MANGO (Philippine 'Carabao' Var.) POWDER MADE FROM

MANGO (Philippine 'Carabao' var.) POWDER MADE FROM

DIFFERENT DRYING SYSTEMS

OFERO ABAGON CAPARIÑO

A dissertation submitted in partial fulfillment of

the requirements for the degree of

DOCTOR OF PHILOSOPHY

WASHINGTON STATE UNIVERSITY

Department of Biological Systems Engineering

AUGUST 2012

To the Faculty of Washington State University:

The members of the Committee appointed to examine the dissertation of OFERO ABAGON CAPARIÑO find it satisfactory and recommend that it be accepted.

______Juming Tang, Ph.D., Chair

______Caleb I. Nindo, Ph.D.

______Shyam S. Sablani, Ph.D.

______Joseph R. Powers, Ph.D.

______John K. Fellman, Ph.D.

ACKNOWLEDGMENTS

The completion of this dissertation and my Ph.D. program would not have been possible without the generous help from many individuals and institutions.

I would like to express my deepest gratitude to my committee chair and academic adviser

Dr. Juming Tang for giving me this great opportunity to work with him throughout my Ph.D. program. I am thankful to him for sharing his expert and scientific advice from conceptualization until completion of my dissertation. His generous support for providing me additional funding beyond my Ford Foundation scholarship program to complete my program, his continuous guidance in ascertaining the quality of this work, and his encouragement and motivation are deeply appreciated.

I am grateful to all my committee members: Dr. Caleb I. Nindo, Dr. Shyam S. Sablani,

Dr. Joseph R. Powers and Dr. John K. Fellman for sharing their time and technical expertise on the different topic areas of my research. I am also thankful to them for allowing me to use their respective laboratories in carrying out various analyses of my samples. Their constructive criticism and comments, editorial comments, enlightening suggestions and advice in coming up with this dissertation are very much appreciated.

My deepest appreciation to the Ford Foundation International Fellowship Program

(IFP)/Institute of International Education (IIE)-New York for granting me scholarships for the first 3.5 years of my Ph.D. program; the Washington State University graduate research assistantship program for awarding me additional funding support for 1.5 years, through the help of my academic adviser Dr. Tang; the IFP-Philippines Social Science Council (IFP-PSCC) for efficiently administering my IFP/IIE scholarship grant; and my home institution, the Philippine

Center for Postharvest Development and Mechanization-PhilMech (formerly the Bureau of

iii

Postharvest Research and Extension or BPRE) for allowing me a 5-year study leave with continued salaries and benefits.

My sincerest gratitude to Director Luisa Fernan and Asst. Director Criselda Doble of IFP-

Philippines and their staff, and the management and staff of IFP/IIE in New York for their ongoing guidance and for ensuring the timely release of all my scholarship benefits; the IFP selection committee for their trust and confidence in me; Ateneo de Manila University and

University of the Philippines for hosting the IFP pre-academic training course, and the WSU

Intensive American Language Center for preparing me to overcome the many challenges in my graduate courses. I also thank the WSU Office of International Students and Scholars for their guidance and advice in maintaining my legal status in the US.

My great appreciation to Executive Director Ricardo L. Cachuela, Deputy Directors

Rosendo Rapusas and Arnel Apaga, and all the Department Directors, Asst. Department

Directors, Division Chiefs and Section Chiefs of BPRE/PhilMech and their respective staff for their unwavering support and trust in allowing me to pursue my doctoral studies in the US. My sincere gratitude to Chief Ronaldo Reyes who has served as my surety and guarantor during my study leave; I am grateful to Engr. Reynaldo Gregorio, Dr. Cristina Gragasin, and Engr. Baldwin

Jallorina for serving as my OICs, at the former Physical Processes Research Division and the newly created Bioprocess Engineering Division where I am the chief; Dr. Manolito Bulaong for overseeing the division's research plans and programs; to my former supervisor at the

Postharvest Engineering Department Engr. Ruben Manalabe for allowing me to pursue higher education, and to my colleagues in the department, Dr. Romy, Lorie, Aileen, Sheila, Arlene,

Beth, James, Andy, Monch, Donald, Rob, Egay, Reagan, JOs and service contractors for their full cooperation and support.

I am especially thankful to Richard E. Magoon and Karin M. Bolland of MCD

Technologies, Inc. for allowing me to use their novel drying facilities in the US; Ramar

International, CA for helping me in importing mango puree samples from the Philippines; Dr.

Boon Chew and Ms. Bridget Mathison for allowing me to use their laboratory and for helping me in the nutritional analysis of my research samples; Frank Younce, Galina Mikhaylenko,

Roopesh Syamaladevi, Binying Ye, Scott Mattinson, and Valerie Lynch-Holm for providing their technical assistance on the proper operation of different laboratory equipment and instruments; Mr. Clifton Coy of KSM Enterprises, WA for providing free aluminum packaging materials; Dr. Claudio Stöckle, chair of the Biological Systems Engineering department for his full support in completing my program; John Anderson, Joan Hagedorn, Pat King and Pat

Huggins for ascertaining the timely delivery of the needed research materials and other administrative matters; Wayne Dewitt for helping in the fabrication of some of my experimental devices and Vincent Himsl for IT support and assistance; Fermin Resurreccion and Medy

Villamor for their invaluable suggestions that helped in improving many aspects of my research; members of Dr. Tang's research group and Food Engineering Club, Galina, Stewart, Dr. Liu, Dr.

Zhongwei, Dr. Shaojin, Huimin, Peter, Yang, Wenjia, Jing, Donglei, Bandar, Shunshan,

Rossanna, Ellen, Sumeet, Rajat, Sunil, Kanishka, Balu and Gopal for the good camaraderie and for sharing some important information relevant to my research; Stewart Bohnet for proof reading/editing of my published article and Jeannie Bagby for painstakingly editing my dissertation.

Special thanks goes to Jeff and Rina and their sons Carl and Cedrick for their care, understanding, and moral support while I stayed with their family throughout my studies. I am also thankful to Kuya Lito and Ate Beth, Richard and Jane, Damian and Weng, Randy and

Michelle, John and Janet, Clark and Ate Nita, Greg and Yeyet, and their respective families; Tito

Benny and Tita Delia, Tito Cesar and Ate Nellie, Dan and Medy, Ed and Julie, Iah, Rhoda, Avi,

Anne, James, Chromewell, Jeff and many other friends in Pullman and Moscow, Idaho. Their

warm welcome, friendship and camaraderie provided much needed relief, especially during occasional potlucks and special gatherings.

My greatest appreciation to my father who inspired me greatly; he died just a couple of days before my final interview with the IFP scholarships program. Due to a difficult situation at that time, I waived my rights as a finalist and I recommended my slot to other applicants. But, in spite of this circumstance, still, I was granted the Ford Foundation fellowship award. My father's memory will live forever in this piece of work. I am also grateful to my loving mother, Sening, for taking care of my children for a while when my wife was also in the US, my brother Pendoy, sisters Jocelyn, Gina and Yeng and their spouses Lilia, Panong and Jong and their respective families for their continuous prayers, moral support and assistance whenever my children needed their help. Special thanks to my mother-in-law Nanay Elen, brothers- and sisters-in-law and their spouses and families (Kuya Rene and Ate Rezie, Kuya Wilie and Ate Nitz, Levnie and Ditse,

Onion and Malou, Eric and Lea) and Jojo for their moral support and prayers. I also thank my many other immediate relatives, friends and colleagues who showed their concerns and prayers towards achieving my degree but whose names were not included in these acknowledgments.

My sincere appreciation also goes to my brothers & sisters in the Couples for Christ and

Oasis of Love-Marriage Encounter, and Fr. Rudy Ibale of CLSU for their moral support and prayers.

Lastly, I would like to express my deepest gratitude to my dearest wife Edith, and my three children, Iris, Dondon and Gelo (whom I saw personally only once throughout my 5-year stay in the US) for their love and concern, understanding, tolerance and constant prayers for the successful completion of my Ph.D. studies.

Above all, to our Lord Almighty for giving me the knowledge, wisdom, skills, patience, strength, good health perseverance and strong determination in the fulfillment of this dissertation and for finally obtaining my doctoral degree.

MANGO (Philippine 'Carabao' var.) POWDER MADE

FROM DIFFERENT DRYING SYSTEMS

ABSTRACT

by Ofero Abagon Capariño, Ph.D. Washington State University August 2012

Chair: Juming Tang

Mango (Mangifera indica L.) is one of the finest tropical fruits in the world with about

75% of the world production coming from Asia. In the Philippines, mango ranks third among fruit crops after banana and pineapple based on export volume and value. Acclaimed to be one of the best worldwide, this fruit has established a good reputation in the international market.

However, huge postharvest losses ranging 87 % have been reported due mainly to inadequate preservation technologies, and improper handling and storage. In this research, a more stable product such as mango powder was investigated using a novel drying technology called

Refractance Window® drying (RW), and three other commonly used drying methods, namely: freeze drying (FD), drum drying (DD) and spray drying (SD). The influence of these four drying methods on the physical properties and microstructures of mango powder was studied. RW drying can produce mango powder with quality that is comparable to freeze drying, and better than the drum and spray-dried mango powders. Water sorption characteristics and glass transition temperatures of mango powder was examined to understand water mobility within the mango solids. Physical and chemical stability of mango powder were analyzed during storage at different temperatures and by subjecting them to different packaging atmospheres. Nitrogen

vii

flush packaging was effective in preserving the ascorbic acid (AA) in mango powder at room temperature and under refrigerated conditions, while educed pe centage l ss β-carotene was observed after 6 months of storage. Regardless of packaging atmosphere, mango powder stored at 45 °C suffered discoloration as well as AA and β-carotene degradation over a period of 6 and

12 months. Mango powder has high concentration of low molecular weight sugars which causes the product to become sticky when exposed to high temperatures and humid conditions. To avoid occurrence of this phenomenon, several sticky-point temperature measuring devices were explored in the past, but there is still a need for further development. A new method to characterize the sticky phenomena in mango powder was investigated using an advanced rheometer. The developed method and protocol was found suitable to characterize the sticky point temperature of mango powder with high degree of repeatability and accuracy.

viii

TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS ...... iii

ABSTRACT ...... vii

LIST OF TABLES ...... xvi

LIST OF FIGURES ...... xviii

CHAPTER ONE

INTRODUCTION ...... 1

1. BACKGROUND ...... 1

2. OBJECTIVES ...... 4

3. SCOPE AND LIMITATIONS ...... 5

4. RESEARCH LAYOUT ...... 5

REFERENCES ...... 8

CHAPTER TWO

REVIEW OF DEHYDRATION IN FRUITS AND VEGETABLES ...... 13

1. INTRODUCTION ...... 13

2. SOME FUNDAMENTAL MATERIAL PROPERTIES RELATED TO FOOD

DEHYDRATION ...... 15

2.1 Concept of water activity ...... 15

2.2 Moisture sorption isotherms ...... 17

2.3 Hysteresis ...... 19

2.4. Sorption isotherm models ...... 20

2.5. Glass transition temperature ...... 22

3. DEHYDRATION METHODS AND SYSTEMS ...... 28

3.1 Sun drying ...... 28

3.2 Solar drying ...... 29

3.3 Hot air drying ...... 31

3.3.1. Cabinet or tray dryers...... 32

3.3.2. Tunnel dryers ...... 33

3.3.3. Belt conveyor dryers ...... 34

3.3.4. Pneumatic conveyor dryers ...... 36

3.3.5. Fluidized bed dryers ...... 37

3.3.8. Spray dryers ...... 38

3.3.9. Drum dryers ...... 41

3.4. Freeze drying ...... 43

3.5. Novel and Innovative Drying Technologies ...... 47

3.5.1. Refractance Window® drying ...... 47

3.5.2. Microwave drying ...... 49

3.5.3. Ultrasonic drying ...... 53

5. CONCLUDING REMARKS ...... 56

REFERENCES ...... 57

CHAPTER THREE

EFFECT OF DRYING METHODS ON THE PHYSICAL PROPERTIES AND

MICROSTRUCTURES OF MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER ...... 68

1. INTRODUCTION ...... 69

2. MATERIALS AND METHODS ...... 72

2.1 Preparation of mango puree ...... 72

2.2. Drying experiment ...... 73

2.2.1. Refractance Window® drying ...... 73

2.2.2. Freeze drying...... 75

2.2.3. Drum drying...... 75

2.2.4. Spray drying...... 76

2.3. Handling and packaging of samples...... 77

2.4. Grinding and sieving ...... 77

2.5. Water content ...... 78

2.6. Water activity ...... 78

2.7. Physical properties of mango powders ...... 78

2.7.1. Color analysis ...... 78

2.7.2. Bulk density ...... 79

2.7.3. Particle density and bulk porosity ...... 80

2.7.4. Solubility ...... 80

2.7.5. Hygroscopicity ...... 81

2.8. Glass transition temperature ...... 81

2.9. X-ray diffraction ...... 82

2.10 Microstructure analyses ...... 82

2.11. Statistical analysis ...... 83

3. RESULTS AND DISCUSSION ...... 83

3.1. Residence time, water content and product temperature ...... 83

3.2. Physical properties of mango powder ...... 84

3.2.1. Color analysis...... 84

3.2.2. Bulk density and porosity ...... 87

3.2.3. Solubility ...... 89

3.2.4. Hygroscopicity ...... 90

3.3. Glass transition temperature ...... 92

3.4. X-Ray diffraction ...... 92

3.5 Microstructures ...... 93

4. CONCLUSIONS...... 95

ACKNOWLEDGMENTS ...... 96

REFERENCES ...... 98

CHAPTER FOUR

WATER SORPTION CHARACTERISTICS, GLASS TRANSITION

TEMPERATURES AND MICROSTRUCTURES OF REFRACTANCE WINDOW®-

AND FREEZE-DRIED MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER ...... 123

1. IINTRODUCTION ...... 124

2. MATERIIALS AND METHODS ...... 126

2.1 Preparation of mango powder and packaging ...... 126

2.2. Measurements of residence time, product temperature and water content ...... 127

2.3. Determination of sorption isotherms ...... 128

2.4. Thermal glass transition ...... 129

2.5. Microstructures of mango powders ...... 131

2.6. Statistical analysis ...... 131

3. RRESULTS AND DISCUSSION ...... 131

xii

3.1. Drying time, water content and product temperature ...... 131

3.2. Water sorption isotherms ...... 131

3.3. Glass transition temperature ...... 133

3.4. Microstructures ...... 137

4. CONCLUSIONS...... 139

ACHKNOWLEDGMENTS ...... 139

REFERENCES ...... 141

CHAPTER FIVE

PHYSICAL AND CHEMICAL STABILITY OF REFRACTANCE WINDOW®-DRIED

MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER DURING STORAGE ...... 156

1. INTRODUCTION ...... 157

2. MATERIALS and METHODS...... 159

2.1. Preparation of mango flakes ...... 159

2.2 Sieving, packaging and storage...... 160

2.3. Water content and water activity ...... 161

2.4. Physicochemical and thermal characteristics ...... 161

2.4.1. Headspace gas ...... 161

2.4.2 Color measurement ...... 162

2.4.3. L-Ascorbic Acid analysis ...... 162

2.4.4. β-carotene analysis ...... 163

2.5. Microstructures of mango powders ...... 164

2.6. Statistical analysis ...... 165

3. RESULTS and DISCUSSION ...... 165

xiii

3.1. Water content and water activity ...... 165

3.2. Headspace gas analysis ...... 166

3.3. Color stability...... 167

3.4. L-Ascorbic Acid...... 170

3.5. β-carotene ...... 173

3.6. Caking and glass transition temperature ...... 175

3.7 Microstructures ...... 176

4. CONCLUSIONS...... 178

ACKNOWLEDGMENTS ...... 179

REFERENCES ...... 181

CHAPTER SIX

RHEOLOGICAL MEASUREMENTS FOR CHARACTERIZING STICKY POINT

TEMPERATURE OF FRUIT POWDERS: AN EXPERIMENTAL INVESTIGATION ...... 201

1. INTRODUCTION ...... 202

2. MATERIALS and METHODS...... 207

2.1 Preparation of samples ...... 207

2.2. Conditioning of samples at different water content ...... 207

2.3. Determination of water content ...... 209

2.4. Characterization of sticky point temperature using the rheological method ...... 209

2.4.1 Rheometer system ...... 209

2.4.2. Fabrication of powder loading device (PLD) ...... 210

2.4.3. Loading of samples ...... 210

2.4.4. Rheological settings and measurements ...... 210

xiv

2.5. Measurement of glass transition temperature ...... 212

2.6. Comparison of the sticky-point temperature in the present study with

published literature...... 212

3.0. RESULTS and DISCUSSION ...... 213

3.1 Fundamental basis for characterizing sticky point temperature using the

rheological method...... 213

3.2. Storage and loss modulus-time profile as it relates to sticky point temperature ...... 215

3.3. Sticky-point temperature of RW-dried mango powder using the rheometer

method...... 216

3.4. Relationship between sticky-point temperature and glass transition

temperature ...... 217

3.5. Comparison of the glass transition temperature of RW-dried mango powder

with published data...... 219

3.6. Comparison of the measured sticky-point temperature with published data ...... 220

4.0. CONCLUSIONS...... 221

ACKNOWLEDGMENTS ...... 222

REFERENCES ...... 223

CHAPTER SEVEN

CONCLUSIONS AND RECOMMENDATIONS ...... 249

LIST OF TABLES

Page

CHAPTER TWO

Table 1. Water activities of saturated salt solutions used in the determination of water

sorption at temperature typical of food storage...... 22

Table 2. Performance of MCSTD dryer for selected fruits, vegetables and fish...... 31

Table 3. General technical data for vegetable dehydration using tunnel dryer...... 34

Table 4. Process conditions and rehydration characteristics of selected vegetables dried

using a pneumatic conveyor dryer...... 36

CHAPTER THREE

Table 1. Drying conditions for production of mango powders using different methods...... 106

Table 2. Hunter color measurements of reconstituted mango powders obtained from

different drying processes...... 106

Table 3. Solubility and hygroscopicity of RW-, freeze-, drum-, and spray-dried mango

powders with particle size 180-250 m...... 107

Table 4. Glass transition temperatures and water activity of RW-, freeze-, drum-, and

spray-dried mango powders...... 107

CHAPTER FOUR

Table 1. Temperature, retention time and water content of mango puree and

Refractance Window- and freeze-dried mango powders...... 147

Table 2. Measured GAB parameters of mango powders and other sugar-rich fruits ...... 147

Table 3. Glass transition temperatures and moisture contents of Refractance Window-

and freeze-d ied p wde s at wate activity (0.113 ≤ aw ≤ 0.860)...... 148

xvi

Table 4. Evaluating water sorption isotherm and glass transition models of Refractance

Window- and freeze-dried mango powders...... 148

CHAPTER FIVE

Table 1. Water content and water activity of stored mango powder after 6 and 12

months of storage at different packaging methods and temperatures...... 189

Table 2. Headspace gas concentration of RW-dried mango flakes after 6 and 12

months of storage...... 190

Table 3. Hunter color measurements of mango flakes or powders stored in air or

nitrogen atmosphere at 5, 22 and 45°C and evaluated after 6 and 12 months...... 191

Table 4. Hunter color measurements of reconstituted mango powders stored in air or

nitrogen atmosphere at at 5, 22 and 45°C and evaluated after 6 and 12 months...... 192

CHAPTER SIX

Table 1. Sticky point temperature of RW-dried mango powder with particle size 180

m and 500 m...... 228

Table 2. data of water content, sticky point temperature and glass transition

temperature of mango powder at different water content...... 228

Table 3. Differences between the sticky point temperature and glass transition

temperature of mango powder at different water content...... 229

Table 4. Glass transition temperatures of selected fruits, unhydrous sugars and

carbohydrate polymers...... 230

xvii

LIST OF FIGURES

Page

CHAPTER TWO

Figure 1. Relationships between water activity and deteriorative reactions on foods ...... 16

Figure 2. A typical moisture isotherms showing different physicochemical state of food ...... 17

Figure 3. A typical water sorption isotherms ...... 18

Figure 4. Water sorption hysteresis ...... 20

Figure 5. Schematic diagram illustrating the transition between a supercooled liquid

state (rubber) and an amorphous solid state (glass) ...... 24

Figure 6. Heat flow influencing in sun drying ...... 28

Figure 7. Schematic diagram of multi-commodity solar tunnel dryer ...... 30

Figure 8. Schematic diagram of a cabinet 'Esteban' dryer ...... 32

Figure 9. Schematic diagram of a concurrent-flow tunnel dryer ...... 33

Figure 10. Schematic diagram of multi-pass conveyor belt dryer ...... 35

Figure 11. Schematic diagram of a fluidized bed dryer ...... 37

Figure 12. Schematic of a typical spray dryer and spray drying system with external

vibrating fluidized bed ...... 39

Figure 13. Schematic diagram of a single drum dryer, a double drum dryer, a twin-drum

dryer, and a vacuum drum dryer ...... 42

Figure 14. Phase diagram for water showing sublimation of ice ...... 44

Figure 15. Schematic diagram of a freeze dryer ...... 45

Figure 16. Schematic of a pilot scale Refractance Window® drying system ...... 48

Figure 17. Microwave-vacuum drying vs. air drying ...... 51

xviii

Figure 18. Microwave and spouted bed combined (MWSB) drying system...... 52

Figure 19. Experimental setup of microwave vacuum dryer...... 52

Figure 20. Schematic illustration of ultrasonic-assisted hot air dryer ...... 54

Figure 21. Experimental dehydration setup for air-borne ultrasonic drying of liquid or

direct contact to a vibrating solid material ...... 55

CHAPTER THREE

Figure 1. Mango puree samples ...... 108

Figure 2. Schematic layout of Refractance Window® dryer ...... 108

Figure 3. Measurements of product temperature of mango puree ...... 109

Figure 4. Monitoring of hot water temperature ...... 109

Figure 5. Refractance Window drying of mango puree ...... 110

Figure 6. Freeze drying of mango puree ...... 111

Figure 7. Drum drying of mango puree ...... 111

Figure 8. Spray drying of mango puree...... 112

Figure 9. Photograph of mango flakes or powders at different particle sizes obtained

from Refractance Window® drying, freeze drying, drum drying, and spray

drying...... 113

Figure 10. Lightness, yellowness of mango flakes or powders at different particle sizes

obtained from Refractance Window® (RW) drying, freeze drying, drum

drying, and spray drying...... 114

Figure 11. Chroma and hue angle of mango flakes or powders at different particle sizes

obtained from Refractance Window® drying, freeze drying, drum drying, and

spray drying ...... 115

xix

Figure 12. Photograph of reconstituted mango powders obtained from Refractance

Window® drying, freeze drying, drum drying, and spray drying...... 116

Figure 13. Bulk density of mango powders obtained from Refractance Window® drying,

freeze drying, drum drying, and spray drying...... 117

Figure 14. Porosity of mango powders obtained from Refractance Window® drying,

freeze drying, drum drying, and spray drying...... 118

Figure 15. X-ray diffraction patterns of Refractance Window®-, freeze-, drum- and

spray-dried mango powders with particle size 180-2 0 μ and aw < 0.2...... 119

Figure 16. Scanning electron micrographs (SEM) of mango powders (180-2 0 μ ) d ied

using Refractance Window® drying...... 120

Figure 17. Scanning electron micrographs (SEM) of individual mango powder particles

(180-2 0 μ ) d ied using Re actance Wind w® drying, freeze drying, drum

drying and spray drying ...... 121

Figure 18. Field emission scanning electron micrographs (FESEM) of Refractance

Window®-dried, freeze-dried, drum-dried and spray-dried mango powders

(180-250 mm) stored for 7 days at 25 oC with RH=75.5 %...... 122

CHPATER FOUR

Figure 1. Experimental set-up for determining water sorption isotherms of RW- and

freeze-dried mango powders...... 149

Figure 2. Differential scanning calorimeter ...... 149

Figure 3. Water adsorption isotherm data for Refractance Window- and freeze-dried

mango powders at 23 °C, with fitted curves using GAB and BET models...... 150

Figure 4. Glass transition temperatures of Refractance Window-dried and freeze-dried

mango powders equilibrated over selected water activity...... 151

Figure 5. State diagram of Refractance Window- and freeze-dried mango powders...... 152

Figure 6. Water plasticization and sorption characteristics of Refractance Window-dried

mango powders showing water activity, water content and glass transition

temperature...... 153

Figure 7. Water plasticization and sorption characteristics of freeze-dried mango

powders showing water activity, water content and glass transition

temperature...... 154

Figure 8. Scanning electron micrographs (SEM) of Refractance Window- and freeze-

dried mango powders ...... 155

CHAPTER FIVE

Figure 1. General procedure followed during storage of Refractance Window®-dried

mango powder...... 193

Figure 2. Nitrogen-flushed packaging and air-packaging of mango powder before

storage ...... 194

Figure 3. Collection of gas sample from headspace ...... 194

Figure 4. Gene al p cedu e du ing analysis β-carotene of mango powder ...... 195

Figure 5. Photographs of RW-dried mango powder and reconstituted mango puree after

6 and 12 months of storage at different packaging atmosphere...... 196

Figure 6. Influence of storage time, temperature and packaging atmosphere on the

ascorbic acid content of RW-dried mango powder...... 197

xxi

Figure 7. In luence st age ti e, te pe atu e and packaging at sphe e n the β-

carotene content of RW-dried mango powder...... 198

Figure 8. Caking of RW-dried mango powder after 6 and 12 months of storage at 45 °C...... 199

Figure 9. Field emission scanning electron micrographs (FESEM) of Refractance

Window®(180-2 0 μ ) st ed 1 yea at di e ent storage conditions ...... 200

CHPATER SIX

Figure 1. Schematic diagram of some methods and devices in characterizing sticky-

point temperature measuring device...... 231

Figure 2. Standardized procedure during moisture conditioning of mango powder

samples ...... 232

Figure 3. A developed standardized loading procedure of samples that were subjected

for conditioning at ~100% RH ...... 233

Figure 4. Schematic diagram of the rheometer system used in determining the sticky

point temperature of mango powder...... 234

Figure 5. A developed fabricated powder loader (PLD) and scraper and procedure of

loading the samples in the lower heated serrated plate of the rheometer...... 235

Figure 6. Proposed mechanism of sticking of sugar-rich amorphous fruit powders

applying storage modulus (G’) and loss modulus (G”)...... 236

Figure 7. Typical rheological profile of loss and storage modulus versus time of

sa ples 2 0 °C at a ramping increment of 10 °C...... 237

Figure 8. End-point sticky point temperature (Tse) of RW-dried mango powder at 180

and 500 m particle size and water content of 0.039 kg water/kg dry solids...... 238

xxii

Figure 9. Average sticky point temperature (Tsa) of RW-dried mango powder at 180 and

500 m particle size...... 239

Figure 10. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.003 kg water/kg dry solids...... 240

Figure 11. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.022 kg water/kg dry solids...... 241

Figure 12. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.029 kg water/kg dry solids...... 242

Figure 13. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.039 kg water/kg dry solids...... 243

Figure 14. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.048 kg water/kg dry solids...... 244

Figure 15. End point (Tse) and average (Tsa) sticky point temperature of samples dried at

water content of 0.067 kg water/kg dry solids...... 245

Figure 16. Relationships of sticky point temperature and glass transition temperature of

RW-dried mango powder as function of water content...... 246

Figure 17. Comparison of glass transition temperature of sucrose-fructose model food,

RW-dried, and vacuum-dried mango powders...... 247

Figure 18. Sticky point temperatures of RW-dried mango powder and other sugar-rich

fruit powders...... 248

xxiii

Dedication

This dissertation is dedicated to my wife Edith and my children, Iris, Dodon and Gelo.

xxiv

CHAPTER ONE

INTRODUCTION

1. BACKGROUND

Fruits and vegetables are abundant sources of human diet and nutrition (Liu, 2003). The

World Health Organization (WHO) estimated that sufficient daily intake of fruits and vegetables could potentially save approximately 2.7 million lives and help prevent major diseases such as cardiovascular diseases (CVDs) and certain cancers (WHO, 2003). The recent health awareness ca paign “Diets ich in uits and vegetables may reduce the risk of some types of cancer and

the ch nic diseases” ec ended by the F d and D ug Ad inist ati n (FDA) d processors has led to the rise in consumption of fruit products including tropical fruits such as mango, banana, pineapple, passion fruits and others (Occena-Po, 2006). According to the Food and Agriculture Organization of the United Nations (FAO), the estimated world production of tropical fruits in 2009 was 82.2 million tonnes with 98 percent produced in developing countries.

It is reported that mango (Mangifera indica L.) is the most dominant tropical fruit comprising a global output of 31.7 million tonnes or 39 % of the world production. About 75 % of mangoes worldwide is produced in Asia, while 14 % and 11 % is produced in Latin America and A ica, espectively. Acc ding t the sa e s u ce, the w ld’s biggest ang p duce is

India (40 %) of the total production followed by China (11 %), Thailand (5.3 %), Pakistan (5.1

%), Indonesia (4.9 %), Philippines (2.9 %) and Viet Nam (1.1 %).

In the Philippines, mango ranks third among the fruit crops next to banana and pineapple in terms of export volume and value, with a total of 1,023,907 metric tons harvested in 2007, supporting about 2.5 million farmers (BAS, 2009). The Carabao variety popularly known as

“Philippine Supe Mang ” acc unts 73 % the c unt y's p ducti n (BAS, 200 ). This variety is acclaimed as one of the best in the world due to its sweetness and non-fibrous flesh.

The fruit is rich in ascorbic acids and β-carotene.

However, mango is highly perishable because of its high moisture (> 80 % wb) and may deteriorate in a short period of time if improperly handled, resulting in quality loss and total spoilage (Liu, 2003). Serrano (2005) reported that losses in fresh mango ranged from 5% to 87% due to improper handling and storage. Gonzalez-Aguilar (2007) also mentioned in his paper that

100% of untreated ripe mango fruits of the 'Hadin' variety showed fungal infection and severe decay damage by the end of 18 days of storage at 25 °C. A wide range of post-harvest losses in fruits and vegetables both in developed and developing countries were reported to vary anywhere from 5 - 80% along the supply chain starting from harvesting, packing, storage, transportation, retailing and consumption. The combined effect of lack of postharvest facilities and technical knowhow by the farmers on proper handling and storage mostly contributed to these losses

(Jayaraman et al., 1992 & 2006; FTTC Database on Asian Agriculture, 1993; Singh, 2001,

Panhwar, 2006; FAO, 2011). As a consequence, availability of mango supplies is reduced; growers and those involved with the food handling chain also suffer major financial losses

Aside from the large postharvest losses, it was also reported that raw fruits including mango served as potential carrier of pathogenic microorganisms that caused human disease worldwide (WHO, 2003). Several incidence of outbreaks were reported associated with eating of raw fruits and vegetables (Creel, 1912; Melick, 1917; Altekruse et al., 1997) (Creel, 1912). Four documented Salmonella outbreaks have been associated with mangoes, occurring in 1998, 1999,

2001, and 2003 from Salmonella Oranienburg, Newport, and Saintpaul and Saint paul, respectively (Sivaapalasingam et al., 2003). The increase in outbreak were being highlighted due

to globalization of food supply mostly traded from developing to developed countries (Potter, et al., 1997; Altekruse et al., 1997; Altekruse and Swerdlow,1996). Recently, Strawn & Danyluk

(2010) reported Salmonella populations can grow on cut mangoes and papayas. They also found survival for at least 180 days of E. coli O157:H7 and Salmonella on frozen cut mangoes and papayas. Hence, these fruits are potential vectors for E. coil O157:H7 and Salmonella transmission.

The continuing advocacy effort by many concerned government and private institutions including their scientists and researchers to reduce postharvest losses and prevent contamination in fruits and vegetables are enormous. But, because of the nature of these commodities being highly perishable, contamination and losses are inevitable. Hence, there is a perceived need to process them into a more stable product that can be stored for a longer period without compromising their quality and safety.

Several preservation techniques can be applied to extend shelf life and add value to fruits and vegetables and one of them is drying and dehydration. Reported to be one of the oldest preservation technique, dehydration is continuously gaining significant interest among researchers, scientists, entrepreneurs and processors both in academia and industry. Dehydration is the process that involves the application of heat to remove moisture from the fresh fruit and vegetable product (Vega-Mercado et al.; 2001; Jarayamanan & Das Gupta, 2006). Its primary purpose is to reduce microbial activity and deterioration, and extend shelf life of the product.

Conversion of fresh fruits and vegetables into dried product has many advantages including substantial reduction of weight and volume and a potential decrease in transportation cost by as much as 85% when compared to its raw counterpart (Ratti, 2001). Depending on the purpose and

application of the dried product, fruits and vegetables can be dried into different forms such as whole, cut, sliced, flakes or powder using appropriate drying technologies.

The demand for dried fruits and vegetables in various forms is increasing. For instance, the estimated world demand for dried, dehydrated, and freeze-dried fruits and vegetables increased from $11.6 billion in 2001 to $16.12 billion in 2011(Parker, 2005). In 2006, the value of dried products was estimated at $12.7 billion, with the largest portion in Asia of $4.1 billion or

31.86 percent, followed by a combined share from Africa, Europe & the Middle East at $4.2 billion or 33.11 percent, and North America & the Caribbean with $3.3 billion or 25.67 percent of the world market (Parker, 2005).

Among the many dried food products, food powders such as fruit drink powders, flours, spices, instant coffee, and powdered milk represent the largest fraction available in the food industry for different applications (Barbosa-Canovas et al., 2005). One specific fruit drink product is mango powder known to be used as ingredients for health drinks, baby foods, sauces, marinades, confections, yogurt, ice cream, nutrition bars, baked goods and cereals (Rajkumar et al., 2007). Mango powder is also used in preparing natural mango-flavored beverages as well as pharmaceuticals and cosmetics (FAO, 2007). As the consumption of dried fruits and vegetables has been growing coupled with the demand for quality and safe products, innovative and novel drying techniques in manufacturing these products as well as their handling and storage methods should be explored.

2. OBJECTIVES

The overall objective of this dissertation was to convert mango fruits into a powder form and analyses them for various quality attributions. Specifically, the dissertation aims to: 1) determine the effect of drying methods on the physical properties and microstructures of mango

(Philippine 'Carabao' var.) powder; 2) investigate the influence of the sorption isotherms, glass transition and microstructures of mango powder dried using Refractance Window and freeze drying methods; 3) determine the stability of RW-dried mango powder during long term storage; and 4) develop a rheological method in characterizing sticky point temperature of mango powder.

3. SCOPE AND LIMITATIONS

The overall research focused on the conversion of mango fruits (Philippine 'carabao' var.) into a powder form. Due to difficulty in importing fresh mango from the Philippines, frozen mango puree with an average water content of 6.5 ± 0.1 kg water/kg dry solids was utilized in the different studies. The mango puree was produced in a commercial processing plant in the

Philippines, transported to the US under refrigerated c nditi n and st ed at 35 °C until the actual experimentation. The research used one novel and three commonly used drying systems, including Refractance Window® drying (RW), freeze drying (FD), drum drying (DD) and spray drying (SD). The dryers used throughout the entire experiments were pilot-scale models. Due to time constraints, not all the mango powders obtained from RW drying, freeze drying, drum drying, and spray drying were utilized in the different studies. Study on sorption isotherms and glass transition temperature used the RW- and freeze-dried mango powders, while the rheological and storage studies focused on the RW-dried mango powder. All the drying experiments were carried out at Washington State University (WSU), Pullman, WA, and MCD

Technologies, Inc., Tacoma, WA., while the subsequent laboratory analyses were done at WSU and University of Idaho, ID.

4. RESEARCH LAYOUT

The dissertation contains seven chapters, as follows:

Chapter One: Introduction - This chapter summarizes the importance of fruits and vegetables in the human diet, and constraints and challenges in the postharvest handling chain.

Potential prospects in developing value-adding products such as mango powder was explored.

This chapter also includes the scope, limitation and research objectives of the dissertation. A general dissertation outlines is also presented.

Chapter Two: Literature Review. This chapter discusses the fundamental material properties related to food dehydration such as water activity, sorption isotherms, glass transition temperture. Recent developments in dehydration of fruits and vegetables and their applications to fruit and vegetable products were reviewed.

Chapter Three: Effect of drying methods on the physical properties and microstructures of mango (Philippine 'Carabao' var.) Powder. The objective of this work was to investigate the influence of four drying methods (Refractance Window® drying, freeze drying, drum drying and spray drying) on the physical properties and microstructures of resulting mango powders to provide better understanding in selecting drying techniques that can be applied for the manufacture of high quality mango powder.

Chapter Four: Water sorption characteristic, glass transition temperatures and microstructures of Refractance Window®- and freeze-dried mango (Philippine 'Carabao' var.) powder. The study was conducted to expand understanding on the water mobility within the mango powder. Specifically, to investigate the influence of the sorption isotherms, glass transition and microstructures in mango powder dried using Refractance Window and freeze drying methods. The results of the study were used as reference on long-term storage of mango powder.

Chapter Five: Physical and chemical stability of Refractance Window®-dried mango

(Philippine 'Carabao' var.) powder during storage. This study was conducted to evaluate the physical and chemical stability of RW-dried mango powder during storage as affected by packaging atmosphere and storage temperature over a period of 12 months to aid in the manufacture of high quality mango powder.

Chapter Six: Rheological method in characterizing sticky point temperature of fruit powder: An experimental investigation. The objectives of the study were to: 1) develop a new method with high repeatability and reliability to characterize sticky point temperature of a model fruit (RW-dried mango powder) using a rheometer; and 2) find the relationships between the obtained sticky point temperatures and glass transition temperature of the model powder sample at different water content, and 3) Compare the sticky point temperature of fruit powders obtained from the proposed method with the existing data reported in the literature.

Chapter Seven: Conclusions and recommendations. The highlights of findings and recommendations of the present study were presented in this chapter.

Chapter three was already published in Journal of Food Engineering while Chapters four, five and six are ready for submission to selected referred journals.

REFERENCES

Altekruse, S. F., & Swerdlow, D. L. (1996). The changing epidemiology of foodborne diseases.

Am. J. Med.Sci. , 311, 23-29.

Altekruse, S. F., Cohen, M., & Swerdlow, D. (1997). Emerging foodborne diseases. Emerg.

Infect. Dis. , 3, 285-293.

Bansal, N. K., & Garg, H. P. (1987). Solar crop drying. In A. S. Mujumdar, Advances in drying

(Vol. 4, pp. 279-299). Washington D.C.: Hemisphere.

Barbosa-Cánovas, G. V., & Vega-Mercado, H. (1996). Dehydration of Foods. New York:

Chapman and Hall.

Bolin, H. R., & Salunkhe, D. K. (1982). Food dehydration by solar energy. Crit. Rev. Food Sci.

Nutri. , 16, 327–354.

Creel, R. H. (1912). Vegetables as a possible factor in dissemination of typhoid fever. Public

Health Rep. , 27, 187-193.

Ekechukwu, O. V. (1999). Review of solar-energy drying systems II: An overview of solar

drying technoloy. Energy Conservation and Mangament , 40, 615-655.

FAO. (2011, May 3-5). Meeting of Intergovernmental Group on Bananas and Tropical Fruits.

Retrieved January 25, 2012, from http://www.fao.org

Flink, J. M. (1977). Energy analysis in dehydation process. Food Technology , 31 (3), 77-79.

FTTC. (1993). Postharvest Losses of Fruits and Vegetables in Asia. Retrieved November 20,

2011, from Agriculture, Food & Fertilizer Technology Center (FTTC) Database on Asian

Agriculture: http://www.agnet.org/library/article

George, J. P., & Datta, A. K. (2002). Development and validation of heat and mass transfer

models for freeze drying of vegetables slices. Journal of Food Engineering , 52, 89-93.

Gonzalez-Aguilar, G. A., Zavaleta-Gatica, R., & Tiznado-Hernandez, M. (2007). Improving

p stha vest quality ang ‘Haden’ by UV-C treatment. Postharvest Biology and

Technology , 45, 108–116.

Grabowski, S., Marcotte, M., & Ramaswamy, H. S. (2003). Drying of fruits, vegetables and

spices. In A. Chakraverty, A. S. Mujumdar, G. S. Raghavan, & H. S. Ramaswamy,

Handbook of Postharvest Technology (pp. 653-696). New York: Marcel Dekker.

Imre, L. (1995). Solar drying. In A. S. Mujumdar, Handbook of Industrial Drying (pp. 373-452).

New York: Marcel Dekker, Inc.

Jayaramayan, K. S., & Das Gupta, D. K. (1992). Dehydration of fruits and vegetables - recent

developments in principles and techniques. Drying Technology , 10 (1), 1-50.

Jayaramayan, K. S., & Das Gupta, D. K. (2006). Drying of Fruits and Vegetables. In A. S.

MUJUMDAR, Handbook of Industrial Drying (3rd ed.). New York: Taylor and Francis

Group.

Kudra, T., & Mujumdar, A. S. (2002). Advanced Drying Technologies. New York: Marcel

Dekker, Inc.

Labuza, T. D. (1982). Shelf-life dating of food. Westport, CT: Food & Nutrition Press.

Labuza, T. K. (1985). Effect of temperature of the moisture sorption isotherm and water activity

of two dehydrated foods. Journal of Food Science , 50, 385-391.

Law, C. L., & & Mujumdar, A. S. (2008). Dehydration of fruits and vegetables. In A. S.

Mujumdar, Guide to industrial drying (pp. 223–249). India: Three S Colors.

Lawland, T. A. (1981). Agricultural and other low temperature applications of solar energy. In S.

E. Handbook, Kreider, J. F., Kreith, F. (p. 18.1). New York: McGraw Hill.

Liu, L. (2003). Entry into supermarket of agricultural products after entering WTO. Agricultural

Products Processing , 6 (5), 4-5.

Liu, Y., Zhao, Y., & Feng, X. (2008). Energy analysis for a freeze-drying process. Applied

Thermal Engineering , 28, 675-690.

Melick, C. O. (1917). The possibility of typhoid infection through vegetables. J. Infect. Dis. , 21,

28-38.

Muhlbauer, B. M., Bauer, W., & Kohler, B. (2002). Solar crop drying in developing countries.

Berne.

Mujumdar, A. S. (2008). Classification and selection of industrial dryers. In A. S. Mujumdar,

Guide to industrial drying (pp. 23–36). India: Three S Colors.

Occena-Po, L. G. (2006). Banana, Mango and Passion Fruits. In Y. H. Hui, Handbook of Fruits

and Fruit Processing (p. 635). Iowa: Blackwell Publishing.

Oetjen, G. W., & Haseley, P. (2004). Freeze-drying (2nd ed.). Germany: Wiley-VCH GmbH &

Coo. KGaA.

Ozgul-Evranuz, E. (2011). Drying vegetables: new technology, equipment, and examples. In N.

K. Sinha, Y. H. Hui, E. Ozgul Evranuz, M. Siddiq, & J. Ahmed, Handbook of vegetables

and vegetable processing (pp. 299-333). Iowa: Blackwell Publishing Ltd.

Panhwar, F. (2006, May). Report on: Post Harvest Technology of Fruits and Vegetables.

Retrieved November 8, 2011, from http://www.eco-web.com/editorial

Parker, B. F. (1991). Solar energy in agriculture. Amsterdam: Elsevier.

Parker, M. P. (2005). World Outlook Report 2006-2011: Dried, Dehydrated & Freeze-Dried

Fruits & Vegetables. Singapore and Fontainebleau, France: ICON Group International,

Inc.

Paterson, A. H., Bronlund, J. E., & Brooks, J. F. (2001). The blow test for measuring the

stickiness of powders. AIChE 2001 Annual Meeting. Reno, Hilton USA.

PhilMech. (2012). Multi-commodity solar tunnel dryer. Retrieved January 15, 2012, from

philmech.gov.ph

Potter, M. E., Motarjemi, Y., & Käferstein, F. K. (1997). Emerging foodborne diseases. World

Health , January-February, 16-17.

Potthast, K. (1978). Influence of water activity on enzymatic activity in biological systems. In J.

H. Crowe, & J. S. Clegg, Dry biological Systems (pp. 324-342). New York: Academic

Press.

Ramaswamy H., M. M. (2006). Food Processing: Principles and Applications. New York:

Taylor & Francis Group.

Ratti, C. (2009). Advances in Food Dehydration. New York: Taylor & Francis Group.

Ratti, C. (2001). Hot air and freeze-drying of high-value foods-a review. Journal of Food

Engineering , 49, 311–319.

Rizvi, S. S. (1995). Thermodynamic properties of foods in dehydration. In M. Rao, & S. S.

Rizvi, Engineering properties of foods (pp. 223-309). New York, USA: Marcel Dekker,

Inc.

Seiler, D. A. (1976). The stability of intermediate moisture foods with respect to mould growth.

In R. Davies, G. G. Birch, & K. J. Parker, Intermediate Moisture Foods (pp. 166-181).

London: Applied Science Publishers Ltd.

Singh, R. (2001). Postharvest Technology : Cereals, Pulses, Fruits and Vegetables. University of

California, Davis: Science Publishers. Plymouth.

Sivapalasingam, S., Barrett, E., Kimura, A., Van Duyne, S., DeWitt, W., Ying, Y., et al. (2003).

A multistate outbreak of Salmonella entericia serotype Newport infection linked to

mango consumption: impact of Water-dip disinfection technology. Clin. Infect. Dis. , 37,

1585-1590.

Sokhansanj, S., & Jaya, D. S. (1995). Drying of foodstuffs. In A. S. Mukumdar, Handbook of

Industrial Drying (pp. 589-625). New York: Marcel and Dekker, Inc.

Strawn, L. K. (2010). Fate of Escherichia coli O157:H7 and Salmonella spp. on fresh and frozen

cut mangoes and papayas. International Journal of Food Microbiology , 138 (1-2), 78-84.

Strumillo, C. (2006). Perspectives in developments in drying. Drying Technology , 24, 1059-

1068.

Vega-Mercado, H., Gongora-Nieto, M. M., & Barbosa-Canovas, G. V. (2001). Advances in

dehydration of foods. Journal of Food Engineering , 49, 271-289.

Weiss, W., & Buchinger, J. (n.d.). Solar drying. Retrieved October 19, 2011, from

http://www.aee-intec.at/

WHO. (2003). Fruit and Vegetable Promotion Initiative. Geneva: World Healt Organization

Technical Report Series, No. WHO/NMH/NPH/NNP/0308).

CHAPTER TWO

REVIEW OF DEHYDRATION IN FRUITS AND VEGETABLES

1. INTRODUCTION

Believed to be one of the oldest preservation techniques, dehydration is continuously gaining significant interest among researchers, scientists, entrepreneurs and processors, both in academia and in industry. Dehydration is a process that involves the application of heat to remove moisture from fresh fruit and vegetable products (Vega-Mercado et al., 2001;

Jarayamanan & Das Gupta, 2006). Its primary objective is to reduce microbial activity and deterioration, and to extend the shelf life of the product. Other benefits of dehydration include reduction of weight, which greatly decreases the cost of packaging, handling, storage and transport (Singh & Heldman, 2009).

Drying of fruits and vegetables offers a challenging task due to the different structural characteristics of these products. The removal of water from these products must be accomplished without compromising the physico-chemical properties and quality of the dried products. To obtain the desired quality products, some fundamental concepts associated with food dehydration such as water activity, sorption isotherms and glass transition temperature, among other concepts need to be understood. A selection of appropriate drying methods and systems must be analyzed based on the entire drying process. For example, there is a need to define the types and characteristics of raw materials to be dried, and the specifications of the finished product. Also, the desired production capacity, shape and particle size distribution, and drying characteristics of the product should also be critically assessed (Vega-Mercado et al.

2001).

Several dehydration methods and types of dryers are commercially available for drying a wide range of food products. Approximately 100 types of dryers of the more than 400 types reported in the literature (Strumillo, 2006) are being utilized in the industry. These dryers are classified into several categories, based on certain specific criteria. Vega-Mercado et al., 2001 divided dehydration technologies into four generations: first generation (i.e kiln, tray, truck tray, rotary flow conveyor and tunnel); second generation (i.e. spray and drum drying); third generation ( i.e. freeze-dying and osmotic dehydration); and fourth generation (i.e. Rrefractance

Window ®(RW), high-vacuum, fluidization, microwaves, and RF). Mujumdar (2008) and Law &

Mujumdar (2008) classified dehydration into four categories, namely, drying strategy, drying medium, method of handling of solids, and mode of heat input. Jarayamanan & Das Gupta

(2006) suggested three basic types of drying processes for fruits and vegetables such as sun or solar drying, atmospheric and continuous, sub-atmospheric drying, and low-temperature and low-energy drying process.

Regardless of the classifications and categorization of dehydration technologies, the application of a particular drying technique for fruits and vegetables depends on the type of raw material and its characteristics and intended used of the dried material. In most cases, the acceptance of drying methods and designs is highly dependent on the capacity, efficiency, investment cost, operational drying cost, and impact on the environment. Several authors have published books and articles focused on different drying techniques and technologies (Barbosa-

Cánovas & Vega-Mercado, 1996; Ekechukwu, 1999; Muhlbauer, et al., 2002; Kudra &

Mujumdar, 2002; Jayaraman & Das Gupta, 1992; Mujumdar, 2006; Ratti, 2009). The present review describes some fundamental concepts related to food dehydration and developments in dehydration of fruits and vegetables.

2. SOME FUNDAMENTAL MATERIAL PROPERTIES RELATED TO FOOD

DEHYDRATION

2.1 Concept of water activity

Water activity (aw) is an important parameter in food dehydration. Water activity is the equilibrium property of water at a given temperature and moisture content. Its relationship is defined as the ratio of the partial pressure of water in the headspace of the product (P) to the vapor pressure of pure water (Po) at the same temperature. When vapor and temperature equilibrium are obtained, the water activity of the sample is equal to the relative humidity of air surrounding the sample in a sealed measurement chamber. Multiplication of water activity by

100 gives the equilibrium relative humidity (ERH) in percent, expressed as (Toledo, 1999;

Ramaswamy and Marcotte, 2006):

The relationship of water activity to stability of equilibrium in a food system is very important. Unlike water content, water activity can be used to predict the stability of the food, e.g. shelf life, safety, texture, flavor, and smell in an equilibrium condition. Also, microorganisms as potential sources of spoilage and infection can be determined through water activity. Although temperature, pH and several other factors can influence the growth of organisms, water activity may be the most important factor in controlling spoilage in the food system and maintaining its chemical stability. In addition to influencing microbial spoilage, water activity can play a significant role in determining the activity of enzymes and vitamins in foods and can have a major impact on color, taste, and aroma. The relationship between water activity and deteriorative reactions on foods is shown in Fig. 1 (Labuza, 1982).

Figure 1. Relationships between water activity and deteriorative reactions on foods (Adapted from Labuza, 1982)

The inc ease wate activity up t 0.3 0.5aw in food decreases the rate of lipid oxidation, but stability of other deteriorative reactions, particularly non-enzymatic browning reactions, increases. Vertucci and Leopoldo (1984) reported that water molecules in soybeans with 0.4aw are tightly bound and not available for the reactions that occur during autoxidation.

Further increase in aw beyond 0.5 resulted in an increase in lipid oxidation. Noticeably, the rapid rate of non-enzymatic browning (NEB) is expected at intermediate water activity levels between

0.5 and 0.7, but slowed down beyond these values. Enzymatic reaction accelerates as the water content of food increases and this can be attributed to the increased mobility of enzymes as a function of water activity. The growth of molds and yeast are inhibited at water activity range of

0.7 and 0.8, while no bacteria can grow at water activity below 0.9 aw.

2.2 Moisture sorption isotherms

The relationship between water content of food and its corresponding water activity at a constant temperature is designated as water sorption isotherms (WSI). Their relationship is directly proportional. An increase in aw corresponds to an increase in the water content, but in a nonlinear trend. Because of the complexity of sorption processes, isotherms cannot be determined by calculation, but must be recorded experimentally for each product. A typical moisture isotherm of food is presented in Fig. 2. It can be seen that the variations in shape depends on the physical-chemical state (e.g. crystalline or amorphous), chemical composition

(e.g. oil, starch and protein) and microstructures of foods (Aguilera, 2003).

25 Anti-caking agent

15 Amorphous sucrose food 10 Crystalline Potato sucrose 5 starch

Water content(% dry basis) Peanut oil

0 0.2 0.4 0.6 0.8 1.0

Water Activity (aw)

Figure 2. A typical moisture isotherms showing different physical-chemical state of food

(Adapted from Aguilera, 2003)

The shape of moisture sorption isotherms is sigmoidal for most foods, and a moisture sorption isotherm prepared by adsorption (starting from the dry state) will not necessarily be the same as an isotherm prepared by desorption (starting from the wet state). A food is more stable against microbial spoilage when its aw is adjusted by absorption rather than by desorption. A hypothetical analysis of MSI can be described in three regions I, II and III (Fig. 3). Generally, foods in region 1 contain water content below 0.1, with g/g dry solids and water activity ranging from 0.2 0.4 (Aguilera, 2003). In this region, the water monolayer molecules present in the product are strongly bound to active sites mainly by hydrogen bonding. In such manner of binding, it is difficult to remove the water, and application of energy in excess of latent heat of vaporization is needed in order to overcome the bonds. It is widely accepted that this water is not available as a solvent to support chemical and microbiological activity, and hence food at this moisture level is very stable.

30 Region I Region II Region III 25

Strongly Less strongly bound Solvent and bound water layers and free water 20 monolayer capillary adsorbed basis) water

15 Desorption

Adsorption

5 Water content(% dry

0 0 0.2 0.4 0.6 0.8 1.0

Water activity (aw )

Figure 3. A typical water sorption isotherms showing three regions (Adopted from Seiler, 1976)

In region II, water begins to fill in the micro- and macropore spaces in the food system. At this condition, the mobility of the solutes increases, thus facilitating chemical and biochemical reactions, such as enzymatic reactions. Water molecules in this region are less bound to the solid to some extent. In region III, excess water is present in the food system as part of the fluid phase in high moisture products, and accounts for the depression of water vapor pressure of the wet material. The growth of microorganisms, molds, yeast and pathogenic bacteria in a food system with water activity ranging from 0.6 to 0.91 normally occurs in this region.

2.3 Hysteresis

Hysteresis in water sorption behavior, exhibited by many foods, is a phenomenon of different values obtained for water activity vs. moisture of foods, and the adsorption and desorption isotherms of the same product shows a different shape. Shown in Fig. 4 is an indicative water activity isotherm displaying the hysteresis often encountered depending on whether the water is being added to the dry material or removed (drying) from the wet material.

In foods containing high sugars and pectin, hysteresis is normally developed at the monolayer region and unlikely to occur at water activity above 0.65. A moderate hysteresis can be observed in high protein food with aw from 0 to 0.85, while a large hysteresis is discernible in starchy foods with a maximum aw of 0.7. The appearance of hysteresis is due to non-reversible structural changes and non-equilibrium effects. There are many empirical equations and tables that attempt to describe this behavior, but, although indicative, none predict with sufficient accuracy, and the water activity isotherm should be experimentally determined for each material.

In the food industry, such empirical equations combine contributions from the ingredients to give an estimate of aw, which is then used to estimate the mold-free shelf life (Seiler, 1976; Rizvi,

1995).

Strongly Less strongly bound Solvent and 20 bound water layers and free water

monolayer capillary adsorbed water basis)

15 Desorption Hysteresis

Adsorption

5 Water content(% dry

0 0 0.2 0.4 0.6 0.8 1.0

Water activity (aw )

Figure 4. Water sorption hysteresis(Adopted from Seiler (1976)

2.4. Sorption isotherm models

Modeling of sorption isotherms is important because it can aid in predicting the temperature ranges and determine the water content at certain critical aw values related to stability and safety of food. It is also important in predicting the right ingredient, such as in making a formulation at high moisture content, which maintains a safe aw level. Many mathematical models have already been developed and fitted to the isotherm curves. Some models are empirical in nature, while others were based on the thermodynamics of adsorption.

Due to complexities in structural and physical-chemical composition of different food matrices, no single isotherm model has been universally adopted to represent the sorption behavior of foods over the whole range of water activity. The most widely quoted isotherm model is the

Brunauer Emmett-Teller (BET) Model (Brunauer et al., 1938). The general equation of BET model is mathematically expressed as:

(1)

where Mw is the water content (kg water/kg dry solids); Mb is the BET monolayer water content

(dry basis); B is a constant related to net heat of sorption. The wider acceptance of the BET isotherm equation is due to its good fit to data for a range of physicochemical systems over the region 0.05 , 0.35 0.5 (Flade and Aworth, 2004; Labuza and Altunakar, 2007; Rahman, 1995).

It can provide a good estimate of the monolayer value of moisture absorbed on the surface. It is also recognized by the IUPAC (International Union of Pure and Applied Chemistry) as the standard for evaluation of monolayer and specific areas of sorbates. Although some assumptions about the BET model have been faulted, it has been found useful in determining the optimum moisture content and stability of foods during drying and storage.

The Guggenheim-Anderson-de Boer (GAB) Model has three simple parameters, which facilitate engineering calculation. The GAB equation is expressed as (Labuza, 1968):

(2)

where Mw is the water content (kg water/kg dry solids); Mg is the GAB monolayer water content

(dry basis); C is a constant related to the monolayer heat of sorption; and K is a factor related to the heat of sorption of the multilayer and the value of K varies from 0.7 to 1. The GAB model has several advantages: It can describe the sorption behavior of almost all foods from zero to 0.9 aw (Rizvi, 1995). The GAB model is also chosen in several isotherm studies because of its proven best fit in the desorption and adsorption of foods over a large range of water activities

(Prothon & Ahrné, 2003), for instance, the model which showed an excellent fit, as reported by

Labuza et al. (1985), where the GAB equation was applied to moisture isotherms for fish flour and corn meal. For food engineering applications related to water-interactions, the model proved to be a reliable and accurate equation for modeling and design work (Rizvi, 1995). Other

theoretical adsorption models are reported elsewhere (Chirife and Iglesias, 1978; van den Berg and Bruin, 1981).

It was reported that these models proved to be useful in predicting water sorption since little experimental data on relative vapor pressures are needed. A widely proven technique in obtaining sorption data is by equilibrating the samples in sealed humidity dessicators containing saturated salt solutions at a constant temperature, and measurement of water content by gravimetric method (Roos, 1995). Several salt solutions with different water activities can be found in Greenspan (1977) and Labuza et al. (1985). Table 1 shows the common salt solutions applied for sorption studies over a temperature range of 5 to 35 °C.

Table 1. Water activities of saturated salt solutions used in the determination of water sorption at temperature typical of food storage (Adapted from Roos, 1995a. Water activity Salt 5 °C 15 °C 25 °C 35 °C LiBr 0.074 0.069 0.064 0.597 LiCl 0.113 0.113 0.113 0.113

KCH3CO2 0.291 0.234 0.225 0.216

MgCl2 0.336 0.333 0.328 0.321

K2CO3 0.431 0.432 0.432 0.436

Mg(NO3)2 0.589 0.559 0.529 0.499

NaNO2 0.732 0.693 0.654 0.628

SrCL2 0.771 0.741 0.709 - NaCl 0.757 0.756 0.753 0.749

(NH4)2SO4 0.824 0.817 0.803 0.803 KCl 0.877 0.859 0.843 0.830

BaCl2 - 0.910 0.903 0.895

K2SO4 0.985 0.979 0.973 0.967

2.5. Glass transition temperature

Glass transition temperature (Tg) is a temperature at which the amorphous region between the glassy-solid and rubbery-liquid state occurs. Tg is one of the important factors

associated with dehydration of foods. It helps in understanding the textural properties of the material and explains the physical and chemical changes that may occur during food dehydration. It is also of great relevance in explaining different phenomena such as stickiness, caking, and softening and hardening of dried products during storage (Ratti, 2001; Kasapis,

2006). Tg is often considered as a second-order phase transition because of its apparent behavior at a very low rate during heating and cooling, which exhibits step changes in heat capacity and thermal expansion coefficient with temperature (Roos, 2005). However, Labuza et al. (2001) suggest consideration of Tg as a kinetic and relaxation transition because it occurs over a temperature range and it is dependent to time and method of measurements.

Glass transition temperature is classified as state transition rather than a phase transition because its inherent property is a non-equilibrium system (Slade & Levine, 1991). The physical state associated with Tg is the amorphous solid state. Amorphous solids are solids with random disoriented molecules such as those in a liquid state, unlike crystalline solids which have a fixed geometric pattern (Rahman, 1995). The glass transition occurs within a range of temperatures between solid state (glassy state) and a supercooled, highly viscous liquid with leathery, rubbery characteristics (Roos, 2003a,b). The transformation between supercooled liquid and amorphous solid known as glass transition is illustrated in Fig. 5. The solid arrow line shows that amorphous solids can be produced normally in two events: a decrease in water content by rapid evaporation of solvent molecules, and a decrease in temperature by a rapid cooling to prevent the formation of an equilibrium crystalline state of the material (Roos, 2003). On the other hand, glass transition can be reversible or bidirectional with hysteresis (Seyler, 1994) depending on the material being measured.

Glass Transition

Decrease in water content and/or temperature

Increase in water content and/or temperature

Supercooled liquid state Amorphous solid state (Rubber) (Glass)

Figure 5. Schematic diagram illustrating the transition between a supercooled liquid state (rubber) and an amorphous solid state (glass)(Adopted from Roos, 2003).

Glass transition temperature can be measured by monitoring the changes in thermal, mechanical, dielectric, volumetric properties and molecular mobility of the food components within a given range of Tg (Sablani et al., 2010). Different analytical techniques have been applied to determine the Tg in food biopolymers. The common methods are differential scanning calorimetry (DSC), thermomechanical analysis (TMA), dynamic mechanical analysis (DMA), dielectric analysis (DEA), and rheological method.

In DSC, the measurement involves a continuous monitoring of heat flow between the reference (sealed empty pan) and the test sample at a predetermined heating and cooling rate, usually from 5 to 10 C/min through the Tg region of the material. The resultant DSC curve provides the location of Tg manifested by the step change in the heat flow of the DSC thermogram. DSC is applicable to a wide range of products and it is convenient to use.

Hoewever, DSC techniques may not be sensitive to weak Tg and hence requires optimization of the procedure to obtain accurate Tg values (Sablani et al., 2010).

Determination of Tg by TMA method is carried out by monitoring the changes in dimensions of the food sample under non-oscillatory stress, including torsion, compression, and tension of flexure at a constant heating rate through its Tg region (Abad et al., 2009). During the measurement, a parallel plate geometry is normally adopted to obtain consistent results. The change in dimension associated from the change of the material from glassy to rubbery state is observed by a sensing probe that is in direct contact with the material during heating. The intersection of the tangent lines along the glassy and rubbery regions determines the glass transition temperature of the material (Earnest, 1994; Seyler, 1994; Roos, 1995; Sablani, et al.

2010). The TMA technique is sensitive in recognizing amorphous and crystalline material better than the DSC. But, this method cannot guarantee accurate results for some porous material due to the system's inability to maintain the water content of the product during heating (Farhat,

2004).

In DMA, a food material with specified dimension is placed in oscillation at either fixed or resonant frequency and the viscoelastic changes of the material are monitored by finding the storage modulus (E"), loss modulus (E") and mechanical loss factor (tan δ = E'/E") as a function of temperature. E' is related to the recoverable elastic energy stored in the material, while E" is related to the deformation energy dissipated as heat (De Graaf et al., 2003). A range of glass transition can be defined when a sharp decrease and increase in E" and E" is observed, respectively. The onset glass transition temperature is located at the temperature where E" reached its peak value, which also indicates the transition from a glassy to a rubbery state. The

DMA method in determining Tg is more sensitive than DSC when using food biopolymers as test material. One limitation in the application of DMA is the extended time for preparing the

samples, particularly the powder product, which may have led to deterioration prior to analysis

(Forssel et al., 1997).

A rheological method was proposed and applied in determining the glass transition temperature of a food material. The procedure was performed using a controlled strain rheometer. The test sample loaded in a circular plate was first scanned to temperature below Tg and followed by heating up above Tg. A master curve was established by plotting the log storage modulus (G') and loss modulus (G") as a function of temperature. The glass transition te pe atu e was dete ined by taking a te pe atu e ange between the inte secti n the Tan δ vs G' and Tan δ vs G", or when the values of G' were higher than G" as affected by the heat loss. This method has been applied in several food products such as gelatin, sugar mixtures, and dehydrated fruits and fish products (Kasapis, 2004).

Measurements of glass transition temperature by DEA thermal analysis involved placement of a food system between two parallel plate capacitors and applying a sinusoidal voltage to one of the plates to establish an electric field in the test sample. The generated field in one plate made the sample electrically polarized and conducted a small charge to the other plate.

The dielectric constant or permittivity and dielectric loss constant were determined by measuring the resultant current (Abiad et al., 2009). The glass transition temperatures were identified when dielectric and dielectric loss constants showed a sharp increase during the thermal scan (Bidstrup

& Day, 1994). The DEA was more sensitive compared to DSC and DMA when applied to wheat doughs (Laaksonen and Roos, 2003).

Detailed descriptions and procedures in determining Tg using the above methods have been extensively reviewed by Abiad et al., 2009, Roos, 2010, and Sablani et al., 2010.

It has been reported that no single Tg value can be obtained for a specific material.

Variations can occur in many factors including principles, methods and procedures, history of the samples, composition, and water content, among other factors (Sablani et al., 2010).

Different Tg results might also be influenced by the different protocols being considered within each method. For instance, in DSC, the Tg values obtained in maltose at a cooling and heating rate of ~100 K/ min and 10 K/min was 43.1 °C, while a Tg value of 41.2 °C was obtained when using the same samples and applied at an equal cooling and heating rate of 10 K / min (Schmidt,

2004). Since the e is n single eth d established as a “t ue” value, any the eth ds can be useful to predict the stability of a product according to specific application.

Glass transition temperature can greatly affect the stability and shelf life of a specific product. In general, low moisture foods are more stable when stored at temperatures below Tg (T

< Tg ), while stability decreases when the material is kept above Tg (T >Tg). (Slade & Levine,

1991), although the Tg measurement on product stability depends on the chemical, physical, biochemical and microbial growth of the material under study (Roos, 1995).

In many cases, Tg alone cannot predict the deterioration, stability and shelf-life of food products, and hence a combined water activity and glass transition temperature concepts has been proposed (Roos, 1995). Glass transition temperature as a function of water activity or water content can be used to construct a state diagram for a particular food system (Sablani et al.,

2010). Studies have shown strong evidence that using the state diagram can better assist the food industry in determining the stability of their products (Sablani et al., 2010). Several studies related to the interactions of water activity, water content and glass transition temperature of agricultural and fishery products have been summarized and reported (Syamaladevi et al., 2009;

Sablani et al., 2007).

3. DEHYDRATION METHODS AND SYSTEMS

3.1 Sun drying

In sun drying, the heat from the surrounding air and diffused radiation from the sun is transferred to the material being dried (Fig. 6). The heat generated is used to increase the temperature of the product causing the moisture or vapor to migrate from interior to the surface of the crop (Muhlbauer, 2002). Part of the radiant energy provides the heat to evaporate the water from the product surface by convection and radiation, which is finally removed from the surrounding air by convection through the application of forced air. In an ambient condition, the equilibrium moisture content of the product is attained when the vapor pressure of the product is equal to the atmospheric pressure. Sun drying is a simple and inexpensive method for drying fruits and vegetables. However, this method is only useful when sun is available. Other limitations include uncontrollable drying conditions, possible contamination from dust and insects, and possible microbial and enzymatic degradation as a result of longer drying time. Non- uniform drying may also lead to deterioration during storage.

Evaporative cooling Solar Heat Convection Reflection radiation radiation Condensation

Product

Heat conduction

Figure 6. Heat flow influencing in sun drying (Adapted from Muhlbauer, 2002)

3.2 Solar drying

To address the many issues confronting sun drying of fruits and vegetables, researchers have applied a solar energy collector or concentrator, calling it a “solar dryer”. The development of various types of solar dryers, their principles and operations can be found in many publications elsewhere (Bansal & Garg, 1987; Imre, 1995; Parker, 1991; Muhlbauer, 2002;

Wiess & Buchinger, n.d.; Jayaramayan & Das Gupta, 1992; Bolin & Salunkhe, 1982). In general, solar dryers are classified according to their utilization of heat generated from solar radiation to dry the product. These classifications include direct, indirect, hybrid, and mixed systems (Lawland, 1981).

In a direct solar drying, the product to be dried directly absorbs radiant energy from the sun through a transparent cover. The product and air temperature inside the enclosures increases as the heat is generated through absorption from solar radiation. This heat is used to evaporate the moisture from the product to the circulating air. On the other hand, in indirect solar dryers, the solar radiation is heated in a solar collector and then blown to the drying chamber to dry the food material. The efficiency of the solar collector is dependent on the design and operating condition of the dryer such as heater design, airflow rate, heat transfer coefficient of absorber- air-insulation inter-phase, among others (Jayaramayan & Das Gupta, 1992). In a mixed or hybrid system, a supplemental heating source is introduced to the product to be dried using another source of energy like fuel or electricity, either direct or indirect mode of heating.

An example of an indirect solar dryer is the Multi-commodity Solar Tunnel Dryer

(MSCTD) being promoted in the Philippines (Fig. 7) (PhilMech, 2012; FAO, 2012). Adapted from the Institute of Agricultural Engineering in Tropics and Subtropics, University of

Hohenheim, Germany, the MCSD is a type of dryer that follows the typical east-west orientation

and uses solar energy. The design allows for natural movement of air by convection, creating a critical velocity that forces air across the tray of wet products to be dried. The whole system is placed horizontally on a raised platform. The drier consists of a small fan to provide the required air flow over the products to be dried; a solar collector using an absorber painted in black; and a tunnel drying unit. Fresh air is forced through the solar collector and the heated air is transferred from the absorber as it passes over the products and absorbs moisture from the products.

Additional heat that passes through the transparent cover helps to increase the temperature of the dryer and hence facilitate the drying process. The system requires only 20 30 W of electric power, which is used to operate the fans. To prevent microbial contamination and improve the quality of the dried product, the product to be dried (normally laid in a single layer) is enclosed with the UV-stabilized plastic sheets throughout the drying period. Table 1 shows some results of the performance of the dryer for selected fruits, vegetables and fish.

Axial fan Fresh air

Solar collector

Hot air

Drying trays

Exhaust

UV-stabilized plastic enclosure

Figure 7. Schematic diagram of multi-commodity solar tunnel dryer (Adapted from Miranda, et al., 2003)

Table 2. Performance of MCSTD dryer for selected fruits, vegetables and fish.

Drying duration, days Product Load, kg MCSTD Sun-drying Mango 160 2.5 3-4 Banana 100 1 1-2 Cassava 200 1.5 2 Tomato 400 3-4 7 Mushroom 60 1 2 Shrimp 60 0.5 1 Squid 40 0.5 1-1.5 Tilapia 60 1-1.5 3-4 Anchovy 50 1 1-1.5 Source: PhilMech, http://www.philmech.gov.ph/

The application of the solar dryer was found to be economical for some fruits and vegetables. But, since some solar energy systems are capital-intensive, the cost of the equipment should be lowered in order to be successful in a full scale of operation.

3.3 Hot air drying

Drying of wet material using a hot air drying method is accomplished mainly by convective heat transfer process facilitated through direct contact of the heated air into the wet material to be dried. Mass transfer of water from the core to the surface of the product and removal of water vapor from the surface are the two main important mechanisms that facilitate the drying process. To obtain dried product at reasonable drying cost, the desired type and quality of dried product, raw material, economy of scale and drying efficiency must be considered.

Hot air drying designs are generally adopted in drying fruits and vegetables. These dryers include cabinet or tray dryer, tunnel dryers, pneumatic conveyor dryers, and belt conveyor dryers, fluidized, spray and drum dryers, among others.

3.3.1. Cabinet or tray dryers

Generally, cabinet dryers are designed for small-scale drying operation (100-2000 kg/day of dried product), normally used for laboratory or pilot-scale drying experiments (Tang & Yang,

2004). The dryer configuration consists of perforated trays located in an insulated chamber. It is equipped with blower and ducts to allow for a controlled circulation of heated air around or across the drying chamber. The heat source can be from a gas burner, electric heater, biomass furnace or by means of a heat exchanger. The energy is supplied from the air stream to provide the heat required to evaporate the water from the product surface, and carried away from the drying chamber (Tang & Yang, 2003; Barbosa-Cánovas & Vega-Mercado, 1996; Simate &

Ahrne, 2006). An example of a cabinet dryer design is illustrated in Fig. 8.

Exhaust air Recycling air Electric motor

Fan

Fresh air

Burner

LPG

Figure 8. Schematic diagram of a cabinet 'Esteban' dryer (Adapted from Miranda, et al., 2003)

Cabinet or tray dryers have been applied to produce a variety of products. Rao & Roy

(1980) produced mango leather using a tray dryer. Mango puree smeared with glycerine was evenly spread to a metal tray at a thickness of 1 mm. Drying of the pu ee t btain the ang leathe with istu e c ntent 1 20 %wb was accomplished in less than 20 h at a drying temperature range of 50-80 °C. Jayaraman (1989) used a cabinet dryer to produce fruit bars with a soft gel-like texture. The fruit pulp was mixed with pectin and sugar and dried by applying a three-stage drying protocol first for 1 h at 80 °C, followed by drying for 2 3 h at 70 °C, and finally for 5 6 h at 65 °C.

3.3.2. Tunnel dryers

Typically, the dryer dimension maybe up to 24 m long with a 2 x 2 m rectangular or square section. It consists of a cabinet equipped with mobile truck and trays with rails that move along a tunnel (Fig. 9). Drying of material is done in continuous mode wherein trays of wet

Fresh air Exhaust air

Circulated air

1 2 3 4 5

1. Blower 2. Heater 3. Exit/entrance for trucks 4. Trucks 5. Entrance/exit for trucks

Figure 9. Schematic diagram of a concurrent-flow tunnel dryer (From Van Arsdel, 1973)

product are stacked in trolleys and periodically introduced into one end of the tunnel, then the dried materials are discharged at the other end at predetermined time intervals. The movement of the product to be dried can be in the opposite direction of the heated air (counter-current flow) or in the same direction (co-current flow). The drying capacity of the dryer depends on the size of the tunnel, number of trays, drying temperatures and airflow rates. Similar to cabinet dryers, tunnel dryers are operated using low temperature, and hence the process is slow (Tang & Yang,

2003; Jayaraman & Das Gupta, 2006). Tunnel dryers have been applied to dry various fruits and vegetables such as carrots, potatoes, onions and others (Table 3).

Table 3. General technical data for vegetable dehydration using tunnel dryer. Commodity Drying condition Finished product Initial Final Water Load, kg/m temperature, temperature, Time, h Yield, % content °C °C Potatoes 8 85 75 4-6 8-10 12-16 Carrots 7 85 65 3-5 4-6 7 Cabbage 6 80 65 3-4 4-7 4-6 Onions 7 70 60 3-5 4-6 8-11 Green peas 5 75 60 3-4 4-6 9-14 Leafy 4-5 65 55 3-4 6-8 5-7 vegetables Herbs 3.4 60 55 3-4 5-7 5-7

3.3.3. Belt conveyor dryers

A conveyor belt dryer consists of several sections in series. Each section contains an independent control system that allows for easy adjustment to a desired belt speed, airflow rate and direction of airstream. During the operation, wet products are conveyed using either a high tensile strength plastic belt or perforated metal. The heated air is introduced directly to the conveyor belt for maximum product exposure. To achieve a higher drying efficiency, the thickness of the material to be dried varied, ranging 6 cm for the first or second secti ns

( ast ving belts) and 1 20 cm deep when transferred to succeeding sections (slower moving belts). For uniform drying, it is recommended that the semi-dried product should be properly mixed while transferring them from one section to the other. Depending on the materials to be dried, this type of dryer is capable of drying a product with water content of 80 90 %wb down to 5 %wb in single pass at a feed rate of 2 7 metric tons per hour. Depending on the water content of the initial and finished product, the yield capacity of the conveyor belt dryer ranges from 300 to 1,800 kg/h. Although relatively expensive, the dryer can produce consistent dried product with higher throughput capacity compared to cabinet or tunnel dryer (Tang & Yang,

2003).

Some belt conveyor designs consist of two or more conveyors in a series, while other designs can be multiple conveyors, one above the other (Fig. 10). The wet product is introduced

Wet product Air outlet

Heater Heater

Dry product Air Air

Figure 10. Schematic diagram of multi-pass conveyor belt dryer (Adapted from Grabowski et al., 2003).

on the top conveyor and progresses downward from one conveyor to the next conveyor until the last conveyor where the dried product is collected. This type of dryer is limited to cut or granulated foods.

3.3.4. Pneumatic conveyor dryers

Pneumatic conveyor dryers are effectively used for powders or granulated products, specifically for potato granules. The drying process begins by feeding the material into a stream of heated air and then conveyed through a duct at a predetermined conveyor speed to attain the desired moisture of the material. The used air from the system is exhausted through a cyclone or filter, while the dried product is separated and conveyed to a discharge outlet. The semi-dried material that comes out from this type of dryer can be re-dried in a cabinet dryer at a faster rate.

Jayaraman et al. (2006) reported that pieces of vegetables were dried down to 50 % wb in 8 min using a temperature of 160 180 °C. The semi-dried product was found to have increased porosity, which was facilitated during finish drying in a conventional cabinet type dryer.

Examples of vegetables dried using the system are presented in Table 4.

Table 4. Process conditions and rehydration characteristics of selected vegetables dried using a pneumatic conveyor dryer. Optimum HTST Rehydration Rehydration Moisture content drying time coefficient Material Cooked HTST Final Temp. Time Raw /blanch Treated dried ( °C) ( min) ed Potatoes 82.2 83.3 59.3 4.1 170 8 5 0.94 Green peas 71.1 72.5 38.3 3.4 160 8 5 1.06 Carrots 89.3 91.0 52.9 4.2 170 8 5 0.5 Yams 76.6 78.3 50.2 3.9 180 8 6 1.01 Sweet 73.6 78.6 53.8 5.3 170 8 2 1.06 potatoes Plantains, 80.8 83.3 58.8 4.6 170 8 4 0.97 raw Source: Jayaraman et al., 1982; Jayaraman & Das Gupta, 2006.

3.3.5. Fluidized bed dryers

The fluidized bed dryer was originally designed for drying of potato granules (Jayaraman et al., 2006). This process requires a small drying space, but it is an effective technique to maximize the surface area of the material to be dried, which facilitates drying of the product.

During the drying operation, the heated air is introduced through a bed of food particles at a high velocity which is enough to overcome the gravitational force of the material and thus allows it to be suspended in the air. As illustrated in Fig. 11, the movement of the product within the system is facilitated by the reduction of mass as moisture is removed.

Moist air outlet

Insulated hood Moist granules in

Fluidized bed

Porous plate

Plenum chamber Dry granules out Hot air in

Figure 11. Schematic diagram of a fluidized bed dryer (From Singh and Heldman, 2006)

Fluidization of food particles results in equal drying of the product and enhances the drying rates because of the intimate contact between the carrier gas (hot air) and solids (Tang &

Yang, 2003, Singh & Heldman, 2009). The air velocity can be controlled according to the

density, particle size, shape and viscosity of the material. Holdsworth (1971) reported that this technology has been used for drying of diced vegetables, whole beans and peas, potato granules, onion flakes and fruits juice powders. One drawback of this technology is its limited application to larger particle size > 10 mm to achieve an effective fluidization (Jayaraman et al., 2006).

3.3.8. Spray dryers

Spray drying is a process wherein the feed such as paste, solution or suspension is sprayed in droplets in a stream of hot air using an atomizer to accomplish evaporation and produce a free flowing dry powder (Barbosa et al., 2005). A typical spray drying system consists of a large cylindrical drying chamber, feed atomizer, hot air supply and a cyclone to collect the dried particle product, and fan (Masters, 1985 & 1991) (Fig. 12a). The geometry of drying chambers depends mainly on the type of atomizer, which dispenses the liquid, slurry or paste- like material into fine droplets. A rotary wheel atomizer is used for a short and wide drying chamber (conical type), while a nozzle type atomizer is used for a long and narrow (tall type) chamber (Ramaswamy & Marcotte, 2006). Conversion of liquid to powder particle during spray drying takes place in two basic steps. The first step is referred to as a drying constant rate period.

Basically, the drying process is controlled at the feed droplets. The heat transfer occurs from the droplet surface through the gas phase, while mass transfer of water vapor occurs from the droplet surface to the gas phase. The second step is the formation of a solid particle in which evaporation is controlled by the diffusion of moisture from the internal core of the particle toward the particle surface. A very fast drying rate, normally measured in seconds is expected due to the high velocity of hot air coupled with direct contact between the product and air. The rapid evaporation maintains a relatively low product temperature, and is hence favorable for drying heat sensitive materials.

Liquid feed Atomizer Air Hot air Droplets Exhaust

Blower

Air Scrubber

Cyclone Drying chamber

Fan

Dried product a b

Liquid feed Atomizer Air Hot air Droplets Exhaust

Blower

Air Scrubber

Cyclone Drying chamber

Fan Fan

Hot air Dried Cold air product Fan

Figure 12. Schematic of a typical spray dryer(a) and spray drying system with external vibrating fluidized bed. (Adapted from Barbosa-Cánovas & Vega-Mercado, 1996 and GEA-Niro, Inc.)

Some innovative spray dryer designs have been reported. A two-stage spray drying system was proposed by adding a fluidized bed in series with the main spray dryer (Moller &

Fredsted, 2009; Barbosa-Cánovas & Vega-Mercado, 1996) (Fig. 12b). The system allows for a very rapid evaporation of surface moisture in the first stage, while supporting accurate control and management of the internal moisture of the dried particles in the second stage. This type of design allows improvement of the physical properties of a powder product with low moisture.

Other innovative designs such as superheated steam spray drying, two-stage horizontal spray drying, low humidity spray drying and spray-freeze drying were also investigated (Amelo &

Gauvin, 1986; Strumillo et al., 1983). However, very few such designs are readily available commercially due to certain limitations of the new systems.

Even though spray drying is a beneficial drying technique because of its ability to produce relatively uniform, spherical particles (Nindo & Tang, 2007), it also demonstrates some drawbacks. Degradation of some of the beneficial physico-chemical properties and sensory qualities, especially vitamin C, beta carotene, flavors and aroma of the products, have also been observed (Dziezak, 1988). High volatile losses also occur during spray drying due to high atomization and air temperatures, which ranges from 150-300 °C (Nindo & Tang, 2006). Spray drying of sugar-rich materials is difficult due to stickiness. The product may remain as syrup and stick in the chamber wall during drying. Also, unwanted agglomeration may happen inside the chamber as well as in the conveying system resulting in lower cyclone recovery (Bhandari et al., 1997a,b ). The most common approach to overcome stickiness in such materials is to add a drying aid with high molecular weight, high glass transition temperature such as maltodextrin, glycerol monostearate, tricalcium phospate and some proteins. Improvement in cyclone recovery was obtained with the addition maltodextrin during spray drying of mango (Jaya & Das, 2004)

and strawberry (Abonyi et al., 2002). The application of certain protein-X and low molecular surfactants also improved product yield during spray drying of model food containing sucrose

(Adhikari et al., 2009a,b; Jayasundera et al., 2011a,b).

3.3.9. Drum Dryers

A drum dryer is made up of either a single-drum, double-drum, twin- drum or vacuum- drum dryer (Fig. 13) (Karel, 1975; Barbosa-Cánovas & Vega-Mercado, 1996). The single-drum dryer (Fig. 13a) consists of one cylinder, while the double-drum dryer (Fig. 13b) design consists of two cylinders rotating in opposite directions. The double-drum design allows adjustment of the thickness of the feed by changing the gap between the two drums. The twin-drum dryer (Fig.

13c) is also comprised of two drums similar to a double-drum, but they are positioned apart and rotating away from each other. A vacuum-drum dryer (Fig. 13d) has similar configuration to the other types except that the entire system is enclosed in a vacuum-tight chamber. This system is normally used when handling heat-sensitive materials (Barbosa-Cánovas & Vega-Mercado,

1996).

During drum drying operation, the product to be dried in slurry, paste or puree form is spread in a thin layer onto the metal surface of the internally heated drums by means of a feeder mechanism. The dried product in the form of flakes is removed after rotating approximately three fourths of the revolution of the drums by a doctor blade (Saravacos, 2006). To ensure good performance of the dryer, the uniform thickness of the material applied to the outer surface of the drum must be considered. The recommended speed of rotation and temperature inside the drum must be followed. When drying heat sensitive products, the drum must be enclosed in a vacuum tight chamber.

Feed

(a) Heated drum

Dried product

Feed

(b)

Heated Heated drum drum Dried Dried product product

Feed

Dried Dried product product

Vacuum chamber

(d)

Heated Feed drum

Dried product

Figure 13. Schematic diagram of a single drum dryer (a); a double drum dryer (b); a twin-drum dryer (c); and a vacuum drum dryer (d) (Adapted from Karel, 1975; Barbosa-Cánovas & Vega- Mercado, 1996 ).

Drum dryers are used for manufacturing different types of products, particularly potato flakes. However, their application to sugar-rich and low fiber content materials is limited because of the high temperature required for drying. Kitson & MacGregor (1982) successfully produced drum-dried apple, banana, apricot, guava, papaya, and cranberry without additives

(high fiber content), while adding a 1% methoxyl pectin to produce raspberry, blue berry and strawberry (low fiber content). Mango flakes with 3% water content was produced using a double-drum dryer at 4.2 kg/cm2 steam pressure and 0.254 drum space for 25 s (Brekke et al.,

1975). Travaglini et al. (1994) found a conventional drum dryer suitable for producing mango flakes at 60 psia pressure, gap spacing of 0.06 inches between two drums and residence time of

20 s. They also improved the drying process by adding 1% glycerol monostearate and 4% corn starch to the mango puree before drying. Abonyi et al. (2002) also reported similar results for drum-dried carrot powder with severe nutrition loss. Drum drying is economical to use in producing food powder. However, its major drawback is that the severity of heat pre-processing can produce undesirable cooked flavors and aromas in the dried product and cause severe quality degradation in the product (Nindo & Tang, 2006).

3.4. Freeze drying

Freeze-drying or lyophilization is a drying process in which the food is first frozen and thereafter sublimed from the solid state directly into the vapor phase without passing through the liquid state, generally under reduced pressure (Ramaswamy & Marcotte, 2006; Athanasios &

Bruttini, 2006; Barbosa-Cánovas & Vega-Mercado, 1996). Fig. 14 represents the phase diagram of water with conditions of atmospheric air pressure, vacuum and freeze drying. The point of intersection of the curves is called the "triple point" at which the three phases exist in equilibrium.

Solids Liquid

Meltin Vaporizati g on

1 atm (760 Pa)

Pressure 4.58 Torr 610.62 Pa Triple (760 Pa) Sublimati point on Gas

0 10 0.0098 0

Temperature, °C

Figure 14. Phase diagram for water showing sublimation of ice (Adapted from Karel, 1975; Barbosa-Cánovas 1996).

The freeze drying process involves three stages, as illustrated in Fig. 15: 1) freezing stage, b) primary drying stage, and c) secondary drying stage. During the freezing stage, the food is frozen to obtain small crystals and an a ph us state. The p duct’s final temperature must be lower than the collapse temperature in order to maintain its structure. Depending on the types of material to be dried and the freezing rates, the freezing behavior may differ in that either the liquid phase immediately solidifies or just becomes a highly viscous liquid. It is reported that the freezing rate greatly influences the heat and mass transfer rates in the dried layer as well as the quality of the dried product. Fast freezing may produce small and discontinuous crystals; hence, mass transfer of vapor in the dried layer is restricted. On the other hand, slow freezing allows for the formation of larger ice crystals and it favors faster drying due to the high mass transfer rate of the water vapor in the dried layer.

Air molecules and other non-condensibles evacuated

Product vapor collected as ice

Low temperature condenser

Water vapor

Dry product Ice interface Product ice

Heat input Vacuum chamber

Figure 15. Schematic diagram of a freeze dryer(Adapted from Singh & Heldman, 2006)

In the second stage (primary drying), the objective is to remove the unbound or easily removed ice from the product. This water is now in the form of free ice, which is removed by converting it directly from a solid to a vapor, in a process called "sublimation." To ensure that sublimation occurs, the pressure of the water vapor remains below its triple point or at pressure

≤ 4. 8 T (Figu e 14). Othe wise, elting ay ccu esulting in sh inkage and st uctu al collapse in the product (Fellows, 1992).

During the last stage (secondary drying), the sorbed water or the water that was strongly bound to the solids in the product is converted to vapor. Evaporation at this stage can be very low since the remaining bound water has a lower pressure than free water at the same temperature, which makes it difficult to remove. Freeze-drying is complete when all of the frozen layer has been removed or there is no more sublimed interface.

Freeze drying has been widely used for dehydrating food products such as fruits, vegetables, coffee, and fishery products. It has also been applied to dehydrate blood plasma,

living cells, bacteria, yeast, viruses, pharmaceuticals (e.g. antibiotics), historical documents and other materials (Liapis & Bruttini, 2006). Many authors reported that the freeze-drying method can maintain the original quality attributes of raw food materials after drying. Woodroof & Luh

(1975) and Bruin & Jongen (2003) reported some benefits of using freeze drying, including high retention of physical and chemical properties of the products, such as color, volatile flavor, aroma, minimal damage to product structure, shape and texture, and high porosity resulting in fast rehydration. Caparino (1999) reported that the sensory qualities of freeze-dried mango powder such as color, aroma, taste and sweetness were retained after freeze drying at different pre-freezing temperatures. The total soluble solids present in the original mango puree were not affected during the freeze drying process. Chang et al. (2006) reported that 90% of vitamin C was retained during freeze drying of tomato cubes, two-folds higher compared to hot-air drying.

Asami et al. (2003) found undetectable changes in vita in in ee e-d ied st awbe ies,

a i nbe ies and c n. The c l , lav and taste ang , pineapple and papaya we e p ese ved with high p sity anging 0.84 0.93, as reported by Marquez et al. (2006).

Abonyi et al. (2002) found a high percentage of etenti n in β-carotene (93.6%) and ascorbic acid (90%) in freeze-d ied st awbe y. Des b y et al. (1 7) als ep ted that 2% β-carotene was preserved in carrot during freeze drying.

However, the freeze drying process is known to be very expensive compared to other typical drying methods (normally 4-8 times higher than air drying) because of its high energy consumption, high operating and maintenance costs and low output (Xu et al., 2005; Fellows,

1990). Ratti (2001) compared the cost of freeze drying of high and low-value raw materials and found that energy consumption becomes insignificant when high-valued raw commodities are processed. Because of these limitations, the freeze drying method is practically applicable only

to high value-added commodities under an industrial scale of operation (Farenzena & Ratti,

2001; Simate & Ahrne, 2006).

3.5. Novel and Innovative Drying Technologies

3.5.1. Refractance Window® drying

The Refractance Window® (RW™) d ye was designed ainly t c nve t uit pu ee and other slurries into powder, flakes, or concentrates (Nindo & Tang, 2007). It is a novel concept which utilizes thermal energy to transfer hot water to a thinly prepared pureed material placed on a transparent plastic conveyor belt that moves on a shallow trough while in contact with the hot water (Bolland, 2000; Nindo et al., 2006). As illustrated in Fig.16, the hot water (about 95 97 oC) circulates within the system and transfers thermal energy to the product samples to be dried.

The product samples are laid out on a conveyor belt and move continuously at a desired speed while in direct contact with the hot water. To achieve uniform spreading of the feed on the plastic conveyor, the juice extract or purees must have the correct consistency. Nindo et al.

(2006) explained that during the drying of liquid foods using the RW dryer, the heat transfer modes of convection, conduction and radiation are utilized (Nindo & Mwithiga, 2010). The latent heat from steam condensing water is transferred by convection to the bottom surface of the plastic sheet. The use of a thin plastic sheet causes its temperature to easily equilibrate with the flowing water underneath. As the system operates, the thermal energy from water is transmitted via the plastic conveyor by conduction and radiation. The amount of energy conducted and radiated varies depending on the resistance and thickness of the plastic sheets being used. The inventor of the RW dryer claim that the use of a thin and transparent plastic (DuPont Plastics,

Wilmington DE) allows efficient transfer of heat radiation into the product compared to a thicker

one. They also reported that by using a high moisture pureed product during drying, more thermal energy is absorbed by the product, which facilitates rapid evaporation.

Exhaust

Wet product application Evaporation Scraper

Recirculating hot water (95 97 Plastic °C) conveyor belt Dried product

Cooling section

Figure 16. Schematic of a pilot scale Refractance Window® drying system (adapted from Ramaswamy & Marcotte, 2006)

This technology gains several advantages and benefits when applied to fruits and vegetables. In an experiment conducted by Nindo & Tang (2007) using carrots, strawberries and squash, it was concluded that the RW drying system maintains the nutritional (vitamins, antioxidants) and sensory (color, aroma) attributes of the dried product. Nindo et al. (2003) also reported that the bright green color of pureed asparagus remains unchanged when dried in the

RW dryer. Abonyi et al. (2002) confirm that dried carrot and strawberry produce from the RW

dryer was found to have no significant difference when compared to the freeze-dried product.

Moreover, it was found that the ascorbic acid present in strawberries was retained after drying using the RW dryer, and it was comparable to the freeze-dried product Nindo et al. (2003). Other benefits in using the RW dryer are that it demonstrates higher thermal efficiency or low energy consumption compared with other conventional dryers. Nindo et al. (2003) also reported that in pumpkin puree, coliform, Escherichia coli and Listeria innocua were significantly reduced to a minimum detection level, and hence food safety is assured.

3.5.2. Microwave drying

Microwave drying has gained popularity in the food industry as an alternative to improve the quality of dried food products. There are two mechanisms that explain heat generation during microwave drying, including dipole rotation and ionic polarization. Food materials contain a random orientation of polar molecules such as water. But, when an electrical field is introduced to the product, the heat generated by the absorption of electromagnetic radiation by dipolar molecules is utilized for drying the product. Unlike in conventional heating where heat is transferred to the product surface by conduction, convection, and/or radiation, and by thermal conduction into the interior of the product, the heat induced by microwave is directly generated inside the material, resulting in higher heat flux and faster temperature rise due to conversion of microwave radiation into kinetic energy (Sutar & Prasad, 2007b & 2008; Yongsawatdigul &

Gunasekaran, 1996a) (Fig. 17).

A number of microwave drying systems have been reported such as microwave spouted bed, microwave-convective and microwave-vacuum drying systems. A combined microwave and spouted bed (MWSB) drying system was developed by Feng, Tang & Cavalieri (1999) and tested for diced apples (Fig. 18). The system provides a mechanism wherein diced apples move

through a macroscopic circulation in upward direction "spouls" and a downward annulus (Feng

& Tang, 1999). Their process allows for uniform heating and temperature distribution on the particles of the product. The MWSB obtained significant improvement in the quality of the finished product, with higher reconstitution capacity and lower bulk density compared to traditional drying methods (Feng & Tang, 1998). Nindo et al. (2003) (2003a) also confirmed that using a combined microwave and spouted bed drying has produced dried asparagus with good rehydration and color characteristics when compared to freeze drying and tray drying.

Modeling in microwave vacuum drying kinetics and moisture diffusivity of carrot slices was carried out by Sutar & Prasad (2007a) using a laboratory scale set up, as shown in Fig. 19.

Carrots were positioned in a single layer of sample slices in a vacuum chamber and placed inside the microwave cavity. Uniform exposure of carrots to microwave radiation was achieved by rotating the samples at 360°. The study found that the drying rates increased with the increase in microwave power density at all pressure levels. Microwaves in combination with vacuum can produce heating rates that are 20–30 times higher than those achieved in freeze drying (Farid,

2002). Lin et al. (1998) reported that vacuum microwave dried carrot slices had a softer texture, high retention in color of the product, and were rated by a sensory panel to be equal to or better than freeze drying in terms of color, texture. It was also found that vacuum microwave dried - cranberries had redder color and softer texture as compared to the hot air dried counterpart

(Yongsawatdigul & Gunasekara, 1996b).

Electrical Energy transfer by radiation Microwave Absorption Energy source energy of energy by water molecules

Water evaporation from surface Vacuum pump Diffusion of / condenser high energy water molecules Water vapor diffusion through surface

Energy transfer by convection Energy transfer

Heated by conduction Energy source air Diffusion of liquid to surface Evaporation from surface

Figure 17. Microwave-vacuum drying vs. air drying

Microwave power controller

Spouted bed Computer Magnetron

T Microwave cavity

F T Sample Temperature controller

Sink Water load D V Water pump Heater

Electric balance Blower

D - Dew point temperature; F - Flowrate: T - Temperature; V - Velocity

Figure 18. Microwave and spouted bed combined (MWSB) drying system (Adapted from Feng & Tang, 1999).

16 1 2 3 4 15 19

5 20 6 17 7

18 8 9

14 13 12 11 10

1. Carrot slices. 2. Vacuum chamber. 3. Fiber optic cables with sensors. 4. Signal conditioner. 5. Microwave timer. 6. Microwave power on=off switch. 7. Turntable controller. 8. Electric power switch. 9. Manual

microwave power control. 10. Microwave power level indicator. 11. Remote power control. 12. Heat extractor fan switch. 13. Moisture condenser. 14. Vacuum pump. 15. Pressure gauge. 16. Pressure control valve in moisture suction line. 17. Heat extractor fan. 18. Turntable. 19. RS-232 cable. 20. Personal computer (From Sutar & Prasad, 2007).

Figure 19. Experimental setup of microwave vacuum dryer.

3.5.3. Ultrasonic drying

Ultrasonic drying is one of the novel drying techniques that has been explored for drying of food products either from a liquid or solid raw materials. Fig. 20 shows a schematic diagram of an acoustic dryer (Carcel et al., 2007). It consists of a vibrating drying chamber driven by a high-intensity ultrasonic transducer. The ultrasonic transducer is connected to an electronic generator through an impedance matching unit and a digital power meter.

For drying of solid material, high-intensity ultrasonic waves are produced and travel rapidly in a series of alternating contractions and expansions of the material. This alternating movement creates microscopic channels within the product, allowing for faster diffusion of moisture from the core to the surface of the product (Gallego-Juarez, 1999). The generated acoustic waves also influence cavitation of water molecules inside the solid matrix, which facilitate diffusion and removal of bound water (Muralidhara et al., 1985). Applying the ultrasonic drying system can significantly increase heat and mass transfer because of the reduction of diffusion boundary as influenced by the micro-streaming and oscillating-velocity effect generated from acoustic energy (Mason & Lorimer, 2002). According to Mason et al.

(1996), the heat transfer rate can be increased to more than two-fold when using this type of drying technique. The ultrasonic drying technique can be applied for air-borne ultrasonic drying of liquid (Fig. 21a) or by direct contact to a vibrating solid material (Fig. 21b).

Ultrasonic drying is suitable for highly sensitive material because it is capable of producing dried products with minimal effect on the structure of the dried product. Ratti (2009) reported that this drying technique was successfully applied for drying of carrots (de la Fuente-

Blanco et al., 2006) and persimmons (Carcel et al., 2007) and was found effective to enhance osmotic drying of banana (Fernandez & Rodriguez, 2007) and papaya (Fernandez et al., 2007). It

was also tested as a pre-treatment before drying of mushrooms and cauliflower (Jambrak et al.,

2007). A combination of solar and ultrasound energy was found to be successful for the drying of different food products (Ramaswamy et al., 2006). However, the application of this drying technique was limited due to some practical difficulties for application in an industrial operation

(Gallego-Juarez, 1999).

Vibrating drying chamber

Ultrasonic Impedance Digital power transducer matching meter unit

Ultrasonic power generator

Air flow

Figure 20. Schematic illustration of ultrasonic-assisted hot air dryer (Adapted from Carcel et al., 2007).

Electronic generator

Ultrasonic transducer Hot air generator Vibrating plate

Samples Support plate

Microphone

Anemometer Thermometer

a b

Laser vibrometer Electronic generator

Ultrasonic transducer

Vibrating plate Samples Support plate (with drain channel) Suction system Pneumatic system

Static pressure

Figure 21. Experimental dehydration setup for air-borne ultrasonic drying of liquid (a) or direct contact to a vibrating solid material (b) (Adopted from Gallego-Juarez & Riera, 2011)

5. CONCLUDING REMARKS

This article covers some relevant fundamental aspects in food dehydration and a brief review of some of the developments in food dehydration of fruits and vegetables. Considered to be an old preservation technique, advances in dehydration processes and development of innovative and novel drying technologies have evolved over several decades. This has resulted in the availability of various types of dried products in different forms, e.g. ready-to-eat, ingredients, instantly reconstituted food powders and many others. This trend will continue to evolve because of the increasing demand for high quality dehydrated fruits and vegetables worldwide. As the cost of energy is becoming prohibitive, researchers in this area should consider offering solutions for reducing the energy consumption of the different drying systems already in existence to make them more attractive and profitable to the end users.

REFERENCES

Abiad, M. G., Carvajal, M. T., & Campanella, O. H. (2009). A review on methods and theories

to describe the glass transition phenomenon: applications in food and pharmaceutical

products. Food Engineering Review , 1 (2), 105-132.

Abonyi, B. I., Feng, B. I., Edwards, C. G., & Tang, J. (2002). Quality retention in strawberry and

carrot purees dried with Refractance Window system. Journal of Food Science , 67,

1051–1056.

Adhikari, B., Howes, T., Bhandari, B., & Langrish, T. G. (2009a). Effect of addition of proteins

on the production of amorphous sucrose powder through spray drying. Journal of Food

Engineering , 94 (2), 144-153.

Adhikari, B., Howes, T., Wood, B. J., & and Bhandari, B. R. (2009b). The effect of low

molecular weight surfactants and proteins on surface stickiness of sucrose during powder

formation through spray drying. Journal of Food Engineering , 94, 135-143.

Aguilera, J. M. (2003). Moisture sorption isotherms. In D. R. Heldman, Encyclopedia of

Agricultural, Food, and Biological Engineering (pp. 676-679). New York: Marcel

Dekker, Inc.

Amelot, M., & Gauvin, W. (1986). Spray drying with plasma-heated water vapor. In A. S.

Mujumdar, Drying (Vol. 1, pp. 285–290.). New York: Hemisphere, McGraw-Hill.

Asami, D. K., Hong, Y., Barret, D. M., & Mitchelle, A. E. (2003). Comparison of the total

phenolic and ascorbic acid content of freeze-dried and air-dried marionberry, strawberry,

and corn grown using conventional, organic, and sustainable agricultural practices.

Journal of Agricultural and Food Chemistry , 51, 1237-1241.

Bansal, N. K., & Garg, H. P. (1987). Solar crop drying. In A. S. Mujumdar, Advances in drying

(Vol. 4, pp. 279-299). Washington D.C.: Hemisphere.

Barbosa-Cánovas, G. V., & Vega-Mercado, H. (1996). Dehydration of Foods. New York:

Chapman and Hall.

Bhandari, B. R., Datta, N., & Howes, T. (1997a). Problems associated with spray drying of

sugar-rich foods. Drying Technology , 15 (2), 671-684.

Bhandari, B. R., Datta, N., Crooks, R., Howes, T., & Rigby, S. (1997b). A semi-emperical

approach to optimize the quantity of drying aids required to spray dry sugar-rich foods.

Drying Technology , 15 (10), 2509-2525.

Bidstrup, S. A., & Day, D. R. (1994). Assignment of the glass trasnition temperature using

dielectric analysis. In R. J. Seyler, Assignment of the Glass Transition (pp. 108-119).

Philadelphia, PA: ASTM.

Bolin, H. R., & Salunkhe, D. K. (1982). Food dehydration by solar energy. Crit. Rev. Food Sci.

Nutri. , 16, 327–354.

Bolland, K. (2000). Refractance Window Drying: A new low temperature, energy efficient

process. Cereal Foods World, 45 (7), 293-296.

Brekke, J. E., Cavaletto, C. G., Stafford, A. E., & Chan, H. T. (1975). Mango: Processed

Products. Washington, DC: USDA.

Brunauer, S., Emmett, P. H., & Teller, E. (1938). Adsorption of gases in multimolecular layers.

Journal of American Chemists' Society , 62, 309-319.

Caparino, O. A. (2000). Characteristics and quality of freeze-dried mango powder pre-frozen at

different temperatures. Philippine Agricultural Scientist Journal , 83 (4), 338-343.

a cel, J. A., Ga cıa-Perez, J. V., Riera, E., & Mulet, A. (2007). Influence of high-intensity

ultrasound on drying kinetics of persimmon. Drying Technology , 25, 185–193.

Chang, C., Lin, H., Chang, C., & Liu, Y. (2006). Comparison on the antioxidant properties of

fresh, freeze-dried and hot-air-dried tomatoes. Journal of Food Engineering , 77, 478-

485.

Chirife, J., & Iglesias, H. A. (1978). Equations for fitting water sorption isotherms of foods: Part

1 - A review. Journal of Food Technology , 13, 159-174.

De Graaf, R. A., Karman, A. P., & Jansen, L. (2003). Material properties and glass transition

temperatures of different thermoplastic starches after extrusion processing. Starch , 55,

80-86. de la Fuente-Blanco, S., Riera-Franco de Saravia, E., Acosta-Aparicio, V. M., Blanco-Blanco,

A., & Gallego-Juares, J. A. (2006). Food drying process by power ultrasound.

Ultrasonics , 44, e523-e527.

Desobry, S. A., Netto, F. M., & Labuza, T. P. (1997). Comparison of spray-drying, drum-drying

and freeze-d ying β-carotene encapsulation and preservation. Journal of Food Science

, 62 (6), 1158-1162.

Dziezak. (1988). Microencapsulation and encapsulated ingredients. Food Technology , 42, 136-

148.

Earnest, C. M. (1994). Assignment of the glass transition temperatures uisng thermomechanical

analysis. In R. J. Seyler, Assignment of the Glass Transition . Pliladelphia: ASTM.

Ekechukwu, O. V. (1999). Review of solar-energy drying systems II: An overview of solar

drying technoloy. Energy Conservation and Mangament , 40, 615-655.

Falade, K. O., & Aworth, O. C. (2004). Adsorption isotherms of osmo-oven dried african star

apple (Chrysophyllum albidum) and african mango (Irvingia gabonensis) slices. Eur

Food Res Technology , 218, 278-283.

Farhat, I. A. (2004). Measuring and modelling the glass trasnsition temperature. In R. Steele,

Understanding and measuring the shelf-life of food. New York: CRC Press.

Fellows, P. j. (1992). Food Processing Technology Principles and Practice. London: Ellis

Horwood. Ltd.

Feng, H., & Tang, J. (1998). Microwave finish drying of diced apples in a spouted bed. Journal

of Food Science , 63 (4), 679-683.

Feng, H., Tang, J., & Cavalieri, R. P. (1999). Combined microwave and spouted bed of diced

apples: Effect of drying conditions on drying kinetics and product temperature. Drying

Technology , 17 (10), 1981-1998.

Fernandes, F. A., & Rodrigues, S. (2007). Ultrasound as pre-treatment for drying of fruits:

Dehydration of banana. Journal of Food Engineering , 82, 261-267.

Fernandes, F. A., Oliveira, F. I., & Rodrigues, S. (2007). Use of ultrasound for dehydration of

papayas. Food Bioprocess Technology , DOI:100.1007/s11947-007-0019-9.

Forssel, P. M., Mikkila, J. M., & Parker, R. (1997). Phase transition behaviour of concentrated

barley starch-glycerol-water mixturs- a model for thermoplastic starch. Carbohydrate

Polymers , 34, 275-282.

Gallego-Jua e , J., R d ıgue , G., Galve , J., & Yang, T. (1 ). A new high-intensity ultrasonic

technology for food dehydration. Drying Technology , 17 (3), 597-608.

Grabowski, S., Marcotte, M., & Ramaswamy, H. S. (2003). Drying of fruits, vegetables and

spices. In A. Chakraverty, A. S. Mujumdar, G. S. Raghavan, & H. S. Ramaswamy,

Handbook of Postharvest Technology (pp. 653-696). New York: Marcel Dekker.

Greenspan, L. (1977). Humidity fixed points of binary saturated aqueous solutions. Journal of

Research and National Bureau of Standards (A): Physics and Chemistry 81A (1) , 89-96.

Holdsworth, S. D. (1971). Dehydration of food products - A review. Journal of Food

TEchnology , 6, 331-370.

Jambrak, A. R., Mason, T. J., Paniwnyk, L., & Vesna, L. (2007). Accelerated drying of button

mushrooms, brussels sprouts and cauliflower by applying ultasound and its rehydration

properties. Journal of Food Engineering , 81, 88-97.

Jaya, S., & Das, H. (2004). Effect of maltodextrin, glycerol monostearate and tricalcium

phosphate on vacuum dried mango powder properties. Journal of Food Engineering , 63,

125-134.

Jayaramayan, K. S., & Das Gupta, D. K. (1992). Dehydration of druits and vegetables - recent

developments in principles and techniques. Drying Technology , 10 (1), 1-50.

Jayaramayan, K. S., & Das Gupta, D. K. (2006). Drying of Fruits and Vegetables. In A. S.

MUJUMDAR, handbook of Industrial Drying (3rd ed.). New York: Taylor and Francis

Group.

Jayasundera, M., Adhikari, B., Adhikari, R., & Aldred, P. (2011a). The effect of protein types

and low molecular weight surfactants on spray drying of sugar-rich foods. Food

Hydrocolloids , 25 (3), 459-469.

Jayasundera, M., Adhikari, B., Adhikari, R., & Aldred, P. (2011b). The effects of proteins and

low molecular weight surfactants on spray drying of model sugar-rich foods: Powder

production and characterisation. Journal of Food Engineering , 104 (2), 259-271.

Karel, M. (1975). Freeze dehydration of foods. In M. Karel, O. R. Fennema, & D. B. Lund,

Principles of Food Science (pp. 359-395). New York: Marcel Dekker, Inc.

Kasapis, S. (2006). Definition and applications of the network glass transition temperature. Food.

Food Hydrocolloids , 20, 218–228.

Kasapis, S. (2004). Definition of a mechanical glass transition temperature for dehydrated foods.

J. Agric. Food Chem. , 52, 2262-2268.

Kitson, J. A., & MacGregor, D. R. (1982). Drying fruit purees on an improved pilot plant drum

dryer. Journal of Food Engineering , 17 (2), 285-288.

Kudra, T., & Mujumdar, A. S. (2002). Advanced Drying Technologies. New York: Marcel

Dekker, Inc.

Laaksonen, T. J., & Roos, Y. H. (2003). Water sorption and dielectric relaxations of wheat

dough (containing succrose, NaCl and their mixtures). Journal of Ceral Science , 37,

319-326.

Labuza, T. D. (1982). Shelf-life dating of food. Westport, CT: Food & Nutrition Press.

Labuza, T. K. (1985). Effect of temperature of the moisture sorption isotherm and water activity

of two dehydrate foods. Journal of Food Science , 50, 385-391.

Labuza, T. P. (1968). Sorption phenomena in foods. Journal of Food Technology , 22 (3), 15-24.

Labuza, T., & Altunakar, B. (2007). Water activity prediction and moisture sorption isotherms.

In G. V. Barbosa-Canovas., & A. J. Fontana, Water Activity in Foods: Fundamentals and

Applications (pp. 109-154). Ames, Iowa: Blackwell Publishing.

Law, C. L., & & Mujumdar, A. S. (2008). Dehydration of fruits and vegetables. In A. S.

Mujumdar, Guide to industrial drying (pp. 223–249). India: Three S Colors.

Lawland, T. A. (1981). Agricultural and other low temperature applications of solar energy. In S.

E. Handbook, Kreider, J. F., Kreith, F. (p. 18.1). New York: McGraw Hill.

Liapis, A. I., & Bruttini, B. (2006). Freeze drying. In A. S. Mujumdar, Handbook of Industrial

Drying (3 ed., pp. 257-281). Boca Raton, FL: Taylor & Francis Group, LLC.

Lin, T., Durance, T., & Scaman, C. (1998). Characterization of vacuum microwave, air and

freeze-dried carrot slices. Food Research International , 31 (2), 111–117.

Marques, L. G., Silveira, A. M., & Freire, J. T. (2006). Freeze-drying characteristics of tropical

fruits. Drying Technology , 4, 457–463.

Mason, T. J., Paniwnyk, L., & Lorimer, J. P. (1996). The uses of ultrasound in food technology.

Ultrasonic Sonochemistry , 3, s253-s260.

Mason, T., & Lorimer, J. (2002). The Uses of Power Ultrasound in Chemistry and Processing.

Weinheim: Wiley-VCH.

Masters, K. (1985). Spray drying handbook (4th ed.). London: Godwin.

Miranda, L. N., Asuncion, N. T., Flores, E. D., & Martinez, R. (2003). Adaptability testing of

existing dryers for nongrain commodities. Science City of Muñoz, Nueva Ecija,

Philippines: Bureau of Postharvest Research and Extension.

Moller, J. T., & Fredsted, S. (2009). A primer on spray drying. Retrieved May 5, 2012, from

GEA Niro: www.niro.com

Muhlbauer, B. M., Bauer, W., & Kohler, B. (2002). Solar crop drying in developing countries.

Berne.

Mujumdar, A. S. (2008). Classification and selection of industrial dryers. In A. S. Mujumdar,

Guide to industrial drying (pp. 23–36). India: Three S Colors.

Mujumdar, A. S. (2006). Handbook of Industrial Drying (3rd ed.). Florida.: Taylor & Francis

Group, CRC Press.

Muralidhara, H., Ensminger, D., & Putnam, A. (1985). Acoustic dewatering and drying (low and

high frequency): State of the art review. Drying Technology , 3 (4), 529-566.

Nindo, C. I., & Tang, J. (2007). Refractance Window dehydration technology: A novel contact

drying method. Drying Technology , 25, 37–48.

Nindo, C. I., Sun, T., Wang, S. W., Tang, J., & Powers, J. R. (2003a). Evaluation of drying

technologies for retention of physical and chemical quality of green asparagus

(Asparagus officinalis, L.). Food Sci. Technol. (LWT) , 36 (5), 507–516.

Nindo, C. I., Tang, J., Cakir, E., & Powers, J. R. (2006). Potential of Refractancec Window

technology for value added processing of fruits and vegetables in developing countries.

ASABE Annual International Meeting (pp. 1-14). Portland: ASABE.

Nindo, C.I. & Mwithiga, G. 2010. Infrared Drying. In Infrared Heating for Food and

Agricultural Processing. Z. Pan and G. Atungulu (Eds.). CRC Press, Boca Raton, FL.

Parker, B. F. (1991). Solar energy in agriculture. Amsterdam: Elsevier.

PhilMech. (2012). Multi-commodity solar tunnel dryer. Retrieved January 15, 2012, from

philmech.gov.ph

Potthast, K. (1978). Influence of water activity on enzymatic activity in biological systems. In J.

H. Crowe, & J. S. Clegg, Dry biological Systems (pp. 324-342). nEW yORK: Academic

Press.

Rahman, M. S. (1995). Food Properties Handbook. Boca Raton, FL: CRC Press.

Ramaswamy H., M. M. (2006). Food Processing: Principles and Applications. New York:

Taylor & Francis Group.

Ratti, C. (2009). Advances in Food Dehydration. New York: Taylor & Francis Group.

Ratti, C. (2001). Hot air and freeze-drying of high-value foods-a review. Journal of Food

Engineering , 49, 311–319.

Rizvi, S. S. (1995). Thermodynamic properties of foods in dehydration. In M. Rao, & S. S.

Rizvi, Engineering properties of foods (pp. 223-309). New York, USA: Marcel Dekker,

Inc.

Roos, Y. H. (2003a). Glass transition temperature. In D. R. Heldman, Encyclopedia of

Agricultural, Food, and Biological Engineering (pp. 440-443). New York: Marcel

Dekker, Inc.

Roos, Y. H. (1995). Phase Transitions in Foods. London: Academic Press.

Roos, Y. H. (2003b). Thermal analysis, state transition and food quality. Journal of Thermal

Analysis Calorimetry , 197-203.

Sablani, S. S., Syamaladevi, R. M., & Swanson, B. G. (2010). A review of methods, data and

applications of state diagram of food systems. Food Eng Rev , 2, 168-203.

Schmidt, S. J. (2004). Water and solids mobility in foods. Advances in Food and Nutrition

Research , 48, 1-101.

Seiler, D. A. (1976). The stability of intermediate moisture foods with respect to mould growth.

In R. Davies, G. G. Birch, & K. J. Parker, Intermediate Moisture Foods (pp. 166-181).

London: Applied Science Publishers Ltd.

Seyler, R. J. (1994). Assignment of the glass transition. Baltimore, MD: ASTM.

Simate, I. N., & Ahrne, L. M. (2006). Dehydration of tpical fuits. In Y. H. Hui, Handbook of

Food Science, Technology and Engineering (Vol. 3rd, pp. 104-1-104-7). FL: Taylor &

Francis.

Singh, R. P., & Heldman, D. R. (2009). Introduction to Food Engineering. Burlington, MA,

USA: Academic Press.

Slade, L., & Levine, L. (1991). A food polymer science approach to structure property

relationships in aqueous food systems: Non-equilibrium behaviour of small carbohydrate-

water system. In L. Levine, & L. Slade, Water Relationships in Food. New York:

Plennum Press.

Strumillo, C. (2006). Perspectives in developments in drying. Drying Technology , 24, 1059-

1068.

Strumillo, C., Markowski, A., & Kaminski, W. (1983). Modern developments in drying of paste-

like materials. In A. S. Mujumdar, Advances in Drying (Vol. 2, pp. 193-232). New York:

McGrawhill.

Sutar, P. P., & Prasad, S. (2007a). Modeling microwave vacuum drying kinetics and moisture

diffusivity of carrot slices. Drying Technology , 25 (10), 1695-1702.

Sutar, P. P., & Prasad, S. (2007b). Advances in microwave assisted heating and drying

technology for foods. Proceedings of Indian Convention of Food Scientist and

Technologists. Kharagpur: Indian Institute of Technology.

Sutar, P. P., & Prasad, S. (2008). Microwave drying technology-recent developments and R&D

needs in India. Proceedings of 42nd ISAE Annual Convention. ISAE.

Tang, J., & T., Y. (2003). Dehydrated vegetables: principles and systems. In D. R. Heldman,

Encyclopedia of Agricultural, Food, and Biological Engineering (pp. 335-373). New

York: Marcel Dekker, Inc.

Toledo, R. T. (2007). Fundamentals of Food Process Engineering (3 ed.). Athens, GA: Springer

Science+Business Media, LLC.

Travaglini, D. A., Carvalho, P. R., Ruiz, F. S., & Shirose, I. (1984). Drum drying of mango

puree. Coletanea-Instituto-de-Technologia-de-Alimentos , 24 (2), 193-202. van den Berg, C., & Bruin, S. (1981). Water activity and its estimation in food systems:

Theoretical aspects. In L. B. Rockland, & G. F. Stewart, Water activity: Influence on

Food Quality, ed (pp. 1-61). New YOrk: Academic Press.

Vega-Mercado, H., Gongora-Nieto, M. M., & Barbosa-Canovas, G. V. (2001). Advances in

dehydration of foods. Journal of Food Engineering , 49, 271-289.

Verticci, C. W., & Leopold, A. C. (1984). Bound water in soybean seed and its relation to

respiration and ambibitional damage. Plant Physio. , 75, 114-117.

Yongsawatdigul, J., & Gunasekaran, S. (1996a). Microwave vacuum drying of cranberries: Part

I: Energy use and efficiency. Journal of Food Processing and Preservation , 20, 121-

143.

Yongsawatdigul, J., & Gunasekaran, S. (1996b). Microwave vacuum drying of cranberries: Part

II: Quality evaluation. Journal of Food Processing and Preservation , 20, 145-156.

CHAPTER THREE

EFFECT OF DRYING METHODS ON THE PHYSICAL PROPERTIES AND

MICROSTRUCTURES OF MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER

ABSTRACT

Mango powders were obtained at water content below 0.05 kg water/kg dry solids using

Refractance Window® (RW) drying, freeze drying (FD), drum drying (DD), and spray drying

(SD). The spray-dried powder was produced with the aid of maltodextrin (DE = 10). The chosen drying methods provided wide variations in residence time, from seconds (in SD) to over 30 h

(in FD), and in product temperatures, from 20 °C (in FD) to 105 °C (in DD). The colors of RW- dried mango powder and reconstituted mango puree were comparable to the freeze-dried products, but were significantly different from drum-dried (darker), and spray-dried (lighter) counterparts. The bulk densities of drum and RW-dried mango powders were higher than freeze- dried and spray-dried powders. There were no significant differences (p ≤ 0.05) between RW and freeze-dried powders in terms of solubility and hygroscopicity. The glass transition temperature of RW-, freeze-, drum- and spray-dried mango powders were not significantly different (P 6 0.05). The dried powders exhibited amorphous structures as evidenced by the X- ray diffractograms. The microstructure of RW-dried mango powder was smooth and flaky with uniform thickness. Particles of freeze-dried mango powder were more porous compared to the other three products. Drum-dried material exhibited irregular morphology with sharp edges, while spray-dried mango powder had a spherical shape. The study concludes that RW drying can produce mango powder with quality comparable to that obtained via freeze drying, and better than the drum and spray-dried mango powders.

1. INTRODUCTION

Mango (Mangifera indica L.) is one of the most appreciated fruits in the world. The 2005 world production of mango was estimated at 28.5 million metric tons, of which 85% was produced in the following 10 countries: India (37.9%), China (12.9%), Thailand (6.3%), Mexico

(5.9%), Indonesia (5.2%), Pakistan (5.9%), Brazil (3.5%), Philippines (3.5%), Nigeria (2.6%), and Egypt (1.3%) (Evans, 2008). In the Philippines, mango ranks third among fruit crops next to banana and pineapple in terms of export volume and value, with a total of 1,023,907 metric tons harvested in 2007. The Carabao variety popularly known as "Philippine Super Mango" accounts for 73% of the country's production (BAS, 2009). This variety is acclaimed as one of the best in the world due to its sweetness and non-fibrous flesh. Fresh mangoes are perishable and may deteriorate in a short period of time if improperly handled, resulting in large physical damage and quality loss, ranging from 5% to 87% (Serrano, 2005). Gonzalez-Aguilar et al. (2007) reported that 100% of untreated ripe mango fruits of the ‘Hadin’ variety showed fungal infection and severe decay damage by the end of 18 days of storage at 25 °C. In order to take advantage of the potential health benefits of mango and add value to the commodity with lesser handling and transport costs, there is a need to develop mango products in forms of mango powders that not only have desired functionality but also are stable over a longer storage time. Mango powder offers several advantages over other forms of processed mango products like puree, juice and concentrate. Besides having a much longer shelf life due to considerable reduction in water content, the transport cost is also significantly reduced. Mango powders may also offer the flexibility for innovative formulations and new markets. For example, mango powders can be used as a convenient replacement for juice concentrates or purees, and as shelf-stable ingredients for health drinks, baby foods, sauces, marinades, confections, yogurt, ice cream, nutrition bars,

baked goods and cereals (Rajkumar et al., 2007). Development of high quality mango powder may match the increasing worldwide demand for more natural mango- flavored beverages either singly flavored or in multi-flavored products (FAO, 2007), and meet the great demand for natural fruit powders by the pharmaceutical and cosmetic industries.

Several drying technologies can be viable commercial options for manufacture of mango powders, including freeze drying, drum drying, spray drying and Refractance Window® drying.

Each has its own advantages and limitations. The final product obtained from these methods may differ in physicochemical or nutritional properties and microstructures. Freeze drying, also known as lyophilization, is a drying process in which the food is first frozen then dried by direct sublimation of the ice under reduced pressure (Oetjen and Haseley, 2004; Barbosa-Cánovas,

1996). To carry out a successful freeze drying operation, the pressure in the drying chamber must be maintained at an absolute pressure of at least 620 Pa (Toledo, 2007). Freeze drying is generally considered as the best method for production of high quality dried products (Ratti,

2001). But, it suffers from high production costs, high energy consumptions, and low throughputs (Ratti, 2001; Hsu et al., 2003; Caparino, 2000).

Drum drying is commonly used in production of low moisture baby foods and fruit powders (Kalogiannia et al., 2002; Moore, 2005). A drum dryer consists of two hollow cylinder drums rotating in opposite directions. The drums are heated with saturated high temperature

(120–170 °C) steam inside the drums. Raw materials are spread in thin layers on the outer drum surface and dry rapidly. The product is scraped from the drum in the form of dried flakes

(Kalogiannia et al., 2002; Saravacos and Kostaropoulos, 2002). A major likely drawback is undesirable cooked aromas and other severe quality losses in the final products caused by the high temperature used in the drying process (Nindo and Tang, 2007).

Spray drying is widely used in commercial production of milk powders, fruits and vegetables (Kim et al., 2009; Kha et al., 2010). This method has several advantages, including rapid drying, large throughput and continuous operation (Duffie and Marshall, 1953). During the drying process, the feed solution is sprayed in droplets in a stream of hot air (Saravacos and

Kostaropoulos, 2002). The liquid droplets are dried in seconds as a result of the highly efficient heat and mass transfers (Toledo, 2007). The finished product can be made in the form of powder, granules or agglomerates (Nindo and Tang, 2007). Spray drying processes can be controlled to produce relatively free flowing and uniform spherical particles with distinct particle size distribution (Barbosa-Canovas et al., 2005; Duffie and Marshall, 1953). However, due to the relatively high temperatures involved in spray-drying processes, this drying technique may cause loses of certain quality and sensory attributes, especially vitamin C, b-carotene, flavors and aroma (Dziezak, 1988). In addition, it is difficult to directly spray dry sugar- rich materials such as mango, because they tend to stick to the walls of the dryer (Bhandari et al., 1997a,b; Masters,

1985). Drying aids, such as maltodextrin, are widely added to the feed to increase glass transition temperature of the dried product and hence overcome the problem of stickiness during spray drying.

Refractance Window® (RW™) is a novel drying technique designed mainly to convert fruit puree into powder, flakes, or concentrates. The technology utilizes circulating hot water

(95–97 °C) to transfer thermal energy to a thinly spread liquid material placed on a polyester conveyor belt that moves at a predetermined speed while in direct contact with hot water. During drying, the thermal energy from hot water is transmitted to foods through the plastic conveyor by conduction and radiation. Water vapor from foods is carried away by aflowof filtered air over the thin layer. This technology offers several benefits when applied to fruits and vegetables. For

example, good retention of nutritional (vitamins), health-promoting (antioxidants) and sensory

(color, aroma) attributes were reported for dried carrots, strawberries and squash (Nindo and

Tang, 2007). The bright green color of pureed asparagus remained virtually unchanged when dried in the RW dryer, and was comparable to the quality of freeze-dried product (Abonyi et al.,

2002). In addition, energy efficiency of RW drying method compares favorably with other conventional dryers (Nindo and Tang, 2007).

Studies were reported that compared the influence of different drying methods on various quality attributes of fruits and vegetables, including the color of dehydrated apple, banana, carrots and potatoes (Krokida et al., 2001), b-carotene and ascorbic acid retention in carrots and strawberry (Abonyi et al., 2002), antioxidants and color of yam flours (Hsu et al., 2003), asparagus (Nindo et al., 2003), and antioxidant activities in soybean (Niamnuy et al., 2011), encapsulated β-carotene (Desobry et al., 1997), and color and antioxidant of beetroots (Figiel,

2010). However, no studies have been conducted to evaluate the effect of drying methods on mango powders in terms of color, bulk density, porosity, hygroscopicity, solubility, and microstructures. Thus, the objective of this work was to investigate the influence of four drying methods (Refractance Window® drying, freeze drying, drum drying and spray drying) on the physical properties and microstructures of resulting mango powders to provide better understanding in selecting drying techniques that can be applied toward the manufacture of high quality mango powder.

2. MATERIALS AND METHODS

2.1 Preparation of mango puree

Frozen mango puree (Philippine 'Carabao' var.) was acquired from Ramar Foods

International (Pittsburg, CA) (Fig. 1). The puree was produced following the manufacturer's

standard process that involved selection of ripened mangoes (95–100% ripeness), washing using chlorinated water, manual trimming, removal of any black portions of the peel and separation of stone/peel. The cleaned mango fruits went through a pulping machine that separated the pulp and discarded excess fibers. A buffer tank was used to standardize the puree at 14–15 °Brix. The mango puree was pasteurized, packed in 5 kg polyethylene (PE) bags, sealed and blast frozen at

–35 °C. Bags of puree were placed in carton boxes and stored at –18 °C. The frozen mango puree was kept at constant temperature while in transit from the Philippines to California and finally to Washington State University (Pullman, WA). This frozen mango was stored at –35 °C until it was ready for drying. The average moisture content of the mango puree was 6.5 ± 0.1 kg water/kg dry solids determined using standard oven method (AOAC, 1998).

2.2. Drying experiment

Frozen mango puree was thawed overnight at room temperature (23 °C), and afterward blended for 5 min to a uniform consistency using a bench top blender (Oster Osterizer, Mexico) with lowest speed setting. The puree was dried to below 0.05 kg water/kg dry solids by

Refractance Window® drying, freeze drying, drum drying, or spray drying. Due to difficulty in spray drying of this sugar-rich material, maltodextrin (DE = 10) (Grains Processing Corporation,

Muscatine, IA) was added to mango puree before spray drying. No addition of carrier was used for the other three drying systems. Detailed procedures for each drying method are described below:

2.2.1. Refractance Window® drying

A pilot scale Refractance Window® dryer at MCD Technologies, Inc. (Tacoma, WA) was used for drying mango puree. The dryer has an effective surface drying area of 1.10 m2 and length of 1.83 m in the direction of belt movement. The main components of the dryer included a

conveyor belt made of "Mylar®" (polyethylene terephthalate) plastic, a water pump, a hot water tank, a heating unit, two water flumes, a hood with suction blowers and exhaust fans, a spreader, and a scraper (Fig. 2). The drying was accomplished by spreading homogenized mango puree on the plastic conveyor belt that moves over the surface of circulating hot water. The thickness of the puree on the belt was 0.5–0.7 mm and was controlled using a spreader bar. The thermal energy from the circulating hot water (transferred to the puree through the belt) was used to remove moisture from the product (Nindo et al., 2003). Previous studies reported that the temperature of the product during drying rarely exceeded 80 °C (Abonyi et al., 2002). During drying operation, the temperature of circulating hot water was maintained between 95 and 97 °C similar to that reported by Abonyi et al. (2002) and Nindo and Tang (2007). The temperature of the circulating hot water was continuously monitored at the flume inlet and outlet section using pre-calibrated Type T thermocouple sensors (Fig. 3). The sensors were connected to a data acquisition unit equipped with monitoring software. Water vapor removal from the samples was facilitated by forcing the suction air (22 °C) with relative humidity (50–52%) over the puree at an average air velocity of 0.7 m/s (Abonyi et al., 2002). The product temperature was monitored along the drying section by scooping sufficient amount of samples and measured using an

Infrared thermometer (Raytek mini-FS, Sta. Cruz, CA) (Fig. 4). The dried product was in the form of thin sheets or flakes which were converted into powder form for analyses (Fig. 5). The residence time to dry the mango puree into flakes or powder was determined by monitoring the time travelled by the thinly spread mango puree from inlet to the outlet section of the plastic conveyor belt. Measurement of the residence time was performed in triplicate.

2.2.2. Freeze drying.

Freeze drying was carried out using a laboratory freeze dryer (Freeze Mobile 24, Virtis

Company, Inc., Gardiner, NY). The thawed mango puree was poured into a stainless pan to form a layer of 15 mm (Fig. 6). The samples were placed at –25 °C for 24 h before transferring to the freeze dryer. The vacuum pressure of the dryer was set at 20 Pa, the plate temperature was

20 °C, and the condenser was at – 60 °C. The residence time needed to dry the mango puree to below 0.05 kg water/kg dry solids was determined when the vacuum pressure had dropped to 30 mTorr (4 Pa).

2.2.3. Drum drying.

Figure 7 shows the illustration and actual drum drying experiment. A laboratory atmospheric double drum dryer (Model no. ALC-5, Blaw-Knox Co., Buffalo, NY) was utilized in this experiment. The dryer has two hollow metal drums with 0.15 m external diameter and

0.19 m length. The drums were internally heated by steam at 379.2 ± 7 kPa producing a temperature of 152 ± 2 °C. Preliminary experiments were conducted at different rotational speed settings in order to obtain dried sheets of below 0.05 kg water/kg dry solids. The clearance between the two drums was fixed at 0.01 mm allowing the puree to flow (forced by rotary action) into a thin layer as it passed through the gap. The drum temperature was allowed to stabilize before feeding the puree. This prepared puree was poured evenly over the hot pool area between the two drums. After traveling approximately three fourths of the revolution of the drums or ~15 cm distance, the dried product was scraped from the drum surface by doctor blades. The residence time for drying was recorded by taking three fourths of the time measured for one complete revolution of the drum. Due to stickiness of mango, the dried product at the exit section of the dryer tended to roll and build up while the drum was rotating forming an extruded-

like product and not the expected thin flakes. Thin sheet or flakes of dried product was obtained by carefully pulling the dried product as it goes out of the exit section of the dryer. The dried product removed from the two drums was mixed together for analysis because their appearance and moisture content were generally similar.

2.2.4. Spray drying.

The thawed mango puree was spray-dried in a pilot scale S-1 spray dryer (Anhydro

Attleboro Falls, MA) (Fig. 8). Before starting the experiment, the dryer was conditioned for 20 min by pumping de-ionized water through the atomizer with the dryer inlet and outlet temperatures set at 180 °C and 80 °C, respectively (Shrestha et al., 2008). The mango puree was pumped into the spray dryer chamber at a flow rate of 50 ± 2 g/min using Masterflex pump

(Cole-Parmer Instruments Co., Chicago, IL). The air temperature was maintained at 190 ± 2 °C

(dryer inlet) and 90 ± 2 °C (dryer outlet) during drying. These air inlet and outlet conditions are within the recommended temperatures of 180 – 220 °C and 90 – 110 °C, respectively, for spray drying of heat sensitive products at atmospheric pressure (Filkova and Mujumdar, 1995; Kim et al., 2009). The outlet temperature determines the thermal exposure of the sample during spray drying. It was observed during preliminary experiments that spray drying of mango puree without any carrier was not possible due to the high content of low molecular weight sugars (e.g. fructose, glucose, sucrose), similar to what had been reported by other authors (Abonyi et al.,

2002; Bhandari et al., 1997). Maltodextrin (DE = 10) having a median glass transition temperature of Tgm = 139.7 °C (Jakubczyk et al., 2010) was added to produce a non-sticky and free flowing powder (Bhandari et al., 1997). Preliminary experiments were carried out to obtain dried product that has better appearance and throughput. Three maltodextrin concentrations of

0.25, 0.35 and 0.45 kg/kg dried mango solids were tested for this purpose (Jaya et al., 2006;

Nindo and Tang, 2007; Sablani et al., 2008). By visual examination, the color and appearance of the dried mango powder from the three treatments showed very little variation. Hence, the spray- dried mango powder with the lowest maltodextrin concentration of 0.25 kg/kg dried mango solids was selected for comparison with other dried powders. The actual residence time to obtain mango powder with a moisture content below 0.05 kg water/kg dry solids was not measured, but the information from previous studies on spray drying of sugar-rich material was used to approximate the time.

2.3. Handling and packaging of samples.

The product from each drying process had unique geometries at the exit point, so different handling procedures were employed. Rectangular cake-like dried products obtained from the freeze drying process were collected and sliced into smaller pieces using a clean stainless steel knife and packed in leak-proof Ziploc® plastic bags. The spray-dried material, which appeared like agglomerated spherical shapes, was immediately packed in the same type of plastic bags after coming out from the dryer. The dried thin sheets collected from the drum and

RW drying processes were handled in a similar manner. All the samples sealed in Ziploc® bags we e placed inside alu inu -c ated p lyethylene bags. T p event idati n, all the packaged sa ples we e lushed with nit gen gas, heat sealed and st ed at 35 °C until further analyses.

2.4. Grinding and sieving

One hundred grams each of dried flakes or sheets obtained from different drying processes were ground using mortar and pestle. Sieving analysis was carried out by stacking and vibrating the sieves in ascending order of mesh sizes of 35, 45, 60 and 80 (American Society for

Testing and Materials, ASTM) to obtain particle sizes of 500, 350, 250 and 180 m

(International Standard for Organization, ISO), respectively (Barbosa-Cánovas 2005). Those 77

with particle sizes ranging between 180 – 500 m and flakes or sheets were evaluated in terms of color, bulk density and bulk porosity, while particle sizes of 180 – 250 m were analyzed for solubility, hygroscopicity and microstructures.

2.5. Water content

The water content of mango puree and dried flakes/powders made from RW, freeze, drum and spray drying methods was determined using the standard oven method at 70 °C and

13.3 kPa for 24 h (AOAC, 1998). The drying, cooling and weighing of samples was continued until the difference between two successive weighing was less than 1 mg.

2.6. Water activity

Water activity of the RW-, freeze-, drum-, and spray-dried mango powders was measured using water activity meter (Aqualab 3TE series, Decagon Devises, Pullman, WA). Duplicate samples were measured at 24.7 ± 1 °C.

2.7. Physical properties of mango powders

2.7.1. Color analysis

The dried mango in flakes or sheet forms and four different particle sizes of 500, 350,

250 and 180 m were evaluated for color comparison. Mango powders or flakes were poured into Petri dishes, slightly shaken to form a layer of 10 mm thickness and covered with transparent film (SaranTM Wrap, SC Johnson, Racine, WI). The International Commission on

Illumination (CIE) parameters L*, a* and b* were measured with a Minolta Chroma CR-200 color meter (Minolta Co., Osaka, Japan). The colorimeter was calibrated with a standard white ceramic plate (L* = 95.97, a* 0.13, b* 0.30) prior to reading. Corresponding L* value

(lightness of color from zero (black) to 100 (white); a* value (deg ee edness (0 60)

g eenness (0 t 60); and b* values (yell wness (0 60) blueness (0 t 60) were measured for all the samples. The average L, a* and b* values were obtained from 6 readings taken from

each of 5 locations. The hue angle, and chroma, expressed as = and =

, respectively were also calculated (Abonyi et al., 2002). Hue is a color attribute by which red, yellow, green and blue are identified, while chroma distinguishes between vivid and dull colors.

For color comparison with the original mango puree, 2 g each of RW-, freeze-, drum-, and spray-dried mango powders (~250 m) with water content of 0.017 ± 0.001, 0.023 ± 0.002,

0.013±0.001 and 0.043±0.003 kg water/kg dry solids were reconstituted by adding an amount of

12.10, 12.04, 11.96 and 11.70 g of distilled water, respectively using material balance. The reconstituted mango powders produced slurries with moisture content of 6.143 kg water/kg dry solids similar as the original mango puree. The reconstitution of mango powder was carried out by mixing the powder and water at 23 °C while vortexing (Fisher Scientific mini vortexer, USA) until the powder was completely dissolved. The L*, a* and b*, H* and C* values were immediately measured and calculated following the same procedure employed for mango flakes and powders. The total change in color of the reconstituted mango powders with reference to the

original puree were computed as: where, subsc ipt “ ” den tes the c l iginal pu ee (Jaya and Das, 2004; Nindo et al., 2003).

2.7.2. Bulk density

The bulk density of the mango powder obtained from different drying processes and particle sizes was measured following the procedure described in previous studies with modification (Barbosa-Cánovas et al., 2005; Goula et al., 2008). Approximately 5 g of mango

powder was freely poured into a 25 ml glass graduated cylinder (readable at 1 ml) and the samples were repeatedly tapped manually by lifting and dropping the cylinder under its own weight at a vertical distance of 14 ± 2 mm high until negligible difference in volume between succeeding measurements was observed. Given the mass m and the apparent (tapped) volume V of the powder, the powder bulk density was computed as m/V (kg/m3). The measurements were carried out at room temperature in three replicates for all samples.

2.7.3. Particle density and bulk porosity

The particle densities of mango powders obtained by different drying methods were calculated by adopting the pycnometer method. A 2.5 ± 0.04 g of each of the RW-, freeze-, drum-, and spray-dried ang p wde s (180 2 0 m) was placed in an empty liquid pycnometer (25 ml), and filled with measured volume of toluene. Toluene was used because of its ability to penetrate the finest external pores connected to surface of the material without dissolving the material. Bulk porosity (εb) was calculated by determining the ratio of particle density (p) and bulk density (b) using the qs. (1) (3) as (Krokida et al., 1997):

 ( 1 )

 ( 2)

 ( 3)  where  is the bulk density of mango solids,  is the particle density of the solids, is the mass of mango solids, and is the total and volume of the dry solids, respectively.

2.7.4. Solubility

Solubility of mango powder was determined using the procedure developed by East and

Moore (1984) as adopted by Cano-Chauca et al. (2005). One gram of the powder (dry basis) was

dispersed in 100 ml distilled water by blending at high speed (~13,000 rpm) for 5 min using an

Osterizer blender (Oster, Mexico). The dispersed mango powder was then centrifuged at 3000 g for 5 min. A 25 ml aliquot of the supernatant was carefully pipetted and transferred to a pre- weighed aluminum dish and then oven-dried at 105 °C for 5 h. Drying was continued and weighed every hour for 2 h. The solubility of the powder (%) was determined by taking the weight difference.

2.7.5. Hygroscopicity

Ten (10) g a s each RW-, ee e-, d u - and sp ay-d ied ang p wde s with pa ticle si es 180 2 0 m and moisture content below 0.05 kg H2O/kg mango solids were placed in an open glass container. Three replicate samples for each product were put separately in three sealed humidity jars containing NaCl saturated solution (75.5 % humidity) and stored at 25 °C for 7 days. Samples were prepared at 20 °C. Hygroscopicity, HG (%) or 1 g of adsorbed moisture per 100 g dry solids (g /100 g) was calculated using the following equation:

(4)

where (g) is the increase in weight of powder after equilibrium, is the initial mass of powder and (% wb) is the free water contents of the powder before exposing to the humid air environment (Jaya and Das, 2004, Sablani et al., 2008, Tonon et al., 2008).

2.8. Glass transition temperature

Glass transition temperature (Tg) of mango powders with water activity below 0.2 was measured using differential scanning calorimeter (DSC, Q2000, TA Instruments, New Castle,

DE), following the procedure described by Syamaladevi et al. (2010). The calorimeter was calibrated for heat flow and temperature using standard indium and sapphire. Twelve to sixteen

illig a s each ang p wde sa ple was sealed in an alu inu pan (v lu e 30 μl), cooled down from 25 t 0 using liquid nit gen and then equilib ated 10 in. The sa ples at 90 °C were scanned to 90 °C then cooled down to 25 °C. Scanning of all samples was carried out using the same heating and cooling rate of 5 °C/min. To avoid condensation on the surface of the powder particles, a nitrogen carrier gas was purged at a flow rate of 50 ml/min.

The onset- (Tgi), mid- (Tgm) and end-point (Tge) values of the mango powders were determined by finding the vertical shift in the heat flow-temperature diagram. All measurements were performed in duplicate.

2.9. X-ray diffraction

X-ray diffraction (XRD) characteristics of mango powders obtained from different drying processes were investigated using a Siemens D-500 diffractometer (Bruker, Karlsruhe,

Ge any). The p wde sa ples (180 2 0 μ ) we e placed and slightly p essed in an aluminum holder using a glass slide. The diffractometer was operated at a wavelength of 0.15 nm and the input energy was set at 30 mA and 35 kV. Diffractograms were taken between 5° and 50° (2θ) with a step angle of 0.02° and scan rate of 1 s per step. The XRD patterns of all the samples were plotted for comparison.

2.10 Microstructure analyses

A s all quantity ang p wde s (180 2 0 m) from different drying systems were mounted on aluminum stubs and coated with a fine layer of gold (15 nm) using a Sputter gold coater (Technics Hummer V, Anatech, San José, CA). All powder samples were examined by

Scanning Electron Microscopy using SEM Hitachi S‐570 camera (Hitachi Ltd., Tokyo, Japan)

operated at an accelerating voltage of 20 kV. Micrographs were photographed at a magnification of 100x, 300x and 1000x at scale bar of 0.30 mm, 100 m and 30 m.

The microstructure of samples prepared for hygroscopicity experiments were also analyzed to identify possible relationships between the obtained hygroscopicity values for each mango powder sample using a Quanta 200F Environmental Scanning Electron Microscope (FEI,

Field Emission Instruments, Hillsboro, Oregon, USA). The low vacuum mode (200 Pa) was used during scanning to allow measurement of samples at their native state. Observations were carried out with an accelerated voltage of 30 kV and magnification of 700x at a scale of 100 m.

2.11. Statistical analysis

All experiments were carried out at least in duplicate, the results analyzed using the general linear model procedure of SAS (SAS Institute Inc., Cary, NC), and the means separated by Tukey-honest significant difference test with a confidence interval of 95% used to compare the means. Mean standard deviations are presented in the results.

3. RESULTS AND DISCUSSION

3.1. Residence time, water content and product temperature

The residence time during drying of mango puree from the initial moisture content of

6.52 kg water/kg mango solids to below 0.05 kg water/kg mango solids was accomplished in 180

± 0.15, 111,600 ± 5100 and 54 ± 0.2 s for RW, FD and FD, respectively, and less than 3 s with

SD (Table 1). It should be noted here that the residence time used for SD was only an approximation based on the data reported by Desobry et al. (1997) and Jayasundera et al.

(2011b). The actual residence time during spray drying of mango powder in our study might be higher than 3 s because of the difference in drying conditions and specifications of the spray

dryer used as compared from the literature. Nevertheless, the estimated residence time for SD is definitely much smaller than for RW, freeze and drum drying. The product temperatures measured for each drying process was 74 ± 2 °C (RW), 20 ± 0.5 °C (FD) , 105 ± 5 °C (DD) and

90 ± 2 °C (SD).

3.2. Physical properties of mango powder

3.2.1. Color analysis

The color of the dried product (mango flakes/sheet) or powders of different particle sizes were affected by the drying methods. Visual examination showed that spray-dried (agglomerate powder particles) and drum-dried mango powder had the lightest and darkest color, respectively.

The color difference between mango powder obtained using RW and FD was not significantly different (p ≤ 0.0 ) (Fig. ). Hunte c l t isti ulus values for mango powder of different particle sizes are presented in Fig. 10 and 11. Overall, the product (flakes) at exit had a significant difference in the L value (lightness) among the RW-, freeze-, drum-, and spray-dried mango flakes or powders, except for the RW and freeze-dried powder with particle size of 500 and 350 m which showed no significant variation (p ≤ 0.0 ) (Fig. 10a). The si ila ity in L- value for RW, FD and DD powders of the smallest particle size (180 m) may be attributed to negligible effect on reflectance.

The mango powder produced by spray drying had the highest L value, while the drum- dried mango powder appeared to have the lowest L value (indicating darkest color). The lighter color in spray drying was due to the addition of maltodextrin carrier which was necessary to reduce the stickiness of the mango to allow the spray drying process to be effective (Abonyi et al., 2002; Jaya et al., 2006). While the outlet temperature during sp ay d ying eached 0 2 , the d ying ti e was ve y sh t (l 3 sec) as reported by Desobry et al. (1997). Hence the color

degradation was limited. On the other hand, the darker color of the drum-dried mango powder can be attributed to high drying temperature. Such effect confirmed previous studies on strawberry puree (Abonyi, et al., 2002) wherein color degradation was greatly influenced by high processing temperatures. The dark color in drum-dried mango flakes or powder can be characterized by browning reaction or Maillard reaction caused by the chemical reactions between sugars and proteins (Potter and Hotchkiss, 1995). Moreover, caramelization of sugars in mango can occur due to high temperature contributing to darkening during drying. The dominant color in mango puree is yellow and hence can be best represented by Hunter color b*

(yellowness) to distinguish the color difference of the resulting mango powders as affected by the drying process. No significant difference was observed in b* value (yellowness) between

RW and freeze-dried mango powder while there was a highly significant difference between spray and drum-dried product (p ≤ 0.0 ) (Fig. 10b). h a value vividness in yell w c l of 250 m particle size RW and freeze dried mango flakes and p wde s sh wed n signi icant di e ence, but RW-d ied ang p wde with pa ticle si e 3 0 00 m were of a more vivid yellow color than freeze dried mango powder having obtained the highest chroma value (Fig.

11a). The hue angle value in spray-dried mango powder was the highest but its chroma value is very low indicating a dull color (Fig. 11b). RW dried mango flakes or powder at all particle sizes obtained a higher hue angle compared to freeze and drum-dried mango powders suggesting that

RW-dried mango powder is more vivid in its yellow color implying that it will be more attractive and appealing to consumers. The overall distinct vivid yellow color of the RW-dried mango may be indicative high β-carotene retention. Abonyi et al. (2002) rep ted that β-carotene in RW and freeze-dried carrot puree was 53% and 55% higher compared to drum-dried products, respectively. Wagner and Warthesen (1995) reported that the yellow and red color of carrot

slices is attributed to the presence of carotenes. Also, the b* (yellow) values for raw and puree sweet p tat we e highly c elated with β-carotene content (Ameny & Wilson, 1997). The minimal color change of product produced by RW and freeze drying suggests the appropriateness of these processes to produce high quality products. The comparable yellow color of RW and freeze-dried mango powder can also be attributed to low product temperature for RW (74 ± 2 °C) and freeze-dried (20 ± 0.5 °C), compared to spray-dried (90 ± 2 °C) and drum-dried (105 ± 5 °C) mango powder.

The reconstituted mango powder was prepared by adding water to achieve the same solid contents as the original mango puree. Visual examination of the color of the reconstituted RW-, freeze-, drum-, and spray-dried mango powders showed variations in comparison with the original mango puree (Fig. 12). Luminosity (L*) values as presented in Table 2 showed no significant difference between reconstituted RW- and freeze-dried mango puree and both are similar in luminosity to the original puree. Reconstituted drum-dried mango puree was darker as expected because of the darker powder. The result is in agreement with the work of Abonyi et al.

(2002) wherein a drum-dried carrot puree was perceived as darker in comparison with powders produced by spray, freeze and RW drying methods. Spray-dried mango powder was darker than

RW- and freeze-dried but lighter than reconstituted drum dried mango puree. The original mango puree and reconstituted RW and freeze-dried mango powders are significantly different but comparable in terms of vividness and saturation of yellow color. On the other hand, the reconstituted drum-dried and spray-dried mango puree had lower chroma values indicating less saturation and dull yellow appearance. A comparable result was also observed for hue angle among the original puree, reconstituted RW- and freeze-dried mango puree while reconstituted drum-dried mango puree had a low hue angle value, which indicates a dull yellow color. The

reconstituted spray-dried mango puree had the highest hue angle value but because of its low lightness and chroma values, it produced a grayish pale color. The reconstituted mango powder from drum drying process showed the highest deviation in color with respect to the original mango puree having a E value of 9.22 ± 0.01 followed by the reconstituted spray-dried mango puree with E value of 6.23 ± 0.02 (e.g. lightest). The reconstituted RW-dried mango puree had the lowest color difference with E value = 1.22 ± 0.02, a value very close to reconstituted freeze-dried mango puree with E value = 1.57 ± 0.02. The distinct superiority of RW drying process against drum and spray drying processes in producing mango powder in the present experiment is in corroboration with previous studies for asparagus (Nindo et al., 2003), and carrots and strawberry (Abonyi et al., 2002).

3.2.2. Bulk density and porosity

For all drying methods, the bulk density of mango powders increased and their porosity decreased with decreasing particle size (Fig. 13 & 14). These results may be attributed to the decrease in the inter-particle voids of smaller sized particles with larger contact surface areas per unit volume. Similar observation was reported for bulk density of ginger powder at different particle sizes (Xiaoyan, 2008). It was also consistent with the explanation by other authors that powder characteristics such as particle size may result in significant changes in bulk density and porosity (Barbosa-Cánovas 2005).

Freeze- and spray-dried mango powders had significantly lower bulk densities and higher porosities compared to drum- and RW-dried products (p ≤ 0.05) (Fig. 13 & 14). It is well recognized that in freeze drying of foods in the form of either puree or as a whole, the material is first frozen allowing it to maintain its structure following sublimation of ice under high vacuum

(Oetjen and Haseley, 2004). Since liquid phase in the material is not present during this process,

there is no transfer of liquid water to the surface, but instead the ice changes to vapor below the collapse temperature without passing the liquid state (Krokida et al., 1997). In effect the collapse and shrinkage of the product is prevented thereby resulting in a porous dried material (Karel,

1975).

The higher porosity or lower bulk density in spray-dried mango powder was due to the addition of maltodextrin (Fig. 14). Shrestha et al. (2007) demonstrated that increasing maltodextrin concentration in tomato pulps led to the decrease in bulk density. Goula and

Adamopoulus (2008) also explained that maltodextrin is considered a skin-forming material and by using it as carrier can induce accumulation and trapping of air inside the particle causing it to become less dense and porous.

On the other hand, the bulk porosity and density of RW- and drum-dried mango powder were significantly lower and higher than freeze and spray dried product, respectively with drum dried product exhibited the lowest porosity (p ≤ 0.0 ) (Fig. 13 & 14). Du ing d u d ying, the mango puree poured inside a pool between the two drums has vapor bubbles bursting at the free surface and spattered along side of the two drum surfaces as triggered by high temperature

(above boiling). The high temperature used in drum drying may have caused collapse which resulted in more compact and rigid product. These characteristics resulted in lower porosity when compared to freeze- or spray-dried mango powder. RW-dried mango powder exhibited low porosity compared to freeze- and spray-dried mango powder but significantly higher than drum- dried powder (p ≤ 0.0 ) (Fig. 13 & 14). RW is categ i ed as a di ect d ying technique si ila t drum-drying (Nindo and Tang, 2007), except that the energy is indirectly transferred via plastic film instead of steel as in drum drying. Apparently, both drying processes seem to produce a similar form of end product.

3.2.3. Solubility

Solubility is the most reliable criterion to evaluate the behavior of powder in aqueous solution. This parameter is attained after the powder undergoes dissolution steps of sinkability, dispersability and wettability (Chen and Patel, 2008). There was no significant difference in the solubility between spray and drum-dried mango powder, while both were significantly higher compared to RW and freeze-dried product (p ≤ 0.0 ) (Table 3). The high solubility of spray- dried mango powder can be attributed to the addition of maltodextrin (DE = 10). This result was in agreement with the study reported by Cano-Chauca et al. (2005) where they concluded that solubility of mango powders increased when maltodextrin was added during spray-drying.

Maltodextrin is a material that serves as coating agent as the particle crust is developed during spray drying resulting in a product that is highly soluble (Desai and Park 2004). Cai and Corke

(2000) also confirmed that maltodextrin as a carrier and coating agent increased the solubility of spray-dried betacyanins. The atomization of mango puree during spray drying may also contribute to solubility of spray-dried product. Fibers present in mango might have been broken into tiny pieces as a result of high atomization of the material resulting in increased solubility.

From the above observations, maltodextrin was proven effective in increasing solubility of spray-dried mango powder. However, spray drying of mango puree containing 25 kg/kg dried mango solids significantly altered the total color change of the resulting mango powder as earlier discussed. Likewise, the cyclone recovery of mango powder at this maltodextrin concentration was only 37.8 ± 1.8 % (data not shown), far below the > 50 % benchmark cyclone recovery for a marginally successful spray drying process of sugar-rich material (Bhandari et al.,1997;

Jayasundera et al., 2011a,b). The application of alternative drying aids such as proteins and low molecular surfactants may improve the recovery and quality of spray-dried mango powders and

help in maintaining higher solubility (Jayasundera et al., 2011a,b,c; Adhikari et al., 2009a,b).

Recently, a type of protein called "Protein X" developed at the University of Sydney, was found to increase the recovery of sugar-rich material of up to 80% by just adding a small amount (<5%) to the sticky fruit juice or puree (Wang, et al., 2011).

Drum dried mango powder also had a high solubility value that was not significantly different from spray-dried mango powder. The higher solubility of drum-dried samples could be attributed to a higher degree of macromolecular disorganization of the material as affected by drying process and condition. However, its inferior dark color may restrict consumer acceptance even though its solubility is high.

The solubilities of freeze- and RW-dried mango powders were similar and significantly lower than that of spray- and drum-dried mango powders. Both drying methods are gentle in terms of product temperature (Table 3). One possible reason for the lower solubility of those samples is that the cell structure of mango puree was not disrupted and smaller amounts of solids were dissolved to become part of the supernatant.

3.2.4. Hygroscopicity

A demarcation or cut-off values for hygroscopicity (HG) of mango powder ranging from

5.13% to 9.38 % were considered as the basis for comparing the results in our study. These figures were based on the average range of hygroscopicity values of instant coffee (lower HG) and tomato soup powder (higher HG) as calculated by Jaya and Das (2004). Table 3 shows the hygroscopicity of mango powders made from Refractance Window, freeze, drum and spray drying. The drum-dried mango powder exhibited the highest hygroscopicity (20.1 ± 0.88 %),

74% higher than the higher limit cut-off HG, indicating its strong capacity to attract water molecules when in contact with the surrounding air. Mujumdar (2007) explained that drum

drying of sugar-rich fruits requires high temperature and usually a dry to very low moisture thin sheet product. This drying condition usually causes the product to be very hygroscopic and often other quality attributes are degraded. The lower hygroscopicity value obtained for spray drying

(16.5 ± 0.06) or 68.91% higher than the benchmark HG can be attributed to the addition of maltodextrin in the mango puree before drying. Tonon et al. (2000) demonstrated that the hygroscopicity of spray-dried acai powder gets lower as the concentration of maltodextrin was increased. Similar observation was confirmed during spray drying of cactus pear juice

(Rodríguez-Hernández et al., 2005), sweet potato powder (Ahmed et al., 2009) and betacyanin pigments (Cai and Corke, 2000), suggesting that maltodextrin is an efficient carrier agent in lowering hygroscopicity of dried material. Consequently, when this carrier is added, other quality att ibutes the ang p wde such as c l and β-carotene (data not shown) were found inferior. There was no significant difference in the hygroscopicity between RW- (18.0 ± 0.36 %) and freeze-d ied (18.0 0.1 %) ang p wde (p ≤ 0.0 ) btaining a si ila inc ease 71. 0

% based on the higher limit cut-off HG, suggesting the superiority of RW over drum and spray- dried mango powder.

The small variation of moisture content of the different samples may have direct relationship with the hygroscopicity as shown in Table 3. Tonon et al. (2000) expound that the low moisture spray-dried acai has the greater capacity to absorb water from the surrounding air and hence is more hygroscopic. However, his indings n the istu e hyg sc picity relationship cannot be generalized for all commodities. Ahmed et al. (2009) reported that hygroscopicity of spray-dried sweet potato was greatly affected by carrier agents with no direct relationship to varying moisture content. The present study was in agreement with his findings wherein maltodextrin greatly influenced the hygroscopicity of the spray-dried mango powder.

3.3. Glass transition temperature

The glass transition temperatures of RW-, freeze-, drum- and spray-dried mango powders were determined in the water activity range from 0.169 to 0.177 and water content below 0.05 kg water/kg mango solids (Table 4). The onset of Tg (Tgi) values of mango powders were slightly lower than the room temperature (25 °C) normally used for long-term storage of food powders.

Adhikari et al. (2009a) reported that water activity below 0.2 is the value commonly applied for processing of spray-dried powders in a commercial scale. The Tgi values of mango powders ranged from 18.7 ± 0.2 to 26.1 ± 0.8 with RW-dried mango powder displaying the lowest Tgi and being significantly different from spray-dried product having the highest Tgi value (p < 0.05).

The higher Tg value of spray-dried product may be attributed to the addition of higher Tg maltodextrin before spray drying. Maltodextrin has high molecular weight and adding this amorphous material to low molecular weight sugar-rich material such as mango will cause it to increase the glass transition temperature of the product (Jaya and Das, 2004). The glass transition temperature of RW- freeze- and drum-dried mango powder produced without the drying aids showed no significant differences among each other (p < 0.05). It can also be seen from the data that the glass transition temperature of mango powder with water activity < 0.2 was not affected by the drying process and condition.

3.4. X-Ray diffraction

X-ray diffraction is a common technique used to confirm the crystalline a ph us state of dried products in a powder form. In general, crystalline material shows a series of sharp peaks, while amorphous product produces a broad background pattern. The X-ray diffraction patterns of RW-, freeze-, drum- and spray-dried mango powders at aw < 0.2 clearly exhibited amorphous characteristics and showed no crystalline peak formation (Fig. 15). The rapid drying

of low molecular weight sugars present in mango (sucrose, fructose and glucose) and organic acids that happened under RW, drum and spray drying processes tend to produce amorphous metastable state dried products because of insufficient time to crystallize (Jayasundera et al.,

2011a,b). The diffractogram of freeze-dried mango powder obtained in this study was similar to the one reported by Harnkarnsujarit and Charoenrein (2011) and Haque & Roos (2005).

The X-ray patterns and shapes for all the mango powders tested were similar to spray- dried sucrose indicating the dominance of sucrose sugars present in mango (Adhikari et al.

(2009a). However, it is interesting to note that the intensity count for drum-dried mango powder as shown in the diffractograms was significantly lower compared to the other three powder products. This could be due to puree gelatinization that occurred before the actual drum drying, resulting in the disorganization of intra- and intermolecular hydrogen bonding between water and starch molecules (Gavrielidou et al., 2002). Anastasiades et al. (2002) confirmed that gelatinization process causes irreversible changes in the physical structure of starch, which is present in mango resulting in degradation of molecular structure and loss of crystallinity. The absence of crystalline peaks confirmed that no substantial changes occurred on the hygroscopicity of RW-, freeze-, drum- and spray-dried mango powders as earlier discussed.

3.5 Microstructure

Scanning elect n ic g aphic studies ang p wde s (180 2 0  ) btained by di e ent d ying p cesses a e sh wn in Fig. 16 18. The microstructure of RW-dried mango powder was smooth, and flaky with uniform thickness (Fig. 16a & b). The uniformity of the flake thickness was the result of a controlled feeding of mango puree using a spreader bar at the inlet section of the RW dryer. During dying, the thinly spread mango puree on the surface of the plastic film conveyor is undisturbed, except for removal of moisture, as it moves toward the

other end of the dryer, hence producing a continuous sheet with thickness nearly equal. Crushing the RW-dried mango flakes into powder form produced irregularly shaped particles while maintaining its thickness. The two sides of a single particle were smooth indicating more flowability and less susceptibility to oxidation because of lesser surface area. Freeze-dried mango powder (Fig. 16c & d), showed a skeletal-like structure and was more porous than the other mango powders. This result happens because the ice in the material during freeze drying helps prevent shrinkage and collapse of the structure and shape resulting in an insignificant change in volume (Ratti, 2001). The microstructure of drum-dried mango powder (Fig. 16e & f) was compact and exhibited irregular particles with sharp edges and considerable indentation as a result of crushing into powder. Caric and Kalab (1987) reported similar structure for drum-dried milk powder. They explained that the compactness of drum-dried milk powder was due to deaeration of raw milk during drum drying. It is also evident that the drum dried sheets are smooth on one side that is in direct contact with the drum surface, while visible corrugation and crinkle was observed on the other side. These observations are in agreement with the microstructure of drum-dried pre-gelatinized maize starches as described by Anastasiades et al.

(2002). Spray-dried mango powder (Fig. 16g & h) has spherical or oval shape and smooth surface particles due to effect of spray-drying condition, which was maintained at inlet temperature of 190 ± 2 °C during drying. Nijdam and Langrish (2005) demonstrated that milk powders spray-dried at inlet temperature of 200 °C have spherical, smooth and larger particles, while particles were smaller and shriveled when the inlet temperature was reduced to 120 °C.

The smooth spherical-shaped mango powder contributed to its high porosity compared to the other three drying methods.

Individual particles of mango powders obtained from different drying processes were further examined (Fig. 17). The RW-dried mango powder clearly showed a composite sheet with distinguishable internal pores within the particle indicating that some emptied space during evaporation is not replaced as the mango puree is dried. These pores might have contributed to the higher porosity of RW-dried compared to drum-dried mango powder. Drum-dried mango powder developed a fine particle surface allowing it to be more compact and rigid. Spray-dried mango powder particle showed a very fine and smooth surface, but it may not be indicative of being rigid and compact as it contains vacuoles forming a hollow spherical shape (Cai and

Corke, 2000). Apparently, external pores were developed within the internal pores of a single particle freeze-dried mango powder. This further explains why the porosity of freeze-dried materials always is higher in comparison with other drying methods.

The ic st uctu es ang p wde s (180 2 m) exposed at 23 °C for 7 days at high relative humidity (75.5 %) showed different water adsorption behavior (Fig. 18). The particle surfaces and edges of RW- and freeze-, and spray-dried mango powders were still visible indicating that the materials adsorbed less when compared to drum-dried mango powder wherein its particles were nearly dissolved with water. This result confirmed the higher hygroscopicity value obtained for drum-dried mango powder compared to the other three powder products.

4. CONCLUSIONS

The physical properties and microstructures of mango powders were significantly affected by drying methods applied. Drying of mango puree to below 0.05 kg/kg dry mango solids was accomplished in 180 ± 0.15, 111,600 ± 5100 and 54 ± 0.2 s for RW, FD and FD, respectively, and less than 3 s with SD. The color of drum-dried mango powder was severely degraded because of high processing temperature, while the spray-dried powder became lighter

due to the addition of maltodextrin. On the other hand, the color of RW-and freeze-dried mango powder was comparable at different particle sizes. The reconstituted RW-dried mango puree showed a slight deviation in comparison with the original puree and was very close to reconstituted freeze-dried mango puree. Reconstituted drum- and spray-dried mango puree suffered discoloration and were respectively darker and lighter than the original puree. Both the drum- and RW-dried mango powders were significantly denser compared to freeze- and spray- dried. Regardless of the particle size and shape, freeze-dried mango powder had the highest bulk porosity compared to the other three drying methods. Drum-dried mango powder was the most hygroscopic while spray-dried was the least hygroscopic. There was no significant difference in hygroscopicity and solubility between RW and freeze-dried material. The glass transition temperatures of RW-, freeze-, drum- and spray-dried powders were not significantly different at water activity just below 0.2. The X-ray diffraction patterns of RW-, freeze-, drum- and spray- dried mango powders (aw < 0.2) clearly exhibited amorphous characteristics and showed no crystalline peak formation. The microstructure analysis verified the variations in bulk density, porosity, solubility and hygroscopicity of mango powders. Also, the microstructures of individual particles played an important role in analyzing the physical properties of mango powders. Overall, our study concludes that the RW drying method can produce superior quality mango powder compared to drum and spray drying, while it is highly comparable to freeze drying. The study provides an opportunity to the powder processing industry in selecting a better drying technique that can be utilized for the manufacture of high quality mango powder.

ACKNOWLEDGMENTS

We thank the Ford Foundation International Fellowship Program (IFP)/Institute of

International Education-New York through IFP-Philippines for providing the financial support,

and the Philippine Center for Postharvest Development and Mechanization (PhilMech) for granting study leave to Ofero Caparino. Special thanks to Richard E. Magoon and Karin M.

Bolland of MCD Technologies, Inc (Tacoma, WA) for allowing the use of their RW drying facilities, and for their assistance in doing the experiments; Eng. Frank Younce and Ms. Galina

Mikhaylenko for assisting with the drying and physical analysis experiments, respectively,

Roopesh Syamaladevi for assisting in X-ray diffraction, and Dr. Valerie Lynch-Holm for helping with the SEM and FESEM imaging.

REFERENCES

Abonyi, B. I., Feng, B. I., Edwards, C. G., & Tang, J. (2002). Quality retention in strawberry and

carrot purees dried with Refractance Window system. Journal of Food Science , 67,

1051–1056.

Adhikari, B., Howes, T., Bhandari, B., & Langrish, T. G. (2009a). Effect of addition of proteins

on the production of amorphous sucrose powder through spray drying. Journal of Food

Engineering , 94 (2), 144-153.

Adhikari, B., Howes, T., Wood, B. J., & and Bhandari, B. R. (2009b). The effect of low

molecular weight surfactants and proteins on surface stickiness of sucrose during powder

formation through spray drying. Journal of Food Engineering , 94, 135-143.

Ameny, M. A., & Wilson, P. W. (1 7). Relati nship between Hunte l Values and β-

Carotene Contents in White-Fleshed African Sweetpotatoes (Ipomoea batatas Lam).

Journal of the Science of Food and Agriculture , 73, 301-306.

Anastasiades, A., Thanou, S., Loulis, D., Stapatoris, A., & Karapantsios, T. D. (2002).

Rheological and physical characterization of pregelatinized maize starches. Journal of

Food Engineering , 52, 57-66.

AOAC. (1998). Official methods of analysis. Washington, DC, USA: Association of Official

Analytical Chemists.

Barbosa-Cánovas, G. V. (1996). Dehydration of Foods. New York: Chapman and Hall.

Barbosa-Canovas, G., V., Ortega-Rivas, E., Juliano, P., & Yan, H. (2005). Food Powders:

Physical Properties, Processing, and Functionality. New York: Kluwer Academic/Plenum

Publishers.

BAS. (2009). Situation Report on Selected Fruit Crops. Retrieved March 1, 2011, from

http://www.bas.bas.gov 98

Bhandari, B. R., Datta, N., & Howes, T. (1997a). Problems associated with spray drying of

sugar-rich foods. Drying Technology , 15 (2), 671-684.

Bhandari, B. R., Datta, N., Crooks, R., Howes, T., & Rigby, S. (1997b). A semi-emperical

approach to optimize the quantity of drying aids required to spray dry sugar-rich foods.

Drying Technology , 15 (10), 2509-2525.

Cai, Y. Z., & Corke, H. (2000). Production and properties of spray-dried amaranthus betacyanin

pigments. Journal of Food Science , 65 (6), 1248-1252.

Cano-Chauca, M., Stringheta, P. C., Ramos, A. M., & Cal-Vidal, C. (2005). Effect of the carriers

on the microstructure of mango powder obtained by spray drying and its functional

characterization. Innovative Food Science and Emerging Technologies , 6, 420-428.

Caparino, O. A. (2000). Characteristics and quality of freeze-dried mango powder pre-frozen at

different temperatures. Philippine Agricultural Scientist Journal , 83 (4), 338-343.

Caric, M., & Kalab, M. (1987). Effects of drying techniques on milk powder quality and

microstructure: A review. Food Microstructure , 6, 171-180.

Chen, X. D., & Patel, K. C. (2008). Manufacturing better quality food powders from spray

drying and subsequent treatments. Drying Technology , 26, 1313-1318.

Desai, K. G., & Park, H. J. (2004). Solubility studies of valdecoxib in the presence of carriers,

co-solvent and surfactants. Drug Development Research , 62, 41-48.

Desobry, S. A., Netto, F. M., & Labuza, T. P. (1997). Comparison of spray-drying, drum-drying

and freeze-d ying β-carotene encapsulation and preservation. Journal of Food Science ,

62 (6), 1158-1162.

Duffie, J., & Marshall, W. (1953). Factors influencing the properties of spray-dried materials.

Chem. Eng. Progr. , 49, 417-423.

Dziezak. (1988). Microencapsulation and encapsulated ingredients. Food Technology , 42, 136-

148.

Eastman, J. E., Moore, C. O. 1984. Cold Water Soluble Granular Starch for Gelled Food

Composition. U.S. Patent 4465702.

Evans, E. A. (2008). Recent Trends in World and U.S. Mango Production Trade, and

Consumption. University fo Florida. Gainesville, Florida: Food and Resource Economics

Department.

FAO. (2007). FAOSTAT database collections, agricultural data, Food and Agriculture

Organization of the United Nations. Retrieved March 10, 2011, from

http://www.faostat.org

Figiel, A. (2010). Drying kinetics and quality of beetroots dehydrated by combination of

convective and vacuum-microwave methods. Journal of Food Engineering , 98, 461-470.

Filkova, I., & Mujumdar, A. S. (1995). Industrial spray drying systems. In A. S. Mujumdar,

Handbook of Industrial Drying (pp. 263-307). New York: Marcel Dekker.

Gavrielidou, M. A., Vallous, N. A., Karapantsios, T. D., & and Raphaelides, S. N. (2002). Heat

transport to a starch slurry gelatinizing between the drums of a double drum dryer. Journal

of Food Engineering , 54, 45-58.

Gonzalez-Aguilar, G. A., Zavaleta-Gatica, R., & Tiznado-Hernandez, M. (2007). Improving

p stha vest quality ang ‘Haden’ by UV-C treatment. Postharvest Biology and

Technology , 45, 108–116.

Goula, A. M., & Adamopoulus, K. G. (2008). Effect of maltodextrin addition during spray

drying of tomato pulp in dehumidified air: II. Powder properties. Drying Journal , 26, 726-

737.

100

Haque, M., & Roos, Y. (2005). Crystallization and X-ray diffraction of spray-dried and freeze-

dried amorphous lactose. Carbohydrate Research , 340, 293–301.

Harnkarnsujarit, N., & Charoenrein, S. (2011). Effect of water activity on sugar crystallization

and Beta-carotene stability of freeze-dried mango powder. Journal of Food Engineering ,

105, 592-598.

Hsu, C. L., Chen, W., Weng, Y. M., & Tseng, C. Y. (2003). Chemical composition, physical

properties, and antioxidant activities of yam flours as affected by different drying methods.

Food Chemistry , 83, 85-92.

Jakubczyk, E., Ostrowska-Ligeza, E., & Gonde, E. (2010). Moisture sorption characteristics and

glass transition temperature apple puree powder. International Journal of Food Science

and Technology , 45, 2515–2523.

Jaya, S., & Das, H. (2004). Effect of maltodextrin, glycerol monostearate and tricalcium

phosphate on vacuum dried mango powder properties. Journal of Food Engineering , 63,

125-134.

Jaya, S., Das, H., & Mani, S. (2006). Optimization of maltodextrin and tricalcium phosphate for

producing vacuum dried mango powder. International Journal of Food Properties , 9, 13--

24.

Jayasundera, M., Adhikari, B., Adhikari, R., & Aldred, P. (2011a). The effect of protein types

and low molecular weight surfactants on spray drying of sugar-rich foods. Food

Hydrocolloids , 25 (3), 459-469.

Jayasundera, M., Adhikari, B., Adhikari, R., & Aldred, P. (2011b). The effects of proteins and

low molecular weight surfactants on spray drying of model sugar-rich foods: Powder

production and characterisation. Journal of Food Engineering , 104 (2), 259-271.

101

Jayasundera, M., Adhikari, B., Howes, T., Aldred, P. 2011c. Surface protein coverage and its

implications on spray-drying of model sugar-rich foods: Solubility, powder production and

characterisation. Food Chemistry, 128 (4), 003-1016.

Kalogiannia, E. P., Xynogalos, V. A., Karapantsios, T. D., & Kostloglou, M. (2002). Effect of

feed concentration on the production of pregelatinized starch in a double drum dryer.

Elsevier Science Ltd.

Karel, M. (1975). Freeze dehydration of foods. In M. Karel, O. R. Fennema, & D. B. Lund,

Principles of Food Science (pp. 359-395). New York: Marcel Dekker, Inc.

Kha, T. C., Nguyen, M. H., & Roach, P. D. (2010). Effects of spray drying conditions on the

physicochemical and antioxidant properties of the Gac (Momordica cochinchinensis) fruit

aril powder. Journal of Food Engineering , 98, 385-392.

Kim, E. H.-J., Chen, X. D., & Pearce, D. (2009). Surface composition of industrial spray-dried

milk powders. 2. Effects of spray drying conditions on the surface composition. Journal of

Food Engineering , 94, 169-181.

Krokida, M. K., & Maroulis, Z. B. (1997). Effect of drying method on shrinkage and porosity.

Drying Technology , 15 (10), 2441-2458.

Krokida, M. K., Maroulis, Z. B., & Saravakos, G. D. (2001). The effect of the method of drying

on the color of dehydrated products. International Journal of Food Science and Technology

, 36, 53-59.

Masters, K. (1985). Spray drying handbook (4th ed.). London: Godwin.

Moore, J. G. (2005). Drum Dryers. In A. S. Mujumdar, Handbook of Industrial Drying (2nd ed.,

Vol. 1, pp. 249-262). New York: Marcel Dekker.

102

Mujumdar, A. S. (2007). Handbook of Industrial Drying (3rd ed.). Florida.: Taylor & Francis

Group, CRC Press.

Niamnuy, C., Nachaisin, M., Laohavanich, J., & Devahastin, S. (2011). Evaluation of bioactive

compounds and bioactivities of soybean dried by different methods and conditions. Food

Chemistry , 899-906.

Nijdam, J. J., & Langrish, T. A. (2005). An investigation of milk powders produced by a

laboratory-scale spray dryer. Drying Technology , 23 (5), 1049–1056.

Nindo, C. I., & Tang, J. (2007). Refractance Window dehydration technology: A novel contact

drying method. Drying Technology , 25, 37–48.

Nindo, C. I., Sun, T., Wang, S. W., Tang, J., & Powers, J. R. (2003). Evaluation of drying

technologies for retention of physical and chemical quality of green asparagus (Asparagus

officinalis, L.). Food Sci. Technol. (LWT) , 36 (5), 507–516.

Oetjen, G. W., & Haseley, P. (2004). Freeze-drying (2nd ed.). Germany: Wiley-VCH GmbH &

Coo. KGaA.

Potter, N. N., & Hotchkiss, J. H. (1995). Food Science (5th ed.). New York: Springer.

Rajkumar, P., Kailappan, R., Viswanathan, R., Raghavan, G. S., & Ratti, C. (2007). Foam mat

drying of alphonso mango pulp. Drying Technology , 25, 357-365.

Ratti, C. (2001). Hot air and freeze-drying of high-value foods-a review. Journal of Food

Engineering , 49, 311–319.

Rodríguez-Hernández, G.R., Gonzalez-Garcia, R., Grajales-Lagunes, A., Ruiz-Cabrera,M.A.,

2005. Spray-drying of cactus pear juice (Opuntia streptacantha): effect onthe

physicochemical properties of powder and reconstituted product. DryingTechnology, 4,

955–973.

103

Sablani, S. S., Shrestha, A. K., & Bhandari, B. R. (2008). A new method of producing date

powder granules: Physicochemical characteristics of powder. Journal of Food Engineering

, 87 (3), 416-421.

Saravacos, G. D., & Kostaropoulos, A. E. (2002). Hannbook of Food Processing Equipment.

New York: Kluwer Academic/Plenum Publishers.

Serrano, E. P. (2005). Postharvest management of fruit and vegetables in the Asia-Pacific region:

Philippines country report. Asian Productivity Organization (APO); Food and Agriculture

Organization of the United Nations (FAO) (pp. 216-226). Japan; Rome: APO and FAO.

Shrestha, A. K., Us-arak, T., Adhikari, B. R., Howes, T., & Bhandari, B. R. (2007). Glass

transition behavior of spray-dried orange juice powder measures by differential scanning

calorimetry (DSC) and thermal mechanical compression test (TMCT). International

Journal of Food Properties , 10, 661-673.

Syamaladevi, R. M., Sablani, S. S., Tang, J., Powers, J., & Swanson, B. G. (2009). State diagram

and water adsorption isotherm of raspberry (Rubus Idaeus). Journal of Food Engineering ,

460-467.

Toledo, R. (2007). Fundamentals of Food Process Engineering (3rd ed.). Athens, GA, USA:

Aspen Publishers, Inc.

Tonon, V. R., Brabet, C., & Hubinger, M. D. (2008). Influence of process conditions on the

physicochemical properties of acai (Euterpe oleraceae Mart.) powder produced by spray

drying. Journal of Food Engineering , 88, 411–418.

Wagner, L. A., & Warthesen, J. J. (1995). Stability of spray- dried encapsulated carrot carotenes.

Journal of Food Science , 60, 1048–1053.

104

Wang, S., Konkol, E., & Langrish, T. G. (2011). Spray drying of fruit juice using proteins as

additives. Drying Technology , 29 (16), 1868-1875.

Xiaoyan, Z. Y. (2008). Effect of superfine grinding on properties of ginger powder. Journal of

Food Engineering , 91 (2), 217-222.

105

Table 1. Drying conditions for production of mango powders using different methods.

Product Temperature Water content Product o Residence time, s C kg water/kg dry solids Fresh puree - - 6.518 ± 0.123 RW 74 ± 2 180 ± 0.15 0.017 ± 0.001 FD 20 ± 0.5 111,600 ± 509 0.023 ± 0.002 DD 105 ± 5 54 ± 0.2 0.013 ± 0.001 SD 90 ± 2 1-3 a 0.043 ± 0.003 Standard deviation from the average value of at least two replicates a The residence time was an approximate value, based on information given in Desobry et al., 1997 and Jayasundera et al. (201a,b,c)

Table 2. Hunter color measurements of reconstituted mango powders obtained from different drying processes.

Drying Method L* a* b* C* Hue Angle *b/*a ΔE

c a c a c Original Puree 45.12±0.02a 4.65±0.01 41.52±0.03 41.78±0.03 83.61±0.01 8.93±0.01 --

b d a b a b d RW 43.95±0.02 4.40±0.01 41.79±0.03 42.02±0.03 83.99±0.01 9.50±0.02 1.22±0.02

c b b d b d c FD 43.74±0.06 4.69±0.01 40.99±0.23 41.26±0.23 83.47±0.04 8.73±0.06 1.57±0.03

d a c e c e a DD 37.73±0.01 6.92±0.02 36.48±0.02 37.13±0.02 79.27±0.03 5.28±0.03 9.22±0.01

e e c a d a b SD 41.59±0.07 3.05±0.01 36.64±0.02 36.77±0.03 85.24±0.02 12.00±0.05 6.23±0.02

ΔE is calculated using the original mango puree as reference . a-e Means with the same superscript letters within a column indicate no significant difference (p ≤ 0.0 ).

106

Table 3. Solubility and hygroscopicity of RW-, freeze-, drum-, and spray-dried mango powders with particle size 180-250 m. Particle Moisture content Drying methods size kg water/kg mango SolubilityA HygroscopicityB m solids % %

RW 180-250 0.017 ± 0.001 90.79 ± (0.394)*a 18.0 ± 0.36a

FD 180-250 0.023 ± 0.002 89.70 ± 0.631a 18.0 ± 0.19a

DD 180-250 0.013 ± 0.001 94.38 ± 0.431b 20.1 ± 0.88b

SD 180-250 0.043 ± 0.003 95.31 ± 0.112b 16.5 ± 0.06c * Standard deviation from the average value a-b Means with the same superscript letters within a column indicate no significant differences (P<0.05). AMeasurement was done at 23 °C BSamples were exposed to 75±1% RH at 25°C for 7 days.

Table 4. Glass transition temperatures and water activity of RW-, freeze-, drum-, and spray-dried mango powders.

Drying Glass transition temperatures Water activity Methods Tgi Tgm Tgi RW 18.7 ± 0.2b 23.1 ± 0.7a 24.6 ± 0.5a 0.177 ± 0.001a FD 20.1 ± 0.8ab 25.8 ± 2.7a 27.6 ± 1.8ab 0.174 ± 0.001a DD 22.4 ± 1.1ab 27.6 ± 1.3a 30.3 ± 1.1ab 0.169 ± 0.002a SD 24.4 ± 0.8a 28.8 ± 4.3a 31.5 ± 0.7a 0.173 ± 0.006a

Means with the same superscript letters within a column indicate no significant differences (P<0.05). Tgi, Tgm and Tge represent the onset, mid- and end-point glass transition temperatures of mango powders (180-250 μm).

107

Ripening - 95-100% Washing of mango (chlorinated & regular water) Trimming / removal of black part of peel Peeling/pulping Blending Pasteurizing Packaging- PE bags Blast Freezing at -35oC Packing in carton boxes Cold Storage at -18oC

Philippine ang ‘ a aba va .’

Figure 22. Mango puree samples

Figure 23. Schematic layout of Refractance Window® dryer (Adapted from Nindo et al, 2007, Abonyi et al., 2002)

108

.Figure 24. Measurements of product temperature of mango puree

Figure 25. Monitoring of hot water temperature

109

Figure 26. Refractance Window drying of mango puree

110

Figure 27. Freeze drying of mango puree

Figure 28. Drum drying of mango puree

111

A B C

Figure 29. Spray drying of mango puree. A, B & C are the spray-dried mango powder with added carrier of 0.25, 0.35 and 0.45 kg maltodextrin / kg mango solids.

112

Photograph of mango flakes or powders at different particle sizes obtained from Refractance Window sizes Refractance particle powders or from obtained of different at Photograph flakes mango

Figure Figure (SD). spray drying (DD), drying and freeze drying drum (FD), drying, (RW)

113

L* Value L* 60

40 0 Flakes 1 500 2 350 3 250 4 180 5

Particle Size, m

RW FD DD SD (a) 60

40 b* Value b*

20 Flakes0 1 500 2 350 3 250 4 180 5 Particle Size, m

RW FD DD SD (b)

Figure 31. Lightness (a), yellowness (b) of mango flakes or powders at different particle sizes obtained from Refractance Window® (RW) drying, freeze drying (FD), drum drying (DD), and spray drying (SD).

114

40 Chroma

20 0 Flakes 1 500 2 350 3 250 4 180 5

Particle Size, m

RW FD DD SD (a) 90

Hue Angle Hue 70

60 0 1 2 3 4 5 Flakes 500 350 250 180

Particle Size, m

RW FD DD SD (b)

Figure 32. Chroma (a) and hue angle (b) of mango flakes or powders at different particle sizes obtained from Refractance Window® (RW) drying, freeze drying (FD), drum drying (DD), and spray drying (SD).

115

(RW) drying, freeze freeze (RW)drying,

Photograph of reconstituted mango powders obtained from Refractance Refractance powders Window obtained mango from reconstituted of Photograph

Figure Figure

spray drying and drying drum (FD), (SD). drying (DD),

116

1000

800

600

400 Bulk Density,k g/m3 Density,k Bulk 200

0 RW FD DD SD Drying method

μ500 μ350 μ250 μ180

Figure 34. Bulk density of mango powders obtained from Refractance Window® (RW) drying, freeze drying (FD), drum drying (DD), and spray drying (SD).

117

0.60

0.50

0.40

0.30 Porosity 0.20

0.10

0.00 RW FD DD SD Drying method

500μ 350μ 250μ μ180

Figure 35. Porosity of mango powders obtained from Refractance Window® (RW) drying, freeze drying (FD), drum drying (DD), and spray drying (SD).

118

250

200

150

100

d Intensity (Counts) Intensity 50 a b

c 0 0 10 20 30 40 50 2 Theta (Deg)

Figure 36. X-ray diffraction patterns of Refractance Window®-dried (a), freeze-dried (b), drum- dried (c) and spray-dried (d) mango powders with particle size 180-2 0 μ and aw < 0.2.

119

h), 20 kV. kV. 20 h),

drying (a & b), & (a drying

2 0 μ ) d ied using Re actance Windw actance Re using ied )d μ 0 2

Scanning electron micrographs (SEM) of mango powders (180 powders mango of (SEM) micrographs electron Scanning

freeze drying (c & d), drum drying (e & f) and spray drying (g & h) (magnification of 100x (a, c, e & g) and 300x (b, d, f & f d, (b, 300x and &g) e c, (a, of 100x (magnification &h) (g drying spray and f) & (e drying drum &d), (c drying freeze Figure Figure

120

Figure 38. Scanning electron micrographs (SEM) of individual mango powder particles (180-250 μ ) d ied using Re actance Wind w® drying (a), freeze drying (b), drum drying (c) and spray drying (d) (magnification of 1,000 x, 20 kV).

121

(a) (b)

Figure 39. Field emission scanning electron micrographs (FESEM) of Refractance Window®- dried (a), freeze-dried (b), drum-dried (c) and spray-dried (d) mango powders (180-250 mm) stored for 7 days at 25 oC with RH=75.5 % (magnification of 300 x, 30 kV).

122

CHAPTER FOUR

WATER SORPTION CHARACTERISTICS, GLASS TRANSITION TEMPERATURES

AND MICROSTRUCTURES OF REFRACTANCE WINDOW®- AND FREEZE-DRIED

MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER

ABSTRACT

Water sorption isotherms, glass transition, and microstructures of Refractance Window

(RW)- and freeze-dried mango powders obtained from Philippine 'Carabao' mango variety were investigated to aid in choosing suitable packaging material and environment conditions during storage. Water adsorption isotherms were established by the isopeistic method. Thermal transition of the powders was determined at various water activities (aw = 0.11-0.86) using differential scanning calorimetry (DSC). The sorption isotherms of RW- and freeze-dried (FD) mango powders exhibited a sigmoidal characteristic of type III showing higher and lower adsorption capacities above and below 0.5 aw, respectively. A significant difference (p < 0.05) in water content of RW- and freeze-dried mango powders for equivalent water activities were obtained above 0.5 aw. The onset glass transition temperature (Tgi) of RW- and freeze-dried mango powders decreased as the water content increased. There were no significant differences in Tgi of RW- and freeze-dried mango powders at constant water activities except for aw = 0.86

(p ≥ 0.05). Microscopic examination of mango powders indicated that freeze-dried mango powders exhibited greater surface area and porosity in comparison to RW-dried mango powders.

123

1. IINTRODUCTION

Mango (Mangifera indica L.) is one of the most important fruit trees in tropical countries.

In the Philippines, mango ranks third among fruit crops produced, after banana and pineapple, based on export volume and value, with a total of 1,023,907 metric tons harvested in 2007, according to the Philippine Bureau of Statistics (BAS, 2009). The Carabao variety popularly kn wn as “Philippine Supe Mang ” is the d inant va iety that acc unts 73 % the country's production (BAS, 2009). Although many food companies produce various products from mango such as puree, juice and concentrate, there is a continuing effort to process it into powder form to make it more stable. Mango powders can be utilized for innovative formulations, thus creating opportunities for a wide range of applications such as in dry beverage mixes, health drinks, baby foods, sauces, marinades, confections, yogurt, ice cream, nutrition bars, baked goods and cereals (Rajkumar et al., 2007). However, production of mango powder embraces many research challenges during drying because of its inherent sticky characteristics attributed to its low molecular weight sugars such as sucrose, fructose and glucose and organic acids

(Bhandari et al., 1997).

Water activity is a widely accepted and utilized concept parameter to predict the stability of food products over time. It influences microbial growth, lipid oxidation, non-enzymatic and enzymatic activities (Rahman & Labuza, 1999). Expressing the relation between water activity and equilibrium water content in a graphical form at a fixed temperature produces a water sorption isotherm (Rizvi, 1995). Sorption isotherms data provide relevant information on the physical, chemical and microbiological stability of foods, and these are used as an important parameter for drying, packaging and storage of certain products (Kumar, 2006).

124

Monolayer water content is a useful value that relates to the amount of bound water in a solid material. It normally varies between 0.04 – 0.11 kg H2O/kg dry solids for several dried foods (Karel, 1975). This low water content does not support microbial reactions, and hence contributes to food product stability (Rahman & Labuza, 1999). Microstructure and morphology of materials are altered in the drying process. Investigation on the microstructures of food powders may provide some useful insights into the sorption isotherm behavior of a specific product.

Many authors have discussed limitations in applying water activity for predicting food stability: 1) not all food products are in an equilibrium state, and measurement of water activity at a specific time might not describe the steady state conditions of these products; 2) critical limits of water activity might provide inaccurate values as affected by pH, salt, anti-microbial agent and pre-treatments; and 3) water activity cannot safely tell whether the water present in food is bound or unbound water (Rahman, 2006; Rahman & Labuza, 1999; Chirife, 1994). In consideration of these limitations, a glass transition temperature concept was applied into the food system. It is hypothesized that glass transition temperature (Tg) can greatly affect the stability of food, as the water in its solid-like state becomes immobilized and cannot support its own weight against flow due to gravity (Rahman & Labuza, 1999; Slade & Levine, 1991). Both the water activity and glass transition concepts have been proposed to predict deterioration, stability and shelf-life of food because, in many instances, glass transition alone does not work

(Roos, 1995). Glass transition temperature as a function of water activity or water content can be used to construct a state diagram for a particular food system (Sablani et al., 2010). Previous studies have shown strong evidence that using the state diagram can better assist the food industry in determining the stability of their products (Sablani et al., 2010) (Table 2). Several

125

studies related to the interactions of water activity, water content and glass transition temperature of agricultural and fishery products have been reported (Syamaladevi et al., 2009; Sablani et al.,

2007).

Glass transition temperature greatly influences the stability of food since below this temperature water is kinetically immobilized, restricting it to participate in the reactions

(Rahman, 1995). Depending on the temperature and rate of water content removal during drying, powder production of sugar-rich fruit may cause degradation during drying and subsequent storage. However, there is a little or no information about sorption isotherms, glass transition and state diagrams of freeze-dried and Refractance Window-dried mango powder. The objective of this work was to investigate the influence of the sorption isotherms, glass transition and microstructures of mango powder dried using Refractance Window and freeze drying methods.

2. MATERIIALS AND METHODS

2.1 Preparation of mango powder and packaging

Frozen mango puree ('Carabao' var.), a dominant mango variety in the Philippines, was acquired from Ramar Foods International (Pittsburg, CA). The puree was thawed overnight at

~22 °C and blended to a uniform consistency for 5 minutes using a bench top blender (Oster

Osterizer, Mexico) with lowest speed setting. The puree with initial water content of 6.52 ± 0.12 kg water/kg dry solids was dried to below 0.03 kg water/kg dry solids using Refractance

Window® drying or freeze drying methods. A pilot scale Refractance Window® dryer developed by MCD Technologies, Inc. (Tacoma, WA) was used in the experiment (Caparino et al., 2012).

The dryer has the following components: a water pump, a hot water tank, a heating unit, a two wate lu es, a h d with sucti n bl we s and e haust ans, a c nvey belt ade “Myla ®”

(polyethylene terephthalate) plastic, a spreader at the inlet section , and a scraper at the end

126

section of the dryer. The dryer has an effective surface drying area of 1.10 m2. During drying operation, circulating hot water between 95 – 97 °C was maintained to continuously transfer thermal energy to the puree through the plastic conveyor interface. An average air velocity of 0.7 m/s with a relative humidity ranging from 50 – 52 % was applied on the surface of the puree to facilitate moisture removal (Abonyi et al., 2002; Nindo &Tang, 2007; Caparino et al., 2012).

Freeze drying method was carried out using a laboratory freeze dryer (Freeze Mobile 24, Virtis

Company, Inc., Gardiner, NY). The thawed mango puree was poured into a stainless pan to form a layer of 15 mm. The frozen samples were placed at – 25 °C for 24 hours before being transferred to the freeze dryer. The vacuum pressure of the dryer was set at 20 Pa, the plate temperature was 20 °C, and the condenser was at – 60 °C (Caparino et al., 2012).

The produced RW-and freeze-dried mango flakes or sheets were collected and packed in leak-proof Ziploc® plastic bags and double packed in aluminum-coated polyethylene bags. All packaged samples were flushed with nitrogen gas to prevent oxidation, heat sealed and stored at

–35 °C until further analyses. One hundred grams of dried mango flakes or sheets obtained from

RW and freeze- drying processes were ground using a mortar and pestle and were sieved using sizes 60 and 80 (American Society for Testing and Materials, ASTM) to obtain particle sizes of

180 − 2 0 (International Standard for Organization, ISO) (Barbosa-Cánovas 2005). This range of particle size was selected for better interpretation of the microstructures of mango powders and for convenience during DSC loading. The prepared samples were used for water sorption, thermal and microstructures experiments.

2.2. Measurements of residence time, product temperature and water content

For Refractance Window drying, the residence time to dry the mango puree to below 0.03 kg water/kg dry solids was determined by monitoring the time travelled by the thinly spread

127

mango puree from the inlet to the outlet section of the plastic conveyor belt. The product temperature was obtained by scraping about 3 grams of the samples along different location of the drying section of RW dryer and measuring them by temperature infrared sensor (Raytek MT6

Mini, CA). In the freeze-drying method, the residence time required to bring down the initial water content to a similar water content level as RW drying was determined when the vacuum pressure of the freeze dryer had dropped to 30 mTorr (4 Pa). The product temperature was measured using pre-calibrated Type T thermocouple sensors, which were connected to a data acquisition device equipped with monitoring software. The water content of mango puree, RW- and freeze-dried mango flakes or powders were determined using the standard oven method at 70

°C and 13.3 kPa for 24 hours (AOAC, 1998). All measurements were performed at least in duplicate.

2.3. Determination of sorption isotherms

An adsorption isotherm was developed using the isopiestic method according to Speiss &

Wolf (1987), Lim et al. (1995), Sablani et al. (2007) and Syamaladevi et al. (2009). RW- and freeze-dried mango powders were placed in airtight humidity jars and equilibrated at room temperature (~23 °C) for 35 days using saturated salt solutions with constant water activity (Fig.

1). The saturated salt solutions used were LiCl, CH3COOK, MgCl, K2CO3, MgNO3, NaNO2,

NaCl, and KCl (Fisher Scientific, Houston, TX), with corresponding known relative humidity of

11.3 %, 22.5 %, 32.8 %, 43.2 %, 52.9 %, 65.8 %, 75 % and 86 % at 23 °C, respectively. To prevent microbial growth in the samples, a small amount of thymol was added in a small uncapped bottle and placed together with the samples inside the airtight humidity jars. The water content of the equilibrated RW- and freeze-dried mango powders were determined using the standard oven method at 13.3 kPa and 70 °C and for 24 h (AOAC, 1998).

128

Several water sorption isotherm model equations are applied to determine monolayer water content in foods such as BET equation (Brunauer et al., 1938), Henderson equation

(Henderson, 1952), Smith equation (Rockland & Stewart, 1981) and GAB equations (Labuza,

1968; Bizot, 1983). For the sorption modeling of RW- and freeze-dried mango powder, we used the Brunauer - Emmetters - Teller (BET) and Guggenheim (Bizot, 1983). The BET isotherm is applicable between water activities of 0.05 and 0.45, while GAB is applicable for a wide range of water activities between 0 and 0.95 (Flade & Aworth, 2004; Labuza & Altunakar, 2007;

Rahman, 1995). The BET equation is expressed as (Labuza, 1968):

(1)

where Mw is the water content (kg water/kg dry solids); Mb is the BET monolayer water content

(dry basis); B is a constant related to net heat of sorption.

The GAB equation is expressed as (Labuza, 1968):

(2)

where Mw is the water content (kg water/kg dry solids); Mg is the GAB monolayer water content

(dry basis); C is a constant related to the monolayer heat of sorption; and K is a factor related to the heat of sorption of the multilayer and the value of K varies from 0.7 to 1.

Estimation and optimization of parameters in BET and GAB equations was done using

Windows Excel® at water content and water activity values. At least duplicate for all samples were measured and analyzed.

2.4. Thermal glass transition

The thermal glass transition analysis for RW- and freeze-dried mango powders with water content ranging from 0.074-0.097 kg solid/kg mango powder) was carried out using a

129

differential scanning calorimeter (DSC, Q2000, TA Instruments, New Castle, USA) (Fig. 2) following the procedure described in Syamaladevi et al. (2009). The calorimeter was calibrated for heat flow and temperature using standard indium and sapphire. An empty aluminum pan was used as a reference for each sample test. Ten to twelve millig a s the equilib ated ang p wde s we e sealed in an alu inu pan (v lu e 30 l), and c led 2 d wn t

90 °C using liquid nitrogen and equilibrated for 10 min. The equilibrated samples were scanned to 70 °C and then cooled down to 25 °C. Scanning of all samples was carried out using the same heating or cooling rate of 5 °C/min (Syamaladevi et al., 2009). To avoid condensation on the surface of the powder particles, a nitrogen carrier gas was purged at a flow rate of 50 ml/min.

The state diagram was determined by DSC thermograms using Universal Analysis 2000 software

(TA Instruments, Newcastle, USA). Glass transition temperature (Tg) of the mango powders was determined by finding the vertical shift in the heat flow-temperature diagram. Glass transition temperatures of mango powder at different water activity levels and equilibrium water content were measured at least in duplicate.

Plasticization behavior of mango constituents by water was predicted using the Gordon and Taylor (1952) equation, expressed as:

(3)

where Xw, Xs Tgm, Tgs and Tgw are the mass fraction of water, total solids (wet basis), glass transition temperatures of the mixture, solids and water, respectively. The k value represents the

Gordon–Taylor constant parameter. The glass transition curve was established by extrapolating the Tg values of the samples using the glass transition temperature of water (Tg −13 ) as the lower end temperature limit. The glass transition temperature of samples at zero water (Tgs) and k values were calculated by applying non-linear regression (Sablani et al., 2007).

130

2.5. Microstructures of mango powders

A s all quantity ee e- and RW-d ied ang p wde s (180 2 0 mm) were mounted on aluminum stubs and coated with fine layer of gold (15 nm) using a Sputter gold coater

(Technics Hummer V, Anatech, San José, CA). The samples were examined by scanning electron microscopy using an SEM Hitachi S‐570 camera (Hitachi Ltd., Tokyo, Japan) operated at an accelerating voltage of 20 KV. Micrographs were photographed at a magnification of 300 x and 1,000 x at a scale of 100 mm and 30 mm.

2.6. Statistical analysis

All experiments were carried out at least in duplicate, and the results were analyzed using the general linear model procedure of SAS (SAS Institute Inc., Cary, NC), and the means separated by Tukey-honest significant difference test with a confidence interval of 95% was used to compare the means.

3. RRESULTS AND DISCUSSION

3.1. Drying time, water content and product temperature

Drying of mango puree from the initial water content of 6.52 ± 0.123kg water/kg mango solids to below 0.05 kg water/kg mango solids was accomplished in 3 ± 0.01 and 1,860 ± 85 min, using RW and freeze drying (Table 1). The product temperatures measured for each drying process were 74 ± 2 °C and 20 ± 0.5 °C for RW and freeze-drying, respectively.

3.2. Water sorption isotherms

The equilibrium condition of the mango powders were achieved after 35 days of storage at 23 °C and at different constant relative humidities or water activities. As shown in Fig. 3, it is evident that water activity and equilibrium water content of the product has a direct relationship,

131

i.e., as water content increases there is a corresponding increase of water activity. Both the RW- and freeze-dried plotted curves showed a convex formation compared to the equilibrium water content axis, or a sigmoidal type III (J-shape), based on the description by Brunauer, Emmett, &

Teller (1938). These observations were similar to sorption curves observed for other products with high sugar contents and amorphous dried materials such as osmo-dried star apple and mango (Falade & Worth, 2004), freeze-dried blue berries (Lim et al., 1995), air-dried grapefruits

(Fabra et al., 2009), apple puree powders (Jakubczyk et al., 2010), and freeze-dried mango pulp

(Rangel–Marrón et al., 2011). RW- and freeze-dried mango powders adsorb small amounts of water at low water activity (< 0.5) due to possible local dissolution of sugars and formation of new active sites (Falade & Aworth, 2004). On the other hand, at higher water activity (aw > 0.5) showed a sharp increase due to gradual dissolution and complete exudation of sugar present in mango as a result of its crystalline structure breakdown and appearance of more active sites

(Rangel–Marrón et al., 2011; Falade, 2004).

A significant difference (p < 0.05) in water content of RW-dried and freeze-dried mango powders for equivalent water activities were obtained above 0.5 aw (Fig. 3). That is, water adsorption capacities of freeze-dried mango powder were higher than RW-dried mango powder at water activity above 0.50, as shown by the significant increase in water content in freeze-dried mango powders. This result might be attributed to the highly porous and hygroscopic nature of freeze-dried mango powder compared to RW-dried powder. The higher product temperature for

RW-drying (74 ± 2 °C) as against freeze-drying (20 ± 0.5°C) might also be responsible for the greater adsorption capacity above 0.5 aw. Crossing of isotherms was observed for freeze-dried, osmo-freeze-dried and osmo-air-dries cherries and blueberries (Maroulis et al., 1988; Yu et al.,

1999), and osmo-oven dried mango (Falade & Worth (2004) when tested at different drying

132

temperatures. These authors observed that the sorbed water at a given water activity increases as temperature increases, caused by the increase in solubility of sugars at higher temperature.

The water activity-water content data of RW- and freeze-dried mango powders were experimentally fitted to BET and GAB models with a high coefficient of determination (R2) values (0. 72 0.986) (Fig. 3). The monolayer (monomolecular) water content is considered as the lower practical limit for most drying applications below which drying processes became inefficient. This limit is also widely accepted as the safest water content for longer storage and for food stability (Gabas et al., 2007). Similar BET monolayer water contents (Mb) were obtained for RW- (0.081 kg H2O/kg dry solids) and freeze-dried (0.087 kg H2O/kg dry solids) mango powders by nonlinear regression analyses. The obtained GAB monolayer water content (Mg) of

RW- and freeze-dried mango powder was 0.078 kg H2O/kg dry solids and 0.045 kg H2O/kg dry solids, respectively (Table 4). The smaller GAB monolayer value of freeze-dried mango powder in comparison to RW mango powder may be attributed to the overlapping of predicted water content values below 0.5 aw and the significant difference in water content values above 0.5 aw

(Fig. 3). The obtained BET and GAB monolayer values in the present experiments are within the range of monolayer values for several dehydrated fruit products between 0.026–0.185 kg H2O/kg dry solids (Table 2) (Falade & Worth, 2004; Talla, 2005; Syamaladevi et al., 2009; Jakubczyk,

2010, Rangel–Marrón et al., 2011). The variations in monolayer water content of sugar-rich materials could be attributed to the different drying processes applied and sugar composition of these fruits.

3.3. Glass transition temperature

The onset, mid and end glass transition temperatures (Tgi, Tgm and Tge) were recorded in the experiment, since there is no consensus definition on either one of these temperatures being

133

the transition point in a DSC curve (Rahman, 2006). However, some authors used Tgm to describe the glass transition of certain products (Roos & Karel, 1991a). On the premise that glass transition temperature starts at Tgi, we considered it as the safest Tg for storage of mango powder, and hence used it to establish the relationship between Tg, water activity and water content. The

Tgi of both the RW- and freeze-dried mango powders obtained by a single scan DSC analysis leaned toward a lower temperature as the water activity increased (Fig. 4). The inverse relationship of glass transition and water activity in this particular study is attributed to the strong plasticizing effect of water (Tg 135 °C) on amorphous components of the food matrix (Roos

& Karel, 1991a; Bhandari, 1999). The thermograms of RW- and freeze-dried mango powders conditioned at water activities (aw 0.11 0.86) showed one transition wherein no crystalline peak was observed in the DSC thermogram. Similar thermogram behavior was reported for dried fruits containing high sugars, such as raspberry (Syamaladevi, 2009), pineapple (Telis & Sobral,

2000), and strawberries (Roos, 1987). No significant differences in the initial glass transition temperatures (Tgi) of RW-dried and freeze-dried mango powders were observed for equivalent water activities, except for aw = 0.86 (Table 3) (< 0.05). The Tgm values RW and ee e-d ied

ang p wde s dec eased 3 . 6 d wn t 67.01 °C and 39.73 °C down to -70.16 °C as the water content increased from 0.027 ± 001 to 0.343 ± 0.005 kg water/kg dry solids and 0.029

± 0.005 to 0.351 ± 0.044 kg water/kg dry solids, respectively (Table 3). The lowest initial glass transition temperatures were (Tgi) easu ed RW-d ied ( 65.3 °C) and freeze-dried ( 72.2

°C) mango powder with water activity of 0.86, due to the plasticization by the large amount of water. The highest initial glass transition temperature was observed for RW-dried (30.6 °C) and

FD-dried (33.4 °C ) mango powder with water activity of 0.113, due to greater dry solid concentration (Sablani & Syamaladevi, 2010; Teles, 2011). The glass transition curve was fitted

134

with the Gordon-Taylor (G-T) equation, and was extrapolated to the glass transition temperature of water (Tg −13 ) (Fig. 5). The glass transition temperature of mango powders at zero water (Tgs) and constant, k values were calculated by applying non-linear regression, as described in Sablani et al. (2007). The glass transition temperature for RW- and freeze-dried mango powders at zero water (Tgs) when fitted to Gordon-Taylor model was 55.8 and 63.6 °C, respectively (Fig. 5). These values were close to the glass transition temperature of vacuum-dried mango (unspecified variety) of 62 °C, as reported by Jaya and Das (2009). Apparently, the low

Tgs RW- and ee e-d ied ang p wde a e due ainly t the high c ncent ati n suc se p esent in ang . Mang is ich in suga s, ainly suc se (0.060 0.0 kg kg pulp),

uct se (0.02 0.027 kg/kg of pulp) and glucose (0.007 0.047 kg /kg of pulp) (Jaya & Das

2009). The glass transition temperatures of dry sucrose, fructose and glucose were reported to be

62 °C, 5 °C and 32 °C, respectively (Roos & Karel, 1991a,b). The higher glass transition temperature of freeze-dried mango powder at zero water compared to RW-dried mango powder could be attributed to the prediction of two parameters using glass transition temperature and water content values by non-linear optimization technique.

Sorption isotherms and state diagrams of RW and freeze-dried mango powders were combined to examine their relationships in terms of product stability of the product during storage (Fig. 6 & 7 and Table 4). It can be seen from the data in Fig. 5 that the relationships among glass transition temperature, water activity and water content for RW and freeze-dried mango powder showed certain degrees of variations. The stable temperature range predicted by the glass transition model based on the sorption isotherms underestimate the stable temperature range. For example, in the case of RW-dried mango powder, the critical glass transition temperature based on the sorption isotherm at 23 oC was smaller than the stable temperature

135

range. The RW-dried mango powders are stable at a BET monolayer water content of 0.081 (Fig.

6). However, by taking the same water content, the Tg value as presented in the glass line (dotted arrow lines) was predicted at lower storage temperature (0 °C). The Tg-water activity relationship presented in Table 4 suggests that RW-dried mango powder at water content of 0.983 kg solids/kg mango pulp is stable when stored at temperatures not higher than 23 °C. The water activity of RW-dried mango powder obtained in sorption isotherms (aw = 0.16) as predicted at 23

°C is lower than the monolayer water activity (aw = 0.34). This result indicates that mango powder with aw ≤ 0.34 is sa e t st e at 23 l we . Othe auth s w king n suga -rich materials observed similar observations (Telis and Sobral, 2000; Roos, 1987; Syamaladevi,

2009; Sablani et al., 2007).

On the other hand, the critical glass transition temperature for freeze-dried mango powder appeared to be similar to the stable temperature at its mono layer water content (Fig. 7). It was observed that freeze-dried mango powder is stable at 0.087 kg water/kg dry solids or 0.913 kg solids/kg mango puree when stored at 23 °C. Apparently, when using the same lower limit water content, the glass transition temperature did not incur significant change, as shown in the glass line or dotted lines (23 °C). Analyzing the water content and glass transition temperature shown in Table 4, it clearly shows that freeze-dried mango powder with a solid concentration of 0.977 kg solids/kg mango puree is stable at 23 °C. At the same temperature, sorption isotherm predicts a water activity of 0.18 slightly lower than the monolayer water activity (aw = 0.33) for safe storage of freeze-dried mango powder. In the current study, RW-dried and freeze-dried mango powders exhibited similar sorption and glass transition characteristics as other high sugar materials. Sablani et al. (2007) reported that the glass transition concept often underestimates the safe temperature for dehydrated fruits with sugar content. Further studies on physicochemical

136

changes such as deg adati n in bi active c p unds (β-carotene and vitamin C) in dehydrated mango stored at the selected water contents/activities may explain the appropriateness of water activity or glass transition temperature concepts.

3.4. Microstructures

The microstructures of mango powders depend on the drying method applied. RW- dried mango powder was in the form of smooth flakes with uniform thickness (Fig. 8a,b). The uniformity of the flake thickness was the result of a controlled feeding of mango puree using a spreader bar at the inlet section of the RW dryer. During the dying, the thinly spread mango puree on the surface of the plastic film conveyor was undisturbed, except for removal of water, as it moved toward the other end of the dryer, hence producing a continuous sheet with thickness nearly equal. Breaking the RW-dried mango flakes into powder form produced irregular shape particles while maintaining the thickness. The two sides of a single particle were smooth, indicating that it was more flowable and less susceptible to oxidation because of lesser surface area.

The microstructures of freeze-dried mango powder had a skeletal-like structure with void spaces previously occupied by ice prior to freeze drying (Fig. 8c,d). This is because the absence of liquid phase in the material during freeze drying process suppressed the transfer of liquid water to the surface and the ice was converted to vapor without passing the liquid state (Krokida et al., 1997). In effect, the collapse and shrinkage of the product is prevented, thereby resulting in a porous dried material (Karel, 1975).

Collapse and shrinkage phenomena have been proposed to have relationships to glass transition temperature (Krokida et al., 1998). Achanta & Okos (1996) hypothesized that shrinkage can be observed only when the drying temperature applied is above the glass transition

137

temperature of the material at a given water content. Also, the concept of glass transition as related to pore formation has been explained by Rahman (2001). According to that study, more pores or negligible collapse can be observed when a material is dried below Tg while fewer pores can be observed when processed at T > Tg. The onset glass transition temperature of RW- and freeze-dried mango powder was 30.61 °C ± 1.5 and (33.35 °C ± 3.4), respectively (Table 3). It can be seen from Figure 6 that freeze-dried mango powder obtained from freeze drying at plate temperature of 20 °C ± 1 (T < Tg) formed large pores, while the application of higher temperature of 74 ± 2 °C during RW- drying (T > Tg ) resulted in a rigid product with lower porosity. Hence, our observations agreed on the above hypothesis. Other studies have revealed that products obtained by freeze drying at T < Tg were in the glassy state, allowing for negligible shrinkage and hence very porous, while in comparison with the product processed by hot air drying (T > Tg) they were in rubbery state resulting in collapse and shrinkage (Sablani &

Rahman, 2007; Krokida et al., 1997).

Individual particles of mango powders obtained from RW and freeze drying process were further examined at higher magnification. The RW-dried mango powder clearly showed a composite sheet with distinguishable internal pores along the cross section of each particle, indicating that some emptied space during evaporation was not replaced as the mango puree was dried (Fig. 6b,c). On the other hand, visual observation of freeze-dried mango powder exhibited external pores within the internal pores of a single particle (Fig. 6e,f). This further explains why the porosity of freeze-dried materials was always higher in comparison with other drying methods.

138

4. CONCLUSIONS

The sorption isotherms for RW- and freeze-dried mango powders showed a sigmoidal characteristics type III (J-shape) as experimentally fitted to both GAB and BET models. This result might be attributed to the highly porous and hygroscopic nature of freeze-dried mango powder compared to RW-dried powder. The obtained GAB monolayer water content (Mg) of

RW- and freeze-dried mango powder was 0.078 kg H2O/kg dry solids and 0.045 kg H2O/kg dry solids, respectively, falls within the range of (Mg) value for several foods between 0.029–0.11 kg

H2O/kg dry solids. Both the Tgi of RW- and freeze-dried mango powders decrease as the water activity increases. There were no significant differences in initial glass transition temperatures

(Tgi) of RW-dried and freeze-dried mango powders at constant water activities, except for aw =

0.86. The glass transition temperature for RW- and freeze-dried mango powders at zero water

(Tgs), when fitted to the Gordon-Taylor model, was 55.82 and 63.61 °C, respectively.

Microscopic examination of mango powders showed that RW-dried mango powder was smooth and flaky with uniform thickness, while freeze-dried mango powder showed a skeletal-like structure with large pores. The generated results provide valuable information in predicting the stability of RW- and freeze-dried mango powders.

ACKNOWLEDGMENTS

We thank the Ford Foundation International Fellowship Program (IFP)-Philippines and

IFP/Institute of International Education-New York for providing the financial support; the

Philippine Center for Postharvest Development and Mechanization (PhilMech) for granting study leave. Special thanks to Richard E. Magoon and Karin M. Bolland of MCD Technologies,

Inc. (Tacoma, WA) for allowing the use of their RW drying facilities and their assistance in conducting the experiments; Engr. Frank Younce for technical assistance on the operation of the

139

freeze dryer; and Dr. Valerie Lynch-Holm of the WSU-Electron Microscopy Center for helping with SEM imaging.

140

REFERENCES

Abonyi, B. I., Feng, B. I., Edwards, C. G., & Tang, J. (2002). Quality retention in strawberry and

carrot purees dried with Refractance Window system. Journal of Food Science , 67,

1051–1056.

AOAC. (1998). Official methods of analysis. Washington, DC, USA: Association of Official

Analytical Chemists.

Atchanta, S., & Okos, M. R. (1996). Predicting the quality of dehydrated foods and biopolymers-

-Research needs and opportunities. Drying Technology , 14 (6), 1329-1368.

Barbosa-Canovas, G., V., Ortega-Rivas, E., Juliano, P., & Yan, H. (2005). Food Powders:

Physical Properties, Processing, and Functionality. New York: Kluwer

Academic/Plenum Publishers.

BAS. (2009). Situation Report on Selected Fruit Crops. Retrieved March 1, 2011, from

http://www.bas.bas.gov

Bhandari, B. R. (1999). Implication of glass transition for the drying and stability of dried

products. Journal of Food Engineering , 40 (1-2), 71-79.

Bhandari, B. R., Datta, N., & Howes, T. (1997). Problems associated with spray drying of sugar-

rich foods. Drying Technology , 15 (2), 671-684.

Bizot, H. (1983). Using the GAB model to construct sorption isotherms. In R. Jowitt, F. Escher,

B. M. Hallstrom, W. E. Spiess, & G. Vos, Physical properties of foods (pp. 42-54).

London: Applied Science.

Brunauer, S., Emmett, P. H., & Teller, E. (1938). Adsorption of gases in multimolecular layers.

Journal of American Chemists' Society , 62, 309-319.

141

Caparino, O. A., Tang, J., Nindo, C. I., Sablani, S. S., Powers, J. R., & Fellman, J. K. (2012).

Effect of drying methods on the physical properties and microstructures of mango

(Philippine carabao var.) powder. Journal of Food Engineering , 00 (0), 00 - 00.

Chirife, J., & Buera, M. P. (1994). Water activity, glass transition, and microbial stability in

concentrated semi-moist food systems. Journal of Food Science , 49, 1604-1608.

Fabra, M. J., Talens, P., Moraga, P., & Martinez-Navarrete, N. (2009). Sorption isotherm and

state diagram of grapefruit as a tool to improve product processing and stability. Journal

of Food Engineering , 93, 52-58.

Falade, K. O., & Aworth, O. C. (2004). Adsorption isotherms of osmo-oven dried african star

apple (Chrysophyllum albidum) and african mango (Irvingia gabonensis) slices. Eur

Food Res Technology , 218, 278-283.

Gabas, A. L., Telis, V. R., Sobral, P. J., & Telis-Romero, J. (2007). Effect of Maltodextrin and

arabic gum in water vapor sorption thermodynamic properties of vacuum-dried pieapple

pulp powder. Journal of Food Engineering , 82, 246-252.

Henderson, S. (1952). A basic concept of equilibrium mositure. Agricultural Engineering , 33,

29-32.

Jakubczyk, E., Ostrowska-Ligeza, E., & Gonde, E. (2010). Moisture sorption characteristics and

glass transition temperature apple puree powder. International Journal of Food Science

and Technology , 45, 2515–2523.

Jaya, S., & Das, H. (2009). Glass transition and sticky point temperatures and stability/mobility

doagram of fruit powders. Food Bioprocess Technol , 2, 89-95.

142

Jayaramayan, K. S., & Das Gupta, D. K. (2006). Drying of Fruits and Vegetables. In A. S.

MUJUMDAR, handbook of Industrial Drying (3rd ed.). New York: Taylor and Francis

Group.

Karel, M. (1975). Freeze dehydration of foods. In M. Karel, O. R. Fennema, & D. B. Lund,

Principles of Food Science (pp. 359-395). New York: Marcel Dekker, Inc.

Krokida, M. K., & Maroulis, Z. B. (1997). Effect of drying method on shrinkage and porosity.

Drying Technology , 15 (10), 2441-2458.

Krokida, M. K., Karathanos, V. T., & Maroulis, Z. B. (1998). Effect of free-drying conditions on

shrinkage and porosity of dehydrated agricultural products. Journal of Food Engineering

, 35, 369-380.

Kumar, P., & and Mishra, H. N. (2006). Moisture sorption characteristics of mango-soy-fortified

yogurt powder. International Journal of Dairy Technology , 59 (1), 22-28.

Labuza, T. P. (1968). Sorption phenomena in foods. Journal of Food Technology , 22 (3), 15-24.

Labuza, T., & Altunakar, B. (2007). Water activity prediction and moisture sorption isotherms.

In G. V. Barbosa-Canovas., & A. J. Fontana, Water Activity in Foods: Fundamentals and

Applications (pp. 109-154). Ames, Iowa: Blackwell Publishing.

Lim, L. T., Tang, J., & He, J. (1995). Moisture sorption characteristics of freeze-dried

blueberrries. Journal of Food Science , 60, 810-814.

Maroulis, Z. B., Tsami, E., & Marino-Kouris, D. (1988). Application of the GAB model to the

soprtion isotherms of for dried fruits. Journal of Food Engineering , 7 (1), 63-78.

Nindo, C. I., & Tang, J. (2007). Refractance Window dehydration technology: A novel contact

drying method. Drying Technology , 25, 37–48.

Rahman, M. S. (1995). Food Properties Handbook. Boca Raton, FL: CRC Press.

143

Rahman, M. S. (2006). State diagram of foods: Its potential use in food processing and product

stability. Trends in Food Science & Technology , 17, 129-141.

Rahman, M. S. (2001). Toward prediction of porosity in foods during drying: A brief review.

Drying Technology , 19 (1), 1–13.

Rahman, M. S., & Labuza, T. P. (1999). Water activty and food preservation. In M. S. Rahman,

Handbook of Food Preservation (pp. 339-382). New York: Marcel Dekker, Inc.

Rajkumar, P., Kailappan, R., Viswanathan, R., Raghavan, G. S., & Ratti, C. (2007). Foam mat

drying of alphonso mango pulp. Drying Technology , 25, 357-365.

Rangel-Marron, M., Welti-Chanes, J., Cordova-Quiros, A. V., Ceron-Breton, J. G., Ceron-

Breton, R. M., & Anguebes-Fransechi, F. (2011). Estimation of sorption isotherms of

mango pulp freeze-dried. International Journal of Biology and Biomedical Engineering ,

5 (1), 18-23.

Rizvi, S. S. (1995). Thermodynamic properties of foods in dehydration. In M. a. Rao,

Engineering properties of foods (pp. 223-309). New York, USA: Marcel Dekker, Inc.

Rockland, L. B., & Stewart, G. F. (1981). Water Activity: Influence of Food Quality. New York:

Academic Press.

Roos, Y. H. (1987). Effect of moisture on the thermal behavior of straberries studied using

differential scanning calorimetry. Journal of Food Science , 1, 146-149.

Roos, Y. H., & Karel, M. (1991a). Phase transition of amorphous sucrose and sucrose solution.

Journal of Food Science , 56, 266-267.

Roos, Y. H., & Karel, M. (1991b). Plasticizing effect of water on thermal behaviour and

crystallization of amorphous food models. Hournal of Food Science , 56, 38-43.

144

Roos, Y. (1995). Water activity and glass transition temperature. In G. V. Barbosa-Canovas, & J.

Welti-Chanes, Food Preservation by moisture control: Fundamentals and applications

(pp. 133-154). Pennsylvania: Technomic.

Sablani, S. S., Kasapis, S., & Rahman, M. S. (2007). Evaluating water activity and glass

transition concepts for food stability. Journal of Food Engineering , 78, 266-271.

Sablani, S. S., Syamaladevi, R. M., & Swanson, B. G. (2010). A review of methods, data and

applications of state diagram of food systems. Food Eng Rev , 2, 168-203.

Slade, L., & Levine, L. (1991). A food polymer science approach to structure property

relationships in aqueous food systems: Non-equilibrium behaviour of small carbohydrate-

water system. In L. Levine, & L. Slade, Water Relationships in Food. New York:

Plennum Press.

Spiess, W. E. (1987). Critical evaluation of methods to determine moisture sorption isothherms.

In L. B. Rockland, & L. R. Beuchat, Water Activity: Theory and applications to foods

(pp. 215-234). New York: Mercel Dekker.

Syamaladevi, R. M., Sablani, S. S., Tang, J., Powers, J., & Swanson, B. G. (2009). State diagram

and water adsorption isotherm of raspberry (Rubus Idaeus). Journal of Food Engineering

, 460-467.

Talla, A., Jannot, Y., Nkeng, G. E., & Puiggali, J.-R. (2005). Experimental determination and

modeling of sorption isotherms of tropical fruits: banana, mango and pineapple. Drying

Technology , 23, 1477-1498.

Telis, V. R., & Sobral, P. J. (2002). Glass transition of freeze-dried and air-dried tomato. Food

Research International , 35, 435-443.

145

Yu, L., Mazza, G., & Jaya, D. S. (1999). Moisture sorption characteristics of freze-dried, osmo-

freeze-dried and 0smo-air-dried cherries and bluberries. Transactions of the ASAE , 42

(1), 141-147.

146

Table 1. Temperature, retention time and water content of mango puree and Refractance Window- and freeze-dried mango powders.

Product Temperature, Retention time Water content Product (°C) (min) (kg water/kg dry solids)

Fresh puree - - 6.518 ± 0.123 RW 74 ± 2 3 ± 0.01 0.017 ± 0.001 FD 20 ± 1 1,860 ± 85 0.023 ± 0.002

Table 2. Measured GAB parameters of mango powders and other sugar-rich fruits BET Model parameters GAB Model parameters Air Mg (kg Mg (kg Product 2 2 Temperat H2O /kg B R H2O /kg C K R ure dry solids) dry solids) RW-dried mango a 23oC 0.081 3.77 0.984 0.078 0.081 3.77 0.984

Freeze-dried mango a 23oC 0.087 3.35 0.986 0.045 0.087 3.35 0.986

Freeze-dried mango b 25oC 0.136 -19.54 0.971 0.124 0.136 -19.54 0.972

Osmo-oven dried mango c 25oC NA NA NA 0.166 NA NA NA

Dehydrated mango d 40oC 0.129 94.30 NA 0.096 0.129 94.30 NA

Dehydrated pineapple d 40oC 0.266 24.45 NA 0.185 0.266 24.45 NA

Dehydrated banana d 40oC 0.181 74.49 NA 0.108 0.181 74.49 NA

Freeze-dried Raspberry e 23oC 0.056 NA NA 0.074 0.056 NA NA

Freeze-dried apple f 25oC NA NA NA 0.120 NA NA NA

Air-dried apple f 25oC NA NA NA 0.125 NA NA NA

a Present study b Rangel–Marrón, et al., 2011 c Falade and Aworth, 2004 d Talla, et al., 2005 e Syamaladevi, et al., 2009 f Jakubczyk, et al., 2010 NA - Not available

Mg - Monolayer mositure content

147

Table 3. Glass transition temperatures and moisture contents of Refractance Window- and freeze-d ied p wde s at wate activity (0.113 ≤ aw ≤ 0.860). Water Content Water Tgi (°C) Tgm (°C) Tge (°C) (kg water Activity /kg dry solids) RW FD RW FD RW FD RW FD 0.113 30.6 ± 1.5a 33.4 ± 3.4a 35.6 ± 1.5j 39.7 ± 1.2j 39.2 ± 1.21 46.1 ± 0.92 0.03 ± 0.00113 0.03 ± 0.00513

0.225 13.9 ± 0.1b 13.6 ± 6.1b 20.3 ± 1.3k 27.2 ± 5.3l 24.6 ± 0.83 35.2 ± 4.24 0.05 ± 0.00114 0.05 ± 0.00214

0.328 -3.3 ± 0.9c 0.2 ± 2.5c 1.8 ± 0.2m 5.9 ± 3.3m 5.9 ± 0.25 10.8 ± 2.66 0.08 ± 0.00115 0.08 ± 0.00115

0.432 -17.5 ± 0.2d -16.6 ±0.5d -12.7 ± 0.5n -13.1 ± 0.5n -10.1 ± 0.67 -9.7 ± 0.37 0.11 ± 0.00116 0.11 ± 0.00116

0.529 -31.9 ± 0.9e -31.4 ± 0.4e -26.7 ± 0.6o -27.3 ± 0.1o -23.3 ± 0.18 -23.9 ± 0.58 0.15 ± 0.00217 0.15 ± 0.00117

0.658 -45.3 ± 3.2f -47.4 ± 0.6f -40.6 ± 2.9p -42.4 ± 0.6p -37.6 ± 3.39 -39.4 ± 0.79 0.20 ± 0.00118 0.22 ± 0.00119

0.750 -56.6 ± 3.5g -60.7 ± 0.1g -52.3 ± 2.8q -57.0 ± 1.7r -49.7 ± 2.710 -53.3 ± 1.010 0.28 ± 0.00320 0.30 ± 0.00421

0.860 -65.3 ± 1.9h -72.2 ± 1.4i -61.4 ± 3.3r -67.8 ± 1.6s -57.6 ± 1.811 -64.7 ± 1.512 0.34 ± 0.00522 0.37 ± 0.01423

(a-s); (1-23)Different superscript letters and numbers represent statistical significant differences of glass transition temperatures of RW- and freeze-dried mango powders at various water activities (aw 0.113 0.86).

Table 4. Evaluating water sorption isotherm and glass transition models of Refractance Window- and freeze-dried mango powders using BET monolayer water content. Sorption isotherm model Glass transition model aw corresponding T from a Temper BET monolayer g Water w to the glass corresponding to Product ature water content T content monolayer transition g the monolayer °C (kg water/kg (°C) (kg/kg water content model water content mango mango (Fraction) (°C) (Fraction)

RW-dried 23 0.081 0.34 0 23 0.017 0.16 mango powder

Freeze-dried 23 0.087 0.33 0 23 0.023 0.18 mango powder

148

(a)

(b)

Figure 1. Experimental set-up for determining water sorption isotherms of RW- (a) and freeze- dried (b) mango powders.

Figure 2. Differential scanning calorimeter (DSC, Q2000, TA Instruments, New Castle, USA).

149

0.45 RW-Observed RW-BET Predicted 0.40 RW-GAB Predicted FD 0.35 FD-Observed FD-GAB Predicted RW FD-BET Predicted 0.30

0.25

0.20

0.15

0.10

0.05 Moisture Content, kg water/kg dry solidsContent,dry water/kg kg Moisture 0.00 0.00 0.20 0.40 0.60 0.80 1.00

Water Activity, aw

Figure 3. Water adsorption isotherm data for Refractance Window- and freeze-dried mango powders at 23 °C, with fitted curves using GAB and BET models.

150

(a)

Tgi

Tgm

Tge

(b) Figure 4. Glass transition temperatures of Refractance Window-dried (a) and freeze-dried (b) mango powders equilibrated over selected water activity (scan rate of 5 °C/min).

151

d 65

45 c

25 a C)

RW Observed T

( gi

gi FD Observed Tgi

T 5 RW G-T model FD G-T model -15

Mango powder Tgi , C Tgs , C

-35 RW dried 30.6 1.5 (a) 55.8 (c) Freeze dried 33.4 3.4 (b) 63.6 (d) -55

-75

Glass Transition Temperataure, Temperataure, Transition Glass -95 Tg of water = − 13 C -115 Solid content (Xs) RW dried =0.983 Freeze dried = 0.977 -135 0.0 0.2 0.4 0.6 0.8 1.0

Solid Content, Xs (kg solid/kg mango pulp)

Figure 5. State diagram of Refractance Window- (RW) and freeze-dried (FD) mango powders. The onset (Tgi) and solids (Tgs) glass transition temperatures of RW- and freeze-dried mango powders are represented by letters (a & b) and (c & d), respectively.

152

Glass transition temperature vs water content 60 Water content vs water activity 1 RW G-T model RWBET Model 40 RW GAB Model 0.9

20 0.8

(

gi 0 0.7 T

-20 0.6

(Fraction)

w a -40 0.5

-60 0.4

-80 0.3 Activity, Water aw = 0.34

Glass Transition Temperature, Temperature, Transition Glass -100 0.2

-120 BET monolayer = 0.1 0.081 kg water / kg aw = 0.16 -140 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Moisture Content, Xs (Kg Water/Kg mango pulp)

Figure 6. Water plasticization and sorption characteristics of Refractance Window-dried (RW) mango powders showing water activity, water content and glass transition temperature.

153

Glass transition temperature vs water content Water content vs water activity 60 FD G-T model 1 FD BET Model FD GAB Modelz 40 0.9

20 0.8

(

gi 0 0.7 T

-20 0.6

(Fraction)

w -40 0.5 a

-60 0.4

-80 0.3 Water Activity, Activity, Water aw = 0.33

Glass Transition Temperature, Temperature, Transition Glass -100 0.2

BET monolayer = 0.087 -120 kg water / kg ds aw = 0.18 0.1

-140 0 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Moisture Content, Xs (Kg Water/Kg mango pulp)

Figure 7. Water plasticization and sorption characteristics of freeze-dried (FD) mango powders showing water activity, water content and glass transition temperature.

154

(a) (b)

(c) (d) 1 Figure 8. Scanning elect n ic g aphs (S M) ang p wde s (180 2 0 m) dried using Refractance Window® drying (a&b) and freeze drying (c&d) at magnification of 300x (a & c) and 1,000 (c & d), 20 Kv.

155

CHAPTER FIVE

PHYSICAL AND CHEMICAL STABILITY OF REFRACTANCE WINDOW®-DRIED

MANGO (PHILIPPINE 'CARABAO' VAR.) POWDER DURING STORAGE

ABSTRACT

Fundamental investigations on the stability of fruit powders produced using emerging novel drying technologies are crucial for assessing the shelf life of dehydrated products and for selecting appropriate packaging and storage conditions needed to preserve quality. Refractance

Window® (RW) drying is one such technology that has been proven to produce high quality food powders from different raw materials. However, information on the stability of RW-dried products during long-term storage is still lacking. In this study, the effect of packaging atmosphere, storage temperature and time on the physical and chemical stability of RW-dried mango powder was evaluated over a period of 12 months. RW-dried mango powder with a water content of 0.037 ± 0.001 kg water/kg dry solids was stored at 5, 22 and 45 °C for 12 months using air packaging or nitrogen flushing. Headspace, water content, color, ascorbic acid,

β-carotene and microstructures were measured at months 0, 6 and 12. The mango powder stored at 45 °C suffered discoloration and ascorbic acid (AA) and β-carotene degradation after 6 and 12 months of storage in both air or nitrogen flush packaging, while powder stored at 5 and 22 °C had reduced nutrient degradation and preserved color of mango powder for both packaging conditions. Replacing the air inside the package with nitrogen gas was effective in preserving

AA in mango powder at room temperature and under refrigerated condition. Nitrogen flushing als educed the pe centage l ss β-carotene after 6 months, but no significant difference was found after 12 months. The initial structure of mango powder stored at 5 and 22 °C was preserved after 6 and 12 months. All of the entire powder particles inside the package formed

156

into a single lump when stored at 45 °C. The microstructures of mango powder at different storage conditions confirmed these results.

1. INTRODUCTION

The Philippine a aba ang va iety kn wn as “Manila Supe Mang ” anks thi d among the fruit crops produced in the country after banana and pineapple, based on export volume and value with a total of over a million metric tons harvested in 2007 (BAS, 2009). In terms of total mango production worldwide, Philippines ranked 8th in 2005 as producer and as exporter (Evans, 2008). Mang is ich in β-ca tene, anging 800 13000 µg per 100 g depending on the cultivar (Hymavathi, 2005). About 60% of all carotenoids found in mango is β- carotene (Jungalwala & Cama, 1963). It also contains ascorbic acid, ranging from 7.8–172 mg/100g of ripe mango pulp (Abdon et al., 1990; Tharanathan et al., 2006). However, the problem confronting the Philippine mango industry is seasonality of production and unstable market situations. Mango is highly perishable, prone to mechanical damage and microbial contamination during postharvest handling and processing (Rajkumar et al., 2007). Serrano reported that physical damage and chemical quality losses in mango ranges between 5–87%, depending on how the product is handled. In fact, untreated ripe mango fruit 'Hadin' suffered

100% fungal infection and severe decay damage during 18 days of storage at 25 °C (Gonzalez-

Aguilar et al., 2007).

Development of alternative products, such as mango powder, are gaining popularity as ingredients of other food products including health drinks, baby foods, sauces, marinades, confections, yogurt, ice cream, nutrition bars, baked goods, cereals (Rajkumar et al., 2007) and in the pharmaceutical and cosmetic industry (FAO, 2007). Currently, several drying technologies for commercial production of fruit flakes or powder are available, including Refractance

157

Window® drying (RW), a novel drying technique designed mainly to convert liquid and semi- liquid foods or puree into powder or flakes (Bolland, 2000). The principles and operations of this drying technology are described in Nindo et al. (2003a), Nindo & Tang (2007) and Caparino et al. (2012).

RW drying offers many advantages and benefits when applied to fruits and vegetables. In experiments using carrots, strawberries and squash, it was shown that the RW drying system maintains the nutritional (vitamins, antioxidants) and sensory (color, aroma) attributes of the dried products (Nindo and Tang, 2007). Nindo et al. (2003a) also reported that the bright green color of RW-dried asparagus remains unchanged. An earlier investigation by Abonyi et al.

(2002) also showed that RW-dried carrots and strawberries were not significantly different when compared to their freeze-dried counterparts. Moreover, the study found retention in ascorbic acid ( 4%) and β-carotene (90%) in strawberries after RW drying, which were comparable to freeze-dried products. Most recently, Caparino et al. (2012) concluded that the physicochemical properties of RW-dried mango powder were comparable to the freeze-dried counterpart and better than drum and spray-dried mango powders. RW dryers also have higher thermal efficiency and low energy consumption compared with conventional dyers. Nindo & Tang (2007) and

Nindo et al. (2003b) reported that the coliforms, Escherichia coli and Listeria innocua in pumpkin puree were significantly reduced to a minimum detection level after RW-drying, indicating that foods dried using this method are safe. Despite the reported positive results of fruit products derived from RW-drying, there is no published literature showing thermal degradation, physical and nutritional changes of RW-dried mango powder during long term storage. The objective of this study was to evaluate the physical and chemical stability of RW-

158

dried mango powder during storage as affected by packaging atmosphere and storage temperature over a period of 12 months to aid in the manufacture of high quality mango powder.

2. MATERIALS and METHODS

Fig. 1 shows the general procedure of the storage study. Mango puree was dried into thin sheets or flakes using Refractance Window® drying. Large flakes RW-dried mango powder with water content of 0.037 ± 0.001 kg water/kg dry solids was stored at 5 °C (refrigerator), 22 °C

(ambient temperature) and 45 °C (oven) for 12 months using air packaging and nitrogen flushing. The physical and nutritional qualities of the samples were analyzed at 0, 6 and 12 months.

2.1. Preparation of mango flakes

Frozen mango puree (Philippine 'Carabao' var.) was acquired from Ramar Foods

International (Pittsburg, California). The mango puree was prepared following the same procedure as described by Caparino et al. (2012). The average water content of the puree determined using a standard oven method (AOAC, 1998) was 6.5 ± 0.1 kg water/kg dry solids, while the sugar content was 14.5 ± 0.5 °Brix. Before drying, the samples were thawed overnight at ~23 °C and blended (Oster blender, Mexico) to a uniform consistency using the lowest speed for 5 minutes. The puree was dried to 0.017 ± 0.001 kg water/kg dry solids using a pilot scale

Refractance Window® (RW) dryer (Caparino et al., 2012).

During the drying operation, the mango puree was uniformly spread on the top surface of a plastic conveyor belt, with the bottom surface being in contact with a e-ci culating h t wate

( 97 °C). Water temperature was maintained to transmit thermal energy directly into mango puree through the plastic conveyor belt (Abonyi et al., 2002; Nindo & Tang, 2007). The thickness of puree samples of 0.5 – 0.7 mm was controlled using a spreader bar located at the

159

inlet section of the dryer. Moisture removal from the samples was facilitated by forcing the suction air of 22° (RH 0 52 %) at an average velocity of 0.7 m/s. The dried mango puree exiting the dryer in the form of flakes or thin sheets was sealed in Ziploc® bags. The sealed bags were then put inside aluminum-coated polyethylene bags. The samples were flushed with nitrogen gas to prevent oxidation, heat sealed and stored at –35 °C until further analysis.

2.2 Sieving, packaging and storage

Dried mango flakes or sheets obtained from RW drying were ground using a mortar and pestle, and then sieved using mesh sizes 10 and 12 (American Society for Testing and Materials,

ASTM) t btain ang p wde with pa ticle si es anging 1.7 − 2.0 (Inte nati nal

Organization for Standardization, ISO), respectively (Barbosa-Cánovas, 2005). A multilayer film supplied by KSM Enterprise, WA designed for maximum barrier with zero oxygen transmission rate (OTR) and water vapor transmission rate (WVTR) (Osiopak, CA) was used for packaging the samples.

Thirty six pouches were prepared, each containing approximately 20 g of the dried mango powder (Fig. 2). One-half of the pouches were air packaged and the other half nitrogen flushed. Air-packaged pouches of mango powder were heat sealed using an impulse bag sealer

(16" Powerseal, CA), while an automatic vacuum sealer (UltraVac®250, KS) was used for the nitrogen-flushed counterparts, by first sucking out the air from the packages at 26 kPa followed by purging with nitrogen gas for 3 s, and finally heat sealing for a standard time of 7 s. Physical properties, nutritional qualities and microstructures were evaluated in duplicate during 0, 6 and

12 months of storage.

160

2.3. Water content and water activity

The water content of mango puree and RW-dried mango flakes/powders was determined using the standard oven method at 70 °C and absolute pressure of 13.3kPa for 24 h (AOAC,

1998). The drying, cooling and weighing of samples was continued until the difference between two successive weighing was less than 1 mg. The water activity of the RW-dried mango powder was measured in duplicate using a water activity meter (Aqualab 4TE series, Decagon Devices,

Pullman, WA) set at 24.7 ± 1 °C.

2.4. Physicochemical and thermal characteristics

2.4.1. Headspace gas

The headspace carbon dioxide, oxygen, and nitrogen gas concentration of the packaged mango flakes was measured with a gas analyzer (Fisher Gas Partitioner 1200, Fisher Scientific

Co., Pittsburgh, PA) equipped with a thermal conductivity detector. During the measurement, a 3 cm2 adhesive septum with a 0.5 cm layer of dried silicon gel was attached to the surface of the package to avoid leakage when the gas sample was collected (Fig. 3). A gas sample of 0.25 ml was withdrawn from the headspace of the package by a syringe through the septum. Gas concentration from each package headspace was determined by injecting the gas sample into a

6.096 m stainless steel column (3.18 mm ID) packed with 80/100 mesh Chromasorb P-AW and a

0.9144 m (4.76 mm ID) aluminum column packed with 60/80 mesh Molecular Sieve 13X connected in series and a thermal conductivity detector inside the gas analyzer. The ultra-pure

-1 He2 carrier gas flow was adjusted to 10 mL.min operated at oven temperature of 40°C (Rudell et al., 2000). All readings were performed in duplicate at 23 °C.

161

2.4.2 Color measurement

The International Commission on Illumination (CIE) color parameters L*, a* and b* values of stored mango powder were measured at different storage time intervals with a color meter (Minolta Chroma CR-200, Minolta Co., Osaka, Japan). The color meter was calibrated with a standard white ceramic plate (L* = 95.98, a* = – 0.13, b* = – 0.30). Prior to taking the readings, mango powder was poured into Petri dish to form a 10 mm thickness layer and covered with transparent film (SaranTM Wrap, SC Johnson, Racine, WI). The average L*, a* and b* values were obtained from 6 readings taken from each of 5 locations of each sample. The hue

angle, and chroma, expressed as = and = , respectively, were

calculated.

For color comparison with the original mango puree, 2 grams each of the stored mango flakes with water content of 0.017 ± 0.001 kg water/kg dry solids was reconstituted by adding

12.10 g of distilled water. The reconstituted mango flakes produced slurries with water content of 6.143 kg water/kg dry solids similar to the original mango puree. The reconstitution of mango powder water mixture was stirred at 23 °C using a vortex mixer (mini vortexer, Fisher Scientific,

CA) until the powder was completely dispersed. The L*, a* and b*, H* and C* values were immediately measured and calculated following the same procedure employed for mango flakes.

The total change in color of the reconstituted mango powders with reference to the original puree

was computed as: whe e subsc ipt “ ” den tes the color of original puree (Nindo et al., 2003a; Jaya & Das, 2004).

2.4.3. L-Ascorbic Acid analysis

The L-Ascorbic acid content of mango powder was determined by the 2,6-

Dichloroindophenol Titrimetric method (AOAC Official Method 967.21) with modification.

162

Filtered indophenol dye solution containing 50 mg 2, 6 dichlorophenolindophenol (Sigma-

Aldrich, USA), standard ascorbic acid solution containing dissolved pure ascorbic acid in distilled water (1 mg/mL), and extracting solutions (3 % metaphosphoric acid (HPO3) were prepared for the analysis. Calibration standard curve for L-Ascorbic acid was established with 6 data points: 0.2, 0.4, 0.6, 0.8 mg/mL (ascorbic acid : HPO3) and blank (HPO3) and undiluted ascorbic acid as end points (Fig. 3) . A 2 mL solution placed in a 25 mL conical flask was titrated with indophenol solution until a faint pink color persisted for 30 s. The ascorbic acid concentration was expressed as mg ascorbic acid equivalent to 1 mL dye solution or mg / mL.

The L-Ascorbic acid from mango powder was determined by blending 2.5 g of mango flakes with 25 mL extracting solution for 3 minutes at 23 °C using a vortex mixer (mini vortexer,

Fisher Scientific, CA). The mixture was centrifuged at 3000 g for 10 min at 4 °C. Aliquots of the supernatant (2 mL) were transferred to 25 mL conical flasks, and titrated rapidly with 2,6 dichlorophenolindophenol solution until a faint pink color persisted for 30 s. All measurements were performed three times and the readings agreed within 0.1 mL deviation.

2.4.4. β-carotene analysis

Fig. 4 shows a general procedure during analysis of β-carotene. An extraction method used by Kaspar et al. (2011) and Abonyi (2000) was adopted in this study with modification.

One gram of newly-dried or stored mango powder (< 2 mm) was first dispersed with 2 mL distilled water in a 50 mL centrifuge tube and mixed thoroughly. Then, 7.5 ml acetone (0.1 %

BHT in ethanol) was added and vortexed (mini vortexer, Fisher Scientific, CA). The resulting slurry was cold saponified by adding 1.2 mL of 40 % methanolic KOH solution (MeOH) and left to stand in the dark for 16 h at 4 °C. A 7.5 mL solvent mixture (hexane/ethyl acetate, (1:1, v/v) was added and vortexed for 1 min, followed by adding 10 mL of 10% Na2SO4 and vortexing for

163

1 min. The solution was allowed to stand in the dark for 1 h until the organic layer was separated. The upper organic layer was collected and transferred to a small tube and centrifuged at 600 g for 8 min (DYNAC Centrifuge, Becton Dickenson, MD). Two mL of supernatant were dried under nitrogen (Analytical Nitrogen Evaporator, Organomation Assoc., Inc. Berlin, MA) for analysis of carotenoids by HPLC. All extraction procedures were performed under dimmed light.

The dried extract was reconstituted in mobile phase (mixture of acetonitrile, methanol and chloroform, 47:47:6; v/v/v), and analyzed using HPLC. The HPLC system consisted of a

Waters 2690 separation module and a Waters 996 photodiode array detector (Waters Corp,

Mil d, MA). A 18, μ × 1 0 c lu n (Res lve, Wate s p ati n, Mil d, MA) was used t sepa ate the sepa ate the β-carotene. The flow rate was 1 mL/min with an injection volume of 10 L. Duplicate samples were analyzed for each treatment and the mean values were reported.

2.5. Microstructures of mango powders

A small quantity of grounded RW-dried mango powder (180 – 250 mm) was obtained by passing through sieve sizes 60 and 80 (American Society for Testing and Materials, ASTM)

(Barbosa-Cánovas et al., (2005) and mounted on aluminum stubs and coated with a fine layer of gold (15 nm) using a sputter gold coater (Technics Hummer V, Anatech, San José, CA). The gold coated samples were examined by a Quanta 200F Environmental Scanning Electron

Microscope (FEI, Field Emission Instruments, Hillsboro, OR) at low vacuum mode (200 Pa).

Observations were carried out at accelerated voltage of 30 kV and magnification of 100x and

500x at a scale of 1 mm and 200 m, respectively.

164

2.6. Statistical analysis

All experiments were carried out at least in duplicate, and the results were analyzed in full factorial using the general linear model procedure of SAS (SAS Institute Inc., Cary, NC), and the means separated by Tukey-honest significant difference test with a confidence interval of

95% used to compare means.

3.0. RESULTS and DISCUSSION

3.1. Water content and water activity

The water content and water activity (aw) of RW-dried mango powders packed in air or nitrogen increased after 6 and 12 months of storage at different temperatures is presented in

Table 1. Prior to storage, the water content and aw values of the powder was 0.037 ± 0.001 kg water/kg dry solids and 0.126 ± 0.003, respectively. Regardless of packaging atmosphere, we observed no significant (p < 0.05) increase in water content of mango powder stored at 5 and 22

°C between 6 and 12 months of storage, but a significant increase (p < 0.05) was observed in the mango powder stored at 45 °C. The nitrogen-flushed mango powder stored at ambient temperature obtained the lowest percentage increase after 6 months (1.4 %) and 12 months (2.1

%), while air-packaged mango powder stored at 45 °C showed the highest moisture gain at 17.5

% and 22.6 % after 6 and 12 months of storage, respectively. The increase in pressure in the headspace of the package at higher temperature could weaken the sealing and might have caused the entrance of moisture from the environment over time. Exposure of amorphous material to high temperature promotes crystallization and subsequently increases the water content within the product. Al Mahdi et al. (2006) reported that the release of water in stored dried skim milk was accelerated during crystallization process resulting in caking caused by inter-particle liquid

165

bridges. The water activity values for all the samples tested did not exceed 0.3. Some implications of this low water activity values are discussed in the succeeding sections.

3.2. Headspace gas analysis

The initial concentration of oxygen and nitrogen inside the air-packaged mango powder was 19.8 % and 80.2 % by volume, respectively (Table 2). The initial oxygen level in the headspace of the air-package mango powder significantly decreased (p < 0.05) over 6 months of storage at 5, 22 and 45 °C. However, the oxygen concentration of stored samples at 5 °C and 22

°C was the same after 6 and 12 months of storage, while significantly lower than samples stored at 45 °C at the same storage period. On the other hand, the oxygen concentration significantly increased (p < 0.05) when the packaged mango powder was stored for 6 and 12 months at 5, 22, and 45 °C using nitrogen-flushed packaging. But, the level of oxygen for all samples (< 4 % by volume) was still acceptable to measure its effect on the physical and chemical stability of mango powder during storage. In a study on the effect of nitrogen flushing and storage temperature of whole milk, the authors discarded nitrogen-flushed samples when the oxygen concentration inside the package exceeded a value of 6.5 % by volume (Lloyd et al., 2009).

A carbon dioxide level of 6.1 % to 2.8 % by volume was detected inside the air and nitrogen-flush packaging after 6 to 12 months of storage, respectively. One possible reason for this was the presence of citric acid, which is the major organic acid found in mango (Jain et al.,

1959). It might be that during storage at higher temperature, the acetyl CoA molecule, which is the fuel for the citric acid cycle, was formed from the breakdown of starch (storage form of glucose) and completely oxidized to CO2 – decarboxylation (Berg et al., 2002). It may also be that the occurrence of carbon dioxide inside the package indicates the presence of heat resistant bacterial spores, which might have been obtained during puree preparation, drying and

166

packaging operation. It has been reported that certain bacterial spores and vegetative cells exist at low water activity. Fine & Gervais (2005) found that heat resistance spores and vegetative cells were strongly improved for initial aw values in the range of 0.3 to 0.5. We did not carry out microbial analysis on our samples because the measured water activity of mango powder in all levels of temperatures did not exceed 0.3 after 6 and 12 months of storage (Table 2). According to Barbosa-Cánovas et al. (2003) and Toledo (2007), dried product with water activity level below 0.3 is safe from various microbial parameters. Likewise, other authors reported that lowering of water activity of a product below 0.75 will inhibit the growth of molds, yeast and bacteria (Labuza et al., 1972).

Extending the storage time of air-packaged samples to 12 months further reduced the level of oxygen by 12.8 %, while the amount of carbon dioxide present in air and nitrogen- flush packaging further increased to 10.9% and 4.1 % by volume, respectively. The replacement of air by nitrogen gas raised the headspace nitrogen level from 80.2 % (initial) to 97.7 % by volume.

Keeping the mango powder at 5 and 22 °C for up to 12 months in nitrogen-flushed technique showed a significant reduction of 3.3 % and 6 % after 6 and 12 months of storage, respectively.

However, the rate of increase for these samples at 5 and 22 °C were not significant. No carbon dioxide was traced in nitrogen-flushed mango powder at 5 and 22 °C up to 12 months of storage, but it was detected at 45 °C with a lower concentration compared to air packaging. It is apparent that regardless of packaging atmosphere, carbon dioxide evolved to a certain level at 45 °C but was completely absent at 5 and 22 °C.

3.3. Color stability

The photographs of RW-dried mango powder showed similar yellow color at 5 and 22 °C after 6 and 12 months of storage in air packaging and nitrogen-flushed bags but the powders

167

turned brown color when stored at 45 °C (Fig. 5a). The L*-values (lightness) of air-package and nitrogen-flushed mango powder as presented in Table 3 remained stable over 12 months of storage at 5 and 22 °C, confirming the visual appearance as shown in Figure 4a. However, the

L*-value was significantly reduced (p < 0.05) when mango powder was stored at 45 °C, beginning at 6 months and further darkened after 12 months of storage. The a* (degree of redness and greenness) values obtained in both air- and nitrogen-flushed packaging during the 12 month storage increased while the corresponding L*-values decreased. This can be explained by the higher oxidation of β-carotene, which affected the natural yellow color of mango powder turning it to brown. Desobry et al. (1997) reported similar relationship between the a*- and L*- values for freeze-dried encapsulated β-carotene during storage.

The darkening or formation of brown color as indicated by reduced L*-value of mango powders stored at 45 °C can be attributed to non-enzymatic browning (NEB). Sagar et al. (2000) reported that the NEB of air-dried mango powder stored at 35 °C was significantly higher compared to mango powder stored at 7 °C. Also, a contributory factor in the formation of brown color in mango powder could be due to the significant reduction of ascorbic acid as further discussed in the succeeding section. Hymavathi & Khader (2005) reported that browning of vacuum-dried mango powder packed in highly oxygen permeable packaging material during 6 months of storage was associated with high ascorbic acid degradation. Ascorbic acid degradation was also reported as the major mechanism for browning in lemon juice (Robertson &

Samaniego, 1986). Moreover, the browning of mango powder stored at 45 °C can also be attributed to caramelization. Although caramelization occurs when a highly concentrated sucrose solution is heated to a high temperature (above boiling) (Eggleston & Vercellotti, 2000; Owusu-

Apaten, 2005; Quintas et al., 2000), our study showed that RW-dried mango powder, which

168

contains high sucrose concentration caramelized after 6 and 12 months of storage when stored at

45 °C (Fig. 5a). Another explanation could also be due to Maillard reaction when the package was stored at 45 °C. Maillard reaction is a type of non-enzymatic browning which involves the chemical reactions between sugars and amino acids (Potter & Hotchkiss, 1995). Mango contains

ee a in acids anging 30 − 126 g 100 g-1 (Kohli et al., 1986). The negligible color change of mango powder stored at ambient or low temperatures for 12 months is a good quality attribute for the RW-dried mango powder product.

The color b* value is a good indicator to distinguish the color difference in yellowness of mango powders as affected by storage condition. The b* value of air-packaged mango powder after 6 months of storage showed significant difference, having a lower values compared to the initial b* value of mango powder, while the b* value of samples stored at 45 °C dropped significantly, indicating its darker color (Table 3). The b* value of nitrogen-flushed mango powder stored at different temperatures was significantly lower than air packaging (p ≤ 0.0 ). It is interesting to note that nitrogen flushing preserved the yellowness of mango powder stored at ambient temperature after 6 and 12 months, while powder color was significantly altered when stored at 5 and 45 °C (p ≤ 0.0 ). Als , the vividness in yell w c l ep esented by ch a values of all stored samples significantly changed (p ≤ 0.0 ) a te 6 nths st age at different temperatures in both packaging atmospheres, except for samples stored under nitrogen at 22 °C.

When the stored mango powders were reconstituted to the same water content as the original puree, their visual color appearance followed a similar trend as in powder form. Slight changes were observed after 6 and 12 months except for samples stored at 45 °C, which turned brown both for air-packaged and nitrogen-flushed (Fig. 5b). Apparently, regardless of

169

temperature and packaging atmosphere, the L*, a*, b* values of mango powder were significantly different among each other with the exceptions of samples stored with nitrogen flushing at 22 °C where yellowness was unchanged (Table 4). Likewise, the yellow vividness

(chroma value) of reconstituted mango powder stored at 22 °C revealed no significant difference with the original puree after 12 months of storage. Moreover, the hue angle of the same samples did not change after 6 months of storage at 22 °C in both packaging. Interestingly, color change was more pronounced in cold storage than in ambient temperature which needs further investigations. However, one study revealed that discoloration of egg yolk during storage is more dominant when stored under refrigeration than in warmer temperature (Abo-ashour & Edwards,

1970). Aydin (2006) believed that discoloration of egg yolk in cold storage could be attributed to yolk hardening. The lower hue angle value obtained for mango powder stored at 45 °C indicates dull color.

The total deviation (E) in color of reconstituted mango powder at different temperatures with respect to the original puree ranged from 1.87 ± 0.09 - 16.02 ± 0.10 in nitrogen-flushed and

5.93 ± 0.12 - 20.82 ± 0.09 in air-packaged samples, respectively (Table 4). The lowest deviation in color of reconstituted mango powder was detected in nitrogen-flushed samples stored at 22 °C

(E = 1.87 ± 0.09), while the highest values were observed in air-packaged samples stored at 45

°C having a E value of 20.82 ± 0.09. In this study, it can be generally concluded that the color of mango powder and its reconstituted form can be well preserved with the aid of nitrogen flushing during storage of up to 12 months.

3.4. L-Ascorbic Acid

The ascorbic acid (AA) content of mango powders obtained from Refractance Window®

(RW) drying and freeze drying process significantly decreased over a period of 12 months of

170

storage regardless of the packaging atmosphere (p < 0.05) (Fig. 6). Packaging atmosphere in this study comprises air-packaging and nitrogen-flushing containing oxygen headspace of 19.8 and

2.4 %, respectively (Table 2). It can be seen from the figure that, as expected, the percentage loss of ascorbic acid increased as the storage temperature increased. At storage temperature of 45 °C, ascorbic acid content of RW-dried mango powder was significantly reduced from 13 ± 0.2 mg/100g dried solids to 1.1 ± 0.1 mg/100g dried solids or 91.3 % loss and 1.6 ± 0.2 mg/100g dried solids or 87.5 % loss after 6 months of storage using air-packaging and nitrogen-flushing, respectively (p < 0.05). This result indicates high storage temperature affects AA retention more than the packaging atmosphere in mango powder. Several kinetic studies confirmed that ascorbic acid is a very heat-sensitive compound and its destruction increases as the temperature of the product in dried or aqueous solution increases (Kirk et al., 1977; Eison-Perchonok & Downes,

1982; Johnson et al., 1995; Kabasakalis et al., 2000; Uddin et al., 2002), Another factor for AA loss may be include the increase in moisture at 45 °C over time, as explained in section 3.1. The slight increase in ascorbic acid degradation observed after 12 months (91.4 to 92.6 %) of storage in both packaging atmospheres may be attributed to the small amount of AA left after 6 months of storage. The effect of packaging atmosphere on the loss of ascorbic acid after 12 months at 45

°C yielded no significant difference (p < 0.05) among all samples.

The combined use of 5 °C refrigerated storage with nitrogen flushing led to the lowest proportional AA loss (10.5 %) among all treatments over the period of 12 months. This small

AA loss might can be attributed to the depression of the aerobic reactions inside the package

(Abonyi et al., 2002). The loss of ascorbic acid in nitrogen-flushed mango powder at room temperature (22 °C) slightly increased (10.5% to 23.0 %) but was significantly lower compared to air-packaged powder (31.7 %) during 6 months storage (p < 0.05). AA loss of mango powder

171

further increased to 31.8 % in nitrogen-flushed and 41.7% in air package at the end of 12 months storage.

The AA loss of RW-dried mango powder stored in nitrogen-package for 12 months at 22

°C (31.8 %) was about one-third lower when compared with a previous studies on freeze-dried mango 'Badami' powder (42.3 %) packed in cans under nitrogen and stored at 24 – 28 °C (Ammu et al., 1977). Also, the AA loss of RW-dried mango powder in both air packaging and nitrogen flushing was more than twice lower than the vacuum-dried mango powder packed in a high oxygen permeable packaging material (82 – 83 %) stored at 27 – 32 °C (Hymavathi & Khader,

2005). By taking freeze-dried mango powder as the standard for a quality product, we can conclude that AA loss of RW-dried mango powder cannot be associated with the drying process, but merely based on the effect of storage conditions and the packaging material used in our study. However, it was proven that RW-dried mango powder can well be preserved compared to freeze-dried and vacuum-dried mango powder. The variations may be attributed to variety, drying preparation and packaging methods. This is a good indication that the physicochemical stability and quality attributes of RW-dried mango powder during storage is superior compared to that of mango powder produced by the widely accepted standard process of freeze drying.

The loss of ascorbic acid in nitrogen-flushed packaging was found to be significantly (p <

0.05) lower than air-packaged mango powder stored at 5 and 22 °C over a 12 month period of storage, indicating that the presence of oxygen in the headspace in the package at room temperature and under refrigeration greatly affect degradation of ascorbic acid. Kirk et al. (1977) reported that the destruction of ascorbic acid in a model dehydrated food system was dependent on oxygen and moisture contents. Dennison and Kirk (1978) further concluded that dissolved and gaseous oxygen are the primary factors that affect the storage of dehydrated food. The effect

172

of nitrogen flushing on the ascorbic acid retention in orange juice during storage at 4, 22 and 37

°C was investigated by Zhang et al. (2002). The study revealed no significant difference (p >

0.05) in the loss of ascorbic acid between air and nitrogen head space, suggesting that nitrogen flushing is not effective for orange juice.

The increasing trend in degradation of ascorbic acid during 6 and 12 months of storage at

5, 22 and 45 °C was directly related to the water activity of mango powder (Table 1). This finding is in agreement with the study of some authors on model food who claim that the rate of ascorbic acid degradation increased as the water activity increased (Lee & Labuza, 1975). In general, the replacement of air inside a sealed aluminized package with nitrogen gas proved to extend the shelf-life of mango powder for up to 12 months at room temperature (22 °C ) and below with ~90% retention in ascorbic acid content. It is also apparent from the present study that temperature and time are the most critical factors that affect ascorbic acid degradation. This observation is in agreement with other authors who reported that packaging technique is not sufficient to prevent ascorbic acid degradation unless accompanied by lowering the storage temperature (Trammell et al., 1986).

3.5. β-carotene

The β-carotene content of the mango powders obtained from the Refractance Window®

(RW) drying process significantly decreased during 12 months of storage regardless of the packaging atmosphere and temperature (Fig. 7). The effect of high temperature at 45 °C severely degraded the β-carotene content of mango powder when using air packaging, with a percentage loss of 82.4 % and 98.5 % after 6 and 12 months of storage, respectively. The β- carotene loss of air-packaged mango powder was slightly lower at 22 °C after 6 months (54.2 %) but was very high after 12 months of storage (92.6 %). Our results after 6 months of storage was

173

similar to others (Hymavathi & Khader, 2005) who reported β-carotene loss ranging from 58 –

62 % in vacuum dehydrated mango stored in high oxygen permeability packaging material during the same period of storage at room temperature (27-32 °C).

When the mango powder was stored at 5 °C at the same air-packaging atmosphere, the β- carotene loss was reduced to 37.2 % after 6 months, but incurred similar destruction with other treatments after 12 months of storage. The high sensitivity of β-carotene to temperature coupled with higher concentration of oxygen inside the package are likely the main cause for the high degradation of β-carotene in mango powder (Desobry et al., 1997). It was reported that the possible cause of carotenoid degradation in dehydrated products is due to an increase in porosity and surface area (Hymavathi & Khader, 2005). Also, loss of β-carotene can generally be associated with auto-oxidation, wherein the high susceptibility of unsaturated chemical structure in the product contributes to thermal degradation and oxidation (Stefanovich & Karel, 1982).

The decrease in β-carotene can also be attributed to the formation of cis-isomers due to degradation of all-trans β-carotene (Khachik et al., 1986; Desobry et al., 1997; Tang & Chen,

2000). Another factor in our study that might have increased the degradation of β-carotene is the low water activity, aw < 0.3 of the mango powder regardless of storage temperature and duration

(Table 1). Lavelli et al. (2007) found a U-shaped response curve when they plotted β-carotene against water activity of freeze-dried carrots stored at 40 °C. The authors reported a minimum loss of β-carotene at water activity ranging between 0.34 – 0.54 while they recorded increased degradation at below and above these values. Likewise, Arya et al. (1979) found the β-carotene of dehydrated carrots more stable at aw range of 0.32 – 0.57.

The replacement of air in nitrogen-flushed mango powder packages yielded significantly lower loss of β-carotene with 26.07 %, 21.8% and 23.38 % at 5, 22 and 45 °C, (p < 0.05) after 6

174

months of storage, respectively. It can be seen here that β-carotene losses at different temperatures are not significantly different, indicating the positive effect of nitrogen flushing in maintaining the stability of β-carotene even at higher temperature. The observation reported by

Talcott & Howard (1999) supports these findings; this group claimed that less degradation of β- carotene in processed carrots was observed when treated with nitrogen compared to oxygen- treated and control samples.

However, there was considerable decline of β-carotene content of RW-dried mango powder over a period of 12 months with losses ranging from 86.5 – 87.1 % for all temperatures applied and no significant difference was found compared to samples stored in air packaging.

The increase of water content measured inside the package after 12 months of storage ranging from 7 - 22.6 % in air packaging and 2.1 - 15.6 % in nitrogen flushing (Table 1) may also have contributed in the degradation of β-carotene at longer storage periods. This finding indicates that the application of nitrogen flushing of RW-dried mango powder is not useful to prevent degradation of β-carotene after 12 months of storage.

3.6. Caking and glass transition temperature

One major observation in this study was the caking phenomenon (Palzer & Sommer,

2001) affecting the quality of mango powder after storage at high temperature. This finding is highly associated with the influence of glass transition temperature on the stability of a certain product. In a separate study conducted by Caparino et al. (2012), they reported that the onset glass transition temperature (Tgi ) for RW-dried mango powder at water content of 0.03 ± 0.001 kg water/kg dry solids was 30.6 ± 1.5 °C. In the present study, it is clearly proven that mango powders stored at 5 and 22 °C ( below Tg) were well preserved and free from structural

175

transformations, while mango powder stored at 45 °C ( > Tg) suffered intense degradation due to caking, resulting in a hard and brittle assemblage of powder particles (Fig. 8).

Caking normally occurs when amorphous sugar initially in the form of a rigid glass-like substance is transformed into a rubbery state (Peleg, 1993; Aguilera, 1995; Wallack & King,

1988; Roos & Karel, 1991b,c). Several authors reported that physical and chemical stability of amorphous food powders are well protected when the product is stored at temperature below Tg, while deteriorative reactions such as stickiness and caking occurred at temperature above Tg

(Arvanitoyannis, 1993; Aguilera et. al., 1995; Fitzpatrick et al., 2007; Syamaladevi et al., 2009;

Sablani et al., 2010). It has also been reported that sugar-rich food powders become sticky at temperature range (T-Tg) between 10 20 °C during storage (Roos & Karel, 1991c;

Arvanitoyannis, 1993). Jaya & Das (2009) found the sticky point temperature of mango powder at 11.5 and 15.5 °C (T-Tg) for 0 % and 5 % water content (db), respectively. This was consistent with our study of RW-dried mango powder stored at 45 °C, which has a (T-Tg) value of 14.4 °C.

3.7 Microstructures

The particle size of RW-dried mango powder prior to and during storage was at the range below 2 mm. For the scanning electron microscopic study, all the samples were converted into smaller particles (180 – 250 m) for better interpretation of the images. The microstructures of

RW-dried mango powders stored at different storage conditions are presented in Fig. 9. At the beginning, the microstructures of RW-dried mango powder can be distinguished by its smooth, and flaky with uniform thickness, nearly the same as reported by Caparino et al. (2012). The powder samples clearly show that their particles were un-agglomerated. The two sides of every single particle were smooth indicating more flowability and less susceptibility to oxidation because of lesser surface area. These particle characteristics of RW-dried mango powder or

176

flakes were achieved with the aid of a spreader bar provided in the RW dryer, coupled with a controlled mechanism that allows uniform feeding of mango puree from the inlet section and as it moves toward the other end of the dryer.

The initial structure of mango powder was well preserved after 6 and 12 months of storage at 5 and 22 °C in both air-packaging and nitrogen-flushing (Fig. 9). This result suggests that the flowability of RW-dried mango powder is very stable during storage at ambient temperature and below, with or without air modification inside the package. From an economic point of view, the powder can be well protected even by just air-sealing the package and storing at ambient temperature, eliminating the need for refrigeration. Except for the chemical stability of the present study, these findings are supportive to the physical properties of mango powder at the same storage temperatures. On the other hand, microscopic examination showed that mango powder stored at 45 °C was severely degraded after 6 and 12 months of storage, regardless of the packaging technique used (Fig. 9 ), although the degradation effect of mango powder stored in air-packaging was more intense than in nitrogen-flushed samples. It is clearly shown in Fig. 8 that the original un-agglomerated powder particles were strongly bonded together forming into large assemblages or lumps of particles after 6 and 12 months of storage. It is worthy to note that mango powders stored at 45 °C formed the lump particles as early as 2 months of storage but with lighter color (image not shown). The flaky structure of the powder particles coupled with its fine and smooth surface characteristics may have contributed to the high magnitude of inter- particle forces allowing them to form strong adherence between particles. Likewise, these product lump formations can be associated with partial melting during storage, which enhances surface contact between particles through diffusion of molecules from one particle to another as affected by the elevated temperature (Barbosa-Cánovas et al., 2005).

177

4. CONCLUSIONS

The physical and nutritional changes of air-packaged and nitrogen-flushed RW-dried mango powder varied after 6 and 12 months of storage at 5, 22 and 45 °C. Regardless of packaging atmosphere, the results showed no significant (p < 0.05) increase in water content of mango powder stored at 5 and 22 °C after 6 and 12 months of storage, while there was significant increase (p < 0.05) when the mango powder was stored at 45 °C. Significant reduction in headspace oxygen concentration was found in both air and nitrogen-flushed samples stored at 45 °C after 6 and 12 months storage (p < 0.05). At this high storage temperature, evolution of carbon dioxide was noted inside all packages, which could be associated with the presence of heat resistant spores and decarboxylation of citric acid present in mango.

Changes in yellow color of RW-dried mango powder was not detected at 5 and 22 °C after 6 and 12 months of storage for both types of packaging. However, the mango powder darkened at 45 °C may be due to caramelization, non-enzymatic browning and Maillard reaction.

Consequently, replacing the air headspace with nitrogen preserved the color of reconstituted mango powder during storage of up to 12 months.

The L-ascorbic acid (AA) content of RW-dried mango powders diminished severely when stored at 45 °C over a period of 12 months of storage regardless of the packaging atmosphere. Replacing the air inside the package with nitrogen gas improved the retention of AA in mango powder over a period of 6 and 12 months of storage at room temperature (22 °C) and under refrigeration (5 °C) ranging from 68.2 % to 89.5 %, respectively.

Similarly, the application of high temperature (45 °C) severely degraded the β-carotene content of mango powder when using air-packaging, obtaining a loss of 82.4 % after 6 months of storage, while the effect of nitrogen flushing proved to obtain significantly lower percentage loss

178

with 26.1 %, 21.8 % and 23.4 % at 5, 22 and 45 °C (p < 0.05) respectively, during the same period. Extending the storage to 12 months showed a considerable decline of β-carotene, ranging from 86.5 – 87.1 % and insignificant differences were found compared to samples stored in air-packaging.

By applying the concept of glass transition temperature (Tg), mango powders stored at 5 and 22 °C (< Tg) were well preserved and free from structural transformations, while mango powder stored at 45 °C ( > Tg) suffered intense degradation due to caking, resulting in a hard and brittle assemblage of powder particles. The RW-dried mango powder with water content of

0.983 kg solids/kg mango puree and with aw ≤ 0.32 e hibited g eate stability when st ed at temperature no greater than 23 °C for both air packaging and nitrogen flushing.

The initial structure of mango powder was well preserved after 6 and 12 months of storage when stored at 5 and 22 °C, with or without replacing the air headspace. On the other hand, microscopic examination of mango powder stored at 45 °C showed severe degradation after 6 and 12 months of storage, regardless of the packaging technique applied.

The results of the present study will serve as a good reference in the production and storage of mango powder, particularly those obtained by Refractance Window® drying process.

ACKNOWLEDGEMENTS

We thank the Ford Foundation International Fellowship Program (IFP)/Institute of

International Education (IIE)-New York through the IFP - Philippine Social Science Council

(IFP-PSSC) for providing financial support; the Philippine Center for Postharvest Development and Mechanization (PhilMech) for granting study leave to Ofero Caparino. Special thanks to

Richard E. Magoon and Karin M. Bolland of MCD Technologies, Inc. (Tacoma, WA) for allowing the use of their RW drying facilities and their assistance in doing the experiments; Mr.

179

Clifton Coy of KSM Enterprises, WA for supplying the packaging material; Dr. Boon P. Chew and Bridget Mathison for providing guidance and technical support in developing and analyzing the beta carotene content of mango powder; Ms. Galina Mikhaylenko, Roopesh Syamaladevi,

Scott Mattinson and Dr. Valerie Lynch-Holm for providing assistance in ascorbic acid analysis, glass transition temperature measurements, gas chromatography and FESEM imaging, respectively.

180

REFERENCES

Abdon, I. C., del Rosario, I. F., Aguinaldo, A. R., Lontoc, A. V., & Alejo, L. G. (1990). Food

Composition Tables. Manila: Food and Nutrition Research, Dept of Science and

Technology.

Abonyi, B. I. (2000). Evaluation of Refractrance Window drying method for fruits and

vegetables-MS Thesis. Pullman: Washington State University.

Abonyi, B. I., Feng, B. I., Edwards, C. G., & Tang, J. (2002). Quality retention in strawberry and

carrot purees dried with Refractance Window system. Journal of Food Science , 67,

1051–1056.

Abou-ashour, A., & Edwards Jr, H. (1970). Effect of dietary sterculia foetida oil on pink-white

discoloration and fatty acid distribution in stored eggs. Journal of Nutrition , 100, 757-

766.

Aguilera, J. M., del Valle, J. M., & Karel, M. (1995). Caking phenomena in amorphous food

powders. Trends in Food Science and Technology , 6, 149-155.

Al Mahdi, R., Nasirpour, A., Banon, S., Scher, J., & Desobry, S. (2006). Morphological and

mechanical properties of dried skimmed milk and wheat flour mixtures during storage.

Powder Technology , 163, 145-151.

Ammu, K., Radhakrishna, K., Subramanian, V., & Sharma, T. R. (1977). Storage behaviour of

freeze-dried fruit juice powders. Journal of Food Technology , 12, 541-554.

AOAC. (1998). Official methods of analysis. Washington, DC, USA: Association of Official

Analytical Chemists.

181

Arvanitoyannis, I., Blanshard, J. M., Ablett, S., Izzard, M. J., & & Lillford, P. J. (1993).

Calorimetric study of the glass transition occurring in aqueous-glucose-fructose solutions.

Journal of Food Science and Agriculure , 63, 177-188.

Arya, S. S., Natesan, V., Parihar, D. B., & Vijayaraghavan, P. K. (1979). Stability of carotenoids

in dehydrated carrots. Journal of Food Technology , 14, 579–586.

Aydin, R. (2006). Effect of storage temperature on the quality of eggs from. South African

Journal of Animal Science , 36 (1), 13-19.

Barbosa-Canovas, G. V., Fernandez-Molina, J. J., Alzamora, S. M., Tapia, M. S., Lopez-Malo,

A., & Chanes, J. W. (2003). Handling and preservation of fruits and vegetables by

combined methods for rural areas: Technical manual. Rome: FAO.

Barbosa-Canovas, G., V., Ortega-Rivas, E., Juliano, P., & Yan, H. (2005). Food Powders:

Physical Properties, Processing, and Functionality. New York: Kluwer

Academic/Plenum Publishers.

BAS. (2009). Situation Report on Selected Fruit Crops. Retrieved March 1, 2011, from

http://www.bas.bas.gov

Berg, J., Tymoczko, J. L., & Stryer, L. (2002). Biochemistry. New York: W H Freeman.

Bolland, K. (2000). Refractance Window Drying: A new low temperature, energy efficient

process. 45 (7), 293-296.

Caparino, O. A., Tang, J., Nindo, C. I., Sablani, S. S., Powers, J. R., & Fellman, J. K. (2012).

Effect of Drying Methods on the Physical Properties and Microstructures of. Journal of

Food Engineering , 00 (0), 00 - 00.

Dennison, D. B., & Kirk, J. R. (1978). Oxygen effect on the degradation of ascorbic acid in

dehydrated food system. Journal of ?food Science , 43 (2), 609-618.

182

Desobry, S. A., Netto, F. M., & Labuza, T. P. (1997). Comparison of spray-drying, drum-drying

and freeze-d ying β-carotene encapsulation and preservation. Journal of Food Science

, 62 (6), 1158-1162.

Eggleston, G., & Vercellotti, J. R. (2000). Degradation of Sucrose, Glucose and Fructose in

Concentrated Aqueous Solutions Under Constant pH Conditions at Elevated

Temperature. Journal of Carbohydrate Chemistry , 19 (9), 1305-1318 .

Eison-Perchonok, M. H., & W., D. T. (1982). Kinetics of ascorbic acid autoxidation as a function

of dissolved oxygen concentration and temperature. Journal of Food Science , 47, 765-

768.

Evans, E. A. (2008). Recent Trends in World and U.S. Mango Production Trade, and

Consumption. University fo Florida. Gainesville, Florida: Food and Resource Economics

Department.

FAO. (2007). FAOSTAT database collections, agricultural data, Food and Agriculture

Organization of the United Nations. Retrieved March 10, 2011, from

http://www.faostat.org

Fine, F., & Gervais, P. (2005). Thermal destruction of dried vegetative yeast cells and dried

bacteria spores in a convective hot air flow: Strong influence of initial water actiivty.

Environmental Microbiology , 7, 40-46.

Fitzpatrick, J. J., Hodnett, M., Twomey, M., Carqueira, P. S., O'Flynn, J., & Roos, Y. H. (2007).

Glass transition and the flowability and caking of food powders containing amorphous

lactose. Powder Technology , 178, 119-128.

183

Gonzalez-Aguilar, G. A., Zavaleta-Gatica, R., & Tiznado-Hernandez, M. (2007). Improving

p stha vest quality ang ‘Haden’ by UV-C treatment. Postharvest Biology and

Technology , 45, 108–116.

Haymavathi, T., & Khader, V. (2005). Carotene, ascorbic and sugar content of vacuum

dehydrated ripe mango powders stored in flexible packaging material. Journal of Food

Composition and Analysis , 18, 181–192.

Hymavathi, T., & Khader, V. (2005). Carotene,ascorbic acid and sugar content of vacuum

dehydrated ripe mango powders stored in flexible packaging material. Journal of Food

Composition and Analysis , 18, 181–192.

Jain, N., Krishnamurthy, G., & Lal, G. (1959). Non-volatile organic acids in unripe pickling

mangoes and salted mango slices by paper chromatography. Food Science , 3, 115–118.

Jaya, S., & Das, H. (2004). Effect of maltodextrin, glycerol monostearate and tricalcium

phosphate on vacuum dried mango powder properties. Journal of Food Engineering , 63,

125-134.

Jaya, S., & Das, H. (2009). Glass transition and sticky point temperatures and stability/mobility

doagram of fruit powders. Food Bioprocess Technol , 2, 89-95.

Johnson, J., Braddock, R., & Chen, C. S. (1995). Kinetics of ascorbic acid loss and non-

enzymatic browning in orange juice serum: Experimental rate constants. Journal of Food

Science , 60, 502-504.

Jungalwala, F., & Cama, H. (1963). Carotenoids in mango (Mangifera indica) fruit. Indian J.

Chem. , 1, 36–41.

Kabasakalis, V., Siopidou, D., & Moshatou, E. (2000). Ascorbic acid content of commercial

fruits juices and its rate of loss upon storage. Food Chemistry , 70, 325-328.

184

Kaspar, K. L., Park, J. S., Brown, C. R., Mathison, B. D., Massimino, S., & Chew, B. P. (2011).

Processing of pigmented flesh potatoes (Solanum tuberosum L.) on retention of bioactive

compounds. Int. J. Food Sci. Technology , (in press).

Khachik, F., Beecher, G. R., & Whittaker, N. F. (1986). Separation, identification, and

quantification of the major carotenoid and chlorophyll constituents in extracts of several

green vegetables by liquid chromatography. Journal of Agric. Food Chem , 34, 603-616.

Kholi, K., Qadry, J. S., & Ali, M. (1986). Protein, amino acids and ascorbic acid in some

cultivars of mango. Journal of the Science of Food and Agriculture , 39, 247-252.

Kirk, J., Dennison, D., Kokoczka, P., & Heldman, D. (1977). 1977. Degradation of ascorbic acid

in dehydrated food system. Journal of food Science , 42 (5), 1724-1729.

Labuza, T. P. (1972). Nutrient losses during drying and storage of dehydrated foods. CRC in

Food Technology , 3, 217.

Lavelli, A., Zanoni, B., & Zaniboni, A. (2007). Effect of water activity on carotenoid

degradation in dehydrated carrots. Food Chemistry , 104, 1705–1711.

Lee, S., & Labuza, T. P. (1975). Destruction of ascorbic acid as a function of water activity.

Journal of Food Science , 40, 370-373.

Lloyd, M. A., Hess, S. J., & Drake, M. A. (2009). Effect of nitrogen flushing and storage

temperature on flavor ans shelf-life of whole milk. Journal of Dairy Science , 92, 2409-

2422.

Nindo, C. I., & Tang, J. (2007). Refractance Window dehydration technology: A novel contact

drying method. Drying Technology , 25, 37–48.

185

Nindo, C. I., Feng, H., Shen, G. Q., Tang, J., & Kang, D. H. (2003b). Energy utilization and

microbial reduction in a new film drying system. Journal of Food Processing &

Preservation , 27, 117-136.

Nindo, C. I., Sun, T., Wang, S. W., Tang, J., & Powers, J. R. (2003a). Evaluation of drying

technologies for retention of physical and chemical quality of green asparagus

(Asparagus officinalis, L.). Food Sci. Technol. (LWT) , 36 (5), 507–516.

Owusu-Apaten, R. (2005). Introduction to Food Chemistry. Boca Raton, FL: CRC Press.

Palzer, S., & Sommer, K. (2001). Caking of water-soluble amorphous and crystaline food

powders. In J. M. Aguilera, J. Simpson, D. Welti-Chanes, D. Bermudez-Aguirre, & G.

Barbosa-Canovas, Food Engineering Interfaces (p. 680). Springer.

Peleg, M. (1993). Glass transition of the physical stability of food powders. In J. Blanshard, & P.

Lillford, The glassy state in foods (pp. 435-451). Leicestershire: Nottingham University

Press.

Potter, N. N., & Hotchkiss, J. H. (1995). Food Science (5th ed.). New York: Springer.

Quintas, M. A., Brandao, T. R., & Silva, C. L. (2007). Modelling colour changes during the

caramelisation reaction. Journal of Food Engineering , 83 (4), 483–491.

Rajkumar, P., Kailappan, R., Viswanathan, R., Raghavan, G. S., & Ratti, C. (2007). Foam mat

drying of alphonso mango pulp. Drying Technology , 25, 357-365.

Robertson, G. L., & Samaniego, C. M. (1986). Effect of initial dissolved oxygen levels on the

degradation of ascorbic acid and the browning of lemon juice during storage. Journal of

Food Science , 51 (1), 184-187.

Roos, Y. H., & Karel, M. (1991c). Phase transitions of mixtures of amorphous polysaccharides

and sugars. Biotechnology Progress , 7, 49-53.

186

Roos, Y. H., & Karel, M. (1991b). Plasticizing effect of water on thermal behaviour and

crystallization of amorphous food models. Hournal of Food Science , 56, 38-43.

Rudell, D. R., Mattinson, D. S., & Fellman, J. K. (2000). The progression of ethylene production

and respiration in the tissues of ripening 'Fuji' apple fruit. HortScience , 35 (7), 1300-

1303.

Sablani, S. S., Syamaladevi, R. M., & Swanson, B. G. (2010). A review of methods, data and

applications of state diagram of food systems. Food Eng Rev , 2, 168-203.

Sagar, V. R., Khurdiya, D. S., & Maini, S. B. (2000). Quality of mango powder as affected by

storage temperature and period. Journal of Food Science & Technology , 37 (2), 165-168.

Serrano, E. P. (2005). Postharvest management of fruit and vegetables in the Asia-Pacific region:

Philippines country report. Asian Productivity Organization (APO); Food and

Agriculture Organization of the United Nations (FAO) (pp. 216-226). Japan; Rome: APO

and FAO.

Stefanovich, A. F., & Karel, M. (1982). Kinetics of b-carotene degradation at temperatures

typical of air drying of foods. Journal of Processing and Preservation , 6, 227-242.

Syamaladevi, R. M., Sablani, S. S., Tang, J., Powers, J., & Swanson, B. G. (2009). State diagram

and water adsorption isotherm of raspberry (Rubus Idaeus). Journal of Food Engineering

, 460-467.

Talcott, S., & Howard, L. R. (1999). Phenolic Autoxidation Is Responsible for Color

Degradation in Processed Carrot Puree. Journal of Agricultual Food Chemistry , 47,

2109-2115.

Tang, Y. C., & Chen, B. H. (2000). Pigment change of freeze-dried carotenoid powder during

storage. Food Chemistry , 69, 11-17.

187

Tha anathan, R. N., Yash da, H., & P abha, T. (2006). Mang (Mangi e a indica L.), “The King

of Fruits" - An Overview. Food Reviews International , 22:95–123.

Toledo, R. T. (2007). Fundamentals of Food Process Engineering (3 ed.). Athens, GA: Springer

Science+Business Media, LLC.

Trammell, D. J., Dalsis, D. E., & Malone, C. T. (1986). Effect of oxygen on taste, ascorbic acid

loss and browning of HTST pasteurized single strength orange juice. Journal of Food

Science , 51 (4), 1400-1408.

Uddin, M. S., Hawlader, M. N., Ding, L., & Mujumdar, A. S. (2002). Degradation of ascorbic

acid in dried guava during storage. Journal of Food Engineering , 51, 21-26.

Wallack, D. A., & King, C. J. (1988). Sticking and agglomeration of hygroscopic, amorphous

carbohydrate and food powders. Biotechnology Progress , 4 (1), 31-35.

Zhang, Q. H., Qiu, X., & Sharma, S. K. (2002). Improving safety and quality of orange juice by

pulse electric field processing. In X. Yang, & J. Tang, Advances in Bioprocess

Engineering (Vol. 1, pp. 45-67). Singapore: World Scientific Publishing Co. Pte. Ltd.

188

Table 1. Water content and water activity of stored mango powder after 6 and 12 months of storage at different packaging methods and temperatures. After 6 months Parameter Initial value Air Packaging Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C Water Content 0.041 ± 0.040 ± 0.046 ± 0.039 ± 0.043 ± (Kg water/kg dry 0.037 ± 0.006a 0.038 ± 0.001a 0.000acf 0.002ac 0.000bcg 0.002ad 0.001cdeh solids) 0.154 ± 0.159 ± 0.190 ± 0.157 ± 0.155 ± 0.171 ± Water Activity 0.126 ± 0.003a 0.000b 0.000b 0.000d 0.000b 0.000b 0.000h

189

After 12 months Parameter Initial value Air Packaging Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C Water Content 0.041 ± 0.041 ± 0.049 ± 0.039 ± 0.039 ± 0.045 ± (Kg water/kg dry 0.037 ± 0.006a 0.001acf 0.000acf 0.001bg 0.000ad 0.001ae 0.001fgh solids) 0.243 ± 0.241 ± 0.290 ± 0.226 ± 0.234 ± 0.279 ± Water Activity 0.126 ± 0.003a 0.001c 0.004c 0.003e 0.005fg 0.003cg 0.003i

a-i Means with the same superscript letters or numbers within a row across 6 and 12 months of storage indicate no significant differences (p < 0.05). Means were obtained from 3 replicates.

189

Table 2. Headspace gas concentration of RW-dried mango flakes after 6 and 12 months of storage, % by volume. After 6 months After 6 months Components Initial value Air Packaging Initial value Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C a a a b 1 1 1 2 CO2 0 0 0 6.06 ± 0.08 0 0 0 2.8 ± 0.01 a bcde de f 1 2 2 2 O2 19.8 ± 0.06 19.0 ± 0.17 18.9 ± 0.13 14.4 ± 0.09 2.4 ± 0.01 2.7 ± 0.01 2.8 ± 0.01 2.7 ± 0.01 a bc c d 1 2 2 4 N2 80.2 ± 0.06 81.0 ± 0.17 81.1 ± 0.13 79.6 ± 0.09 97.7 ± 0.01 97.3 ± 0.01 97.2 ± 0.01 94.6 ± 0.01

190 After 12 months After 12 months

Components Initial value Air Packaging Initial value Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C a a a c 1 1 1 3 CO2 0 0 0 10.9 ± 0.028 0 0 0 4.1 ± 0.03 a ce e g 1 3 3 3 O2 19.8 ± 0.06 19.4 ± 0.07 19.0 ± 0.16 12.8 ± 0.03 2.4 ± 0.01 3.7 ± 0.15 3.6 ± 0.04 3.7 ± 0.04 a ab cb e 1 3 3 5 N2 80.2 ± 0.06 80.6 ± 0.17 81 ± 0.06 76.3 ± 0.03 97.7 ± 0.01 96.3 ± 0.15 96.4 ± 0.04 92.2 ± 0.04

a-f & 1-5 Means with the same superscript letters or numbers within a row across 6 and 12 months of storage indicate no significant differences (p < 0.05). Means were obtained from 3 replicates.

190

Table 3. Hunter color measurements of mango flakes or powders stored in air or nitrogen atmosphere at 5, 22 and 45°C and evaluated after 6 and 12 months. After 6 months Hunter Color Initial value Air Packaging Nitrogen Flushing parameters 5°C 22°C 45°C 5°C 22°C 45°C L 63.68 ± 0.85a 64.06 ± 0.05a 64.26 ± 0.18a 48.90 ± 0.13c 63.63 ± 0.08a 62.47 ± 0.07a 49.17 ± 0.57e a* 6.54 ± 0.22a 4.79 ± 0.06b 5.06 ± 0.08b 6.70 ± 0.08a 6.31 ± 0.07ac 6.26 ± 0.14ac 8.29 ± 0.13f b* 32.91 ± 0.97a 28.82 ± 0.25b 28.25 ± 0.14b 19.47 ± 0.61c 35.64 ± 0.18e 31.29 ± 1.324a 24.69 ± 0.25f Chroma 33.56 ± 0.98aefh 29.22 ± 0.38bj 28.69 ± 0.53bj 22.95 ± 0.21c 36.19 ± 0.19dg 31.91 ± 1.41fh 26.04 ± 1.00ik Hue 78.74 ± 0.23a 80.55 ± 0.06b 79.84 ± 0.13b 73.02 ± 0.09c 79.95 ± 0.05b 78.66 ± 0.36a 71.44 ± 0.43e

191 After 12 months Components Initial value Air Packaging Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C L 63.68 ± 0.85a 67.37 ± 0.59b 65.09 ± 0.11a 40.19 ± 0.10d 64.33 ± 0.72a 64.47 ± 0.59a 42.73 ± 0.81f a* 6.54 ± 0. 22a 5.82 ± 0.29c 5.95 ± 0.21c 9.09 ± 0.25d 7.46 ± 0.25e 7.36 ± 0.20e 10.61 ± 0.24g b* 32.91 ± 0.97a 32.18 ± 0.43a 28.82 ± 0.50b 21.95 ± 0.20c 35.93 ± 0.77e 34.30 ± 1.36ae 25.24 ± 0.99f Chroma 33.56 ± 0.98a 32.71 ± 0.38af 29.43 ± 0.14bfj 21.48 ± 0.65c 35.93 ± 0.77ad 35.08 ± 1.32ag 27.38 ± 0.20jk Hue 78.74 ± 0.23a 79.74 ± 0.62b 78.34 ± 0.27a 64.97 ± 0.35d 78.26 ± 0.15a 77.87 ± 0.67a 67.19 ± 0.34f

a-k Means with the same superscript letters within a row from initial color values and after 6 and 12 months of storage indicate no significant differences (p < 0.05). Means were obtained from 6 readings taken from each of 5 locations per sample.

191

Table 4. Hunter color measurements of reconstituted mango powders stored in air or nitrogen atmosphere at at 5, 22 and 45°C and evaluated after 6 and 12 months. After 6 months Components Initial value Air Packaging Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C L 43.97 ± 0.04a 46.89 ± 0.11b 45.90 ± 0.22c 38.68 ± 0.06e 45.69 ± 0.18c 46.20 ± 0.02h 39.44 ± 0.03j a* 4.23 ± 0.04a 2.50 ± 0.06b 3.49 ± 0.05d 5.77 ± 0.01f 4.69 ± 0.05h 4.36 ± 0.01a 6.84 ± 0.24j b* 39.53 ± 0.22a 34.68 ± 0.08bc 33.74 ± 0.17d 27.49 ± 0.12f 39.20 ± 0.23g 41.74 ± 0.12i 32.48 ± 0.49j Chroma 39.75 ± 0.21a 34.77 ± 0.08b 33.92 ± 0.17b 29.31 ± 0.11d 38.57 ± 0.22f 40.75 ± 0.12h 33.92 ± 0.52i Hue Angle 83.89 ± 0.09a 85.88 ± 0.09b 77.89 ± 0.10a 77.89 ± 0.05e 83.01 ± 0.11g 83.86 ± 0.03a 77.60 ± 0.26e E 0.00a 5.93 ± 0.12b 6.15 ± 0.10b 13.79 ± 0.13d 2.19 ± 0.06fgh 1.87 ± 0.10g 9.98 ± 0.35i

192

After 12 months Components Initial value Air Packaging Nitrogen Flushing 5°C 22°C 45°C 5°C 22°C 45°C L 43.97 ± 0.04a 45.86 ± 0.11c 45.24 ± 0.09d 28.06 ± 0.07j 44.53 ± 0.10gi 44.39 ± 0.11i 30.60 ± 0.06k a* 4.23 ± 0.04a 2.74 ± 0.08c 3.72 ± 0.03e 10.18 ± 0.04g 5.99 ± 0.02i 5.60 ± 0.04f 11.45 ± 0.02k b* 39.53 ± 0.22a 34.03 ± 0.22cd 31.04 ± 0.07e 26.88 ± 0.13f 38.29 ± 0.77h 40.52 ± 0.42a 31.04 ± 0.33bcd Chroma 39.75 ± 0.21a 34.14 ± 0.21b 32.69 ± 0.8c 27.49 ± 0.19e 39.66 ± 0.76g 42.11 ± 0.41a 32.69 ± 0.32j Hue Angle 83.89 ± 0.09a 85.40 ± 0.16c 83.47 ± 0.03d 69.67 ± 0.05f 82.37 ± 0.08h 81.31 ± 0.12i 71.61 ± 0.13j E 0.00a 5.99 ± 0.16b 7.18 ± 0.07c 20.82 ± 0.09e 2.70 ± 0.61fh 2.45 ± 0.06h 16.02 ± 0.10j

192

Figure 1. General procedure followed during storage of Refractance Window®-dried mango powder.

193

5 °C 22 °C 45 °C 5 °C 22 °C 45 °C

Nitrogen-flushed packaging Air-packaging

Figure 2. Nitrogen-flushed packaging and air-packaging of mango powder before storage

Figure 3. Collection of gas sample from headspace

194

Figure 4. General procedure during analysis of β-carotene of mango powder

195

)

(

(a)

dried mango powder (a) and reconstituted mango puree (b) after 6 and (b) and 6 months 12 after mango of powder and puree reconstituted mango dried (a)

. Photographs of RW of Photographs .

Figure Figure packaging different atmosphere. at storage

196

Figure 6. Influence of storage time, temperature and packaging atmosphere on the ascorbic acid content of RW-dried mango powder. a-h Different letters indicate significant differences (P < 0.05).

197

Figure 7. In luence st age ti e, te pe atu e and packaging at sphe e n the β-carotene content of RW-dried mango powder. a-f Different letters indicate significant differences (P < 0.05).

198

Month 0 Month 6 Month 12

Air packaging, 45 45 packaging, Air

Nitrogen flushing, 45 45 flushing, Nitrogen

Figure 8. Caking of RW-dried mango powder after 6 and 12 months of storage at 45°C.

199

2 2 μ 0 )

Field emission scanning electron micrographs (FESEM) of Refractance Refractance Window®(180 (FESEM) of electron micrographs scanning Fieldemission

Figure Figure and 100x (magnification conditions of different kV). 30 500x, storage year 1 at for stored

200

CHAPTER SIX

RHEOLOGICAL MEASUREMENTS FOR CHARACTERIZING STICKY POINT

TEMPERATURE OF FRUIT POWDERS: AN EXPERIMENTAL INVESTIGATION

ABSTRACT

Stickiness is one of the common problems frequently encountered during production, handling and storage of fruit powders with high concentration of low molecular weight sugars.

Several techniques and devices were developed in the past to determine the level of stickiness of some of those products. Nevertheless, there is still a need for a simple, more accurate and reliable method. In this study, a new method to quantify and characterize the sticky point temperature (Ts) of fruit powder was explored using a rheometer technique. The rheometer system utilized a serrated parallel plate to hold the samples and was operated in dynamic oscillation mode at a frequency of 1 Hz and a constant strain amplitude of 2 %. The samples of a model fruit powder (Refractance Window (RW)-dried mango powder) were scanned from 25 to

95 °C at an increment of 10 °C and a holding time of 180 s for each increment. A crossover between the storage modulus (G') and loss modulus (G") of a powder product was established and was denoted as sticky point temperature of the model fruit powder. Results showed that the sticky point temperature obtained using the new method agreed well with the published data and can be considered as a suitable technique to characterize the sticky point temperature of sugar- rich materials. The procedure for sample conditioning and rheometric measurements to determine the sticky point temperature are straightforward. This new technique can measure the sticky point temperature of fruit powders with a high degree of repeatability and accuracy

(SD 0. 8 1.73 °C).

201

1. INTRODUCTION

In recent years, many researchers have carried out studies on stickiness of food powders because of its wide practical applications to the food industry particularly during drying, handling and storage. However, one common problem that is frequently encountered in food powder processing, particularly those powders with a high concentration of low molecular weight sugars, e.g. glucose, fructose and sucrose, is stickiness (Dowton et al., 1982; Bhandari et al., 1997). These materials tend to form hard lumps or cakes, and they become less porous and non-flowable when they are exposed to high temperature or humidity.

Stickiness is a surface property which, in a practical sense, can be felt "by hand," thus allowing one to easily recognize the sticky characteristics of a certain product. Schubert (1987) reported that the mechanisms that may influence the tendency of particles to stick together include liquid bridging, solid bridging, inter-particle attraction forces and mechanical attraction.

Liquid bridging happens when a sufficient amount of moisture is present between particles, while solid bridging is a result of solid diffusion or condensation within the solid matrix, normally at elevated temperatures (Barbosa-Cánovas et al., 2005). The appearance of liquid bridges between particles allows for the transformation of the material from the glassy to rubbery phase and greatly influences the strength of the powder product, causing it to be sticky (Ozkan et al., 2002).

Stickiness of food powders has also been described as the sticking between particles or

"cohesion" and sticking of particles to a wall surface, or "adhesion" (Papadakis & Bahu, 1992).

Cohesion between similar particles or surfaces and adhesion between dissimilar particles or surfaces is influenced by particle size. It has been reported that particles less than 200 m are

202

susceptible to cohesion, causing difficulty for the powders to flow, while particles of sizes larger than 300 m are more flowable (Teunou et al., 1999).

Stickiness is also influenced by increases in temperature, which affects flow and deformation properties. For instance, the viscosity of an amorphous food material can dramatically drop from 1012-14 106-8 Pa.s when exposed to higher temperature, creating stickiness of the material (Wallack & King, 1988). Roos (1993) reported that the sticky point temperature of sugars at Ts-Tg is 10 20°C higher than the glass transition temperature (Tg) of a product. Roos & Karel (1991) also reported that if the temperature of a glassy amorphous material exceeds the glass transition temperature, such material is transformed into a rubbery state and may become sticky. These results imply that stickiness can be prevented if the temperature of amorphous material is below glass transition temperature. Previous experiments on tomato, pineapple and mango powders (Jaya, 2009), and milk powder (Hennings et al., 2001) showed that the sticky point temperature and glass transition temperatures do not show the same values, but clearly revealed a constant range of differences between the two temperatures at water content. Stickiness in amorphous form of materials is also due to water plasticization and subsequent depression of glass transition temperature during processing and storage (Goula et al., 2007).

Rao (1999) reported that the transformation of glassy amorphous material to rubbery state was associated with the storage modulus (G') and loss modulus (G”). The G' value is a measure of deformation energy stored in the samples during the shear process (elastic behavior), while

G” is a measure of dissipation of energy (viscous behavior) (Mezger, 2002). It has been reported that when the material was in a glassy solid form, the value of G' was expected to be higher than the G”, while the materials are transformed into rubbery or "liquid-like" material when the G”

203

exceeded the G' (Rao, 1999). Using the mechanism of liquid-bridging related to elasticity and viscous behavior of the material at the interface, the observed values wherein G' and G" are crossing at certain temperature can also be characterized as sticky point temperature of powder as a function of water activity and water content.

Several direct techniques have been proposed to characterize stickiness of food products, as reviewed by Adhikari et al. (2001), Boonyai et al. (2004), and Chen & Ozkan, (2007). Some of those techniques are presented in Fig. 1. The oldest published method is the manually-driven propeller method initially developed by Lazar et al. (1956) to measure the sticky-point temperature of spray-dried tomato powder, and which was later used to test dairy products, orange, mango and other powders (Figure 1a). The instrument consisted of a test tube containing a sample of known moisture content submerged in a water bath. An impeller embedded in the sample is used to manually stir the sample intermittently. A sticky point temperature is recorded when the force required to rotate the impeller dramatically increases (Dowton et al., 1982;

Wallack & King, 1998; Jaya & Das, 2009). However, this method was physically demanding, and yielded non-reproducible, and inaccurate results due to moisture evaporation during testing at elevated temperature was a problem. In addition, this method heavily relied on the personal judgment of the user (Brennan et al., 1971).

Brennan et al. (1971) modified the Lazar design by adding a motor-driven impeller, which was further improved by Hennings et al. (2001) by connecting the impeller to a data logging system to record electrical resistance output (Figure 1b). The reading of the sticky point temperature is indicated when the electrical current drawn by the system sharply increased.

Another method for measuring stickiness was proposed by Okzan et al., (2002) by using a viscometry technique, with an L-shape spindle inserted into the powder product (Figure 1c). The

204

torque required to rotate the spindle or propeller at a given temperature is recorded using a data logging system that is connected to the viscometer. Sticky point temperature is determined when a sharp increase in torque is observed. Boonyai et al. (2004) commented that the disadvantage of using this method is the relatively large quantity of material required for measurement (52 g), and possible variable packing densities, which could affect the torque measurement.

Kudra (2002) reported a semi-automatic sticky-point tester with humidity control (Fig.

1d). It was claimed that the device can prevent excessive moisture evaporation and condensation from the product during testing, and provide accurate results. The use of an optical probe was explored by Lockman (1999) (Fig. 1e). The device consists of a glass test tube containing a sample with known moisture content, mounted to a driver rotating at approximately 50 rpm and immersed in a temperature controlled oil bath at an angle of about 30° (Boonyai et al., 2004).

Temperature of the sample is recorded by a resistance thermometer, while the sample motion is observed and recorded using a fiber-optic sensor. The sensor records reflection of the light directed from the illuminator towards the sample through the integrated light tube. The temperature and the sensor signal are recorded using a computer-based data logger. Changes in reflectance property of a given material are monitored to determine the sticky point temperature of the product.

Adhikari et al. (2003) developed an in situ method for measuring sticky point temperature. This method was principally designed to measure sticky point temperature during drying of a single droplet of low molecular weight sugars. As illustrated in Fig. 1f, the instrument consists of a drying chamber, an image acquisition system and a weighing balance connected to a data logging device. Measurements of surface sticky point temperature was carried out by mechanically raising the samples until a good contact of droplet with the probe

205

was observed and then the probe was withdrawn at the same speed. Images of bonding and de- bonding taken by the video camera and the maximum tensile force measured during withdrawal of probe were recorded. The obtained data were used to analyze the sticky point temperature of the sample (Boonyai et al., 2004). One disadvantage of this method is that a custom-made device is not readily accessible to other laboratories.

There are other methods reported in the literature, such as the Jenike shear cell method

(Jenike, 1964), the ampule method (Tsourouflis et al., 1976), the fluidization test (Dixon et al.,

1999), the blow test (Peterson et al., 2001), the cyclone sticky test (Boonyai et al., 2004), the tack method (Green, 1941), and the contact probe test method (Kilscast & Roberts, 1998; Adhikari et al., 2001; Adhikari et al., 2003).

A critical analysis conducted by Boonyai, Bhandari and Howes (2004) suggests that the above measuring devices and techniques are all empirical in nature. They also perceived that due to the inaccuracy and difficulty of application of these devices to actual processing and handling operations, there are continuing demands and challenges to develop accurate, simple and easy to use methods to characterize the onset of stickiness of food powders. Thus, the objectives of this study were to: 1) develop a new method with high repeatability and reliability to characterize the sticky point temperature of a model fruit (RW-dried mango powder) using a rheometer; 2) find the relationships between the obtained sticky point temperatures and glass transition temperature of the model powder at different water contents; and 3) compare the obtained sticky point temperature using the new method with the existing data reported in literature.

206

2. MATERIALS and METHODS

2.1 Preparation of samples

Refractance Window® (RW)-dried mango powder was chosen as the model sample for the entire experiment. The powder with water content of 0.039 kg water/kg dry solids was produced using the RW drying process following the procedure described by Caparino et al.

(2012). One hundred grams of RW-d ied ang p wde with pa ticle si e anging 180

350 µm was prepared by slight grinding using a mortar and pestle, and sieving using mesh sizes of 80 and 45 (American Society for Testing and Materials, ASTM) (Barbosa-Cánovas, 2005).

The selected range of particle sizes was based on the approximate size of the serration of the two parallel plate geometries used during the sticky point measurements. The preliminary measurements of sticky point temperature using the new procedure as described in the results and discussion revealed no significant effect of particles sizes at 180 and 500 µm. The prepared powder samples were put inside aluminum-coated polyethylene bags, flushed with nitrogen gas, heat sealed and kept at 35 °C for use in later measurements.

2.2. Conditioning of samples at different water content

Overall, samples with six different water contents of 0.003, 0.022, 0.029, 0.039, 0.048,

0.066 kg water/kg dry solids were prepared and used in the experiment. The conditioning of samples was carried out by drying or by water absorption methods. The prepared RW-dried mango powders with water content of 0.039 kg water/kg dry solids served as the starting reference sample prior to the application of the conditioning procedures. The samples with water content below 0.039 kg water/kg dry solids were obtained by drying the reference samples, while samples above 0.039 kg water/kg dry solids were prepared by conditioning them in thin-layers under 100% relative humidity at different times. Specifically, the sample with water content of

207

0.003 kg water/kg dry solids was achieved by putting 5 g of the reference sample in a sealed jar containing P2O5 solution and storing it for 30 days at room temperature.

The samples with water content of 0.022 and 0.029 kg water/kg dry solids were obtained by placing 2 g of the reference sample in an aluminum pan with approximate thickness of 1 mm and oven-dried at 50 °C for 30 and 60 min, respectively. To obtain samples with higher water contents of 0.048 and 0.067 kg water/kg dry solids, the reference samples were held for 20 min in a sealed dessicator (Polylab® Plasticware, Hyderabad, Andhra Pradesh, India) above 500 ml of dionized water and held for 20 and 40 min at room temperature, respectively (Fig. 2). The relative humidity inside the dessicator was assumed to be ~100%. To obtain uniform moisture adsorption during conditioning at ~100% RH, an aluminum sample holder with a dimension of 2 x 2 inch and thickness of 2 mm was fabricated (Fig. 3). The size of the holder allows for spreading the samples evenly in a 90 mm diameter filter paper # 2 (WhatmanTM, NJ), obtaining a

2 mm thickness of samples before loading into the dessicator. Apparently, the powder formed a lump of particles or agglomerated after conditioning at ~100% RH, and hence they were dispersed quickly using a stainless spatula, sealed in a small bottle and kept for at least 2 h inside a dessicator to equilibrate prior to measurement and analysis. Samples with water content greater than 0.067 kg water/kg dry solids were not included in the measurement because the material turned very sticky at room temperature and they were no longer practical to use for the study. We note here that some researchers were able to measure the sticky point temperature of a material up to 0.08 kg water/kg dry solids for higher molecular weight samples such as mango, orange and pineapple powders with added maltodextrin carrier (Jaya & Das, 2009).

208

2.3. Determination of water content

Water content of mango powder samples was determined using an automatic Karl Fischer titrator (Mettler Toledo, Columbus, OH) following the procedure described in the instrument manual. The water content of the sample was determined by adding approximately 1 g of mango powder into the titration vessel through a funnel with the aid of a glossy paper. The exact weight of the sample was measured to 0.1 mg accuracy by weighing the funnel before and after the sample was delivered to the titration vessel. The water content of the samples was determined by a built-in software program of the instrument. Measurements were performed in triplicate at room temperature.

2.4. Characterization of sticky point temperature using the rheological method

2.4.1 Rheometer system

The sticky point temperature measurement was performed using a controlled stress rheometer (AR2000 Advanced Rheometer, TA Instruments, New Castle, DE). The rheometer system was fitted with a lower and upper stainless serrated plate with a diameter of 65 mm and

20 mm, respectively (Fig. 4). The lower plate was fixed directly above the Peltier plate (the temperature heating system that controls the temperature of the samples), while the upper plate was free to oscillate with no supplemental heating. The serrated 'groove' parallel plate geometry was selected as a design criterion to prevent slippage between particles and the parallel plate during oscillation, and to minimize moisture evaporation from the samples during the experiment. A roughened or serrated surface plate was reported to be effective in preventing slippage in oscillatory testing of cookies and crackers dough (Menjivar & Faridi, 1994). The temperature and moduli responses of the samples were continuously monitored via the rheometer sensors that were connected to a computer equipped with proprietary monitoring software.

209

2.4.2. Fabrication of powder loading device (PLD)

Proper loading of samples was critical in testing new methods for characterizing the sticky point temperature of the samples. To allow uniformity and consistency on the thickness and quantity of the loaded samples for rheological measurements, a simple circular powder loading device (PLD) made from polycarbonate plastic material was fabricated (Fig. 5). With a thickness of 1 mm, the PLD has an outside diameter of 65 mm and a central hole with a diameter of 20 mm designed to fit into the standard serrated lower and upper plates of the rheometer used in the experiment. A scraper made from the same material as the PLD was also fabricated.

2.4.3. Loading of samples

The fabricated PLD was positioned at the center of the serrated lower plate of the rheometer. Approximately 300 mg of mango powder samples with particle sizes ranging from

180 350 µm was poured to fill the circular hole of the PLD, using a stainless spatula, followed by carefully scraping the excess samples (without tapping) with the aid of a straight-fine edge cut polycarbonate material. After scraping, the PLD was removed by first holding the side portion of the loader while gently lifting the other side in a slightly inclined motion, leaving a circular- molded shape of the powders with a dimension of 20 mm diameter and 1 mm thickness. The diameter and thickness of the loaded powder samples was designed to match with the diameter of the serrated upper plate and the set gap of 1mm between the two parallel plates, respectively.

2.4.4. Rheological settings and measurements

Rheometric measurement was carried out immediately after loading the samples to minimize changes of the sample moisture. The rheometer was operated in dynamic oscillation mode with a 1 mm gap between the parallel plate geometries (upper and lower plate). A time

210

sweep dynamic test for the model sample was applied at a frequency of 1 Hz and a constant strain amplitude of 2%. Samples with different water content were exposed to varying temperatures by ramping from 25 to 95 °C in steps of 10 °C. This temperature step procedure was adopted because our preliminary experiment on continuous scanning at 3, 5 and 10 C/min of powder samples resulted in a delayed temperature response and hence produced a non- reproducible output. To ensure that the temperature readings of samples represented their actual temperature, the samples were held constant for 180 s at each temperature, generating 20 data points before ramping to the next higher temperature level. We did not carry out an actual measurement on the temperature variation within the loaded bulk powders due to difficulty of inserting temperature sensors inside a 1 mm gap between parallel plates. However, considering that the thickness of the powders was only 1 mm, and the estimated amount of powders lodged in each serration of the two parallel plates was less than 2 mg, the temperature gradient in all sections of the loaded samples during heating was assumed to be insignificant. Also, due to the very high heat capacity and heat transfer coefficient of the serrated stainless steel plates, the temperatures recorded in this plate were assumed to be the same as the actual temperature of the samples.

Two sticky point temperatures measured in the proposed method were denoted as the end-point sticky point temperature (Tse) and average sticky point temperature (Tsa). Tse for each temperature setting were obtained from the end-point value during a holding time of 180 s during ramping, while Tsa was taken from the average of 20 data points recorded during the holding time period. The measured sticky point temperatures at different water content were used to establish a ‘‘sticky-p int cu ve’". All easu e ents we e pe ed in t iplicate at temperature.

211

2.5. Measurement of glass transition temperature

Previous studies comparing sticky point temperature and glass transition temperature (Tg) of sugar-rich materials showed a linear direct relationship with almost parallel trend (Ts Tg) at different water content (Jaya & Das, 2009; Downton et al., 1982). Thus, determination of Tg for the current investigation was also carried out to confirm this relationship. The Tg of RW-dried mango powders with different water content levels (0.003, 0.022., 0.029, 0.039, 0.048 and 0.067 kg water/kg dry solids) was measured using a differential scanning calorimeter (DSC, Q2000,

TA Instruments, New Castle, USA), following the procedure described by Syamaladevi et al.

(2009). The calorimeter was calibrated for heat flow and temperature using standard indium and sapphire. Twelve to fourteen milligrams of each mango powder sample was sealed in an alu inu pan (v lu e 30μl), c led d wn 2 t – 90 °C using liquid nitrogen and then equilibrated for 10 min. The samples at – 90 °C were scanned to 90 °C and then cooled down to 25 °C. Scanning of all samples was carried out at the same heating and cooling rate at 5

°C/min. To avoid condensation on the surface of the powder particles, a nitrogen carrier gas was purge at a flow rate of 50 ml/min. The onset- (Tgi), mid- (Tgm) and end-point (Tge) values of the mango powders was determined by finding the vertical shift in the heat flow-temperature diagram. All measurements were performed in duplicate.

2.6. Comparison of the sticky-point temperature in the present study with published literature

The obtained sticky-point temperature curves of RW-dried mango powder at different water content using the proposed method were compared with reported sticky-point temperatures of sugar-rich powders published in the literature, such as mango powder (Jaya & Das, 2009); orange juice powder (Brennan et al., 1971); and tomato juice powder (Lazar et al., 1956). The

212

samples with water content ranging from 0.003 to 0.067 kg water/kg dry solids were plotted and superimposed with the existing published data.

3.0. RESULTS and DISCUSSION

3.1 Fundamental basis for characterizing sticky point temperature using the rheological method

The assumption in characterizing sticky point temperature of the material focused on particle-to-particle stickiness or 'cohesion' and was independent of the types of wall surfaces, e.g. particle-to-wall stickiness or 'adhesion'. The parallel serrated 'groove surface' geometry used in the experiment was considered as a design criterion for the measurements to prevent any mobility at the interface between the plates and the loaded samples. In such conditions, it was assumed that resistance to particle mobility happens only in the interface between the surfaces of the individual particles.

When applying the proposed rheological method to characterize sticky point temperature, we considered the powder sample as a viscoelastic material by means of the measured storage modulus (G') and loss modulus (G"). The presupposition to determine viscoelastic response for food powders where G' is higher than G" (solid-like) was adopted as a valid reciprocal of the viscoelastic response in food gel where G" is higher than G' (liquid-like), as reported by Tabilo-

Munizaga & Barbosa-Cánovas (2005). Fig. 6a to 6h illustrate the relationships between the G’ and the G”, and the proposed sticking mechanism as the material is exposed to varying temperatures. The intersection or crossover that may be developed between G' and G" over a range of temperatures was proposed to be the criterion in characterizing the sticky point temperature of the material. It is noted here that the scanning electron micrographs of RW-dried

213

mango powder shown in Fig. 6f,g,h provide visual illustration only and do not necessarily represent the actual image taken during the conduct of experiment.

As presented in Fig. 6, we postulated that when the model powder was in its glassy/dry amorphous state, the surfaces of individual particles do not cohere with each other, and hence the materials are free flowing. It can be seen that at the onset of the measurement (a), the values of

G' of the powder sample was higher than G" or the material behaves like a solid, indicating a negligible liquid-bridging or inter-particle forces. The dominance of G’ against G” was due to high surface viscosity of sugar-rich material such as mango powder in its glassy form. This assumption is supported in the study of Downton et al. (1982), who reported that fruit powders containing high sugars have very high viscosity at the glassy stage. White & Cakebread (1966) also described that amorphous solids are often considered to be "metastable supercooled liquids" below their glass transition temperature with extremely high viscosity in the order of 1012 Pa.s.

When the particles are exposed to higher temperature, their surfaces start to plasticize (b)

[thermal plasticization, Roos (2003)], while the viscosity gradient between G’ and G” decreases.

When the solid sample particles are further heated up, they attract from each other and behave more like a liquid "rubbery" with lower viscosity (Downton et al., 1982) until a certain temperature where the two moduli intersect with each other, creating a crossover. The lowering of viscosity at this stage allows for a liquid bridging between particles, and hence the powder starts to become sticky and to cohere or adhere to each other (c). When advancing the temperature above the crossover point, a reversal of viscoelastic response is observed wherein the G” becomes dominant over G’, suggesting a steady sticky condition of the sample (d & e).

Further increase in temperature may result in loss of the stability of particle shape and structure, resulting in cake formation and, finally, in melting of the material. This sticking mechanism that

214

is being postulated is similar to the one reported by other authors based on the expanded caking mechanism as described by Peleg (1983) and Palzer & Sommer (2011).

Other sticking observations may support further explanations about the above presumptions. The sticking phenomena is attributed to inter-particle attraction as affected by moisture and temperature, either by electrostatic force, liquid bridges, solid bridges or mechanical interlocking (Adhikari et al., 2001). Downton et al. (1982) reported a mechanistic model suggesting that when two moistened particles come in contact, they build a bridge between them that is sufficient to overcome subsequent mechanical deformation leading to reduction of viscosity and hence stick to one another. Specifically, when an initial dry amorphous particle is exposed to higher temperature, its flowability diminishes due to a drastic lowering of viscosity (Wallack & King, 1988). Also, due to sensitivity of the amorphous sample towards free and bound water, the water that diffused from the internal core of the powder particles driven by the rise in temperature may promote steric effects, leading to the development of liquid-bridges (Provent et al.,1993). Once liquid-bridges are developed, the powder begins to transform from its glassy state to a rubbery state, making them to be more sticky and non- flowable (Ozkan et al., 2002). From the above premise, we can state that cohesiveness and stickiness of food powders are affected through the formation of liquid-bridging and lowering of viscosity of samples as defined by the behavior of the G’ and G” over a range of temperatures.

3.2. Storage and loss modulus - time profile as it relates to sticky point temperature

The storage modulus (G') and loss modulus (G") profile of the samples as a function of time was established to determine the end-point and average values of the two moduli. The value of G' represents the energy stored in a viscoelastic mass when a strain is applied, while G" indicates the viscous behavior of the samples that leads to the dissipation of the mechanical

215

energy into heat (Hernandez et al., 2006). The typical plots of G' and G" sa ples at each inc e ental te pe atu e within the te pe atu e ange between 2 95 °C are presented in Fig.

7. It can be seen that at the onset temperature of 25 °C for a given sample with specific water content, the G' values recorded for 180 s holding time were always higher than the G". This indicates that the solid-like behavior of the material at this point was discernible. Although both

G' and G" follow a similar path, the difference between the two moduli widened with oscillation time. Ramping the temperature to 35 °C reduced the difference between values of G' and G". We observed that at the moment the temperature was ramped to 45 °C, differences between the two parameters narrowed even further until a certain point, where the values of G' and G" intersected each other and eventually G" values became higher than G'. When the temperature was ramped to 55 °C and further up to 95 °C, the values of G" were higher, suggesting that the material remains in a sticky condition. Depending on the water content of the initial samples and particle size, the magnitude of the two moduli may appear interchangeably.

3.3. Sticky-point temperature of RW-dried mango powder using the rheometer method

RW-dried mango powder with particles size ranging from 180 3 0 µm was selected for measuring the sticky point temperature at different water content. The sticky point temperature at this range of particle size was insignificant, as evidenced by the preliminary experiment comparing particles sizes of 180 & 500 µm (Fig. 8 & 9; Table 1).

Graphical plots of G' and G" as a uncti n te pe atu e anging 2 °C at different water contents of 0.003, 0.022, 0.029, 0.039, 0.048, and 0.067 kg water/kg dry solids are shown in Fig. 10 to 15. Each figure contains two sets of graphs arranged in columns. The first set on the left column represents the three replicates of end-point sticky point temperature

(Tse), while the second set in the right column represents the average sticky point temperature

216

(Tsa). The obtained Tse and Tsa values are summarized in Table 2. It can be seen from the figures that an inverse relationship exists between the sticky point temperature and water content. For example, the measured Tse at the lowest water content samples (0.003 kg water/kg dry solids) is shown in Fig. 8 was 83.8 ± 0.8, while the Tse at highest water content (0.067 kg water/kg dry solids) is presented in Fig. 13 was 26.3 ± 0.6. A similar trend was observed for the values for Tsa.

Overall, the Tsa values were found to be higher than Tse by 2.3 11.7 °C over a range of water content. This result could be attributed to the large variation of G' and G" values from the start to the last data point during the 180 s holding time at each temperature setting. Using the end-point value wherein the samples were at their steady state condition or thermal equilibrium

(Fig. 7) resulted in lower Tse values. It is interesting to note that regardless of the differences in magnitude for G' and G", which could be attributed to slight variations during loading and packing density of the samples (e.g., shown in Fig. 13), the intersection or crossover observed from the G' and G" cu ves den ted as the sticky p int te pe atu e was und ve y c nsistent with a high deg ee epeatability (SD 0. 8 1.73 °C). This suggests that the temperature at which the crossover occurs is independent of the magnitude of G' and G" values.

3.4. Relationship between sticky-point temperature and glass transition temperature

The experimental values obtained from the current investigation for sticky point and glass transition temperatures between the lower and higher water contents showed a high degree of linearity having a coefficient of determination (R2) ranging from 0.907 – 0.966 (Fig. 14). The high consistency and repeatability of measurements using the new eth d (SD 0. 8 1.73 °C) or > 98% accuracy, coupled with a highly controlled and standardized preparation, suggests the proposed method as a reliable technique to quantify and characterize mango powder.

217

Roos & Karel (1991c) and Jaya & Das (2009a) reported a linear relationship between sticky point temperature (Ts) and glass transition temperature (Tg) of sugar-rich materials at different water content. The proposed method strongly supports these findings to describe the sticky phenomenon, as presented in Fig. 14. It can be seen from the figure that the sticky point temperatures (Tse and Tse) of mango powder was always higher than the glass transition temperature at different water content. Both temperatures decrease as the water content increases, which is a typical pattern of the plasticization effect of water in suppressing the glass transition temperature in amorphous solids (Slade & Levine, 1991a). Similar behavior was reported by Roos & Karel (1991b) on the effect of water plasticizers on amorphous sucrose- fructose model foods. At the lowest water content of 0.003 kg water/kg dry solids, the Tse and Tse were 83 ± 0.76 °C and > 95 ± 0.00 °C, respectively. On the other hand, the onset (Tgi), mid-

(Tgm) and end- (Tge) glass transition temperature of the same samples was recorded at 49.2 ±

0.78, 51.3 ± 0.57 °C and 54 ± 0.28 °C, respectively. At the higher water content of 0.067 kg water/kg dry solids, the values of Tse and Tsa were reduced to 26.3 ± 0.58 °C and 29.7 ± 0.58 °C, respectively, with corresponding Tgi, Tgm and Tge to be 4.0 ± 0.74 °C, -3.9 ± 1.07 °C and -0.6 ±

1.87 °C. The differences between the Tse versus the Tgi, Tgm and Tge was calculated t be 21 44

°C, while a highe ange di e ence 2 50 °C was observed between the Tsa versus Tgi, Tgm and Tge (Table 3). These variations are attributed to the lower values of Tse compared to the average of data points calculated for Tsa as explained in Section 3.2. The trend in the sticky point and glass transition temperature of the present investigation over a range of water content was in agreement with previous studies. It was reported that the sticky point temperature was 4

11 °C higher than glass transition temperature in sucrose : fructose (87.5:12.5% w/w mixture)

(D wnt n et al., 1 82), 11. 15.5 °C in vacuum dried mango powder with 0.093 kg

218

maltodextrin/kg dry solids (Jaya & Das, 2009a), 2. 3.5 °C in pineapple with 0.065 kg maltodextrin/kg dry solids (Jaya & Das, 2009a) 10 20 °C for a mixtures of amorphous sucrose- fructose sugars (7:1 ratio) (Roos & Karel, 1991b).

The higher range of differences between sticky point temperature and glass transition temperature obtained from the current investigation (2 0 °C) in comparison with the combined data obtained from the existing literature enti ned that ab ve (4 20 °C) could be attributed to various factors including variations of chemical composition (e..g sugar concentration) and molecular weight of fruit components used in the experiment, drying methods, measurement of water content (oven versus Karl Fisher methods), powder particle sizes, and surface structures of the dried product. The methods applied in measuring the sticky point temperature for each product could also influence these variations. Although the sticky point temperature and glass transition temperatures obtained from the present study were more than two-fold over the previous reports, it was clearly revealed that there was a defined constant range of differences between two temperatures (Ts Tg) over a range of water content of the samples. Also, based on the sticky point temperature being affected by the inter-particle bridging

(solid and liquid), and viscosity as described by loss and storage moduli and water content of the product, the new protocol proposed here is strongly justified as a good alternative to quantify and characterize sticky point temperature of mango powders.

3.5. Comparison of the glass transition temperature of RW-dried mango powder with published data.

The Tg values of vacuum-dried mango powder with 0.093 kg maltodextrin, DE 36 / kg dry solids (Jaya & Das, 2009a) and a sucrose : fructose powder (7:1 ratio) (Roos & Karel,

1991b) were compared to the Tg of RW-dried mango powder measured in the present study (Fig.

219

15). It was clearly noticed that the Tg of vacuum-dried mango powder appears to be greater than

RW-dried mango powder and the sucrose: fructose model powder over a range of water content

(e.g. 0 0.075 kg water/kg dry solids). This trend is obviously due to the higher glass transition temperature for maltodextrin, DE 36 (Tgs = 100 °C) that was added to the vacuum-dried powder as against the sucrose : fructose powder (Tgi = 57.1 and Tge = 68 °C) (Roos & Karel, 1991b) and

RW-dried mango powder (Tgs = 55.8 °C) (Caparino et al., 2012) (Table 4). It is a known fact that adding a high molecular weight carrier such as maltodextrin to a low molecular weight material raises the glass transition temperature of the latter (Jaya & Das, 2004). The slight deviation of glass transition temperature between the sucrose : fructose model powder and RW-dried mango powder could be attributed to similar amount of sugar in both product.

3.6. Comparison of the measured sticky-point temperature with published data

The measured sticky point temperature of RW-dried mango powder using the rheometer was compared with the sticky point temperatures of other sugar-rich fruit powders measured by a sticky-point tester developed by Lazar et al. (1956). The sticky point temperature of the RW- dried mango powder decreases with increase in water content and shows a similar trend as the measured glass transition temperatures. This results is in agreement with the Tg obtained in other food powders such as vacuum-dried mango and pineapple juice powder (Jaya & Das, 2009), orange juice powder (Brennan et al., 1971), model powder containing sucrose : fructose

(Downton et al., 1982), and tomato juice powder (Lazar et al., 1956) (Fig. 16). Fig. 16 indicates that the measured sticky point temperature using the new method lies within the experimental values obtained for the different fruits mentioned. Evidently, the orange juice powder showed the lowest sticky point temperature among the other products presented. One obvious reason for this is due to the low Tg of orange juice powder without a carrier, which measured only 5.7 °C, as

220

reported by Goula & Adamopoulos, 2010. Although a liquid glucose was added during the processing of orange juice powder, the lower Tg obtained is justified because this sugar has low

Tg of 31 °C (Table 4). This is far below the result compared to the Tg of mango powder without a carrier obtained from RW drying (Tgs = 55.8 °C) and freeze drying (Tgs = 63.6 °C), as reported by Caparino et al. (2012). On the other hand, the sticky point temperature of the vacuum-dried mango powder is higher compared to that of RW-dried mango powder due to the addition of maltodextrin to the former product. The sticky point temperature of RW-dried mango powder showed little variation compared to the sucrose : fructose (87% : 12.5 %) mixture powder, which could also be attributed to similar sugar concentration of the product.

Comparing the sticky point testing device developed by Lazar et al. (1956) to determine the sticky point temperature, the proposed method is more accurate and precise. This claim is justified because, unlike the former method, which applied human perception or feeling by hand, the present method is automated and uses computer software to generate data directly from a rheometer that provides both controlled stress and strain data. In addition, the sample preparation is standardized and carefully managed, a procedure which adds consistency and reliability for this new method.

4.0. CONCLUSIONS

The results of the experimental investigation clearly demonstrate that the new proposed rheometer method is suitable as an alternative technique to quantify and characterize sticky point temperature of sugar-rich powder materials. A crossover between the storage modulus (G') and loss modulus (G") of a powder product was established and denoted as sticky point temperature of RW-dried mango powder at different water content. The obtained data using the new method was highly correlated to the glass transition temperature of the product. A direct relationship

221

between the sticky point and glass transition temperatures of RW-dried mango powder agreed with other published sugar-rich powder materials. The developed procedure and protocol developed for sample conditioning and rheometric measurement to determine the sticky point temperature of the material is easy to implement. The new technique can measure the sticky point temperature of a food powder with a high degree of repeatability, reliability and accuracy.

While there is strong evidence to support these claims, more experiments using the new developed method are needed to further confirm its application to a wider range of food powders.

Measuring the crystallization and melting point of fruit powders using this method is also recommended. More understanding of other theoretical aspects of the sticky phenomena using the proposed method is recommended.

ACKNOWLEDGMENTS

We thank the Ford Foundation International Fellowship Program (IFP)/Institute of

International Education (IIE)-New York through the IFP - Philippine Social Science Council

(IFP-PSSC) for providing financial support; the Philippine Center for Postharvest Development and Mechanization (PhilMech) for granting study leave to Ofero Caparino. Special thanks to

Richard E. Magoon and Karin M. Bolland of MCD Technologies, Inc. (Tacoma, WA) for allowing the use of their RW drying facilities and their assistance in doing the experiments;

Wayne Dewitt for helping in the fabrication of the powder loading device, and Binying Ye for providing technical assistance on the operation of rheometer, differential scanning calorimetry and Karl Fisher instruments.

222

REFERENCES

Adhikari, B. (2003). Drying Kinetics and Stickiness of Single Drop of Sugar and Acid-Rich

Solutions. PhD thesis, Chemical Engineering,University of Queensland .

Adhikari, B., Howes, T., Bhandari, B. R., & Truong, V. (2001). Stickiness in foods: A review of

mechanisms and test methods. International Journal of Food Properties , 4 (1), 1-33.

Adhikari, B., Howes, T., Bhandari, B., & Truong, V. (2003). In situ characterization of stickiness

of sugar-rich foods using a linear actuator driven stickiness testing device. Journal of

Food Engineering , 58, 11–22.

Barbosa-Canovas, G., V., Ortega-Rivas, E., Juliano, P., & Yan, H. (2005). Food Powders:

Physical Properties, Processing, and Functionality. New York: Kluwer

Academic/Plenum Publishers.

Bhandari, B. R., Datta, N., & Howes, T. (1997). Problems associated with spray drying of sugar-

rich foods. Drying Technology , 15 (2), 671-684.

Boonyai, P., Bhandari, B., & Howes, T. (2004). Stickiness measurement techniques for food

powders: A review. Powder Technology , 145, 34-46.

Brennan, J. G., Herrera, J., & & Jowitt, R. (1971). A study of some of the factors affecting the

spray drying of concentrated orange juice on a laboratory scale. Journal of Food

Technology , 6 (3), 295–307.

Caparino, O. A., Tang, J., Nindo, C. I., Sablani, S. S., Powers, J. R., & Fellman, J. K. (2012).

Effect of drying methods on the physical properties and microstructures of mango

(Philippine 'carabao' var.) powder. Journal of Food Engineering , 111, 135–148.

Chen, X. D., & Ozkan, N. (2007). Stickiness, functionality, and microstructures of food powders.

Drying Technology , 25, 969-979.

223

Dixon, A., Bloore, M., & Rhodes, M. (1999). Correlating food powder stickiness with

composition, temperature, and relative humidity - thesis. Melbourne: Department of

Chemical Engineering, Monash University.

Downton, G. E., Flores-Luna, J. L., & King, C. J. (1982). Mechanisms of stickiness in

hygroscopic, amorphous powders. Ind. Engineering Chem. Fundamental , 21, 447-451.

Goula, A. M., Karapantsios, T. D., & Adamopoulos, K. G. (2007). Characterization of tomato

pulp stickiness during spray drying using a contact probe method. Drying Technology ,

25, 591–598.

Green, H. (1941). The tackmeter, an instrument for analyzing and measuring tack. Indistrial Eng.

Chem. Anal. , 13, 632-639.

Hennigs, C., Kockel, T. K., & & Langrish, T. A. (2001). New measurements of the sticky

behavior of skim milk powder. Drying Technology , 19 (3–4), 471–484.

Hernandez, L., Gurruchaga, M., & Goni. (2006). Influence of powder particle size distribution

on complex viscosity and other properties of acrylic bone cement for vertebroplasty and

kyphoplasty. Journal of Biomedical Materials Research Part B: Applied Biomaterials ,

77B (1), 98–103.

Jaya, S., & Das, H. (2004). Effect of maltodextrin, glycerol monostearate and tricalcium

phosphate on vacuum dried mango powder properties. Journal of Food Engineering , 63,

125-134.

Jaya, S., & Das, H. (2009). Glass transition and sticky point temperatures and stability/mobility

doagram of fruit powders. Food Bioprocess Technol , 2, 89-95.

Jineke, A. W. (1964). Storage and flow of solids (Vol. 123). University of Utah: Utah

Engineering Experimental Station.

224

Kilcast, D., & Roberts, C. (1998). Perception and measurement of stickiness in sugar-rich foods.

Journal of Texture Studies , 29, 81–100.

Kudra, T., & Mujumdar, A. S. (2002). Advanced Drying Technologies. New York: Marcel

Dekker, Inc.

Lazar, W. E., Brown, A. H., Smith, G. S., Wong, F. F., & Lindquist, F. E. (1956). Experimental

production of tomato powder by spray drying. Food Technology , 129-134.

Lockman, C. (1999). A new laboratory method to characterize the sticking properties of free-

flowing solids. Chemical Engineering and Processing , 38, 301–306.

Menjivar, J. A., & Faridi, H. (1994). Rheological properties of cookie and cracker doughs. In H.

Faridi, The Science of Cockie and Cracker Production. (pp. 283-322). New York:

Chapman & Hall.

Mezger, T. G. (2002). In U. Zorll, The rheology handbook for users of rotational and oscillatory

rheometers. Hannover, Germany.

Ozkan, N., Walisinghe, N., & Chen, X. D. (2002). Characterization of stickiness and cake

formation in whole and skim milk powders. Journal of Food Engineering , 55, 293–303.

Palzer, S., & Sommer, K. (2011). Caking of water-soluble amorphous and crystaline food

powders. In J. M. Aguilera, J. Simpson, D. Welti-Chanes, D. Bermudez-Aguirre, & G.

Barbosa-Canovas, Food Engineering Interfaces (p. 680). Springer.

Papadakis, S., & Bahu, R. (1992). The sticky issues of drying. Drying Technology , 10, 817–837.

Paterson, A. H., Bronlund, J. E., & Brooks, J. F. (2001). The blow test for measuring the

stickiness of powders. AIChE 2001 Annual Meeting. Reno, Hilton USA.

Peleg, M. (1983). Physical characteristics of food powders. In M. Peleg, & E. B. Bagley,

Physical Properties of Foods (pp. 293-323). Westport, CT: AVI Publishing Co.

225

Provent, B., Chulia, D., & Cary, J. (1993). Particle size and the caking tendency of a powder. Eu.

J. Pharma. Biopharm , 39 (5), 2002-207.

Rao, M. A. (1999). Rheological behavior of processed fluid and semisolid foods. In M. A. Rao,

Rheology of fluid and semisolid foods: principles and operations (pp. 105-108, 244-254).

Roos, Y. H. (2003). Thermal analysis, state transition and food quality . Journal of Thermal

Analysis Calorimetry , 197-203.

Roos, Y. H. (1993). Water activity and physical state effects on amorphous food stability.

Journal of Food Processing and Preservation , 16, 433-437.

Roos, Y. H., & Karel, M. (1991c). Phase transitions of mixtures of amorphous polysaccharides

and sugars. Biotechnology Progress , 7, 49-53.

Roos, Y. H., & Karel, M. (1991b). Plasticizing effect of water on thermal behaviour and

crystallization of amorphous food models. Hournal of Food Science , 56, 38-43.

Schubert, H. (1987). Food Particle Technology. Part I: Properties of particles and particulate

food systems. Journal of Food Engineering , 6, 1.

Slade, L., & Levine, L. (1991). A food polymer science approach to structure property

relationships in aqueous food systems: Non-equilibrium behaviour of small carbohydrate-

water system. In L. Levine, & L. Slade, Water Relationships in Food. New York:

Plennum Press.

Syamaladevi, R. M., Sablani, S. S., Tang, J., Powers, J., & Swanson, B. G. (2009). State diagram

and water adsorption isotherm of raspberry (Rubus Idaeus). Journal of Food Engineering

, 460-467.

Tabilo-Munizaga, G., & Barbosa-Canovas, G. V. (2005). Rheology for food industry. Journal of

Food Engineering , 67, 147-156.

226

Teunou, E., Fitzpatrick, J., & Synnott, E. (1999). Characterisation of food powder Flowability.

Journal of Food Engineering , 39, 31-37.

Tsourouflis, S., Flink, J. M., & & Karel, M. (1976). Loss of structure in freeze-dried

carbohydrates solutions: effect of temperature, moisture content and composition.

Journal of the Science of Food and Agriculture , 27, 0 − 1 .

Wallack, D. A., & King, C. J. (1988). Sticking and agglomeration of hygroscopic, amorphous

carbohydrate and food powders. Biotechnology Progress , 4 (1), 31-35.

White, G. W., & Cakebread, S. H. (1966). The glassy state of certain sugar-containing food

products. Journal of Food Technology , 1, 73-82.

227

Table 5. Sticky point temperature of RW-dried mango powder with particle size 180 m and 500 m. Water content, kg Particle size, m Replicates Tse, °C Tsa, °C water/kg dry solids 180 R1 36.0 36.0 R2 3.9 ± 0.07 36.0 36.5 R3 34.2 35.5 500 R1 36.0 40.0 R2 3.9 ± 0.07 36.0 38.0 R3 35.6 39.0

Table 6. data of water content, sticky point temperature and glass transition temperature of mango powder at different water content.

Sample Sticky point Water content, Glass transition temperature temperature kg water/kg dry Replic solids ation Tse Tsa Tgi Tgm Tge 1 0.0.003 ± 0.001 R1 84.5 > 95.0 49.7 51.7 54.2 R2 83.0 > 95.0 48.6 50.9 53.8

R3 84.0 > 95.0

2 0.022± 0.008 R1 65.0 70.0 20.5 22.9 26.4 R2 63.0 69.0 20.4 23.3 27.3

R3 64.0 70.0

3 0.029 ± 0.003 R1 59.0 65.0 15.4 18.8 23.2 R2 59.5 68.0 15.8 19.4 23.9

R3 57.0 65.0

4 0.039 ± 0.007 R1 41.0 45.0 12.4 16.1 19.9 R2 43.0 46.0 12.8 16.3 20

R3 40.0 43.0

5 0.048 ± 0.003 R1 36.0 39.0 1.9 5.1 8.6 R2 36.0 38.0 2.1 5.3 8.7

R3 35.0 37.0

6 0.067 ± 0.008 R1 26.0 30.0 -7.02 -3.19 0.75 R2 26.0 29.0 -0.98 -4.7 -1.89

R3 27.0 30.0

1 Samples were stored in P2O5 for 30 days2 Reference sample was oven-dried for 1 h at 50 °C 3 Reference RW-dried mango powder was oven-dried for 30 min at 50 °C 4 Reference RW-dried mango powder with water content of 3.89 kg water / kg dry solids 5 Reference RW-dried mango powder was placed inside a dessicator with 100% RH and conditioned for 20 min. 6 Reference RW-dried mango powder was placed inside a dessicator with 100% RH and conditioned for 40 min.

228

Table 7. Differences between the sticky point temperature and glass transition temperature of mango powder at different water content.

Water Difference (Ts Tg), °C Content, kg water / kg Tse Tgi Tse Tgm Tse Tge Tsa Tgi Tsa Tgm Tsa Tge dry solids 0.003 ± 0.001 35 33 30 46 44 41

0.022± 0.008 44 41 37 49 47 43

0.029 ± 0.003 43 39 35 50 47 42

0.039 ± 0.007 29 25 21 32 28 25

0.048 ± 0.003 34 30 27 36 33 29

0.067 ± 0.008 30 30 27 34 34 30

Differences 2 44 2 41 21 37 32 0 28 47 2 43

Ts - Sticky point temperature Tg - Glass transition temperature Tse - End-point sticky point temperature Tsa - Average-point sticky point temperature Tgi - Onset glass transition temperature Tgm - Mid-point glass transition temperature Tge - End-point glass transition temperature

229

Table 8. Glass transition temperatures of selected fruits, unhydrous sugars and carbohydrate polymers.

Organic Sucrose, Glucose, Fructose, Tgs, Melting acid Product Kg per kg Kg per kg Kg per kg References °C point, °C Kg per kg pulp (%) pulp (%) pulp (%) pulp (%) Mango - RW-dried 55.8 - - - - - Caparino et al., 2012 - Freeze-dried 63.6 - - - - - Caparino et al., 2012 - Vacuum-dried 61 - 0.064 0.047 0.026 0.003 Jaya & Das, 2009 Pineapple - Vacuum-dried 61 - 0.062 0.037 0.016 0.002 Jaya & Das, 2009 Orange - Without carrier 5.7 - (43.1) - - (43.1) Goula & Adamopoulos, 2010 - With DE 6* maltodextrin (50:50) 66 - (48 2) - - ( .3 6.3) Goula & Adamopoulos, 2010 - With DE 6 maltodextrin (50:50) 78 - (43.1) - - (4.3) Shresta et al., 2007 Tomato Bhandari & Howes, 1997 31 - 0.006 0.010 0.010 0.006

230 Jaya & Das, 2009 Amorphous food model

57 - - - - - Roos & Karel, 1991b Sucrose-Fructose (7:1) Sucrose-Fructose solution - - (87.5) - (12.5) - Downton et al., 1982 (87.5 : 12.5%) mixture 62 - - - - - Jaya & Das, 2009 Sucrose - 173 - - - - Roos, 1993 31 - - - - - Jaya & Das, 2009 D-Glucose - 163 - - - - Roos, 1993 5 - - - - - Jaya & Das, 2009 D-Fructose - 108 - - - - Roos, 1993 11 - - - - - Jaya & Das, 2009 Organic acid - 153 - - - - Roos, 1993b Starch 250 - - - - - Roos, 2003 Maltodextrin DE 5 188 240 - - - - Roos & Karel, 1991c DE 10 160 - - - - - Roos & Karel, 1991c DE 20 141 - - - - - Roos & Karel, 1991c DE 36 100 - - - - - Roos & Karel, 1991c * DE: Dextrose equivalents

230

(a) (b)

(e) (f)

Figure 1. Schematic diagram of some methods and devices in characterizing sticky-point temperature measuring device (Adapted from Boonyai et al., 2004). (a) Sticky-point tester (Lazar et al., 1956; Wallack & King, 1988); (b) Motor-driven sticky point temperature device (Brennan, 1971); (c) Viscometer method (Okzan et al., 2002 ); (d) Mechanically stirred with relative humidity controller (Kudra, 2002); (e) optical probe method; ( f) In situ stickiness test device (Boonyai et al., 2004).

231

Figure 2. Standardized procedure during moisture conditioning of mango powder samples

232

Figure 3. A developed standardized loading procedure of samples that were subjected for conditioning at ~100% RH

233

Figure 4. Schematic diagram of the rheometer system used in determining the sticky point temperature of mango powder.

234

Circular powder loader Scraper

(a)

(b)

Figure 5. A developed fabricated powder loader (PLD) and scraper (a) and procedure of loading the samples in the lower heated serrated plate of the rheometer (b).

235

a b c d e

f g h

Solid-like Liquid-like (< Ts ) ( > Ts )

G’

G' G" G' Crossover

G”

0 0 Temperature

- Cross section of powder particle - Thermal plasticization - Interface

Ts - Sticky point temperature

Figure 6. Proposed mechanism of sticking of sugar-rich amorphous fruit powders applying storage modulus (G’) and loss modulus (G”). Letters (a) to (h) are explained in the results and discussion.

236

8000 25 C 40000 65 C G' G'

6000 30000 G' Pa

(Pa) G'' G'' G'' Pa

4000 (Pa) 20000

G' G" G'

2000 G" G' 10000

0 0 0 50 100 150 200 0 50 100 150 200

Time, s Time, s

15000 35 C 30000 75 C

G' Pa G' Pa 10000 20000 G'' Pa G'' Pa 5000

G' G" (Pa) G" G' 10000 G' G" (Pa) G" G' 0 0 0 50 100 150 200 0 50 100 150 200 Time, s Time, s

45 C 15000 20000 85 C G' Pa 15000 10000 G' Pa G'' Pa 10000 G'' Pa 5000

G' G" (Pa) G" G' 5000 G' G" (Pa) G" G'

0 0 0 50 100 150 200 0 50 100 150 200 Time, s Time, s

15000 55 C 15000 95 C

G' Pa G' Pa

10000 10000 (Pa)

G'' Pa G'' Pa

(Pa)

5000 5000

G' G" G' G' G" G'

0 0 0 50 100 150 200 0 50 100 150 200 Time, s Time, s

Figure 7. Typical he l gical p ile l ss and st age dulus ve sus ti e sa ples 2 0 °C at a ramping increment of 10 °C.

237

Tse (particle size of 500 m) Tse (particle size of 180 m)

R1 R1 100000 G' 80000 G' G'' 80000 36°C G'' 36°C

60000

(Pa)

40000

G' G" G' G' G" G' 20000 20000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Axis Title

R2 R2

100000 G' 250000 G' 36°C

80000 G'' 200000 36°C G''

150000 (Pa)

(Pa) 60000

40000 100000

G' G" G' G' G" G' 20000 50000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

R3 R3 60000 G' 120000 34.2°C 35.6°C G'

50000 G'' 100000 G''

40000 80000

(Pa)

30000 60000 G' G" G' G' G" G' 20000 40000 10000 20000 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, °C Temperature, C

Figure 8. End-point sticky point temperature (Tse) of RW-dried mango powder at 180 and 500 m particle size and water content of 0.039 kg water/kg dry solids.

238

Tsa (particle size of 500 m) Tsa (particle size of 180 m)

R1 R1

100000 G' 100000 G' 36°C 40°C

80000 G'' 80000 G''

60000 60000

(Pa)

40000 40000 G' G" G' G' G" G' 20000 20000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

R2 R2 100000 36.5°C G' 300000 G' G''

80000 250000 38°C G''

200000

60000 (Pa)

(Pa)

150000

40000 G' G" G' G' G" G' 100000 20000 50000 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

R3 R3 100000 140000 39°C G' 36°C G' 120000

80000 G'' G''

100000 (Pa)

60000 (Pa)

80000

40000 60000 G' G" G' G' G" G' 40000 20000 20000 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 9. Average sticky point temperature (Tsa) of RW-dried mango powder at 180 and 500 m particle size.

239

G' G' Tse (R1) Tsa (R1) 1600 G'' 8000 G''

1200 6000

(Pa)

800 4000 G' G" G' G' G" G' 85°C > 95°C 400 2000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

G' G' Tse (R2) Tsa (R2) 500 G'' 5000 G''

400 4000

300 (Pa) 3000

(Pa)

83°C

200 2000 > 95°C

G' G" G' G' G" G' 100 1000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R3) G' Tsa (R3) G'

1000 G'' 5000 G''

800 4000

(Pa)

(Pa) 600 3000

84°C G" G' 400 > 95°C G' G" G' 2000

200 1000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 10. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.003 kg water/kg dry solids.

240

G' Tse (R1) Tsa (R1) G'' 2500 3200

2000

2400

(Pa) (Pa)

1500 70°C 65°C 1600 G'

1000 G' G" G' G' G" G' G'' 500 800

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R2) Tsa (R2)

3200 4000

2400 3000

(Pa)

69°C

1600 63°C G' 2000 G' G' G" G' G' G" G' G'' G'' 800 1000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R3) Tsa (R3)

2000 3200

1600

2400

(Pa) 70°C 1200 64°C (Pa) G' 1600 G'

800 G' G" G'

G'' G" G' G'' 400 800

0 0 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 11. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.022 kg water/kg dry solids.

241

Tse (R1) Tsa (R1)

4000 6000

5000

3000

4000

(Pa) (Pa)

2000 59°C G' 3000 65°C G' G' G" G' G' G" G' 2000 1000 G'' G'' 1000 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R2) Tsa (R2)

4000 6000

5000 3000

(Pa) 4000

(Pa)

68°C

2000 61°C G' 3000 G' G' G" G' G'' G" G' 2000 G'' 1000 1000 0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R3) Tsa (R3)

2500 4000

2000 3000

65°C

(Pa)

(Pa) (Pa) 1500 57°C G' 2000 G'

1000 G" G' G' G" G' G'' 1000 G'' 500

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 12. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.029 kg water/kg dry solids.

242

Tse (R1) Tsa (R1)

16000 16000 41°C 45°C

12000 12000

(Pa)

8000

(Pa) (Pa) 8000 G' G'

G'' G'' G' G" G' 4000 G' G" G' 4000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R2) Tsa (R2)

20000 20000 16000

15000

46°C

12000 43°C

(Pa) (Pa) 10000 G' G' 8000

G'' G'' G' G" G' G' G" G' 5000 4000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R3) Tsa (R3) 20000 20000

16000 16000

12000 40°C 12000 43°C

(Pa)

G' G' 8000 8000

G'' G'' G' G" G' 4000 G" G' 4000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 13. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.039 kg water/kg dry solids.

243

Tse (R1) Tsa (R1)

16000 16000

12000 12000

36°C 39°C

(Pa)

(Pa) 8000 8000 G' G'

G'' G" G' G'' G' G" G' 4000 4000

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R2) Tsa (R2)

16000 20000 16000

36°C 12000 38°C

12000 (Pa) G' " (Pa) " 8000 G'

G'' 8000 G'' G' G G' 4000 G" G' 4000

0 0 0 20 40 60 80 100 0 20 40 60 80 100

Temperature, C Temperature, C

Tse (R3) Tsa (R3)

20000 20000 37°C 35°C

15000 15000

(Pa)

10000 G' 10000 G'

G' G" G' G' G" G' 5000 G'' 5000 G''

0 0 0 20 40 60 80 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 14. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.048 kg water/kg dry solids.

244

Tse (R1) Tsa (R2)

50000 100000

40000

80000 (Pa)

(Pa) 30000

60000 26°C G' G'

20000 40000 G' G" G' G' G" G' G" 30°C G'' 10000 20000

0 0 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 Temperature, C Axis Title

Tse (R2) Tsa (R2)

250000 200000 200000

150000

(Pa) 150000

(Pa) G' 100000 G' 100000

G' G" G' 29°C 26°C G'' G" G' G'' 50000 50000

0 0 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 Temperature, C Temperature, C

Tse (R3) Tsa (R3)

100000 160000

80000

120000 (Pa)

60000

(Pa)

G' 80000 G' G' G" G' 40000 G'' G" G' G'' 20000 27°C 40000 30°C

0 0 0 10 20 30 40 50 60 70 80 90 100 0 20 40 60 80 100 Temperature, C Temperature, C

Figure 15. End point (Tse) and average (Tsa) sticky point temperature of samples dried at water content of 0.067 kg water/kg dry solids.

245

R² = 0.967 Tse, C R² = 0.951 120 Tsa, C R² = 0.927 Tgi R² = 0.933 Tgm 100 R² = 0.943 Tge

40 Temperature, ºC Temperature,

-20 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Water content, kg water / kg dry solids

Figure 16. Relationships of sticky point temperature and glass transition temperature of RW- dried mango powder as function of water content.

246

RW-dried mango powder without additives, Tgi (present study) RW-dried mango powder without additives, Tgm (present study) RW-dried mango powder without additives , Tge (present study) Vacuum-dried mango powder + maltodextrin (Jaya & Das, 2009) Sucros-fructose (7:1) (Roos & Karel, 1991b) 70

40 R² = 0.995

C 30

10 Temperature,

R² = 0.9806 -10 R² = 0.943 -20 R² = 0.933 R² = 0.927 -30 0.000 0.020 0.040 0.060 0.080 0.100

WAter content, kg water/kg dry solids

Figure 17. Comparison of glass transition temperature of sucrose-fructose model food, RW- dried, and vacuum-dried mango powders.

247

RW-dried mango powder without additives (present study) Vacuum-dried mango powder + 0.093 kg maltodextrin/kg pulp (Jaya & Das, 2009) Sucros-fructose powder (Downton et al., 1982) 90 pineapple juice powder + 0.065 kg maltodextrin/kg pulp (Jaya & Das, 2009) Orange juice powder + liquid glucose, 39-43 D.E (Brennan et al., 1971) Tomato Powder (Lazar et al., 1956) 80

40 Temperature, ºC Temperature,

R² = 0.968 10 RW-dried mango powder 0 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

Water content, kg water/kg dry solids

Figure 18. Sticky point temperatures of RW-dried mango powder and other sugar-rich fruit powders.

248

CHAPTER SEVEN

CONCLUSIONS AND RECOMMENDATIONS

CONCLUSIONS

Production of mango (Philippines 'Carabao' var.) powders with water content below

0.05 kg water/kg dry solids using Refractance Window drying (RW), freeze drying (FD), drum drying (DD) and spray drying (SD) methods were found to be technically feasible. Despite the presence of low molecular weight sugars in mango, e.g. sucrose, fructose and glucose, RW-, FD- and DD-dried mango powders were produced without the addition of any carrier, while it was found to be difficult to produce spray-dried mango powder without the addition of a drying aid such as maltodextrin. The lowest concentration of carrier added during spray drying of mango

(e.g. 25 kg maltodextrin/kg dry mango solids) significantly altered the original color of mango puree during spray drying and resulted in poor recovery. The color of RW-dried mango powder and reconstituted mango puree were comparable to the freeze-dried powder, but significantly different from the drum-dried and spray-dried counterparts, with the latter two appearing to be darker and lighter, respectively. RW- and drum-dried mango powders were more bulky and less porous compared to freeze-dried and spray-dried powders. The difference in solubility and hygroscopicity of RW- and freeze-dried mango powder was insignificant.

The glass transition temperatures of RW-, freeze-, drum- and spray-dried mango powders were not significantly different. All of the powders exhibited amorphous structures, as evidenced by X-ray diffractograms. Microscopic examination showed that RW-dried mango powder was smooth and flaky, with uniform thickness, while freeze-dried mango powder showed a skeletal- like structure with large pores. Drum-dried mango powder exhibited irregular morphology with sharp edges, while spray-dried mango powder had a spherical shape. The RW drying technique

249

can produce mango powder with quality comparable to that obtained via freeze drying, and better than the drum and spray-dried mango powders.

The sorption isotherms for RW- and freeze-dried mango powders showed sigmoidal characteristics when fitted to both GAB and BET models. No significant differences were observed in the initial glass transition temperatures of RW-dried and freeze-dried mango powders at constant water activities below 0.86. The sorption isotherms and glass transition temperatures generated for RW-dried mango powder should provide valuable information for predicting the stability of RW- and freeze-dried mango powders during long term storage.

The effects of packaging atmosphere (air and nitrogen flushing), storage temperatures (5,

22 and 45 °C) and time on the physical and chemical stability of RW-dried mango powder were evaluated over a period of 12 months. The study showed that mango powder stored at 45 °C suffered discoloration as well as and asc bic acid (AA) and β-carotene degradation after 6 and

12 months of storage in both air or nitrogen flush packaging, while powder stored at 5 and 22 °C preserved the color with minimal changes in nutrients. Nitrogen flush packaging was ineffective in preserving AA in mango powder, but helped in educing the pe centage l ss β-carotene when stored at 5 and 22 °C over a period of 6 months. Regardless of the packaging methods applied, the initial structure of mango powder stored at 5 and 22 °C was preserved after 6 and 12 months, while all of the powder particles inside the package formed a single lump of particles when stored at 45 °C.

The sticky phenomenon in RW-dried mango powder was investigated using a newly developed method with the application of a rheometer system. Results showed that the sticky point temperature obtained using the new method agreed well with the published data and can be considered as a suitable technique to characterize sticky point temperature of sugar-rich

250

materials. The protocol is easy to handle and operate and can measure sticky point temperature of mango powder with high degree of repeatability and accuracy.

RECOMMENDATIONS

Based on the outcomes of the present research, the following recommendations are offered:

 The analyses of RW-, freeze-, drum-, and spray-dried mango powders in the present study

focused on their physical properties. Further investigation on the retention of taste, flavor and

aroma of different dried mango powders is recommended.

 The longer residence time (e.g. 32 h) to obtain freeze-dried mango powder with a desired

water content, coupled with the prohibitive cost of freeze drying equipment may not be

profitable for processing mango powder. Studies to reduce the drying time will attract

processors to utilize the technology for drying mango as well as for non- high valued

commodities.

 To improve the color of the spray-dried mango powder and obtain good recovery, there is a

need to optimize the drying conditions (e.g. inlet temperature, outlet temperature and relative

humidity) during spray drying operation. The application of drying aids other than

maltodextrin, such as proteins and low molecular surfactants should also be explored to

improve recovery.

 The darker color of drum dried mango powder and difficulty in scraping the dried material

was due to high processing temperature and sticky characteristics of mango pulp,

respectively. Studies to optimize the drum speed, thickness of the samples, and addition of

drying aid to mango puree may help improve the quality and recovery of the finished

product.

251

 The RW drying technology performed very well in obtaining good quality mango powder

comparable to freeze drying method. This technology has great potential for adaptation and

commercialization in developing or tropical countries like the Philippines, where the supply

of raw mango is abundant. But, in order to attract possible investors, there is a need to

conduct a feasibility study using a commercial RW drying unit in a specific location. Actual

processing the drying mango puree in commercial quantity will help validate the various

quality parameters generated from the present laboratory scale experiments. Conduct of a

feasibility study is recommended to determine the profitability and viability of the RW

drying technology for commercial production of mango powder.

 When adapting RW drying technology in developing countries, there is a need to study the

lowering of energy costs, especially in countries where the cost of electricity is very high.

One possible option is to retrofit the RW dryer with a boiler powered by a biomass furnace or

any other alternative source of energy that is capable of delivering the required hot

circulating water. Since the RW drying technology is under proprietary ownership, this

option could only be possible with arrangement with the inventor.

 During the storage experiment for RW- and freeze-dried mango powder, only one type of

packaging material (e.g. multi-layer film pouch) was used. The application of other types of

packaging materials during short and long term storage is recommended, with analyses

focused on physicochemical, nutritional and sensorial factors.

 The proposed method to characterize the sticky point temperature of mango powder revealed

a high degree of repeatability, reliability and accuracy. However, there is still a need to

conduct further investigations to validate its application to a wider range of food powders and

particle sizes.

252

~ ~ ~ END ~ ~ ~