Sustainable Cultivation of Microalgae Using Diluted Anaerobic Digestate for Biofuels

Production

A dissertation presented to

the faculty of

the Russ College of Engineering and Technology of Ohio University

In partial fulfillment

of the requirements for the degree

Doctor of Philosophy

Husam A. Abu Hajar

August 2016

This dissertation titled

Sustainable Cultivation of Microalgae Using Diluted Anaerobic Digestate for Biofuels

Production

HUSAM A. ABU HAJAR

has been approved for

the Department of Civil Engineering

and the Russ College of Engineering and Technology by

R. Guy Riefler

Associate Professor of Civil Engineering

Dennis Irwin

Dean, Russ College of Engineering and Technology

ABSTRACT

ABU HAJAR, HUSAM A., Ph.D., August 2016, Civil Engineering

Sustainable Cultivation of Microalgae Using Diluted Anaerobic Digestate for Biofuels

Production

Director of Dissertation: R. Guy Riefler

Microalgae cultivation has gained considerable interest recently as a potential source for the production of biofuels. Nevertheless, many obstacles still face the industrial application of microalgal biofuels such as the high production costs due to nutrient requirements and the high energy input for cultivating and harvesting the microalgal biomass. In this study, the utilization of the anaerobic digestate as a nutrient medium for the cultivation of two microalgal species was investigated. The anaerobic digestate was initially characterized and several pretreatment methods such as hydrogen peroxide treatment, filtration using polyester filter bags, and supernatant extraction were applied to the digestate. It was found that the supernatant extraction was the simplest and most effective method in decreasing the turbidity and COD of the diluted anaerobic digestate while maintaining sufficient nutrients (particularly nitrogen) for microalgae cultivation. The microalga Neochloris oleoabundans was cultivated using the diluted anaerobic digestate supernatant on a bench-scale. It was found that 100 mg N/L dilution was sufficient to maximize the biomass concentration of this microalga. It was attempted to scale up the cultivation of the microalga N. oleoabundans to 100 L raceway ponds; however, the culture was contaminated with other algal species.

The microalga Scenedesmus dimorphus was then cultivated using the anaerobic digestate supernatant on a bench-scale. The highest biomass concentration recorded was

654 mg/L. Furthermore, it was found that 50 – 100 mg N/L dilutions were sufficient to maximize the specific growth rate of this microalga while still producing relatively high biomass concentrations. As a result, the microalgae cultivation was scaled up to 100 L raceway ponds using 100 mg N/L dilution at 454 and 317 µmol/m2/s light intensities and

50 mg N/L dilution at 384 and 234 µmol/m2/s light intensities. The highest biomass concentration achieved was 432 mg/L in the 100 mg N/L – 454 µmol/m2/s culture.

Nitrogen removal efficiencies were in the 65 – 72% range with complete ammonia removal. Phosphorus removal efficiencies were in the 63 – 100% range while COD removal efficiencies as a result of the bacteria presence in the unsterilized nutrient media were in the 78 – 82% range.

The effect of mixing on the growth of the microalga S. dimorphus was evaluated by cultivating this microalga in a raceway pond at 0.1, 0.2, and 0.3 m/s water surface velocities. It was concluded that the biomass concentration and growth rate increased with an increase in the mixing velocity. However, by balancing the power required to operate the pond at different velocities with the potential energy yield from biodiesel synthesis, it was found that operating the pond at 0.1 m/s surface velocity was the only case with a positive net energy.

Finally, two biomass growth models were developed and tested on the growth of the microalga S. dimorphus. It was found that the logistic model, which assumes a maximum bearing capacity of the culture, represented the biomass growth better than the

5 exponential growth model, which assumes that the microalgae grow exponentially at specific growth rates equal to or less than the maximum specific growth rate depending on the culture conditions.

PREFACE

This dissertation is divided into two parts. Part I (chapters 1 through 7) titled

“Sustainable Cultivation of Microalgae Using Diluted Anaerobic Digestate for Biofuels

Production” addresses my research work over the past 3 years of my Ph.D. program, while Part II (Appendix A) titled “Selective Precipitation of Aluminum and Iron in Acid

Mine Drainage” covers my earlier Ph.D. research which focused on the treatment of acid mine drainage in eastern and southeastern Ohio.

DEDICATION

Dedicated to my beloved family in Jordan and to all my friends in Athens, Ohio. My

success is because of your love, support, and encouragement

ACKNOWLEDGMENTS

I would like to express my sincere gratitude to my adviser Dr. R. Guy Riefler for his full support, guidance, and encouragement throughout my Ph.D. program. I am also very thankful to Dr. Ben J. Stuart for his endless support and for giving me the chance to work on the “Sustainable Housing through Holistic Waste Stream Management and

Algal Cultivation” project. I would like also to thank my committee members, Dr. David

Bayless, Dr. Morgan Vis, and Dr. Kurt Rhoads for agreeing to serve on my Ph.D. committee and for their valuable feedback and guidance.

I would like also to thank all Civil Engineering faculty, staff, graduate and undergraduate fellows who helped and supported me throughout my graduate career at

Ohio University. Also, I would like to extend my gratitude to the faculty, staff, and students of the Institute for Sustainable Energy and the Environment. To all my friends in

Athens, thank you for being there for me, I have been lucky enough to have met wonderful people like you, and you will always be my family.

Finally, I would like to acknowledge the National Science Foundation (NSF) for funding the “Sustainable Housing through Holistic Waste Stream Management and Algal

Cultivation” project through the Sustainable Energy Pathways (SEP) program (Award #

1230961) and the Wayne National Forest, U.S. Forest Service and the Ohio Department of Natural Resources, Division of Mineral Resources Management for funding the

“Selective Precipitation of Aluminum and Iron in Acid Mine Drainage” project.

TABLE OF CONTENTS

Page Abstract ...... 3 Preface...... 6 Dedication ...... 7 Acknowledgments...... 8 List of Tables ...... 13 List of Figures ...... 15 Chapter 1. Introduction ...... 18

1.1 Energy Sources and the Environment ...... 18 1.2 Microalgal Biofuels ...... 19 1.3 Research Objectives ...... 20 1.4 Dissertation Outline ...... 20

Chapter 2. A Review of the Cultivation of the Microalga Neochloris oleoabundans for Biofuels Production and other Industrial Applications ...... 24

2.1 Abstract ...... 24 2.2 Introduction ...... 25 2.3 Microalgae Cultivation ...... 27

2.3.1 Microalgae definition and growth conditions ...... 27 2.3.2 Applications and products from microalgae ...... 28 2.3.3 Lipids accumulation in microalgae ...... 30 2.3.4 Challenges facing microalgae cultivation and potential solutions .... 31

2.4 Neochloris oleoabundans ...... 33

2.4.1 Introduction to the microalga Neochloris oleoabundans ...... 33 2.4.2 N. oleoabundans biomass composition...... 35 2.4.3 Stress conditions in N. oleoabundans and lipids accumulation ...... 39 2.4.4 Harvesting ...... 40 2.4.5 Applications ...... 42

2.5 Factors Affecting the Growth of N. oleoabundans ...... 43

2.5.1 Nutrients ...... 43 2.5.2 Light ...... 46 2.5.3 Temperature ...... 48 2.5.4 pH ...... 49 2.5.5 Dissolved oxygen concentration ...... 50 2.5.6 Salinity ...... 50 2.5.7 Circadian clock ...... 51 2.5.8 Other factors...... 51

2.6 Biomass, Lipids, and Fatty Acids ...... 52

2.6.1 Biomass ...... 52 2.6.2 Lipids ...... 53 2.6.3 Fatty acids profile ...... 54

2.7 Other Cultivation Conditions ...... 58

2.7.1 Heterotrophic cultivation ...... 58 2.7.2 Mixotrophic cultivation ...... 58

2.8 Alternatives for the Cultivation of N. oleoabundans ...... 59

2.8.1 Wastewater as a nutrient medium ...... 59 2.8.2 Cultivation of N. oleoabundans using flue gas ...... 61

2.9 Conclusions and Recommendations ...... 61

Chapter 3. Anaerobic Digestate as a Nutrient Medium for the Growth of the Green Microalga Neochloris oleoabundans ...... 63

3.1 Abstract ...... 63 3.2 Introduction ...... 64 3.3 Materials and Methods ...... 69

3.3.1 Anaerobic digestate ...... 69 3.3.2 Analytical methods ...... 69

3.3.3 Pretreatment methods...... 70 3.3.4 Microalgae selection ...... 71 3.3.5 Microalgae cultivation using diluted AD ...... 72 3.3.6 Statistical analysis ...... 72

3.4 Results ...... 72

3.4.1 AD characterization ...... 72 3.4.2 Zeta potential ...... 75 3.4.3 Filtration using polyester filter bags ...... 77 3.4.4 Hydrogen peroxide treatment ...... 80 3.4.5 Supernatant characterization ...... 82 3.4.6 Microalgae cultivation ...... 87

3.5 Conclusions ...... 95

Chapter 4. Cultivation of Scenedesmus dimorphus Using Anaerobic Digestate as a Nutrient Medium ...... 96

4.1 Abstract ...... 96 4.2 Introduction ...... 97 4.3 Materials and Methods ...... 102

4.3.1 Anaerobic digestate ...... 102 4.3.2 Analytical methods ...... 102 4.3.3 Microalgae selection ...... 103 4.3.4 Bench-scale microalgae cultivation ...... 103 4.3.5 Microalgae cultivation in the raceway ponds ...... 104

4.4 Results and Discussion ...... 106

4.4.1 Biomass and optical density calibration ...... 106 4.4.2 Bench-scale cultivation of S. dimorphus ...... 107 4.4.3 Cultivation of S. dimorphus in raceway ponds ...... 112

4.5 Conclusions ...... 121

Chapter 5. Hydrodynamic Characterization of Algae Raceway Ponds Using Tracer Tests ...... 123

5.1 Introduction ...... 123 5.2 Materials and Methods ...... 125

5.2.1 Theory ...... 125 5.2.2 Experimental work ...... 130

5.3 Results ...... 132 5.4 Conclusions ...... 145

Chapter 6. Modeling the Growth of Microalgae in Raceway Ponds Using Diluted Anaerobic Digestate as a Nutrient Source ...... 147

6.1 Introduction ...... 147 6.2 Materials and Methods ...... 148

6.2.1 Microalgae selection ...... 148 6.2.2 Biomass growth model ...... 148 6.2.3 Nutrient medium suspended solids concentration ...... 155

6.3 Results and Discussion ...... 156

6.3.1 Light intensity effect ...... 156 6.3.2 Light attenuation ...... 162 6.3.3 Nutrient medium suspended solids concentration ...... 167 6.3.4 Models validation...... 169

6.4 Conclusions ...... 177

Chapter 7. Conclusions and Recommendations ...... 178 References ...... 185 Appendix A. Selective Precipitation of Aluminum and Iron in Acid Mine Drainage .... 206

LIST OF TABLES

Page

Table 2.1: Biomass and Lipids Concentrations and Productivities for the Microalga N. oleoabundans ...... 56 Table 3.1: AD Characterization (average ± SD; n = 3) ...... 75 Table 3.2: Characterization of the Unfiltered and 10, 5, and 1 μm Filtered Diluted AD (average ± SD; n = 3) ...... 79 Table 3.3: Target Nitrogen and Phosphorus Concentrations in the Diluted AD Supernatant and Filtrate (average ± SD; n = 3) ...... 88 Table 4.1: Specific Growth Rates Corresponding to Different Nutrients Concentrations in the Diluted AD Supernatant ...... 112 Table 4.2: Initial Characterization of the Nutrient Media Used for Microalgae Cultivation in the Raceway Ponds (Average ± SD; n=3) ...... 114 Table 5.1: Mean Velocities Estimated by Method 1 ...... 138 Table 5.2: Mean Velocities Estimated by Method 2 ...... 140 Table 5.3: Dispersion Coefficients Values (m2/s) ...... 144 Table 5.4: Total Head Losses and Power Requirements Estimated as a Function of Average Velocity ...... 144 Table 6.1: Specific Growth Rates Corresponding to Each Light Intensity (Exponential Model) ...... 157 Table 6.2: Specific Growth Rates Corresponding to Each Light Intensity According to the Logistic Model ...... 159 Table 6.3: Light Attenuation Coefficients (K) for Different Biomass Concentrations .. 162 Table 6.4: Light Attenuation Coefficients (K) for Different AD TSS Concentrations .. 165 Table 6.5: Light Attenuation Coefficients (K) for Various Microalgae and AD Concentrations (Actual vs. Predicted Values) ...... 167 Table A. 1: PHREEQC Simulation Results ...... 212 Table A. 2: PHREEQC Simulation Results (Gibbsite is Present) ...... 213 Table A. 3: PHREEQC Simulation Results (Amorphous is Present) ...... 214

Table A. 4: PHREEQC Simulation Results (Iron is in the Ferric Form) ...... 214 Table A. 5: Comparison between Laboratory Results and PHREEQC Simulation Results ...... 216 Table A. 6: Iron and Aluminum Results for the Aluminum Settling Tank Influent and Effluent ...... 218

LIST OF FIGURES Page

Figure 2.1: Application and products from microalgae biomass ...... 29 Figure 2.2: N. oleoabundans cells (adapted from the Culture Collection of Algae at The University of Texas at Austin website) ...... 34 Figure 3.1: Zeta potential (♦) and conductance (●) values for the diluted AD (average ± SD; n = 3) ...... 77 Figure 3.2: N/P comparison between the unfiltered diluted AD and the 10 μm filtrate for sources A and B (average ± SD; n = 3) ...... 79 Figure 3.3: TSS and COD concentrations for various hydrogen peroxide dosages (average ± SD; n = 3)...... 81 Figure 3.4: Nitrogen and ammonia concentrations for various hydrogen peroxide dosages (average ± SD; n = 3) ...... 82 Figure 3.5: TSS and OD 750 values in the supernatant as a function of time (average ± SD; n = 4) ...... 83 Figure 3.6: COD concentration in the supernatant as a function of settling time with respect to the initial COD concentration at time 0 (average ± SD; n = 4) ...... 85 Figure 3.7: Nutrients concentrations in the supernatant as a function of settling time with respect to the initial concentrations at time 0 (a) total N source A, (b) total N source B, (c) total P source A, and (d) total P source B (average ± SD; n = 4) ...... 86 Figure 3.8: N/P ratio in the supernatant as a function of settling time (average ± SD; n = 4) ...... 87 Figure 3.9: Supernatant AD as a nutrient medium (average ± SD; n = 5) ...... 92 Figure 3.10: Filtered AD as a nutrient medium (average ± SD; n = 5) ...... 92 Figure 3.11: Supernatant AD (4 mL) and DI water (1 mL) (average ± SD; n = 5) ...... 93 Figure 4.1: Microalgae cultivation (a) bench-scale experiment (b) raceway ponds ...... 106 Figure 4.2: Biomass concentration and optical density calibration (a) OD 680 and TSS calibration for the microalga S. dimorphus (b) OD 680/OD 750 vs. actual microalgae biomass fraction ...... 107

Figure 4.3: Microalgae growth using diluted anaerobic digestate supernatant (a) biomass concentration (mg/L) (b) OD 680 ...... 111 Figure 4.4: S. dimorphus growth in the raceway ponds under different dilutions and light intensities (a) biomass concentration (mg/L) (b) OD 680 ...... 115 Figure 4.5: Nutrients concentrations in the S. dimorphus cultures grown in the raceway pond (a) nitrogen (b) phosphorus (c) ammonia (d) COD ...... 116 Figure 4.6: N/P ratio in the S. dimorphus cultures grown in the raceway pond ...... 117 Figure 5.1: Raceway pond illustration ...... 131 Figure 5.2: Conductivity and sodium chloride concentration calibration ...... 133 Figure 5.3: 0.1 m/s surface velocity (point 1) ...... 133 Figure 5.4: 0.1 m/s surface velocity (point 2) ...... 134 Figure 5.5: 0.1 m/s surface velocity (point 3) ...... 134 Figure 5.6: 0.2 m/s surface velocity (point 1) ...... 135 Figure 5.7: 0.2 m/s surface velocity (point 2) ...... 135 Figure 5.8: 0.2 m/s surface velocity (point 3) ...... 136 Figure 5.9: 0.3 m/s surface velocity (point 1) ...... 136 Figure 5.10: 0.3 m/s surface velocity (point 2) ...... 137 Figure 5.11: 0.3 m/s surface velocity (point 3) ...... 137 Figure 5.12: S. dimorphus growth at different mixing velocities ...... 145 Figure 6.1: Biomass growth of S. dimorphus at various light intensities in µmol/m2/s . 156 Figure 6.2: Specific growth rate (exponential model) as a function of light intensity (measured vs. predicted) ...... 158

Figure 6.3: Measured and predicted μu values for the logistic model according to Tamiya’s model ...... 159 Figure 6.4: Measured and predicted biomass concentrations by the logistic model at different light intensities ...... 161 Figure 6.5: Correlation between microalgae biomass concentration and light attenuation coefficient ...... 163 Figure 6.6: Light attenuation coefficient measured values (markers) vs. predicted values (dotted lines) for the microalga S. dimorphus ...... 164

Figure 6.7: Correlation between AD concentration and light attenuation coefficient .... 165 Figure 6.8: Light attenuation coefficient measured values (markers) vs. predicted values (dotted lines) for the AD supernatant...... 166 Figure 6.9: Nutrient media TSS (measured vs. predicted)...... 168 Figure 6.10: Measured and predicted time course biomass concentrations when the microalga S. dimorphus was cultivated in the raceway pond using BG-11 medium ..... 170 Figure 6.11: Measured and predicted biomass concentrations when the microalga S. dimorphus was cultivated in the raceway pond using BG-11 medium ...... 170 Figure 6.12: Measured and predicted time course biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant ...... 172 Figure 6.13: Measured and predicted biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant ...... 173 Figure 6.14: Measured and predicted time course biomass concentrations using the logistic model when the microalga S. dimorphus was cultivated using AD supernatant 175 Figure 6.15: Measured and predicted biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant ...... 176 Figure A. 1: Concentration variation with pH (a) Aluminum, (b) Iron [1]: Initial filtered metals' concentration after adding FAS and AS to the acidic solution (pH= 2.9), [2]: Filtered metals' concentration at the desired pH (time= 0), [3]: Filtered metals' concentration at the desired pH (time= 15 min), and [4]: Unfiltered metals concentration at the desired pH (time= 15 min)...... 215 Figure A. 2: Effluent/influent concentration ratio for aluminum settling tank ...... 219

CHAPTER 1. INTRODUCTION

1.1 Energy Sources and the Environment

Until today, more than 80% of the global energy demand is supplied by nonrenewable fossil fuels such as oil, coal, and natural gas. Fossil fuels combustion is the primary contributor to the excessive production of greenhouse gases (GHGs); which has posed environmental concerns such as air pollution, ozone depletion, and acid precipitation (Chen et al., 2011; Dincer, 2000; Höök & Tang, 2013; Mata et al., 2010;

Williams & Laurens, 2010). Furthermore, the nonrenewable energy reserves will eventually be depleted due to the increasing energy demand associated with the continuous growth in the world population and the economic and industrial development.

These reasons have necessitated exploring renewable energy sources to gradually replace the nonrenewable fossil fuels. Since the early 1970s, considerable research and studies have been dedicated to the development of renewable energy sources such as wind, solar, geothermal and biofuels (Dincer, 2010; Greenwell et al., 2009; Lam & Lee, 2012; Mata et al., 2010).

Biofuels such as bioethanol and biodiesel have been studied widely; however, their contribution is still limited to less than 1% of the worldwide energy demand.

Challenges facing the large scale production of biofuels include the high production costs and the utilization of edible feedstocks and arable lands (Brennan & Owende, 2010; Chen et al., 2011; Lam & Lee, 2012; Mubarak et al., 2015; Williams & Laurens, 2010). In order to make biofuels production sustainable, these challenges have to be overcome.

1.2 Microalgal Biofuels

Microalgal biomass has emerged as a promising feedstock for the production of biofuels such as biodiesel, hydrogen, bioethanol, and methane as well as other valuable products. Microalgae are unicellular microorganisms that utilize light energy to produce biomass and storage compounds. The most common cultivation mode of microalgae is phototrophic; however, some microalgae species can grow heterotrophically or mixotrophically. Advantages of microalgae cultivation for biofuels production include the high photosynthetic efficiency, growth rate, and oil content of microalgae compared to other energy crops, besides the fact that microalgae cultivation will eliminate the competition for arable lands, edible feedstocks, and even freshwater. One of the drawbacks of microalgal biofuels production is the high production cost associated with biomass cultivation, harvesting, and lipid extraction (Amin, 2009; Giovanardi et al.,

2014; Mata et al., 2010; Milledge, 2011; Villa et al., 2014; Williams & Laurens, 2010).

Therefore, it is desirable to investigate alternative cultivation approaches that can reduce the capital and operating costs of microalgae cultivation in order to make the microalgal biofuels more economically feasible.

Microalgae cultivation using wastewater as a nutrient medium is an attractive approach to reduce the production cost of microalgal biofuels. Moreover, microalgae biomass can function as a sink for the removal of nutrients and heavy metals from wastewater (Baldisserotto et al., 2014; Brennan & Owende, 2010; Kim et al., 2014; Wang

& Lan, 2011a). For example, anaerobic digestion, which is a biological process used widely to treat organic wastes, produces a nutrient-rich liquid waste (digestate) that can

20 be utilized as a nutrient medium for the growth of microalgae. Even though the organic loading is reduced through the anaerobic digestion process, nutrients such as nitrogen and phosphorus are not eliminated (Franchino et al., 2013; Levine et al., 2011; Olguín et al.,

2015a).

1.3 Research Objectives

The general objective of this research was to investigate the cultivation of two microalgal species using the anaerobic digestate as a nutrient medium. The experiments were initially conducted on a bench-scale and the cultivation of one species was then scaled up to 100 L raceway ponds. In addition, a hydrodynamic characterization of the flow in the raceway ponds was conducted at different water surface velocities in order to determine the actual flow velocities, dispersion coefficients, energy losses, and power required to operate the ponds at different velocities. Finally, two biomass growth models were developed and tested on the growth of one microalgae species in the raceway ponds.

The individual tasks are explained further in the section below.

1.4 Dissertation Outline

 Chapter 1: Introduction

This chapter provides a brief description on nonrenewable energy sources and the environmental consequences of fossil fuels exploitation which necessitated exploring renewable and clean energy sources such as biofuels. Then the microalgae cultivation for biofuels production was addressed where the pros and cons of microalgal biofuels and the challenges facing the industrial production of microalgal biofuels were discussed concisely. Additionally, the utilization of wastewater such as the anaerobic digestate as a

21 nutrient medium for the cultivation of microalgae was introduced as an alternative approach to reduce the production cost of microalgal biofuels. Finally, research objectives and tasks as well as the outline of this dissertation were addressed in this chapter.

 Chapter 2: A Review of the Cultivation of the Microalga Neochloris

oleoabundans for Biofuels Production and other Industrial Applications

This chapter is a comprehensive literature review that covers several key aspects associated with the cultivation and applications of microalgae in general and the microalga N. oleoabundans in particular. This species is a promising candidate for the production of biodiesel due to its high lipid content and biomass growth rate compared to other species cultivated for biofuels synthesis. Biomass composition, factors affecting the growth, and biomass and lipid productivities of this species were addressed. In addition, different growth conditions as well as alternative readily available nutrient media to support the growth of N. oleoabundans were addressed in this review.

 Chapter 3: Anaerobic Digestate as a Nutrient Medium for the Growth of the

Green Microalga Neochloris oleoabundans

In this chapter, the microalga Neochloris oleoabundans was cultivated in a sustainable manner using diluted anaerobic digestate to produce biomass as a potential biofuel feedstock. Prior to microalgae cultivation, the anaerobic digestate was characterized and several pretreatment methods that are different from the ones typically used in previous studies were investigated and their impact on the removal of suspended solids as well as other organic and inorganic matter was evaluated. The feasibility of

22 using the anaerobic digestate as a nutrient medium for the microalga N. oleoabundans was assessed by a bench-scale experiment which was conducted using multiple dilutions of the supernatant and filtered anaerobic digestate.

 Chapter 4: Cultivation of Scenedesmus dimorphus Using Anaerobic Digestate as a

Nutrient Medium

In this chapter, the microalga Scenedesmus dimorphus was cultivated phototrophically using unsterilized anaerobic digestate as a nutrient medium. A bench- scale experiment was conducted by inoculating the microalga S. dimorphus with a range of dilutions of the anaerobic digestate supernatant. The microalgae cultivation was then scaled up to 100 L open raceway ponds using two anaerobic digestate dilutions selected based on the bench-scale experiment. Two light intensities were tested for each dilution and the growth of the microalgae as well as the nutrients removal efficiencies were evaluated by frequent monitoring. To our best knowledge, this is the first study to address the cultivation of S. dimorphus in open raceway ponds using the anaerobic digestate as a nutrient medium.

 Chapter 5: Hydrodynamic Characterization of Microalgae Raceway Ponds Using

Tracer Tests

In this chapter, the average flow velocities and dispersion coefficients corresponding to variable water surface velocities were evaluated in a 100 L microalgae raceway pond. Furthermore, the effect of mixing velocity was assessed on the growth of the microalga S. dimorphus by testing three water surface velocities. Finally, the energy losses and power required to operate the ponds at different velocities were calculated and

23 compared with the biomass and lipid productivities so as to determine the feasibility of different degrees of mixing.

 Chapter 6: Modeling the Growth of Microalgae in Raceway Ponds Using Diluted

Anaerobic Digestate as a Nutrient Source

In this chapter, two biomass growth models were developed and tested to simulate the growth of the microalga Scenedesmus dimorphus in open raceway ponds using two different types of nutrient media. The main focus of the models was on the effect of light intensity on the growth of the microalga S. dimorphus and the light attenuation in deep microalgae cultures due to the microalgal biomass and the suspended solids in the nutrient media.

 Chapter 7: Conclusions and Recommendations

In this chapter, the findings of the previous studies were summarized and recommendations for future research were presented.

CHAPTER 2. A REVIEW OF THE CULTIVATION OF THE MICROALGA

NEOCHLORIS OLEOABUNDANS FOR BIOFUELS PRODUCTION AND

OTHER INDUSTRIAL APPLICATIONS1

2.1 Abstract

Microalgae cultivation for biofuels production and other applications has gained considerable interest recently. Despite their simple structures, microalgae can accumulate significant amounts of neutral lipids per dry cell weight compared to other energy crops.

Neochloris oleoabundans is a promising microalga known for its high lipid content and biomass growth rate compared to other species cultivated for biofuels synthesis; therefore, it is considered to be a suitable candidate for biodiesel synthesis. This review paper covers several key aspects associated with the cultivation and applications of the microalga N. oleoabundans. Biomass composition, factors affecting the growth, and biomass and lipids productivities of this species were addressed. In addition, different growth conditions as well as alternative readily available nutrient media to support the growth of N. oleoabundans were presented in this review.

Keywords: Neochloris oleoabundans, Microalgae, Biofuels, Biodiesel, Growth

Factors

1 This chapter will be submitted to the Algal Research Journal after the committee review.

2.2 Introduction

Environmental concerns have increased recently as a result of the exploitation of fossil fuels. In 2015, more than 80% of the global energy demand is supplied by fossil fuels. This has adversely affected the environment due to the significant increase in greenhouse gas (GHG) emissions (Chen et al., 2011; Mata et al., 2010; Williams &

Laurens, 2010). Hence, it was recognized that the dependence on fossil fuels has to be decreased gradually, although the complete abandoning of liquid hydrocarbons is unrealistic especially for aviation and shipping (Mata et al., 2010; Williams & Laurens,

2010). Renewable and sustainable energy sources such as solar, wind, hydro, and biomass have gained intensive interest as means of mitigating the environmental consequences linked to fossil fuels utilization. Nevertheless, combustible energy sources such as biomass have higher potential compared to other renewable sources (Greenwell et al., 2009; Lam & Lee, 2012; Mata et al., 2010).

Biofuels have been developed and produced on a large scale; yet, they contribute less than 1% of the energy demand worldwide (Williams & Laurens, 2010). The most common biofuels are bioethanol and biodiesel. Bioethanol is an alternative to the petroleum-derived transportation fuels and is produced by the fermentation of biomass feedstocks such as corn and sugar cane. Biodiesel is a renewable alternative to conventional diesel. It is produced by the transesterification of neutral lipids from plant or animal sources to form fatty acid methyl esters (FAME), namely biodiesel (Balat et al.,

2008; Chisti, 2008; Murray et al., 2012; Pruvost et al., 2009; Yang et al., 2011).

The first generation biofuels were extracted from edible oils such as vegetable oils and animal fats, and have been commercially produced in many countries. However, due to the controversy associated with the competition for arable lands as well as utilizing edible sources, first generation biofuels were considered unsustainable. Second generation biofuels emerged using alternative feedstocks such as agricultural residues, wood processing wastes, plants, and other non-edible sources. The commercial production of second generation biofuels has not been attained due to cultivation challenges such as irrigation and heavy fertilization. Third generation biofuels such as microalgae-derived biofuels arose as sustainable alternatives for biodiesel and bioethanol synthesis (Brennan & Owende, 2010; Chen et al., 2011; Lam & Lee, 2012; Mubarak et al., 2015).

Microalgae biomass is a promising feedstock for several applications such as the production of biofuels (biodiesel, hydrogen, bioethanol, and methane), valuable biochemical and medical products, nutritional supplements, and as feed for aquaculture

(Milledge, 2011; Williams & Laurens, 2010). Cultivating microalgae for biofuels synthesis has several advantages such as eliminating the need for fertile soil and freshwater (especially for marine algae) and higher growth rates, productivities, oil content, and photosynthetic efficiency compared to other energy crops (Amin, 2009;

Mata et al., 2010; Williams & Laurens, 2010). However, biofuels in general and algal biodiesel in particular are not economically feasible yet in comparison to conventional fossil fuels. To reduce the cost of biodiesel and make it competitive, the biomass feedstock production and processing costs have to be decreased significantly (Mata et al.,

2010; Milledge, 2011). Hence, it is desirable to explore cost effective cultivation techniques that will improve the feasibility of algal biodiesel. It is also essential to investigate strains with high biomass and oil productivities.

This review paper focuses on the promising microalga Neochloris oleoabundans and its applications for biodiesel production besides other applications. This species has attracted considerable attention recently as a candidate for biodiesel synthesis due to its remarkable biomass and lipids productivities compared to other well-studied microalgae species for biofuels production. Hence, this study addresses various aspects of the cultivation of this microalga such as biomass composition, factors affecting growth, reported biomass concentrations, lipids, and fatty acids contents. Moreover, this study covers different cultivation modes of this microalga while addressing several alternatives that will potentially improve the feasibility of cultivation.

2.3 Microalgae Cultivation

2.3.1 Microalgae definition and growth conditions

Microalgae are unicellular microorganisms that convert solar energy to chemical energy via photosynthesis (Giovanardi et al., 2014; Villa et al., 2014). Microalgae utilize light energy to release electrons from water to temporary carriers such as NADPH. These carriers function as reducing agents to fix carbon in the functional biomass and storage compounds. Any excess energy is dissipated in the photosystems through chlorophyll fluorescence (da Silva et al., 2009; Klok et al., 2013b).

The most popular cultivation mode of microalgae is phototrophic. Heterotrophic and mixotrophic cultivation modes are used as well to grow microalgae but are less

28 common (Giovanardi et al., 2014; Morales-Sánchez et al., 2013; Morales-Sánchez et al.,

2014). In fact, mixotrophic cultivation of microalgae using organic carbon along with light supplement may improve biomass growth rates and lipids accumulation in comparison to other cultivation modes (Baldisserotto et al., 2014; Giovanardi et al.,

2014).

2.3.2 Applications and products from microalgae

Microalgae biomass can be a feedstock for the production of biofuels, pharmaceuticals, food for mussels and some fish, pigments, poly-unsaturated fatty acid

(PUFA), and proteins as shown in Figure 2.1 (Barnhart, 2006; Baldisserotto et al., 2014; de Winter et al., 2014; Gatenby et al., 2003; Klok et al., 2013b; Morales-Sánchez et al.,

2014; Santos et al., 2012). In addition, microalgae cultivation can play an important role in carbon dioxide sequestration from flue gases to mitigate GHG emissions (da Silva et al., 2009; Mata et al., 2010). Microalgae biodiesel, a third generation biofuel produced via the transesterification of the algal neutral lipids, is a promising future alternative to the conventional diesel (Murray et al., 2012; Pruvost et al., 2009; Santos et al., 2012;

Yang et al., 2011). Advantages of microalgal biodiesel over conventional diesel include higher cetane number, shorter ignition delay, lower NOx emissions, lower sulfur content, and higher oxidative stability (Mata et al., 2010; Sun et al., 2014). The remainder of the algal biomass after oil extraction can be used as a fertilizer or further processed for other products (Mata et al., 2010). Bioethanol can also be produced from the microalgal biomass by the fermentation of carbohydrates; however, biodiesel production is more

29 desirable as the heating value of biodiesel is 37.3 MJ/kg compared to 26.7 MJ/kg for ethanol (Chisti et al., 2008; Murray et al., 2012; Sun et al., 2014).

The large-scale microalgae cultivation for biofuels production is promising due to the high efficiency of microalgae in carbon fixation, solar energy conversion, and oil accumulation compared to other energy crops (Franchino et al., 2013; Gouveia &

Oliveira, 2009; Murray et al., 2011; Villa et al., 2014; Wang & Lan, 2011a). Even though the current trend in microalgae cultivation is biofuels synthesis, other byproducts from microalgal biomass may have greater value per kg compared to biofuels (Murray et al.,

2012).

Figure ‎2.1: Application and products from microalgae biomass

2.3.3 Lipids accumulation in microalgae

Microalgal lipids are classified as storage (non-polar) lipids and structural (polar) lipids. Storage lipids are mostly made of saturated and some unsaturated fatty acids whereas the polyunsaturated fatty acids content is high in the structural lipids. Storage lipids such as triacylglycerol (TAG) which is the most predominant form of non-polar lipids are targeted for the production of biodiesel. On the other hand, polar lipids such as phospholipids and sterols are key structural components of the cell membranes (Sharma et al., 2012).

Lipids accumulation in microalgae occurs due to a metabolism shift that results from stress conditions such as nitrogen starvation, high pH, phosphate limitation, salinity, carbon supplement, light intensity, photo-oxidative stress, and iron concentration

(Garibay-Hernández et al., 2013; Giovanardi et al., 2013; Santos et al., 2013; Sun et al.,

2014). Generally, the abundance of carbon accompanied with shortage of at least one nutrient (typically nitrogen) triggers lipids accumulation in microalgae (Morales-Sánchez et al., 2014).

The most common strategy to increase the content of storage compounds such as

TAG and starch is nitrogen starvation. It is hypothesized that under nitrogen limited conditions, some of the excess energy produced in the photosystems is directed towards the synthesis of storage compounds instead of protein or chlorophyll (Bona et al., 2014;

Breuer et al., 2012; Klok et al., 2013b). As a result, a two-stage microalgae cultivation is often recommended, where the cell density is maximized initially by providing sufficient nutrients followed by a second stage in which the harvested microalgae biomass is further

31 cultivated under stress conditions to optimize lipids/carbohydrates synthesis (Gouveia et al., 2009; Klok et al., 2013a; Sun et al., 2014).

2.3.4 Challenges facing microalgae cultivation and potential solutions

One of the major obstacles facing the large scale cultivation of microalgae is the high capital and operational costs due to the nutrients requirements and the high energy input necessary for providing light, mixing, and pumping (Levine et al., 2011; Li et al.,

2008; Franchino et al., 2013; Santos et al., 2013; Urreta et al., 2014; Wang & Lan, 2011a;

Yang et al., 2011). Harvesting the algal biomass adds to the production cost as it accounts for 20 – 30% of the total production cost (Davis et al., 2012; Salim et al., 2012). Hence, it is vital to seek alternative means of cultivation and harvesting to reduce the production costs of algal biomass.

Process design, operation, and optimization are key aspects to overcome the economical barrier of microalgae cultivation. Improving the photosynthetic efficiency and light supply and screening microalgal strains based on their lipid content, productivity, and composition improve the economic feasibility of microalgae cultivation

(Breuer et al., 2012; Li et al., 2008). Furthermore, utilizing the carbon dioxide portion of flue gas can decrease the production cost considerably besides the environmental advantages of carbon sequestration (Franchino et al., 2013; Urreta et al., 2014; Wang &

Lan, 2011a). Cultivating microalgae using freshwater is probably not sustainable; due to the huge demand on freshwater as a result of world population increase. Therefore, microalgal species that can tolerate brackish and saline media or utilize wastewaters such as secondary wastewater effluent, industrial, municipal, and agricultural wastewaters, and

32 anaerobic digestion (AD) liquid waste as a nutrient source are attractive (Baldisserotto et al., 2014; Franchino et al., 2013; Levine et al., 2011; Urreta et al., 2014; Wang & Lan,

2011a; Yang et al., 2011). Additionally, the use of exhausted culture media can be viable for some microalgae species. According to Sabia et al. (2015), it was feasible to grow the microalga Neochloris oleoabundans phototrophically using exhausted growth medium with the supplement of nitrate and phosphate; thus, preserving valuable freshwater resources.

Outdoor cultivation systems are widely applied as a way to reduce the capital and operational costs related to microalgae cultivation; however, the main disadvantage of outdoor systems is the contamination risk by other algal species, bacteria, zooplankton, and viruses (Chen et al., 2013; Peng et al., 2015; Sabia et al., 2015). On the other hand, closed cultivation systems are adopted when the target is the production of valuable compounds (Sabia et al., 2015). One potential solution to the contamination risk is the cultivation of native species or species that tolerate extreme conditions such as

Arthrospira and Dunaliella, and fast-growing species such as Chlorella, Scenedesmus, and Nannochloropsis which can resist predation and cross-contamination (Menetrez,

2012; Muller-Feuga et al., 2012).

Harvesting techniques such as bio-flocculation, auto-flocculation, and electro- coagulation-flocculation are promising approaches that are more energy efficient than traditional methods such as centrifugation (Olguín et al., 2015b; Salim et al., 2012). Bio- flocculation is a sustainable pre-harvesting technique that has several advantages such as the lower energy requirement and no chemical addition (Salim et al., 2012). Improving

33 the amount of extractable lipids is another essential goal in the viability of microalgal biodiesel. For example, utilizing some bacterial strains that produce extracellular algicidal substances may help increase the amount of extractable lipids (Lenneman et al.,

2014).

Finally, exploitation of the entire algal biomass, not just the lipid fraction, should be considered for the production of biofuels as well as other high value products so as to improve the economics of microalgal cultivation (Klok et al., 2013b; Tibbetts et al.,

2015; Wang & Lan, 2011a).

2.4 Neochloris oleoabundans

2.4.1 Introduction to the microalga Neochloris oleoabundans

Neochloris oleoabundans, also known as Ettlia oleoabundans, is an oleaginous freshwater unicellular microalga from the Chlorophyceae class and the Chlorococcaceae family. It can also be cultivated in saline media with salt concentrations similar to seawater (Baldisserotto et al., 2014; Garibay-Hernández et al., 2013; Gouveia et al.,

2009; Popovich et al., 2012; Rismani-Yazdi et al., 2012; Sun et al., 2014; Yang et al.,

2013). In terms of growth modes, this microalga is capable of growing under phototrophic (most common), heterotrophic, or mixotrophic conditions (Morales-Sánchez et al., 2014).

The reproduction method of the N. oleoabundans is by zoospores (da Silva et al.,

2009). The cell shape of this microalga is generally spherical and 3 – 3.5 μm in diameter

(Figure 2.2), but this depends on the cultivation mode as well as the growth phase

(Baldisserotto et al., 2014; Giovanardi et al., 2014). Davis et al. (2012) indicated that the

34 average cell diameter increased to 5.2 μm under nitrate starvation conditions. Similarly,

Peng et al. (2015) stated that the cell size increased from 3.53 to 6.31 μm when the sodium bicarbonate concentration in the nutrient medium was increased from 0 to 160 mM, which was also thought to cause stress on the cells. Moreover, the cell shape may flatten out under different growth conditions such as those reported by Giovanardi et al.

(2014) and Giovanardi et al. (2013) when the cells were in the stationary phase of mixotrophic cultivation.

Figure 2‎ .2: N. oleoabundans cells (adapted from the Culture Collection of Algae at The University of Texas at Austin website)

Despite not being studied widely, this microalga has received attention since the

1980s due to its high growth rates and ability to accumulate high quantities of neutral lipids, mainly TAG dominated by monounsaturated fatty acids, making it a promising candidate for biofuels production (Giovanardi et al., 2014; Sousa et al., 2013a; Yang et

35 al., 2013). According to Li et al. (2008), most literature on this topic focused on this microalga as a feeding source for aquaculture species such as mussels.

2.4.2 N. oleoabundans biomass composition

As with all organisms, the major biomass components of the microalga N. oleoabundans are proteins, carbohydrates, and lipids. The first priority in the electron utilization throughout the photosynthetic process is cell maintenance, followed by functional biomass synthesis. Starch is the primary storage compound, whereas TAG production is limited to nitrogen starvation conditions (Klok et al., 2013b). Additionally, factors such as the nitrogen phosphorus ratio for phototrophic growth and carbon nitrogen ratio for heterotrophic growth affect the composition of the algal biomass; as this can shift the production towards one storage compound over the other (Morales-Sánchez et al., 2013).

2.4.2.1 Proteins

Proteins are structural and metabolic components of the algal biomass. Protein content in the algal cells is typically high during the exponential growth phase (Mata et al., 2010; Williams & Laurens, 2010). The overall protein content in N. oleoabundans ranges from 30 to 45% per dry cell weight (DCW), and it depends on the growth phase as well as nutrient availability (Gatenby et al., 2003; Tibbetts et al., 2015). Less protein is often produced throughout the stationary phase (Gatenby et al., 2003; Morales-Sánchez et al., 2013). Protein content tends to decrease in saline media and under insufficient nitrogen concentrations and can get as low as 10% of the DCW (Garibay-Hernández et al., 2013; Popovich et al., 2012).

2.4.2.2 Carbohydrates

Carbohydrates are the photosynthetic early products that are utilized for synthesizing other cellular components (Williams & Laurens, 2010). Carbohydrate content depends on stress conditions; as microalgae tend to accumulate more carbohydrate towards the stationary phase or under starvation conditions (Mata et al.,

2010). Gatenby et al. (2003) reported that more carbohydrate per DCW was produced in later stationary phase compared to earlier stages unlike protein production. The overall carbohydrate content of N. oleoabundans according to Gatenby et al., (2003) was approximately 20%. Similarly, Garibay-Hernández et al. (2013) demonstrated that low nitrate concentrations (≤ 1 mM) enhanced the accumulation of more carbohydrates towards the end of cultivation period. The carbohydrate contents at these low nitrogen concentrations ranged from 23% to 38%, 85% of which was starch. However, other researchers reported that there was no significant impact of nitrogen concentration on the carbohydrate content. Morales-Sánchez et al. (2013) stated that the carbohydrate content of the microalga N. oleoabundans was approximately 32% under heterotrophic growth which was not affected significantly by nitrogen limitation. Popovich et al. (2012) revealed that the average carbohydrate content per DCW was 38% in saline media, which did not differ significantly based on nitrogen availability.

2.4.2.3 Lipids

Lipids are molecules soluble in organic solvents. Lipids are major components of the algal biomass and can function as an energy reserve. Furthermore, cell membranes are made of phospholipids and glycolipids (Mubarak et al., 2015; Williams & Laurens,

2010). Gatenby et al. (2003) stated that analogous to carbohydrate content, lipid content can vary significantly depending on the growth conditions and nutrients availability.

They also reported that the total lipid content per DCW in the microalga N. oleoabundans increased from 19% in the early lag phase to 37% in the late stationary phase, while the overall total lipid content throughout all growth phases was 28.5%. Other researchers indicated significantly higher lipid contents. For instance, Morales-Sánchez et al. (2013) reported that the lipid content of the microalga N. oleoabundans increased from 23.3 to

51.7% under heterotrophic growth conditions, when this microalga was cultivated at high

C/N ratio (278) compared to a ratio of 17. Gouveia et al. (2009) found that the highest lipid content per DCW was 56% which was obtained when this microalga was cultivated in a nitrate limited solution at 30 °C and without carbon dioxide enrichment. Tornabene et al. (1983) revealed that under nitrogen limited conditions, the microalga N. oleoabundans accumulated 35 – 54% of its dry weight as lipids, 80% of which was TAG. They were also able to detect seven individual sterols; the major one was C-28 sterol. According to

Gatenby et al. (2003), the average sterol concentration found in N. oleoabundans was 6.4

μg/mg lipids, which was in the form of ergostatrienol, ergostadienol, and ergostenol.

2.4.2.4 Pigments (chlorophyll and carotenoids)

Chlorophyll is an essential element in plants due to its role in light harvesting

(Hosikian et al., 2010). Chlorophyll content in the microalga N. oleoabundans depends on several factors such as nitrogen concentration, light intensity, and growth mode. Sun et al. (2014) indicated that the chlorophyll content in this microalga decreased from an initial value of 3% (per DCW) to 1% under nitrogen starvation. A similar range of 1% to

3% was reported by Sousa et al. (2013a) and Sousa et al. (2013b); however, under different testing conditions where the length of light exposure and light intensity were found to be inversely related to the chlorophyll content. Li et al. (2008) concluded that low nitrate concentrations in the nutrient medium decrease the total chlorophyll content in the microalga N. oleoabundans, especially at initial nitrate concentrations less than 15 mM. It was hypothesized that when the nitrate is completely exhausted in the nutrient medium, cells tend to consume the intercellular nitrogen in the nitrogen-rich chlorophyll.

Urreta et al. (2014) reported similar effects under lower nitrate concentrations; as the chlorophyll content decreased three fold when the initial nitrate concentration decreased from 9 to 3 mM. Under heterotrophic conditions, Morales-Sánchez et al. (2013) stated that the total chlorophyll content was 1 – 1.5% per DCW which is significantly lower than those values under phototrophic conditions.

Carotenoids are protective pigments, and their content in the cells is directly related to the light intensity (Sousa et al., 2013b). Urreta et al. (2014) revealed that carotenoid content in the microalga N. oleoabundans depends on nitrate concentration. In their experiment, the initial carotenoids content was 38.7 mg/g DCW when the nitrate concentration in the nutrient medium was 9 mM which dropped to 20.9 mg/g DCW at 3 mM nitrate concentration. 60% of the total carotenoid content was lutein and astaxanthin.

β-carotene was also detected and its content was independent of nitrate concentration.

Sousa et al. (2013a) reported that the N. oleoabundans carotenoid content ranged from

2.3 to 4.9 mg/g DCW at 200 and 500 μmol/m2/s, respectively. Much lower

39 concentrations were reported by Goiris el al. (2012) who indicated that the content of carotenoids in the industrially cultivated N. oleoabundans was 1.56 mg/g DCW.

2.4.3 Stress conditions in N. oleoabundans and lipids accumulation

Similar to other microalgae species, nitrogen starvation is the most common stress condition leading to lipid accumulation in this microalga (Baldisserotto et al., 2014).

Under nitrogen limited conditions, N. oleoabundans tends to dissipate the excess energy as heat and fluorescence as a predominant way of protecting the photosystems, and a fixed 8.6% of the excess electrons are utilized for the production of energy sinks such as

TAG (Klok et al., 2013a; Klok et al., 2013b). The microalga N. oleoabundans tends to accumulate carbohydrate as a first energy reserve; while lipids (mainly TAG) are the secondary energy sink are be produced as a long-term reserve with the potential of the conversion of carbohydrates to TAG along the starvation process (Breuer et al., 2012;

Sun et al., 2014). It was suggested that a period of 2 d is sufficient for lipids/carbohydrates synthesis (Sun et al., 2014). Moreover, nitrogen depletion in the microalgae culture is not the sole reason for lipids accumulation; Urreta et al. (2014) indicated that nitrogen level prior to starvation is also a key factor for lipids synthesis.

Many studies concluded that the majority of the neutral lipids accumulated in the starved microalga N. oleoabundans are TAG unlike the non-starved cultures where the majority of the neutral lipids are mono- and diglyceride (Beal et al., 2010). For instance,

Tornabene et al. (1983) reported that up to 80% of the total lipids accumulated in the microalga N. oleoabundans under nitrogen limited conditions were TAGs; whereas

Urreta et al. (2014) indicated that more than 95% of total lipids accumulated in the

40 nitrogen-starved N. oleoabundans cells were TAG or neutral lipids. Davis et al. (2012) compared gradual to punctuated nitrate limitation on the microalga N. oleoabundans and found that the highest neutral lipid content coupled with an increase in the cell size was achieved under the latter starvation approach. Bona et al. (2014) concluded that a semi- continuous cultivation approach with limited nitrogen content (24.75 mg N/L) resulted in higher biomass productivity and total fatty acids content compared to nitrogen depleted cultures.

Besides nitrogen starvation, Santos et al. (2012) stated that elevated pH values as high as 10 accompanied with nitrogen limitation resulted in more lipids accumulation

(35% per DCW) compared to lower pH values. Santos et al. (2014) studied the impact of salinity, pH, and light intensity on the lipids accumulation of this microalga. They recommended a two-stage cultivation process under low light intensity (58 μmol/m2/s) where biomass is produced in the first stage under alkaline-saline conditions (pH 8), while in the second stage the microalgae cells are stressed under nitrogen limited conditions at pH 10. Increased sodium bicarbonate concentration has been linked with stress on N. oleoabundans cells featured by an increase in the cell size (Peng et al., 2015).

2.4.4 Harvesting

General techniques used for dewatering and harvesting the microalgal biomass in general and the microalga N. oleoabundans in particular include flocculation, gravity sedimentation, floatation, filtration, electrophoresis, and centrifugation. Centrifugation and filtration are not sustainable due to their high capital and operation costs.

Flocculation on the other hand is a more feasible option due to lower energy input

(Brennan & Owende, 2010; Chen et al., 2011; Lam & Lee, 2012; Lam et al., 2015). The microalga N. oleoabundans is a non-self-flocculating microalga; thus, it is required to add a flocculant for effective harvesting of the microalgal biomass (Olguín et al., 2015b).

Beach et al. (2012) reported that chitosan, a biopolymer derived from the exoskeletons of crustaceans, was the most effective flocculant for the microalga N. oleoabundans compared to ferric sulfate and alum. In addition, this species can be bio-flocculated with the addition of other algal species such as the marine alga Tetraselmis suecica (Salim et al., 2012). Lam et al. (2015) investigated the use of commercially available polyacrylamide-based flocculants that are typically used for wastewater treatment. They found that 43 – 109 mg flocculant/g biomass was the optimum dosage for the effective harvesting of N. oleoabundans. However, they concluded that using those commercial flocculants adds a considerable cost to the harvesting of this microalga. On the other hand, Davis et al. (2012) indicated that when the N. oleoabundans cells were placed under sharp (as opposed to gradual) nitrate starvation, cells which contained the highest quantities of TAG had the largest diameter, suggesting size selective membranes for harvesting.

One of the techniques that can be used to improve the lipids extraction from this microalga is by utilizing algicidal bacteria. Algicidal strains such as Pseudomonas pseudoalcaligenes and Aeromonas hydrophil were able to utilize the N. oleoabundans cells as a sole carbon and nitrogen source. This process resulted in deflating and withering the algal cells, as well as a six fold increase in the extractable lipids. de Winter et al. (2013) indicated that the circadian clock plays an important role in the harvesting

42 time; as the peak lipid, starch, and protein contents are typically observed prior to cell division.

2.4.5 Applications

Besides the potential of utilizing the microalga N. oleoabundans for biofuels and biodiesel synthesis, there are additional applications that make the cultivation of this microalga attractive. One of the important functions of microalgae is nutrient removal.

Wang and Lan (2011a) noticed that this microalga was able to remove the entire N and P

- 3- from an artificial wastewater media up to 218 mg N-NO3 /L and 47 mg P-PO4 /L. These results emphasized the importance of this species as these removal efficiencies were significantly higher than other well-studied species such as Chlorella vulgaris, Spirulina platensis, and Chlorella pyrenoidosa (Wang & Lan, 2011a). When cultivated using diluted AD liquid waste; Franchino et al. (2013) reported that this microalga was capable of removing ammonium and phosphate at 99% and 94% removal efficiencies, respectively. Furthermore, microalgae have a role in nitrogen fixation. According to Villa et al. (2014), Azotobacter vinelandii bacteria fix and supply the atmospheric nitrogen to agricultural crops under aerobic conditions in the forms of siderophores, ammonia, urea, or proteins, and the microalga N. oleoabundans was able to utilize the A. vinelandii siderophores as a nitrogen source (Villa et al., 2014).

Antioxidants, which are food additives used to extend the shelf life of food products, can be produced from unicellular microalgae such as N. oleoabundans as a replacement of the synthetic antioxidants. This is due to the cellular content of carotenoids and other phenolic compounds; as carotenoids are well known for their

43 contribution to antioxidant functions within the cell as well as their applicability in food industry as dyes and food additives which can make the production of this microalga more economical. However, simultaneous optimization of TAG and carotenoids might not be achievable with this microalga (Goiris el al., 2012; Urreta et al., 2014).

Additionally, biopolymers such as polysaccharides can be produced from the microalgae biomass. Wu et al. (2011) revealed that under mixotrophic conditions and using lactose as a carbon source, the microalga N. oleoabundans produced up to 5 g/L of high viscosity polymers. Medical applications of biopolymers, sulfated polysaccharides in particular, include anticoagulation as well as preventing HIV replication in cell culture. A more common application of biopolymers is polymer flooding for oil recovery enhancement.

Adding polysaccharides to water increases its viscosity, which can increase the amount of crude oil recovered from drilling operations.

2.5 Factors Affecting the Growth of N. oleoabundans

2.5.1 Nutrients

Nutrient availability (mainly carbon, nitrogen, and phosphorus) is a key factor for the growth of microalgae (Lam & Lee, 2012). The approximate stoichiometry of the macro-nutrients uptake and biomass conversion of the microalga N. oleoabundans is given by Equation 2-1 (Pruvost et al., 2009):

CO2 + 0.148HNO3 + 0.14H2SO4 + 0.012H3PO4 + 0.751H2O Equation 2-1 → CH1.715O0.427N0.148S0.014P0.012 + 1.437O2

In general, carbon makes up approximately 50% of the algal biomass (Milledge,

2011). Carbon is the most important building block for the microalga N. oleoabundans, which has a significant impact on the total and neutral lipid content as well as the chlorophyll content (Gouveia et al., 2009; Uretta et al., 2014). It has been reported that the growth rate and biomass productivity can be doubled with carbon dioxide enrichment

(5% v/v) compared to cultivation without enrichment (Gouveia et al., 2009; Kwak et al.,

2015; Urreta et al., 2014).

Nitrogen is also an essential nutrient for microalgae biomass growth and lipids accumulation. Additionally, nitrogen affects the nitrogen-containing chlorophyll content; as the chlorophyll may function as a nitrogen reservoir under nitrogen limited conditions

(Li et al., 2008; Rismani-Yazdi et al., 2012; Urreta et al., 2014). Regarding nitrogen form,

Franchino et al. (2013) indicated that in general, ammonium is the preferred form of nitrogen for microalgae. Wang and Lan (2011a) hypothesized that consuming ammonium is easier than nitrate for N. oleoabundans; as the majority of ammonium from enriched secondary municipal wastewater effluent was consumed within the first day of cultivation unlike nitrate, which took longer time. On the contrary, other researchers stated that nitrate is the preferred form of nitrogen for this microalga. For instance, Li et al. (2008) found that sodium nitrate was the most favorable nitrogen source when compared to ammonium bicarbonate and urea. Levine et al. (2011) revealed that the microalga N. oleoabundans grew significantly better with sodium nitrate compared to ammonium chloride. They also indicated that ammonium concentrations as high as 100 mg/L may have toxic effects; however, the toxicity can be avoided by following a fed-batch or

45 continuous flow cultures. Kwak et al. (2015) also emphasized that the growth of N. oleoabundans was significantly better with nitrate compared to ammonium.

Several studies discussed the optimum nitrogen concentration for the growth of N. oleoabundans. Sun et al. (2014) found that the optimum sodium nitrate was 5.9 mM (82 mg N/L). Li et al. (2008) reported that the highest biomass concentration was observed at

10 mM (140 mg N/L) sodium nitrate concentration whereas the highest lipid content was achieved with 3 mM (42 mg N/L) sodium nitrate concentration; thus, they recommended a nitrogen concentration of 5 mM to optimize biomass and lipids production. According to Levine et al. (2011), the optimum nitrogen concentration for biomass was 7.3 mM

(100 mg N/L). Likewise, Kwak et al. (2015) found that 7 mM (91 mg N/L) nitrogen concentration was optimum whether nitrogen was in the form of ammonium or nitrate.

Urreta et al. (2014) recommended a potassium nitrate concentration of 3 mM to achieve both maximum biomass and TAG productivities. Nitrogen concentrations higher than 10 mM may inhibit the algal growth (Li et al., 2008).

Phosphorous is an important element in plants in general and is found in the ATP, nucleic acids, phospholipids, and other molecules (Schachtman et al., 1998). The preferred phosphorous form to algae is orthophosphate (Franchino et al., 2013). Besides,

N/P ratio is a key aspect in the growth of microorganisms in general and microalgae in particular. Phytoplankton generally exhibit flexibility in their stoichiometry depending on nutrients availability. Based on the ecological conditions, the optimum atomic N/P ratio for phytoplankton varies from 8.2 to 45 (Klausmeier et al., 2004). Wang and Lan (2011a) studied the impact of the N/P ratio on the growth of the microalga N. oleoabundans based

3- on constant N or P concentrations in artificial wastewater. At a constant 108 mg P-PO4

/L phosphorus concentration, N was varied by enriching the media with NaNO3 so as the resulting N/P ratios were in the 0.42 – 2.02 range. Under these conditions of excessive

- phosphorus concentrations, the optimal cell growth was observed at 140 mg N-NO3 /L

- (N/P = 1.33) with a DCW of 3 g/L. When N was held constant at 140 mg N-NO3 /L, and the N/P ratio was varied between 3 and 26.4 by changing the phosphorus concentration,

3- the highest biomass concentration of 2.1 g/L was observed at 47 mg P-PO4 /L (N/P = 3).

Olguín et al. (2015b) reported that the N/P ratio in the nutrient medium, which was the effluent from the AD of vinasse, was 5.88, which promoted the growth of N. oleoabundans.

2.5.2 Light

Light is the most important limiting factor in the microalgae cultivation (Mata et al., 2010). Microalgae in general adapt to varying light intensities in what is called photoacclimation by adjusting their pigment content as well as the absorptive cross- sectional area (Klok et al., 2013a; Yang et al., 2013). According to Klok et al. (2013b), the hypothetical minimum absorptive cross section for the microalga N. oleoabundans is

0.023 m2 per g of functional biomass in order to be able to capture light under severe nitrogen starvation or in a steady state conditions with extremely high light intensities.

Chlorophyll a is the primary photosynthetic pigment which absorbs light within the 400 –

450 nm and 650 – 700 nm wavelengths (Williams & Laurens, 2010). Storage compounds such as lipids do not absorb light, they scatter it instead. Light scattering results in lengthening the light path. A common way of evaluating transmission in a suspension is

47 by using Lambert-Beer’s law, which depends on the absorption characteristics of the biomass, as well as the length of the light path. However, it does not take light scattering into account (Klok et al., 2013b). Photoinhibition occurs when the microalgae are cultivated under high or over-saturating light intensities. High irradiance results in the release of excess electrons; thus, forming highly reactive oxygen species such as H2O2 and the highly reactive oxygen singlet. The latter damages the photosynthetic activity of the microalgae cells (Sousa et al., 2012; Sousa et al., 2013a).

Under low average light supply rates combined with nitrogen repletion, more functional biomass such as protein can be produced, while for higher starch content, high average light intensities and limited nitrogen conditions are required. However, the conditions for a higher volumetric starch production are low light and sufficient nitrogen.

TAG optimization will always take place under nitrogen limited conditions and low light intensities (Klok et al., 2013b). Several studies have indicated that the saturation light intensity for the growth of the microalga N. oleoabundans is in the 180 – 220 μmol/m2/s range. For instance, Sun et al. (2014) revealed that 200 μmol/m2/s light intensity was optimal for the growth, 100 μmol/m2/s was optimal for carbohydrate production, and that a slight improvement in the total lipid content was observed when the light intensity was increased from 50 to 200 μmol/m2/s. They recommended 200 μmol/m2/s light intensity for both the growth and TAG production due to fatty acids profile consideration. Klok et al. (2013a) stated that for the highest TAG content, light intensity should be in the 50 -

150 μmol/m2/s range, whereas the optimum yield can be obtained in the 50 – 100

μmol/m2/s range. Sousa et al. (2012) reported that the saturation light intensity for this

48 microalga is 218 μmol/m2/s, whereas Wahal and Wiamajala (2010) indicated that photosaturation occurs at 180 μmol/m2/s. Klok et al. (2013a) revealed that an increase in the specific growth rate from 1.15 to 1.74 d-1 was witnessed when the light intensity was increased from 70 to 200 μmol/m2/s under surplus nitrogen concentrations. Urreta et al.

(2014) pointed out that there was no significant difference in the biomass concentration and TAG content when the microalga N. oleoabundans was cultivated under 240 and 400

μmol/m2/s light intensities.

Even though most studies in the literature report light intensity in terms of quantum, Zhao et al. (2015) tested different combinations of red and blue LED mixing ratios. They concluded that mixed LED wavelengths yielded higher growth rates compared to a single wavelength. In particular, when the microalga N. oleoabundans was cultivated in filtered and sterilized AD slurry and under 200 μmol/m2/s light intensity, a

50:50 red/blue combination yielded the highest growth rate as well as the highest COD and total phosphorus removal efficiencies. Nevertheless, the highest total nitrogen removal efficiency was achieved with 70:30 red/blue ratio.

2.5.3 Temperature

Temperature is also a key limiting factor that affects the growth of microalgae

(Mata et al., 2010). According to Yang et al. (2013), temperature is a more significant factor to the growth of N. oleoabundans than light, and it affects nutrients consumption; as microalgae tend to consume more nitrates at higher temperatures. However, in their experimental design, the microalgae were cultivated at ambient carbon dioxide concentrations, which can explain the lower need to light due to the lower

49 photosynthesis. Wang and Lan (2011a) investigated the effect of temperature on the microalga N. oleoabundans when cultivated in an artificial wastewater. They concluded that a range of 25 – 30 °C provided the optimum growth; while inhibition occurred at 32

°C. Gouveia et al. (2009) revealed that the growth of this microalga was faster at 30 °C compared to 26 °C. On the other hand, Yang et al. (2013) reported that the highest biomass concentration was observed at 10 °C and 70 μmol/m2/s light intensity, whereas the highest maximum specific growth rate was observed at 15 °C, and no growth was observed at 5 °C. It was speculated that the microalga N. oleoabundans tolerated temperatures as low as 5 °C due to the fact that this species was originally isolated from

Rub al Khali’s desert in Saudi Arabia from the sand dunes surface. As the desert conditions fluctuate between day and night times, the frigid conditions at night may explain the tolerance to such low temperatures. Further, the growth of N. oleoabundans was inhibited after 11 days at 35 °C. Despite the poor biomass growth at 35 °C, the quality of raw biodiesel at this temperature was enhanced (Yang et al., 2013).

2.5.4 pH

According to Murray et al. (2012), the growth of the microalga N. oleoabundans was significantly better at pH values in the 7 – 7.56 range compared to fluctuating pH

(6.75 – 9.89). Santos et al. (2013) concluded that an increase in the pH from 8.2 to 10 in a saline medium resulted in a lower efficiency in light utilization; hence, a lower biomass concentration. Kwak et al. (2015) reported that at pH > 8, the growth of this microalga was inhibited. Santos et al. (2012) on the other hand pointed out that the growth was inhibited at pH 10.5 coupled with 720 mM salt concentration in an artificial seawater.

2.5.5 Dissolved oxygen concentration

In general, high dissolved oxygen concentrations have negative impact on the algal growth; as high O2/CO2 ratio favors photorespiration over photosynthesis (Sousa et al., 2012; Sousa et al., 2013b). It is therefore a general practice in microalgae cultivation to use a degassing unit in order to reduce the elevated oxygen concentrations. For N. oleoabundans, Sousa et al. (2013b) reported that the specific growth rate was not affected by varying the oxygen partial pressures from 0.21 to 0.63 bar followed by a rapid degassing under 200 μmol/m2/s light intensity. However, a significant impact was observed on the growth of this microalga by the length of dark period in the degasser.

Sousa et al. (2013a) found that the optimum growth of N. oleoabundans grown in

2 artificial seawater under 500 μmol/m /s was achieved at PO2 = 0.21. Higher oxygen partial pressures resulted in decreased growth rates. Sodium bicarbonate has been successfully employed to mitigate the inhibition resulting from elevated oxygen concentrations for the microalga N. oleoabundans, especially at sub-saturation light intensities (Sousa et al.,

2012; Sousa et al., 2013a).

2.5.6 Salinity

Salinity affects the growth of microalgae and can impose osmotic/salt stress; thus, affecting the selective ion permeability (Mata et al., 2010). According to Santos et al.

(2013), salinity does not affect the microalga N. oleoabundans efficiency in light utilization. On the other hand, salinity has an impact on the cell wall; as N. oleoabundans tend to have a much thicker wall under high salinity compared to freshwater media

(Lenneman et al., 2014). Salinity may also help provide better control over outdoor

51 cultures. It has been reported that a combination of 160 mM sodium bicarbonate concentration and pH 9.5 prevented the growth of protozoa in the outdoor N. oleoabundans cultures, besides promoting high lipid accumulation. This was referred to the effect of salinity caused by the increased sodium bicarbonate concentration rather than the high pH; due to the lack of protective cell walls in the protozoa; thus, increasing the sensitivity of these cells to osmotic pressure (Peng et al., 2015).

2.5.7 Circadian clock

de Winter et al. (2014) indicated that besides temperature, light, and the availability of nutrients, circadian clock is a key factor affecting biomass composition.

Circadian clock is an endogenous biochemical rhythm governed by the natural day and night times which controls the schedule of metabolic and physiological processes such as photosynthesis and cell division within the microalgae cell on a daily basis. For instance, photosynthesis takes place during the day whereas cell division is a night time activity

(de Winter et al., 2014; de Winter et al., 2013). de Winter et al. (2014) studied the effect of circadian clock under steady state conditions with a continuous light supply and constant temperature (30 °C). They found that higher growth rates, starch, and total fatty acids contents were observed during the natural day time.

2.5.8 Other factors

Concentration of other macro and micronutrients can affect the biomass growth and lipid accumulation of microalgae. For instance, iron and magnesium may affect the uptake of other nutrients as well as lipid production, whereas sulfur is an essential element for protein biosynthesis (Wang & Lan, 2011b). Sun et al. (2014) indicated that

52 the optimum iron concentration for the cultivation of N. oleoabundans was 0.037 mM.

On the other hand, Wang and Lan (2011b) reported that there was no significant effect of the micronutrients on the growth of this microalga.

2.6 Biomass, Lipids, and Fatty Acids

2.6.1 Biomass

The microalga N. oleoabundans has gained significant interest due to its ability to accumulate high lipid content besides its remarkable biomass productivities. Table 2.1 summarizes some of the highest biomass concentrations, productivities, areal productivities, and growth rates reported in the literature along with the main culture parameters for each study. It is noteworthy to mention that the biomass concentrations and productivities reported in Table 2.1 represent the maximum found for each study.

However, this does not necessarily mean that the highest biomass concentrations and productivities or growth rates were achieved under the same cultivation conditions. For instance, the maximum biomass concentration reported by Yang et al. (2013) was achieved at 10 °C and 70 μmol/m2/s, whereas the highest growth rate and biomass productivity were achieved at higher temperature and light intensity.

It is clear that phototrophic is the main cultivation mode in the vast majority of the studies conducted on N. oleoabundans. However, when the microalga N. oleoabundans was cultivated heterotrophically or mixtotrophically, significant high biomass concentrations were reported such as the study conducted by Morales-Sánchez et al. (2013); who reported a maximum biomass concentration of 14.2 g/L when the microalga N. oleoabundans was cultivated in a fed batch culture with nitrate addition

53 under strict heterotrophic conditions, and using glucose as a carbon source. Furthermore,

Giovanardi et al. (2014) stated that the biomass productivity of this microalga was 0.202 g/L/d under mixotrophic conditions using glucose as a carbon source.

Despite the relatively high biomass concentrations in the literature for this microalga, most of the studies were conducted on a bench-scale level using flasks or reactors less than 1 L in volume. Very few studies discussed the outdoor cultivation either in closed or open photobioreactors. For example, da Silva et al. (2009) was one of the few researchers who reported the outdoor cultivation of N. oleoabundans in an open raceway pond during summer time; thus, solar energy was the light source and there was no control over the culture’s temperature. The highest biomass concentration achieved in their study was 2.8 g/L with a maximum specific growth rate of 0.18 d-1. Hence, it is necessary to conduct more studies on the cultivation of this microalga in outdoor cultures and on a larger scale in order to evaluate the suitability of this species for industrial applications.

2.6.2 Lipids

Neutral lipid content is one of the important parameters in microalgae cultivation; as this indicates the potential for biodiesel production. As shown in Table 2.1, Yang et al.

(2013) found that the highest FAME yield was 11.52 mg/L, which was observed at 10 °C

/70 μmol/m2/s. On the other hand, some studies reported low lipid content of this microalga; however, when coupled with the high biomass concentration, lipids yield was promising. For instance, da Silva et al. (2009) estimated that the oil yield can reach

21,523 L/ha when this microalga is cultivated in an outdoor raceway pond which is

54 higher than the yield of other energy crops even though the lipid content in their study did not exceed 11% per DCW. Giovanardi et al. (2014) reported that the highest cell lipid content per DCW was 46% when the microalga N. oleoabundans was cultivated in a mixotrophic growth mode with 2.5 g/L glucose. Morales-Sánchez et al. (2013) revealed that lipids productivity of the microalga N. oleoabundans was 528 mg/L.d under high

C/N ratio of 278 which was accompanied with a high lipid content of 51.7% per DCW.

Furthermore, Wang and Lan (2011b) optimized the composition of the modified Bristol

Medium to increase lipids yield and productivity from 143 mg/L and 55 mg/L/d to 635 mg/L and 176 mg/L/d, respectively. Several studies discussed lipids accumulation in N. oleoabundans in saline media cultures. Popovich et al. (2012) stated that the total lipid content was 15% (per DCW) in a seawater medium without nutrient limitation, 64% of which was neutral lipids. Under nitrogen limited conditions, the total lipid content increased to 27%, 78% of which was neutral lipids with a 56.4 mg/L/d productivity.

Santos et al. (2013) indicated that the total fatty acids content increased from 18 to 29% by weight when the pH was increased from 8.2 to 10 in saline media with a maximum productivity of 112.4 mg/L.d.

2.6.3 Fatty acids profile

The majority of fatty acids produced by the microalga N. oleoabundans are palmitic acids (C16:0) oleic acids (C18:1), linoleic acid (C18:2), stearic acid (C18:0), and alpha-linolenic acid (C18:3). The less common fatty acids are C14:0, C15:0, C16:2, and

C16:3 (Gatenby et al., 2003; Morales-Sánchez et al., 2013; Popovich et al., 2012;

Tornabene et al., 1983; Yang et al., 2011; Yang et al., 2013). According to Sun et al.,

(2014), C16/C18 is approximately 90%, which makes this microalga a suitable biodiesel feedstock. Yang et al. (2011) reported that there was no significant difference in the fatty acids profile when the microalga N. oleoabundans was cultivated using synthetic culture media or AD liquid waste. Temperature affects the fatty acids profiles, as the degree of unsaturation is inversely related to the temperature, particularly for C16:3 and C18:3. At extreme temperatures such as 10 and 35 °C, the yield of the shorter chain fatty acids

C12:0 and C14:0 approaches the palmitic acids yield, unlike temperatures in the 15 – 25

°C range (Yang et al., 2013).

Table 2.1: Biomass and Lipids Concentrations and Productivities for the Microalga N. oleoabundans Growth T (°C) Light Biomass Bioreactor Nutrient medium Lipids Reference mode / pH intensity X PX PX,A μm Sunlight in Bold (1949) Serpentine the day and medium with added 0.12 12 Muller-Feuga P glass (3-in 26 / - 140 Ca and Mg to 2.4 g/L - - g/L/d g/m2/d et al. (2012) diameter) μmol/m2/s in account for reverse the night osmosis 70 10 / Modified Bold's 0.039 11.5 mg/L FAME Yang et al. P Shake flasks μmol/m2/s 2.2 g/L - 0.34 d-1 6.4 Basal Medium g/L/d concentration (2013) (continuous) 1280 Artificial Cylinder 30 / Lumens- 0.35 Wang and P wastewater with 3.15 g/L - - - flasks 6.8 fluorescent g/L/d Lan (2011a) 1.33 N/P ratio (continuous) 80 Erlenmeyer μmol/m2/s Brackish medium 79 × 106 0.202 46% lipid content Giovanardi et M 24 / - - - flasks (16:8 with 2.5 g/L glucose cells/mL g/L/d (per DCW) al. (2014) light/dark) Bold's Basal 5-L reactor 52% lipid content Morales- Medium with 50 g/L 1.42 H (Biolafitte, 25 / 7 Dark 14.2 g/L - 1.2 d-1 (per DCW), 0.53 Sánchez et al. glucose (278 C/N g/L/d USA) g/L/d productivity (2013) ratio) Modified soil Glass 360 extract with sodium 0.63 0.133 g/L/d Li et al. P columniform 30 / - μmol/m2/s 3.2 g/L - - nitrate as nitrogen g/L/d productivity (2008) flasks (continuous) source 150 Glass bubble 0.15 56% lipid content Gouveia et al. P 31 / - μmol/m2/s Bristol medium - - 0.5 d-1 column g/L/d (per DCW) (2009) (continuous) 150 29% lipid content Gouveia and Polyethylene 0.09 P - / - μmol/m2/s - 2 g/L - - (per ash free dry Oliveira bags g/L/d (continuous) weight) (2009)

Table 2.1: continued 37% lipid content Flat panel 270 25 / Modified Bold's 0.55 16.5 (per DCW), 3.8 Pruvost et al. P airlift μmol/m2/s 6 g/L - 7.5 Basal Medium g/L/d g/m2/d g/m2/d total lipids (2009) bioreactor (continuous) productivity 168 26% lipid content Seawater enriched 25 - 26 μmol/m2/s (per DCW), 0.056 Popovich et al. P Flasks with soil extract and 1.5 g/L - - 0.73 d-1 / 8 (16:8 g/L/d total lipids (2012) salts light/dark) productivity Outdoor 9.2% total lipid da Silva et al. P raceway pond - / - - Bristol medium 2.8 g/L - - 0.18 d-1 content (per DCW) (2009) (Summer) 25% lipid content 6000 lux- Cylinder Modified Bristol 0.71 (per DCW), 0.176 Wang and Lan P 28 / - fluorescent 2.54 g/L - - flasks medium g/L/d g/L/d total lipids (2011b) (continuous) productivity 400 30% lipid content 24 / < μmol/m2/s Modified Bold's 0.47 (per DCW), 0.154 Urreta et al. P Glass reactors 5.17 g/L - - 8 (16:8 Basal Medium g/L/d g/L/d TAG (2014) light/dark) productivity 147 μmol/m2/s in Open trough 25.6 / 0.109 Murray et al. P the day and BG-11 1.14 g/L - - - system 7 - 7.5 g/L/d (2011) ≤ 10 in the night 60 Erlenmeyer 25 / Anaerobic digestion 2.8 - 2.9 10 - 13% total lipid Yang et al. P μmol/m2/s - - - flasks 6.5 effluent supernatant g/L content (per DCW) (2011) (continuous) P: Phototrophic; H: Heterotrophic; M: Mixotrophic; X: Biomass concentration; PX: Biomass productivity; PX,A: Areal biomass productivity; μm: Maximum specific growth rate.

2.7 Other Cultivation Conditions

2.7.1 Heterotrophic cultivation

Heterotrophic cultivation of microalgae has advantages over phototrophic cultivation such as improving biomass, protein, and lipids productivities as well as simpler operation and maintenance. Heterotrophic microalgae utilize organic carbon and oxygen in the absence of light, and generate carbon dioxide unlike phototrophic microalgae which consume carbon dioxide to produce oxygen (Morales-Sánchez et al.,

2013). Morales-Sánchez et al. (2014) investigated the growth of the microalga N. oleoabundans heterotrophically using glucose as a carbon source and under nitrate limited conditions. The highest lipid content found was 53.8% per DCW. However, when they compared the production of bioethanol via the fermentation of sugars to the heterotrophic pathway of converting sugars into algal biomass, it was found that the former is more feasible. Morales-Sánchez et al. (2013) indicated that the microalga N. oleoabundans can utilize glucose and cellobiose, with glucose being the favorite source for higher growth rates. Nevertheless, this microalga could not utilize xylose, arabinose, sucrose, fructose, lactose, glycerol, or acetate as organic carbon sources when cultivated under strict heterotrophic conditions.

2.7.2 Mixotrophic cultivation

Mixotrophy is a metabolic way in which microorganisms utilize both light and organic carbon (Baldisserotto et al., 2014; Giovanardi et al., 2013). Mixotrophic growth can also provide a cheap way of managing wastes from agro-food industries (Giovanardi et al., 2013). According to Giovanardi et al. (2014), 2.5 g/L glucose was optimum for the

59 growth and lipids accumulation of N. oleoabundans in a brackish medium under mixotrophic conditions. The highest biomass productivity encountered under this mixotrophic cultivation was 202 mg/L/d, which is higher than most reported values in the literature. Additionally, the cell lipid content per DCW was 46%. Baldisserotto et al.

(2014) compared the phototrophic and mixotrophic growth of N. oleoabundans using a brackish medium (17% salinity) and by adding apple to vinegar waste products as organic carbon supplement to the medium. They found that mixotrophic cultivation resulted in four times higher cell densities compared to phototrophic cultivation under the same conditions. Giovanardi et al. (2013) concluded that adding apple vinegar production waste not only provided the microalga N. oleoabundans with an organic carbon source, but it also indirectly activated the metabolism for nitrate uptake. On the contrary, Wu et al. (2011) reported that this microalga grew much better under phototrophic conditions compared to heterotrophic and mixotrophic conditions using glucose or lactose as a carbon source.

2.8 Alternatives for the Cultivation of N. oleoabundans

2.8.1 Wastewater as a nutrient medium

Microalgae cultivation using wastewaters as nutrient media is one of the most economically attractive options to produce microalgal biomass. Microalgae can function as a sink for the removal of organic and inorganic nutrients as well as heavy metals from a variety of wastewaters such as the secondary wastewater effluent; hence, reducing the high cost associated with synthetic nutrient media in addition to exploiting non-fresh water and reducing the disposal costs of these wastewaters (Baldisserotto et al., 2014;

Brennan & Owende, 2010; Kim et al., 2014; Wang & Lan, 2011a). Wang and Lan

(2011a) indicated that the maximum cell densities of the microalga N. oleoabundans were 0.68 and 2 g/L when cultivated using secondary municipal wastewater effluent

- without N enrichment and with 70 N-NO3 /L enrichment, respectively.

AD is a biological process that is used to treat many categories of organic wastes such as animal and industrial wastes (Olguín et al., 2015a). This process decreases the organic loading, while nutrients are not reduced; instead, some nutrients become more bioavailable such as nitrogen in the form of ammonium, so a further treatment of the effluent is required (Franchino et al., 2013; Levine et al., 2011; Olguín et al., 2015a).

Many studies reported the use of AD liquid waste as a nutrient medium for the growth of the microalga N. oleoabundans. However, there are some challenges related to adopting this approach in microalgal cultivation such as nutrients unbalanced concentrations, high turbidity, and other competing microorganisms (Levine et al., 2011).

Many studies indicated the use of sterilized AD liquid waste for microalgal cultivation. Franchino et al. (2013) reported that by autoclaving this liquid waste, ammonium concentration decreased by 60%, nitrate concentration increased by 22%

(although nitrate was significantly lower than ammonium), while pH remained unchanged. Levine et al. (2011) found that there was no significant difference in the biomass productivity and lipid content of N. oleoabundans when cultivated using raw or autoclaved diluted AD liquid waste. The highest growth rate was observed with 50-fold diluted liquid waste compared to 100- and 200-fold dilutions. Yang et al. (2011) reported that the highest biomass concentration was 2.9 g/L using 2% diluted effluent from the

AD of soybean waste. Further, the effluent from the AD of rice hull and catfish waste was a suitable nutrient medium for this microalga. It was found that combining the three waste categories did not improve the growth; however, the microalgae oil content increased six times. Franchino et al. (2013) revealed that 4 – 10% diluted cattle slurry and raw cheese whey AD supernatant supported the growth of N. oleoabundans with no signs of inhibition. Olguín et al. (2015b) stated that the microalga N. oleoabundans was cultivated successfully using the effluent of the anaerobically digested vinasse, which is a strong wastewater that contains high concentrations of organic and toxic compounds such as phenolic compounds.

2.8.2 Cultivation of N. oleoabundans using flue gas

Besides utilizing a readily available nutrient medium, utilizing the carbon dioxide portion of the flue gas can help offset the high production cost of microalgae biofuels.

Yoon et al. (2015) reported that the microalga N. oleoabundans was cultivated phototrophically using solar energy as well as combusted liquefied natural gas comprising 22 ppm NOx, 1.4 ppm CO, 5% CO2, and 12% O2. It was concluded that adaptation was necessary for the successful cultivation of N. oleoabundans using flue gas. Factors such as the concentration of the algal inoculum as well as the composition of the nutrient medium may limit the inhibitive effects of NOx on the metabolic activity of microalgae.

2.9 Conclusions and Recommendations

In this review, several aspects related to the cultivation of the promising microalga Neochloris oleoabundans were covered. This species has received significant

62 attention recently due to its high biomass concentration, lipid content, and biomass and lipid productivities. Furthermore, this species can be cultivated under phototrophic, heterotrophic, or mixotrophic conditions, which makes it an attractive candidate for biofuels synthesis as well as managing organic wastes from agro-food industries. This species can also be cultivated using a variety of wastewaters such as the secondary effluent and anaerobic digestion liquid waste; hence, the production cost can be considerably reduced. The highest biomass concentration reported in the literature for this species was 14.2 g/L which was achieved under heterotrophic conditions. Under phototrophic conditions, the highest biomass concentration achieved was 6 g/L. Several studies reported the total and neutral lipid content of this microalga. It has been indicated that under nitrogen limited conditions, the total lipid content per dry cell weight can exceed 50%, with triacylglycerol making up the majority of the total lipid content.

However, the vast majority of the studies conducted on this microalga were on a bench- scale level. Very few studies reported large scale cultivation such as outdoor open raceway ponds. Therefore, it is necessary to conduct more studies in the future on the large-scale outdoor cultivation of this microalga in order to assess the potential of this species as a candidate for the industrial biodiesel production and other applications.

CHAPTER 3. ANAEROBIC DIGESTATE AS A NUTRIENT MEDIUM FOR THE

GROWTH OF THE GREEN MICROALGA NEOCHLORIS

OLEOABUNDANS2

3.1 Abstract

In this study, the microalga Neochloris oleoabundans was cultivated in a sustainable manner using diluted anaerobic digestate to produce biomass as a potential biofuel feedstock. Prior to microalgae cultivation, the anaerobic digestate was characterized and several pretreatment methods including hydrogen peroxide treatment, filtration, and supernatant extraction were investigated and their impact on the removal of suspended solids as well as other organic and inorganic matter was evaluated. It was found that the supernatant extraction was the most convenient pretreatment method and was used afterwards to prepare the nutrient media for microalgae cultivation. A bench- scale experiment was conducted using multiple dilutions of the supernatant and filtered anaerobic digestate in 16 mm round glass vials. The results indicated that the highest growth of the microalga N. oleoabundans was achieved with a total nitrogen concentration of 100 mg N/L in the 2.29% diluted supernatant in comparison to the filtered digestate and other dilutions.

Keywords: Neochloris oleoabundans, microalgae, anaerobic digestion, biofuels.

2 This chapter was published in the Environmental Engineering Research Journal (Abu Hajar et al., 2016).

3.2 Introduction

The need for unconventional fuel feedstocks such as biofuels emerges due to the environmental consequences of utilizing conventional fossil fuels such as the gaseous emissions (Gouveia et al., 2009; Gouveia & Oliveira, 2009; Levine et al., 2011).

Biodiesel is a biofuel that is typically produced from oleaginous crops via the transesterification of their oils with methanol or ethanol to produce fatty acid methyl esters (FAME). These crops include rapeseed, soybean, sunflower, and palm (Gouveia &

Oliveira, 2009; Levine et al., 2011; Pruvost et al., 2009).

Microalgae are eukaryotic aquatic photosynthetic microorganisms capable of harvesting the solar energy effectively to produce biomass that is ideal for the production of biodiesel (Gouveia & Oliveira, 2009; Gour et al., 2014; Li et al., 2008; Meti & Sailaja,

2014). Microalgae, similar to plants, convert solar energy to chemical energy via photosynthesis, producing organic biomass from carbon dioxide and water.

Approximately, 1.83 kg carbon dioxide is fixed for each 1 kg biomass produced (James

& Boriah, 2010). There are many advantages to producing biofuels from microalgae such as the high areal yield, along with avoiding competition for fertile soil; since microalgae can grow on non-arable lands, using non-potable water, in a continuous rather than seasonal mode with a significantly lower water consumption rate and in an aqueous suspension system that provides more access to nutrients and water (Gouveia et al., 2009;

Gouveia & Oliveira, 2009; Levine et al., 2011; Pruvost et al., 2009; James & Boriah,

2010; Terry & Raymond, 1985). Additionally, biodiesel derived from microalgal lipids is sustainable and renewable in nature with lower harmful emissions such as CO,

65 hydrocarbons, and particulate matter, and no SOx emissions, besides the potential of utilizing the carbon dioxide portion of the flue gas from power plants as a carbon source for the growth of microalgae (Gouveia et al., 2009; James & Boriah, 2010; Li et al.,

2008).

Neochloris oleoabundans is a freshwater and saline media microalga from the

Chlorophyceae class and the Chlorococcaceae family (Gouveia et al., 2009; Santos et al.,

2013). Li et al. (2008) pointed out that this microalga is promising; as the lipid productivity of this species was nearly twice that of any other microalga studied for the purpose of biodiesel production. Moreover, the majority of the fatty acids produced by this microalga are saturated fatty acids in the range of 16 – 20 carbons, which makes this microalga ideal for biodiesel synthesis, even though most of the previous studies reported utilizing this species as a feeding source for aquaculture species such as mussels (Li et al., 2008). The highest biomass concentrations reported in the literature for the phototrophic cultivation of this microalga were in the 2 – 5.17 g/L range L (da Silva et al., 2009; Giovanardi et al., 2014; Li et al., 2008; Urreta et al., 2014; Wang & Lan,

2011a; Yang et al., 2013).

A major challenge to the production of biodiesel from microalgae is the relatively high capital and operational costs compared to conventional fossil fuels due to cultivation requirements (Gour et al., 2014; Terry & Raymond, 1985). Nutrients availability, mainly nitrogen and phosphorus, is a key factor for the growth of microalgae, which adds to the production cost. Industrial, municipal, and agricultural wastewaters may provide the necessary nutrients for the growth, where microalgae remove nitrogen and phosphorus

66 from wastewater via direct uptake. Moreover, the oxygen produced by microalgae can be utilized by the aerobic bacteria for further reduction in the organic matter (Choi & Lee,

2012; Levine et al., 2011; Meti & Sailaja, 2014). Microalgae such as Chlamydomonas,

Botryococcus, Chlorella, Haematococcus, Spirulina, and Scenedesmus have been utilized for wastewater treatment (Choi & Lee, 2012). For instance, Tam and Wong (1990) reported that the microalga Chlorella pyrenoidosa grew well when cultivated using supernatant from the preliminary and primary sedimentation and the secondary effluent from an activated sludge process. In addition, , Choi and Lee (2012) stated that nitrogen and phosphorus removal efficiencies of 81 – 85% and 32 – 36%, respectively, were achieved when the microalga Chlorella vulgaris was cultivated with wastewater from the preliminary sedimentation of a sewage plant. However, growth inhibition caused by elevated concentrations of ammonia, urea, and volatile fatty acids may limit the use of microalgae as means of secondary wastewater treatment (Park et al., 2009; Tam & Wong,

1996).

Anaerobic digestion of animal manure is an often-used approach to reduce the biological oxygen demand in the waste. However, nutrients are not eliminated via this route; in fact they become more bioavailable in the forms of ammonium and phosphate.

Using the anaerobic digestate (AD) to cultivate microalgae has some challenges including the potential high and unbalanced concentrations of nutrients, turbidity, other competing microorganisms, as well as the potential toxicity cause by elevated COD and ammonia concentrations; therefore, diluting the AD may become necessary before microalgae inoculation (Levine et al., 2011; Meti & Sailaja, 2014). Furthermore,

67 pretreatment methods such as autoclaving, filtration, or other techniques are often applied to wastewaters in general and AD in particular prior to microalgae cultivation in order to reduce the suspended solids concentration as well as prevent interference from other microorganisms such as bacteria or protozoa (Choi & Lee, 2012; Park et al., 2010).

Several studies indicated the potential of growing the microalga N. oleoabundans using the anaerobic digestion effluent or digestate (Franchino et al., 2013; Levine et al.,

2011; Yang et al., 2011). However, the range of dilutions covered by each individual study was narrow. For instance, Levine et al. (2011) inoculated N. oleoabundans in 50-,

100-, and 200-fold diluted anaerobic digestion effluent under 200 μmol/m2/s light intensity. They concluded that 50-fold dilution which was equivalent to 60 mg N/L total nitrogen; 2.6 mg P/L phosphorus; and 42 mg N/L ammonia, yielded the highest growth rates. Yang et al. (2011) analyzed the effluent supernatant from the anaerobic digestion of rice hull, soybean, and catfish wastes. The ammonium concentrations for these waste categories were 258 – 293 mg/L, 743 – 787 mg/L, and 3,105 – 3,684 mg/L, respectively.

They found that the growth in the 50-fold diluted effluent yielded the highest biomass growth rate compared to other dilutions regardless of the waste category Franchino et al.

(2013) reported the cultivation of N. oleoabundans using 10-fold to 25-fold diluted cattle slurry and raw cheese whey anaerobic digestion effluent with an initial (undiluted) ammonium concentration of 1,634 mg N/L. They found that there was no significant difference in the growth among all dilutions. The aforementioned studies on the anaerobic digestion liquid waste as a nutrient medium focused on filtration, autoclaving, or centrifugation as necessary steps ahead of microalgae cultivation in order to eliminate

68 the suspended solids as well as the interference from other microorganisms. According to

Levine et al. (2011), there was no significant difference between raw and autoclaved diluted effluent on the microalgae biomass productivity. It was thought initially that the native bacteria within the effluent produce vitamin B12, which is necessary for the growth of microalgae. Additionally, the heterotrophic microorganisms may contribute to the growth of microalgae by remineralizing nutrients and producing carbon dioxide as a result of organic carbon oxidation (Levine et al., 2011). On the other hand, centrifugation is an energy intensive process, which may limit its applicability as a pretreatment step to the anaerobic digestate (Sturm & Lamer, 2011).

In this study, the AD from two sources was characterized in order to be used as a nutrient source for microalgae. Several pretreatment methods other than those discussed in the previous studies were explored including filtration with variable mesh sizes, hydrogen peroxide oxidation, and supernatant extraction. The purpose of these treatment methods was to reduce the suspended solids in the nutrient medium as well as eliminate the potential toxicity caused by elevated concentrations of organic and inorganic matter.

Further, the growth of the microalga N. oleoabundans using diluted AD was evaluated by a bench-scale experiment using 16 mm round glass vials. This experiment covered unfiltered and filtered digestate as well as a wider range of dilutions compared to the previous studies in literature.

3.3 Materials and Methods

3.3.1 Anaerobic digestate

AD was sampled from two sources in Columbus, OH. Source A is a commercial digester where animal manure and other organic wastes are digested anaerobically to produce biogas, while source B is an anaerobic digester processing waste activated sludge from a domestic wastewater treatment plant. Samples were brought to the laboratory and stored at 4 °C until the time of analysis.

3.3.2 Analytical methods

Measurements of total solids (TS), volatile solids (VS), total suspended solids

(TSS), volatile suspended solids (VSS) were performed according to the APHA Standard

Methods for the Examination of Water and Wastewater (methods 2540 B, D, and E)

(Clesceri et al., 1998). Chemical oxygen demand (COD), total nitrogen, total phosphorus, and ammonia nitrogen were determined according to the colorimetric methods in compliance with APHA Standard Methods for the Examination of Water and Wastewater and EPA methods (HACH methods 8000, 10072, 10127, and 10031) using HACH DR

3900 spectrophotometer. Cations such as iron, calcium, magnesium, manganese, and potassium were measured using Thermo Scientific iCAP 6300 ICP spectrometer.

Samples were analyzed in triplicates and expressed as mean ± standard deviation. Total and filtered concentrations as shown in Table 1 refer to well-mixed samples and filtered samples through 0.45 μm syringe filters, respectively. Zeta potential measurements were conducted on diluted digestate using Brookhaven ZetaPlus analyzer. Finally, microalgae

70 biomass concentration was quantified as the optical density at 750 nm (OD 750) using

HACH DR 3900 spectrophotometer.

3.3.3 Pretreatment methods

3.3.3.1 Filtration using polyester filter bags

The AD’s from sources A and B were diluted with deionized (DI) water to 1% and 2%, respectively. These dilutions were selected based on the differences between the two sources in terms of nutrients and solids content (Table 3.1). For instance, nitrogen concentration in source A was twice that of source B; therefore, the dilution was selected so as to yield total nitrogen concentration within the detection limits of the colorimetric method. The diluted digestate was then filtered using 10, 5, and 1 μm welded polyester filter bags (16" length and 7" diameter, Duda Energy) and the filtrate was sampled and analyzed for OD 750, TSS, COD, total nitrogen, and total phosphorus.

3.3.3.2 Hydrogen peroxide treatment

Hydrogen peroxide (H2O2) is a strong oxidant that can be used alone or combined with other oxidation techniques such as UV light and ozone or with a catalyst such as iron (Crittenden et al., 2005; Ksibi, 2006). The AD’s from sources A and B were diluted with DI water to 1% and 2%, respectively for the same reason mentioned in Section

2.3.1. The stoichiometric hydrogen peroxide dosage is 2 moles 30% H2O2 per 1 mole of

COD. The experiments were conducted using three dosages of 30% H2O2: 0.5, 1.0, and

1.5 times the calculated dosage based on COD. These dosages were 1.96, 3.92, and 7.84 mL 30% H2O2/L of source A 1% diluted digestate and 2.22, 4.44, and 8.89 mL 30%

H2O2/L of source B 2% diluted digestate. The solutions were kept in Erlenmeyer flasks

71 on a shaker plate at 200 RPM for 2 h. In addition, a combination of UV/H2O2 treatment was applied to both sources with 1.0 × calculated H2O2 dosage as an advanced oxidation process, which is expected to improve the oxidation efficiency due to the generation of hydroxyl radicals (Crittenden et al., 2005). After the hydrogen peroxide dosage was added, the solution was recirculated for 30 min through a Turbo-Twist 3x UV unit

(Coralife products) with a wavelength of 253.7 nm and an irradiation of 9,580 μW/cm2.

Samples were collected at the end of the mixing period and were analyzed for OD 750,

TSS, COD, total nitrogen, and ammonia.

3.3.3.3 Supernatant characterization

Three dilutions from each AD source were tested. These dilutions were most likely to be used for microalgae cultivation. 5%, 3.5%, and 2% dilutions of source A, and

10%, 7%, and 4% dilutions of source B were prepared and poured in eight 50 mL centrifuge vials per dilution; hence, the total number of vials was 48. These vials were left undisturbed to settle and exactly 30 mL supernatant was extracted from two vials per dilution every 45 min for a total sampling time of 3 h using a pipette and the remaining sludge in the bottom of each vial (15 mL) was disposed. The supernatant was analyzed for OD 750, TSS, COD, total nitrogen, and total phosphorus and the results were expressed as C/Co, where C is the concentration at time t and Co is the initial well-mixed concentration.

3.3.4 Microalgae selection

The microalga N. oleoabundans (UTEX 1185) was purchased from the algae culture collection at the University of Texas in Austin. The culture was maintained at 25

°C in Bristol medium (James, 1978). BG-11 medium (James, 1978) was used later as it provided faster growth rates.

3.3.5 Microalgae cultivation using diluted AD

A bench-scale experiment was conducted to evaluate the growth of N. oleoabundans using diluted AD by inoculating 4 mL of nutrient medium with 1 mL of microalgae culture that was previously cultivated phototrophically using BG-11 medium

(average OD 750 of the microalgae inoculum was 0.15) and in the same cultivation conditions listed below. The control nutrient medium was BG-11 medium; and therefore, the highest concentrated dilutions were prepared to match the nitrogen concentration in

BG-11 medium (250 mg N/L). All dilutions were conducted in five replicates in 16 mm round glass vials and placed inside an incubator at 25 °C with a 14:10 light:dark cycle using two built-in fluorescent lamps providing an average light intensity of 50 μmol/m2/s.

The biomass concentration was monitored frequently by placing each vial directly in the spectrophotometer and obtaining an OD 750 value.

3.3.6 Statistical analysis

Statistical analysis was performed using IBM SPSS software in order to compare the growth under different nutrient media.

3.4 Results and Discussion

3.4.1 AD characterization

Samples of AD from sources A and B were diluted with DI water and characterized as shown in Table 3.1. It is clear that source A is richer in most of the parameters than source B. For instance, total nitrogen in source A is twice that of source

B. Approximately 65% of the total nitrogen in source A exists in the form of ammonia, while ammonia is only 49% of the total nitrogen in source B. The difference in the nutrients concentrations between the two sources is reflected on the selection process.

Source A, for example, contains higher concentrations of nitrogen and phosphorus compared to source B; thus, it will require more dilution to achieve a target nitrogen or phosphorus concentrations as opposed to source B. Another difference is that source A contains higher ammonia concentration compared to source B; thus, if the two sources are diluted to the same nitrogen concentration, source A might not be favorable for some microalgae species due to the higher ammonia content, and the growth can be inhibited beyond certain threshold. Unfiltered phosphorus was also significantly higher in source

A, but the filtered phosphorus in source B was higher than that in source A. Moreover, the N/P ratio was 4 for the unfiltered source A, 4.6 for the unfiltered source B, 8.1 for the filtered source A, and 3.1 for the filtered source B. The variation of the N/P ratio may affect the biomass productivity of microalgae, but this is species-dependent. For instance,

Wang and Lan (2011a) studied the impact of N/P ratio on the growth of the microalga N. oleoabundans in artificial wastewater and under surplus phosphorus concentrations. The

- nitrogen concentration was varied by enriching the media with 45 – 218 N-NO3 /L sodium nitrate concentration, and the corresponding N/P ratios were in the 0.42 – 2.02

- range. The optimal cell growth was observed at 140 mg N-NO3 /L (N/P = 1.33). Under

- constant nitrogen concentration of 140 mg N-NO3 /L, and N/P ratio in the 3 – 26.4 range,

3- the highest biomass concentration was observed at 47 mg P-PO4 /L initial P concentration (N/P = 3). They also concluded that low N/P ratio is necessary for

74 complete nitrogen removal. For example, at 140 mg N/L, they found that the N/P ratio should be less than or equal to 3 for complete nitrogen removal from the artificial wastewater. In contrast, phosphorus removal was independent of N/P ratio (Wang and

Lan, 2011a). Finally, source A contained higher Fe, K, and Na concentrations compared to source B, whereas the concentrations of Ca and Mg were higher in source B.

Although filtration using 0.45 μm syringe filters reduces turbidity and removes suspended solids from the diluted digestate, it does not necessarily improve the media to support the growth of microalgae. For example, the elevated ammonia concentrations in filtered compared to unfiltered digestate may have negative impact on the growth of many microalgae species. Besides, 0.45 μm filtration would be expensive on a larger scale.

Table 3.1: AD Characterization (average ± SD; n = 3) Source Parameter Unit A B TS g/L 53.3 ± 1.7 37.3 ± 1.4 VS g/L 31.3 ± 1.0 23.3 ± 0.7 TSS g/L 34.9 ± 1.4 32.1 ± 2.0 VSS g/L 22.8 ± 0.8 21.0 ± 0.9 Total 55,300 ± 300 31,367 ± 202 COD mg/L Filtered 12,700 ± 141 3,567 ± 491 Total 5,567 ± 58 2,867 ± 29 Nitrogen mg N/L Filtered 3,700 ± 200 1,667 ± 202 Total 1,381 ± 11 626 ± 12 Phosphorus mg P/L Filtered 459 ± 4 545 ± 66 Total 3,593 ± 25 1,393 ± 28 Ammonia mg N/L Filtered 3,113 ± 21 1,377 ± 98 Ca mg/L 116 ± 1 220 ± 16 Fe mg/L 6.8 ± 0.1 2.4 ± 0.3 K mg/L 661 ± 13 185 ± 12 Mg mg/L 27.7 ± 0.3 45.6 ± 4.2 Na mg/L 2,916 ± 79 691 ± 38

3.4.2 Zeta potential

Zeta potential is an appropriate way to measure the electrophoretic mobility of the solids in a suspension, and is often used to determine the surface properties of sludge flocs (Forster, 1986; Pere et al., 1993). According to Liao et al. (2001), the hydrophobicity of sludge decreases with an increase in the surface charge. It was observed in our experiments that diluting the AD resulted in improved settleability of the suspended solids; therefore, zeta potential measurements were conducted on 0.1%, 0.2%,

1%, 2%, and 5% diluted digestate from both sources as shown in Figure 3.1. The higher concentration digestate (> 5% dilution) resulted in unstable readings from the instrument; thus, they were excluded from the analysis.

No clear correlations between the dilution% and zeta potential were found

(Figure 3.1). The values ranged from -35.71 to -25.02 mV for source A digestate, whereas the corresponding values for source B ranged from -25.95 to -23.17 mV. Su et al. (2014) investigated the impact of dilution on the zeta potential of aerobic granular sludge. They concluded that concentrations higher than 10 g TSS/L resulted in unstable zeta potential readings. Furthermore, they reported that there was no clear relationship between TSS concentration and zeta potential over the range of 0.1 – 8.0 g TSS/L. In our analysis, the range of TSS concentrations tested for zeta potential was 0.032 – 1.7 g/L.

Morgan et al. (1990) indicated that sludge solids carry a negative charge regardless of the sludge type. Moreover, they postulated that the amount of negative charge carried by the sludge solids was due to the extracellular polymers (ECPs) yield; as the activated sludge solids had higher negative charge and ECPs yield compared to the AD sludge. The lower

ECPs generated in the AD sludge can be related to the potential degradation of these biopolymers by bacteria to form methane and carbon dioxide. On the other hand, there was a strong correlation between the dilution% and conductance in both sources; which relates to the total dissolved solids concentration.

(Source A) (Source B) -40 5,000 -30 1,600

S) 4,000 -28 S) μ -35 1,200 μ 3,000 -26 -30 800 2,000 -24 -25 400 1,000 -22 Conductance ( Conductance Conductance ( Zeta Zeta Potential(mV) Zeta Zeta Potential(mV) -20 0 -20 0 0 1 2 3 4 5 0 1 2 3 4 5 Dilution % Dilution %

Figure ‎3.1: Zeta potential (♦) and conductance (●) values for the diluted AD (average ± SD; n = 3)

3.4.3 Filtration using polyester filter bags

Elevated concentrations of COD and organic carbon in particular may negatively impact the growth of microalgae due to rapid microbial growth (Roudsary et al., 2014).

For instance, Travieso et al. (2006) reported that COD concentrations as high as 1,100 mg COD/L inhibited the growth of Chlorella vulgaris when cultivated using piggery wastewater.

Filtration using polyester filter bags is a convenient and relatively low cost method for the removal of solids from liquid media such as wastewater. There was a significant reduction in all parameters when comparing the unfiltered diluted digestate with the filtered liquid (Table 3.2). OD 750, which reflects the turbidity of the liquid, decreased by 0.398 units and 0.514 units for the 10 μm filtrate for sources A and B diluted digestates, respectively. Further reductions were not significant with finer filtration using the 5 and 1 μm filter bags. For the cultivation of microalgae using the AD, the initial turbidity of the nutrient media can become a limiting factor due to light

78 attenuation in the highly concentrated solutions. TSS results follow the same pattern; as

89% and 85% of the TSS were retained on the 10 μm filter bags for the diluted sources A and B digestates, respectively. COD was reduced by 61% and 76% when the diluted sources A and B digestates were filtered through the 10 μm filter bags, respectively.

Filtration using smaller mesh sizes did not result in significant additional reductions in

COD. Total nitrogen was reduced by 27% and 39% when the diluted sources A and B digestates were filtered through the 10 μm filter bags, respectively. Finer filtration did not yield any additional reductions in nitrogen concentrations. The reduction in phosphorus was more apparent by filtering the diluted digestate through the 10 μm filter bags; as 55% and 68% of the total phosphorus was removed from sources A and B diluted digestates, respectively. Similar to total nitrogen, finer filtration did not yield any additional significant reduction in the total phosphorus content in both digestates.

The impact of filtration on the N/P ratio was higher in the case of source B digestate as the ratio increased from 4.6 to 8.5, whereas the initial ratio in source A digestate was 4 which increased to 6.5 after filtration as shown in Figure 3.2. In addition, the remainder of the nitrogen in the filtrate was 80 – 90% ammonia; which may not be the favorable nitrogen form for many microalgae species in comparison to nitrates and organic nitrogen. Overall, filtration using 10 μm filter bags is an attractive option to reduce the suspended solids and COD contents of the diluted AD. However, elevated ammonia concentrations have to be taken into consideration when preparing nutrient media for the growth of microalgae.

Table 3.2: Characterization‎of ‎the‎ Unfiltered‎and‎10,‎5,‎and‎1‎μm ‎Fil tered‎Dilut ed‎A D‎ (average‎±‎SD; ‎n‎ =‎3 )

Filter Mesh Size OD 750 TSS (mg/L) COD (mg/L) N (mg N/L) P (mg P/L) (μm) A 1% B 2% A 1% B 2% A 1% B 2% A 1% B 2% A 1% B 2%

Unfiltered 0.520 ± 0.011 0.675 ± 0.014 417 ± 8 530 ± 4 553 ± 5 627 ± 12 55.7 ± 0.9 57.3 ± 1.3 13.8 ± 0.4 12.5 ± 0.2

10 0.122 ± 0.005 0.161 ± 0.008 46 ± 2 78 ± 2 214 ± 9 149 ± 4 40.5 ± 0.3 35.0 ± 0.1 6.2 ± 0.1 4.1 ± 0.1

5 0.113 ± 0.002 0.157 ± 0.002 43 ± 1 75 ± 2 207 ± 3 146 ± 7 40.4 ± 0.7 34.5 ± 0.2 6.0 ± 0.2 4.0 ± 0.1

1 0.098 ± 0.001 0.146 ± 0.004 31 ± 1 59 ± 1 186 ± 3 142 ± 2 39.5 ± 0.5 34.5 ± 0.3 6.0 ± 0.1 4.0 ± 0.1 A 1%: 100-fold diluted AD source A B 2%: 50-fold diluted AD source B

N/P ratio 4

0 A 1% UnfilteredA 1% 10 μ FiltrateB 2% UnfilteredB 2% 10 μ Filtrate

Figure ‎3.2: N/P comparison between the unfiltered diluted AD and the 10 μm filtrate for sources A and B (average ± SD; n = 3)

3.4.4 Hydrogen peroxide treatment

The purpose of hydrogen peroxide treatment was to evaluate the potential of oxidizing the COD and ammonia in the diluted AD. Chemical oxidation is a process that has many applications in the water and wastewater treatment. Taste and odor control, disinfection, hydrogen sulfide removal, and color removal are amongst the applications of chemical oxidation in water treatment (Crittenden et al., 2005).

The theoretical dosage of 30% H2O2 was estimated based on the COD content.

Then 0.5, 1.0, and 1.5 times the theoretical dosages were applied for each source as indicated in the Materials and Methods section. As shown in Figure 3.3, TSS concentrations decreased as a result of hydrogen peroxide pretreatment. The maximum reduction observed in the case of source A was 29% with 1.5 × dosage, while the maximum reduction was 13% for source B with the same dosage. Similar pattern can be noticed with OD 750 as the initial OD 750 value decreased from 0.383 with no hydrogen peroxide to 0.293 with 1.5 × dosage for source A, whereas for source B, OD 750 decreased from an initial value of 0.556 to 0.484 when the 1.5 × dosage was applied.

There was a slight decrease in the COD (< 5%) when half the theoretical dosage of hydrogen peroxide was used for both sources. However, for higher dosages (1 and 1.5 × dosage), the COD values were higher than the initial, which suggests that the residual hydrogen peroxide interfered with the spectrophotometric COD measurement.

Similar trends were observed when the digestate samples were treated with a combination of UV and hydrogen peroxide. This may indicate that hydrogen peroxide is not an effective oxidant for the COD in the diluted AD. The same interference was

81 observed with the total nitrogen measurement, particularly for source B (Figure 3.4). On the other hand, ammonia concentrations decreased with an increase in the hydrogen peroxide dosage. The maximum reduction was 22% for source A with 1.0 dosage + UV whereas 1.0 × dosage without UV resulted in 15% reduction in ammonia. The maximum reduction of ammonia for source B was 16% when the 1.5 × dosage of hydrogen peroxide was added to the solution. It is not clear however, whether the ammonia that is lost has been converted to nitrogen gas or to nitrate, as the total nitrogen tests were inconclusive due to the interference of hydrogen peroxide residual.

In summary, hydrogen peroxide pretreatment was not an effective way to decrease the COD of the diluted AD or to oxidize the organic nitrogen and ammonia.

This can be related to the high alkalinity in the AD which decreases the efficiency of advanced oxidation processes such as ozone or peroxide; as bicarbonate is a radical scavenger (Park et al., 2010).

TSS Source A TSS Source B COD Source A COD Source B 600 1,000

500 800

400 600 300 400 200 TSS (mg/L) TSS COD (mg/L) COD 100 200 0 0 0 0.5 1.0 1.5 1.0 + UV

H2O2 Dosage ( × Theoretical Required Dosage)

Figure ‎3.3: TSS and COD concentrations for various hydrogen peroxide dosages (average ± SD; n = 3)

Total N Source A Total N Source B Ammonia Source A Ammonia Source B 80 70 60 50 40 30 20 10

Concentration (mg N/L) Concentration 0 0 0.5 1.0 1.5 1.0 + UV

H2O2 Dosage ( × Theoretical Required Dosage)

Figure ‎3.4: Nitrogen and ammonia concentrations for various hydrogen peroxide dosages (average ± SD; n = 3)

3.4.5 Supernatant characterization

This experiment was conducted based on a previous observation that solids tend to settle over time in the diluted AD, unlike the raw digestate in which solids remain in suspension. It is thought that the dilution process helps in improving the settleability of the digestate. Pere et al. (1993) indicated that the heavily loaded sludges tend to have higher zeta potential; i.e., they were more hydrophilic, which could be due to the increased content of ECPs, which affects the viscosity as well as the settling characteristics due to bioflocculation (Morgan et al., 1990; Pere et al., 1993).

The optical density values presented in Figure 3.5 indicate that the more concentrated dilutions exhibited the highest reductions of OD 750 as the overall OD 750 reductions were 1.91, 1.29, and 0.76 for the 5%, 3.5%, and 2% source A dilutions, respectively and 2.25, 1.66, and 1.00 for the 10%, 7%, and 4% source B dilutions, respectively. This is equivalent to 83 – 87% reduction in TSS concentration for both sources; as the final TSS concentrations after 3 h were 320, 191, and 81 mg/L for the 5%,

3.5%, and 2% source A dilutions, respectively and 352, 235, and 120 mg/L for the 10%,

7%, and 4% source B dilutions, respectively. However, the high OD 750 values observed in the supernatant after 3 h of settling may not be suitable for the growth of microalgae due to the higher turbidity and thus the lower light penetration, suggesting additional dilution might be required.

( Source A) (Source B) TSS 5.0% TSS 3.5% TSS 10% TSS 7% TSS 2.0% OD 750-5.0% TSS 4% OD 750-10% OD 750-3.5% OD 750-2.0% OD 750-7% OD 750-4% 3.5 2,000 3 2,400 3.0 1,600 2,000

2.5

2 1,600 1,200 2.0 750 1,200 1.5 OD OD 800 750OD TSS (mg/L) TSS TSS (mg/L) TSS 1 800 1.0 400 400 0.5 0 0 0 0.0 0 1 2 3 0 1 2 3 Time (h) Time (h) Figure ‎3.5: TSS and OD 750 values in the supernatant as a function of time (average ± SD; n = 4)

Figure 3.6 and Figure 3.7 show the COD, nitrogen, and phosphorus concentrations in the supernatant as a function of time. As indicated in the Materials and

Methods section, supernatant samples were extracted every 45 min and analyzed, and the

C/Co ratios were plotted for each parameter over time; where Co is the initial concentration for the well-mixed sample, while C is the measured concentration at time t.

It is clear from Figure 6 that the major reduction in COD occurred within the first hour for both sources. Regardless of dilution, COD decreased by approximately 55 – 60% for

84 source A and 65 – 70% for source B in the first hour. The overall COD reductions after 3 h of monitoring were 64 – 82% for source A and 74 – 78% for source B. Approximately

30% of total nitrogen was lost in the first hour in source A due to solids settling whereas the reduction in source B was in the range of 27 – 52% (Figure 3.7a and Figure 3.7b).

The reduction in phosphorus was more apparent than nitrogen as approximately 70 –

80% of source A’s total phosphorus was lost in the first hour compared to 70 – 73% for source B (Figure 3.7c and Figure 3.7d). As a result, the N/P ratio varied over time as shown in Figure 3.8.

In summary, the supernatant extraction method appears to be the most attractive and least expensive pretreatment method. Compared to hydrogen peroxide treatment, supernatant extraction was more effective in decreasing the TSS and COD contents of the digestate. On the other hand, the removal efficiencies of TSS and COD of the supernatant extraction method were comparable to those achieved with polyester filter bags; however, the cost of filters decreases the feasibility of filtration in comparison to supernatant extraction. Additionally, the supernatant experiments can be useful in evaluating the settling per unit depth and/or area over time. For example, if we consider the 5% dilution of source A, the COD removal in the first 45 min was 491 mg/L, from which the COD flux for the control volume of 30 mL and over the depth of 5.5 cm is equivalent to 3.57 mg/cm/h. Therefore, if the same removal efficiency is desirable and the initial COD content is known, the required retention time can be estimated based on the reactor depth assuming a discrete settling where solids do not tend to flocculate and particles settle without interaction (Carlsson, 1998).

In general, after diluting the two sources, both behaved similarly in terms of nutrients as well as the suspended solids removal. However, nitrogen decreased to lower levels in source B compared to source A. This means that in order to match a certain nitrogen concentration, the supernatant extracted from source B has to be diluted even less than source A. As a result, the turbidity of the target dilution will be higher, which will affect the phototrophic growth of microalgae due to light attenuation. Moreover, source B dilutions appeared to have separation in the form of foamy layer on the surface, which interfered with the process of supernatant extraction and in fact contributed to more solids in the decanted liquid. Accordingly, source A was selected as a nutrient source for the growth of the microalga N. oleoabundans.

Source A Source B 5.0% 3.5% 2.0% 10% 7% 4% 1.0 1.0 0.8 0.8

0.6 0.6 C/Co 0.4 C/Co 0.4 0.2 0.2 0.0 0.0 0 1 2 3 0 1 2 3 Time (h) Time (h)

Figure ‎3.6: COD concentration in the supernatant as a function of settling time with respect to the initial COD concentration at time 0 (average ± SD; n = 4)

(a) (b) 5.0% 3.5% 2.0% 10% 7% 4% 1.0 1.0 0.8 0.8

0.6 0.6 C/Co C/Co 0.4 0.4 0.2 0.2 0.0 0.0 0 1 2 3 0 1 2 3 Time (h) Time (h)

Figure ‎3.7: Nutrients concentrations in the supernatant as a function of settling time with respect to the initial concentrations at time 0 (a) total N source A, (b) total N source B, (c) total P source A, and (d) total P source B (average ± SD; n = 4)

Source A Source B 5.0% 3.5% 2.0% 10% 7% 4% 12 14 10 12 8 10

6 8 N/P N/P 6 4 4 2 2 0 0 0 1 2 3 0 1 2 3 Time (h) Time (h)

Figure ‎3.8: N/P ratio in the supernatant as a function of settling time (average ± SD; n = 4)

3.4.6 Microalgae cultivation

The microalga N. oleoabundans was cultivated using diluted AD as stated earlier.

Source A AD was used, however, it was a different batch from the earlier one characterized in the previous sections. For instance, the total nitrogen in the well-mixed digestate was 4,775 mg N/L compared to the initial batch value of 5,567 mg N/L. Total phosphorus in this batch was 1,507 mg P/L compared to 1,381 mg P/L in the initial batch.

Similarly, 0.45 μm syringe filtered nitrogen and phosphorus concentrations were 2,525 mg P/L and 483 mg P/L, respectively. Finally, the supernatant, which was extracted after

3 h of gravity settling, had undiluted nitrogen, phosphorus, and ammonia concentrations of 4,460 mg N/L, 538 mg P/L, and 2,490 mg N/L respectively. These concentrations were actually close to those obtained in the supernatant extraction experiment discussed earlier, where the majority of nitrogen remained in suspension regardless of the dilution.

For instance, the average undiluted nitrogen concentration for the last 2 h of the supernatant extraction experiment was 4,633 mg N/L for the 5% dilution, 4,386 mg N/L

88 for the 3.5% dilution, and 4,229 mg N/L for the 2% dilution. The corresponding undiluted phosphorus concentrations averaged 517 mg P/L for the 5% dilution, 510 mg

P/L for the 3.5% dilution, and 511 mg P/L for the 2% dilution. As a result, the two batches resembled each other considerably, but it is always important to characterize the digestate prior to microalgae cultivation; as the anaerobic digestion process and its digestate can vary significantly depending on the feedstock and the operational conditions.

A range of 5 – 250 mg N/L was targeted by extracting the supernatant of the

5.71% diluted digestate which yielded a nitrogen concentration of 250 mg N/L and sequentially diluting this supernatant as shown in Table 3. The 0.45 μm filtered digestate was diluted as well to a range of 0.2 – 10% (Table 3.3).

Table 3.3: Target Nitrogen and Phosphorus Concentrations in the Diluted AD Supernatant and Filtrate (average ± SD; n = 3) Target Nitrogen Supernatant Filtered Concentration (mg N/L) Dilution % Total P (mg P/L) Dilution % Total P (mg P/L) 250 ± 13.2 5.71 30.7 ± 0.69 10 48.3 ± 0.52 200 ± 8.2 4.57 24.6 ± 0.54 8 38.6 ± 0.42 100 ± 4.4 2.29 12.3 ± 0.27 4 19.3 ± 0.21 50 ± 2.2 1.14 6.1 ± 0.14 2 9.7 ± 0.11 25 ± 1.1 0.57 3.1 ± 0.07 1 4.8 ± 0.05 10 ± 0.4 0.23 1.2 ± 0.03 0.4 1.9 ± 0.02 5 ± 0.2 0.11 0.6 ± 0.01 0.2 1.0 ± 0.01

As shown in Figure 3.9, the nutrient media contributed to the initial turbidity of some cultures, mainly the 5.71%, 4.57%, and 2.29% supernatant, as the OD 750

89 measurements of these dilutions (without microalgae inoculation) were 0.780, 0.619, and

0.285, respectively. Shortly after the cultures were inoculated, OD 750 of these three dilutions in addition to the 1.14% dilution cultures decreased, but after 4 – 6 days, OD

750 started to increase again. This may indicate a lag phase of the microalgal growth concurrent with direct consumption of the particulate substrate or bacterial production of extracellular enzymes to hydrolyze and solubilize the particulates. In order to verify this, additional vials were inoculated with 4 mL of the supernatant dilutions and 1 mL deionized water. The OD 750 of the new vials was monitored in order to assess if the particulate matter dissolution was a result of direct consumption by microalgae or another reason such as bacterial activity. The results indicated that after 5 days, the OD 750 of the

5.71%, 4.57%, 2.29%, and 1.14% decreased by 0.072, 0.045, 0.017, and 0008, respectively (Figure 3.11). As a result, it is likely that both native bacteria and microalgae contribute to the decrease or dissolution of the particulate matter.

The growth in the lower dilutions resembled that in the BG-11 culture in terms of continuous increase in optical density, although the growth provided by BG-11 was significantly higher than the 0.11 – 0.57% dilutions (P < 0.05). After 16 days of monitoring, the OD 750 values for the 2.29 – 5.71% dilutions were not significantly different (P < 0.05) and the three curves were approaching the same point, regardless of the significant differences in the initial OD 750 readings amongst these three dilutions.

However, after 36 days of monitoring, the 2.29% dilution had the highest OD 750, which was significantly higher than any other dilution (supernatant or filtered) except the 4.57% supernatant. Considering the lower initial OD 750 reading, the 2.29% dilution was

90 selected in order to maximize the N. oleoabundans biomass concentration, despite the insignificant differences between the 2.29% and 4.57% curves by the end of the monitoring period.

For the filtered digestate cultures, the 4% dilution achieved the highest biomass concentration expressed as OD 750 as of day 16 followed by BG-11, 2%, and 1% dilutions, respectively (Figure 3.11). In summary, it appears that a nitrogen concentration of 100 mg N/L (7.3 mM) provided the optimum growth for both categories (supernatant and filtered). However, the higher optical density observed in the supernatant dilution may be due to the initial turbidity of the supernatant media compared to the filtered media. Nitrogen form may have contributed to the differences among the two categories; since nitrogen exists entirely as ammonia in the filtered digestate, whereas both ammonia and organic nitrogen exist in the supernatant. Li et al. (2008) indicated that nitrate is the favorable nitrogen form to the microalga N. oleoabundans followed by urea and ammonium. Moreover, they compared 3, 5, 10, 15, and 20 mM nitrate concentrations and concluded that the optimum nitrate concentration for cell growth was 10 mM while 5 mM was optimum for high lipid production. They also suggested that 15 mM or higher may inhibit cell growth. Even though our testing configuration differs from that used by Li et al. (2008), the suggested optimum dilution is well within the range of 5 – 10 mM for optimum biomass and lipid productivities. Levine et al. (2011) obtained similar results when they observed a significantly better growth with nitrate compared to ammonium. In fact, they indicated that high ammonium concentrations may cause growth inhibition, but the toxic effects could be reduced by following a fed-batch or continuous flow patterns;

91 thus, reducing the initial inhibition. On the other hand, Tam and Wong (1990) indicated that microalgae in general prefer ammonium over nitrate or organic nitrogen, especially under continuous cultivation conditions. Tam and Wong (1996) revealed that 20 – 250 mg N/L ammonia concentrations did not inhibit the growth of Chlorella vulgaris.

Additionally, pH affects the inhibition due to ammonia elevated concentrations. At pH values less than 8, nitrogen exists mostly in the non-toxic ammonium form as opposed to the toxic ammonia which exists under alkaline conditions. Park et al. (2010) reported that the microalga Scenedesmus accuminatus experienced inhibition at ammonium concentration up to 100 mg NH4-N/L. 200 mg NH4-N/L or higher resulted in decreasing the final biomass concentration, but the inhibition became severe with an increase to

1000 mg NH4-N/L. Wang and Lan (2011a) found that the microalga N. oleoabundans consumed ammonium faster than nitrate, and they hypothesized that ammonium is the preferred nitrogen form for this microalga. In our experiment, the highest ammonia concentration was 249 mg N/L for the filtered 10% dilution. At this dilution, the biomass concentration as expressed by OD 750 was relatively low in the early stages, and in fact it was the lowest after around 23 d. After that, the growth increased and towards the end of the monitoring period, the OD 750 of the 10% filtered exceeded the 1%, 0.4%, and 0.2% filtered cultures. As a result, there was no clear sign of ammonia inhibition at the concentrations tested; however, the growth at high ammonia concentrations appeared to be slower in the early stages.

BG-11 5.71% 4.57% 2.29% 1.14% 0.57% 0.23% 0.11% 1.6

1.4

1.2

1.0

0.8 OD 750 0.6

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 Time (d)

Figure ‎3.9: Supernatant AD as a nutrient medium (average ± SD; n = 5)

BG-11 10% 8% 4% 2% 1% 0.4% 0.2% 1.4

1.2

1.0

0.8

OD 750 0.6

0.4

0.2

0.0 0 5 10 15 20 25 30 35 40 Time (d) Figure ‎3.10: Filtered AD as a nutrient medium (average ± SD; n = 5)

5.71% 4.57% 2.29% 1.14% 0.57% 0.23% 0.11%

0.7

0.6

0.5

0.4 750

OD 0.3

0.2

0.1

0 0 5 10 15 20 25 30 Time (d)

Figure 3‎ .11: Supernatant AD (4 mL) and DI water (1 mL) (average ± SD; n = 5)

There were several attempts to scale-up the cultivation of the microalga N. oleoabundans in a 100-L raceway pond by using the 2.29% AD supernatant as a nutrient medium. However, it has been observed that the culture was not purely N. oleoabundans, as Scenedesmus sp. and cyanobacteria cells were identified within the culture. Filtering of the invasive species was attempted to maintain a unialgal culture using filter bags with opening sizes of 10 and 5 μm. This technique was effective in removing cyanobacteria cells; however, Scenedesmus sp. cells were still dominant in the 5 μm filtrate. This was attempted several times; however, due to the relatively long cultivation period, other species grew in addition to N. oleoabundans.

Closed photobioreactors generally have the advantage of less contamination risks and easier contamination control when compared to open systems (Mata et al., 2010).

While widely used for the cultivation of high oil content microalgae, open ponds

94 susceptibility to contaminants invasion such as bacteria, viruses, and other algae may limit their applicability on a commercial scale (Mata et at., 2010; Menetrez, 2012). Some of the practices suggested in the literature to mitigate microalgal culture contamination include growing species that have been identified with less contamination risk when cultivated outdoor even for longer periods due to their tolerance to extreme conditions such as high alkalinity or salinity. These species include Dunaliella, Chlorella, Spirulina, and Arthrospira (Menetrez, 2012; Rodolfi et al., 2009). Menetrez (2012) indicated that genetically modified algae may become a method to limit contamination. Moreover, allowing native invasive species that are acclimated to the local conditions to take over and grow instead can reduce the contamination concerns (Rodolfi et al., 2009; Sheehan et al., 1998). Consequently, closed photobioreactors might be more suitable for the cultivation of the microalga N. oleoabundans. Future growth experiments in the raceway ponds will employ Scenedesmus sp. under shorter cultivation periods as maintaining a unialgal culture in the open system for multiple weeks is extremely difficult. This will reduce the likelihood that other invasive species such as cyanobacteria will have time to dominate the ecosystem (Mata et al., 2010; Rodolfi et al., 2009). Finally, subjecting the culture to extreme conditions in temperature, light, or pH may favor the growth of the target microalgae strain if it is tolerant to these extreme conditions (Mata et al., 2010).

In general, the combination of anaerobic digestion and microalgae cultivation is an attractive solution for biofuels production. The former produces biogas while decreasing the COD of the waste, and generates digestate that is rich with nutrients. This digestate is widely used as a fertilizer but it was proven that it can support the growth of

95 microalgae. Moreover, combining the two processes has other promising potentials such as scrubbing the carbon dioxide from the biogas using the microalgae and the anaerobic digestion of the microalgae biomass to produce biogas.

3.5 Conclusions

In this study, several pretreatment methods were applied to the anaerobic digestate from two sources. These pretreatment methods included hydrogen peroxide oxidation, filtration, and supernatant extraction. It was found that diluting the digestate and allowing it to settle for a certain time resulted in decreasing the COD as well as the turbidity in the supernatant. However, N/P ratio increased in the supernatant; as the reduction of phosphorus was considerably higher than nitrogen. Finally, the microalga N. oleoabundans was cultivated using diluted anaerobic digestate supernatant and filtrate. It was found that the supernatant provided slightly better growth compared to the filtered media, which is thought to be due to the high ammonia concentrations in the filtered digestate. Moreover, the 2.29% diluted supernatant, which is equivalent to a total nitrogen concentration of 100 mg N/L, appeared to yield the maximum biomass concentration of the microalga N. oleoabundans. It was attempted to scale up the microalgae cultivation to a raceway pond, but the culture was invaded by other microalgae species. Several attempts were made in order to eliminate the contamination and maintain a unialgal culture, but the extended cultivation time allowed other species to prevail.

CHAPTER 4. CULTIVATION OF SCENEDESMUS DIMORPHUS USING

ANAEROBIC DIGESTATE AS A NUTRIENT MEDIUM3

4.1 Abstract

In this study, the microalga Scenedesmus dimorphus was cultivated phototrophically using unsterilized anaerobic digestate as a nutrient medium. A bench- scale experiment was conducted by inoculating the microalga S. dimorphus with 0.05 –

10% dilutions of the anaerobic digestate supernatant. It was found that 1.25 – 2.5% dilutions, which is equivalent to 50 – 100 mg N/L total nitrogen concentrations and 6 –

12 mg P/L total phosphorus concentrations, provided sufficient nutrients to maximize the growth rate along with achieving high concentrations of algal biomass. The microalgae cultivation was scaled up to 100 L open raceway ponds, using 2.5% diluted anaerobic digestate at 317 and 454 μmol/m2/s average incident light intensities and 1.25% diluted anaerobic digestate at 234 and 384 μmol/m2/s average incident light intensities. The maximum biomass concentration was 432 mg/L which was achieved in the 2.5% dilution and 454 μmol/m2/s light intensity culture. Moreover, nitrogen, phosphorus, and COD removal efficiencies from the nutrient media were 65 – 72%, 63 – 100%, and 78 – 82%, respectively, whereas ammonia was completely removed from all cultures. For a successful and effective cultivation in open raceway ponds, light intensity has to be increased considerably to overcome the attenuation caused by the algal biomass as well as the suspended solids from the digestate supernatant.

3 This chapter was submitted to the Algal Research Journal.

Keywords: Scenedesmus dimorphus, microalgae, anaerobic digestion, biofuels, nutrients removal.

4.2 Introduction

In the past, algal cultivation was limited to laboratory scale experiments which aimed at food and feed purposes. After 1948, outdoor cultivation began by utilizing flue gases from industries to provide algae with carbon dioxide and the harvested biomass was used for protein synthesis. Later, outdoor algal cultivation systems were employed for wastewater treatment through a combination of microalgae and bacteria. In 1979, the production of biofuels from aquatic plants began by the Solar Energy Research Institute

(SERI) for the United States Department of Energy (Terry & Raymond, 1985).

Large scale cultivation of microalgae for biofuel and other purposes can be achieved in open or closed photobioreactors. While closed photobioreactors offer the advantage of easier contamination control, they are more expensive to construct and operate. On the other hand, open systems are widely used for the cultivation of microalgae due to the ease of construction and scale-up (Chen et al., 2013; Mata et al.,

2010). Amongst the open cultivation systems, raceway ponds are the most attractive choice for algae cultivation. A raceway pond is a closed loop that is built with concrete, compacted earth material, or plastic. The typical operating depth is 0.1 – 0.3 m and the biomass and nutrients are recirculated within the pond with the aid of a paddlewheel.

Nutrient medium is often introduced ahead of the paddlewheel, while harvesting takes place behind it. The introduction of carbon dioxide to the culture via bubbling is optional; as it enhances the growth of microalgae (James & Boriah, 2010). One of the biggest

98 challenges facing the cultivation of microalgae in raceway ponds is the photosynthesis inefficiency due to light attenuation in the water column, especially in deep cultures

(Chen et al., 2013). Practices such as flashing lights and forced light/dark cycles have been suggested to improve the photosynthetic efficiency and biomass productivity (Chen et al., 2013; Lunka & Bayless, 2013). Nevertheless, these practices might be impractical or hard to implement for outdoor cultures.

Microalgae cultivation for wastewater treatment has been studied broadly. Species such as Chlamydomonas, Botryococcus, Chlorella, Haematococcus, Spirulina, and

Scenedesmus have been utilized for wastewater treatment (Choi & Lee, 2012). For instance, Tam and Wong (1990) revealed that sewage samples from preliminary and primary sedimentation as well as activated sludge secondary effluent supported the growth of Chlorella pyrenoidosa. Anaerobic digestion, which is a widespread approach used to manage organic solid wastes such as animal manure, produces nutrient-rich slurry

(digestate) besides biogas. This nutrient-rich liquid waste can be used to cultivate microalgae; however, there are some challenges in utilizing the anaerobic digestate (AD) to cultivate microalgae such as high nutrients concentrations and suspended solids.

Furthermore, growth of microalgae can be inhibited by high concentrations of constituents such as ammonia and chemical oxygen demand (COD) (Abu Hajar et al.,

2016). Several studies have indicated the potential of cultivating microalgae using anaerobic digestion effluent or slurry. For example, the microalga N. oleoabundans has been cultivated using the digestate and the effluent from digesting animal manures, raw cheese whey, rice hull, soybean, and catfish wastes (Abu Hajar et al., 2016; Franchino et

99 al., 2013; Levine et al., 2011; Yang et al., 2011). Park et al. (2010) reported that the microalga Scenedesmus accuminatus was cultivated using filtered and autoclaved anaerobic digestion effluent from a local piggery farm. However, they observed inhibition to the growth at 200 mg NH4-N/L or higher concentrations. Tam and Wong

(1996) revealed that ammonia concentrations up to 250 mg NH4-N/L did not inhibit the growth of Chlorella vulgaris. However, they indicated that ammonia inhibition is a function of pH; as nitrogen exists in the non-toxic ammonium form at pH values less than

8. Travieso et al. (2006) stated that the growth of the microalga Chlorella vulgaris was not inhibited when cultivated using piggery wastewater at COD concentrations less than

1,100 mg/L. They speculated that the inhibition may be caused by high suspended solids resulting in light attenuation in the water column. It is noteworthy to mention that many studies conducted on the cultivation of microalgae using wastewaters reported using sterilized instead of raw wastewaters in order to prevent interference from other microorganisms. Common sterilizing techniques are filtration and autoclaving (Park et al., 2010).

Scenedesmus dimorphus is a freshwater microalga from the Chlorophyceae class that is known for its high growth rates and total fatty acids content (Renaud et al., 1994).

This microalga can grow in four distinguishable forms depending on the growth conditions such as temperature, radiation, and nutrients availability. The typical form is a colony of four cells with crescent-shaped outer cells. Other forms include single cells, rounded four-cell colonies resembling the shape of S. obliquus, and rounded single cells

(Oron et al., 1981).

100

Several studies have been conducted on the microalga S. dimorphus and its applicability for biofuels production. Besides biodiesel synthesis, anaerobic digestion of the S. dimorphus biomass has been suggested as an effective way to produce bio-methane due to the high methane yield of this species which can reach up to 400 mL methane/g

VS (Frigon et al., 2013). Moreover, this microalga is a potential candidate for carbon dioxide sequestration produced from fossil fuels burning as well as flue gases due to its high tolerance to elevated concentrations of carbon dioxide, a range of pH values, and salinity concentrations (Vidyashankar et al., 2013). It has been indicated that S. dimorphus can tolerate up to 20% CO2, 500 ppm NO, and up to 100 ppm SO2 concentrations when fed with flue gas. These conditions might impose toxic effects for many other algal species (Jiang et al., 2013). Giannetto et al. (2015) demonstrated that the tolerance of this species to elevated carbon dioxide concentrations can work as a tool to mitigate the contamination risk by other invasive species; thus, maintaining the abundance of S. dimorphus in the culture. Additionally, S. dimorphus is known for its ability to remove ammonia and phosphorus from a variety of wastewaters (González et al., 1997). In fact, Kang and Wen (2015) revealed that S. dimorphus was able to remove gaseous ammonia when blended with carbon dioxide and air. High biomass concentrations in the range 2.5 – 6.5 g/L were reported in the literature for this microalga

(Jiang et al., 2013; Ruangsomboon et al., 2013; Wang et al., 2013). In addition to its remarkable growth rates and high biomass concentrations, this species can accumulate high lipid content which can range from 18 to 32% by weight (Goswami & Kalita, 2011;

Welter et al., 2013). The most abundant fatty acids in this microalga are palimitic acid,

101 alpha-linolenic acid, cis-9-oleic acid, and cis-9, 12-linoleic acid (Ruangsomboon et al.,

2013; Vidyashankar et al., 2013).

One of the major obstacles facing the large scale cultivation of microalgae is the expensive production cost as a result of nutrients and energy requirements (Abu Hajar et al., 2016; Wang & Lan, 2011a). One way to offset the production costs of microalgae biomass is by utilizing readily available nutrient media including a variety of wastewaters

(Abu Hajar et al., 2016). Although studied widely, the large scale industrial application of microalgae in wastewater treatment is still limited (Pittman et al., 2011). Many studies have focused on cultivating S. dimorphus for nutrients removal from wastewaters; however, very few studies focused on the controlled cultivation of this microalga in open raceway ponds using wastewater as a nutrient medium. In fact, to our best knowledge, this is the first study to discuss the cultivation of S. dimorphus in an open raceway pond using diluted AD.

In this study, the microalga S. dimorphus was cultivated in a sustainable manner using diluted AD. The nutrient medium was not sterilized or filtered in our experiment as opposed to most of the studies on this topic. A bench-scale experiment was conducted by inoculating the microalga S. dimorphus with diluted AD supernatant at a range of dilutions in order to determine the optimum dilution or range of dilutions that will provide sufficient nutrients to optimize the growth of this microalga. The microalgae cultivation was then scaled up to 100 L raceway ponds. The microalga S. dimorphus was cultivated using two dilutions of the AD supernatant and each dilution was tested under

102 two light intensities. The growth of the microalgae as well as the nutrients removal efficiencies were evaluated by frequent monitoring.

4.3 Materials and Methods

4.3.1 Anaerobic digestate

AD was sampled from a commercial digester in Columbus, OH, that typically accepts food waste and animal manure for anaerobic digestion. The unsterilized digestate was stored in the laboratory at 4 ͦ C until the time of analysis.

4.3.2 Analytical methods

Total nitrogen, total phosphorus, ammonia, and COD concentrations were determined according to the colorimetric methods in compliance with APHA Standard

Methods for the Examination of Water and Wastewater and EPA methods (HACH methods 10072, 10127, and 10031, and 8000) using HACH DR 3900 spectrophotometer.

The algal biomass concentration was quantified by measuring the optical densities at 680 nm and 750 nm wavelengths (OD 680 and OD 750). A correlation was established between the S. dimorphus biomass concentration as total suspended solids (TSS) and the

OD 680. However, this correlation was not valid when this microalga was grown with turbid media such as the AD supernatant. It was observed during the experimental work that the ratio between OD 680 and OD 750 was almost constant for the microalga S. dimorphus, but this ratio tends to decrease as the microalgae solution was mixed with AD supernatant. Thus, a correlation between the OD 680/OD 750 ratio and the actual portion of microalgae biomass concentration in the mixture of microalgae and AD supernatant was established. It was indicated by de Winter et al. (2013) that the OD 680/OD 750 ratio

103 is a relative measure of pigment/chlorophyll content of the algal cells. For this purpose, three supernatant dilutions were prepared (TSS concentrations: 40; 80; and 142 mg/L) and mixed with S. dimorphus solution (biomass concentration: 330 mg/L) at four mixing ratios (20%, 40%, 60%, and 80% microalgae by volume); thus, a total of 12 solutions were tested. The OD 680 and OD 750 were measured for each separate component of the mixture (microalgae or AD supernatant) as well as the 12 combinations of microalgae and AD supernatant. In addition, the actual TSS concentrations of microalgae and supernatant were predetermined prior to mixing. The correlation was established between the OD 680/OD 750 ratio and the fraction of the actual microalgae biomass concentration with respect to the mixture of microalgae and AD supernatant. Measurement of TSS was performed according to the APHA Standard Methods for the Examination of Water and

Wastewater (method 2540 D) (Clesceri et al., 1998).

4.3.3 Microalgae selection

The microalga S. dimorphus, which was previously obtained from Algaeventure

Systems (Marysville, OH) and cultivated in our laboratory, was selected for cultivation in the bench-scale experiments as well as in the raceway ponds. The microalgae were cultivated in BG-11 medium (James, 1978) and kept in an incubator at 25 °C with 14:10 light/dark cycle using two built-in fluorescent lamps with an average light intensity of 50

μmol/m2/s.

4.3.4 Bench-scale microalgae cultivation

The microalga S. dimorphus was cultivated on a bench-scale in order to determine the optimum dilution or range of dilutions which will provide sufficient nutrients to

104 optimize the growth of this microalga on a larger scale. The nutrient medium used was the diluted and unsterilized AD supernatant which was prepared in a similar manner as explained in Abu Hajar et al. (2016). A range of 0.05% – 10% dilutions were tested, that is equivalent to 2 – 400 mg N/L and 0.23 – 48 mg P/L. Each dilution was tested in 5 replicates by mixing 4 mL of nutrient medium with 2 mL microalgae (initial OD 680 =

0.125) in 16 mm round glass vials (Figure 4.1a). The vials were kept inside an incubator at 25 °C with an average light intensity of 50 μmol/m2/s and 14:10 light-dark cycle. The biomass concentration was measured by placing each vial inside the spectrophotometer, and OD 680 and OD 750 were recorded for each vial. Specific growth rates were calculated for each dilution using the least-squares regression method, and statistical analysis was performed using IBM SPSS software in order to compare the algal growth under different dilutions.

4.3.5 Microalgae cultivation in the raceway ponds

The microalga S. dimorphus was cultivated in 100 L raceway ponds which were purchased from MicroBio Engineering (http://www.microbioengineering.com). The length and width of each pond were 95 and 40 cm, respectively, and the radius of the bend was 20 cm. The water depth equivalent to 100 L was 20 cm. Temperature and pH of the microalgae cultures in each pond were monitored using Neptune APEX controller

(http://www.neptunesystems.com), while pH was regulated by carbon dioxide bubbling to maintain a pH value of 7.5 ± 0.1. Light was supplied by a 760W UFO LED Grow

Light providing red, blue, orange, and white colors (Figure 4.1b). Temperature was maintained at 25 ± 1 °C using Eheim Jager 200 W aquarium heater, and the paddlewheel

105 was operated to achieve a surface water velocity of 0.3 m/s. The nutrient medium used was AD supernatant; where two dilutions were selected based on the bench-scale experiments which provide sufficient nutrients for the growth of the microalga S. dimorphus. The inoculation ratio was 10% microalgae (by volume) which yielded initial microalgae concentrations of 40 – 60 mg/L. Four average incident water surface light intensities were tested: 234 and 384 μmol/m2/s for one dilution and 317 and 454

μmol/m2/s for the other dilution (14:10 light/dark cycle). The difference in the light intensity between the two dilutions was to account for the difference in the initial turbidity between the two dilutions. This variation in the initial turbidity may not have a pronounced effect on the bench-scale experiment; however, in raceway ponds, light will likely be the limiting factor. Light intensity was measured using an LI-COR 190 quantum sensor connected to LI-250 A light meter (https://www.licor.com). Microalgal biomass concentration was quantified in triplicate by the frequent measurement of OD 680 and

OD 750 as indicated earlier. The samples were then stored in the fridge at 4 °C and allowed to settle for approximately 12 h, and the nutrients concentrations were measured in triplicates from the supernatant.

106

Figure ‎4.1: Microalgae cultivation (a) bench-scale experiment (b) raceway ponds

4.4 Results and Discussion

4.4.1 Biomass and optical density calibration

As indicated in section 4.3.2, 12 solutions were prepared by mixing different concentrations of the microalga S. dimorphus and AD supernatant. The TSS of each component (microalgae or AD supernatant) was measured separately and the OD 680 and

OD 750 values were measured for the 12 solutions of combined microalgae and AD supernatant. The calibration between OD 680 and the overall solution TSS (microalgae and AD supernatant) is shown in Figure 4.2a. The calibration shown in Figure 4.2b was established between OD 680/OD 750 and the ratio between the microalgae biomass concentration and the overall TSS concentration estimated by the equation in Figure 4.2a.

107

In other words, the microalgae biomass concentration can be determined according to

Equation 4-1:

OD 680 2 OD 680 Biomass concentration (mg/L)= [−232.8 ( ) +1158.4 ( ) OD 750 OD 750 Equation 4-1 − 1014.1] OD 680

(a) (b) 350 1.0

300 0.8 250 200 0.6 150 y = 402.09x 0.4 2 TSS (mg/L) 100 R² = 0.98 y = -0.579x + 2.881x - 2.522 AlgaeFraction 0.2 R² = 0.995 50 0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2 2.2 OD 680 OD 680/OD 750

Figure ‎4.2: Biomass concentration and optical density calibration (a) OD 680 and TSS calibration (b) OD 680/OD 750 vs. actual microalgae biomass fraction

4.4.2 Bench-scale cultivation of S. dimorphus

The growth of S. dimorphus using diluted AD supernatant on the bench-scale is shown in Figure 4.3a. In the beginning of the experiment, all vials were not significantly different from each other at 95% confidence level. It was clear that the initial dark turbidity expressed as OD 680 was caused by the AD supernatant especially in the highly concentrated solutions as shown in Figure 4.3b. This brown turbidity decreased in the first 4 – 8 d and then the color began turning green as a result of the algal growth. After 4 d of monitoring, the highest biomass concentration observed was for the 0.6% dilution

108 which was significantly higher than all other dilutions except the 1.25% dilution. Two days later, the maximum biomass concentration was observed in the 1.25% dilution which was significantly higher than all other dilutions and this remained to be the case until 8 d. After 8 d, the growth in the less concentrated dilutions (≤ 1.25%) started to slow down while the more concentrated solutions continued to grow under either exponential or linear growth mode. From 10 d to 13 d, the maximum biomass concentration was observed in the 2.5% dilution; however, the biomass concentration at this dilution was not significantly different from those observed in the 1.25%, 3.75%, and 5% dilutions. As a result, it appears that the lower diluted solutions resulted in faster growth at an earlier stage and reached the stationary phase faster than the more concentrated dilutions, which can be explained by the potential chemical inhibition in the more concentrated solutions due to elevated concentrations of ammonia and COD. Furthermore, by increasing the concentration of the supernatant solution, the turbidity increased as shown in Figure 4.3b, where the initial OD 680 ranged from 0.1 for the 0.05% dilution to 0.76 for the 10% dilution. This turbidity can decrease the light penetration by more than 30% in the highly concentrated solutions (10% and 7.5% dilutions) while the light attenuation is significantly lower in the less concentrated solutions.

The maximum biomass concentrations corresponding to each dilution are shown in Table 4.1. These biomass concentrations were achieved after 20 to 28 d for the 2.5 –

10% dilutions and after 1 to 11 d for the 0.05 – 1.25% dilutions. The highest biomass concentration of 654 mg/L which was observed in the 10% dilution was significantly higher than all other dilutions except the 5% and 7.5%. It is clear also from Table 4.1 that

109 there was essentially no growth in the 0.05% and 0.1% dilutions and that nutrients may become limited below the 0.6% dilution. Even though the maximum specific growth rate of 0.414 d-1 was observed in the 0.6% dilution, the exponential growth phase only extended until 4 d, after which the growth entered the stationary phase and the biomass concentration declined thereafter. Furthermore, the low biomass concentration observed in this dilution necessitates utilizing a more concentrated dilution. For instance, the

1.25% dilution, which had the second highest specific growth rate, resulted in 80% higher biomass concentration compared to the 0.6%, and the exponential growth extended to 6 d. The 2.5% dilution resulted in further increase in the biomass concentration and longer growth. Higher biomass concentrations were achieved in the more concentrated dilutions; however, the specific growth rates were lower and the initial turbidities were higher compared to the 1.25% and 2.5% dilutions. As a result, it was hypothesized that 1.25 –

2.5% dilutions (50 – 100 mg N/L total nitrogen) provide sufficient nutrients to maximize the specific growth rate of the microalga S. dimorphus while still producing a high concentration of biomass. This conclusion is in agreement with many studies in the literature. For example, Wang et al. (2013) indicated that high nitrogen concentrations such as those witnessed in a full strength BG-11 may have inhibitory effects on the growth of S. dimorphus. They also recommended an optimum nitrogen concentration range of 64 – 128 mg N/L for biomass and lipid. On the other hand, Xu et al. (2015) reported that the optimum nitrogen concentration for the microalga S. dimorphus was 152 mg N/L when a manure wastewater was used as a nutrient medium. Goswami and Kalita

(2011) cultivated the microalga S. dimorphus using BG-11 medium and by replacing

110 nitrate with urea as a nitrogen source. They concluded that urea concentration of 0.1 g/L, that is 46.6 mg N/L, was optimum for the biomass growth with a maximum specific growth rate of 0.54 d-1.

111

(a) 10% 7.5% 5% 3.75% 2.5% 1.25% 0.6% 0.3% 0.1% 0.05% 700

600

500

400

300

200

100 Biomass Biomass concentration (mg/L) 0 0 5 10 15 20 25 30 Time (d)

(b) 10% 7.5% 5% 3.75% 2.5% 1.25% 0.6% 0.3% 0.1% 0.05% 2.0 1.8 1.6 1.4

1.2 1.0

OD 680 0.8 0.6 0.4 0.2 0.0 0 5 10 15 20 25 30 Time (d)

Figure 4‎ .3: Microalgae growth using diluted anaerobic digestate supernatant (a) biomass concentration (mg/L) (b) OD 680

112

Table 4.1: Specific Growth Rates Corresponding to Different Nutrients Concentrations in the Diluted AD Supernatant Dilution Total N Total P Ammonia N COD TSS μ Max Biomass Conc (%) (mg N/L) (mg P/L) (mg N/L) (mg/L) (mg/L) (d-1) mg/L 10.00 400 48.0 292.0 2,048 326 0.107 654 7.50 300 36.0 219.0 1,536 242 0.191 600 5.00 200 24.0 146.0 1,024 153 0.213 595 3.75 150 18.0 109.5 768 118 0.218 484 2.50 100 12.0 73.0 512 78 0.280 449 1.25 50 6.0 36.5 256 38 0.359 290 0.60 24 2.9 17.5 123 18 0.414 159 0.30 12 1.4 8.8 61 8 0.280 80 0.10 4 0.5 2.9 20 3 0.171 37 0.05 2 0.2 1.5 10 1 0.162 28

4.4.3 Cultivation of S. dimorphus in raceway ponds

Based on the results of the bench-scale experiment, the microalga S. dimorphus was cultivated in 100 L raceway ponds on four separate tests using 2.5% and 1.25% diluted AD supernatant; that is equivalent to 100 and 50 mg N/L, respectively. As indicated in the Materials and Methods section, two light intensities were tested for each dilution: 317 and 454 μmol/m2/s for the 2.5% dilution and 234 and 384 μmol/m2/s for the

1.25% dilution. The initial characterization of the nutrient media for the four cultivation conditions is presented in Table 4.2, and the growth of the microalga S. dimorphus under these four conditions is shown in Figure 4.4a.

As expected, light intensity clearly impacted the growth of the microalga S. dimorphus, particularly in the first 4 d; as higher biomass concentrations were achieved under higher light intensity regardless of the dilution. This is in agreement with Wang et al. (2013) who indicated that for the microalga S. dimorphus, the specific growth rate

113 increased linearly with an increase in the light intensity until reaching a saturation light intensity of 510 μmol/m2/s. On the other hand, it was reported by Xu et al. (2015) that the optimal light intensity for the growth of S. dimorphus is 238 μmol/m2/s; however, growth under the 1.25% dilution at 234 μmol/m2/s light intensity in this experiment yielded the lowest biomass concentrations amongst all other conditions, which suggests that this light intensity is far from the optimum for the growth of this microalga in a raceway pond using a turbid media such as the diluted AD supernatant. It was demonstrated earlier that remarkable biomass concentrations have been reported in the literature for this microalga.

However, the experimental conditions differ from those adapted in our experiments. For example, Wang et al. (2013) stated that the highest S. dimorphus biomass concentration achieved was 6.5 g/L. However, their experiments were conducted outdoors in 1.9 – 7.6 cm deep glass panel photobioreactors with a peak light intensity of 1600 μmol/m2/s.

Similarly, Jiang et al. (2013) reported that the highest S. dimorphus biomass concentration was 5.17 g/L which was achieved in 3 – 5 cm diameter columns. The low biomass yield in our experiments compared to the previous studies can be explained simply by the high depth of liquid (20 cm), and the lower light intensities combined with the high turbidity caused by the suspended solids in the nutrient medium especially when the microalgae were cultivated using the 2.5% dilution with OD 680 as high as 1.2

(Figure 4.4b). For the 1.25% dilution, growth rate was likely limited by either low light penetration from increasing biomass concentration, nitrogen limitation, and/or phosphorus limitation particularly after 4 d when the nitrogen and phosphorus concentrations dropped below 30 mg N/L and 3 mg P/L, respectively (Figure 4.4,

114

Figure 4.5a, and Figure 4.5b). This conclusion matches the findings in Table 4.1 where it appears that at 0.6% dilution or below, which is equivalent to 24 mg N/L nitrogen concentration and 2.9 mg P/L phosphorus concentration, nutrients became limiting. It is not clear however whether a single or multiple nutrients were limiting. Many studies have investigated the effect of limiting nutrients on the algal growth. Kunikane et al. (1984) and Kunikane and Kaneko (1984) reported that it is necessary to consider the multiplicative effect of nitrogen and phosphorus on the growth of the microalga S. dimorphus. On the other hand, Rhee (1978) reported that the multiplicative or additive effect of nitrogen and phosphorus was not applicable to the microalga S. dimorphus.

Table 4.2: Initial Characterization of the Nutrient Media Used for Microalgae Cultivation in the Raceway Ponds (Average ± SD; n=3) Dilution 2.5% 1.25% 2 Light intensity (μmol/m /s) 317 454 234 384 Total N (mg N/L) 109 ± 3 109 ± 2 57 ± 1 53 ± 4 Total P (mg P/L) 10.8 ± 0.1 10.2 ± 0.1 4.8 ± 0.1 4.2 ± 0.1 Ammonia N (mg N/L) 78.8 ± 3 78.8 ± 3 41.8 ± 1 36.6 ± 2 COD (mg N/L) 487 ± 16 509 ± 14 272 ± 9 283 ± 5 TSS (mg/L) 93 ± 2 92 ± 1 45 ± 3 52 ± 4

115

(a)

2.5% (454) 2.5% (317) 1.25% (384) 1.25% (234) 500

400

300

200

100 Biomass Biomass concentration (mg/L)

0 0 2 4 6 8 10 12 14 Time (d)

(b)

2.5% (454) 2.5% (317) 1.25% (384) 1.25% (234) 1.5

1.2

0.9

OD 680 0.6

0.3

0.0 0 2 4 6 8 10 12 14 Time (d)

Figure ‎4.4: S. dimorphus growth in the raceway ponds under different dilutions and light intensities (a) biomass concentration (mg/L) (b) OD 680

116

(a) (b) 2.5% (454) 2.5% (317) 2.5% (454) 2.5% (317) 1.25% (384) 1.25% (234) 1.25% (384) 1.25% (234) 125 10

100 8 75 6 50 4 25 2 0 0 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Concentration P/L) (mg Concentration N/L) (mg Time (d) Time (d)

400 75 300 50 200 25 100 0 0

Concentration N/L) (mg 0 2 4 6 8 10 12 0 2 4 6 8 10 12 Time (d) Time (d) Concentration COD/L) (mg Figure 4‎ .5: Nutrients concentrations in the S. dimorphus cultures grown in the raceway pond (a) nitrogen (b) phosphorus (c) ammonia (d) COD

117

2.5% (454) 2.5% (317) 1.25% (384) 1.25% (234) 50 40

30 20 N/P ratio 10 0 0 2 4 6 8 10 12 Time (d)

Figure 4‎ .6: N/P ratio in the S. dimorphus cultures grown in the raceway pond

The highest biomass concentration in our experiments was 432 mg/L which was observed after approximately 10 d of cultivation using the 2.5% dilution and under a light intensity of 454 μmol/m2/s. The second highest biomass concentration was 367 mg/L for the same dilution but with 317 μmol/m2/s which was close to the biomass concentration achieved using the 1.25% dilution at 384 μmol/m2/s (352 mg/L). Finally, the highest biomass concentration achieved in the 1.25% dilution and in 234 μmol/m2/s light intensity was 145 mg/L. The biomass productivities were calculated between the initial inoculation time and the time the maximum biomass concentration was achieved. These productivities were 41.1, 32.3, 28.5, and 8.4 mg/L/d for the 2.5% dilution (454), 2.5% dilution (317), 1.25% dilution (384), and 1.25% dilution (234), respectively. However, the corresponding productivities for the first 4 d of cultivation were 63.6, 38.8, 47.3, and

27.2 mg/L/d, respectively, which suggests that a significant improvement to the biomass productivity can be achieved by a continuous or semi-continuous cultivation mode

118 compared to the batch cultivation which was followed in our experiments where nutrients were added only in the beginning of cultivation.

Besides the cultivation of S. dimorphus for biofuels production, this microalga has gained considerable interest for its ability to remove nutrients from wastewater. As shown in Figure 4.5a, the total nitrogen removal efficiencies were in the 65 – 72% range for all four conditions. Ammonia was completely removed in all four conditions.

However, it appears that higher light intensity resulted in faster ammonia uptake; as there was a sharp drop in the ammonia concentration in the first four days for the 384 and 454

μmol/m2/s cultures compared to the other two cultures (Figure 4.5c). Ammonia can be removed from wastewater via different methods such as volatilization (stripping), nitrification and denitrification, or assimilation by microalgae. For stripping to occur, the algae culture must be alkaline and under high temperature conditions, while nitrification and denitrification are characterized by a relatively long generation time of bacteria; besides a rapid decrease in the pH when nitrification occurs (Tam & Wong, 1990; Tam &

Wong, 1996). Throughout the cultivation in our system, the pH was maintained at 7.5 ±

0.1 and temperature was 25 °C; thus, the stripping effect can be excluded. Further, no nitrate was detected in the samples, which concludes that nitrification did not contribute to ammonia removal. Hence, it can be hypothesized that ammonia removal was the result of microalgae uptake. This can also suggest that ammonia is a favorable form of nitrogen for this microalga compared to organic nitrogen in the nutrient medium or even residual nitrate in the microalgae inoculum. This is in agreement with Tam and Wong (1990), who indicated that microalgae generally prefer ammonium over other nitrogen forms.

119

Wang and Lan (2011a) revealed that ammonium consumption by the microalga N. oleoabundans was faster than nitrate. However, Vidyashankar et al. (2013) reported that the microalga S. dimorphus grew better with sodium nitrate and ammonium acetate as nitrogen sources compared to ammonium carbonate and urea. Arumugam et al. (2013) reported similar results when they observed that the microalga S. bijugatus grew better with potassium nitrate and sodium nitrate as nitrogen sources compared to urea and ammonium nitrate.

N/P ratio is an important aspect in the microalgae cultivation. According to

Klausmeier et al. (2004), the atomic N/P ratio for phytoplankton varies between 8.2 and

45, and that in general, the cellular stoichiometry is flexible based on culture conditions.

Rhee (1978) reported that for Scenedesmus sp., the optimal atomic N/P ratio for cell density was 30, below which nitrogen becomes limited, while above this ratio phosphorus becomes limited. Kunikane et al. (1984) found that the specific growth rate of the microalga S. dimorphus was independent of phosphorous concentration at TN/TP ratios of 2 – 10, that is equivalent to an atomic ratio of 4.4 – 22.1. In our experiments, the atomic N/P ratios were in the range 14.4 – 28.3 except for the 1.25% dilution (384) when the phosphorus was entirely consumed towards the end of cultivation which increased the ratio to over 50 (Figure 4.6). Similarly, when the phosphorus was nearly depleted in the

1.25% dilution (234), the N/P ratio increased to 40. This may indicate that phosphorus was actually the limiting nutrient towards the end of cultivation period in the 1.25% cultures. The phosphorus removal efficiencies were 63%, 63%, 76%, and 100% for the

2.5% dilution (317), 2.5% dilution (454), 1.25% dilution (234), and 1.25% dilution (384),

120 respectively. This suggests that lower initial phosphorus concentrations in the diluted supernatant yielded higher removal efficiencies.

The existence of BOD or COD in a water sample implies the requirement of oxygen by microorganisms to oxidize the organic matter; thus, high BOD or COD concentrations indicate the potential of dissolved oxygen depletion in the receiving waters. The COD removal efficiencies ranged from 78 – 82% in all four cultures

(Figure 4.5d). Given the fact that microalgae provide oxygen, COD can be removed as a result of the mixed culture of microalgae and bacteria in an open pond using unsterilized digestate as a nutrient medium (Choi & Lee, 2012). As a result, the use of unsterilized

AD was advantageous in decreasing the COD levels of the nutrient medium; since high

COD have been reported to inhibit the growth of several microalgae species (Roudsary et al., 2014; Travieso et al. 2006).

Utilizing the microalga S. dimorphus can be an effective way of removing nutrients from wastewaters as well as producing a biomass for biofuels synthesis and other applications. However, in order to increase the efficiency of the process, light should be provided at higher intensities compared to those provided in this study. It is apparent from the nutrients analysis that light is likely to be the limiting factor when the microalga S. dimorphus was cultivated in 100 mg N/L AD supernatant dilution.

However, at least one nutrient was limiting after 4 d of cultivation when the 1.25% dilution was used as a nutrient medium. Thus, by increasing the light intensity while using 100 mg N/L, it is expected that nitrogen and phosphorus will be removed at higher efficiencies concurrent with higher biomass concentrations and growth rates. On the other

121 hand, continuous or semi-continuous cultivation as opposed to batch can be achieved effectively using the 1.25% dilution and at 384 μmol/m2/s light intensity; which is expected to produce comparable biomass productivities to those achieved at the 2.5% -

454 μmol/m2/s culture.

Outdoor cultivation is an attractive option for summer cultivation in a greenhouse where higher light intensities are available for the algal growth. Light intensities during summer days can peak at 1,500 μmol/m2/s or higher in Ohio. Further, the increased temperature can enhance the growth rate of this microalga. Even though many studies focused on the 20 – 25 ͦ C temperature as the optimal range for biomass and lipid accumulation, Benider et al. (2001) reported that the optimum growth for the microalga

S. dimorphus was achieved at 35 ͦ C. Further, they revealed that the optimal light intensity for the cultivation of this microalga increases with an increase of the temperature up to

35 ͦ C.

4.5 Conclusions

In this study, it was proven that the microalga S. dimorphus can be cultivated phototrophically using diluted anaerobic digestate supernatant. A bench-scale experiment was conducted over a wide range of dilutions, and it was determined that 50 – 100 mg

N/L dilutions provided sufficient nutrients to optimize the microalgal growth. The cultivation was scaled up to 100 L raceway ponds. Two dilutions (100 and 50 mg N/L) were selected and each dilution was tested under two light intensities. Regardless of the dilution, increasing the light intensity improved the biomass growth significantly. The highest biomass concentration achieved in the bench-scale experiment was 654 mg/L,

122 whereas in the raceway ponds the highest biomass concentration was 432 mg/L.

Furthermore, it was found that the nitrogen removal efficiency was 65 – 72%, whereas ammonia was completely removed from the nutrient media. Phosphorus was removed at

63 – 100% efficiencies while COD removal efficiencies were in the 78 – 82% range. It is recommended to use 1.25% dilution (50 mg N/L) with 384 μmol/m2/s or higher light intensities if the microalgae cultivation mode is continuous or semi-continuous. However, for a batch cultivation, 2.5% dilution (100 mg N/L) with 454 μmol/m2/s or higher light intensities should be used so as to ensure sufficient nutrients.

123

CHAPTER 5. HYDRODYNAMIC CHARACTERIZATION OF ALGAE

RACEWAY PONDS USING TRACER TESTS

5.1 Introduction

Reactors typically used for water and wastewater treatment can be categorized based on the operation pattern as batch and continuous-flow reactors. In batch reactors, the operation is non-continuous, where reactants are mixed and reactions proceed, while in continuous-flow reactors the operation is continuous with flow entering and leaving the system constantly. The reactions taking place in a reactor determine whether the system will function as ideal or non-ideal. Ideal reactor conditions are featured by uniform hydraulic conditions; i.e., no dispersion or diffusion. Several factors cause non- ideal flow such as short circuiting which results in a portion of inflow leaving the reactor in a significantly short time compared to the mean residence time; dispersion caused by longitudinal mixing which is independent of the constituent; wind, density, and temperature driven mixing; turbulence and eddies from turns, bends, constrictions, or expansions in flow; and molecular diffusion which occurs as a result of concentration gradient (Crittenden et al., 2005).

In order to characterize the non-ideal flow in reactors, it is essential to conduct residence time distribution (RTD) analysis by either a pulse or step change in tracer concentration (Teefy, 1996; Watten et al., 2000). A tracer is a conservative element that passes through a system without changing in total mass, such that no reacting or accumulating takes place (Crittenden et al., 2005). The application of sodium and chloride ions as tracers has been used widely; as these ions are non-reactive in addition to

124 the ease of analysis. However, using sodium chloride as a tracer may become challenging due to the high background concentrations in the water being tested such as tap water. As a result, the initial concentration in the tracer input must be increased to yield significantly higher concentrations than the background concentration (Teefy, 1996). The other alternative is to use deionized water instead of tap water, which may not be feasible for large scale testing. When conducting a tracer test, the effluent concentration is monitored with the aid of a conductivity meter in the case of salt addition or a spectrophotometer if dye is introduced. A tracer curve is a plot of the tracer concentration in the effluent as a function of time, and is considered as a helpful tool in evaluating the mean residence time and the dispersion coefficient (Crittenden et al., 2005; Teefy, 1996;

Watten et al., 2000).

In algal raceway ponds, paddlewheel mixing is necessary in order to maintain algal suspension throughout the depth of light penetration and prevent thermal stratification. For these reasons, the optimal surface velocity according to Green et al.

(1995) is 0.15 m/s, while Hadiyanto et al. (2013) proposed a range of 0.1 – 0.3 m/s to maintain a good mixing. Lower velocities result in poor mixing which leads to stagnant zones and may eventually result in anaerobic bacteria propagation and potentially toxic compounds. On the other hand, higher velocities may harm the growth especially for cells that are sensitive to hydrodynamic stress (Hadiyanto et al., 2013). In order to improve the economics of algal biodiesel production, it is important to quantify energy losses due to friction and bends and the energy required to operate the raceway ponds and compare that with the microalgal biomass and lipid productivities (Green et al., 1995).

125

In this chapter, tracer tests were conducted in a 100 L raceway pond at different surface velocities in order to evaluate the mean residence times necessary for a complete revolution of the liquid in the pond and to calculate the dispersion coefficients which can be incorporated in the algal growth modeling. Furthermore, the effect of paddlewheel mixing on the microalgal growth was evaluated by testing the growth of the microalga

Scenedesmus dimorphus at three different water surface velocities. Finally, energy losses due to friction and bending and the power required to operate the ponds were computed for the different mixing velocities and compared with the potential biodiesel yield at each velocity.

5.2 Materials and Methods

5.2.1 Theory

The theoretical hydraulic residence time (τ) is the volume of the reactor divided by the flow, while the mean hydraulic residence time (t̄), which is always less than τ due to the existence of stagnant regions (SR), is computed by Equation 5-1 (Crittenden et al,

2005; Watten et al., 2000):

∞ 퐶푡 푑푡 ∫0 푡̅ = ∞ Equation 5-1 퐶 푑푡 ∫0

Where C = concentration exiting the reactor at time t (mg/L). Note that residence time strictly applies to continuous flow reactors and is the average time that water spends in the reactor. We will apply this concept to raceway ponds which are batch systems, and

126 in this configuration, residence time refers to the average time it takes the water to circulate around the raceway pond once.

In general, diffusion is negligible compared to dispersion, and therefore, the longitudinal dispersion is the main reason for deviating from ideal flow, and the dispersed-flow model (DFM) as shown in Equation 5-2 describes the transport in one spatial dimension under these conditions with constant cross-sectional area and no short circuiting (Crittenden et al., 2005).

휕퐶 1 휕2퐶 휕퐶 − + = Equation 5-2 휕푧̅ 푃푒 휕푧̅2 휕(푡/휏)

Where z̅ is a dimensionless length = z/L; L is the reactor length (m); Pe = Peclet number = vL/E; v = average fluid velocity (Q/A); and E = dispersion coefficient (m2/s).

The average velocities were estimated using two methods. In method 1 (integral method), the mean hydraulic residence time for one complete lap was estimated according to Equation 5-1, and then the average velocity was computed by dividing the average travel distance by this calculated time. Method 2 (peaks method) was applied for estimating the average velocity based on the time between the first two distinguished peaks in the tracer response curve. This time represents the travel time required for the slug to complete one revolution in the pond, and the average velocity was calculated by dividing the average travel distance by this time.

Additionally, the tracer concentrations obtained from all the tests were fit to the

DFM model in order to estimate the dispersion coefficients. To solve for DFM, boundary conditions can be either closed or open. In the closed-system, dispersion is assumed only

127 to occur within the system, while plug flow takes place in and out of the reactor. For the open system, the reactor is assumed as a segment of the flow. Although not significantly different, the closed-system may provide better approximation (Crittenden et al., 2005).

Thomas and McKee (1944) developed a solution for the closed system (Equation 5-3) which relies on repetition until finding a Peclet number that approximately matches the tracer curve (Crittenden et al., 2005):

∞ 퐶 푎 sin 푏 + 푏 cos 푏 (푎2 + 푏2) 휃 = ∑ [ 2 2 ] exp [푎 − ] Equation 5-3 퐶표 푎 + 2 푎 + 푏 2 푎 푖=1

Where a = Pe/2, b= cot-1 [0.5 (b/a – a/b)], and θ is the normalized detention time.

An easier approximation based on the variance (σ) of the tracer curve is given by

Equation 5-4 (Crittenden et al., 2005; Watten et al., 2000):

2 2 휎푡 2 1 휎2 = = ( ) − [2 ( ) (1 − 푒−푃푒)] Equation 5-4 휃 (푡)̅ 2 푃푒 푃푒

For an open system, the solution can be obtained by Levenspiel and Smith (1957) formula as shown in Equation 5-5 (Crittenden et al., 2005):

푃푒(1−휃2) 1 − 퐸(휃) = 푒 4휃 Equation 5-5 √4휋 휃 (1/푃푒)

Where E(θ)‎ is the normalized exist age distribution. Similarly, the variance can be used to estimate Pe as shown in Equation 5-6:

2 2 휎푡 2 1 휎2 = = + 8 ( ) Equation 5-6 휃 (푡)̅ 2 푃푒 푃푒

128

Coefficient of dispersion can also be predicted as in Davies (1972) formula for an open channel with high Reynolds number as shown in Equation 5-7 (Crittenden et al,

2005):

퐸 = 1.01 휐 (푅푒)0.875 Equation 5-7

Where υ is the kinematic viscosity (m2/s); Re is Reynolds number = 4 v R/υ; v is the velocity (m/s); and R is the hydraulic radius (m). This approach may not provide accurate values, but it will be used as a comparison to the values obtained using the DFM model with closed and open boundary systems.

One-way ANOVA was performed using the IBM SPSS software in order to compare the mean velocities and dispersion coefficients computed at different tracer input locations as well as to compare the two methods (integral and peaks) used to estimate the mean hydraulic residence time.

The head loss due to the 180 ̊ curve is calculated as shown in Equation 5-8 (Green et al., 1995; Lundquist et al., 2010; Rogers et al., 2014):

퐾푣2 ℎ = Equation 5-8 퐾 2푔

Where hK is the head loss due to the 180 ̊ curve (m); v is the mean velocity (m/s); g is the gravitational acceleration (m/s2); and K is the kinetic loss coefficient

(theoretically equals 2 for 180 ̊ bends). For one pond, the hK will be multiplied by 2 to account for two bends.

129

Manning equation can be used to calculate the frictional head losses along straight lengths of the raceway pond as shown in Equation 5-9 (Green et al., 1995; Lundquist et al., 2010; Rogers et al., 2014):

푣2푛2퐿 ℎ퐿 = 4 Equation 5-9 푅3

Where hL is the frictional head loss along the straight length of the raceway pond

(m); v is the mean channel velocity (m/s); n is Manning’s roughness coefficient (0.015);

R is the hydraulic radius (m); and L is the channel length (m). For one pond, the channel length is double the length of the pond in order to account for the full revolution. The summation of hK and hL yields the total head loss.

The power required to operate the raceway ponds is calculated based on

Equation 5-10 (Rogers et al., 2014):

푄. 휌. ℎ 푊 = 9.8 Equation 5-10 푒

Where W is the power required to operate the pond (W); Q is the flow rate (m3/s);

훒 is the density of water (kg/m3); h is the total head loss (m); e is the system efficiency

(40% assumed); and 9.8 is a conversion factor (W.s/kg/m). It is important to mention that the estimated power according to Equation 5-10 takes into account only the energy required to operate the paddlewheel at variable speeds while other operations such as lighting and harvesting are not considered.

130

5.2.2 Experimental work

Four raceway ponds (100 L each) with the dimensions shown in Figure 5.1 were purchased from MicroBio Engineering (http://www.microbioengineering.com). The tracer experiments were conducted by filling one pond with 100 L DI water, which was equivalent to 0.2 m water depth. Three surface velocities that yield turbulent flow in the pond were tested: 0.1, 0.2, and 0.3 m/s. The surface velocity was measured with the aid of a ping-pong ball and a stop watch. A concentrated sodium chloride solution was used as a tracer and added to the flow at approximately one third the depth from the bottom at three points: point 1 upstream of the paddlewheel, point 2 downstream of the paddlewheel right before the 180 ̊ bend, and point 3 after the bend (Figure 5.1). The tracer slug volume used was 40 mL which when added to the 100 L resulted in 40 mg/L salt concentration increment. Sodium chloride was oven-dried at 105 °C for an hour and was desiccated afterwards for 30 minutes prior to the tracer solution preparation. The tracer concentration was measured downstream using a conductivity meter that was measuring at one third the depth from the bottom (Figure 5.1). The average water temperature throughout the test was 20 ± 0.7 °C, so the conductivity readings were corrected to 25 °C, and a calibration was established between conductivity and salt concentration.

131

Figure 5‎ .1: Raceway pond illustration

Finally, the effect of paddlewheel mixing on the microalgal growth was examined by cultivating the microalga S. dimorphus in the raceway ponds using BG-11 medium at

0.1, 0.2, and 0.3 m/s water surface velocities. Temperature was maintained at 25 ± 1 °C using Eheim Jager 200 W aquarium heater and the average incident water surface light intensity was 454 μmol/m2/s (14:10 light-dark cycle) supplied by a 760W UFO LED

Grow Light providing red, blue, orange, and white colors. The growth of microalgae was evaluated by measuring the OD 680 which was correlated with dry biomass concentration in mg/L as shown in Chapter 4.

132

5.3 Results and Discussion

The calibration that was established between the conductivity measurements at 25

°C in μS/cm and the sodium chloride concentration in mg/L is shown in Figure 5.2.

Tracer concentrations as function of time are shown in Figure 5.3 through Figure 5.11 for the three surface velocities and three tracer input locations. Each figure shows three replicates, and the tracer is released at time zero. Figure 5.3 through Figure 5.5 are for 0.1 m/s surface velocity but at different tracer release points, where point 1 is when the tracer is added before the paddlewheel; point 2 is when the tracer is added before the bend; and point 3 in when the tracer is added after the bend. Figure 5.6 through Figure ‎5.8 are for

0.2 m/s surface velocity, and Figure 5.9 through Figure 5‎ .11 are for 0.3 surface velocity.

The application of Equation 5-1 to the raceway ponds is somewhat different from the traditional water and wastewater applications; as the initial concentrations in the pond are not restored at the other side of the peak, and the impulse peak make 3 laps before reaching the steady state concentration, unlike a plug flow reactor where the effluent completely exits the reactor and is replaced by incoming flow. Equation 5-1, which is used to estimate the mean hydraulic residence time, is in fact the total moments in the time direction. Since the initial concentrations in the reactor are not restored, it is assumed that a linear function extends from the lowest point on the left to the lowest point on the right (between the two minimums across the peak). The main function to be integrated will be expressed as the difference between the two functions (the upper curve minus the straight line).

133

1200

1000 y = 0.45x - 9.49 R² = 1.00

800

600

NaCl (mg/L) 400

200

0 0 500 1000 1500 2000 2500 Conductivity @ 25 °C (μS/cm)

Figure 5‎ .2: Conductivity and sodium chloride concentration calibration

T-1 T-2 T-3

90 80 70 60 50 40 30 20 10

NaCl Concentration (mg/L) 0 0 20 40 60 80 100 120 Time (s)

Figure 5‎ .3: 0.1 m/s surface velocity (point 1)

134

T-1 T-2 T-3

140 120 100 80 60 40 20

NaCl Concentration (mg/L) 0 0 20 40 60 80 100 Time (s)

Figure 5‎ .4: 0.1 m/s surface velocity (point 2)

T-1 T-2 T-3

120 100 80 60 40 20

NaCl Concentration (mg/L) 0 0 20 40 60 80 Time (s)

Figure ‎5.5: 0.1 m/s surface velocity (point 3)

135

T-1 T-2 T-3

70 60 50 40 30 20 10

NaCl Concentration (mg/L) 0 0 10 20 30 40 50 60 Time (s)

Figure 5‎ .6: 0.2 m/s surface velocity (point 1)

T-1 T-2 T-3

120 100 80 60 40 20

NaCl Concentration (mg/L) 0 0 10 20 30 40 50 60 Time (s)

Figure ‎5.7: 0.2 m/s surface velocity (point 2)

136

T-1 T-2 T-3

80 70 60 50 40 30 20 10

NaCl Concentration (mg/L) 0 0 10 20 30 40 50 60 Time (s)

Figure 5‎ .8: 0.2 m/s surface velocity (point 3)

T-1 T-2 T-3

70 60 50 40 30 20 10

NaCl Concentration (mg/L) 0 0 10 20 30 40 50 Time (s)

Figure ‎5.9: 0.3 m/s surface velocity (point 1)

137

T-1 T-2 T-3

100

NaCl Concentration (mg/L) 0 0 10 20 30 40 50 Time (s)

Figure 5‎ .10: 0.3 m/s surface velocity (point 2)

T-1 T-2 T-3

90 80 70 60 50 40 30 20 10

NaCl Concentration (mg/L) 0 0 10 20 30 40 Time (s)

Figure ‎5.11: 0.3 m/s surface velocity (point 3)

A similar trend was observed in all the runs as shown in Figure 5.3 through

Figure 5.11; as the concentration tends to reach a steady state after three complete revolutions. The average velocities computed using method 1 (integral method) are shown in Table 5.1.

138

Table 5.1: Mean Velocities Estimated by Method 1 Surface velocity Tracer input location m/s Point 1 Point 2 Point 3 0.1 0.086 ± 0.010 0.119 ± 0.008 0.126 ± 0.001 0.2 0.150 ± 0.004 0.203 ± 0.002 0.212 ± 0.012 0.3 0.194 ± 0.004 0.277 ± 0.033 0.249 ± 0.024

Due to natural velocity gradients, the average velocity in an open channel flow matches the velocity at approximately 0.6 the depth measured from the free surface, and is often 5 – 20% less than the surface velocity (Subramanya, 2009). As shown in

Table 5.1, the average velocities estimated at point 1 were 65 – 86% of the surface velocity. The average velocities estimated at points 2 and 3 exceeded the surface velocities at 0.1 and 0.2 m/s surface velocities. Moreover, estimating the average velocity using this method produced a significant deviation depending on the tracer input location.

For instance, the values obtained for point 1 were less than other values computed for the other locations as well as the corresponding surface velocities. This may be explained by the flow non-uniformity across the paddlewheel, which produced tracer curves different from other locations. This is shown statistically as the average velocities computed by the integral method were compared among the three tracer input locations using one-way

ANOVA. It was found that at a surface velocity of 0.1 m/s, there was a significant difference between point 1 and the other two locations at 95% confidence levels, and that the average velocity estimated using points 2 and 3 were not significantly different at the same confidence level. Similar results were observed for the 0.2 m/s surface velocity as the difference in the mean velocities was not statistically different at 95% confidence levels between points 2 and 3, unlike point 1 which was significantly different from the

139 other two points at the same confidence level. Finally, for the 0.3 m/s surface velocity, the results were slightly different as the average velocity estimated at point 1 was not significantly different from that obtained at point 3, while it was significantly different from the average velocity estimated at point 2, and the velocities estimated using points 2 and 3 were not significantly different at 95% confidence level.

Table 5.2 shows the average velocities estimated using method 2 (method of peaks). One-way ANOVA was also conducted on the average velocities computed using method 2. Unlike method 1, the only significant difference was at 0.1 m/s surface velocity between points 2 and 3 at 95% confidence level. All other comparisons between different points at the three surface velocities yielded insignificant differences. Finally, a comparison was made between the two methods by comparing the average velocities corresponding to each surface velocity at a specific point between the integral and the peaks methods. It was found that the two methods provided average velocities that are significantly different at 95% confidence level except in the cases of points 1 at 0.1 m/s surface velocity and points 2 and 3 at 0.3 m/s surface velocity. Overall, there was a range of average velocities computed by the integral and peaks methods. Relying on point 1 using method 1 appears to underestimate the average velocity while using points 2 and 3 produced average velocities that are greater than the surface velocities. On the other hand, the only significant difference when method 2 was applied was between points 2 and 3 at 0.1 m/s. The ratio of average velocity computed by method 2 to the surface velocity ranged from 0.76 – 0.93, which is a more consistent range compared to the integral method, even though values less than 0.8 appear to be relatively low according to

140

Subramanya (2009). As a result, it was concluded that method 2 (peaks method) provides better representation of the flow in the raceway pond compared to the integral method.

The average velocities computed based on the peaks method were 0.086, 0.174, and

0.239 m/s for the 0.1, 0.2, and 0.3 m/s surface velocities, respectively, and these velocities were averaged at the three tracer input locations, and were used for energy losses estimations.

Table 5.2: Mean Velocities Estimated by Method 2 Surface velocity Tracer input location m/s Point 1 Point 2 Point 3 0.1 0.087 ± 0.001 0.091 ± 0.004 0.079 ± 0.005 0.2 0.172 ± 0.011 0.185 ± 0.003 0.166 ± 0.005 0.3 0.228 ± 0.005 0.242 ± 0.016 0.246 ± 0.010

As explained earlier, the longitudinal dispersion is the main reason for deviating from the ideal flow, so the dispersion coefficients were evaluated assuming closed and open boundary systems, and these values were compared to Davies (1972) approximate formula (Table 5.3). It appears that the dispersion coefficient depends heavily on the tracer input location; as the range of the dispersion coefficients computed using the closed system approach was 0.0027 to 0.0128 m2/s for the 0.1 m/s surface velocity,

0.0041 to 0.0159 m2/s for the 0.2 m/s surface velocity, and 0.0039 to 0.0186 m2/s for the

0.3 m/s surface velocity while for the open system approach the range was 0.0025 to

0.108 m2/s for the 0.1 m/s surface velocity, 0.0039 to 0.0140 m2/s for the 0.2 m/s surface velocity, and 0.0038 to 0.0168 m2/s for the 0.3 m/s surface velocity. Nevertheless,

141 dispersion coefficients estimated using the closed system approach were slightly higher than those estimated using the open system approach but the differences between the two approaches were not statistically significant at 95% confidence level regardless of the surface velocity and tracer input location. Davies (1972) formula (Equation 5-7), which was also used in order to compute approximate values of the dispersion coefficients in the pond, produced values close to those estimated at tracer input location 3. There were no significant differences at 95% confidence level between the dispersion coefficients evaluated using Davies (1972) formula, and those estimated using the closed and open boundary system approaches when the comparison was made at location 3 regardless of the surface velocity and at location 2 at 0.3 m/s. It is therefore possible that the values obtained at location 3 may actually be the most accurate, and could be incorporated in the algal growth to account for the degree of mixing.

Finally, the head losses based on one revolution of flow were estimated based on the computed average velocities as shown in Table 5.4. It is clear that the difference in the head loss indicates that besides the additional energy required to operate at a higher velocity, there will be additional power required by the paddlewheel in order to overcome this energy loss. This additional power is dependent upon the flow rate, density of the culture, and head loss (Hadiyanto et al., 2013). This raises the question whether a higher paddlewheel velocity is desired, and whether the additional power consumed by the paddlewheel will yield a significant improvement in the algal growth and biodiesel yield.

For this reason, the effect of paddlewheel mixing was evaluated by testing the growth of

S. dimorphus at 0.1, 0.2, and 0.3 m/s water surface velocities. Sufficient mixing is

142 necessary in order to maintain the algal suspension throughout the depth of light penetration and prevent thermal stratification (Green et al., 1995; Hadiyanto et al., 2013).

As shown in Figure 5.12, there was a clear impact of the mixing velocity on the growth of S. dimorphus; as the maximum specific growth rates corresponding to the 0.1, 0.2, and

0.3 m/s velocities were 0.198, 0.272, and 0.294 d-1, respectively. Additionally, the maximum biomass concentrations achieved were 450, 669, and 761 mg/L for the 0.1, 0.2, and 0.3 m/s velocities, respectively. The biomass productivities computed from time 0 until the maximum biomass concentration was achieved were 24.0, 32.0, and 52.8 mg/L/d, for the 0.1, 0.2, and 0.3 m/s velocities, respectively, and for the 100 L raceway pond, the corresponding algal biomass productivities were 2.4, 3.2, and 5.3 g/d.

Assuming a 25% lipid content, and 80% extraction efficiency, and 0.86 kg/L density of algal oil (Lundquist et al., 2010; Rogers et al., 2014; Schlagermann et al., 2012), the amounts of microalgal oil produced are estimated as 0.56, 0.74, and 1.23 mL oil/day for the 0.1, 0.2, and 0.3 m/s velocities, respectively. Assuming a heating value of 41 MJ/kg algal oil (Schlagermann et al., 2012), the estimated energy generated from the microalgal oil is 5.47 × 10-3, 1.20 × 10-2, 7.29 × 10-3 kwh/d for the 0.1, 0.2, and 0.3 m/s velocities, respectively. By comparing these values with the power required to operate the pond at different velocities (Table 5.4), running the pond at 0.1 m/s surface velocity appears to be the only case with a positive net energy balance. This implies that even though higher biomass productivities and biomass concentrations were achieved with the 0.2 and 0.3 m/s surface velocities, operating the ponds at 0.1 m/s seems to be the only feasible scenario given the previous assumptions.

143

Paddlewheel mixing is a key aspect in the microalgae cultivation; however, their main disadvantage is the high operating cost associated with the paddlewheel as well as other aspects such as lighting and harvesting. One of the solutions that may decrease the operating costs is to decrease the mixing velocity during the night, whereas the mixing velocity is increased during the day to optimize the photosynthetic yield and provide sufficient light penetration and reduce residence time in the dark zones (Rogers et al.,

2014).

144

Table 5.3: Dispersion Coefficients Values (m2/s) Tracer input location Surface velocity (m/s) Point 1 Point 2 Point 3

Dispersion Closed Open Davies Closed Open Davies Closed Open Davies coefficient formula System System (1972) System System (1972) System System (1972) 0.0033 ± 0.0030 ± 0.0066 ± 0.0049 ± 0.0046 ± 0.0088 ± 0.0113 ± 0.0097 ± 0.0092 ± 0.1 0.0006 0.0006 0.0007 0.0012 0.0010 0.0005 0.0013 0.0010 0.0001 0.0043 ± 0.0041 ± 0.0107 ± 0.0084 ± 0.0078 ± 0.0140 ± 0.0135 ± 0.0121 ± 0.0145 ± 0.2 0.0041 0.0003 0.0003 0.0010 0.0009 0.0001 0.0028 0.0024 0.0007 0.0052 ± 0.0049 ± 0.0135 ± 0.0135 ± 0.0123 ± 0.0184 ± 0.0125 ± 0.0114 ± 0.0167 ± 0.3 0.0012 0.0011 0.0002 0.0049 0.0042 0.0019 0.0009 0.0007 0.0014

Table 5.4: Total Head Losses and Power Requirements Estimated as a Function of Average Velocity Average Velocity Head loss (m) Power required (m/s) Frictional Bend Total kWh/d 0.086 1.17 × 10-4 1.51 × 10-3 1.62 × 10-3 3.28 × 10-3 0.174 4.79 × 10-4 6.17 × 10-3 6.65 × 10-3 2.72 × 10-2 0.239 9.03 × 10-4 1.16 × 10-2 1.25 × 10-2 7.04 × 10-2

145

0.3 m/s 0.2 m/s 0.1 m/s 800

700 600 500 400 300 200 Concentration (mg/L) 100 0 0 5 10 15 20 Time (d)

Figure ‎5.12: S. dimorphus growth at different mixing velocities

5.4 Conclusions

In this chapter, the average flow velocity in an algae raceway pond was evaluated using tracer tests. It was found that evaluating the hydraulic residence time using the time between the first two consecutive peaks and then calculating the average velocity provided more consistent results than computing the average velocity using the integral method. The computed average flow velocities which were used for energy losses computation were 76 – 93% of the surface velocities. Moreover, dispersion coefficients evaluated using the integral method indicated that this parameter depends on the tracer input location; however, the values estimated when the tracer was added after the bend

(location 3) were close to those estimated using Davies (1972) formula. If the dispersion coefficient is to be included in the algal growth model, it is recommended to evaluate the dispersion coefficients based on tracer input location 3. Finally, the microalga

Scenedesmus dimorphus was cultivated in the raceway pond at 0.1, 0.2, and 0.3 m/s water

146 surface velocities. It was found that 0.3 m/s surface velocity yielded the highest biomass concentrations and biomass productivity; however, 0.1 m/s velocity was the only velocity that produced a positive net energy.

147

CHAPTER 6. MODELING THE GROWTH OF MICROALGAE IN RACEWAY

PONDS USING DILUTED ANAEROBIC DIGESTATE AS A NUTRIENT

SOURCE4

6.1 Introduction

Research efforts have been recently directed towards the production of biofuels as a potential replacement to conventional fuels. Microalgae biodiesel is an attractive sustainable third generation biofuel produced from the microalgal neutral lipids

(Greenwell et al., 2009; Lam & Lee, 2012; Williams & Laurents, 2010). Nevertheless, producing biodiesel from microalgae is not yet economically feasible (Milledge, 2011).

Hence, microalgae growth modeling and optimization are necessary in order to make the large-scale cultivation of microalgae sustainable and feasible (Béchet et al., 2013;

Huesemann et al., 2013; Lee et al., 2015).

Most of the existing models on microalgae growth are either complex with many input parameters, non-validated, or difficult to generalize. In order to validate models with many parameters, it is necessary to conduct rigorous experimental measurements

(Huesemann et al., 2013; Lee et al., 2015).

In this chapter, two biomass growth models were developed and tested to simulate the growth of the microalga Scenedesmus dimorphus in raceway ponds using BG-11 medium and diluted anaerobic digestate (AD) as nutrient media. The main focus of the models was on light attenuation due to the algal biomass as well as the suspended solids in the nutrient medium.

4 This chapter will be submitted to the Applied Biochemistry and Biotechnology Journal.

148

6.2 Materials and Methods

6.2.1 Microalgae selection

The microalga S. dimorphus which was cultivated in our laboratory using BG-11 medium and AD supernatant as indicated in Chapters 4 and 5 was selected for modeling.

6.2.2 Biomass growth model

Two models for predicting the microalgal biomass growth were tested. The first model which is widely used is based on the assumption that microalgae grow exponentially at the maximum specific growth rate if all cultivation conditions are optimal. However, under sub-optimal conditions, the specific growth rate decreases. The other model is the logistic model which has gained interest recently as a good fit to batch microbial cultures.

Assuming that basal metabolism and predation are neglected, that microalgal cells are contained within the pond boundaries and that the mixing velocity is sufficient to keep the cells in suspension; hence, settling is neglected, the microalgae exponential growth model is given according to Equation 6-1 (Cerco & Cole, 1995):

휕퐶 = µ퐶 Equation 6-1 휕푡

Where:

C = algal biomass concentration (mg/L).

t = time (d).

µ = specific growth rate (d-1) and is given by Equation 6-2:

149

µ = µ푚 푓(푁)푓(퐼)푓(푇) Equation 6-2

Where:

-1 µm = maximum specific growth rate (d ).

f(N) = effect of suboptimal nutrient concentrations (0 ≤ f ≤ 1).

f(I) = effect of suboptimal illumination (0 ≤ f ≤ 1).

f(T) = effect of suboptimal temperature (0 ≤ f ≤ 1).

For cyanobacteria, an additional function can be added to account for the toxicity resulting from salinity, f(S).

In this model, mixing is assumed to be sufficient so that cells will be under continuous travel vertically and that there will be instantaneous adaptation to the various light conditions across the water depth. This assumption has been made and verified by many researchers for multiple microalgal species (Huesemann et al., 2013). As a result, the exponential growth model can be solved as shown in Equation 6-3:

퐶(푡 + ∆푡) = 퐶(푡). 푒휇 ∆푡 Equation 6-3

The logistic model, which has a sigmoidal shape, has gained popularity recently as a good fit to batch microorganisms’ growth. Unlike other models, the logistic model predicts the lag phase, exponential phase, and the stationary phase; however, it is independent of the substrate concentration (Surendhiran et al., 2015; Kargi, 2009; Nath et al., 2008). Equation 6-4 represents the microalgal biomass growth according to the logistic model (Rao et al., 2008):

150

푑퐶 퐶 = µ푢 (1 − ) 퐶 Equation 6-4 푑푡 퐶푚

Where C is the biomass concentration (mg/L); t is the time (d); µu is the

-1 maximum specific growth rate (d ); and Cm is the maximum biomass concentration

(mg/L). By rearranging the terms, Equation 6-4 becomes:

푑퐶 = µ 푑푡 퐶 푢 Equation 6-5 퐶 (1 − ) 퐶푚

The integration of Equation 6-5 over (t1, t2) and (C1, C2) yields:

퐶푚 퐶2 = 퐶푚 − 퐶1 Equation 6-6 1 + µ 훥푡 퐶1푒 푢

Both models were validated by comparing the growth curves of S. dimorphus when cultivated using BG-11 medium and AD supernatant as nutrient media in the raceway ponds with the predicted growth curves and the goodness of fit was evaluated with the coefficient of determination R2.

The following sections address the suboptimal cultivation conditions and how they were incorporated in both models.

6.2.2.1 Nutrients

Carbon, nitrogen, and phosphorus are the major nutrients for algal growth (Cerco

& Cole, 1995). In our model, the microalgae cultures were assumed nutrient-sufficient, and the sub-optimal nutrient function f(N) was assigned a value of 1.0.

151

6.2.2.2 Temperature

Temperature is also an important parameter that has a significant impact on the growth of microalgae. Many studies have indicated that the optimal temperature for the growth of microalgae falls within the 20 – 25 °C. For instance, Xin et al. (2011) demonstrated that the highest specific growth rate of 0.76 d-1 for the microalga

Scenedesmus sp. was achieved at 25 °C while they suggested that the optimal temperature for both biomass production and lipid accumulation for the same microalga was 20 °C. Benider et al. (2001) reported that the optimal light intensity for the microalga

S. dimorphus is a function of temperature. All of our experiments were conducted at 25

°C; as a result, this temperature was assumed to be the optimal temperature and the factor f(T) was assigned a value of 1.0 in the model.

6.2.2.3 Light

Light is a key factor in the phototrophic cultivation of microalgae. Light is generally delivered in the form of a single quantum, each called a photon. While a portion of the photon flux energy is being absorbed by the pigment molecules in the algal cells, the other portion is dissipated as heat or lost via reflection (Richmond, 2004).

Several models exist in the literature to predict the effect of light intensity on the microalgal growth. One of the most popular models is Tamiya’s (1953) model. This model, which resembles Monod’s equation, assumes microalgae exist in individual cells and predicts the growth rate as a function of incident light intensity with no photoinhibition as shown in Equation 6-7 (Lee et al., 2015):

퐼 휇 = 휇푚 Equation 6-7 퐾퐼 + 퐼

152

-1 Where μm is the maximum specific growth rate (d ), I is the light intensity

2 2 (µmol/m /s), and KI is the saturation constant (µmol/m /s).

On the other hand, the most popular model that assumes photoinhibition is

Steele’s (1962) model which accounts for photoinhibition at higher light intensities as shown in Equation 6-8 (Lee et al., 2015):

퐼 퐼 1− 퐼 휇 = 휇푚 푒 푚 Equation 6-8 퐼푚

-1 Where μm is the maximum specific growth rate (d ), I is the light intensity

2 2 (µmol/m /s), and Im is the optimal light intensity (µmol/m /s). To determine which light intensity model governs the growth of the microalga S. dimorphus, a bench-scale experiment was conducted in 250 mL Erlenmeyer flasks (100 mL working volume) using

BG-11 medium as a nutrient medium. Light was supplied by a 760W UFO LED Grow

Light providing red, blue, orange, and white colors at 14:10 light/dark cycle and the light intensities tested were 25, 80, 190, 245, 311, and 388 µmol/m2/s where three replicates were tested for each light intensity. All flasks were placed on a shaker table at 200 RPM and the temperature was 25 ± 2 °C. The growth was monitored by measuring the optical density at 680 nm (OD 680 nm) which was correlated with the dry biomass concentration in mg/L as shown in Chapter 4. For the exponential model, the specific growth rates corresponding to each light intensity were determined for the exponential growth phase by the nonlinear least-squares regression method and were fit to Tamiya’s and Steele’s models. On the other hand, the effect of light was incorporated in the logistic model by fitting the entire growth curves at different light intensities to the logistic model by the

153 non-linear least-squares regression method, and the obtained µu values were fit to

Tamiya’s and Steele’s models.

For a deep microalgal culture, light varies significantly over the depth of the culture. The overall light attenuation in a microalgae culture column depends on water color, suspended solids, species, and pigment content of the microalgal biomass. It is a common practice to express the light attenuation as a function of chlorophyll content, but for a single species it is reasonable to express the light attenuation in terms of biomass

TSS; as the chlorophyll to TSS ratio is relatively constant (Benson & Rusch, 2006; Cerco

& Cole. 1995; Grima et al., 1994).

Beer-Lambert’s law (Equation 6-9), which is widely applied in the microalgal growth models, states that light is attenuated exponentially with depth in a microalgae culture column. Moreover, the penetration depth of the light in the culture column is inversely related to the biomass concentration (Benson & Rusch, 2006; Cornet et al.,

1992; Huesemann et al., 2013; Richmond, 2014).

−퐾.푧 퐼 = 퐼표 푒 Equation 6-9

2 Where I is the light intensity at depth z (µmol/m /s), Io is the incident light intensity at the surface (µmol/m2/s), and K is the attenuation coefficient (m-1) which encompasses the attenuation caused by water, algal biomass, and other suspended solids

(Benson & Rusch, 2005; Cerco & Cole, 1995). Although Beer-Lambert’s law assumes negligible scattering compared to absorption, it still can be applied for microalgae concentrations up to 3 g/L (Huesemann et al., 2013).

154

For the determination of light attenuation coefficients, seven biomass concentrations of the microalga S. dimorphus (0.941, 0.681, 0.452, 0.180, 0.092, 0.048, and 0.027 g/L), five concentrations of the AD supernatant (0.173, 0.114, 0.078, 0.037, and 0.012 g/L), and four solutions of combined microalgae and AD supernatant (0.034 g/L microalgae & 0.103 g/L AD, 0.086 g/L microalgae & 0.058 g/L AD, 0.173 g/L microalgae & 0.019 g/L AD, and 0.259 g/L microalgae & 0.003 g/L AD) were tested.

Corning® rectangular canted neck cell culture flasks were used (path length = 0.03 m) and for each biomass/AD concentration, the flasks were stacked starting with one flask until reaching six (path length = 0.18 m). The attenuation coefficients were corrected to account for the attenuation caused by flasks (0.54 m-1). Light was supplied by 760W

UFO LED Grow Light providing red, blue, orange, and white colors light orthogonal to the surface of the flasks. The incident surface light intensities measured at the surface of the top flask ranged from 300 to 730 μmol/m2/s and light intensity was measured underneath the bottom flask using an LI-COR 190 quantum sensor, which was connected to LI-250 A light meter. The light attenuation coefficients were determined according to

Beer-Lambert’s law using the nonlinear least-squares regression method.

The light intensity determined by the attenuation model was integrated over the depth (d) of the pond in order to determine the average light intensity over the depth as shown in Equation 6-10 (Huesmann et al., 2013):

푑 퐼표 퐼푎푣푔 = ∫ 퐼(푧). 푑푧 Equation 6-10 푑 0

155

6.2.3 Nutrient medium suspended solids concentration

It is important to model the variation of the nutrient medium turbidity in the microalgae cultures over time. Since the nutrient medium used in the previous experiments was unsterilized anaerobic digestate supernatant; it is expected that the bacteria in the nutrient medium produce extracellular enzymes that hydrolyze the particulate matter into soluble substrate (Tchobanoglous et al., 2004). Therefore, an experiment analogous to the vials experiment in Chapter 4 was conducted by testing 0.3 –

5% AD supernatant dilutions (10 – 130 mg/L TSS concentrations) but without microalgae addition. These were the dilutions that were likely to be used for microalgae cultivation on a larger scale. Each dilution was tested in triplicate by mixing 4 mL of the diluted supernatant with 2 mL DI water. The vials were placed in an incubator at 25 °C and under an average illumination of 50 μmol/m2/s illumination at 14:10 light/dark cycle, and the TSS concentration was monitored over time by measuring OD 680. The overall hydrolysis of the particulate substrate has been traditionally modeled as a first-order reaction with respect to the particulate substrate concentration as shown in Equation 6-11

(Luo et al., 2012; Tomei et al., 2008):

푑푃 = −푘 푃 Equation 6-11 푑푡 푝

Where:

P = particulate matter concentration (mg/L).

-1 kp = hydrolysis rate for the particulate matter (d ), which corresponds to the slope of the ln(P) vs. t line.

156

6.3 Results and Discussion

6.3.1 Light intensity effect

The time course of S. dimorphus biomass concentration under different light intensities is shown in Figure 6.1. It is clear that both the maximum biomass concentration and specific growth rate increased by increasing the light intensity from 25 to 388 µmol/m2/s. The maximum biomass concentration observed was 777.3 mg/L after 8 d of cultivation at 388 µmol/m2/s. Similarly, the highest specific growth rate (for the exponential growth phase) of 0.838 d-1 was observed at the same light intensity

(Table 6.1).

388 311 245 190 80 25 800

600

400 TSS TSS (mg/L)

200

0 0 2 4 6 8 Time (d)

Figure 6‎ .1: Biomass growth of S. dimorphus at various light intensities in µmol/m2/s

157

Table 6.1: Specific Growth Rates Corresponding to Each Light Intensity (Exponential Model) Light intensity (μmol/m2/s) μ (d-1) 388 0.838 ± 0.011 311 0.808 ± 0.005 245 0.792 ± 0.012 190 0.766 ± 0.011 80 0.615 ± 0.005 25 0.306 ± 0.004

A relationship was established between the light intensity and the specific growth rate according to Tamiya’s model (no photoinhibition) and Steele’s model

(photoinhibition). The model parameters were calculated by the nonlinear least-squares

-1 regression method. Tamiya’s model parameters were µm = 0.944 d and KI = 47.19

2 2 -1 µmol/m /s with an R value of 0.99 while Steele’s model parameters were µm = 0.853 d

2 2 and Im = 245.74 µmol/m /s with an R value of 0.94. According to Tamiya’s model, the specific growth rate increases with an increase in the light intensity until reaching an asymptotic value; hence, there is no photoinhibition at high light intensities (Figure 6.2).

On the other hand, Steele’s model predicts a maximum growth rate of 0.853 d-1 at 245.74

µmol/m2/s light intensity, and the growth rate decreases with an increase in the light intensity beyond this point. Since there were no signs of photoinhibition in our experiment, it was concluded that Tamiya’s model was a better fit to our data and was used further for the microalgal growth model.

158

Steele Tamiya Measured 0.9

0.8 ) 1 - 0.7 0.6 0.5 0.4 0.3 0.2 Specific growthrate(d 0.1 0.0 0 100 200 300 400 μmol/m2/s

Figure 6‎ .2: Specific growth rate (exponential model) as a function of light intensity (measured vs. predicted)

The data from Figure 6.1 were also modeled using the logistic model. The model was fit initially to each light intensity using the nonlinear least-squares regression

2 method, and the highest Cm value recorded was 747 mg/L at the 388 µmol/m /s light intensity culture. This value was then assigned to all light intensities, and the μu values were recalculated for each light intensity to correspond with the different logistic model form as shown in Table 6.2. Equation 6-12 was obtained by fitting Tamiya’s model to the calculated μu values. Figure 6.3 shows the measured and predicted μu values by

Equation 6-12.

159

Table 6.2: Specific Growth Rates Corresponding to Each Light Intensity According to the Logistic Model 2 -1 Light intensity (μmol/m /s) μu (d ) 388 1.248 ± 0.008 311 1.135 ± 0.052 245 1.058 ± 0.083 190 0.970 ± 0.043 80 0.648 ± 0.009 25 0.350 ± 0.011

퐼 휇 = 1.53 Equation 6-12 푢 퐼 + 103.4

Predicted Measured 1.4

) 1.2 1 - 1.0

0.8

0.6

0.4

Specific growthrate(d 0.2

0.0 0 100 200 300 400 μmol/m2/s

Figure ‎6.3: Measured and predicted μu values for the logistic model according to Tamiya’s model

The measured biomass concentrations at the different light intensities were plotted against the predicted concentrations by the logistic model as shown in Figure 6.4. It is clear that the logistic model fits the 388, 311, and 245 μmol/m2/s measured data well with

0.99 R2 values. However, the concentrations predicted by the model exceeded those

160 measured towards the end of cultivation period in the 190 and 80 μmol/m2/s cultures with

R2 values of 0.97 and 0.92, respectively, whereas the measured concentrations towards the end of cultivation in the 25 μmol/m2/s culture were higher than the predicted ones even though the R2 value was 0.99.

161

388 µmol/m2/s 311 µmol/m2/s

Measured Predicted Measured Predicted

1,000 1,000 750 750 500 500 250 250 0 0 Concentration (mg/L) 0 2 4 6 8 Concentration (mg/L) 0 2 4 6 8 Time (d) Time (d)

245 µmol/m2/s 190 µmol/m2/s

Measured Predicted Measured Predicted

1,000 1,000 750 750 500 500 250 250 0 0 Concentration (mg/L) 0 2 4 6 8 Concentration (mg/L) 0 2 4 6 8 Time (d) Time (d)

80 µmol/m2/s 25 µmol/m2/s

Measured Predicted Measured Predicted

750 500

500 250 250

0 0 Concentration (mg/L) 0 2 4 6 8 Concentration (mg/L) 0 2 4 6 8 Time (d) Time (d)

Figure ‎6.4: Measured and predicted biomass concentrations by the logistic model at different light intensities

162

6.3.2 Light attenuation

As indicated earlier, light was measured at the surface and the bottom of each flask or stack of flasks for each microalga biomass/AD concentration and at different depths. Table 6.3 presents the attenuation coefficients for each microalgae biomass concentration. It is clear that the attenuation coefficient depends on the biomass concentration, so that there is a not a single coefficient that represents the entire biomass range tested. As a result, it was necessary to correlate the light attenuation coefficient with the biomass concentration (Figure 6.5) and use this correlation in the overall attenuation formula.

Table 6.3: Light Attenuation Coefficients (K) for Different Biomass Concentrations Biomass concentration (g/L) Light attenuation coefficient (1/m) 0.941 175 0.681 150 0.452 107 0.180 50 0.092 27 0.048 16 0.027 10

163

200

y = 198.942x0.829 160 R² = 0.998

120

40 Light attenuation coefficient(1/m) 0 0.0 0.2 0.4 0.6 0.8 1.0 B (g/L)

Figure 6‎ .5: Correlation between microalgae biomass concentration and light attenuation coefficient

The correlation obtained in Figure 6.5 was used to determine the light intensity at any depth inside the bioreactor as shown in Equation 6-13:

−199 퐶0.829 푧 퐼 = 퐼표 푒 Equation 6-13

Where C is the biomass concentration and z is the depth. The measured I/Io for the different biomass concentrations are illustrated in Figure 6.6, where the dotted lines represent the predicted I/Io for each biomass concentration using Equation 6-13. Note that all of the curves were modeled with a single equation and a single set of parameters.

Benson and Rusch (2006) found that the attenuation coefficient for the microalga

Selenastrum capricornutum was a function of the biomass TSS, and that the relationship was linear. Other researchers indicated that the relationship between the biomass concentration and the light attenuation coefficient was hyperbolic, especially at biomass

164 concentrations higher than 350 mg/L (Benson & Rusch, 2006). Grima et al. (1994) reported that the overall absorption coefficient was variable when testing different biomass concentrations of the microalga Isochrysis glabana, which could be attributed to both the dry biomass weight as well as the pigment content. They concluded that the light attenuation coefficient is linearly related to the pigment content of the biomass. In our experiments, the power function provided the best fit for the light attenuation coefficient as a function of biomass concentration; and hence, it was used in the final formula.

0.941 g/L 0.681 g/L 0.452 g/L 0.180 g/L 0.092 g/L 0.048 g/L 0.027 g/L 0.8

0.6

0.4 I/Io

0.2

0.0 0.00 0.05 0.10 0.15 0.20 Depth (m)

Figure 6‎ .6: Light attenuation coefficient measured values (markers) vs. predicted values (dotted lines) for the microalga S. dimorphus

A similar approach was followed for the determination of the attenuation coefficient in the AD supernatant. The light attenuation coefficients were determined for

165 each concentration as shown in Table 6.4. A linear function was used to correlate the AD concentration and the attenuation coefficients as shown in Figure 6.7.

Table 6.4: Light Attenuation Coefficients (K) for Different AD TSS Concentrations AD Supernatant Concentration P (g/L) Light Attenuation Coefficient (1/m) 0.173 50 0.114 35 0.078 26 0.037 15 0.012 9

60 y = 253.505x + 6.093 50 R² = 1.00

10 Light attenuation coefficient(1/m) 0 0.00 0.05 0.10 0.15 0.20 TSS (mg/L)

Figure 6‎ .7: Correlation between AD concentration and light attenuation coefficient

The light attenuation equation for the AD supernatant can be expressed as shown in Equation 6-14:

−(254 푃+6.09) 푧 퐼 = 퐼표 푒 Equation 6-14

166

Where P is the AD supernatant TSS concentration. Figure 6.8 shows the measured (markers) and predicted (dotted lines) I/Io for the different AD concentrations using Equation 6-14. Note that a single equation and a single set of parameters were used to match all curves in Figure 6.8.

0.183 g/L 0.106 g/L 0.064 g/L 0.024 g/L 0.005 g/L 0.8

0.6

0.4 I/Io

0.2

0.0 0.00 0.05 0.10 0.15 0.20 Depth (m)

Figure ‎6.8: Light attenuation coefficient measured values (markers) vs. predicted values (dotted lines) for the AD supernatant

For the combined effect of microalgae and AD supernatant, four solutions were prepared by mixing predetermined portions of microalgae and AD TSS as shown in

Table 6.5. The attenuation coefficients for microalgae and AD supernatant were determined separately using Equation 6-13 and Equation 6-14 and were correlated with the actual measured attenuation coefficient of the combination and the factors were determined using linear regression as shown in Equation 6-15.

167

퐾 푐표푚푏𝑖푛푒푑 = 0.93 퐾 푎푙푔푎푒 + 0.87 퐾 퐴퐷 Equation 6-15

Table 6.5: Combined Light Attenuation Coefficients (K) for Various Microalgae and AD Concentrations (Actual vs. Predicted Values) Microalgae K microalgae AD TSS K AD K measured K regression (g/L) (1/m) (g/L) (1/m) (1/m) (1/m) 0.034 12.2 0.103 32.1 39.9 39.3 0.086 26.1 0.058 20.9 41.8 42.5 0.173 46.4 0.019 10.8 51.4 52.6 0.259 65.0 0.003 6.9 67.4 66.4

The combined attenuation equation was integrated over the depth of the pond (d) in order to determine the average light intensity over the depth as shown in Equation 6-16 and Equation 6-17:

푑 퐼표 −[185.02 퐶0.829+220.55 푇푆푆+5.3]푧 퐼푎푣푔 = ∫ 푒 . 푑푧 Equation 6-16 푑 0

−[185.02 퐶0.829+220.55 푇푆푆+5.3]푑 −퐼표(푒 − 1) 퐼 = Equation 6-17 푎푣푔 푑(185.02 퐶0.829 + 220.55 푇푆푆 + 5.3)

6.3.3 Nutrient medium suspended solids concentration

The hydrolysis of the particulate matter in the nutrient media was represented by a first-order reaction. TSS concentrations were monitored in the vials containing AD dilutions with no microalgae added. The first-order rate (kp) was determined as the slope of the ln(P) vs. t for the 10 – 130 mg/L concentrations . It was found that the kp values ranged from 0.021 to 0.030 d-1 with no clear relationship between the TSS concentration and the value of kp. Tomei et al. (2008) reported that the first-order rate of anaerobic

168 degradation of the particulate matter in the sewage sludge was in the 0.024 – 0.039 d-1 range, and that the food/inoculum ratio did not have a significant impact on the rate. On the other hand, Luo et al. (2012) stated that literature values for the first-order hydrolysis rate of the primary sludge ranged from 0.169 to 0.4 d-1 whereas for the waste activated sludge, the values ranged from 0.017 to 0.8 d-1. As a result, an average value of 0.0255 d-

1 based on the range obtained in our experiment was used to model the hydrolysis of the particulate matter in the nutrient media as shown in Figure 6.9 where the markers represent the measured concentrations and the dotted lines represent the predicted concentrations. Note that all of the curves were modeled with a single equation and

-1 parameter (kp = 0.0255 d ).

140 130 mg/L 100 mg/L 66 mg/L 35 mg/L 18 mg/L 10 mg/L

120

100

60 TSS TSS (mg/L)

0 0 5 10 15 20 25 30 Time (d)

Figure 6.9: Nutrient media TSS (measured vs. predicted)

169

6.3.4 Models validation

The models developed using the previous bench-scale experiments were validated by comparing the measured biomass concentrations of S. dimorphus grown in a 100 L raceway pond with the concentrations predicted by both models. Both the exponential growth model (Equation 6-3) and the logistic growth model (Equation 6-6) were tested using the average light intensity model (Equation 6-17). Figure 6.10 shows the measured

(markers) and predicted (dotted lines) biomass concentrations when the microalga S. dimorphus was cultivated using BG-11 medium in the raceway pond (Chapter 5). The average incident water surface light intensity was 454 µmol/m2/s. Temperature and pH were maintained at 25 °C and 7.5, respectively, throughout the cultivation period. All parameters used in the model were obtained from the bench-scale experiments as indicated earlier. It is clear that the logistic model provides a better fit to the growth data

(R2 = 0.98) compared to the exponential model (R2 = 0.80) (Figure 6.11). This is in agreement with the findings of Yaoyang and Boeing (2014) who reported that the logistic model provided significantly better fit to the growth of the genus Nannochloropsis compared to the exponential model. This can be explained by the fact that the growth under the logistic model is limited by the maximum bearing capacity of the culture (Cm).

On the other hand, under the exponential model the biomass growth will continue; since there is no maximum biomass concentration, and even though the specific growth rate decreases over time due to light limitation (and nutrients if not sufficient), the value of µ will always be greater than zero; thus, the growth continues.

170

Exponential/ BG-11 (454 µmol/m2/s) Logistic/ BG-11 (454 µmol/m2/s)

Predicted Measured Predicted Measured

3000 800

600 2000 400 1000 200

0 0 Concentration (mg/L) 0 2 4 6 8 10 12 14 16 Concentration (mg/L) 0 2 4 6 8 10 12 14 16 Time (d) Time (d)

Figure ‎6.10: Measured and predicted time course biomass concentrations when the microalga S. dimorphus was cultivated in the raceway pond using BG-11 medium

Exponential/ BG-11 (454 µmol/m2/s) Logistic / BG-11 (454 µmol/m2/s) y = x 3000 y = x 800

2500 R² = 0.8032 600 2000 R² = 0.9762 1500 400 1000 200 Predicted (mg/L) 500 Predicted (mg/L) 0 0 0 1000 2000 3000 0 200 400 600 800 Measured (mg/L) Measured (mg/L)

Figure ‎6.11: Measured and predicted biomass concentrations when the microalga S. dimorphus was cultivated in the raceway pond using BG-11 medium

The models were then tested on the microalgae cultivated using AD supernatant.

As indicated in Chapter 4, the microalga S. dimorphus was cultivated using 2.5% and

1.25% diluted AD supernatant, that is corresponding to 93 and 46 mg/L initial TSS concentrations, respectively, in the raceway ponds at 25 °C temperature and pH 7.5, while the light intensities were 454 and 317 µmol/m2/s for the 2.5% dilution and 384 and

171

234 µmol/m2/s for the 1.25% dilution. Clearly, the exponential model did not provide a good fit to any of the growth data with R2 = 0.63, 0.73, 0.58, and 0.40 for the 2.5% – 454

µmol/m2/s, 2.5% – 317 µmol/m2/s, 1.25% – 384 µmol/m2/s, and 1.25% – 234 µmol/m2/s cultures, respectively (Figure 6.12 and Figure 6.13). Even though the predicted biomass concentrations were close to the measured concentrations in the first 4 – 5 d (all cultures except the 2.5% – 454 µmol/m2/s), the predicted values deviated significantly from the measured concentrations regardless of the dilution afterwards. This in part can be explained by nutrients limitation in the 1.25% cultures as indicated in Chapter 4; however, this is not the sole reason for the model deviation.

172

2.5% dilution (454 µmol/m2/s) 2.5% dilution (317 µmol/m2/s)

Predicted Measured Predicted Measured

1200 1200

900 900

600 600

300 300

0 0 Concentration (mg/L) 0 2 4 6 8 10 12 Concentration (mg/L) 0 2 4 6 8 10 12 Time (d) Time (d)

1.25% dilution (384 µmol/m2/s) 1.25% dilution (234 µmol/m2/s)

Predicted Measured Predicted Measured

1500 1000 1200 800 900 600 600 400 300 200 0 0 Concentration (mg/L) 0 2 4 6 8 10 12 14 Concentration (mg/L) 0 2 4 6 8 10 12 Time (d) Time (d)

Figure ‎6.12: Measured and predicted time course biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant

173

2.5% dilution (454 µmol/m2/s) 2.5% dilution (317 µmol/m2/s)

1200 1200 R² = 0.7265 R² = 0.6337 y = x y = x 900 900

600 600

300 300 Predicted (mg/L) Predicted (mg/L)

0 0 0 300 600 900 1200 0 300 600 900 1200 Measured (mg/L) Measured (mg/L)

1.25% dilution (384 µmol/m2/s) 1.25% dilution (234 µmol/m2/s)

1500 y = x 900 y = x

1200 R² = 0.5799 600 900 R² = 0.4001

600 300

Predicted (mg/L) 300 Predicted (mg/L)

0 0 0 300 600 900 1200 1500 0 300 600 900 Measured (mg/L) Measured (mg/L)

Figure 6‎ .13: Measured and predicted biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant

The logistic model representation of the microalgae cultivated using the AD supernatant is presented in Figure 6.14. Clearly, the logistic model was more realistic in estimating the biomass concentrations. For instance, the logistic model predicted the biomass concentrations well in the case of 2.5 – 317 µmol/m2/s culture with an R2 value of 0.87 (Figure 6.15). On the other hand, the measured concentrations of the 2.5% – 454

µmol/m2/s culture exceeded the predicted values in the first 6 d but the concentrations towards the end of cultivation time were close. For the 1.25% dilution, the model

174 overestimated the stationary phase biomass concentrations even though the measured and predicted values were close in the first 4 d. This suggests that for more accurate determination of the microalgae growth, the maximum biomass concentration Cm in the logistic model has to be varied based on the nutrients concentrations in the media; since the logistic model does not account for sub-optimal nutrients conditions. It was attempted to fit the logistic model to the data from the vials experiment in Chapter 4 in order to express the maximum biomass concentration as a function of dilution; however, the values obtained did not produce a good fit to the experimental data in this chapter.

175

2.5% dilution (454 µmol/m2/s) 2.5% dilution (317 µmol/m2/s)

Predicted Measured Predicted Measured

600 600

400 400

200 200

0 0 Concentration (mg/L) 0 2 4 6 8 10 12 Concentration (mg/L) 0 2 4 6 8 10 12 Time (d) Time (d)

1.25% dilution (384 µmol/m2/s) 1.25% dilution (234 µmol/m2/s)

Predicted Measured Predicted Measured

600 600

400 400

200 200

0 0 Concentration (mg/L) 0 2 4 6 8 10 12 14 Concentration (mg/L) 0 2 4 6 8 10 12 Time (d) Time (d)

Figure ‎6.14: Measured and predicted time course biomass concentrations using the logistic model when the microalga S. dimorphus was cultivated using AD supernatant

176

2.5% dilution (454 µmol/m2/s) 2.5% dilution (317 µmol/m2/s)

600 500 R² = 0.816 y = x y = x

400 R² = 0.8708 400 300

200 200

Predicted (mg/L) Predicted (mg/L) 100

0 0 0 200 400 600 0 100 200 300 400 500 Measured (mg/L) Measured (mg/L)

1.25% dilution (384 µmol/m2/s) 1.25% dilution (234 µmol/m2/s)

600 y = x 500 y = x

R² = 0.762 400 R² = 0.4784 400 300

200 200

Predicted (mg/L) Predicted (mg/L) 100

0 0 0 200 400 600 0 100 200 300 400 500 Measured (mg/L) Measured (mg/L)

Figure 6‎ .15: Measured and predicted biomass concentrations using the exponential model when the microalga S. dimorphus was cultivated using AD supernatant

In summary, the logistic model provides more realistic biomass concentration estimates of the microalga S. dimorphus than the exponential model; however, in order to increase the accuracy, the maximum biomass concentration (Cm) has to be varied based on the nutrients availability. Furthermore, the effect of light intensity on the growth of S. dimorphus was evaluated based on a bench-scale experiment using BG-11 nutrient medium, which may not be representative to all nutrient media such as the AD supernatant. Hence, further investigation of the effect of light intensity on the growth of

177

S. dimorphus using a variety of nutrient media is necessary. On the other hand, the exponential model can be useful in the case of continuous or semi-continuous microalgal cultivation where biomass is harvested and nutrient medium is introduced continuously unlike batch cultivation where nutrients are added in the beginning of cultivation.

6.4 Conclusions

In this chapter, two models were investigated in order to predict the growth of the microalga S. dimorphus using BG-11 medium and AD supernatant as nutrient media. It was found that the logistic model provided more realistic estimates of the biomass concentrations especially in the case of BG-11 nutrient medium. For a more accurate prediction of biomass concentrations, the logistic model maximum biomass concentration has to be varied based on the nutrients availability. Furthermore, the effect of light intensity on the growth rate of S. dimorphus has to be investigated using different types of nutrient media.

178

CHAPTER 7. CONCLUSIONS AND RECOMMENDATIONS

The main objective of this dissertation was to evaluate the sustainable cultivation of microalgae using anaerobic digestate as a nutrient medium. Several studies in the literature have utilized anaerobic digestate as a nutrient medium to cultivate microalgae; however, most of these studies were conducted at the bench-scale. Furthermore, these studies focused on pretreatment steps for anaerobic digestate prior to microalgae cultivation such as filtration, centrifugation, and autoclaving. In our study, the anaerobic digestate was brought from a commercial digester which typically accepts organic wastes such as animal manure and food waste. Several pretreatment methods that were not addressed in the literature such as hydrogen peroxide treatment, filtration using polyester filter bags, and supernatant extraction methods were applied. It was found that diluting the digestate and using the supernatant after allowing the mixture to settle was the simplest and least expensive pretreatment method. The supernatant was collected after settling of the diluted digestate to reduce the turbidity and the organic and inorganic matter while still maintaining sufficient nutrients (particularly nitrogen) to support the growth of microalgae.

The growth of microalgae using the anaerobic digestate supernatant was assessed initially by a bench-scale experiment where the microalga Neochloris oleoabundans was cultivated using a wide range of supernatant and filtered dilutions. It was found that the highest growth was achieved using the 2.29% dilution, which was equivalent to 100 mg

N/L. There were several attempts to scale-up the cultivation of N. oleoabundans to a 100

L raceway pond using the 2.29% diluted digestate, but the culture was contaminated with

179 other algal species such as Scenedesmus and cyanobacteria. It was attempted to filter off the invasive species using 10 and 5 μm filter bags; however, this technique was only useful for removing cyanobacteria cells while Scenedesmus cells were still dominant in the filtrate. As a result, we elected to test the growth of the microalga Scenedesmus dimorphus using the anaerobic digestate as a nutrient medium.

The microalga Scenedesmus dimorphus was cultivated on a bench-scale using a range of the anaerobic digestate supernatant dilutions. The results indicated that 1.25 –

2.5% dilutions provided sufficient nutrients to maximize the growth rate and produce relatively high biomass concentrations. The microalgae cultivation was then scaled-up to

100 L raceway ponds. Two dilutions were selected based on the bench-scale experiment and two light intensities were tested for each dilution. For the 2.5% cultures, the average incident light intensities were 317 and 454 μmol/m2/s; whereas for the 1.25% cultures, the average incident light intensities were 234 and 384 μmol/m2/s. The difference in the light intensity between the two cultures was to account for the difference in the turbidity of the nutrient media. It was found that increasing the light intensity improved the growth of the microalga S. dimorphus significantly regardless of the dilution. Furthermore, it was found that nutrients availability had a pronounced effect on the growth of this microalga in the raceway ponds. This was clear in the 1.25% cultures where the growth slowed down after 3 – 4 d when the nitrogen and phosphorus concentrations were below 30 mg

N/L and 3 mg P/L, respectively. The highest biomass concentration recorded was 432 mg/L which was achieved after 10 d of cultivation in the 2.5% – 454 μmol/m2/s culture.

The highest biomass productivity during the exponential growth phase was 63.6 mg/L/d

180 for the 2.5% – 454 μmol/m2/s culture. Finally, nutrients were monitored frequently in order to assess the removal efficiencies by the microalga S. dimorphus. It was found that nitrogen was removed at 65 – 72% efficiencies whereas ammonia was completely removed in all cultures. Phosphorus removal efficiencies were in the 63 – 100% range and the COD removal efficiencies were in the 78 – 82% range.

Tracer tests were conducted in the raceway ponds by testing the flow at three different water surface velocities. The purpose of those tests was to determine several flow parameters such as the actual average flow velocities corresponding to different surface velocities in the 0.1 – 0.3 m/s range. Furthermore, energy losses as a result of the friction across the straight channel and the flow around the bends have been estimated and the power required to operate the ponds at different velocities was computed. Then the effect of mixing on the microalgal growth was assessed by cultivating the microalga

S. dimorphus in the raceway ponds at 0.1, 0.2, and 0.3 m/s water surface velocities. It was concluded that cultivating the microalgae at 0.3 m/s improved the biomass growth rates and concentrations compared to the 0.1 and 0.2 m/s velocities. Based on the energy losses and power calculations, it was found that cultivating the microalgae at 0.1 m/s velocity was the only scenario that produced a positive net energy balance compared to the 0.2 and 0.3 m/s velocities, if biodiesel is to be produced. However, this balance was based on assumptions that can vary significantly based on the species and the cultivation conditions.

Finally, two biomass growth models were developed and tested on the microalga

S. dimorphus. The first model was the exponential model which assumes that microalgae

181 grow exponentially at the maximum specific growth rate under optimal culture conditions such as temperature, light, and nutrients. However, microalgae grow at specific growth rates that are less than the maximum specific growth rate under sub-optimal culture conditions. The second model used was the logistic model, which has gained popularity recently in simulating the microbial growth in batch cultures. Lag, exponential, and stationary phases are well represented by the logistic model unlike the exponential model; however, the logistic model is independent of nutrients concentrations or other suboptimal conditions. It was found that the logistic model provided more realistic estimates of the biomass concentrations when the microalga S. dimorphus was cultivated using BG-

11 or AD supernatant.

The combination of anaerobic digestion and microalgae cultivation is an attractive solution for biofuels production. The anaerobic digestion produces biogas while the COD of the organic waste is reduced. A nutrient-rich digestate is also generated, which is widely used as a fertilizer; however, it was proven that this digestate can be used effectively as a nutrient medium for multiple microalgae species. Combining the anaerobic digestion and microalgal cultivation has other promising potentials such as scrubbing the carbon dioxide from the biogas using microalgae and the anaerobic digestion of the microalgae biomass to produce biogas. Hence, it is recommended to design and construct a pilot plant that combines anaerobic digestion and microalgae cultivation for sustainable housing. The system will include an anaerobic digester that accepts household organic municipal solid wastes such as food waste, yard trimmings, and paper waste. The biogas from the anaerobic digester can be combusted directly to

182 produce heat and electricity for the household, whereas the digestate flowing out of the digester is diluted with water which does not necessarily have to be fresh water, and the diluted digestate will be directed to a settling tank where the supernatant flows to a microalgae pond. Furthermore, the flue gas from the biogas combustion can be utilized as a source of carbon dioxide for the microalgae culture. It is also recommended to place these elements in a greenhouse in order to provide more control on the culture conditions such as temperature as well as to reduce the contamination risk by other invasive species.

Other methods for reducing the contamination risk include cultivating species with low contamination risk such as Dunaliella, Chlorella, and Spirulina. Allowing native invasive species to take over will also reduce the contamination risk; since these species are acclimated to the local conditions.

Light intensity was proven to be a key factor in the cultivation of microalgae in raceway ponds. It was shown in the cultivation of S. dimorphus in the raceway ponds that light was limiting towards the end of cultivation period; hence, summer cultivation in a greenhouse where higher light intensities are available is an attractive option; besides the fact that it is more sustainable to rely on solar energy. Also, it is recommended to cultivate the microalga S. dimorphus on a continuous or semi-continuous basis using

1.25% AD dilution (50 mg N/L) at 384 µmol/m2/s light intensity or higher. However, for batch cultivation, the 2.5% dilution should be used with 454 µmol/m2/s light intensity or higher in order to ensure nutrient sufficiency.

Further research has to be conducted on sustainable harvesting techniques such as bio-flocculation or studying species that are self-flocculating; since harvesting the

183 microalgal biomass makes up a considerable portion of the biodiesel production cost.

Additionally, mixing is necessary for the successful cultivation of microalgae. Although operating the pond at 0.2 and 0.3 m/s water surface velocities yielded higher growth rates and biomass concentrations in our experiments compared to the 0.1 m/s, the net energy balance was positive only in the case of the 0.1 m/s velocity. This balance was based on assumptions such as 25% lipid content and 80% lipid extraction efficiency; therefore, if the lipid content increases, it may be feasible to operate the ponds at higher velocities.

Nevertheless, it is recommended to investigate the effect of variable mixing velocities; where the velocity is higher during the daytime to maximize the photosynthesis, while the velocity is decreased during the night to reduce energy costs. Additionally, the effect of mixing has to be coupled with other cultivation conditions such as nitrogen availability; since this has a significant effect on the biomass growth rate as well as the lipid content which in turn affects the energy balance.

Finally, modeling the microalgal growth is necessary in order to optimize both biomass and lipid productivities. The exponential model can be useful in representing the biomass growth during the exponential growth phase; however, it tends to predict biomass concentrations that are significantly higher than the actual concentrations during the stationary phase. Therefore, this model is better used when a continuous or semi- continuous cultivation is adopted. On the other hand, the logistic model was found to better represent the microalgal growth in our experiments. However, since this model is independent of the nutrients concentrations, the model parameters, particularly the maximum biomass concentration (Cm), have to be determined for different nutrient

184 concentrations. Furthermore, it is recommended to investigate the effect of light intensity on the microalgal growth using multiple nutrient media such BG-11 enriched with ammonia or filtered AD in order to obtain specific growth rates that better represent the microalgal growth using the AD supernatant.

185

REFERENCES

Abu Hajar, H. A., Riefler, R. G., & Stuart, B. J. (2016). Anaerobic digestate as a nutrient

medium for the growth of the green microalga Neochloris oleoabundans. Environ.

Eng. Res. In press.

Acevedo, M. F. (2012). Simulation of Ecological and Environmental Models. CRC Press.

Amin, S. (2009). Review on biofuel oil and gas production processes from microalgae.

Energ Convers Manage, 50(7), 1834-1840.

Andrews, J. F. (1968). A mathematical model for the continuous culture of

microorganisms utilizing inhibitory substrates. Biotechnol Bioeng, 10(6), 707-

723.

Arumugam, M., Agarwal, A., Arya, M. C., & Ahmed, Z. (2013). Influence of nitrogen

sources on biomass productivity of microalgae Scenedesmus bijugatus.

Bioresource Technol, 131, 246-249.

Balat, M., Balat, H., & Öz, C. (2008). Progress in bioethanol processing. Prog Energ

Combust, 34(5), 551-573.

Baldisserotto, C., Giovanardi, M., Ferroni, L., & Pancaldi, S. (2014). Growth,

morphology and photosynthetic responses of Neochloris oleoabundans during

cultivation in a mixotrophic brackish medium and subsequent starvation. Acta

Physiol Plant, 36(2), 461-472.

Barnhart, M. C. (2006). Buckets of muckets: A compact system for rearing juvenile

freshwater mussels. Aquaculture, 254(1), 227-233.

186

Beach, E. S., Eckelman, M. J., Cui, Z., Brentner, L., & Zimmerman, J. B. (2012).

Preferential technological and life cycle environmental performance of chitosan

flocculation for harvesting of the green algae Neochloris oleoabundans.

Bioresource Technol, 121, 445-449.

Beal, C. M., Webber, M. E., Ruoff, R. S., & Hebner, R. E. (2010). Lipid analysis of

Neochloris oleoabundans by liquid state NMR. Biotechnol Bioeng, 106(4), 573-

583.

Béchet, Q., Shilton, A., & Guieysse, B. (2013). Modeling the effects of light and

temperature on algae growth: State of the art and critical assessment for

productivity prediction during outdoor cultivation. Biotechnol Adv, 31(8), 1648-

1663.

Benider, A., Tahiri, M., Belkoura, M., & Dauta, A. (2001). Interaction des facteurs

héliothermiques sur la croissance de trois espèces du genre Scenedesmus. Ann.

Limnol., 37(4), 257-266.

Benson, B. C., & Rusch, K. A. (2006). Investigation of the light dynamics and their

impact on algal growth rate in a hydraulically integrated serial turbidostat algal

reactor (HISTAR). Aquacult Eng, 35(2), 122-134.

Bona, F., Capuzzo, A., Franchino, M., & Maffei, M. E. (2014). Semicontinuous nitrogen

limitation as convenient operation strategy to maximize fatty acid production in

Neochloris oleoabundans. Algal Res, 5, 1-6.

187

Brennan, L., & Owende, P. (2010). Biofuels from microalgae—a review of technologies

for production, processing, and extractions of biofuels and co-products. Renew

Sust Energ Rev, 14(2), 557-577.

Breuer, G., Lamers, P. P., Martens, D. E., Draaisma, R. B., & Wijffels, R. H. (2012). The

impact of nitrogen starvation on the dynamics of triacylglycerol accumulation in

nine microalgae strains. Bioresource Technol, 124, 217-226.

Carlsson, B. (1998). An introduction to sedimentation theory in wastewater treatment.

Systems and Control Group, Uppsala University.

Cerco, C. F., & Cole, T. (1995). User's Guide to the CE-QUAL-ICM Three-dimensional

Eutrophication Model: Release Version 1.0. US Army Engineer Waterways

Experiment Station.

Chen, Y., Wang, J., Zhang, W., Chen, L., Gao, L., & Liu, T. (2013). Forced light/dark

circulation operation of open pond for microalgae cultivation. Biomass Bioenerg,

56, 464-470.

Chen, C. Y., Yeh, K. L., Aisyah, R., Lee, D. J., & Chang, J. S. (2011). Cultivation,

photobioreactor design and harvesting of microalgae for biodiesel production: A

critical review. Bioresource Technol, 102(1), 71-81.

Chisti, Y. (2008). Biodiesel from microalgae beats bioethanol. Trends Biotechnol, 26(3),

126-131.

Choi, H. J., & Lee, S. M. (2012). Effects of microalgae on the removal of nutrients from

wastewater: Various concentrations of Chlorella vulgaris. Environ Eng Res,

17(S1), S3–S8.

188

Clesceri, L., Greenberg, A., & Eaton, A. (1998). Standard Methods for the Examination

of Water and Wastewater. American Public Health Association, Washington D.C.

Cornet, J. F., Dussap, C. G., & Dubertret, G. (1992). A structured model for simulation of

cultures of the cyanobacterium Spirulina platensis in photobioreactors: I.

Coupling between light transfer and growth kinetics. Biotechnol Bioeng, 40(7),

817-825.

Crittenden, J. C., Trussell, R. R., Hand, D. W., Howe, K. J., & Tchobanoglous, G. (2005).

Water Treatment Principles and Design. Hoboken, NJ: John Wiley & Sons. da Silva, T. L., Reis, A., Medeiros, R., Oliveira, A. C., & Gouveia, L. (2009). Oil

production towards biofuel from autotrophic microalgae semicontinuous

cultivations monitorized by flow cytometry. Appl Biochem Biotech, 159(2), 568-

578.

Davis, R. W., Volponi, J. V., Jones, H. D., Carvalho, B. J., Wu, H., & Singh, S. (2012).

Multiplex fluorometric assessment of nutrient limitation as a strategy for

enhanced lipid enrichment and harvesting of Neochloris oleoabundans.

Biotechnol Bioeng, 109(10), 2503-2512. de Winter, L., Klok, A. J., Franco, M. C., Barbosa, M. J., & Wijffels, R. H. (2013). The

synchronized cell cycle of Neochloris oleoabundans and its influence on biomass

composition under constant light conditions. Algal Res, 2(4), 313-320. de Winter, L., Schepers, L. W., Cuaresma, M., Barbosa, M. J., Martens, D. E., &

Wijffels, R. H. (2014). Circadian rhythms in the cell cycle and biomass

189

composition of Neochloris oleoabundans under nitrogen limitation. J Biotechnol,

187, 25-33.

Dincer, I. (2000). Renewable energy and sustainable development: a crucial review.

Renew Sust Energ Rev, 4(2), 157-175.

Forster, C. F. (1968). The surface of activated sludge particles in relation to their settling

characteristics. Water Res, 2(11), 767-776.

Franchino, M., Comino, E., Bona, F., & Riggio, V. A. (2013). Growth of three

microalgae strains and nutrient removal from an agro-zootechnical digestate.

Chemosphere, 92(6), 738-744.

Frigon, J. C., Matteau-Lebrun, F., Abdou, R. H., McGinn, P. J., O’Leary, S. J., & Guiot,

S. R. (2013). Screening microalgae strains for their productivity in methane

following anaerobic digestion. Appl Energ, 108, 100-107.

Garibay-Hernández, A., Vazquez-Duhalt, R., Serrano-Carreón, L., & Martinez, A.

(2013). Nitrogen limitation in Neochloris oleoabundans: A reassessment of its

effect on cell growth and biochemical composition. Appl Biochem Biotech,

171(7), 1775-1791.

Gatenby, C. M., Orcutt, D. M., Kreeger, D. A., Parker, B. C., Jones, V. A., & Neves, R. J.

(2003). Biochemical composition of three algal species proposed as food for

captive freshwater mussels. J Appl Phycol, 15(1), 1-11.

Giannetto, M. J., Retotar, A., Rismani-Yazdi, H., & Peccia, J. (2015). Using carbon

dioxide to maintain an elevated oleaginous microalga concentration in mixed-

culture photo-bioreactors. Bioresource Technol, 185, 178-184.

190

Giovanardi, M., Baldisserotto, C., Ferroni, L., Longoni, P., Cella, R., & Pancaldi, S.

(2014). Growth and lipid synthesis promotion in mixotrophic Neochloris

oleoabundans (Chlorophyta) cultivated with glucose. Protoplasma, 251(1), 115-

125.

Giovanardi, M., Ferroni, L., Baldisserotto, C., Tedeschi, P., Maietti, A., Pantaleoni, L., &

Pancaldi, S. (2013). Morphophysiological analyses of Neochloris oleoabundans

(Chlorophyta) grown mixotrophically in a carbon-rich waste product.

Protoplasma, 250(1), 161-174.

Goiris, K., Muylaert, K., Fraeye, I., Foubert, I., De Brabanter, J., & De Cooman, L.

(2012). Antioxidant potential of microalgae in relation to their phenolic and

carotenoid content. J Appl Phycol, 24, 1477-1486.

González, L. E., Cañizares, R. O., & Baena, S. (1997). Efficiency of ammonia and

phosphorus removal from a Colombian agroindustrial wastewater by the

microalgae Chlorella vulgaris and Scenedesmus dimorphus. Bioresource

Technol, 60(3), 259-262.

Goswami, R. C. D., & Kalita, M. C. (2011). Scenedesmus dimorphus and Scenedesmus

quadricauda: two potent indigenous microalgae strains for biomass production

and CO2 mitigation – A study on their growth behavior and lipid productivity

under different concentration of urea as nitrogen source. J. Algal Biomass

Utln, 2(4), 42-49.

Gour, R. S., Kant, A., & Chauhan, R. S. (2014). Screening of micro algae for growth and

lipid accumulation properties. J. Algal Biomass Utln, 5(1), 38–46.

191

Gouveia, L., Marques, A., da Silva, T., & Reis, A. (2009). Neochloris oleabundans

UTEX #1185: a suitable renewable lipid source for biofuel production. J Ind

Microbiol Biot, 36(6), 821-826.

Gouveia, L., & Oliveira, A. (2009). Microalgae as a raw material for biofuels production.

J Ind Microbiol Biot, 36(2), 269-274.

Green, F. B., Lundquist, T. J., & Oswald, W. J. (1995). Energetics of advanced integrated

wastewater pond systems. Wat. Sci. Tech, 31, 9-20.

Greenwell, H. C., Laurens, L. M. L., Shields, R. J., Lovitt, R. W., & Flynn, K. J. (2009).

Placing microalgae on the biofuels priority list: a review of the technological

challenges. J R Soc Interface, rsif20090322.

Grima, E. M., Camacho, F. G., Pérez, J. A., Sevilla, J. M., Fernandez, F. G., & Gomez,

A. C. (1994). A mathematical model of microalgal growth in light‐limited

chemostat culture. J Chem Technol Biot, 61(2), 167-173.

Hadiyanto, H., Elmore, S., Van Gerven, T., & Stankiewicz, A. (2013). Hydrodynamic

evaluations in high rate algae pond (HRAP) design. Chem Eng J, 217, 231-239.

Hosikian, A., Lim, S., Halim, R., & Danquah, M. K. (2010). Chlorophyll extraction from

microalgae: A review on the process engineering aspects. Int J Chem Eng, 2010.

Höök, M., & Tang, X. (2013). Depletion of fossil fuels and anthropogenic climate

change—A review. Energ Policy, 52, 797-809.

Huesemann, M. H., Van Wagenen, J., Miller, T., Chavis, A., Hobbs, S., & Crowe, B.

(2013). A screening model to predict microalgae biomass growth in

photobioreactors and raceway ponds. Biotechnol Bioeng, 110(6), 1583-1594.

192

James, D. E. (1978). Culturing algae. Carolina Biological Supply Company.

James, S. C., & Boriah, V. (2010). Modeling algae growth in an open-channel raceway. J

Comput Biol, 17, 895-906.

Jiang, Y., Zhang, W., Wang, J., Chen, Y., Shen, S., & Liu, T. (2013). Utilization of

simulated flue gas for cultivation of Scenedesmus dimorphus. Bioresource

Technol, 128, 359-364.

Kang, J., & Wen, Z. (2015). Use of microalgae for mitigating ammonia and CO2

emissions from animal production operations—Evaluation of gas removal

efficiency and algal biomass composition. Algal Res, 11, 204-210.

Kargi, F. (2009). Re‐interpretation of the logistic equation for batch microbial growth in

relation to Monod kinetics. Lett Appl Microbiol, 48(4), 398-401.

Kim, H. C., Choi, W. J., Maeng, S. K., Kim, H. J., Kim, H. S., & Song, K. G. (2014).

Ozonation of piggery wastewater for enhanced removal of contaminants by S.

quadricauda and the impact on organic characteristics. Bioresource Technol, 159,

128-135.

Klausmeier, C. A., Litchman, E., Daufresne, T., & Levin, S. A. (2004). Optimal nitrogen-

to-phosphorus stoichiometry of phytoplankton. Nature, 429(6988), 171-174.

Klok, A. J., Martens, D. E., Wijffels, R. H., & Lamers, P. P. (2013a). Simultaneous

growth and neutral lipid accumulation in microalgae. Bioresource Technol, 134,

233-243.

193

Klok, A. J., Verbaanderd, J. A., Lamers, P. P., Martens, D. E., Rinzema, A., & Wijffels,

R. H. (2013b). A model for customising biomass composition in continuous

microalgae production. Bioresource Technol, 146, 89-100.

Ksibi, M. (2006). Chemical oxidation with hydrogen peroxide for domestic wastewater

treatment. Chem Eng J, 119(2), 161-165.

Kunikane, S., Kaneko, M., & Maehara, R. (1984). Growth and nutrient uptake of green

alga, Scenedesmus dimorphus, under a wide range of nitrogen/phosphorus ratio—

I. Experimental study. Water Res, 18(10), 1299-1311.

Kunikane, S., & Kaneko, M. (1984). Growth and nutrient uptake of green alga,

Scenedesmus dimorphus, under a wide range of nitrogen/phosphorus ratio—II.

Kinetic model. Water Res, 18(10), 1313-1326.

Kwak, H. S., Kim, J. Y. H., & Sim, S. J. (2015). A microscale approach for simple and

rapid monitoring of cell growth and lipid accumulation in Neochloris

oleoabundans. Bioprocess Biosyst Eng, 38(10), 2035-2043.

Lam, M. K., & Lee, K. T. (2012). Microalgae biofuels: A critical review of issues,

problems and the way forward. Biotechnol Adv, 30(3), 673-690.

Lam, G.P. ‘t; Zegeye, E. K., Vermuë, M. H., Kleinegris, D. M. M., Eppink, M. H. M.,

Wijffels, R. H., & Olivieri, G. (2015). Dosage effect of cationic polymers on the

flocculation efficiency of the marine microalga Neochloris oleoabundans.

Bioresource Technol, 198, 797-802.

Lee, E., Jalalizadeh, M., & Zhang, Q. (2015). Growth kinetic models for microalgae

cultivation: A review. Algal Res, 12, 497-512.

194

Lenneman, E. M., Wang, P., & Barney, B. M. (2014). Potential application of algicidal

bacteria for improved lipid recovery with specific algae. FEMS Microbiol

Lett, 354(2), 102-110.

Levine, R. B., Costanza-Robinson, M. S., & Spatafora, G. A. (2011). Neochloris

oleoabundans grown on anaerobically digested dairy manure for concomitant

nutrient removal and biodiesel feedstock production. Biomass Bioenerg, 35, 40-

49.

Liao, B. Q., Allen, D. G., Droppo, I. G., Leppard, G. G., & Liss, S. N. (2001). Surface

properties of sludge and their role in bioflocculation and settleability. Water Res,

35(2), 339-350.

Li, Y., Horsman, M., Wang, B., Wu, N., & Lan, C. (2008). Effects of nitrogen sources on

cell growth and lipid accumulation of green alga Neochloris oleoabundans. Appl

Microbiol Biot, 81(4), 629-639.

Lundquist, T. J., Woertz, I. C., Quinn, N. W. T., & Benemann, J. R. (2010). A realistic

technology and engineering assessment of algae biofuel production. Energy

Biosciences Institute, 1.

Lunka, A. A., & Bayless, D. J. (2013). Effects of flashing light-emitting diodes on algal

biomass productivity. J Appl Phycol, 25(6), 1679-1685.

Luo, K., Yang, Q., Li, X. M., Yang, G. J., Liu, Y., Wang, D. B., Zheng, W., & Zeng, G.

M. (2012). Hydrolysis kinetics in anaerobic digestion of waste activated sludge

enhanced by α-amylase. Biochem Eng J, 62, 17-21.

195

Mata, T. M., Martins, A. A., & Caetano, N. S. (2010). Microalgae for biodiesel

production and other applications: A review. Renew Sust Energ Rev, 14, 217-232.

Menetrez, M. Y. (2012). An overview of algae biofuel production and potential

environmental impact. Environ Sc. Technol, 46, 7073-7085.

Meti, B. S., & Sailaja, B. (2014). Treatment of sugar process waste water and biogas

production using algal biomass. Int J Eng Res Technol, 3(19), 61–67.

Milledge, J. J. (2011). Commercial application of microalgae other than as biofuels: a

brief review. Rev Environ Sci Biotechnol, 10(1), 31-41.

Morales-Sánchez, D., Tinoco-Valencia, R., Caro-Bermúdez, M. A., & Martinez, A.

(2014). Culturing Neochloris oleoabundans microalga in a nitrogen-limited,

heterotrophic fed-batch system to enhance lipid and carbohydrate accumulation.

Algal Res, 5, 61-69.

Morales-Sánchez, D., Tinoco-Valencia, R., Kyndt, J., & Martinez, A. (2013).

Heterotrophic growth of Neochloris oleoabundans using glucose as a carbon

source. Biotechnol Biofuels, 6(1), 1-13.

Morgan, J. W., Forster, C. F., & Evison, L. (1990). A comparative study of the nature of

biopolymers extracted from anaerobic and activated sludges. Water Res, 24(6),

743-750.

Mubarak, M., Shaija, A., & Suchithra, T. V. (2015). A review on the extraction of lipid

from microalgae for biodiesel production. Algal Res, 7, 117-123.

196

Muller-Feuga, A., Lemar, M., Vermel, E., Pradelles, R., Rimbaud, L., & Valiorgue, P.

(2012). Appraisal of a horizontal two-phase flow photobioreactor for industrial

production of delicate microalgae species. J Appl Phycol, 24, 349-355.

Murray, K. E., Healy, F. G., McCord, R. S., & Shields, J. A. (2011). Biomass production

and nutrient uptake by Neochloris oleoabundans in an open trough system. Appl

Microbiol Biot, 90(1), 89-95.

Murray, K. E., Shields, J. A., Garcia, N. D., & Healy, F. G. (2012). Productivity, carbon

utilization, and energy content of mass in scalable microalgae systems.

Bioresource Technol, 114, 499-506.

Nath, K., Muthukumar, M., Kumar, A., & Das, D. (2008). Kinetics of two-stage

fermentation process for the production of hydrogen. Int J Hydrogen Energ,

33(4), 1195-1203.

Olguín, E. J., Castillo, O. S., Mendoza, A., Tapia, K., González-Portela, R. E., &

Hernández- Landa, V. J. (2015a). Dual purpose system that treats anaerobic

effluents from pig waste and produce Neochloris oleoabundans as lipid rich

biomass. New Biotechnol, 32(3), 387-395.

Olguín, E. J., Dorantes, E., Castillo, O. S., & Hernández-Landa, V. J. (2015b). Anaerobic

digestates from vinasse promote growth and lipid enrichment in Neochloris

oleoabundans cultures. J Appl Phycol, 1-10.

Oron, G., Shelef, G., & Levi, A. (1981). Environmental phenotypic variation of

Scenedesmus dimorphus in high‐rate algae ponds and its relationship to

197

wastewater treatment and biomass production. Biotechnol Bioeng, 23(10), 2185-

2198.

Park, J., Jin, H. F., Lim, B. R., Park, K. Y., & Lee, K. (2010). Ammonia removal from

anaerobic digestion effluent of livestock waste using green alga Scenedesmus

sp. Bioresource Technol, 101(22), 8649-8657.

Park, K., Lim, B., & Lee, K. (2009). Growth of microalgae in diluted process water of the

animal wastewater treatment plant. Water Sci Technol, 59(11), 2111–2116.

Peng, L., Lan, C. Q., Zhang, Z., Sarch, C., & Laporte, M. (2015). Control of protozoa

contamination and lipid accumulation in Neochloris oleoabundans culture:

Effects of pH and dissolved inorganic carbon. Bioresource Technol, 197, 143-

151.

Pere, J., Alen, R., Viikari, L., & Eriksson, L. (1993). Characterization and dewatering of

activated sludge from the pulp and paper industry. Water Sci Technol, 28(1), 193-

201.

Pittman, J. K., Dean, A. P., & Osundeko, O. (2011). The potential of sustainable algal

biofuel production using wastewater resources. Bioresource Technol, 102(1), 17-

25.

Popovich, C. A., Damiani, C., Constenla, D., Martínez, A. M., Freije, H., Giovanardi, M.,

Pancaldi, S., & Leonardi, P. I. (2012). Neochloris oleoabundans grown in

enriched natural seawater for biodiesel feedstock: Evaluation of its growth and

biochemical composition. Bioresource Technol, 114, 287-293.

198

Pruvost, J., Van Vooren, G., Cogne, G., & Legrand, J. (2009). Investigation of biomass

and lipids production with Neochloris oleoabundans in photobioreactor.

Bioresource Technol, 100(23), 5988-5995.

Rao, C. S., Sathish, T., Pendyala, B., Kumar, T. P., & Prakasham, R. S. (2009).

Development of a mathematical model for Bacillus circulans growth and alkaline

protease production kinetics. J Chem Technol Biot, 84(2), 302-307.

Renaud, S. M., Parry, D. L., & Thinh, L. V. (1994). Microalgae for use in tropical

aquaculture I: Gross chemical and fatty acid composition of twelve species of

microalgae from the Northern Territory, Australia. J Appl Phycol, 6(3), 337-345.

Rhee, G. Y. (1978). Effects of N:P atomic ratios and nitrate limitation on algal growth,

cell composition, and nitrate uptake. Limnol Oceanogr, 23(1), 10-25.

Richmond, A. (Ed.). (2004). Handbook of microalgal culture: biotechnology and applied

phycology. John Wiley & Sons.

Rismani-Yazdi, H., Haznedaroglu, B. Z., Hsin, C., & Peccia, J. (2012). Transcriptomic

analysis of the oleaginous microalga Neochloris oleoabundans reveals metabolic

insights into triacylglyceride accumulation. Biotechnol Biofuels, 5(1), 74.

Rodolfi, L., Zittelli, G. C., Bassi, N., Padovani, G., Biondi, N., Bonini, G., & Tredici, M.

R. (2009). Microalgae for oil: Strain selection, induction of lipid synthesis and

outdoor mass cultivation in a low-cost photobioreactor. Biotechnol Bioeng, 102,

100-112.

Rogers, J. N., Rosenberg, J. N., Guzman, B. J., Oh, V. H., Mimbela, L. E., Ghassemi, A.,

Betenbaugh, M. J., Oyler, G. A., & Donohue, M. D. (2014). A critical analysis of

199

paddlewheel-driven raceway ponds for algal biofuel production at commercial

scales. Algal Res, 4, 76-88.

Roudsari, F. P., Mehrnia, M. R., Asadi, A., Moayedi, Z., & Ranjbar, R. (2014). Effect of

microalgae/activated sludge ratio on cooperative treatment of anaerobic effluent

of municipal wastewater. Appl Biochem Biotech, 172(1), 131-140.

Ruangsomboon, S., Ganmanee, M., & Choochote, S. (2013). Effects of different

nitrogen, phosphorus, and iron concentrations and salinity on lipid production in

newly isolated strain of the tropical green microalga, Scenedesmus dimorphus

KMITL. J Appl Phycol, 25(3), 867-874.

Sabia, A., Baldisserotto, C., Biondi, S., Marchesini, R., Tedeschi, P., Maietti, A.,

Giovanardi, M., Ferroni, L., & Pancaldi, S. (2015). Re-cultivation of Neochloris

oleoabundans in exhausted autotrophic and mixotrophic media: the potential role

of polyamines and free fatty acids. Appl Microbiol Biotechnol, 99(24), 10597-

10609.

Salim, S., Vermuë, M. H., & Wijffels, R. H. (2012). Ratio between autoflocculating and

target microalgae affects the energy-efficient harvesting by bio-flocculation.

Bioresource Technol, 118, 49-55.

Santos, A. M., Janssen, M., Lamers, P. P., Evers, W. A. C., & Wijffels, R. H. (2012).

Growth of oil accumulating microalga Neochloris oleoabundans under alkaline–

saline conditions. Bioresource Technol, 104, 593-599.

200

Santos, A. M., Lamers, P. P., Janssen, M., & Wijffels, R. H. (2013). Biomass and lipid

productivity of Neochloris oleoabundans under alkaline–saline conditions. Algal

Res, 2(3), 204-211.

Santos, A. M., Wijffels, R. H., & Lamers, P. P. (2014). pH-upshock yields more lipids in

nitrogen-starved Neochloris oleoabundans. Bioresource Technol, 152, 299-306.

Schachtman, D. P., Reid, R. J., & Ayling, S. M. (1998). Phosphorus uptake by plants:

from soil to cell. Plant Physiol, 116(2), 447-453.

Schlagermann, P., Göttlicher, G., Dillschneider, R., Rosello-Sastre, R., & Posten, C.

(2012). Composition of algal oil and its potential as biofuel. J Combust, 2012.

Sharma, K. K., Schuhmann, H., & Schenk, P. M. (2012). High lipid induction in

microalgae for biodiesel production. Energies, 5(5), 1532-1553.

Sheehan, J., Dunahay, T., Benemann, J., & Roessler, P. (1998). A look back at the U.S.

Department of Energy’s Aquatic Species Program – Biodiesel from algae. U.S.

Report NREL/TP-580-24190. National Renewable Energy Laboratory, Golden,

CO.

Sousa, C., Compadre, A., Vermuë, M. H., & Wijffels, R. H. (2013a). Effect of oxygen at

low and high light intensities on the growth of Neochloris oleoabundans. Algal

Res, 2(2), 122-126.

Sousa, C., Valev, D., Vermuë, M. H., & Wijffels, R. H. (2013b). Effect of dynamic

oxygen concentrations on the growth of Neochloris oleoabundans at sub-

saturating light conditions. Bioresource Technol, 142, 95-100.

201

Sousa, C., De Winter, L., Janssen, M., Vermuë, M. H., & Wijffels, R. H. (2012). Growth

of the microalgae Neochloris oleoabundans at high partial oxygen pressures and

sub-saturating light intensity. Bioresource Technol, 104, 565-570.

Sturm, B. S., & Lamer, S. L. (2011). An energy evaluation of coupling nutrient removal

from wastewater with algal biomass production. Appl Energ, 88(10), 3499-3506.

Subramanya, K. (2009). Flow in Open Channels. Tata McGraw-Hill Education.

Su, B., Qu, Z., Song, Y., Jia, L., & Zhu, J. (2014). Investigation of measurement methods

and characterization of zeta potential for aerobic granular sludge. J Environ Chem

Eng, 2(2), 1142-1147.

Sun, X., Cao, Y., Xu, H., Liu, Y., Sun, J., Qiao, D., & Cao, Y. (2014). Effect of nitrogen-

starvation, light intensity and iron on triacylglyceride/carbohydrate production

and fatty acid profile of Neochloris oleoabundans HK-129 by a two-stage

process. Bioresource Technol, 155, 204-212.

Surendhiran, D., Vijay, M., Sivaprakash, B., & Sirajunnisa, A. (2015). Kinetic modeling

of microalgal growth and lipid synthesis for biodiesel production. 3 Biotech, 5(5),

663-669.

Tam, N. F. Y., & Wong, Y. S. (1990). The comparison of growth and nutrient removal

efficiency of Chlorella pyrenoidosa in settled and activated sewages. Environ

Polllut, 65(2), 93-108.

Tam, N. F. Y., & Wong, Y. S. (1996). Effect of ammonia concentrations on growth of

Chlorella vulgaris and nitrogen removal from media. Bioresource Technol, 57(1),

45-50.

202

Tchobanoglous, G., Burton, F. L., & Stensel, H. D. (2004). Wastewater engineering:

Treatment and Reuse. New York, NY: McGraw Hill.

Teefy, S. M. (1996). Tracer Studies in Water Treatment Facilities: A Protocol and Case

Studies, published by AWWARF.

Terry, K. T., & Raymond, L. P. (1985). System design for the autotrophic production of

microalgae. Enzyme Microb. Technol., 7, 474-487.

The Culture Collection of Algae at the University of Texas at Austin. (2016). Retrieved

from www.utex.org

Tibbetts, S. M., Milley, J. E., & Lall, S. P. (2015). Chemical composition and nutritional

properties of freshwater and marine microalgal biomass cultured in

photobioreactors. J Appl Phycol, 27(3), 1109-1119.

Tomei, M. C., Braguglia, C. M., & Mininni, G. (2008). Anaerobic degradation kinetics of

particulate organic matter in untreated and sonicated sewage sludge: Role of the

inoculum. Bioresource Technol, 99(14), 6119-6126.

Tornabene, T. G., Holzer, G., Lien, S., & Burris, N. (1983). Lipid composition of the

nitrogen starved green alga Neochloris oleoabundans. Enzyme and Microb Tech,

5(6), 435-440.

Travieso, L., Benítez, F., Sánchez, E., Borja, R., Martín, A., & Colmenarejo, M. F.

(2006). Batch mixed culture of Chlorella vulgaris using settled and diluted

piggery waste. Ecol Eng, 28(2), 158-165.

Urreta, I., Ikaran, Z., Janices, I., Ibañez, E., Castro-Puyana, M., Castañón, S., & Suárez-

Alvarez, S. (2014). Revalorization of Neochloris oleoabundans biomass as source

203

of biodiesel by concurrent production of lipids and carotenoids. Algal Res, 5, 16-

22.

Vidyashankar, S., Deviprasad, K., Chauhan, V. S., Ravishankar, G. A., & Sarada, R.

(2013). Selection and evaluation of CO2 tolerant indigenous microalga

Scenedesmus dimorphus for unsaturated fatty acid rich lipid production under

different culture conditions. Bioresource Technol, 144, 28-37.

Villa, J. A., Ray, E. E., & Barney, B. M. (2014). Azotobacter vinelandii siderophore can

provide nitrogen to support the culture of the green algae Neochloris

oleoabundans and Scenedesmus sp. BA032. FEMS Microbiol Lett, 351(1), 70-77.

Wahal, S., & Viamajala, S. (2010). Maximizing algal growth in batch reactors using

sequential change in light intensity. Appl Biochem Biotech, 161, 511-522.

Wang, B., & Lan, C. Q. (2011a). Biomass production and nitrogen and phosphorus

removal by the green alga Neochloris oleoabundans in simulated wastewater and

secondary municipal wastewater effluent. Bioresource Technol, 102(10), 5639-

5644.

Wang, B., & Lan, C. Q. (2011b). Optimising the lipid production of the green alga

Neochloris oleoabundans using Box–Behnken experimental design. Can J Chem

Eng, 89(4), 932-939.

Wang, L., Li, Y., Sommerfeld, M., & Hu, Q. (2013). A flexible culture process for

production of the green microalga Scenedesmus dimorphus rich in protein,

carbohydrate or lipid. Bioresource Technol, 129, 289-295.

204

Watten, B. J., Honeyfield, D. C., & Schwartz, M. F. (2000). Hydraulic characteristics of a

rectangular mixed-cell rearing unit. Aquacult Eng, 24, 59-73.

Welter, C., Schwenk, J., Kanani, B., Van Blargan, J., & Belovich, J. M. (2013). Minimal

medium for optimal growth and lipid production of the microalgae Scenedesmus

dimorphus. Environ Prog Sust Energy, 32(4), 937-945.

Williams, P. J. L. B., & Laurens, L. M. (2010). Microalgae as biodiesel & biomass

feedstocks: Review & analysis of the biochemistry, energetics & economyics.

Energ Environ Sci, 3(5), 554-590.

Wu, N., Li, Y., & Lan, C. Q. (2011). Production and rheological studies of microalgal

extracellular biopolymer from lactose using the green alga Neochloris

oleoabundans. J Polym Environ, 19(4), 935-942.

Xin, L., Hong-Ying, H., & Yu-Ping, Z. (2011). Growth and lipid accumulation properties

of a freshwater microalga Scenedesmus sp. under different cultivation

temperature. Bioresource Technol, 102(3), 3098-3102.

Xu, X., Shen, Y., & Chen, J. (2015). Cultivation of Scenedesmus dimorphus for C/N/P

removal and lipid production. Electron J Biotechn, 18(1), 46-50.

Yang, Y., Mininberg, B., Tarbet, A., & Weathers, P. (2013). At high temperature lipid

production in Ettlia oleoabundans occurs before nitrate depletion. Appl Microbiol

Biot, 97(5), 2263-2273.

Yang, Y., Xu, J., Vail, D., & Weathers, P. (2011). Ettlia oleoabundans growth and oil

production on agricultural anaerobic waste effluents. Bioresource Technol,

102(8), 5076-5082.

205

Yaoyang, X. U., & Boeing, W. J. (2014). Modeling maximum lipid productivity of

microalgae: Review and next step. Renew Sust Energ Rev, 32, 29-39.

Yoon, S. Y., Hong, M. E., Chang, W. S., & Sim, S. J. (2015). Enhanced biodiesel

production in Neochloris oleoabundans by a semi-continuous process in two stage

photobioreactors. Bioprocess Biosyst Eng, 38(7), 1415-1421.

Zhao, Y., Sun, S., Hu, C., Zhang, H., Xu, J., & Ping, L. (2015). Performance of three

microalgal strains in biogas slurry purification and biogas upgrade in response to

various mixed light-emitting diode light wavelengths. Bioresource Technol, 187,

338-345.

206

APPENDIX A. SELECTIVE PRECIPITATION OF ALUMINUM AND IRON IN

ACID MINE DRAINAGE5

A.1 Abstract

Acid Mine Drainage (AMD) is a wastewater generated from abandoned mining activities, and is featured by the relatively high concentrations of metals and acidity.

Eastern and southeastern Ohio has witnessed substantial mining between the mid-1800s and mid-1900s. As a result, a legacy of pollution flows into waterways, damaging vegetation, soil, groundwater, rivers, and aquatic life. In this study, a selective sequential precipitation approach was tested for the removal of aluminum and iron from AMD. By precipitating aluminum first, it is hoped that separate streams of pure aluminum and iron can be produced suitable for sale as recovered products. PHREEQC Interactive Software was used to model the AMD and predict the dominant metal species over a pH range while keeping the solution anaerobic. The model findings were coupled with a series of bench-top laboratory tests in order to find the optimal operating pH for aluminum precipitation. Laboratory tests resulted in a 95% reduction in dissolved aluminum concentration over a pH range of 4 to 4.75, whereas dissolved ferrous iron concentration remained relatively high in solution. Actual AMD from Snake Hollow, Ohio, was tested in a pilot treatment plant that was constructed in Nelsonville, Ohio, by the Wayne

National Forest, US Forest Service. Aluminum was removed at pH=4.5 while keeping anaerobic conditions in a pretreatment tank through nitrogen gas sparging. Then the

5 This chapter is a proceeding of the National Association of Abandoned Mine Land Programs Conference. Daniels, WV (Abu Hajar & Riefler, 2013).

207 effluent was directed to an electrolysis reactor to oxidize ferrous iron to ferric iron, and subsequently, precipitate as ferric hydroxide. It was found that 90% of ferrous iron was kept in solution, while 67% of aluminum precipitated as aluminum hydroxide in the pretreatment stage. The results indicate aluminum and iron can be separated through the process described, although refinement is required at the pilot-scale to produce aluminum and iron precipitates of high purity.

Keywords: Acid mine drainage, selective precipitation, active treatment.

208

A.2 Introduction

Acid Mine Drainage (AMD) is an acidic wastewater generated from abandoned mining sites, and contains relatively high concentrations of metals such as iron, aluminum, and manganese as well as sulfuric acid (Cheng et al., 2011; Cui et al., 2012;

Tabak et al., 2003; Wei et al., 2005). Sulfuric acid results from the reaction of water and oxygen with sulfide containing minerals such as pyrite; thus, the resulting acidic solution causes the leaching of metals from the rocks. Pyrite (FeS2) oxidation can occur spontaneously and can be catalyzed by iron oxidizing bacteria along with the presence of ferric iron (Chandra & Gerson, 2010; Cuie et al., 2012; Marcello et al., 2008). The key step in the formation of AMD is the oxidation of ferrous iron in pyrite which exists in deep underground mines according to the following equations (Chandra & Gerson, 2010;

Chartrand & Bunce, 2003):

2+ 2− + FeS2 + 7⁄2 O2 + H2O → Fe + 2SO4 + 2H Equation A. 1

2+ + 3+ Fe + 1⁄4 O2 + H → Fe + 1⁄2 H2O Equation A. 2

3+ 2+ 2− + FeS2 + 14 Fe + 8H2O → 15Fe + 2SO4 + 16H Equation A. 3

The environmental impact of AMD includes damage to vegetation, soil, groundwater, receiving water bodies such as rivers and aquatic life (Chartrand & Bunce,

2003; Cui et al., 2012; Tabak et al., 2003). It has been estimated that 19,300 km (12,000 miles) of rivers and streams and 72,000 ha of lakes and reservoirs were impaired by

209

AMD in the United States (Caraballo et al., 2011). Eastern and southeastern parts of Ohio have been heavily impacted by the AMD activities which resulted in 1,300 miles of polluted waterways currently (“Acid mine drainage,” n.d.).

The conventional treatment of AMD is achieved through the oxidation of metals, neutralizing the pH through alkaline addition, precipitation of metals as hydroxide products, and the separation and processing of the produced sludge (Chartrand & Bunce,

2003; Cui et al., 2012; Tabak et al, 2003). Two types of AMD treatment were widely used: 1) passive treatment such as wetlands, steel slag beds, and ponds which requires both space and frequent maintenance; and 2) active treatment, featured by a treatment plant style, in which the chemical oxidation of metals is the primary step after the pH is neutralized through alkaline addition (Caraballo et al., 2011; Cheng et al, 2011; Wei et al., 2005).

Chemical oxidation of ferrous iron could be a rate-limiting step; due to the slow reaction at pH< 4; thus, alternative oxidizing methods may be desirable (Wei et al.,

2005). In their study on the recovery of iron and aluminum from acid mine drainage by selective precipitation; Wei et al. (2005) indicated that the oxidation of ferrous iron through aeration resulted in a high purity ferric hydroxide precipitate when the operating pH was less than 3.5. However, at higher pH, simultaneous precipitation of aluminum and iron was observed, which would affect the purity of the precipitate (Wei et al., 2005).

Nevertheless, if less aerobic conditions exist in the source, ferrous iron will remain in the reduced state, where ferrous iron is more soluble than ferric iron at pH below 11. This

210 will allow other metals to precipitate in earlier stages (pH< 6) while keeping ferrous iron in solution for further oxidation and precipitation at pH range of 7-8 (Jensen, 2003).

Electrolysis can be used for oxidizing ferrous iron in AMD as an example of active treatment methods. An electric potential is forced across an electrolyte in an electrochemical cell, where the ferrous iron is oxidized at the anode to produce ferric iron

(Arthur, 2011; Chartrand & Bunce, 2003). Electrochemical cell set-up can be either divided or undivided. In case of undivided, a pH range between 7 and 8 should be maintained in the electrolysis unit in order to prevent the back reduction to ferrous iron at the cathode (Chartrand & Bunce, 2003; Cheng et al., 2011; Jensen, 2003).

In case of high concentrations of iron and other metals in the AMD, it is desirable to precipitate other metals in prior stages; thus, improving the purity of the thickened iron sludge, which can be processed further into marketable product such as iron paint pigments (Marcello et al., 2008).

This study discusses the selective sequential precipitation of metals (iron and aluminum) from AMD as hydroxides; thus, improving metals’ precipitate purities. A pilot treatment plant was constructed by Wayne National Forest in Nelsonville, Ohio, and tested for this purpose. The main feature of this plant was ferrous iron oxidation through electrolysis. The results of the plant were compared with a PHREEQC simulation and a series of bench-top tests.

A.3 Materials and Methods

A synthetic AMD sample was modeled using PHREEQC Interactive software, version 2.18.3.5570. Aluminum and iron concentrations were set equal to 80 and 15

211 mg/L, respectively. A solution-spread option was used through PHREEQC by specifying the above concentrations at pH values ranging from 4 to 5.25; thus, simulating the impact of titration on the species of interest (Al and Fe). The minteq version 4 database was applied and parameters such as saturation indices (SI) were computed in order to investigate multiple alternatives.

Artificial AMD sample corresponding to that simulated in the model was prepared by acidifying ultra-pure water with 0.02 N H2SO4 to reduce the pH to below 3, and then sparged with nitrogen gas for 2 hr to eliminate oxygen. Ferrous ammonium sulfate (FAS) and aluminum sulfate (AS) were added to the solution for ferrous and aluminum target concentrations of 15 and 80 mg/L, respectively. Each 250-mL trial was bubbled with nitrogen gas and titrated with 0.02 N NaOH simultaneously to reach the desired pH. Samples were collected immediately after metal addition, immediately after pH adjustment, and 15 min after pH adjustment. Samples were preserved with 20%

HNO3, and dissolved samples were filtered with 0.45 µm syringe filters. Analyses were performed by flame atomic absorption spectrometer (Perkin Elmer AAnalyst 300).

Finally, AMD from the Snake Hollow mining site, Nelsonville, OH was examined for sequential removal of heavy metals. This treatment was achieved through a pilot treatment plant that operated at 10 gpm. The main feature of this plant was the use of electrolysis for oxidizing ferrous iron and precipitating as ferric hydroxide in the subsequent settling tank. In order to achieve high purity in the sludge accumulating in the iron settling tank, aluminum should be precipitated first as aluminum hydroxide in a prior aluminum settling tank and ferrous iron should be kept in solution as indicated earlier.

212

The volume of the pretreatment tank was 2,800 L (740 gallons) and the actual measured retention time corresponding to an inflow of 10 gpm was 66 min. Two six-inch diffusers connected to a nitrogen gas tank were placed at the front end of the aluminum settling tank bottom, with a nitrogen gas flow rate of approximately 20 L/min (0.71 ft3/min).

Additionally, the surface of the water in the aluminum settling tank was covered with a floating cover in order to reduce the exposure to atmospheric oxygen.

A.4 Results and Discussion

A.4.1 PHREEQC simulations

PHREEQC simulations were run on AMD at different pH’s to determine when metals will precipitate. Starting from pH= 4, ferrous iron and aluminum remain dissolved in solution, whereas by increasing the pH, aluminum begins to precipitate as aluminum hydroxide (Table A. 1).

Table A. 1: PHREEQC Simulation Results Saturation Index (SI) [Al] [Fe] [Al3+] [Fe2+] pH Total Total Al(OH) 3(s) Fe(OH) µmol/kg µmol/kg µmol/kg µmol/kg Amorphous Gibbsite 2 (s) 4.00 2965 269 2803 269 -1.82 0.69 -9.37 4.25 2965 269 2677 269 -1.08 1.43 -8.87 4.50 2965 269 2454 269 -0.36 2.15 -8.37 4.75 2965 269 2078 269 0.34 2.85 -7.86 5.00 2965 269 1514 269 0.98 3.49 -7.34 5.25 2965 269 867 269 1.54 4.05 -6.81

The saturation index values for the three solid forms indicate that aluminum would likely precipitate as gibbsite at pH slightly lower than 4 while amorphous Al(OH)3

213 does not form at pH< 4.5. At low pH’s, aluminum was primarily found as the free Al3+ ion with the ligand AlOH2+ the second most common form. Iron was entirely found as the free Fe2+ ion. Two additional runs were performed by assuming equilibrium with gibbsite and amorphous (Table A. 2 and Table A. 3).

Based on equilibrium water chemistry, raising the pH from 4.0 to 4.75 would reduce Al concentration by 99.9% (gibbsite equilibrium) and 66.3% (amorphous equilibrium) while the ferrous iron should remain totally in solution. To evaluate the effect of ferrous oxidation to ferric on the precipitation of aluminum, another set of solutions was simulated with iron concentration being entirely in the ferric form (Table

A. 4). Goethite and hematite replaced ferrous hydroxide in the solubility indices.

Table A. 2: PHREEQC Simulation Results (Gibbsite is Present)

+3 +2 Saturation Index (SI) [Al] Total [Fe] Total [Al ] [Fe ] % Final/Initial pH Al(OH)3(s) Fe(OH)2 (s) µmol/kg µmol/kg µmol/kg µmol/kg Al Fe Amorphous Gibbsite 4.00 331.1 268.6 305.6 268.6 11.2 100 -2.51 0 -9.24 4.25 54.2 268.6 46.3 268.6 1.8 100 -2.51 0 -8.70 4.50 10.6 268.6 7.9 268.6 0.4 100 -2.51 0 -8.20 4.75 2.4 268.6 1.4 268.6 0.1 100 -2.51 0 -7.70 5.00 0.6 268.6 0.2 268.6 0 100 -2.51 0 -7.20 5.25 0.2 268.6 0 268.6 0 100 -2.51 0 -6.70

214

Table A. 3: PHREEQC Simulation Results (Amorphous is Present)

3+ 2+ Saturation Index (SI) [Al] Total [Fe] Total [Al ] [Fe ] % Final/Initial pH Al(OH)3(s) Fe(OH)2 (s) mmol/kg mmol/kg mmol/kg mmol/kg Al Fe Amorphous Gibbsite 4.00 315.1 0.3 276.4 0.3 100 100 0 2.51 -11.48 4.25 161.2 0.3 151.0 0.3 100 100 0 2.51 -10.34 4.50 15.9 0.3 14.2 0.3 100 100 0 2.51 -8.69 4.75 1.0 0.3 0.7 0.3 33.7 100 0 2.51 -7.77 5.00 0.2 0.3 0.1 0.3 6.7 100 0 2.51 -7.21 5.25 0.1 0.3 0 0.3 3.4 100 0 2.51 -6.70

Table A. 4: PHREEQC Simulation Results (Iron is in the Ferric Form)

+3 +3 Saturation Index (SI) [Al] Total [Fe] Total [Al ] Dissolved [Fe ] Dissolved pH Al(OH)3(am) Goethite Hematite µmol/kg µmol/kg µmol/kg µmol/kg Amorphous Gibbsite 4.00 2965 269 2802 0.261 -1.81 0.70 4.47 11.33 4.25 2965 269 2675 0.083 -1.07 1.44 4.73 11.85 4.50 2965 269 2450 0.026 -0.35 2.16 4.98 12.36 4.75 2965 269 2071 0.008 0.35 2.86 5.23 12.87 5.00 2965 269 1505 0.002 0.99 3.50 5.49 13.38 5.25 2965 269 857 0.001 1.55 4.05 5.74 13.89

Clearly, once the iron oxidizes into the ferric state, equilibrium chemistry predicts that the solution is supersaturated and that iron will precipitate at all pH’s tested here.

This suggests that a separation method is required for the removal of iron at very low pH’s while keeping aluminum in solution. However, ferrous iron oxidation is severely rate limiting at low pH, requiring the use of strong oxidants that do not yield high value paint pigments.

215

A.4.2 Artificial AMD sample

An AMD sample was prepared in the lab and tested at the six pH values analyzed with PHREEQC. Unfiltered and filtered aluminum and iron concentrations were measured at each pH (Figure A. 1).

(a) Aluminum Concentration Variation with (b) Iron Concentration Variation with pH Al [1] pH Al [2] Fe [1] Fe [2] Al [3] Al [4] Fe [3] Fe [4] 90 16

80 14 70 12 60 10 50 8 40 6 30 20 4 Concentration (mg/L) Concentration (mg/L) 10 2 0 0 4.00 4.25 4.50 4.75 5.00 5.25 4.00 4.25 4.50 4.75 5.00 5.25 pH pH

Figure A. 1: Concentration variation with pH (a) Aluminum, (b) Iron [1]: Initial filtered metals' concentration after adding FAS and AS to the acidic solution (pH= 2.9), [2]: Filtered metals' concentration at the desired pH (time= 0), [3]: Filtered metals' concentration at the desired pH (time= 15 min), and [4]: Unfiltered metals concentration at the desired pH (time= 15 min).

Aluminum concentration dropped significantly in the pH range of 4 to 4.75. For instance, the initial filtered aluminum concentration for pH= 4.75 solution (prior to titration) was 70.76 mg/L, while the concentration dropped to 3.68 mg/L immediately after titration to the desired pH (4.75) with approximately 95% reduction. In contrast, ferrous concentration remained relatively high throughout the pH range with only moderate precipitation at pH< 5.25. These observations can be explained better by

216 finding the percentages of concentrations between filtered and unfiltered samples at 15 min and comparing them to PHREEQC simulation findings (Table A. 5). Furthermore, the difference in filtered metal concentrations at t= 0 and 15 min is insignificant, which suggests that time is not a controlling factor in the precipitation of either aluminum or iron; i.e., formation of aluminum hydroxide or ferric hydroxide occurs rapidly.

Table A. 5: Comparison between Laboratory Results and PHREEQC Simulation Results Filtered/Unfiltered PHREEQC pH Al Al Fe Al Fe (Gibbsite equilibrium) (Amorphous equilibrium) 4.00 1.012 0.996 1.000 0.11 1.00 4.25 0.908 0.544 1.000 0.02 1.00 4.50 0.731 0.176 1.000 0 1.00 4.75 0.681 0.054 1.000 0 0.34 5.00 0.714 0.034 1.000 0 0.07 5.25 0.651 0.028 1.000 0 0.03

PHREEQC predicted the solution to become supersaturated with aluminum at all pH values in the 4 – 5.5 range when gibbsite is present and at pH values between 4.5 and

4.75 when amorphous is present. Based on laboratory test results, ferrous iron concentrations remained unchanged in the simulations throughout the pH range, unlike the lab tests, in which ferrous iron concentration was reduced by 35% at pH= 5.25 and aluminum started precipitation earlier. Moreover, at pH= 4.5, 82.4% of aluminum precipitated while only 26.9% of ferrous iron precipitated. Thus, a pH= 4.5 was adopted as the precipitation condition for the aluminum settling tank as described below.

217

A.4.3 Pilot treatment plant results

The aluminum settling tank was the first step of the sequential precipitation of metals in AMD achieved through an electrolysis reactor and a series of settling tanks in the pilot plant discussed earlier. The test lasted for 10 h with ten samples collected hourly

(Table A. 6). The target pH for the settling tank was set equal to 4.5, while the actual pH measured through the effluent of the tank averaged 4.44  0.34. The average influent unfiltered concentrations of aluminum and iron were 30.0 and 69.4 mg/L, respectively, while the average influent filtered concentrations were 29.0 and 62.1 mg/L. On the other hand, the average effluent unfiltered concentrations of aluminum and iron were 20.2 and

67.5 mg/L, respectively, whereas the average effluent filtered concentrations were 10.8 and 57.0 mg/L.

Maintaining the majority of iron in solution was successful in the aluminum settling tank, while a significant portion of aluminum precipitated as aluminum hydroxide. In other words, 33% of the influent dissolved aluminum remained in solution in the effluent while 67% was converted to aluminum hydroxide. Also, 90% of ferrous iron remained in solution while the remaining 10% was converted to a precipitate form

(this might be ferrous precipitate such as ferrous hydroxide or ferrous sulfate, or this portion has been oxidized to form ferric precipitate). Filtered concentrations were used for analysis because of the considerable variation in the unfiltered metals concentrations

(Figure A. 2).

218

Table A. 6: Iron and Aluminum Results for the Aluminum Settling Tank Influent and Effluent Influent Effluent Sampling Sampling Fe Conc (mg/L) Al Conc (mg/L) Fe Conc (mg/L) Al Conc (mg/L) Time Time Unfiltered Filtered Unfiltered Filtered Unfiltered Filtered Unfiltered Filtered 0:30 64.8 54.0 29.8 23.7 1:36 66.6 53.6 20.3 7.9 1:30 62.8 62.8 30.3 30.1 2:36 77.2 52.8 34.1 - 2:30 62.5 61.9 29.8 29.0 3:36 63.5 53.4 17.6 10.2 3:30 59.8 58.8 30.3 28.7 4:36 73.4 53.5 30.6 9.5 4:30 58.8 64.3 30.1 30.4 5:36 66.9 58.3 21.6 9.1 5:30 77.8 - 29.6 - 6:36 64.7 62.0 14.8 17.8 6:30 74.5 - 29.3 - 7:36 76.2 59.2 25.6 11.2 7:30 78.4 64.2 30.3 30.1 8:36 63.0 57.8 12.3 10.0 8:30 77.0 67.0 30.1 29.9 9:36 60.6 58.9 12.1 10.7 9:30 77.6 63.6 30.6 29.9 10:36 62.5 60.5 12.8 10.5 Mean 69.4 62.1 30.0 29.0 - 67.5 57.0 20.2 10.8 St. Dev 7.9 3.7 0.4 2.1 - 5.7 3.2 7.4 2.7

By comparing the settling tank results to the PHREEQC simulations and lab tests at pH= 4.5, it can be concluded that PHREEQC results were not achieved in both experimental sets; i.e., the final/initial aluminum ratios were 1, 0.176, and 0.37 for

PHREEQC (amorphous is present), lab tests, and settling tank results, respectively.

Moreover, PHREEQC predicted a significant drop in aluminum concentrations in the case of gibbsite equilibrium at all pH values, whereas aluminum starts to precipitate at pH > 4.5 in the case of amorphous equilibrium, which indicates that amorphous Al(OH)3 is more likely to form instead of gibbsite. Therefore, aluminum precipitated earlier than the simulated results in the lab tests compared to the PHREEQC simulations. However, the results indicated a better ferrous iron performance in the settling tank compared to lab tests (0.92 compared to 0.73), yet, both were lower than PHREEQC simulation which predicted a ratio of 1.0 for unfiltered ferrous iron.

219

One of the reasons for the differences between the model and the experiments may have been the failure to maintain a continuous anaerobic environment in the aluminum settling tank to prevent the oxidation of ferrous iron. During the transportation of AMD from Snake Hollow mines, the AMD was carried in a partially filled water truck travelling on dirt roads for several hours and pumped into onsite holding tanks. This likely aerated the AMD that emerged from the mine oxygen free. The nitrogen gas sparging may have been insufficient to remove all accumulated oxygen (dissolved oxygen readings were inconclusive).

Fe-Unfiltered Fe-Filtered Al-Unfiltered Al-Filtered 1.4 1.2

1.0 0.8 0.6 Conc. Ratio 0.4 0.2 0.0 0 1 2 3 4 5 6 7 8 9 10 Sample

Figure A. 2: Effluent/influent concentration ratio for aluminum settling tank

A.5 Conclusions

In some active AMD treatment facilities, selective precipitation of heavy metals should be applied, especially when the precipitated sludge is to be further investigated for pigments production. For this purpose, a pilot treatment plant utilizing electrolysis for

220 oxidizing ferrous iron was designed. One challenge faced with this source of AMD was the combination of dissolved metals which will reduce the purity of the resulting pigment. Therefore, aluminum should be precipitated anaerobically in the pretreatment stage at pH= 4.5, while ferrous iron should be oxidized in the electrolysis reactor at pH=

7.5 and precipitated in the subsequent settling tank. A PHREEQC model was developed to simulate a synthetic AMD sample and investigate the dominant species over a range of pH values. Analogous bench-top tests were performed including titration and nitrogen sparging for anaerobic solution assurance. Finally, aluminum settling tank results from actual Snake Hollow AMD in the pilot treatment plant were used for comparison with the model and the laboratory tests. Based on the comparison between the model and the bench-top tests, pH= 4.5 was suitable for removing a significant percentage of aluminum from solution while keeping the majority of dissolved ferrous in solution (assuming anaerobic conditions). This was applied in the pilot treatment plant and the results indicated that two thirds of aluminum was precipitated as aluminum hydroxide, meanwhile, only 10% of ferrous iron was precipitated as ferrous or ferric precipitate.

221

A.6 References

Acid mine drainage abatement program. (n.d.). Retrieved on February 15, 2013, from

http://www.ohiodnr.com/mineral/acid/tabid/10421/Default.aspx

Arthur, B. (2011). The electrolysis of mine drainage (Technical report). Columbus: Ohio

Department of Natural Resources.

Caraballo, M. A., Macías, F., Rötting, T. S., Nieto, J. M., & Ayora, C. (2011). Long term

remediation of highly polluted acid mine drainage: A sustainable approach to

restore the environmental quality of the Odiel river basin. Environ Pollut,

159(12), 3613-3619.

Chandra, A. P., & Gerson, A. R. (2010). The mechanisms of pyrite oxidation and

leaching: A fundamental perspective. Surf Sci Rep, 65(9), 293-315.

Chartrand, M. M. G., & Bunce, N. J. (2003). Electrochemical remediation of acid mine

drainage. J Appl Electrochemi, 33(3-4), 259-264.

Cheng, S., Jang, J. H., Dempsey, B. A., & Logan, B. E. (2011). Efficient recovery of

nano-sized iron oxide particles from synthetic acid-mine drainage (AMD) water

using fuel cell technologies. Water Res, 45(1), 303-307.

Cui, M., Jang, M., Cho, S. H., Khim, J., & Cannon, F. S. (2012). A continuous pilot-scale

system using coal-mine drainage sludge to treat acid mine drainage contaminated

with high concentrations of Pb, Zn, and other heavy metals. J Hazard Mater, 215,

122-128.

Jensen, J.N. (2003). A Problem-Solving Approach to Aquatic Chemistry. Hoboken, New

Jersey: John Wiley & Sons, Inc.

222

Marcello, R. R., Galato, S., Peterson, M., Riella, H. G., & Bernardin, A. M. (2008).

Inorganic pigments made from the recycling of coal mine drainage treatment

sludge. J Environ Manage, 88(4), 1280-1284.

Tabak, H. H., Scharp, R., Burckle, J., Kawahara, F. K., & Govind, R. (2003). Advances

in biotreatment of acid mine drainage and biorecovery of metals: 1. Metal

precipitation for recovery and recycle. Biodegradation, 14(6), 423-436.

Wei, X., Viadero Jr, R. C., & Buzby, K. M. (2005). Recovery of iron and aluminum from

acid mine drainage by selective precipitation. Environ Eng Sci, 22(6), 745-755.

! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !

Thesis and Dissertation Services ! !