand ( COD ) 5220 B . Open Reflux Method. 1997. [136] HANNA Instruments HI-83099 COD and Multiparameter Bench Photometer n.d. https://hannainst.com/ (accessed March 1, 2018). [137] 5220 Chemical Oxygen Demand ( COD ) 5220 D . Closed Reflux Method. n.d. [138] Demir ME, Chehade G, Dincer I, Yuzer B, Selcuk H. Synergistic effects of advanced oxidization reactions in a combination of TiO2 photocatalysis for hydrogen production and wastewater treatment applications (Under Review). J Electrochem Soc 2018. [139] Bak T, Nowotny J, Rekas M, Sorrell C. Photo-electrochemical hydrogen generation from water using solar energy. Materials-related aspects. Int J Hydrogen Energy 2002;27:991–1022. doi:10.1016/S0360-3199(02)00022-8. [140] Meriç S, Kaptan D, Ölmez T. Color and COD removal from wastewater containing Reactive Black 5 using Fenton’s oxidation process. Chemosphere 2004;54:435–41. doi:10.1016/j.chemosphere.2003.08.010. [141] Pitwell RL. Standard COD. Chem Brit 1983 19: 907. n.d. [142] Gähr F, Hermanutz F, Oppermann W. Ozonation- An important technique to comply with new German laws for textile waste-water treatment. Water Sci Technol 1994;30:255–63. [143] HI93754C-25 COD High Range Reagent Vials, Dichromate Method - Hanna Instruments n.d. https://hannainst.com/hi93754c-25-cod-high-range-reagent-vials- dichromate-method-25-tests.html (accessed March 2, 2018). [144] Iron Test Kit - 3347-01 - -LaMotte n.d. http://www.lamotte.com/en/browse/3347- 01.html (accessed March 2, 2018). [145] Omega-Mass and Volumetric Flow Meters n.d. https://www.omega.ca/pptst_eng/FMA1600.html (accessed June 12, 2018). [146] Fuel Cells Etc n.d. http://www.fuelcellsetc.com/store/Nafion/N115 (accessed June 22, 2017).

162

[147] Tracker Video Analysis and Modeling Tool for Physics Education n.d. http://physlets.org/tracker/ (accessed June 23, 2017). [148] Instruments G. Reference 3000TM Potentiostat/Galvanostat/ZRA Operator’s Manual 2012. [149] Silicon-cell Pyranometers and Meters | Apogee Instruments n.d. https://www.apogeeinstruments.co.uk/silicon-cell-pyranometers/ (accessed June 22, 2017). [150] Pyranometer Vernier 2018. https://www.vernier.com/products/sensors/solar- radiation-sensors/pyr-bta/ (accessed June 12, 2018). [151] Type K Thermocouple Wire - Duplex Insulated:PFA Insulated Thermocouple Wire - TT-K-24-25 n.d. http://www.omega.ca/ (accessed June 22, 2017). [152] Thermocouple-Types of thermocouples n.d. http://www.thermocoupleinfo.com/ (accessed May 22, 2017). [153] Omega DAQPRO-5300 Portable Handheld Data Logger 2017. [154] Standard Heatblock- VWR Scientific 2018. https://us.vwr.com/store/product/582406/block-heaters (accessed March 20, 2018). [155] Corning® pH/ISE/mV/T meters Model 450, AC/DC input 120 V AC | Sigma- Aldrich n.d. https://www.sigmaaldrich.com/catalog/product/aldrich/z283096?lang=en®ion= CA (accessed June 12, 2018). [156] SpetronicTM 200 Spectrophotometer 2018. https://www.thermofisher.com/order/catalog/product/840-281700 (accessed June 12, 2018). [157] Taylor BN. Guidelines for evaluating and expressing the uncertainty of NIST measurement results. Gaithersburg, MD: 1994. doi:10.6028/NIST.TN.1297. [158] Dincer I, Rosen MA. Chapter 2 - Exergy and Energy Analyses. Exergy, Amsterdam: Elsevier; 2013, p. 21–30. doi:10.1016/B978-0-08-097089-9.00002-4. [159] Bicer Y, Dincer I. Energy and exergy analyses of an integrated underground coal gasification with SOFC fuel cell system for multigeneration including hydrogen production. Int J Hydrogen Energy 2015;40:13323–37. doi:10.1016/j.ijhydene.2015.08.023. [160] Demir ME, Dincer I. Development Of A Hybrid Solar Thermal System With TEG And PEM For Hydrogen And Power Production. Int J Hydrogen Energy 2017. [161] Meriche IE, Beghidja A, Boumedjirek M. Energetic and exergetic analysis of solar gas turbine power plant in South Algeria. IREC 2014 - 5th Int Renew Energy Congr 2014. doi:10.1109/IREC.2014.6826902. [162] European Regional Development Fund and European Neighbourhood and Partnership Instrument. Radiant heating, Convection Heating Systems and Wall Tempering. WP5 Educ Econ Promot (CO2OL Bricks) 2014;1. [163] Cengel YA, Boles MA. Thermodynamics: An Engineering Approach. 8th ed. Mcgraw-Hiill Education; 2015. [164] Zhang X, Zhao L-D. Thermoelectric materials: Energy conversion between heat and electricity. J Mater 2015;1:92–105. doi:10.1016/j.jmat.2015.01.001. [165] Demir ME, Dincer I. Investigation of a solar driven integrated TEG with PEM electrolyzer system for hydrogen production. 3rd Int. Hydrog. Technol. Congr., Antalya: 2018.

163

[166] Sohrabi M, Ghavami M. Photocatalytic degradation of Direct Red 23 dye using UV/TiO2: Effect of operational parameters. J Hazard Mater 2008;153:1235–9. doi:10.1016/j.jhazmat.2007.09.114. [167] Ghiselli G, Jardim WF, Litter MI, Mansilla HD. Destruction of EDTA using Fenton and photo-Fenton-like reactions under UV-A irradiation. J Photochem Photobiol A Chem 2004;167:59–67. doi:10.1016/j.jphotochem.2004.02.005. [168] Zuo Y, Hoigne J. Formation of hydrogen peroxide and depletion of oxalic acid in atmospheric water by photolysis of iron(III)-oxalato complexes. Environ Sci Technol 1992;26:1014–22. doi:10.1021/es00029a022. [169] Lunar L, Sicilia D, Rubio S, Pérez-Bendito D, Nickel U. Degradation of photographic developers by Fenton’s reagent: condition optimization and kinetics for metol oxidation. Water Res 2000;34:1791–802. doi:10.1016/S0043- 1354(99)00339-5. [170] Pignatello JJ. Dark and photoassisted iron(3+)-catalyzed degradation of chlorophenoxy herbicides by hydrogen peroxide. Environ Sci Technol 1992;26:944–51. doi:10.1021/es00029a012. [171] Machulek A, Quina FH, Gozzi F. Fundamental Mechanistic Studies of the Photo- Fenton Reaction for the Degradation of Organic Pollutants. Org Pollut Intech Publ Inc, Rijeka 2012:271–92. doi:10.5772/30995. [172] Bensalah N, Chair K, Bedoui A. Efficient degradation of tannic acid in water by UV/H2O2 process. Sustain Environ Res 2018;28:1–11. doi:10.1016/J.SERJ.2017.04.004. [173] Zamfirescu C, Dincer I, Naterer GF, Banica R. Quantum efficiency modeling and system scaling-up analysis of water splitting with Cd1−xZnxS solid-solution photocatalyst. Chem Eng Sci 2013;97:235–55. doi:10.1016/j.ces.2013.04.020. [174] Almahdi M, Dincer I, Rosen M a. Analysis and assessment of methanol production by integration of carbon capture and photocatalytic hydrogen production. Int J Greenh Gas Control 2016;51:56–70. doi:10.1016/j.ijggc.2016.04.015. [175] Lazzaretto A, Tsatsaronis G. SPECO: A systematic and general methodology for calculating efficiencies and costs in thermal systems. Energy 2006;31:1257–89. doi:10.1016/j.energy.2005.03.011. [176] DOE H2A Production Analysis. H2A Cent. Hydrog. Prod. Model. n.d. [177] Toffolo A, Lazzaretto A. Evolutionary algorithms for multi-objective energetic and economic optimization in thermal system design. Energy 2002;27:549–67. doi:10.1016/S0360-5442(02)00009-9. [178] Holland JH. Adaptation in natural and artificial systems. Ann Arbor: University of Michigan Press.; 1975. [179] Goldberg DE (David E, E. D. Genetic algorithms in search, optimization, and machine learning. Addison-Wesley Longman Publishing Co., Inc.; 1989. [180] Grefenstette JJ, Naval Research Laboratory (U.S.) J, Texas Instruments Incorporated. Proceedings of the First International Conference on Genetic Algorithms and Their Applications : July 24-26, 1985 at the Carnegie-Mellon University, Pittsburg, PA. Lawrence Erlbaum Associates; 1988. [181] Ezzat M. Conceptual Development and Analysis of Sustainable Powering Options for Hybrid Vehicles. University of Ontario Institute of Technology, 2018. [182] Stender J. Parallel genetic algorithms : theory and applications. IOS Press; 1993.

164

[183] Momeni MM, Ghayeb Y. Photoelectrochemical water splitting on chromium-doped titanium dioxide nanotube photoanodes prepared by single-step anodizing. J Alloys Compd 2015;637:393–400. doi:10.1016/j.jallcom.2015.02.137. [184] Mascaretti L, Ferrulli S, Mazzolini P, Casari CS, Russo V, Matarrese R, et al. Hydrogen-treated hierarchical titanium oxide nanostructures for photoelectrochemical water splitting. Sol Energy Mater Sol Cells 2017;169:19–27. doi:10.1016/j.solmat.2017.04.045. [185] Zhao Y, Hoivik N, Wang K. Recent advance on engineering titanium dioxide nanotubes for photochemical and photoelectrochemical water splitting. Nano Energy 2016;30:728–44. doi:10.1016/j.nanoen.2016.09.027. [186] Fan W, Yu X, Lu H-C, Bai H, Zhang C, Shi W. Fabrication of TiO2/RGO/Cu2O heterostructure for photoelectrochemical hydrogen production. Appl Catal B Environ 2016;181:7–15. doi:10.1016/j.apcatb.2015.07.032. [187] Tripathi M, Chawla P. Optimization of semiconductor ns-TiO2-CuO admixed photoelectrode for photoelectrochemical solar cell in regard to hydrogen production. Int J Hydrogen Energy 2016;41:7993–6. doi:10.1016/j.ijhydene.2015.11.163. [188] Jumeri F, Lim H, Zaina Z, Huang N, Pandikumar A. Titanium dioxide-reduced graphene oxide thin film for photoelectrochemical water splitting. Ceram Int 2014;40:15159–65. doi:10.1016/j.ceramint.2014.06.130. [189] Arakelyan VM, Aroutiounian VM, Shahnazaryan GE, Khachaturyan EA. Thin film n-titanium oxide photoanodes for photoelectrochemical production of hydrogen. Renew Energy 2008;33:299–303. doi:10.1016/j.renene.2007.05.012. [190] Skoumal M, Rodríguez RM, Cabot PL, Centellas F, Garrido JA, Arias C, et al. Electro-Fenton, UVA photoelectro-Fenton and solar photoelectro-Fenton degradation of the drug ibuprofen in acid aqueous medium using platinum and boron-doped diamond anodes. Electrochim Acta 2009;54:2077–85. doi:10.1016/j.electacta.2008.07.014. [191] Panizza M, Oturan MA. Degradation of Alizarin Red by electro-Fenton process using a graphite-felt cathode. Electrochim Acta 2011;56:7084–7. doi:10.1016/j.electacta.2011.05.105. [192] Karcher S, Kornmüller A, Jekel M. Anion exchange resins for removal of reactive dyes from textile wastewaters. Water Res 2002;36:4717–24. doi:10.1016/S0043- 1354(02)00195-1. [193] Demir ME, Dincer I. Development of an integrated hybrid solar thermal power system with thermoelectric generator for desalination and power production. Desalination 2017;404:59–71. doi:10.1016/j.desal.2016.10.016. [194] Demir ME, Dincer I. Performance assessment of a thermoelectric generator applied to exhaust waste heat recovery. Appl Therm Eng 2017;120:694–707. doi:10.1016/j.applthermaleng.2017.03.052. [195] TEG Cascade 800°C Hot Side Thermoelectric Power Modules - Thermoelectric Generator n.d. http://thermoelectric-generator.com/teg-cascade-800c-hot-side- thermoelectric-power-modules/ (accessed May 17, 2017). [196] Sahin AZ, Yilbas BS, Shuja SZ, Momin O. Investigation into topping cycle: Thermal efficiency with and without presence of thermoelectric generator. Energy 2011;36:4048–54. doi:10.1016/j.energy.2011.04.044. [197] Aygun A, Yilmaz T, Nas B, Berktay A. Effect of temperature on fenton oxidation

165

of young landfill leachate: Kinetic assessment and sludge properties. Glob Nest J 2012;14:487–95. [198] Sun J-H, Sun S-P, Fan M-H, Guo H-Q, Qiao L-P, Sun R-X. A kinetic study on the degradation of p-nitroaniline by Fenton oxidation process. J Hazard Mater 2007;148:172–7. doi:10.1016/j.jhazmat.2007.02.022. [199] Plappally AK, Lienhard JH. Costs for water supply, treatment, end-use and reclamation. Desalin Water Treat 2013;51:200–32. doi:10.1080/19443994.2012.708996. [200] Tzen E. Renewable energy sources for sea water desalination– present status and future prospects. Desalin Methods, Cost Technol Agric Issues Policy 2010:325– 339.

166