Optimization of processing technology for commercial
Transcription
Optimization of processing technology for commercial
FORSCHUNGSBERICHT AGRARTECHNIK des Fachausschusses Forschung und Lehre der Max-Eyth-Gesellschaft Agrartechnik im VDI (VDI-MEG) 530 John Bosco Kawongolo Optimization of Processing Technology for Commercial Drying of Bananas (Matooke) Dissertation Witzenhausen 2013 Universität Kassel Fachbereich Ökologische Agrarwissenschaften Fachgebiet Agrartechnik Prof. Dr. sc. agr. Oliver Hensel Optimization of Processing Technology for Commercial Drying of Bananas (Matooke) Dissertation zur Erlangung des akademischen Grades Doktor der Agrarwissenschaften (Dr. agr.) Fachbereich Ökologische Agrarwissenschaften von John Bosco Kawongolo, B.Sc. (Mech.) Eng, M.Sc. (Agric. Eng.) Agric. aus Uganda 2013 Die vorliegende Arbeit wurde vom Fachbereich für Ökologische Agrarwissenschaften, Fachgebiet Agrartechnik der Universität Kassel als Dissertation zur Erlangung des akademischen Grades Doktor der Agrarwissenschaften angenommen. Tag der mündlichen Prüfung: 13.12.2013 Erster Gutachter : Prof. Dr. sc. agr. Oliver Hensel Zweiter Gutachter : Prof. Dr.-Ing. Werner Hofacker Mündliche Prüfung: PD Dr. Johannes Kahl Prof. Dr. Florence I. Muranga Alle Rechte vorbehalten. Die Verwendung von Texten und Bildern, auch auszugsweise, ist ohne Zustimmung des Autors urheberrechtswidrig und strafbar. Das gilt insbesondere für Vervielfältigung, Übersetzung, Mikroverfilmung sowie die Einspeicherung und Verarbeitung in elektronischen Systemen. © 2013 Im Selbstverlag: John Bosco Kawongolo Bezugsquelle: Universität Kassel Fachbereich Ökologische Agrarwissenschaften Fachgebiet Agrartechnik Nordbahnhofstrasse 1a 37213 Witzenhausen Acknowledgements The work in this manuscript was accomplished under the inspiring guidance and dynamic supervision of Prof. Dr. Oliver Hensel, Department of Agricultural Engineering, Faculty of Organic Agricultural Sciences, University of Kassel, Germany, who kept his door always open for consultation and advice at all times when required. I want to thank the staff of the Department of Agricultural Engineering, University of Kassel, for their support. I want to acknowledge Prof. Dr. Werner Hofacker, Environmental Process Engineering, Thermal Processing Engineering, University of Applied Sciences, Konstanz, Germany, for his kind advice and guidance. Special thanks to Dr. Ing. Albert Esper, M/s Innotech, Germany, for kind academic and professional guidance during my research. I want to give special thanks and would like to express my gratitude to Rev. Dr. Florence I. Muranga for both academic and parental guidance to ensure that I keep track of my research. I would like to thank all the professionals working at PIBID, the laboratory technician and all other staff of PIBID for their continued support and encouragement in my academic career. I am deeply grateful to the BMC of PIBID for taking me on as a research fellow and giving me the opportunity to undertake my research in order to contribute to the development of the banana processing industry in Uganda. Special thanks to the Government of the Republic of Uganda for the financial support, through PIBID, during the course of my study to enable me accomplish my academic endeavors. Special thanks to my brothers in our Lord Jesus Christ, and to my mother and father (deceased), who tirelessly prayed and encouraged me to be focused in life and in my career. Finally, I have no words to use to express my gratitude to my beloved wife Mary and children Israel, Rebecca and John, for their continued sacrifice at a time when they needed me most, for their patience, eternal affection, support and encouragement during my academic career. Glory be to The Almighty God for the gift of life and success. “The fear of the Lord is the beginning of wisdom and the knowledge of the Holy One is insight” (Proverbs 9:10) Amen. 2013 John Bosco Kawongolo List of Tables Table of Contents I II List of Tables List of Figures List of Figures III IV List of Figures List of Figures V VI List of Tables List of Tables List of Tables VII Nomenclature VIII Nomenclature A m2 Area aw dimensionless Water activity c, c1, c2 & c3 cv. dimensionless Empirical constant Cultivar DD days Degree-days Deff m2/s Effective diffusivity Do m2/s Empirical constant df Numerical Degree of freedom (n-1) Ea kJ/mol Activation energy i Numerical 1, 2, 3, … k dimensionless Empirical constant k /s Drying rate constant L mm Half thickness of slice MR dimensionless Moisture ratio n Numerical 1, 2, 3, … P degrees Total rotation p kPa Equilibrium partial vapor pressure of water in the system P & P΄ degrees Rotation po kPa Equilibrium partial vapor pressure of pure liquid water at the same temperature qo dimensionless Empirical constant Qst kJ/kg Total isosteric heat of sorption qst kJ/kg Net isosteric heat of sorption R spp. kJ/kg K, kJ/mol K Universal gas constant Species (plural) sp. Species (single) t s Time T ˚C/K Temperature xe kg water/kg dry matter Equilibrium moisture content Experimental equilibrium moisture Content at ith xi data xm kg water/kg dry matter Monolayer xo dimensionless Empirical constant xpred kg water/kg dry matter Predicted moisture content α α degrees degrees Hue angle angular rotation λ kJ/kg Latent heat of pure water Introduction 1 Chapter One Introduction Bananas and plantains (Musa spp.) are grown worldwide in 120 -130 tropical countries and are the fourth most important food crop after rice, wheat, and maize (Qi et al., 2000; Nelson et al., 2006; Nyombi, 2010; Akcoaz, 2011). Bananas and plantains are a source of nutrition and household income for 400 – 500 million people in Africa, Asia and South America (Nelson et al., 2006; Adeniji et al., 2010; IITA 2009). The East African highland banana (EAHB) is a distinct group of AAA bananas which are found in East African highlands and a staple food for 80 million people in the Great Lakes region of East Africa (Robinson & Saúco, 2010; Adewole et al., 2012). Banana is an important crop in Uganda; it is a staple food for more than 70% of Uganda’s population and contributes to about 42% of household income in rural areas (Karugaba & Kimaru, 1999). According to FAO, Uganda is the second largest producer of bananas and plantains after India, with a production of 10.5 million metric tons of plantains and 0.5 million metric tons of dessert bananas, giving a total production of 11 million metric tons (FAOSTAT, 2011). The most common of these plantains is the cooking type, locally called matooke, a Musa sp. triploid acuminate genome group (AAA-EAHB), steamed and served as a main course (Karamura et al., 1998; Nyombi et al., 2009). Matooke are perishable and are traded in fresh form (Vinzenz, 2011). This leads to high postharvest losses ranging from 22-45%, worth $411.9 – 842.5 million (Muranga et al., 2010). These high postharvest losses are attributed to many silent factors including non-uniform level of maturity at harvest, poor handling, bulk transportation and lack of value addition/processing technologies, which are currently the main challenges for trade and export, and its diversified utilization. This is evidenced by the fact that even though Uganda is the second world producer of bananas and plantains, it does not appear anywhere among the world exporters of bananas (Robinson & Saúco, 2010). Drying of fruits is a widely applied process to increase shelf life and reduce bulk transportation; thus value addition/processing by drying would contribute 2 Introduction to solving the above-mentioned challenges. Muranga (1998) demonstrated the stabilization of raw matooke flour and showed that it possesses good properties for the food industry which include among others: more than 80% starch (db), extensive shelf life due to low sugar and fat, high levels of potassium and negligible levels of condensed tannins (therefore, non-astringent). In addition, research on banana flour has shown that the properties of banana flour are important for their utilization in industrial food products and non-food products as stabilizers and tablet binders in pharmaceutical industries (Adewole et al., 2012; Vatanausuchart et al., 2012; Niba et al., 2002). Drying is one of the oldest technologies employed in the processing of agricultural produce and a lot of research has been carried out on the drying of different crops. The objective of drying is to remove water to a level such that microbial spoilage and deterioration reactions are greatly minimized (Doymaz, 2007). Drying is a complex process in which heat and mass transfer occur simultaneously. It is important to understand the parameters controlling this complex process and to understand the drying behavior of a given product for use in improving the existing drying systems, controlling the drying process and improving on the quality of the dried product. A number of studies have been conducted in order to understand the drying behavior of bananas. Baini & Langrish, (2007) reported that the diffusion model suited the moisture content better than the empirical model. It has been reported that drying of bananas takes place in three drying regimes, i.e., one warming-up and two falling periods, implying that diffusion was the dominant physical mechanism controlling the moisture movement during the drying process. The diffusion model was, therefore, used to evaluate the drying (Demirel & Turhan, 2003; Prachayawarakorn et al., 2008; Kaddumukasa et al., 2005). Prachayawarakorn et al., (2008) also reported that drying temperature significantly affects the quality attributes of ripe bananas: color, shrinkage and texture of dried banana slices. Banana maturity (ripeness) has little influence on the drying kinetics, despite there being a significant difference in morphology and chemical composition between green and ripe bananas (Nguyen & Price, 2007; Baini & Langrish, 2007). However, there is little information available on the drying behavior of matooke (Musa sp. triploid acuminate AAA EAHB). Also no information is available relating the drying conditions to the physiochemical properties and quality attributes of matooke flour for industrial application. Introduction 3 Due to the increased market demand of quality products (Sablani, 2006) and consistency of product quality on the market, drying of fruits should comply with certain specifications like degree of maturity and use of healthy fruits, since drying does not improve the initial quality (Leite et al., 2007). Matooke, like any other fruit, undergoes a developmental process by which the fruit attains maturity. During the fruit development, starch grains are deposited initially in the pulp cells which form in the vicinity of the vascular bundles, and thereafter starch deposition moves centripetally and continues until fruit maturity (Robinson & Saúco, 2010). This indicates that maturity directly affects the quality of the final product. As Kader (1999) reported, maturity at harvest is the most important factor that determines storage life and final fruit quality. Traditionally, matooke maturity is judged by the visual appearance of the fingers, particularly the angularity (the fingers being full and rounded), the color (the color of the fingers being lighter) and the finger tips turning black (Dadzie, 1998). At present there are no specifications and procedures for estimating the harvest maturity age for commercial processing of matooke. On the other hand, moisture sorption isotherms of dried products are important for understanding the uptake of moisture during storage and distribution of dried commodities (Yan et al., 2008). They are also employed in process design and control, such as in predicting the end-point of drying (Sablani et al., 2007; Wang & Brennan, 1991). The moisture sorption isotherms can be used to obtain information on microbiological and shelf-stability of a product, as it was reported that microorganisms do not grow on food products with water activity below 0.6 (Iglesias and Chirife, 1982; Yan et al., 2008; Bezerra et al., 2013). In addition, moisture sorption isotherm characteristics of a product are important when introducing a new product on the market (Mujumdar, 2000). No moisture sorption isotherm has been reported for dried matooke. The functional application of the flour for use in food and non food industry depends on the physicochemical properties and quality attributes of the flour such as pasting properties and starch content (Niba et al., 2002; Zhang et al., 2005). Little information is available on the influence of maturity on starch content and pasting properties of matooke flour. The Government of Uganda in 2005 commissioned the pilot plant for the processing of matooke to flour under the Presidential Initiative on Banana Introduction 4 Industrial Development (PIBID). PIBID is centered on adopting the research findings on matooke flour (Muranga 1998), in order to diversify its utilization. PIBID has gone a step further to brand the matooke product as Tooke, received two patents: Patent No. AP/P/2005/003308, for Raw Tooke Flour (RTF), which is used for bakery, confectionery and extruded products; and Patent No. UG/P/04/00011, for Instant Tooke Flour (ITF), which is used for porridge and as a vehicle food for soups and instant foods for infants. This study has focused on Raw Tooke Four (referred to as matooke flour) because of its good starch properties and diverse utilization. However, there is still limited information available on the processing of matooke into flour. It was therefore imperative to carry out in-depth study to bridge the following gaps: • Lack of information on estimating the maturity window, in order to process flour of consistent quality at all times on the market. • Lack of information on moisture sorption isotherm. • Lack of information on drying behavior. • Lack of standardized process parameters in relation to physicochemical properties of flour. The main objective of the study was therefore to establish the optimum harvest maturity window and optimize the processing parameters for obtaining microbiologically shelf-stable matooke flour with good starch quality attributes. The specific objectives of the study on matooke were: 1. To establish the optimum harvest maturity window. 2. To establish the moisture sorption isotherms. 3. To establish the effect of process parameters on drying characteristics. 4. To optimize the process parameters. 5. Validation of the models of maturity and optimum process parameters. Literature review Literature Review 5 Chapter Two Chapter Two Literature Review Literature Review 2.1 Banana fruit maturity The physiological maturity is the stage of development when a plant or part of the plan 2.1 Banana fruit maturity continue ontogeny even if detached (Shewfelt, 2009; Kader, 1999). Harvest maturity, The physiological maturity is the stage of development when a plant or referred to as horticultural maturity or commercial maturity, is defined as the stag part of the plant will continue ontogeny even if detached (Shewfelt, 2009; development when a maturity, plant or part the plant possesses the maturity prerequisites for utilizatio Kader, 1999). Harvest also of referred to as horticultural or consumers maturity, for a particular purpose (Kader, The harvest maturity in this commercial is defined as the stage1999). of development when a plant or study refers part the when plant possesses theemerges prerequisites for stem utilization consumers timeoffrom inflorescence from the to theby time when the for sample is harves ahas particular purpose (Kader, 1999). The harvest maturity in this study refers been reported that matooke and bananas are generally harvested between three-quart tofull thematurity time from when inflorescence emerges from the stem to the time when period is depe (Muranga et al., 2007; Robinson & Saúco, 2010). The maturity the sample is harvested. It has been reported that matooke and bananas are on various environmental conditions that include: temperatures, rainfall, soil moisture, lat generally harvested between three-quarters to full maturity (Muranga et al., altitude and nutrients (nitrogen, phosphorous and potassium-(NPK), type of soil, cultiva 2007; Robinson & Saúco, 2010). The maturity period is dependent on various many other factors). Temperature hastemperatures, a strong influence on soil the moisture, physiological maturity p environmental conditions that include: rainfall, (Karugaba & Kimaru, 1999). (nitrogen, phosphorous and potassium-(NPK), latitude, altitude and nutrients type of soil, cultivar and many other factors). Temperature has a strong influence on the physiological maturity period (Karugaba & Kimaru, 1999). Chilletetetal., al.,(2006), (2006),reported reported aa close close relationship relationship between between the temperatures accumulated b Chillet accumulated by the during growth and physiological The fruit during growth andfruit its physiological age. Theitsphysiological ageage. of the fruit is calculated physiological age of the fruit is calculated from equation 2.1. equation 2.1. 𝑃ℎ𝑦𝑠𝑖𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝑎𝑔𝑒 (𝑑𝑎𝑦𝑠) = �00 𝑑𝑒𝑔𝑟𝑒𝑒−𝑑𝑎𝑦𝑠 2.1 𝐷𝑎𝑖𝑙𝑦 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑑𝑒𝑔𝑟𝑒𝑒−𝑑𝑎𝑦𝑠 Where: Where: Daily average degree-days = (Average daily temperature – Base temperature) Daily average degree-days = (Average daily temperature – Base temperature) The total degree-days are defined as heat/energy required for complete fruit The total degree-days are defined as heat/energy required for complete fruit growth. Ban growth. Bananas under normal growth conditions accumulate a total of 900 under normal accumulate a total of 2010; 900 Robinson degree-days (DD) to degree-days (DD)growth to reachconditions physiological maturity (Chillet et al., physiological maturity (Chillet et al., 2010; Robinson & Saúco, 2010; Chillet et al., 2006). daily degree-day is calculated from the daily mean temperature sums accumulated by the o above the temperature threshold (base temperature) 14 C (Bagaud et al., 2010; Nyombi Literature Review 6 & Saúco, 2010; Chillet et al., 2006). The daily degree-day is calculated from the daily mean temperature sums accumulated by the fruit above the Literature review temperature threshold (base temperature) 14oC (Bagaud et al., 2010; Nyombi et al., 2009; Chillet et al., 2006; Robinson & Saúco, 2010). 6 Literature 6 There arereview a number of changes in maturity indices both visual and fruitThere are a number of changes maturity indices both visual characteristic that occur duringinthe fruit growth period dueand to fruit-characteristic morphological that occur during theThe fruit visual growthchanges period due to morphological changes. The visual changes changes. include: size, shape, angularity, skin color andinclude: size, shape, angularity, skin color and nature of the stylar end; and the fruit characteristics changes nature of the stylar end; and the fruit characteristics changes include: fruit There are a number of changes in maturity indices both visual and fruit-characteristic that occur include:length, fruit weight, length, girth/circumference, and pulp-to-peel ratio (Dadzie, 1998; weight, girth/circumference, diameter,diameter, and pulp-to-peel ratio (Dadzie, during the fruit growth period due to morphological changes. The visual changes include: size, 1998; Muranga, 1998; Dhatt & Mahajan, 2007; Ramma, 1999; Singh,&2004; Muranga, 1998; Dhatt & Mahajan, 2007; Ramma, 1999; Singh, 2004; Dadzie Orchard, 1997). shape, angularity, skin color and nature of the stylar end; and the fruit characteristics changes Dadzie & Orchard, 1997). include: fruit weight, length, girth/circumference, diameter, and pulp-to-peel ratio (Dadzie, 1998; Muranga, 1998; Dhatt & Mahajan, 2007; Ramma, 1999; Singh, 2004; Dadzie & Orchard, 1997). The color of the pulp is also used as one of the maturity indices; it is traditionally The color of famers the pulpand is also usedwho as one indices; is traditionally used by the traders, lookofatthe thematurity intensity of theit yellowness of used by the famers traders, who lookcolor at the intensity the been yellowness of the pulp. There are several the pulp.and There are several scales thatofhave developed. However, The color of the pulp is also used as one of However, the maturitythe indices; it is“Lab” traditionally used by the color scales that have been developed. Hunter and its variant the Hunter “Lab” and its variant CIE L*a*b*, are the commonly used scales for CIE L*a*b*, famers and traders, who look at the intensity of the yellowness of the pulp. There are several are the commonly used(Shewfelt, scales for fruits andThe vegetables 2009). The fruits and vegetables 2009). Hunter (Shewfelt, L, a, b color scale isHunter more L, a, b color color scales that have been developed. However, the Hunter “Lab” and its variant CIE L*a*b*, scale is more visually uniform than the XYZ color scale. The Hunter L, a, b color space is visually uniform than the XYZ color scale. The Hunter L, a, b color space is are the commonly used scales for fruits and vegetables (Shewfelt, 2009). The Hunter L, a, b color organized in color values recorded as “L” = Black; 100 = 100 White), “a” (-a = organized in aa cube cubeform, form,with with color values recorded as(0“L” (0 = Black; scale is more visually uniform than the XYZ color scale. The Hunter L, a, b color space is =Green; White), (-a = and Green; +a ==Blue; Red),+band “b” (-b The = Blue; +b =(Saturation) Yellow). The +a“a” = Red), “b” (-b = Yellow). Chroma (C*) represents organized in a cube form, with color values recorded as “L” (0 = Black; 100 = White), “a” (-a = Chroma (Saturation) (C*) represents “richness of color or color saturation”, or “richness color or color saturation”, or=color intensity; hue angle depicts how an average person Green; +a of = Red), and “b” (-b = Blue; +b Yellow). The Chroma (Saturation) (C*) represents color intensity; hue angle depicts how an average person will perceive that will perceive that color (Hunter Lab, 2008, Siddiq et al., 2010). According to Falade & Olugbuyi “richness of color or color saturation”, or color intensity; hue angle depicts how an average person color (Hunter Lab, 2008, Siddiq et al.,Siddiq 2010). According to Falade & Olugbuyi (2010), the that hue angle is one parameter thatet isal., frequently used toto characterize color in food will perceive color (Hunter Lab, 2008, 2010). According Falade & Olugbuyi (2010), the hue angle is one parameter that is frequently used to characterize (2010), theThe huesaturation angle is one that isinfrequently used to characterize colorasinfollows: food products. (C*)parameter and hue angle degree (α) values are calculated color in food products.(C*) The and(α)hue angle in degreeas(α) values products. The saturation andsaturation hue angle in(C*) degree values are calculated follows: are calculated as follows: 𝐶 ∗ = √𝑎∗2 + 𝑏 ∗2 2.2 𝐶 ∗ = √𝑎∗2 + 𝑏 ∗2 −1 𝑏 ∗𝑏 2.2 ∗ 𝛼= = 𝑡𝑎𝑛 𝑡𝑎𝑛−1 � � �∗� 𝛼 ∗𝑎 2.3 2.3 𝑎 Shewfelt, (2009) reported that in human terms an increase in yellowness is signaled by the closeness of the hue angle (α) to 90o. Shewfelt, thatthat in human termsterms an increase in yellowness is signaled by the by the Shewfelt,(2009) (2009)reported reported in human an increase in yellowness is signaled o closeness ofofthe angle (α)program to 90 The Visual-Basic-based closeness thehue hue angle (α) to .90 . used to analyze images, transforms the RGB-values to XYZ-values in the first step, then in transforms the XYZ-values o The Visual-Basic-based program used to analyze images, transforms the RGB-values to XYZ- The Visual-Basic-based program used to analyze images, transforms the RGB-values to XYZ- values in the first step, then in transforms the XYZ-values to the CIE-L*a*b* color space in the values in the first step, then in transforms the XYZ-values to the CIE-L*a*b* color space in the Literature Review 7 to the CIE-L*a*b* color space in the second step (Sturm & Hofacker, 2009; Literature review 7 HunterLab, 2008; Sturm et al., 2012). The color difference ΔE* has beentoreported toimportant be the inmost important The total total color difference ΔE* has been reported be the most determining the color * inchange determining the color change in food (Sturm et al., 2012). The total color in food (Sturm et al., 2012). The total color difference ΔE is calculated as follows: difference ΔE* is calculated as follows: ΔE ∗ = �(ΔL∗2 + Δa∗2 + Δb ∗2 ) 2.4 The color change of dried bananas is estimated by subtracting the color of the The color changefrom of dried is estimated subtracting(Bani the color of the dried 2007). bananas dried bananas thebananas color of the freshbybananas & Langrish, from the color of the fresh bananas (Bani & Langrish, 2007). 2.2 Physicochemical properties Raw matooke flour is a generic developed to improve shelf stability of the fruit and to find alternative uses. It is rich in starch (80 - 85% db) and subsequently has a high potential as a calorie resource base (Muranga, 1998). Table 2.1 shows the average chemical composition of matooke. Table 2.1 Chemical composition of matooke from different cultivars Chemical Components Cultivar and Chemical Composition (%) 2.2 Physicochemical properties Nandigobe Bukumu Embururu Raw matooke flour is a generic developed to improve shelf stability of the fruit and to find (AAA-AE) (AAA-AE) (AAA-AE) Moisture Content 9.6 8.1 8.8 Starch matooke. 81.8 82.5 82.9 Protein 4.71 5.1 4.01 alternative uses. It is rich in starch (80 - 85% db) and subsequently has a high potential as a calorie resource base (Muranga, 1998). Table 2.1 shows the average chemical composition of Fat ND cultivars Table 2.1 Chemical composition0.87 of matooke from different 0.56 Crude Fiber 1.33 Ash Chemical Components 1.25 4.34 ND Cultivar and Chemical Composition (%) 3.58 Nandigobe Bukumu Embururu 0.0058 Potassium (K) Moisture Content 1.9 Magnesium (Mg) Starch 0.09 81.8 0.09 Tannin (Abs at 500nm) Protein 0.111 4.71 0.181 5.1 0.01 82.9 0.012 4.01 0.87 ND 0.56 1.25 ND 1.33 4.34 3.58 4.1 Calcium (Ca) 0.0058 0.0044 0.0052 Potassium (K) 1.9 1.82 1.84 (Source: Fat Muranga, 1998) Crude Fiber Ash (AAA-AE) 9.6 0.0044 4.1 Calcium (Ca) (AAA-AE) 1.82 8.1 82.5 0.0052 (AAA-AE) 1.84 8.8 8 Literature Review Physicochemical properties of flour and starchy products are important for their use in the food and non-food industry (Niba et al., 2002). Raw matooke flour consists of native starch which exists as microscopic granules consisting of two distinct polymers: the amylose, which are linear in structure, and amylopectin, which exhibits substantial branching in its structure as shown in Figure 2.1. Figure 2.1: Amylose & amylopectin molecular structures (source: Damodaran et al., 2008) Functional properties Pasting properties of flour are important in the sense that they give an indication of the cooking and baking qualities of the flour (Oluwalana et al., 2011). Peak viscosity is a measure of the water holding capacity of the starch in terms of the resistance of swollen granules to shear and the swelling performance of granules (Newport Scientific, 2006; Daramola & Osaosnylusi, 2006). It is often correlated with the final product quality (Kolawole et al., 2012). Final viscosity is the most commonly used parameter to define a particular sample’s quality, as it indicates the ability of the material to form a viscous paste or gel after cooking and cooling (Iglett et al. 2012; Osungbaro et al., 2010; Ikegwu et al., 2009; Niba et al. 2002; Muranga, 1998). Therefore, the integrity of the starch granules and hydration properties of starch can easily be investigated by measuring the pasting behavior of flour (Lai & Cheng, 2004). If the starch Literature Review 9 granules are damaged, they will take in more water and will reach the maximum swelling early and will collapse at a lower peak, compared to starch granules of the same sample which have not been damaged (Oluwalana et al., 2011). Figure 2.2 shows the standard rapid viscosity analyzer’s temperature profile and pasting curve. Table 2.2 shows the pasting properties of raw matooke flour from the given sources. Figure 2.2: Typical rapid visco-analyzer temperature profiles and pasting curve (Source: Newport Scientific, 2006). Table 2.2: Pasting properties of raw matooke flour (Source: Muranga et al., 2010) Peak (RVU) Trough (RVU) Breakdown (RVU) Final Viscosity (RVU) Setback (RVU) Peak Time (sec) Pasting Temp (oC) 375.92 201.29 173.63 244.7 43.38 4.7 74.73 2.3 Moisture sorption isotherms “Moisture sorption isotherms” is defined as a graphic relationship between water activity (equilibrium relative humidity) and the equilibrium moisture content of the material at a constant temperature (Thys et al., 2010; Veg-Galvez et al., 2007). Moisture content of the food product is very important for the design 10 Literature Review and optimizing of the drying equipment, the design of packages, the prediction of quality, stability, and shelf-life, and the calculation of moisture changes during storage (Andrade et al., 2011, Johnson & Brennan, 2000, Moreira et al., 2008). In order to preserve and stabilize the main food properties such as texture and microbiological stability, it is important to control its moisture content (Ayala et al., 2011). The state of moisture in a food product is expressed in terms of water activity (aw). Water activity (aw) is defined as the ratio of vapor pressure of water in food to vapor pressure of pure water at a constant temperature (Sahu & Tiwari, 2007). It is essential for describing water availability and mobility in foods (Ayala et al., 2011). Most processing operations are concerned with foods having a water activity (aw) range below 0.95, which are depicted by the general types of moisture sorption isotherms (Bell & Labuza, 2000). Moisture sorption isotherms are classified according to their shape in five different types: I, II, III, IV and V. However, Types II, IV and V are similar, with one or more inflection points (Yan et al., 2008; Andrade et al., 2010). Previous researches reported that the moisture sorption isotherms for bananas and plantains exhibited the Type II curves (Johnson & Brennan, 2000; Yan et al., 2008). Figure 2.3 shows the chemical ‘stability map’, relating the effect of water activity (aw) in the food material on the microbial growth. From the stability map, it is evident that during the intermediate moisture range, many reactions occur causing deterioration (Labuza, 1972). It has been reported that microorganisms do not grow on food products with water activity aw below 0.6 (Labuza et al., 1972; Sahu & Tiwari, 2007; Yan et al., 2008). The moisture sorption isotherm was used to determine the desired water activity from the stability map, below which the dried product would be microbiologically safe. Literature Review 11 Figure 2.3: Stability map of foods as a function of water activity (Source: Labuza et al., 1972). 2.3.1 Moisture sorption models Moisture sorption isotherms can be constructed by either an adsorption process, which starts from the dry state with water activity (aw) = 0, for which it takes in moisture to reach equilibrium moisture content at any given water activity to (aw) ≈ 1, or by a desorption process, which starts from the initial wet state with water activity (aw) ≈ 1 for which moisture is removed to reach equilibrium moisture content at any given water activity to (aw) ≈ 1 (Bell & Labuza, 2000, Yan et al., 2008). The shapes of the adsorption and desorption isotherms are sometimes different from each other for a given product; this phenomenon is referred to as moisture hysteresis (Bell & Labuza, 2000, Yan et al 2008). The hysteresis phenomenon indicates that the water adsorption and desorption processes are irreversible because fresh foods hold more moisture than dry foods (Vega-Gálvez et al., 2008). There are several models reported in literature. However, the following models have been fitted for bananas and plantains: Yan et al., (2008) reported that the GAB (1981), Henderson (1952), modified Hasley (1948), and modified Oswin fitted well the equilibrium moisture content data over 10 – 70% relative humidity. Ajibola, (1986) fitted Henderson, Chung-Pfost, modified Halsey, and modified Hasley (1948), and modified Oswin fitted well the equilibrium moisture content data over 10 – 70% relative humidity. Ajibola, (1986) fitted Henderson, Chung-Pfost, modified Halsey, and Chen-Clayton, the results of the standard error was less than 3%, indicating a good fit. Johnson Literature Review and and BET. Of the 12Brennan (2000) fitted five models: GAB, Henderson, Halsey, Iglesias-Chirfe five only Henderson fitted best the isotherm of fresh material. Phoungchandang & Woods (2000) Chen-Clayton, the results of the standard wasOswin, less than 3%, indicating fitted modified Chung-Pfost, modified Hernderson, anderror modified all of which gave a good good fit. and Brennan models: fit. a Falade and Johnson Awoyele (2005) fitted eight(2000) modelsfitted whichfive included: BET,GAB, GAB, Henderson, Oswin, Hasley, Halsey, Chung-Pfost, Iglesias-Chirfe BET. (1947). Of theOffive only fitted the best the Henderson, Chenand and Smith these onlyHenderson GAB best described sorption of fresh material. Phoungchandang & Woods fitted of isotherm fresh and pre-osmosed oven-dried banana slices. Aguire-Cruz et (2000) al. (2010) fittedmodified BET, GAB, Chung-Pfost, modified Hernderson, andGAB modified of which a Smith and Iglesias-Chirife. Of these only BET and gave theOswin, best fit. all Therefore, for gave this study thegood following models were selected to(2005) describe the eight moisture sorption isotherms for matooke: fit. Falade and Awoyele fitted models which included: BET, BET, GAB, Hernderson, Smith. Therefore, the Smith models(1947). selectedOffor GAB, Oswin, Hasley,Iglesias-Chirife Henderson, and Chung-Pfost, Chen and describing moisture isotherms for matookethe aresorption given in the these only GAB best described of Table fresh2.3. and pre-osmosed ovendried banana slices. Aguire-Cruz et al. (2010) fitted BET, GAB, Smith and Iglesias-Chirife. Of these only BET and GAB gave the best fit. Therefore, for this study the following models were selected to describe the moisture sorption isotherms for matooke: BET, GAB, Hernderson, Iglesias-Chirife and Smith. Therefore, the models selected for describing moisture isotherms for matooke are given in the Table 2.3. Table Moisture Model Equations to sorption describe Table 2.3: 2.3: Moisture IsothermIsotherm Model Equations used to describeused moisture isotherms for moisture matooke sorption isotherms for matooke Author Model 𝑎𝑤 𝑥𝑚 𝑐 (1 − 𝑎𝑤 )(1 + (𝑐 − 1)𝑎𝑤 ) BET (1938) 𝑥𝑒 = GAB (1981) 𝑥𝑒 = Henderson (1952) 𝐼𝑛(1 − 𝑎𝑤 𝑥𝑒 = � � −c2 Chung-Pfost (1967) Iglsias-Chirife (1976) Smith (1947) 𝑥𝑚 𝑐𝑘𝑎𝑤 (1 − 𝑘𝑎𝑤 )(1 − 𝑘𝑎𝑤 + 𝑐𝑘𝑎𝑤 ) 1� 𝑐1 −1 𝑇 ∗ 𝑙𝑛𝑎𝑤 𝑥𝑒 = � � ∗ 𝑙𝑛 � � (−𝑐1 ) 𝑐2 𝑥𝑒 = 𝑐1 + 𝑐2 ∗ � 𝑎𝑤 � (1 − 𝑎𝑤 ) 𝑥𝑒 = 𝑐1 − 𝑐2 ∗ 𝑙𝑛(1 − 𝑎𝑤 ) 2.3.2 Isosteric heat of sorption The isosteric heat of sorption is used to estimate the energy requirements for drying and provides important information on the state of water in food products (Yan et al., 2008). In addition, it provides valuable data for energy consumption important information on the state of water in food products (Yan et al., 2008). In addition, it 2.3.2 Isosteric heat ofon sorption important information the state of water in food products (Yan et al., 2008). In addition, it provides valuable data for energy consumption calculations and subsequent design of drying provides valuable data for energy consumption calculations and subsequent design of drying equipment (McMinn & Magee, 2003). The isosteric heat requirements of sorption for The isosteric heat of sorption is used to estimate the energy for desorption drying and isotherm provides equipment (McMinn & Magee, 2003). The isosteric heat of sorption for desorption isotherm represents the energy required to break the Van der Waals force between the molecules Literature Review important information on the state of water in food products (Yan et al., 2008).13 In addition, ofit represents the energy required to break the Van der Waals force between the molecules of moisture waterdata absorbent the during dehydration process and (Yang et al., 2012). Theofisosteric provides and valuable for energy consumption calculations subsequent design drying moisture and and watersubsequent absorbent the duringofdehydration process (Yang et al., 2012). The isosteric calculations design drying equipment (McMinn & Magee, heat of sorption is calculated from the experimental data using the Clausius-Clapeyron equation equipment (McMinn & Magee, 2003). The isosteric heat of sorption for desorption isotherm heat of sorption is calculated from the experimental data using the Clausius-Clapeyron equation 2003). TheBrennan, isosteric heatMcMinn of sorption for desorption isotherm represents the (Johnson 2000, & Magee, YanWaals et al., 2008) given in Equation 2.5: of represents& the energy required to break the 2003, Van der force as between the molecules (Johnsonrequired & Brennan, 2000, McMinn Magee, 2003, Yanbetween et al., 2008) given in Equation 2.5: energy to break the Van&der Waals force the as molecules of moisture and water absorbent the during dehydration process (Yang et al., 2012). The isosteric moisture and water absorbent the during dehydration process (Yang et al., heat of sorption is calculated from the experimental data using the Clausius-Clapeyron equation 2012). The isosteric heat of sorption is calculated from the experimental data (Johnson 2000, McMinn & Magee, 2003, Yan et al., 2008) as given in Equation 2.5: 𝑑(𝑙𝑛𝑎 ) & Brennan, 𝑄 −𝜆 q using equation (Johnson & Brennan, 2000, McMinn � Clausius-Clapeyron = 𝑠𝑡 = − �� 2.5& � 1 𝑤the 𝑑(𝑙𝑛𝑎 𝑄𝑠𝑡𝑅−𝜆 q𝑅�� 𝑑� �𝑤 ) 𝑇 =− 2.5 � 1 �𝑥2003, 𝑒 = Magee, Yan et al., 2008) as given in Equation 2.5: 𝑅 𝑅 𝑑� � 𝑇 𝑥𝑒 Where: 𝑑(𝑙𝑛𝑎 ) 𝑄 −𝜆 q Where: � 1 𝑤 � = 𝑠𝑡 = − �� 𝑅 𝑅 𝑑� � 𝑇 Q 𝑥𝑒st – Total isosteric heat of sorption (kJ/kg) Qst – Total isosteric heat of sorption (kJ/kg) 2.5 Where: qst – Net isosteric heat of sorption (kJ/kg) Where: qst – Net isosteric heat of sorption (kJ/kg) Total isosteric heatheat of sorption (kJ/kg) λ stQ - –Latent heat of water (kJ/kg) – Total isosteric of sorption (kJ/kg) Q st λ - Latent heat of water (kJ/kg) – Net isosteric heat of sorption (kJ/kg) R Net isosteric of sorption (kJ/kg) qstq-–stUniversal gasheat constant R - Universal gas constant λ - Latent heat of water (kJ/kg) λ -Equation Latent heat water (kJ/kg) Integrating 2.5,ofwe obtain the following Equation 2.6: REquation - Universal gasobtain constant Integrating 2.5, we the following Equation 2.6: R - Universal gas constant Integrating Equation 2.5, we obtain the following Equation 2.6: 𝑎 �𝑠𝑡 1 1 Integrating 𝑙𝑛 � 𝑤2� = −Equation � + 2.5, � we obtain the following Equation 2.6: 𝑎𝑤1 𝑇11 �𝑅𝑠𝑡 𝑇12 𝑤2 𝑙𝑛 � � = − � + � 𝑎𝑤1 𝑅 𝑇2 𝑇1 2.6 2.6 𝑎 net isosteric � 1 heat 1 The of sorption is determined from the slope of the graph 𝑙𝑛 � 𝑤2 � =isosteric − 𝑠𝑡 � heat + �of sorption 2.6) versus (1/T) The net is determined from the slope of the graph of ln(a 𝑎𝑤1 𝑅 𝑇2 𝑇1 w versusheat (1/T)of for a specific moisture content. The ofofthe of Theln(a netw)isosteric sorption is determined from the slope of values the graph ln(anet w) versus (1/T) for a specific moisture content. The values of the net isosteric heat of sorption obtained are fitted isosteric heat of sorption obtained are fitted by the empirical relationship for a specific moisture content. The values of the net isosteric heat of sorption obtained are fitted by the empirical relationship between the isosteric heat of sorption and moisture content between the isosteric heat of sorption moisture (Equation 2.6) by the empirical relationship between theand isosteric heatcontent of sorption and moisture content (Equation 2.6) asheat proposed by Tsami (1991). from The the energy required for breaking the Van der The net isosteric of sorption isThe determined slope of the graph ln(a w) versus (1/T) as proposed by Tsami (1991). energy required for breaking the of Van der (Equation 2.6) as proposed by Tsami (1991). The energy required for breaking the Van der Waals force drying is obtained by integrating 2.7: for a specific moisture content. The values the netEquation isosteric heat of sorption obtained are fitted Waals forceduring during drying is obtained byofintegrating Equation 2.7: Waals force during drying is obtained by integrating Equation 2.7: by the empirical relationship between the isosteric heat of sorption and moisture content −� � eas � proposed by Tsami (1991). The energy required for breaking 2.7 the Van der q �� = q� exp (Equation 2.6) � −� 2.7 q�� = q� exp � �e � �� Waals force during drying is obtained by integrating Equation 2.7: −� 2.4 exp � e � process for bananas q = q Drying �� � �� 2.7 This section describes the drying behavior of bananas. It is defined by moisture diffusivity and drying rate constant. These were then used to compare the effect of the processing parameters. The activation energy was also described 14 Literature Review to estimate the energy requirement for drying as compared to the net isoteric heat of sorption for desorption isotherm. The engineering, quality properties and drying methods have also been discussed in order to establish how they are affected during drying. Drying is the oldest technology used for food preservation to extend shelf life, minimizing postharvest losses, bulk transportation and storage space. Drying is simultaneously a heat and mass transfer process, in which heat is transferred to the product by drying air and moisture is transferred from the product by diffusion to the product surface and the drying air stream picks the moisture from the surface (Lykov 1958 as reported by Corzo et al., 2008). A successful drying process requires the conditions of the drying air to have the capacity to transfer heat to the drying product and also pick up the moisture from the product surface. In the convection drying system, drying kinetics is affected by drying air temperature, air humidity, air velocity and characteristic particle size (sample thickness) (Krokida et al., 20002; Nguyen & Price, 2007; Corzo et al., 2008; Sturm et al., 2012). However, Krokida et al. (2000) reported that of all those factors, the drying kinetics is greatly affected by drying air temperature and the characteristic particle size. Drying affects the properties of the dried product because of the physicochemical changes that occur within the product during drying (Krokida et al., 2000; Leite et al., 2007). Krokida et al. (2000) further reported that the properties of dried products are classified between two major categories, the engineering and quality properties. The engineering properties include: effective moisture diffusivity, effective thermal conductivity, drying kinetics, specific heat and equilibrium moisture content; meanwhile, the quality properties include: thermal properties, structural properties, textural properties, optical properties, sensory properties, nutritional characteristics and rehydration properties. However, Krokida et al. (2000) reported that of all those factors, the drying kinetics is greatly affected by drying air temperature and the characteristic particle size (thickness of banana slice). Studies on banana drying characteristics and quality attributes have been carried out considering the utilization of the dried product and quality product requirements. In this study effective moisture diffusivity, drying kinetics and equilibrium moisture content were the engineering properties considered; color and physicochemical properties (starch content and pasting properties) were considered for quality properties as described in Literature Review 15 Section 2.2. From the study on the effect of drying temperature on the quality of dried bananas (ripe bananas), Leite et al. (2007) reported that drying did not affect the chemical composition of the product. Previous researchers on bananas reported that the drying time decreases with increased temperature, and the drying rate decreases with increase in thickness of slice (Nguyen & Rice, 2007; Islam et al., 2012). The drying method/technology employed in drying has a direct effect on the final dried product (Nguyen & Price, 2007; Bani & Langrish, 2007). Convectional hot-air drying and sun drying have been reported to be largely used in the drying of unripe bananas (Johnson et al., 1998; Islam et al., 2012). Pacheco-Delahaye et al., (2008), reported the effect of drying methods/ technologies employed in drying on starch content (% db) of dried bananas as follows: freeze drying (74.65±2.08), drum drying (63.50±0.55), microwave drying (64.52±0.25) and drying chamber (hot-air drying) (74.30±2.32). These results showed that although freeze drying was superior compared to the convectional hot–air drying, the starch content for both (freeze drying and hot-air drying) were not significantly different. Also, Krokida et al. (1999) reported that dehydrated products do not keep their viscoelastic behavior after rehydration due structural damages that occur during drying. However, hot-air and vacuum-dried products kept their viscoelastic characteristics close to those of dehydrated products before rehydration. In addition, hot-air drying is considered as one of the simplest and most economic ways for commercial processing of fruit and vegetables and appropriate for developing countries (Johnson et al., 1998; Kaddumukasa et al., 2005). Therefore, the hot-air drying method was selected for this study. Drying of food products is not limited to selection of dryer, but rather there is also need to understand the physicochemical concepts associated with food drying in order to appropriately assess the drying phenomena of any food product (Vega-Mercado et al., 2001). In terms of food applications, starch functionality is largely related to its gelatinization and pasting characteristics (Zhang et al., 2005). It has been reported that gelatinization temperature of matooke is 72˚C (Muranga, 1998). Drying of bananas dominantly takes place in the falling rate periods, implying that diffusion was the mechanism controlling the moisture migration Literature Review 16 (Prachayawarakon et al., 2008; Demirel & Turhan, 2003; Kaddumukasa etLiterature al., 2005; Bani & Langrish, 2007). The diffusion model based on Fick’s review 16 Literature review 16 second law is used to describe the transport of moisture inside a banana Literature review 16 (Thuwapanichayaan et al., 2011; Langrish, Fick’s law2007). of diffusion has been used slice (Thuwapanichayaan et Bani al., &2011; Bani2007). & Langrish, Fick’s law of by a number researchers describe drying kinetics bananas, and diffusion hasofbeen used by a number of researchers to describe the vegetables drying (Thuwapanichayaan et al., to 2011; Bani the & Langrish, 2007).of Fick’s lawfruits of diffusion has been used (Thuwapanichayaan et al., 2011; Bani & Langrish, 2007). Fick’s law ofKasheninejd diffusion has been used (Karim & Hawlader, 2005; Baini & Langrish, 2007; Nguyen & Price, (2007); et&al., kinetics of bananas, fruits and vegetables (Karimkinetics & Hawlader, 2005; Baini by a number of researchers to describe the drying of bananas, fruits and vegetables Literature review 16 by a number of researchers to describe the drying kinetics of bananas, fruits and vegetables (2007); Doymaz, Katekawa & Silva (2007);Kasheninejd Islam et al., 2012).al., (2007); Doymaz, Langrish, 2007;(2007); Nguyen Price, (2007); (Karim & Hawlader, 2005; &Baini & Langrish, 2007; Nguyenet & Price, (2007); Kasheninejd et al., (Karim & Hawlader, 2005; Baini & Langrish, 2007; Nguyen & Price, (2007); Kasheninejd et al., (2007); & Silva (2007);& Islam et al., Islam 2012).et al., 2012). (2007); Katekawa Doymaz, (2007); Katekawa Silva (2007); (Thuwapanichayaan et al., Katekawa 2011; Bani & & Silva Langrish, 2007). Fick’s law2012). of diffusion has been used (2007); Doymaz, (2007); (2007); Islam et al., by a number of law researchers to describe the expressed drying kinetics of bananas, fruits and vegetables Fick’s second of diffusion be & Fick’s second law of diffusion can becan expressed as Equationas 2.8Equation (Nguyen &2.8 Price(Nguyen 2007, Doymaz (Karim & Hawlader, 2005; Baini & Langrish, 2007; Nguyen & Price, (2007); Kasheninejd et al., Price 2007):2007, Doymaz 2007): (2007); Doymaz,law (2007); Katekawa & Silva (2007); Islam et al., 2012). Fick’s second of diffusion can be expressed as Equation 2.8 (Nguyen & Price 2007, Doymaz Fick’s second law of diffusion can be expressed as Equation 2.8 (Nguyen & Price 2007, Doymaz 2007): �2 � �� = 𝐷𝑒𝑓𝑓 2 2007): �𝑡 �𝐿 2.8 Fick’s second �2 � law of diffusion can be expressed as Equation 2.8 (Nguyen & Price 2007, Doymaz �� =analytical 𝐷𝑒𝑓𝑓 �2 �2 solution of Fick’s second law of diffusion was obtained on the assumption that 2.8 The �� �𝑡 2007): = 𝐷𝑒𝑓𝑓 �𝐿2 �𝑡 �𝐿 The analytical solution of diffusion, Fick’s second law of diffusion was obtained on 2.8 moisture migration was only by negligible shrinkage, constant temperature and diffusion the assumption that From moisture migration was only by negligible constant (Crank, 1975). the analytical Equation 2.9,diffusion, the dimensionless amount The analytical solution of Fick’s second solution law of of diffusion was obtained on the assumption that �2 � �� = 𝐷analytical 2.8 𝑒𝑓𝑓 �𝐿2constant The solution of Fick’s second law of diffusion was obtained on the assumption that shrinkage, temperature and diffusion constant (Crank, 1975). From �𝑡 of diffusing water (Moisture Ratio – MR) in the slab was deduced Crank, 1975 to Equation 2.9: moisture migration was only by diffusion, negligible shrinkage, constant temperature and diffusion moisture migration was only by diffusion, shrinkage, constant temperature and diffusion the analytical solution of Equation 2.9,negligible the dimensionless amount of diffusing constant (Crank, 1975). From the analytical solution of Equation 2.9, the dimensionless amount 2� 2 𝐷 of𝑡 diffusion was obtained on the assumption that The analytical solution of1–Fick’s second law −(2𝑛+1) 𝑥− 𝑥(Crank, 8 water MR) slab deduced Crank,2.9, 1975 Equation amount constant 1975). From thethe analytical of Equation theto dimensionless 𝑒𝑓𝑓was 𝑒 � = 2 ∑Ratio 𝑒𝑥𝑝 � in �solution 2.9 𝑀𝑅 = (Moisture 𝑛=1 of diffusing (Moisture 4 𝐿2 in the slab was deduced Crank, 1975 to Equation 2.9: 𝑥𝑜 − 𝑥𝑒 water � (2𝑛+1)2 Ratio – MR) moisture migration was only by diffusion, negligible shrinkage, constant Crank, temperature diffusion 2.9: 2.9: of diffusing water (Moisture Ratio – MR) in the slab was deduced 1975and to Equation constant (Crank, 1975). From the analytical solution of Equation 2.9, the dimensionless amount 2 �2 𝐷 Equation It has been a long time drying 2.9 Crank, is expressed Equation −(2𝑛+1) 𝑡 deduced 𝑥− 𝑥water 8(Moisture 1Ratio 𝑒 reported � that for of diffusing – MR) in the slab𝑒𝑓𝑓was 1975 toas Equation 2.9:2.10 𝑀𝑅 = 𝑥− 𝑥 = 82 ∑𝑛=1 1 2 𝑒𝑥𝑝 �−(2𝑛+1)2 �22 𝐷𝑒𝑓𝑓 𝑡� 4𝐿 � 2 𝑥𝑒𝑒 (2𝑛+1) (Nguyen 2007): ∑� = �2007; 𝑀𝑅 = 𝑥𝑜&− Price 𝑛=1Doymaz 2 2 𝑒𝑥𝑝 � 𝑥𝑜 − 𝑥𝑒 𝑥− 𝑥 8 � (2𝑛+1) 1 4𝐿 −(2𝑛+1)2 �2 𝐷 2.9 2.9 𝑡 𝑒𝑓𝑓 𝑒 = 2 ∑� 𝑒𝑥𝑝 � � 2.9 𝑀𝑅 = 𝑛=1 (2𝑛+1) 4 𝐿2time drying Equation 2.9 is expressed as Equation 2.10 −�𝑥𝑒𝑒 �8 − that �2 𝐷2𝑒𝑓𝑓for 𝑡 a long 𝑜 It has 𝑥�− been reported = �𝑒𝑥𝑝 � �� 2.10 as Equation 2.10 𝑀𝑅 = 2 2 hasbeen reported that 2.92.9 is expressed 4 𝐿 for �been � 𝑜 − �𝑒 reported ItIthas that for aa long longtime timedrying dryingEquation Equation is expressed as (Nguyen & Price 2007; Doymaz 2007): (Nguyen &2.10 Price 2007; 2007): Equation (Nguyen & Price 2007; Doymaz 2007): It has been reported thatDoymaz for a long time drying Equation 2.9 is expressed as Equation 2.10 Integrating Equation 2.10, for low xe values and MR<0.6, can be written in logarithmic form as (Nguyen & Price 2007; Doymaz 2007): 2 − � 𝐷𝑒𝑓𝑓 𝑡 �− �𝑒 8 2 𝐷2 �− �2012): �� 2.10 𝑀𝑅 = �−2.11 Equation (Islam et al., �𝑒 = 82 �𝑒𝑥𝑝 𝑒𝑓𝑓 𝑡 𝑀𝑅 = �𝑜 − �𝑒 �𝑜 − �𝑒 �− �𝑒 = �2 �𝑒𝑥𝑝 � 8 � 4𝐿 4 𝐿2 − �2 𝐷𝑒𝑓𝑓 𝑡 �� 2.10 = 2 �𝑒𝑥𝑝 � �� 2.10 𝑀𝑅 = �𝑜 − �8 � �2 𝐷𝑒𝑓𝑓 𝑡 4 𝐿2 𝑒 � = 𝐿𝑛 2 Equation − � � 2.11 𝐿𝑛 Integrating 4 𝐿2 2.10, for low xe values and MR<0.6, can be written in logarithmic form as �𝑜 � Integrating Equation 2.10, for low x values and MR<0.6, can be written in logarithmic form as e Integrating Equation for low x values and MR<0.6, can be written in Equation 2.11 (Islam et2.10, al., Integrating Equation 2.10, for 2012): low x valuese and MR<0.6, can be written in logarithmic form as Equation 2.11 (Islam et al., 2012):e logarithmic form as Equation 2.11 (Islam et al.,time, 2012): Plotting the natural logarithm of moisture ratio versus the effective diffusivity is calculated Equation 2.11 (Islam et al., 2012): from� the slope �2 𝐷𝑒𝑓𝑓2.12): 𝑡 8 (Equation 𝐿𝑛 � = 𝐿𝑛2 82 − ��2 𝐷𝑒𝑓𝑓 2 𝑡� 4𝐿 � �𝐷𝑒𝑓𝑓 − ���2 𝐷𝑒𝑓𝑓 � 𝐿𝑛 ��𝑜 = 𝐿𝑛 𝑡 𝑆𝑙𝑜𝑝𝑒 � 8�2− 4 𝐿2� 𝐿𝑛 �𝑜 ==𝐿𝑛 4 𝐿2 � 2 2 �𝑜 4𝐿 � 2.12 2.11 2.11 2.11 Plotting the natural logarithm of moisture ratio versus time, the effective diffusivity is calculated The drying rate constant has been usingtime, a first order approach described Plotting the natural logarithm ofempirically moisture ratio versus the effective diffusivity is by calculated Plotting natural logarithm of moisture ratiopredicted versus time, the effective diffusivity is calculated from thethe slope (Equation 2.12): Equation 2.13 (Baina & Langrish, 2007; Hofsetz et al., 2007; Doymaz, 2007): from the slope (Equation 2.12): from the slope (Equation 2.12): �2 𝐷 𝑆𝑙𝑜𝑝𝑒 = ��2�2𝐷 𝐷𝑒𝑓𝑓 2 � 𝐿𝑒𝑓𝑓 � 𝑆𝑙𝑜𝑝𝑒 == �� 4 4𝐿𝑒𝑓𝑓 𝑆𝑙𝑜𝑝𝑒 2 2� 4𝐿 2.12 2.12 2.12 Integrating Equation 2.10, for low xe values and MR<0.6, can be written in logarithmic form as Equation 2.11 (Islam et al., 2012): Literature Review2 𝐿𝑛 � �𝑜 = 𝐿𝑛 8 �2 − � 17 � 𝐷𝑒𝑓𝑓 𝑡 � 4 𝐿2 2.11 Plotting the natural logarithm of moisture ratio versus time, the effective Plotting theisnatural logarithm of moisture ratio(Equation versus time, the effective diffusivity is calculated diffusivity calculated from the slope 2.12): from the slope (Equation 2.12): 𝑆𝑙𝑜𝑝𝑒 = � �2 𝐷𝑒𝑓𝑓 4 𝐿2 � 2.12 The drying constant has been predicted using a first order approach described The dryingrate rate constant hasempirically been empirically predicted using a first order by Literature review 17 Equation 2.13 (Baina & Langrish, 2007; 2.13 Hofsetz et al., 2007; Doymaz, 2007; 2007): Hofsetz et al., approach described by Equation (Baina & Langrish, Literature review 2007): 17 2007; Doymaz, �−� 𝑒 Literature review = exp(−𝑘𝑡) 𝑀𝑅 = � −� 𝑀𝑅 = 17 2.13 �𝑜 −�𝑒 �−�𝑒 = exp(−𝑘𝑡) Where:�𝑜−� k 𝑒– drying rate constant 2.13 Where: k – drying rate constant �−�Where: k – drying 𝑒 = exp(−𝑘𝑡) 𝑀𝑅 = rate constant � −� 2.13 𝑜 𝑒 The drying rate constant can be obtained by integrating Equation 2.13 to get Equation 2.14. The Where: k –rate drying rate constant The drying constant can be obtained by integrating Equation 2.13 to get Equation 2.14. The The drying constant can befrom obtained by integrating Equation 2.13 to get drying rate rate constant is obtained the slope of a plot of natural logarithm against time drying rate constant is obtained from the slope of a plot of natural logarithm against time Equation (Equation 2.14. 2.15). The drying rate constant is obtained from the slope of a plot of The drying2.15). rate constant can be obtained by integrating Equation 2.13 to get Equation 2.14. The (Equation natural logarithm against time (Equation 2.15). drying rate constant is obtained from the slope of a plot of natural logarithm against time (Equation 2.15). (𝑀𝑅)==−𝑘𝑡 𝐿𝑛 (𝑀𝑅) −𝑘𝑡 𝐿𝑛 2.14 𝐿𝑛 (𝑀𝑅) = −𝑘𝑡 2.14 2𝐷 �2�𝐷 𝑒𝑓𝑓 𝑒𝑓𝑓 4 𝐿2 2 𝑘 𝑘= = 𝑆𝑙𝑜𝑝𝑒 𝑆𝑙𝑜𝑝𝑒==� � 𝑘 = 𝑆𝑙𝑜𝑝𝑒 = � 2.14 4𝐿 �2 𝐷𝑒𝑓𝑓 4 𝐿2 � � � 2.15 2.15 2.15 The activation energy Ea represents the energy required to break bananathe energy required to break banana-moisture bonding of a The activation energy E represents the energy required to break banana-moisture bonding of a The activation energy aEa represents moisture bonding of a water molecule and to bring it to the surface where(Bani the & water molecule and to bring it to the surface where the molecule finally evaporates water molecule and to bring it to the surface where the molecule finally evaporates (Bani & molecule finally (Bani & Langrish, 2007). Ininterpreted the caseasofthe drying, theactivation energy required banana-moisture bonding of a The activation energy can be minimum Langrish, 2007). In evaporates theEcase of drying, energy to E a represents a,break Langrish, 2007). In the case of drying, activation energy E a, can be interpreted as the minimum water to the surface where the water-water molecule finally evaporates (Bani & , canittobe interpreted as the minimum energy that activation Eabring energymolecule that energy mustand be to supplied break water-solid and/or interactions and must to move energy that must be supplied to break water-solid and/or water-water interactions and to move be&interpreted as the minimum Langrish, the case drying, activation energy the water 2007). molecules from oneofpoint to another inwater-water the solidE(Demirel Turhan, and 2003). be supplied to In break water-solid and/or interactions to Arrhenius’ move a, can the water from one point to another in the solid (Demirel &diffusivity Turhan, 2003). Arrhenius’ energy that must to bedescribe supplied to break water-solid and/or interactions to move Equation ismolecules used temperature on moisture (Nguyen & the water molecules fromthe one point to dependence another in water-water thethe solid (Demirel &and Turhan, Equation is Doymaz, used tofrom describe the to temperature dependence on &the moisture diffusivity (Nguyen & the water molecules one point another in the solid (Demirel Turhan, 2003). Arrhenius’ Price, 2007; 2007) (Equation 2.16): 2003). Arrhenius’ Equation is used to describe the temperature dependence Price, Doymaz, 2007)the (Equation 2.16): Equation is used to diffusivity describe temperature dependence on Doymaz, the moisture2007) diffusivity (Nguyen & on the2007; moisture (Nguyen & Price, 2007; (Equation Price, 2007; Doymaz, 2007) (Equation 2.16): 2.16): 𝐸𝑎 𝐷𝑒𝑓𝑓 = 𝐷𝑜 𝑒𝑥𝑝 �− 𝐷𝑒𝑓𝑓 = 𝑅𝑇 � 𝐸 𝐷𝑜 𝑒𝑥𝑝 �− 𝑎 � 𝐸𝑎𝑅𝑇 𝐷𝑒𝑓𝑓 = 𝐷𝑜 𝑒𝑥𝑝 �− 𝑅𝑇 � 2.16 2.16 2.16 Equation 2.16 can be written in logarithmic form as Equation 2.17: Equation 2.16 can be written in logarithmic form as Equation 2.17: 𝐸 Equation 𝑙𝑛𝐷𝑒𝑓𝑓 = 2.16 𝑙𝑛𝐷𝑜 can − 𝑎be written in logarithmic form as Equation 2.17: 𝑅𝑇 𝐸𝑎 𝑎 𝑙𝑛𝐷𝑒𝑓𝑓 == 𝑙𝑛𝐷 𝑙𝑛𝐷𝑜−−𝑎 𝑙𝑛𝐷 𝑅𝑇 is calculated from the slope of the plot on ln(D ) versus 1/(T). 𝑒𝑓𝑓 The activation𝑜 energy 𝑅𝑇 eff 𝐸 2.17 2.17 2.17 𝐸 𝐷𝑒𝑓𝑓 = 𝐷𝑜 𝑒𝑥𝑝 �−𝐸𝑎 𝑎 � 𝐷𝑒𝑓𝑓 = 𝐷𝑜 𝑒𝑥𝑝 �− 18 𝑅𝑇 � 𝑅𝑇 2.16 2.16 Literature Review Equation 2.16 can be written in logarithmic form as Equation 2.17: Equation 2.16 can be written in logarithmic form as Equation 2.17: Equation 2.16 can be written in logarithmic form as Equation 2.17: 𝐸 𝐸𝑎 𝑎 𝑙𝑛𝐷 𝑙𝑛𝐷𝑒𝑓𝑓 = 𝑙𝑛𝐷 𝑙𝑛𝐷 𝑜 𝑜−− 𝑒𝑓𝑓 = 2.17 𝑅𝑇𝑅𝑇 The activation energy is calculated the slope of ln(D the plot on ln(D The is is calculated from thefrom slope of the versus 1/(T).eff) versus The activation activationenergy energy calculated from the slope ofplot theon plot oneff)ln(D eff) versus 1/(T). 1/(T). 𝐸𝑎 = 𝑅 ∗ 𝑆𝑙𝑜𝑝𝑒 𝐸𝑎 = 𝑅 ∗ 𝑆𝑙𝑜𝑝𝑒 2.18 2.18 2.17 Establishment of optimum harvest maturity window for matooke 19 Chapter Three: Objective: Establishment of optimum harvest maturity window for matooke 3.1 Background Matooke has now become both a food and cash crop in Uganda. In the event of commercial processing of it into flour, it requires the flour to have a uniform quality on the market at all times. This calls for processing matooke of uniform physicochemical properties, which can be reproduced on the market. It was therefore, imperative to establish the harvest maturity window as an initial input to achieving flour of the same physicochemical properties. The objective of this study was to establish the optimum harvest maturity window for matooke through the following specific objectives: To establish the effect of harvest maturity with respect to: 1. Maturity indices. 2. Pasting properties of matooke flour. 3. Starch content. 3.2 Materials and methods Samples were obtained from thr banana plantation established on 20th December 2007 at Presidential Initiative on Banana Industrial Development (PIBID) field station in Bushenyi, Western Uganda. A completely randomized design (CRD) was employed in selecting the banana stools from which samples for the experiments were picked. The banana stools were monitored and coded as soon as the flower shot up (inflorescence emerging from the stem) (Figure 3.1), in order to establish harvest maturity at the time of picking samples of the banana fruit. The cultivar Mbwazirume (Musa sp triploid acuminate genome group AAA-EAHB) which is soft cooking and commonly grown was selected for the study. 20 Establishment of optimum harvest maturity window for matooke Figure 3.1: Illustration of (from left to right) inflorescence emerging from the stem, coding on the matooke stool, growing bunch on the matooke stem in the banana plantation. The maturity indices considered in these investigations were: circumference, finger weight, pulp-peel ratio, color of the pulp, moisture content and drymatter content of the fingers for matooke at different harvest maturity (10 - 23 weeks). The samples for determining pasting properties and starch content of flour were dried at 55˚C in a cabinet dryer. A total of five fingers were picked at random from a bunch from middle to top, because fingers from the bottom are 30 – 40% smaller than those from top (Robinson & Saúco, 2010). The fruit circumference was measured by a tape measure around the middle of the fruit. The fruit weights were measured on an electronic balance. The fingers were hand- peeled, and then the pulp and peel were weighed separately on the electronic balance. The same data was used to calculate the pulp-peel ratio. The moisture content of fresh matooke was determined using the oven method at 105˚C for 24 h (Sablani et al., 2007; Bani & Langrish, 2007). Five fingers from each sample bunch of a given harvest maturity were sliced in the middle and a total of five slices for each finger were photographed. The photographs were taken using a digital camera with a ring of light to provide clear photographs. Establishment of optimum harvest maturity window for matooke 21 Table 3.1 Specification of the measuring equipment used in the study Type of measurement Equipment Model Manufacturer Accuracy Finger circumference Tape measure Finger, pulp & peel weight Electronic balance Explorer Pro RS232 OHAUS Corporation ±0.01 g Pulp color IC Capture 2.0 camera Sony ICX204AK Sony Corporation 1024x768 pixel Slicing Slicing machine V20, Gabr. Graef GmbH & Co. KG ±0.1mm Milling flour Starch Mill IKA M20 IKA WERKER GmbH & Co. Pasting properties Rapid Viscosity Analyzer Model Series 4 Newport Scientific Pty Limited, Warriewood, Australia Starch content Polarimeter Model AP 300 Air velocity Hot wire anemometer ±0.02 m/s Air and dew temperatures Pt-100 Thermal resistor ±0.2K OHAUS Corporation, Parspary, USA ±0.01˚ Table 3.1 above shows the measuring equipment used in these investigations. All experiments were carried out at the PIBID - Technical Business Incubator (TBI). The samples for determining pasting properties and starch content were harvested at harvest maturity from Week 10 – 23, washed, peeled and sliced to specified uniform thickness using a slicing machine. The sliced samples were pretreated with 1% sodium metabisulphite solution as per Patent No. AP/P/2005/003308 (Muranga 2010) and dried at 55oC. The dried matooke chips were milled to flour using the starch mill with double-walled grinding chamber cooled with running water. 22 Establishment of optimum harvest maturity window for matooke Pasting Properties The pasting properties were determined using the Rapid Visco Analyzer (RVA). The following pasting properties were recorded: peak time, peak viscosity, pasting temperature, breakdown, holding strength, setback and final viscosity. The pasting properties were determined using the following procedure: 3 g of flour (adjusted to 14% wb) and 25 g of distilled water (a total of 28 g) were placed in an RVA canister. The moisture content of the sample was used to determine the amount of distilled water to be added in order to maintain the same amount of solids in the sample. The RVA pasting curve was obtained by using a 13-minute test profile which included heating to 90oC for 5 minutes, holding at 90oC for 3 minutes, and cooling to 50oC for 5 minutes. The peak viscosity, breakdown, set-back, peak time and final viscosity were obtained using the RVA software. Starch Content The starch content was determined using a polarimeter by employing the general polarimeter method (Kirk & Sawyer, 1991) in two parts: total optical rotation (P) and optical rotation (P’) in duplicates, plus the blank samples. 2.5 g of matooke flour was weighed in flask of 100 ml, and 50 ml of HCL (11.28 g/l) was added mixed well then placed in boiling water bath for exactly 15 min (the flask was continuously shaken while in bath for the first 3 min). The flask was removed from the bath, added 20 ml of cold water and left it cool to room temperature. The cold sample was then put in 100 ml volumetric flask, 5 ml of Correz 1 solution was added and shaken for 1 min then 5 ml of Correz II was added and shaken well, it was then diluted with distilled water to 100 ml mark, and filtered. The filtrate was used to determine the optical rotation (P) using the polarimeter cell 200 mm (Range 4o – 8o) 5 g of matooke flour sample was weighed and placed in a volumetric flask of 100 ml, 40% ethanol was added to 100 ml mark and shaken for 6 times in 1 hr then filtered. 50 ml of filtrate was placed in 250 ml quick-fit conical flask, 2.1 ml HCL (density = 1.126), was added and shaken, then put on boiling plate with reflux cooler for 15 min and allowed to cool to room temperature. After cooling, 50 ml of solution was put in volumetric flask of 100 ml, 5 ml of Carrez I was added and shaken for 1 min and 5 ml of Carrez II was added then diluted with distilled to the 100 ml mark, shaken and filtered. The filtrate Establishment harvest maturity window for matooke Establishmentofofoptimum optimum harvest maturity window for matooke 23 Establishment of optimum harvest maturity window for matooke 2 23 Establishment of optimum harvest maturity window for matooke ′) was to determine the optical2000(𝑃−𝑃 rotation (P’), using the polarimeter cell 200 mm ′) 3.1 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑜𝑓 𝑆𝑡𝑎𝑟𝑐ℎ (𝑑𝑏) = 2000(𝑃−𝑃 o 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 (𝑑𝑏) = 𝛼 ).𝑆𝑡𝑎𝑟𝑐ℎ (Range 4o – 8𝑜𝑓 𝛼 Establishment of optimum harvest maturity window for matooke 23 3.1 23 Angular rotation (α) at 20˚C for matooke flour was approximated to 184.0 for mixed flour (Kirk & The percentage starch of the sample was calculated from the optical rotation Angular rotation (α) at 20˚C for matooke flour was approximated to 184.0 for mixed flour (K 2000(𝑃−𝑃′ ) Sawyer, 1991). according mass as = follows (Manufacture’s manual): 3.1 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒to𝑜𝑓the 𝑆𝑡𝑎𝑟𝑐ℎ (𝑑𝑏) 𝛼 Sawyer, 1991). 2000(𝑃−𝑃′ ) 3.1 flour (Kirk & 𝑃𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑜𝑓 𝑆𝑡𝑎𝑟𝑐ℎ Angular rotation (α) at (𝑑𝑏) 20˚C=for matooke flour was approximated to 184.0 for mixed 𝛼 Sawyer, 1991). readings for determining the optical rotation were carried out at room The polarimeter Angular rotation (α) at 20˚C for matooke flour was approximated to 184.0 for mixed flour (Kirk & o Angular rotation (α)the at recommended 20˚C for was to 184.0 for therefore, the specific angular rotation read on the at temperature above 20 C,flour The polarimeter readings formatooke determining the approximated optical rotation were carried out Sawyer, 1991). o mixed flour (Kirk & Sawyer, 1991). o C using rotation the equation polarimeter was converted to the equivalent20specific angular rotation at 20angular C, therefore, the specific read o temperature above the recommended oout at room belowpolarimeter (Manufacture’s manual): The readings for determining thespecific opticalrotation rotationwere werecarried carried using the equ polarimeter wasreadings converted todetermining the equivalent angular rotation at 20 The polarimeter for the optical outC o othe specific angular rotation read on the temperature above the above recommended 20 C, therefore, C, therefore, the specific at room temperature the recommended 20 below (Manufacture’s manual): The polarimeter readings for determining the optical rotation were carried out at room o C using the equation polarimeter was converted the polarimeter equivalent specific angular rotation at 20 angular rotation read ontothe was converted to the equivalent o C, therefore, the specific angular rotation read on the temperature above the recommended 20 �����ar r��a���� a� �e�pera��re � below (Manufacture’s manual): using the equation below (Manufacture’s specific angular at 20oC 3.2 𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 rotation 𝑎𝑡 20˚C = {1+0.000144×(�e�pera��re o �−20˚C)} polarimeter was converted to the equivalent specific angular rotation at 20 C using the equation manual): below (Manufacture’s manual): �����ar r��a���� a� �e�pera��re � 3.2 𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20˚C = {1+0.000144×(�e�pera��re �−20˚C)} �����ar r��a���� a� �e�pera��re � 𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20˚C =2.�×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 𝑇𝑜𝑡𝑎𝑙 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑖𝑜𝑛 𝑃 = {1+0.000144×(�e�pera��re �−20˚C)} 𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20˚C = {1+0.000144×(�e�pera��re 𝑇𝑜𝑡𝑎𝑙 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑖𝑜𝑛 𝑃 = 𝑇𝑜𝑡𝑎𝑙 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑖𝑜𝑛 𝑃 = 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑃′ = 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑃′ = Therefore, Therefore, 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑃′ = Therefore, 𝑇𝑜𝑡𝑎𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 2.�×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 3.3 3.4 3.4 𝑓𝑜𝑟 40% �×𝐴𝑛𝑔𝑢𝑙𝑎𝑟𝑆𝑎𝑚𝑝𝑙𝑒 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛𝑚𝑎𝑠𝑠 𝑎𝑡 20𝐶 𝑓𝑜𝑟 40%𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑎𝑚𝑝𝑙𝑒 3.4 �×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 3.4 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 184.0 3.3 3.3 𝑇𝑜𝑡𝑎𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 �×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 �2000�𝑃−𝑃′�� Therefore, x 100 𝑆𝑡𝑎𝑟𝑐ℎ 𝐶𝑜𝑛𝑡𝑒𝑛𝑡 (% 𝑑𝑏) = Therefore, 3.2 �−20˚C)} 2.�×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 2.�×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 𝑅𝑜𝑡𝑎𝑖𝑜𝑛 �×𝐴𝑛𝑔𝑢𝑙𝑎𝑟 𝑃= 𝑟𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑎𝑡 20𝐶 𝑓𝑜𝑟 40% 𝐸𝑡ℎ𝑎𝑛𝑜𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑇𝑜𝑡𝑎𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 𝑇𝑜𝑡𝑎𝑙 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑂𝑝𝑡𝑖𝑐𝑎𝑙 𝑅𝑜𝑡𝑎𝑡𝑖𝑜𝑛 𝑃′ = 3.2 3.3 𝑇𝑜𝑡𝑎𝑙 𝑆𝑎𝑚𝑝𝑙𝑒 𝑚𝑎𝑠𝑠 �����ar r��a���� a� �e�pera��re � 3.5 The data was analyzed using MINTAB Software (version 14 for Windows ′ �� ′ �� �2000�𝑃−𝑃 �2000�𝑃−𝑃 (2003) Minitab Inc., Pennsylvania, The Pearson Correlation was used (% (% xSoftware 100 3.5 Inc., 𝑆𝑡𝑎𝑟𝑐ℎ 𝐶𝑜𝑛𝑡𝑒𝑛𝑡 𝑑𝑏) = 184.0 xUSA). 100 3.5 Minitab 𝑆𝑡𝑎𝑟𝑐ℎ 𝐶𝑜𝑛𝑡𝑒𝑛𝑡 𝑑𝑏) =using The data was analyzed MINTAB (version 14 for Windows (2003) 184.0 to determine the significant correlation between harvest maturity and maturity Pennsylvania, USA). The �2000�𝑃−𝑃 Pearson′��Correlation was used to determine the significant correlation x 100 is greater than 0.5 and a p-value 3.5 𝑆𝑡𝑎𝑟𝑐ℎ 𝐶𝑜𝑛𝑡𝑒𝑛𝑡 = indices. When (% the𝑑𝑏) Pearson correlation 184.0 between harvest maturity and maturity indices. When the Pearson correlation is greater than 0.5 (p<0.05), it indicated that the relationship was statistically significant at α-level and p-value indicated that Software the relationship was significant at α-level of The adata was (p<0.05), analyzed itusing MINTAB (version 14statistically for Windows (2003) Minitab data was analyzed ofusing MINTAB Software (version 14 for Windows (2003) Inc., Minitab ofThe 0.05 (95% confidence the relationship). 0.05 (95% confidence of the relationship). Pennsylvania, USA). The Pearson Correlation was used to determine the significant correlation Pennsylvania, USA). The Pearson was used to determine the significant corre The data was analyzed using MINTABCorrelation Software (version 14 for Windows (2003) Minitab Inc., between harvest maturity and maturity indices. When the Pearson correlation is greater than 0.5 between harvest maturity and maturity indices. Pearson greater tha Pennsylvania, USA). The Pearson Correlation was When used tothe determine thecorrelation significant is correlation and a p-value (p<0.05), it indicated that the relationship was statistically significant at α-level of between harvest(p<0.05), maturity and maturity indices. the Pearson is greater than at 0.5α-le and a p-value it indicated that theWhen relationship wascorrelation statistically significant 0.05 (95% confidence of the relationship). Establishment of optimum harvest maturity window for matooke 24 3.3 Results and Discussion 3.3.1 Physiological maturity The physiological maturity was determined using the meteorological data for Bushenyi located at the Latitude: 0.59˚S, Longitude: 30.21˚E and Altitude: 1570 m. It was obtained by substituting meteorological data at Bushenyi during the period of data collection from January 2011 to January 2013 in equations 2.1 and employing the base Temperature for banana = 14˚C. Average daily maximum temperature = 25.44 ± 0.35˚C Average daily minimum temperature = 14.81 ± 0.25˚C Substitute the above temperatures in equation 2.1 the physiological maturity was 147 days (21 weeks). This implies that physiological maturity for matooke cv. Mbwazirume (Musa sp triploid acuminate genome group AAA-EAHB) at Bushenyi is 21 weeks. 3.3.2 Effect of harvest maturity on maturity indices The effect of harvest maturity of matooke on maturity indices was studied by taking harvest maturity (weeks) as independent variable and the maturity indices as dependent variables. The maturity indices considered in this study include the following: • Finger weight (g) • Pulp weight (g) • Peel weight (g) • Pulp/Peel ratio • Finger length (mm) • Finger circumference/girth (mm) The Pearson correlation was used to determine the correlation of different indices to harvest maturity and amongst themselves as shown in Table 3.2. It was observed from Table 3.2, that all maturity indices were statistically significant. This was in agreement with Dadzie (1998), Muranga (1998) and Robinson & Saúco (2010). From the Table 3.2, the maturity indices with the highest Pearson correlation above 0.90 were selected as the best maturity Establishment of optimum harvest maturity window for matooke 25 indices for estimating the harvest maturity and these include: Finger, peel and pulp weight. However, comparing the level of significance, finger length had the lowest which indicated that the finger had almost the maximum length before week 10; implying that there was slow development in finger length after week 10. This was in agreement with Robinson & Saúco (2010), who reported that the finger length increases rapidly for 30 days after which growth in length slows down and is completed in 40 – 80 days after emergence depending on the area and climate. Table 3.2: Pearson correlation for maturity indices Harvest Maturity (weeks) Finger Weight (g) Finger Weight (g) Finger Length (mm) Finger Circum1 (mm) Peel Weight (g) Pulp Weight (g) Pulp/ Peel Ratio 0.939* 0.000** Finger Length (mm) 0.276 0.199 0.055 0.170 Finger Circum (mm) 0.881 0.916 -0.006 0.000 0.000 0.965 Peel Weight (g) 0.808 0.896 0.060 0.834 0.000 0.000 0.649 0.000 Pulp Weight (g) 0.958 0.978 0.121 0.919 0.897 0.000 0.000 0.143 0.000 0.000 Pulp/ Peel Ratio 0.953 0.900 0.332 0.847 0.691 0.933 0.000 0.000 0.021 0.000 0.000 0.000 Moisture Content (%wb) 0.794 0.727 0.186 0.694 0.599 0.774 0.766 0.000 0.000 0.201 0.000 0.000 0.000 0.000 Key: Cell Contents - * Pearson correlation ** p - value (p = 0.000 => p < 0.0001) 1 Circumference Establishment of optimum harvest maturity window for matooke 26 Figure 3.2 shows the effect of harvest maturity on finger, pulp and peel weights. Figure 3.2 300 Finger Peel Pulp Weight (g) 200 100 0 10 15 20 25 Harvest maturity (weeks) Figure 3.2: Effect of harvest maturity on finger, pulp and peel weights Weight (g) It was observed from Figure 3.2 that both finger and pulp weight increased with Figure 3.5 increase in time until week 22 then decreased. This indicated that matooke 300 at week 22 had reached Finger its maximum harvest maturity. The peel weight also increased with increasing time until week 21 and then decreased to week Peel 22. This reduction in peel weights after week 21 could be attributed to cell Pulp 200 wall degradation which starts underneath the peel (Robinson & Saúco, 2010). After week 22, matooke, being a climacteric fruit, it enters the post climacteric phase at which point dramatic physiological changes occur; these include the peel color and 100 pulp texture among others. In advanced stages of maturity, slight peel color change occurs to the initial stages of ripening. The reduction in finger and pulp weight is attributed to loss of starch, which plays a major 0 role in textural changes in ripening. 10 15 20 Harvest maturity (weeks) 25 Establishment of optimum harvest maturity window for matooke 27 Figure 3.3, shows that the pulp/peel ratio increased with time from week 10 to weekEstablishment 22 and then decreased. of optimum harvest maturity window for matooke 29 2.00 Pulp/Peel Ratio 1.50 y = 0.0707x + 0.2437 R² = 0.9829 1.00 0.50 0.00 10 15 20 25 Harvest Maturity (weeks) Figure 3.3: Effect of harvest maturity on pulp/peel ratio. 3.3: Effect on pulp/peel ratio. FromFigure Figure 3.3of itharvest wasmaturity further observed that matooke attains maximum maturity the3.3harvest atobserved 21 weeks and attains if it ismaximum not harvested starts From at Figure it was further that matooke maturity at the itharvest at going 21 weeks if it is not harvested itstage starts going through the stage (Muchui et al., 2012). through the and post-climacteric (Muchui et post-climacteric al., 2010, Tapre & Jain, Tapre & Jain, 2012). These results are within comparable range to those reported by These2010, results are within comparable range to those reported by Robinson & Robinson & Saúco, (2010), that the pulp/peel ratio of 1.0 was achieved at about 70 days, in this Saúco, (2010), that the pulp/peel ratio of 1.0 was achieved at about 70 days, study was about 77 days (11 weeks). in this study was about 77 days (11 weeks). Finger Circumference (mm) Figure 3.43.4shows thatthethe circumference of increased the fingers increased time Figure shows that circumference of the fingers with time and this waswith in agreement with Muchui et al., (2010) and Robinson &al., Saúco, (2010), and when Robinson they reported that and this was in agreement with Muchui et (2010) & Saúco, Establishment of optimum harvest maturity window for matooke 30 diameter of the fingers increases with bunch age. (2010), when they reported that diameter of the fingers increases with bunch 18 age. 12 6 10 15 20 Harvest Maturity (weeks) 25 Figure 3.4: harvest maturity on finger circumference. Figure Effect 3.4: Effectof of harvest maturity on finger circumference. The Pearson Correlation was used to determine the correlation of different CIE L*a*b* color Establishment of optimum harvest maturity window for matooke 28 The Pearson Correlation was used to determine the correlation of different CIE L*a*b* color space to harvest maturity and amongst themselves. It was observed from Table 3.3, that all the different color space were statistically significant, except the lightness (L*). It was further observed that L* and a* showed inverse relationship with harvest maturity, whereas the saturation (C*), b* and hue angle in degrees (α) values showed a direct relationship with harvest maturity. Table 3.3: Pearson Correlation for Color Space Harvest maturity (weeks) L* a* b* L* -0.134* 0.287** a* -0.617 0.000 0.442 0.000 b* 0.549 0.000 -0.700 0.000 -0.946 0.000 C* 0.556 0.000 -0.685 0.000 -0.953 0.000 1.000 0.000 α 0.540 0.000 0.491 0.000 -0.395 0.001 0.144 0.252 Key: Cell Contents - C* 0.161 0.201 * Pearson correlation ** p - value (p = 0.00 => p < 0.0001) The CIE-Lab parameters were used to express the color of fresh pulp; Lightness (L*), Saturation (C*) and Hue angle in degrees (α) values are calculated from equations 2.2 and 2.3. It was observed from Figure 3.5 and Table 3.3 that lightness (L*) was not statistically significant. The hue angle (α > 90˚) showed that the color of the pulp lies between green and yellow space with low saturation (C*). Both hue angle (α˚) and saturation (C*) increased with harvest maturity. It was noted that hue angle (α) was relatively constant between weeks 17-20 which indicated that the color of the pulp was relatively uniform within that harvest maturity range. Because of nearly constant lightness and low saturation, one cannot easily differentiate between the differences of colors at different harvest maturity. This indicates that color of the pulp is not a good index for estimating the harvest maturity. Harvest maturity (weeks) Establishment of optimum harvest maturity window for matooke 29 Figure 3.5 300 Finger Peel Pulp Weight (g) 200 100 0 10 15 20 25 Harvest maturity (weeks) Figure 3.5 Effect of harvest maturity on given L*, C* & α 3.3.3 Effect of harvest maturity on the physicochemical properties of matooke flour The effect of harvest maturity in weeks of matooke on physicochemical properties of the flour was studied by taking harvest maturity (weeks) as independent variable and the physicochemical properties of flour as dependent variables (starch content and pasting properties). In order to establish the optimum harvest maturity to harvest matooke it was imperative to determine the starch development from third week after the inflorescence emerges from the stem. Figure 3.6, shows the effect of harvest maturity on starch content of matooke flour. The starch data was collected from week 3 then at intervals of three weeks to week 9, then collected every week up to week 22. Starch Content (%DB) 75 Fitted curve Establishment of optimum harvest maturity window for matooke 30 Establishment of optimum harvest maturity window for matooke 50 y= 100 33 -0.34x2 + 10.62x + 1.72 R² = 0.99 Experiment 25 0 0 Starch Content (%DB) 75 Fitted curve 50 5 y = -0.34x2 + 10.62x + 1.72 15 R² = 0.99 20 10 25 Maturity (weeks) 25 Figure 3.6: Effect of harvest maturity on starch content (%db) of matooke flour. 0 0 5 10 15 20 25 Maturity (weeks) Figure 3.6: Effect of harvest maturity on starch content (%db) of matooke flour. Figurethat 3.6: Effect harvest content maturity on increased starch content with (%db) time of matooke Figure 3.6, shows the ofstarch fromflour. week 3 - 12 and it became relatively constant until week 20, when it started to decrease with age. This was an indication that the Figure starch accumulates in the ascontent the fruit continues tofrom grow begins from 3.6,Figure shows that the starch increased with time from week 3 - 12 3.6, shows that pulp the starch content increased with time week until 3 - 12 maturation and it became relatively constant untilmatooke week 20, when it started to decrease with age. This indication that and became relatively constant until week 20, when it 12 started decrease week 12.it This indicated that can be harvested from –was21anto weeks because starch the starch accumulates in the pulp as the fruit continues to grow until maturation begins from withofage. This was indication content matooke flouran is above 80%that (db).the starch accumulates in the pulp as the week 12. This indicated that matooke can be harvested from 12 – 21 weeks because starch fruit continues grow flour untilis above maturation content to of matooke 80% (db). begins from week 12. This indicated that matooke can be harvested from 12 – 21 weeks because starch content of matooke flour is above 80% (db). The quadratic model fitting thefitting starch datadata could auseful useful for estimating the starch content The quadratic model the starch couldbe be a tooltool for estimating the starch content of matooke flourfitting at a given maturity, which to starch as given in The quadratic theharvest starch datawhich couldwas bereferred areferred useful tool for estimating of matooke flour atmodel a given harvest maturity, was to model starch model as given in equation 3.6. the starch equation 3.6. content of matooke flour at a given harvest maturity, which was referred to starch model as given in equation 3.6. 𝒀 = 𝟏. 𝟕𝟐 + 𝟏𝟎. 𝟔𝟐𝒙 − 𝟎. 𝟑𝟒𝒙𝟐 𝒀 = 𝟏. 𝟕𝟐 + 𝟏𝟎. 𝟔𝟐𝒙 − 𝟎. 𝟑𝟒𝒙𝟐 2 R = 0.99 2 R = 0.99 3.6 3.6 Where: Where: Where: Y = Starch content (%db) Y = Starch content (%db) x = Harvest maturity (weeks) Y = Starch content (%db) Table 3.4 showed that pasting properties were statistically significant and also the interactions of all pasting properties have statistically significant relationship. Table 3.5 shows pasting properties for the matooke. The results obtained for the pasting properties were in agreement with previous work on Establishment of optimum harvest maturity window for matooke 31 matooke (Table 2.2). Figure 3.7, shows the pasting curves were grouped with respect to peak viscosities being relatively within the same range and selected harvest maturity 10 &12, 16 & 20 and 22 weeks which are representatives for range of harvest maturity of 10-15, 15-21 and 22 respectively. It was further observed from Figure 3.7, that there are three peaks as per the above groupings week 10-15, 15-21 and 22 from the highest to lowest respectively. The lower peak suggests that at week 22, the starch granules were transformed by the initial stages of ripening. This was also observed for effect of harvest maturity on pulp/peel ratio and other maturity indices. Table 3.4: Pearson Correlation for Pasting Properties Peak Viscosity (RVU) Harvest maturity (weeks) -0.395* Peak Viscosity (RVU) Trough (RVU) Break- Final down Viscosity (RVU) (RVU) Setback (RVU) Peak Time (min) 0.001** -0.274 0.634 0.039 -0.073 0.000 0.247 -0.583 0.568 -0.436 0.049 0.665 0.000 0.967 -0.519 0.001 Setback (RVU) -0.701 0.000 0.242 0.000 0.117 0.000 0.108 0.366 0.000 -0.482 0.054 0.165 0.356 0.675 0.396 -0.679 0.003 0.711 0.308 0.002 -0.321 0.192 -0.106 0.000 0.380 0.000 -0.594 0.000 0.425 0.013 0.268 0.852 0.010 0.402 0.002 0.000 0.000 0.032 0.000 Trough (RVU) Breakdown (RVU) Final Viscosity (RVU) Peak Time (min) Pasting Temperature (˚C) Key: Cell Contents - * Pearson correlation ** p - value (p = 0.00 => p < 0.0001) 32 Table 3.5: Pasting property of Matooke flour (cv Mbwazirume) Peak Viscosity (RVU) Trough Viscosity (RVU) Breakdown Viscosity (RVU) Final Viscosity (RVU) Setback Viscosity (RVU) Peak Time (min) Pasting Temperature (oC) 10 427.83±23.02 213.33±16.03 214.50±14.73 277.53±5.99 64.19±10.44 4.84±0.08 78.63±0.42 11 407.56±14.74 223.20±16.55 175.36±24.17 281.82±20.56 49.63±7.87 4.92±0.13 79.94±1.52 12 442.43±11.19 296.93±31.78 145.50±34.25 347.73±28.59 50.80±4.21 5.04±0.20 78.69±4.49 13 448.07±32.63 299.14±28.21 148.99±7.91 346.36±21.80 47.22±9.29 5.17±0.04 82.12±1.46 14 445.49±7.61 290.61±20.73 154.88±25.37 342.96±20.47 52.35±4.57 4.92±0.06 79.98±0.34 15 401.87±33.52 279.28±19.24 122.58±16.64 327.25±11.11 47.97±8.53 5.15±0.04 83.88±1.03 16 408.90±53.53 257.28±48.10 151.62±11.40 297.20±47.90 39.92±0.86 4.96±0.08 79.31±2.73 17 414.26±16.92 243.83±3.73 170.43±15.52 279.14±4.13 35.31±3.91 4.68±0.02 74.76±0.59 18 399.07±11.22 244.40±39.53 154.67±35.54 286.79±41.15 42.39±2.63 4.84±0.21 78.98±2.02 19 404.34±59.87 228.13±19.98 176.21±79.85 269.92±10.02 41.79±9.96 4.80±0.28 79.60±3.96 20 400.00±6.55 235.67±7.60 164.33±4.45 274.58±7.69 38.92±1.31 4.73±0.07 76.48±1.46 21 412.89±6.69 240.50±10.90 172.39±12.11 278.68±8.63 38.18±4.35 4.72±0.10 77.26±0.75 22 366.44±13.74 214.22±4.57 152.22±9.43 238.14±2.30 23.92±2.39 4.80±0.00 79.37±0.41 Establishment of optimum harvest maturity window for matooke Harvest maturity (weeks) Establishment of optimum harvest maturity window for matooke 10 33 12 16 20 22 Figure 3.7: Pasting curves for Matooke. (.....) Peak viscosity in three groups represented by harvest maturity 10 &12, 16 & 20 and 22 weeks which are representatives for range of harvest maturity of 10-15, 15-21 and 22 respectively. Figure 3.8 shows the effect of harvest maturity on peak viscosity and final viscosity. Figure 3.8 500 Viscosity (RVU) 400 300 200 Peak Viscosity (RVU) Final Viscosity (RVU) 100 10 15 20 25 Maturity (weeks) Figure 3.8: Effect of harvest maturity on peak viscosity and final viscosity. 34 Establishment of optimum harvest maturity window for matooke It was observed from Figure 3.8 that there were two levels of peak viscosity between weeks 12-14 and weeks 15- 21. The first peak between was weeks 12 – 15. This indicated a third stage of banana growth referred to as maturation (Stover & Simmonds, 1987, Robinson & Sauco, 2010). This implied within this period starch is being deposited and it had not yet stabilized, in other words, it was still in transition. Robinson & Sauco, (2010) reported that starch grains are initially deposited in the pulp cells which form in the vicinity of the vascular bundles, and thereafter starch deposition moves centripetally and continues until harvest maturity. This peak between 12 – 15 weeks could be associated with amylose and amylopectin content were not yet fully developed or stabilized which is responsible of the functional characteristics of flour. Therefore, the high peak viscosity could be attributed to fiber and components in the flour causing the resistance to the shearing leading to high viscosity. The second peak viscosity from 15 – 21 weeks was rather uniform, which was an indication that period was beyond the maturation stage, and starch deposition had stabilized. This implied that both amylose and amylopectin molecules had reached the equilibrium point between swelling and polymer leaching, at nearly the same viscosity, which gave a relatively equal peak viscosity and final viscosity. 3.3.4 Prediction Models The following models were developed from the effect of harvest maturity on the maturity indices by employing the regression equation using the MITAB Software. Considering the fact that week 22 had a unique response for all maturity indices, it was omitted from the regression analysis for predicting the harvest maturity for matooke to be harvested for processing flour. Table 3.6 shows the models which could be used to predict harvest maturity (weeks) given the finger weight (g), or pulp/peel ratio, or a combination of finger weight and pulp/peel ratio. However, it was observed that, the model of a combination was the best followed by the pulp/peel model. Establishment of optimum harvest maturity window for matooke 35 Table 3.6 Models for predicting harvest maturity Model Parameters Model R2 r P< Harvest maturity (weeks) & Finger weight (g) Y = 3.29 + 0.07x1 0.87 0.939 0.01 Harvest maturity (weeks) & Pulp/Peel ratio Y = 11.9x2 – 0.298 0.91 0.958 0.01 Harvest maturity (weeks) & Finger weight (g) + Pulp/ Peel ratio (Combined) Y = 0.387 + 0.031x1 + 7.27x2 0.94 Where: Y = Harvest maturity (weeks) x1 = Finger weight (g) x2 = Pulp/Peel ratio 0.01 The model equation given in Table 3.7 can be used to predict the amount of starch at a given harvest maturity (weeks) of matooke after inflorescence emergence. Table 3.7 Models for predicting starch content Model Parameters Model R2 p< Starch content (%db) & Harvest maturity (weeks) Y = 1.72 + 10.62x – 0.34x2 0.99 0.016 Where: Y = Starch content in flour (%db) x = Harvest maturity of matooke (weeks) 3.3.5 Validation of prediction models The above models given in Table 3.6 were validated by using a new set of experiments to determine the actual values for selected maturity indices (finger weight, pulp/peel ratio) at different harvest maturity from week 10 - 22. The analysis of variance (ANOVA) was employed to determine the significance of the relationship between the actual experimental data and the model predicted values (Table 3.8). It was observed that the selected models were significant (p<0.01), it indicated that the relationship was statistically significant at α-level of 0.01(99% confident of the estimated harvest maturity). The combined Finger weight with Pulp/peel ratio was the best followed by the Pulp/peel ratio model. Therefore, the models in Table 3.6 can be adapted for estimating the harvest maturity of matooke at Bushenyi. Establishment of optimum harvest maturity window for matooke 36 Table 3.8 Significance of prediction model for estimating harvest maturity Model R2 p< Finger weight model 0.87 0.01 Pulp/peel ratio model 0.89 0.01 Combined finger weight and pulp/peel ratio model 0.92 0.01 3.4 Conclusions and Recommendations: The physiological maturity of matooke (Mbwazirume cv. Musa sp. AAA EAHB) at Bushenyi located at the Latitude: 0.59˚S, Longitude: 30.21˚E and Altitude 1570m is 21 weeks. Considering the results of starch content and pasting properties, the uniform physicochemical properties were obtained between weeks 15 – 21. This indicates that the optimum harvest maturity window for commercial processing of matooke at Bushenyi is between 15-21 weeks for processing standard raw matooke flour or Raw Tooke Flour (RTF). However, in case of natural disaster like storm, matooke from the harvest maturity from week 12-15 weeks can be processed into matooke flour since at that harvest maturity it already has starch content above 80% (db). In addition, matooke harvested beyond the above maturity window after 21 weeks is recommended for processing Instant Tooke Flour (ITF). This is because beyond the recommended harvest maturity window, it would be in the initial stages of starch loss which plays a major role in textural changes in ripening which would not affect the functional properties of instant flour significantly. Theoretically the results of the study can be used elsewhere. Nevertheless, it is recommended that similar studies be carried out in other the matooke growing areas to determine the applicable models for those particular regions since development of matooke and bananas in general is subject to location (Robinson & Saúco, 2010) Recommended Characteristics within the Maturity Window The finger weight model is recommended for farmers to estimate harvest maturity for matooke and the combined model of finger weight and pulp peel Establishment of optimum harvest maturity window for matooke 37 ratio is recommended for commercial processors. Matooke cv mbwazirume (Musa sp. AAA-EAHB) at Bushenyi, harvested within the optimum harvest maturity window from week 15 – 21, should have the following characteristics given in Table 3.10 and expected starch content ≥ 80 % (db). Table 3.10 Characteristics of matooke cv mbwazirume (Musa sp. AAAEAHB) Characteristics Range Harvest maturity (weeks) 15 – 21 Finger weight (g) 190-234 Pulp/peel ratio 1.4 -1.8 Establishment of moisture sorption isotherms for matooke 38 Chapter Four Objective: Establishment of moisture sorption isotherms for matooke 4.1 Background The commercial processing of matooke requires having microbiologically shelf-stable products on the market. Moisture content is an important criterion for judging food quality and water activity (aw), which is essential for describing the water availability and mobility in food (Ayala et al., 2011). Previous research on moisture isotherms focused on other types of bananas and plantains, but as reported by Robinson and Saúco, (2010), the East African Highland bananas are a distinct ‘AAA group’. No prior moisture sorption isotherms have been reported on the East African Highland bananas (matooke, Musa sp. AAAEAHB). In order to achieve the objective of having a microbiologically shelfstable product it was important to establish its moisture sorption isotherm. The main objective of this study was therefore to establish the moisture sorption isotherms for matooke. The specific objectives were to determine: • The effect of temperature on the moisture sorption isotherm • The existing models which best predict the moisture sorption isotherms • The hysteresis behavior • The sorption heat. 4.2 Materials and methods The static gravitation method recommended by COST 90 Project (Wolf et al., 1985), was used for establishing both adsorption and desorption isotherms. 3 g of samples were put in air-sealed glass jars maintained at equilibrium relative humidity using saturated salt solution. The glass jars were placed in an incubator (BINDER CB Series, ILMACK, Basel, Switzerland) to maintain the required set temperature (Sablani et al., 2007). • Dried samples used to determine the adsorption isotherm at 30 and 60˚C in were maintained at relative humidity ranging from 10 to 90% by the static method employing standard saturated salt solutions. The dry Establishment of moisture sorption isotherms for matooke 39 samples were kept in desiccators with silica gel at room temperature for 2 days prior to the beginning of experiments (Pahlevanzadeh & Yazdani, 2004; Yazdani et al., 2006). • Fresh samples used to determine the desorption isotherm at 50 and 60˚C were maintained at relative humidity ranging from 10 to 90% by the static method employing standard saturated salt solutions. • The final equilibrium moisture content of samples was determined by the oven method at 105˚C for 24 h and results expressed on dry weight basis (Baini & Langrish, 2007). The temperatures selected for desorption isotherm determination are normally used for drying of bananas (Queiroz & Nebra, 2001; Karim & Hawlader, 2005; Nguyen & Price, 2007). The adsorption isotherm temperature of 30˚C was selected because it corresponds with the maximum ambient conditions in Uganda in terms of considering the storage conditions. The adsorption and desorption data at 60˚C were selected for establishing the hysteresis phenomenon. The salt solutions used to obtain constant relative humidity of the surrounding air from 10 to 90% are given in the Table 4.1; all the salts used were of analytical grade. Table 4.1 Relative humidity over saturated salts at 30, 50 and 60˚C Salt Solution Relative Humidity (%) 30˚C 50˚C 60˚C LiCl 11.28 11.21 10.95 CH3COOK 21.61 19.20 18.00 MgCl2 32.44 30.54 29.26 K2CO3 43.17 42.65 42.11 Mg(NO3)2 51.40 46.16 45.44 NaNO2 73.14 69.04 67.27 NaCl 75.09 74.34 74.20 KCl 83.62 81.20 80.25 Statistical Analysis The experiments were carried out in triplicates and the results were analyzed by using SPSS (version 16) for Windows (© 2010 University of Bristol, UK). It minimizes the sum of squares of derivatives between experimental and theory in a series of iterative steps. It also evaluates the parameters/ constants of Establishment of moisture sorption isotherms for matooke 40 the model and goodness-of-fit (R2). The suitability of the models was further evaluated and compared using the following statistical errors: • The root mean square error (RMSE, equation 4.1), gives the fitting ability of a model in relation to number of data points, the smaller the value the Establishment of moisture sorption isotherms for matooke better the fitting ability of an equation (Hassian et al., 2001; Mehta & Singh, 2006; Siripatrawan & Jantawat, 2006;for Moreira et al., 2008; Seid & Establishment of moisture sorption isotherms matooke 2 Hensel, 2012). ∑𝑛 �𝑥 −𝑥 � 𝑝𝑟𝑒𝑑 𝑖=1 𝑖 𝑅𝑀𝑆𝐸 = � 𝑅𝑀𝑆𝐸 = � • 4.1 𝑛 2 ∑𝑛 𝑖=1�𝑥𝑖 −𝑥𝑝𝑟𝑒𝑑 � 4.1 𝑛 Mean relative percentage deviation modulus (E: equation 4.2). This has Mean relative percentage deviation modulus (E: equation 4.2). This has been widely use been widely used and it has been recommended that the E value less been the E less than 10% is (Mehta indicative a good fit for practic than 10% recommended is indicative ofthat a good fit value for practical & of Singh, Mean relative percentage deviation modulus purposes (E: equation 4.2). This has been widely use (Mehta & Singh, 2006; Siripatrawan Jantawat, Yan et al., 2008; Moreira et al., 20 2006; Siripatrawan & Jantawat, 2006;& Yan et al.,2006; 2008; Moreira et al., been recommended that the E value less than 10% is indicative of a good fit for practic 2008). (Mehta & Singh, 2006; Siripatrawan & Jantawat, 2006; Yan et al., 2008; Moreira et al., 20 𝐸 = 100 𝑛 ∑𝑛𝑖=1 �𝑥𝑖 − 𝑥𝑝𝑟𝑒𝑑 � 4.2 𝑥𝑖 Establishment of moisture sorption isotherms for matooke 46 �𝑥 − 𝑥 � ∑𝑛𝑖=1 𝑖 𝑝𝑟𝑒𝑑 𝑥𝑖 𝑛 4.3 Results and Discussion 𝐸 = 100 4.2 4.3 Results and Discussion 4.3.1 Effect of temperature on moisture isotherm for matooke 4.3.1 Effect of temperature on moisture isotherm for matooke Figures 4.1 and 4.2 show the adsorption and desorption isotherms Figures 4.1 and 4.2 show the adsorption and desorption isotherms respectively. respectively. 0.3 Equilibrium moisture content (kg water/kg drymatter) 30˚C 60˚C 0.2 0.1 0.0 0 0.2 0.4 0.6 Water activity, aw Figure 4.1 Experimental data adsorption isotherm Figure 4.1 Experimental data adsorption isotherm 0.8 1 Establishment of moisture sorption isotherms for matooke Establishment of moisture sorption isotherms for matooke 47 41 0.3 Equilibrium moisture content (kg water/kg dry matter) 50˚C 60˚C 0.2 0.1 0.0 0 0.2 0.4 0.6 Water activity, aw 0.8 1 Figure 4.2Experimental Experimental data isotherm desorption isotherm Figure 4.2 data desorption It was observed from Figures 4.1 and 4.2, that the moisture sorption isotherms was observed Figures 4.1 and 4.2, that the moisture sorption isotherms in bothmoisture cases were at a in Itboth casesfrom were temperature dependant. The equilibrium temperature dependant. The equilibrium moisture at a given water activity, decreased with increase in given water activity, decreased with increase in temperature. This was in temperature. This was in agreement with the theory of physical sorption (Iglesias et al., 1975; Hassian et agreement withthat thematooke theory of physical sorption (Iglesias et al.,This 1975; Hassian al., 2001). Implying became less hygroscopic with increasing temperature. was similar the 2001). report on starch (Al-Muhtaseb et al., 2004), became pitahya fruit (Ayala al., 2011) and coffee et et toal., Implying that matooke less et hygroscopic with(Corrêa increasing al., 2010). Both adsorption and desorption isotherms for matooke exhibited Type II curves which is more temperature. This was similar to the report on starch (Al-Muhtaseb et al., pronounced for desorption moisture isotherm. It was consistent with previous researchers on bananas 2004), pitahya fruit (Ayala et al., 2011) and coffee (Corrêa et al., 2010). Both and plantains (Johnson & Brennan, 2000; and Yan et al., 2008). adsorption and desorption isotherms for matooke exhibited Type II curves which is more pronounced for desorption moisture isotherm. It was consistent 4.3.2 Fitting Models to Moisture Sorption Isotherms for Matooke with previous researchers on bananas and plantains (Johnson & Brennan, The fitting the moisture isotherms was done using the nonlinear regression analysis. Figures 2000; andof Yan et al.,sorption 2008). 4.3, 4.4, 4.5 and 4.6 show the best fitted models given in Table 4.2, for both adsorption and desorption moisture isotherms for matooke. 4.3.2 Fitting Models to Moisture Sorption Isotherms for Matooke The fitting of the moisture sorption isotherms was done using the nonlinear regression analysis. Figures 4.3, 4.4, 4.5 and 4.6 show the best fitted models given in Table 4.2, for both adsorption and desorption moisture isotherms for matooke. Establishment of moisture sorption isotherms for matooke 42 Establishment of moisture sorption isotherms for matooke 48 0.3 Xexp Equilibrium Equilibrium Moisture ContentMoisture Content (kg water/kg drymatter) (kg water/kg drymatter) Establishment of moisture sorption isotherms for matooke GAB 48 Chung-Pfost 0.3 0.2 Oswin Xexp Smith GAB Chung-Pfost 0.2 0.1 Oswin Smith 0.1 0 0 0.2 0.4 0.6 Water Activity, aw 0.8 1 Figure 4.3 Best fitted models for adsorption isotherms of matooke at 30˚C Figure 4.3 Best fitted models for adsorption isotherms of matooke at 30˚C 0 0.3 0 0.2 Xexp 0.4 0.6 Water Activity, aw 0.8 1 Equilibrium Equilibrium Moisture ContentMoisture Content (kg water/kg drymatter) (kg water/kg drymatter) BET Figure 4.3 Best fitted models for adsorption isotherms of matooke at 30˚C GAB 0.2 0.3 Xexp Chung-Pfost BET Smith GAB 0.1 0.2 0.0 0.1 Chung-Pfost Smith 0 0.2 0.4 0.6 0.8 1 Water Activity, aw 0.0 Figure 4.4 Best fitted models0 for adsorption of matooke at 0.2 isotherms0.4 0.660˚C 0.8 1 Water Activity, aw Figure 4.4 Best fitted models for adsorption isotherms of matooke at 60˚C Figure 4.4 Best fitted models for adsorption isotherms of matooke at 60˚C Establishment of moisture sorption isotherms for matooke 49 Establishment of moisture sorption isotherms for matooke 43 Establishment of moisture sorption isotherms for matooke 49 0.3 Xexp Equilibrium Moisture Content Equilibrium(kg Moisture Content water/kg dry matter) (kg water/kg dry matter) 0.3 0.2 0.2 GAB Xexp Henderson GAB Smith Henderson Smith 0.1 0.1 0 0 0.2 0 0.4 0.6 Water Activity, aw 0.8 0 0.2 0.4 0.6 Figure 4.5 Best fitted models for desorption isotherms of matooke at 50˚C Water Activity, a 1 0.8 1 w Figure 4.5 Best fitted models for desorption isotherms of matooke at 50˚C Figure 4.5 Best fitted models for desorption isotherms of matooke at 50˚C 0.3 Equilibrium Moisture Content Equilibrium(kg Moisture Content water/kg drymatter) (kg water/kg drymatter) Xexp 0.3 0.2 0.2 BET Xexp GAB BET Iglesias-Chirife GAB Iglesias-Chirife 0.1 0.1 0.0 0 0.0 0.2 0.4 0.6 Water Activity, aw 0.8 0 0.2 0.4 0.6 Figure 4.6 Best fitted models for desorption isotherms of matooke Water Activity, at a 60˚C 0.8 1 1 w Figure 4.6 Best fitted models for desorption isotherms of matooke at 60˚C Figure 4.6 Best fitted models for desorption isotherms of matooke at 60˚C Establishment of moisture sorption isotherms for matooke 44 Table 4.2 Model constants and statistical errors for adsorption and desorption isotherms Model Constants Adsorption 30˚C BET Mm C R2 Desorption 60˚C 50˚C 60˚C 0.044 13.721 0.960 15.064 0.0104 0.038 22.020 0.977 5.294 0.0074 0.0519 24.845 0.829 19.810 0.0262 0.035 17.330 0.971 5.028 0.0131 0.087 4.104 0.737 0.983 3.091 0.0022 0.050 16.051 0.971 0.978 3.653 0.0072 0.059 20.572 0.862 0.999 1.255 0.0018 0.031 0.627 0.800 0.997 5.868 0.0040 E (%) RSME 91.743 14.780 0.997 3.032 0.0028 192.427 14.696 0.931 1.366 0.0127 208.049 14.814 0.963 12.543 0.0122 73.618 8.762 0.995 4.733 0.055 Smith C1 C2 R2 E (%) RSME 0.017 0.108 0.934 8.712 0.0053 0.021 0.110 0.964 2.404 0.0093 0.031 0.097 0.997 2.857 0.0035 -0.032 0.140 0.999 2.937 0.0030 IglesiasChirife C1 C2 R2 0.035 0.046 0.934 25.075 0.0134 0.039 0.049 0.966 9.011 0.0089 0.065 0.027 0.945 17.414 0.0149 0.034 0.034 0.970 6.657 0.1488 1.495 5.433 0.961 15.824 0.0336 2.282 3.311 0.905 15.978 0.0149 0.740 15.146 0.976 10.998 0.00984 9.806 0.879 0.963 20.012 3.2300 E (%) RSME GAB Mm C k R2 E (%) RSME Chung-Pfost C1 C2 R2 E (%) RSME Henderson C1 C2 R2 E (%) RSME Establishment of moisture sorption isotherms for matooke 45 Table 4.2 shows the model constants and statistical errors for adsorption and desorption isotherms. It was observed from Table 4.2, that the relationship between monolayer for BET and GAB for all cases, both adsorption and desorption are consistent with Timmermann et al., (2001), for which the BET monolayer was less than the GAB monolayer; and the constant (c) for BET was greater than that of GAB in both adsorption and desorption isotherms. All the BET and GAB monolayer values fall within the monolayer for starchy foods which generally range from 0.032 – 0.16 (kg water/kg dry matter) (Siripatrawan & Jantawal, 2006). Also, the GAB constants for adsorption are 0.737 and 0.971 at 30˚C and 60˚C respectively, while those of desorption are 0.862 and 0.800 at 50˚C and 60˚C respectively. This shows that they all fall within the recommended range of: 0.7 < k < 1 (Timmermann et al., 2001). The GAB monolayer has been reported to be indicative of optimum moisture content for storage conditions for hydrated foods (Yan et al., 2008; Moreira et al., 2008; Veg-Galvez et al., 2007; Pahlevanzadeh & Yazdani, 2005; Awoyele, 2004). For the case of matooke, the GAB monolayer for adsorption isotherm at 30˚C was 0.087, which was below the recommended safe water activity level of 0.6 (Labuza et al., 1972; Yan et al., 2008; Sahu & Tiwari, 2007). The water activity level of 0.6 corresponds with equilibrium moisture content of 0.11 (kg water/kg dry-matter), equivalent to 10% (wb), on the adsorption isotherm at 30˚C. This implies that the equilibrium moisture content for microbiologically shelf-stable dried matooke is 10% (wb). In this study, all the subsequent drying experiments, the samples were dried up to moisture content of 10 % (wb) which was used as basis for analysis. The six models fitted to describe both the adsorption and desorption isotherms for matooke were: BET, GAB, Henderson, Iglesias-Chirife, Smith and ChungPfost. The models with R2 > 0.95 and E ≤ 10% were considered to be the best fitting models to describe matooke moisture isotherm. Figure 4.3 shows the best fitted models for adsorption isotherms which include: GAB, Chung-Pfost, and Smith for adsorption at 30˚C and Figure 4.4 shows that BET, GAB, ChungPfost, Smith and Iglesiau-Chirife are the best fitting models for adsorption at 60˚C. While Figure 4.5 shows the best fitting for desorption isotherms at 50˚C which include: BET, GAB and Henderson, and Figure 4.6 shows that BET, GAB, and Iglesias-Chirife were the best fitting models for desorption isotherms at 60˚C. The GAB model best described all the adsorption and desorption Establishment of moisture sorption isotherms for matooke 46 isotherms. The results for statistical errors used to evaluate the best fitting models together with model constants are given in Table 4.2. It was observed from Table 4.2, that all models fitted had a small value for RMSE as computed using equation 4.1, which indicated that all models could be used to describe both adsorption and desorption moisture isotherms for matooke. However, the computed values of E% using equation 4.2 indicated that the BET and Iglesias-Chirife model, the E-values were greater than 10% for adsorption and desorption moisture isotherms at 30˚C and 50˚C respectively. Chung-Pfost model, the E-value was greater than 10% for desorption moisture isotherm at 50˚C. The Henderson model, the E-value was greater than 10% for all the adsorption and desorption moisture isotherms for matooke. This indicated that although the RMSE had small values for all models, the E-value eliminated those models whose values were greater than 10% as not being good for describing the moisture sorption isotherms for matooke. Therefore only the GAB model best described all the adsorption and desorption moisture isotherms for matooke. Establishment of moisture sorption isotherms for matooke 53 4.3.3 Hysteresis for matooke 4.3.4 Hysteresis for matooke Figure 4.7 shows the hysteresis behavior for matooke. Figure 4.7 shows the hysteresis behavior for matooke. 0.3 Equilibrium moisture content (kg water/kg dry matter) Desorption 60˚C Adsorption 60˚C 0.2 0.1 0.0 0 0.2 0.4 0.6 0.8 1 Water activity, aw Figure 4.7 Hysteresis for matooke Figure 4.7 Hysteresis for matooke From Figure 4.7 it was observed that matooke exhibits hysteresis which was consistent with previous work on bananas and starch containing foods (Johnson & Brennan, 2000, Al-Muhtaseb et al., 2004), which indicated that the fresh matooke holds more moisture than dry ones over the entire range of water activity (aw). The hysteresis phenomenon was exhibited by the moisture sorption isotherms at 60˚C. However, it has been reported by previous researchers that the hysteresis can be used as a food quality index, with increased hysteresis being indicative of reduced stability and reduced or absence of Establishment of moisture sorption isotherms for matooke 47 From Figure 4.7 it was observed that matooke exhibits hysteresis which was consistent with previous work on bananas and starch containing foods (Johnson & Brennan, 2000, Al-Muhtaseb et al., 2004), which indicated that the fresh matooke holds more moisture than dry ones over the entire range of water activity (aw). The hysteresis phenomenon was exhibited by the moisture sorption isotherms at 60˚C. However, it has been reported by previous researchers that the hysteresis can be used as a food quality index, with increased hysteresis being indicative of reduced stability and reduced or absence of hysteresis being indicative of improved stability of stored product (Caurie, 2007). Also it was reported that the hysteresis is dependent on the sorption temperatures (Sun & Byrne, 1998; Aviara et al., 2006; Vega-Gálvez et al., 2008). It was further reported that the loop becomes wider with increase in temperature (Aviara et al., 2006; Vega-Gálvez et al., 2008; Raji & Ojedrian, 2011). It is therefore, recommended that further investigations on hysteresis at different temperatures should be carried out in order to establish the temperature range within which there is increased hysteresis (being indicative of reduced stability) and reduced hysteresis (being indicative of improved stability) of stored dried matooke. 4.3.4 Isosteric Heat of Sorption for matooke The isosteric heat of sorption for adsorption isotherm is a measure for the energy released during adsorption. While that for desorption isotherm is the energy required for breaking the intermolecular forces between the molecules of water vapor and the surface of adsorbent (Ciro et al., 2008). The GAB model was used to predict the water activity (aw) at different equilibrium moisture content, which was used in the analysis. The isosteric heat of sorption for adsorption and desorption isotherms were calculated from the slope of equation 2.5 and 2.6. Figure 4.8 Isosteric heats of sorption for adsorption and desorption of matooke as a function of moisture content. forces between the molecules of water vapor and the surface of adsorbent (Ciro et al., 2008). The GAB model was used to predict the water activity (aw) at different equilibrium moisture content, which was used in the analysis. The isosteric heat of sorption for adsorption and desorption isotherms were calculated from the slope of equation 2.5 and 2.6. Figure 4.8 Isosteric heats of sorption for adsorption and 48 desorption of matooke as a function of moisture content. Establishment of moisture sorption isotherms for matooke 20,000 Isosteric Heat of Sorption (kJ/kg) Desorption Qst (kJ/kg) Adsorption Qst (kJ/kg) 15,000 Latent heat of Vaporization (kJ/kg) 10,000 5,000 - 0.1 0.2 0.3 Equilibrium Moisture Content (kg water/kg drymatter) 0.4 4.8 Isosteric heatsof of sorption sorption for adsorption and desorption of matooke as a function of moisture Figure 4.8Figure Isosteric heats for adsorption and desorption of matooke content. as a function of moisture content. It was observed from Figure 4.8 that the isosteric heat of sorption decreases It was observed from Figure 4.8 that the isosteric heat of sorption decreases with increasing moisture with increasing moisture content for both adsorption and desorption, this was content for both adsorption and desorption, this was consistent with literature on bananas (Yan et al., consistent2008). withTheliterature onof bananas (Yan etwere al.,higher 2008). The heat of isosteric heat sorption for desorption than that for isosteric adsorption, indicating that moredesorption energy is required for desorption adsorption process. The decrease of isosteric sorption for were higherprocess than than thatfor for adsorption, indicating that heat of sorption with increasing moisture content could be related to the decrease of active sites leading more energy is required for desorption process than for adsorption process. to reduced energy required to break the bond between water molecules and product surface, as the water The decrease of sorption with increasing moisture content occupiesof theisosteric active sites atheat high moisture content. This trend has been reported for bananas (Johnson & Brennan, 2000; Pahlevanzadeh & Yazdani, 2005; Falade & Awoyele, 2005; Ciro et al., 2008; Yan et al., could be related to the decrease of active sites leading to reduced energy 2008; Hossain et al., 2001) and for pitahaya fruits (Yala et al., 2011). required to break the bond between water molecules and product surface, as the water occupies the active sites at high moisture content. This trend has been reported for bananas (Johnson & Brennan, 2000; Pahlevanzadeh & Yazdani, 2005; Falade & Awoyele, 2005; Ciro et al., 2008; Yan et al., 2008; Hossain et al., 2001) and for pitahaya fruits (Yala et al., 2011). As the moisture further increases, xe ≥ 0.3, the sorption heat tends to latent heat of vaporization of pure water, which indicates that the moisture exists in its free form (Oluwamukomi et al., 2008; McLaughlin & Magee, 1998). The values of isosteric heat of sorption reported in this study revealed the existence of large amounts of bound water. The total sorption heat, Qst, are higher than the latent heat of vaporization of pure water, indicating that the energy of binding between the water molecules and the sorption sites is higher than the energy which holds the water molecule of pure water together in liquid phase (Al-Muhtaseb et al., 2004). higher than the energy which holds the water molecule of pure water together in liquid phas Muhtaseb et al., 2004). Establishment of moisture sorption isotherms for matooke 49 The relationship between the net isoteric heat sorption and equilibrium moisture content was determ The relationship between the net isoteric heat sorption and equilibrium using the SPSS 16 in order to determine the constants of equation 2.7. The exponential relationshi moisture content was determined using the SPSS 16 in order to determine the total desorption and adsorption isosteric heatrelationships of sorption, equation 4.3 was used to calculate constants of equation 2.7. The exponential for total desorption desorption and adsorption isosteric heat of sorption for matooke. and adsorption isosteric heat of sorption, equation 4.3 was used to calculate total desorption and adsorption isosteric heat of sorption for matooke. 𝑄𝑠𝑡 = 20870 𝑒𝑥𝑝 � −𝑥𝑒 0.036 �+𝜆 , 2 R = 0.987 4.3 The exponential relationships for both adsorption and desorption isotherm of The exponential relationships for both adsorption and desorption describe w matooke, describe well the dependences of isosteric heat of isotherm sorptionofonmatooke, the dependences of isosteric heat of sorption on the equilibrium moisture content. This was consisten equilibrium moisture content. This was consistent with previous work on fruits previous work on(Ciro fruits et andal., vegetables (Ciro et et al., 2011;2012; Seid & Henel, 2012; and vegetables 2008; Ayala al., 2008; 2011;Ayala Seidet&al., Henel, Gálvez et al.,et2008). It was Itobserved from Figure that there was there inverse exponential relatio Vega-Gálvez al., 2008). was observed from4.8 Figure 4.8 that was inverse exponential relationship theand totalequilibrium isosteric heat of sorption between the total isosteric heat between of sorption moisture contentand of matooke. The equilibrium moisture content of matooke. The total isosteric heat of sorption isosteric heat of sorption (Qst) increased with decreased equilibrium moisture content. The total iso ) increased with The total (Qheat of sorption fordecreased adsorption equilibrium isotherm formoisture matookecontent. ranged from 4,586isosteric – 2,386kJ/kg for equil st heat of sorption for dry matooke fromWhile 4,586the– total isosteric h moisture content for fromadsorption 0.01 – 0.3 isotherm (kg water/kg matter) ranged respectively. 2,386kJ/kg equilibrium moisture contentranged from 0.01 0.3 (kg water/kg sorption forfor desorption isotherm for matooke from –18,194 – 2,391 kJ/kgdry for equilibrium mo matter) respectively. While the total isosteric heat of sorption for desorption content from 0.01 – 0.3 (kg water/kg dry matter) respectively. These results are comparable with isotherm for matooke ranged from 18,194 – 2,391 kJ/kg for equilibrium range of what was reported for bananas and plantain, the difference is attributed to the mo moisture content from 0.01 – 0.3 (kg water/kg dry matter) respectively. These dependence of moisture sorption isotherms. The moisture content used in obtaining the isosteric h results are comparable within the range of what was reported for bananas and sorption should be put in consideration when comparing with values in literature, Ciro et al., (2 plantain, the difference is attributed to the moisture dependence of moisture reported (qst ≈1670 - 215 kJ/kg for 0.05 – 0.26 kg water/kg dry matter), Johnson & Brennan, ( sorption isotherms. The moisture content used in obtaining the isosteric heat reported (Q ≈ 4117 – 2665 kJ/kg (73 – 47 (kJ/g mol) for 0.02 - 0.18 kg water/kg dry matter). of sorption stshould be put in consideration when comparing with values in literature, Ciro et al., (2008), reported (qst ≈1670 - 215 kJ/kg for 0.05 – 0.26 kg water/kg dry matter), Johnson & Brennan, (2000) reported (Qst ≈ 4117 – 2665 kJ/kg (73 – 47 (kJ/g mol) for 0.02 - 0.18 kg water/kg dry matter). It has been reported by previous researchers that the total isosteric heat of sorption for desorption isotherm is an estimate of the energy requirement for dehydration (Kiranoudis et al., 1993; Pahlevanzadeh & Yazdani, 2004; Kaymak-Ertekni & Gedik, 2004; Yazdani et al., 2006; Yang et al., 2012). The energy required to dry matooke was obtained by integrating equation 4.3 which gives the total isosteric heats of sorption for desorption of matooke from 2012). Establishment of moisture sorptionequation isotherms matooke energy required to dry matooke was obtained by integrating 4.3for which gives the total 50 The isosteric heats of sorption for desorption of matooke from 0.11 to 3.5 (kg water/kg dry-matter) using the 0.11 3.5 of (kgvaporization water/kg dry-matter) latentof heat of vaporization pure55˚C latenttoheat of pure water using at the the average the temperatures in theofstudy water at the average of the temperatures in the study 55˚C (λ = 2,386.13kJ/ 2,386.13kJ/kg). kg). � 3.� �20870 𝑒𝑥𝑝 � 0.11 (λ = −𝑥𝑒 � + 2386.13� 𝑑𝑥 = 8,124 𝑘𝐽⁄𝑘𝑔 𝑜𝑓 𝑤𝑎𝑡𝑒𝑟 𝑟𝑒𝑚𝑜𝑣𝑒𝑑 0.036 Therefore, the energy required to dry matooke from 3.5 to 0.11 (kg water/ Therefore, the energy required to80 dry –matooke 0.11 (kg water/kg dry-matter) (equivalent 80 – kg dry-matter) (equivalent 10 % from wb)3.5 is to8,124kJ/kg of water removed. 10 % wb) is 8,124kJ/kg of water removed. Therefore, the total energy required to dry kg of fresh Therefore, the total energy required to dry 1 kg of fresh matooke from 801 – matooke from 80 – 10% (wb) is 5,768 kJ. 10% (wb) is 5,768 kJ. 4.4 Conclusion and Recommendations 4.4 Conclusion and Recommendations This study has bridged the gap of lack of information on the sorption isotherm study has bridged the gap of lack of information on the sorption isotherminformation of matooke. The ofThis matooke. The study has therefore, provided very useful on study has therefore, provided very useful information on moisture sorption isotherm of matooke, which moisture sorption isotherm of matooke, which can be used for prediction can of be used for prediction of quality, design, modeling and optimization of many processes such as drying, quality, design, modeling and optimization of many processes such as drying, aeration and storage. In addition, it is also important for predicting quality changes during packaging and storage aeration and storage. In addition, it is also important for predicting quality of dried matooke. changes during packaging and storage of dried matooke. This study has specifically provided the following information on moisture This studyisotherm has specifically provided the on the moisture sorptionhave isotherm for matooke sorption for matooke forfollowing the firstinformation time, and following been for the first time, and the following have been revealed: revealed: • Matooke Matookeisotherms isotherms exhibited type II curves behavior which iswhich characteristic of foodstuffs. exhibited type II curves behavior is characteristic of foodstuffs. The equilibrium moisture content decreases with increase in temperature at a given water activity (a ), indicating that it becomes less hygroscopic at high temperatures giving a clear stability • The wequilibrium moisture content decreases with increase in temperature at a given water activity (aw), indicating that it becomes less hygroscopic at high temperatures giving a clear stability domain of matooke after drying leading to extensive shelf life. This compliments to Muranga, (1998), who attributed the extensive shelf life to low sugar and fat. • The GAB model best described all the adsorption and desorption moisture isotherms for matooke. It is recommended to be used by the processors to determine the moisture content during storage. • For microbiologically and shelf stability of dried matooke, it is recommended to dry it to moisture content below or equal to 0.11 (kg water/kg dry matter) equivalent to 10% (wb). Establishment of moisture sorption isotherms for matooke 51 • The storage condition should be below or equal to 30˚C, but due to fluctuations in temperatures and humidity in Uganda because of wet and dry seasons, it is recommended to seal the dried matooke to minimize moisture intake. • The hysteresis phenomenon was exhibited by the moisture sorption isotherms and it could be used as a food quality index. It is recommended that further investigations on hysteresis at different temperatures be carried out to establish the ranges of increased hysteresis being indicative of reduced stability and reduced hysteresis being indicative of improved stability of stored dry matooke. • The isosteric heat of sorption for both isotherms increased with decreased moisture content. The total isosteric heat of sorption for adsorption isotherm ranged from 4,586 – 2,386 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (kg water/kg dry-matter) respectively. While the total isosteric heat of sorption for desorption isotherm ranged from 18,194 – 2,391 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (kg water/ kg dry-matter) respectively. • The total isosteric heat of sorption for desorption isotherm is an estimate of the energy requirement for breaking the intermolecular forces between solid and moisture during the drying process to drying. Thus energy required to dry matooke from 3.5 to 0.11 (kg water/kg dry-matter) (equivalent 80 – 10 % (wb) is 8,124kJ/kg of water removed. Therefore, the total energy required to dry 1 kg of matooke from 80 - 10% (wb) is 5,768kJ. Drying Characteristics of Matooke 52 Chapter Five Objective: Drying Characteristics of Matooke 5.1 Background Hot-air drying is considered as one of the simplest, most appropriate and most economical technology for processing of fruits and vegetables in developing countries (Johnson et al., 1998; Baini and Langrish, 2007). Understanding of phenomena involved in the drying process is a primary requirement when using hot-air drying in order to predict the drying times, and the influence of the processing parameters such as: air temperature, dew point temperature, thickness (size) of the slices and air velocity on the drying characteristics (Johnson et al., 1998). These processing parameters have been reported to have a significant effect on the drying characteristics (Krokida et al., 2003). Application of effective diffusivity method has been reported to be a convenient and practical method for describing moisture changes during the drying process (Kayacier et al., 2004; Baini and Langrish, 2007). In addition, diffusion properties are important for understanding of the drying behavior of food during drying and also for the design and optimization of dryers for specific crops (Rasouli et al., 2011). In order to control and optimization of the drying process for matooke it was important to understand the effect of processing parameters on the drying characteristics. The objective of the investigations described in this chapter was to establish the effect of processing parameters on the drying characteristics of matooke and the specific objectives were to determine the effect of: 1. Process parameters on drying characteristics. 2. Drying rate constant, effective moisture diffusivity and activation energy. 3. Processing parameters on total color change (ΔE*). 4. Drying air temperature on surface product temperature. 5. Product temperature on the drying time. Drying Characteristics of Matooke 5.2 53 Materials and Methods The research dryer was developed in corporation with M/s Innotech Ingeniersgesellschaft mbH, Altdorf (Böblingen), Germany and University of Applied Science and Technology, Konstanz, Germany. It comprises of: drying tower for both overflow and through flow unit, water conditioning unit, humidifying unit, fan, pyrometer, digital camera and data-logger. The processing parameters (independent variables): air velocity, drying air temperature, dew-point temperature were set and continuously monitored and data-logged during the drying process. Similarly, the dependent variable: mass of samples were measured continuously using a precision balance and product temperature were measured by a precision infrared pyrometer. The dependent variables were monitored and automatically recorded by datalogged at a 10 sec interval during the drying process. The photographs of the samples were taken by the digital camera at 30 min time intervals. Figure 5.1, shows both the pictorial view of major components of the research dryer and Figure 5.2, shows the system layout of the research dryer. All matooke cv. Mbwazirume (musa sp. AAA-EABH) samples used for the experiments were at harvest maturity age of 17 weeks based from the findings of Chapter 3 of this study. Over flow duct Humidifying Unit Manometer Monitoring scale Figure 5.1 Pictorial view of the research dryer unit Fan Drying Characteristics of Matooke 54 Preparation of Samples The fresh samples were harvested, from the marked banana as described in Chapter 3 of this study, washed, peeled and sliced to specified uniform thickness using slicing machine. The thicknesses of sliced samples were continuously monitored by physical measuring using a Vanier Clippers to confirm the actual thickness (Chapter 3). After slicing the samples were pretreated with 1% sodium metabisulphite solution according to patent No. AP/P/2005/003308. The dried sample were sealed in airtight bags and kept in the fridge at 4˚C for subsequent tests in the laboratory. Samples for determining the pasting properties and starch content were milled (Chapter 3). All experiments were carried out at the Presidential Initiative on Banana Industrial Development (PIBID) - Technology Business Incubator at Bushenyi, Uganda. Table 5.1 shows the measuring equipment used in these investigations. Table 5.1 Measuring equipment Type of measurement Equipment Model Manufacturer Recording experimental data Data-logger Agilent 34970A, Agilent Technologies, Inc. Sample weight Precision balance Sartorius LA16001S Sartorius AG, Goettingen, Germany ±0.01g Capacity ≤12 kg Product temperature Precision infrared pyrometer Heitronics Heitronics Infrarot Messtechnik GmbH, Wiesbden, Germany ±0.05˚C +0.7% of temp difference between room & object KT15 II Accuracy Air temperature Pt 100 Thermocouple ±0.02˚C Dew pint temperature Pt 100 Thermocouple ±0.02˚C The independent variables (processing parameters) were: • Air temperature (50, 60 and 70˚C) • Dew point temperature (15, 25 and 35˚C). • Air velocity 3, 4.5 and 6 m/s for the overflow mode • Slice thickness (3, 5 and 7 mm) Drying Characteristics of Matooke 55 The dependent variables were: • Drying time (min). • Color of the sample. • Sample weight The experiments were carried out in triplicates and the results were analyzed using SPSS 16 for Windows to fit the experiment data. Figure 5.2: Complete system layout of the units for the research dryer. Key: W: water conditioning unit, V: air fan, H: humidifier, C: computer, D: data-logger, T: drying chamber, X: through flow, U: overflow, H1 = heating unit (heaters), a: air temperature, b, g: air velocity (hot air manometer), c: a i r temperature, d: pyrometer, e: camera, f: balance, h: air temperature, i : heaters, j:fan, k: water condition unit Figure 5.3 shows the representation of the whole process of drying on the psychrometric chart. The air is sucked by the fan at point A at ambient conditions, it is cooled in the humidifying unit to point B, humidified to the set dew point temperature point C, and it is then heated by heaters to a set temperature D. Drying Characteristics of Matooke 56 Figure 5.3: The drying process as represented on the psychrometric chart. 5.3 Results and Discussion 5.3.1 Effect of thickness on drying characteristics of matooke Figure 5.4 shows the effect of thickness of the slices on the drying characteristics of matooke. The investigations were carried out for slice thicknesses of 3, 5, and 7 mm at air temperature of 60˚C, dew-point temperature of 15˚C and air velocity of 4.5 m/s. Figure 5.4 5 Mosture Content (kg water/kg dry metter) Thickness 4 3 mm 3 5 mm 7 mm 2 1 0 - 100 200 300 400 Time (min) Figure 5.4 Effect of thickness on the drying characteristics of matooke (air temperature 60˚C, dew-point temperature 15˚C and air velocity 4.5 m/s) Figure 5.5 1.2 Figure 5.4 5 Thickness Drying Characteristics of Matooke 4 57 Mosture Content (kg water/kg dry metter) 3 mm 5 mm It was observed from Figure 5.4 that thickness of the samples has a significant 3 effect on the drying time, in that the drying time increased with increase in 7 mm 2 thickness. This was in line with previous researchers who reported that as thickness increases, the diffusion path, becomes longer with increase in 1 thickness and vice versa (Krokida et al., 2000; Nguyen and Price, 2007; Islam et al., 2012). The drying time required to reach the recommended 0 microbiologically shelf-stable 4)300of 0.10 (kg 100product (Chapter 200 400 water/kg dry Time (min) matter) were 115, 250 and 370 min for 3, 5 and 7 mm respectively. Figure 5.5 shows the plot of moisture ratio of different thickness versus drying time at constant air temperature of 60˚C, dew point temperature of 15˚C and air velocity of 4.5m/s. Figure 5.5 1.2 Thickness Moisture Ratio (x/xo) 1 3 mm 5 mm 0.8 7 mm 0.6 0.4 0.2 0 0 100 200 300 400 Time (min) Figure 5.5 Plot of moisture ratio of different thickness versus the drying time (air temperature 60oC, dew point temperature 15oC, air velocity 4.5 m/s). It was observed from Figure 5.5 that there was also a direct relationship between thickness and moisture ratio, in that the moisture ratio increased with increase in thickness at any given time during the drying process. Figure 5.6 shows the effect of drying rate constant at different thickness on the moisture content of matooke at constant air temperature of 60˚C, dew point temperature of 15˚C, air velocity of 4.5 m/s. Drying Characteristics of Matooke 58 Figure 5.6 0.1 Thickness Drying Rate (kg water/kg dry matter)/min 3 mm 0.08 Warm-up 5 mm 7 mm Falling rate perids 0.06 2nd 1st 0.04 0.02 0 0 1 2 3 4 5 Moisture Content (kg water/kg dry matter) Figure 5.6 Effect of drying rate constant of different thickness on the moisture content of matooke (air temperature 60oC, dew point temperature 15oC, air Figure 5.7 velocity 4.5 m/s) Time (min) Ln(Moisture Ratio) 0 it was observed that the drying of matooke takes place in three From Figure 5.6 0 100 200 400 steps: the warm-up, the two falling rate periods:300 the first falling rate and second falling rate. However, the second falling rate was more pronounced on the 3 -1 mm curve. This was in agreement with the previous research on bananas and plantains (Prachayawarakon et al., 2008). Since drying dominantly took place -2 during the first and second falling rate periods, it implied that diffusion was the Thicknesscontrolling the moisture migration from the samples. dominant mechanism -3 The diffusion model could fit the moisture content better than the empirical 3 mm model for matooke (Baini and Langrish, 2007). 5 mm -4 7 mm Figure 5.7 shows the plot of natural logarithm of moisture ratio at different -5 thickness versus the drying time, at air temperature of 60˚C, dew point temperature of 15˚C and air velocity of 4.5 m/s. The effective diffusivity and drying rate constant were calculated from the slope of the plot of the natural logarithm of moisture ratio versus time employing equations 2.12 and 2.15 respectively. Moisture Content (kg water/kg dry matter) Drying Characteristics of Matooke 59 Figure 5.7 Time (min) 0 0 100 200 300 400 Ln(Moisture Ratio) -1 -2 -3 Thickness 3 mm -4 5 mm 7 mm -5 Figure 5.7 Plot of natural logarithm of moisture ratio at different thickness versus the drying time of matooke (air temperature 60˚C, dew point temperature 15˚C and air velocity 4.5 m/s). Table 5.2 shows the effective diffusivity and drying rate constant of different slice thickness (3, 5 and 7 mm) at air temperature 60˚C, dew point temperature 15˚C, and air velocity 4.5 m/s. The fitted plots are shown in Figures 5.1A and 5.2A (Appendix). Table 5.2: Effective diffusivity and drying rate constant for matooke of different thickness (air temperature 60˚C, dew point temperature 15˚C, and air velocity4.5 m/s). Thickness (mm) Drying Time (min) Drying Rate Constant k (/h) Effective Diffusivity Deff (m2/s) R2 3 115 2.10 5.32E-10 0.984 5 250 0.96 6.75E-10 0.987 7 370 0.60 8.27E-10 0.990 It was observed from Table 5.2 that there was an inverse relationship between slice thickness and drying rate constant, in that the drying rate constant decreased with increase in slice thickness. This was in line with previous researchers on bananas and plantains (Johnson et al., 1998; Maskan, 2000; Nguyen and Price, 2007; Islam et al., 2012). This trend is attributed to the dependence of water removal from high moisture fruit on the migration of Drying Characteristics of Matooke 60 water from internal sites of fruit cells towards the surface. Therefore, the smaller the thickness, the shorter the pathway and the higher the drying rate constant, thus taking a shorter drying time to reach the desired moisture content and vice versa (Nguyen and Price, 2007). It was further observed from Table 5.2, that there was a direct relationship between the thickness and effective moisture diffusivity, implying that the effective moisture diffusivity increased with increase in thickness. This was also in line with previous works on bananas and plantains (Johnson et al., 1998; Nguyen and Price, 2007; Rasouli et al., 2011; Islam et al., 2012). The increase in effective moisture diffusivity with increased slice thickness is attributed to the pathway in that the longer the pathway (thickness) the higher the effective moisture diffusivity and vice versa (Nguyen and Price, 2007; Rasouli et al., 2011; Islam et al., 2012). Even if the effective moisture diffusivity increases with thickness, but because of the longer pathway for the moisture to move from the cells to the surface, it takes longer time to reach the desired moisture content. From practical point view, considering that 12 hours are available for working in a day, the number of batches and amount of product which can be processed in a day are given in Tables 5.3 for single layer drying and 5.4 for multi layer. Table 5.3 Single layer drying Thickness (mm) Time (h) Capacity (kg/m²) Batches in 12h Capacity/ day(kg/m²) 3 1.92 2.33 6 13.98 5 4.17 4.11 3 12.33 7 6.17 5.56 2 11.12 Table 5.4 Multi-layer drying (Thickness - 3mm) Capacity/ batch (kg/m²) Time (h) Batches in 12h Capacity/ day (kg/m²) 3.53 3.48 3 10.59 6.78 5.07 2 13.56 10.52 6.26 1 10.52 It was observed from Table 5.3 and 5.4 that for single layer drying, thickness of 3 mm was the best. In actual practical terms of commercial processing, Drying Characteristics of Matooke 61 thickness is considered in terms of equivalent loading capacities (kg/m²). Considering the loading capacities for single layer drying and multi-layer drying, it is more practical to run 2 batches than 6 batches per day. For commercial processing of matooke, it is recommended to employ multi-layer drying of loading capacity equal or less than 7 kg/m². 5.3.2 Effect of drying air temperature on drying characteristics of matooke Figure 5.8 shows the effect of drying air temperature on the drying characteristics. The investigations were carried out for three different drying air temperatures: 50, 60 and 70˚C at constant slice thicknesses of 3 mm, dewpoint temperature of 15˚C and air velocity of 4.5 m/s. Figure 5.8 5 Drying air temperature Moisture Content (kg water/kg dry matter) 4 50˚C 60˚C 3 70˚C 2 1 0 0 50 100 150 200 250 Time (min) Figure 5.8 Effect of drying air temperature on the drying characteristics of matooke (at dew point temperature 15oC, air velocity 4.5 m/s and thickness 3mm) Figure 5.9 Moisture Ratio (x/xo) It was observed from Figure 5.8 that there was an inverse relationship 1.2 between drying air temperature and drying time, in that the drying time Drying air temperature decreased with increase in drying air temperature. This was in agreement with 50˚C previous research on bananas and plantain (Nguyen and Price, 2007, Islam 0.8 60˚C et al., 2012). This implied that the moisture removal increased with increase in drying air temperature. Since drying is a simultaneously heat and mass 70˚C transfer process, the higher the drying air temperature, the higher the capacity 0.4 0 - 50 100 150 200 250 Figure 5.8 5 Drying air temperature 62 Drying Characteristics of Matooke 4 Moisture Content (kg water/kg dry matter) 50˚C 60˚C to transfer heat to the drying product leading to increase in moisture removal 3 70˚C from the product. In addition, the higher the drying air temperature, the lower the relative humidity in the stream of air. This implies increased capacity of 2 the air stream to pickup moisture from the product surface thus enhancing moisture removal.1 All these factors coupled together resulted in reduced time required to reach the desired moisture content at a higher air temperature and vice versa as shown in Figure 5.8. The drying time required to reach moisture 0 0 100 200 content of 0.1 (kg water/kg50dry matter) was150 200, 115 and 90250min. for 50, 60 Time (min) and 70˚C respectively. Figure 5.9 shows the effect of moisture ratio at different drying air temperature on drying time. The investigations were carried out at constant dew point temperature of 15˚C, air velocity of 4.5 m/s and thickness of 3 mm. Figure 5.9 1.2 Drying air temperature Moisture Ratio (x/xo) 50˚C 0.8 60˚C 70˚C 0.4 0 - 50 100 150 Time (min) 200 250 Figure 5.9 Plot of moisture ratio at different drying air temperature versus drying time of matooke (dew point temperature 15oC, air velocity 4.5 m/s and thickness 3mm) It was observed from Figure 5.9, that there was an inverse relationship between air temperature and moisture ratio in that the moisture ratio decreased with increase in drying air temperature at a given drying time. This trend was observed by previous researchers on bananas and plantains (Islam et al., 2012). Figure 5.10 shows the effect of drying rate constant at different drying air temperature on moisture content, at constant dew point temperature of 15˚C, air velocity of 4.5 m/s and thickness of 3 mm. Drying Characteristics of Matooke Figure 5.10 63 0.12 Drying Rate (kg water/kg dry matter)/min Drying air temperature 50˚C 60˚C 0.08 70˚C 0.04 0 0 2 4 Moisture Content (kg water/kg dry matter) 6 Figure 5.10 Effect of drying rate constant at different drying air temperature on moisture content of matooke (dew point temperature 15oC, air velocity4.5 Figure 5.11 m/s and thickness 3mm) 0 Time (min) Ln(Moisture Ratio) It was also observed in Figure 5.10, 100.00 that the drying of matooke 200.00 took place in Drying air temperature three steps: the warm-up and two falling rate periods: the first falling rate and -1 second falling rate. It was further observed from50˚C Figure 5.10 that the drying 60˚C rate constant for matooke was directly related to the air temperature, in that -2 70˚C drying rate constant increased with increase in drying air temperature. This was in line with -3 the previous work done on plantains and bananas (Krokida et al., 2000; Nguyen and Price, 2007; Rasouli et al., 2011; Islam et al., 2012). This implied that -4 the rate of moisture removal increased with temperature at any given moisture until the product was relatively dry. -5 Figure 5.11 shows the plot of natural logarithm of moisture ratio at different drying air temperatures versus the drying time of matooke at constant dew point temperature of 15˚C, air velocity of 4.5 m/s and thickness of 3 mm. Table 5.5 shows there was a direct relationship for both drying rate constant and effective diffusivity with drying air temperature. This was in line with previous researchers on bananas (Nguyen and Price, 2007). The fitted plots are shown in Figures 5.3A and 5.4A in appendix A. 0 0 2 4 Moisture Content (kg water/kg dry matter) 6 Drying Characteristics of Matooke 64 Figure 5.11 Time (min) 0 - 100.00 -1 Ln(Moisture Ratio) 200.00 Drying air temperature 50˚C 60˚C -2 70˚C -3 -4 -5 Figure 5.11 Plot of natural logarithm of moisture ratio at different drying air temperatures versus the drying time of matooke (dew point temperature 15˚C, air velocity 4.5 m/s and thickness 3 mm) Table 5.5 Effective diffusivity and drying rate constant for matooke at different drying air temperature (dew point temperature 15˚C, air velocity 4.5 m/s and thickness 3mm) Air Temperature (oC) Drying Rate Constant k (/h) Effective Diffusivity Deff (m2/s) R2 50 1.34 3.40E-10 0.988 60 2.10 5.32E-10 0.984 70 2.49 6.31E-10 0.984 Table 5.6 shows the comparisons of the results of effective diffusivity from the study with what has been reported in literature on bananas and plantains. However, though there some differences, the effective diffusivity of matooke falls within the range reported for bananas and plantains ranging in the order of 10-10 m2/s. Drying Characteristics of Matooke 65 Table 5.6: Comparisons of current results with literature Effective Moisture Diffusivity (m²/s) Author 1.25E-10 to 2.19E-10, Islam et al., (2012) 3.16E-10 to 18.01E-10 Johnson et al., (1998) 3.40E-10 - 6.31E-10 Results of the study (Kawongolo, 2013) 4.0E-10 to 7.0E-10 Phoungchandang and Woods, (2000) 4.6E-10 to 13.0E-10 Dandamrongrak et al., (2002) 6.61E-11 to 2.41E-10 Karim and Harlader, (2004) Calculation of the Activation Energy Figure 5.12 shows the plot of natural logarithm of effective diffusivity versus the reciprocal of drying air temperature, at constant dew point temperature of 15˚C, air velocity of 4.5 m/s and thickness of 3 mm. The dependence of effective diffusivity (Deff) was described by Arrhenius as given in equation 2.16. The activation energy was calculated from the slope equation 2.18 (Doymaz, 2007). The activation energy (Ea) for matooke in 28.5 kJ/mol (equivalent to1,605 kJ/kg). The activation energy of matooke lies within the general range of 12.7 – 110 kJ/mol for food materials (Zogzas et al., 1996, Babalis and Drying characteristics 74 Belessiotis, 2004).for matooke 1/(abs. Tempeature) (1/K) 0.0029 0.003 0.0031 0.0032 Ln(Deff) -21.2 y = -3433.2x - 11.135 R² = 0.9456 -21.6 -22 Figure5.12 5.12 Plot of of natural logarithm of effective diffusivity versus the reciprocal of drying airthe temperature Figure Plot natural logarithm of effective diffusivity versus reciprocal ˚ (dew point temperature 15 C, air velocity 4.5 m/s and thickness 3 mm) ˚ of drying air temperature (dew point temperature 15 C, air velocity 4.5 m/s and thickness 3 mm) The activation energy has been reported to be the energy required to break the intermolecular bonding and bring the molecule to the surface where it acquires the latent heat of evaporation and it evaporates from the surface (Demirel and Turhan, 2003; Baini and Langrish, 2007). This indicates that the activation energy (Ea) is analogous to the net isoseteric heat of sorption for desorption isotherms (qst). The activation energy (Ea) for matooke (1,605kJ/kg) falls within the range of net isoseteric heat of sorption for Drying Characteristics of Matooke 66 The activation energy has been reported to be the energy required to break the intermolecular bonding and bring the molecule to the surface where it acquires the latent heat of evaporation and it evaporates from the surface (Demirel and Turhan, 2003; Baini and Langrish, 2007). This indicates that the activation energy (Ea) is analogous to the net isoseteric heat of sorption for desorption isotherms (qst). The activation energy (Ea) for matooke (1,605kJ/ kg) falls within the range of net isoseteric heat of sorption for desorption (qst) for moisture content ranging from 0.05 to 0.1 (kg water/kg dry matter) which is (5,204 - 1,298 kJ/kg) respectively. However, the net isoseteric heat of sorption for desorption isotherms (qst) at 0.1 (kg water/kg dry matter), (the moisture considered in the drying experiments), (1,298 kJ/kg) is smaller than the activation energy Ea (1,605 kJ/kg). This was in line with what was observed by Demirel and Turhan, (2003), who reported that the comparison between Ea and qst suggests that moisture molecules travels effectively in liquid form in the slices during the drying process. 5.3.3 Effect of dew point temperature on drying characteristics of matooke Figure 5.13 shows the effect of dew temperature on the drying characteristics. The investigation was carried out at three different dew point temperatures: 15, 25 and 35˚C, at constant slice thicknesses of 3 mm, air temperature of 60˚C and Figure air velocity of 4.5 m/s. 5. 13 5 Dew point temperature Moisture Content (kg water/kg dry matter) 4 15˚C 25˚C 3 35˚C 2 1 0 - 50 100 Time (min) 150 200 Figure 5.13 Effect of dew point temperature on the drying characteristics of matooke (air temperature 60˚C, air velocity 4.5 m/s and thickness 3mm). Figure 5.14 5 Dew point temperature 4 Drying Characteristics of Matooke 67 Moisture Content (kg water/kg dry matter) Figure 5. 13from Figure 5.13 that the dew point temperature had low It was observed effect on the time5 of drying; however, there was direct relationship between point temperature dew point temperature and drying timeDew in that the drying time increases with 4 increase in dew point temperature. The direct relationship between dew point 15˚C temperature and drying time could be attributed to high relative humidity 25˚C 3 at higher dew point temperature in the stream 35˚C of air which decreased the capacity of the air stream to pickup moisture from the product surface, hence 2 reducing moisture removal. As a result the time required to reach the desired moisture content at higher dew point temperature increased and vice versa 1 as shown in Figure 5.13. The drying time required to reach moisture content of 0.1 (kg water/kg dry matter) were approximately 115, 135 and 145 min. for 0 50 100 150 200 15, 25 and 35˚C respectively. Time (min) Figure 5.14 shows the plot of moisture ratio of different dew point temperatures versus drying time of matooke at constant air temperature of 60˚C, air velocity of 4.5 m/s and thickness of 3 mm. Figure 5.14 5 Dew point temperature Moisture Content (kg water/kg dry matter) 4 15˚C 25˚C 3 35˚C 2 1 0 - 50 100 Time (min) 150 200 Figure 5.14 Plot of moisture ratio at different dew point temperature versus drying time of matooke (air temperature 60˚C, air velocity 4.5 m/s and thickness 3mm). It was observed from Figure 5.14 that there was a direct relationship between dew point temperature and moisture ratio in that the moisture ratio increased with increased dew point temperature at any given drying time during drying. Drying Characteristics of Matooke 68 Figure 5.15 shows the effect of drying rate constant of different dew point temperature temperatures on moisture content, at constant air temperature of 60˚C, air velocity of 4.5m/s and thickness of 3 mm. Figure 5.15 0.08 Drying Rate (kg water/kg dry matter)/min Dew point temperature 15˚C 0.06 25˚C 35˚C 0.04 0.02 0 0 1 2 3 4 Moisture Content (kg water/kg dry matter) 5 Figure 5.15 Effect of drying rate constant at different dew point temperature on Figure 5.16 the moisture content (air temperature 60˚C, air velocity 4.5 m/s and thickness Time (min) 3 mm). 0 - 50 100 150 200 Ln(Moisture Ratio) It was observed from Figure 5.15 that the drying rate constant for matooke -1 was inversely related to the dew point temperature, in that the drying rate constant decreased with increased dew point temperature. -2 Figure 5.16 shows the plot of natural logarithm of moisture ratio at different dew point temperatures versus drying time, at constant air temperature of -3 point temperature 60˚C, air velocity ofDew 4.5m/s and thickness of 3 mm. -4 -5 15˚C 25˚C 35˚C 0 0 1 2 3 4 Moisture Content (kg water/kg dry matter) 5 Drying Characteristics of Matooke 69 Figure 5.16 Time (min) 0 - 50 100 150 200 Ln(Moisture Ratio) -1 -2 -3 -4 -5 Dew point temperature 15˚C 25˚C 35˚C Figure 5.16 Plot of natural logarithm of moisture ratio at different dew point temperatures verses drying time (air temperature 60˚C, air velocity 4.5 m/s and thickness 3mm). Table 5.7 shows there was a direct relationship between drying rate constant and drying air temperature. Also, a direct relationship was observed between effective diffusivity and thickness. This was consistent with Rasouli et al., (2011); Nguyen and Price, (2007). The fitted plots are shown in Figures 5.5A and 5.6A in appendix A. Table 5.7 Effective diffusivity and drying rate constant for matooke at different dew point temperatures (air temperature 60˚C, air velocity 4.5 m/s and thickness 3mm). Dew Point Temperature (oC) Drying Rate Constant, k (/h) Effective Diffusivity Deff (m2/s) R2 15 2.10 5.32E-10 0.984 25 1.76 4.45E-10 0.988 35 1,72 4.36E-10 0.972 5.3.4 Effect of air velocity on drying characteristics of matooke Figure 5.17 shows the effect of air velocity on the drying characteristics. The investigations were carried out at three different airflows: 3, 4.5 and 6 m/s at constant: thicknesses of 3 mm, air temperature of 50˚C and dew point temperature of 15˚C. Drying Characteristics of Matooke 70 Figure 5.17 4 Air velocity Moisture Content (kg water/kg dry mmatter) 3 m/s 3 4.5 m/s 6m/s 2 1 0 - 100 200 300 Time (min) 400 500 Figure 5.17 Effect of air velocity on the drying time of matooke (air temperature of 50oC, dew point temperature of 15oC and the thickness of 3 mm). Figure 5.18 Moisture Ratio (x/xo) It was observed from Figure 5.17 that there was an inverse relationship 1.2 between air velocity and drying time in that the drying time decreased with velocity increased airflow. The inverse relationship Air between air velocity and drying time could be attributed to the fact that drying is a simultaneously heat and 3 m/s 0.9 4.5 m/s mass transfer process, the higher the air velocity of air stream the greater 6 m/s the capacity of air stream to pickup moisture from the product surface hence 0.6 enhancing moisture removal. In addition, at high air velocity, the air stream has a higher heat transfer coefficient. As a result, the time required to reach 0.3 the desired moisture content at higher air velocity reduced and vice versa as shown in Figure 5.17. The drying time required to reach moisture content of 0.1 (kg water/kg 0dry matter) was approximately 420, 200 and 92 min for 3, 4.5 0 100 200 300 400 500 and 6 m/s respectively. Time (min) Figure 5.18 shows the plot of moisture ratio at different air velocity versus drying time at constant air temperature of 50˚C, dew point temperature of 15˚C and the thickness of 3mm. Time (min) Drying Characteristics Figure 5.18 of Matooke 71 1.2 Air velocity 3 m/s Moisture Ratio (x/xo) 0.9 4.5 m/s 6 m/s 0.6 0.3 0 0 100 200 300 400 500 Time (min) Figure 5.18 Plot of moisture ratio at different air velocity versus drying time (thicknesses of 3mm, air temperature 50˚C and dew point temperature 15˚C) It was observed from Figure 5.18 that there was an inverse relationship between air velocity and moisture ratio in that the moisture ratio decreased with increased air velocity at any given drying time during drying. This indicated that as the air velocity increased, the moisture removal increased. Figure 5.19 shows the effect of drying rate on the moisture ratio of matooke at different air velocity at thicknesses of 3mm, air temperature 50˚C and dew Figure 5.19 15˚C. point temperature Drying Rate (kg water/kg dry matter)/min 0.08 Air velocity 0.06 3 m/s 4.5 m/s 6 m/s 0.04 0.02 Constant drying rate period 0 0 1 2 3 Moisture Content (kg water/kg dry matter) 4 Figure 5.19 Effect of drying rate on the moisture ratio of matooke at different air velocity at thicknesses of 3mm, air temperature 50˚C and dew point temperature 15˚C Figure 5.20 Time (min) 0 0 100 200 300 400 500 Figure 5.19 0.08 Drying Characteristics of Matooke Drying Rate (kg water/kg dry matter)/min 72 Air velocity 0.06 It was observed from Figure3 m/s 5.19 it was observed that drying took place in 4.5 m/s three steps: the warm-up and two falling rate periods. At lower air velocity 6 m/s 0.04 (3 m/s), it was observed that there was a constant drying rate period up to approximately a moisture content of 2 (kg water/kg dry matter), then the falling rate period was observed. It was further observed from Figure 5.19, that there 0.02 was a direct relationship between the dryingConstant rate drying constant and air velocity. rate period This implies that the rate of moisture removal increased with increase in air 0 velocity at any given moisture content. 0 1 2 3 4 Moisture Content (kg water/kg dry matter) Figure 5.20 shows plot of natural logarithm of moisture ratio at different air velocities versus drying time, at constant thicknesses of 3mm, air temperature of 50˚C and dew point temperature of 15˚C. Figure 5.20 Time (min) 0 0 100 200 300 400 500 Ln(Moisture Ratio) -2 -4 Air velocity -6 3 m/s 4.5 m/s 6m/s -8 Figure 5.20 Plot of natural logarithm of moisture ratio at different air velocities versus drying time (drying air temperature 50˚C, dew point temperature 15˚C and the thickness 3mm). The slope was calculated from the slope of the linear segment (Rasouli et al., 2011), and the results obtained from the slope as given in Table 5.8, show that the effective diffusivity of matooke increased with increase in air velocity as shown in Figure 5.38, (Nguyen and Price, 2007). The fitted plots are shown in Figures 5.7A and 5.8A in appendix A. Drying Characteristics of Matooke 73 Table 5.8 Effective diffusivity and drying rate constant for matooke at different air velocities (drying air temperature 50˚C, dew point temperature of 15˚C and the thickness of 3mm). Air velocity (m/s) Drying Rate Constant k (/h) Effective Diffusivity Deff (m2/s) R2 3 0.594 1.50E-10 0.995 4.5 1.344 3.40E-10 0.988 6 1.614 4.09E-10 0.991 5.3.5 Effect of processing parameter on color of matooke Table 5.7 shows the effect of processing parameters on the total color difference (ΔE*). The total color difference (ΔE*) of dried matooke was calculated by subtracting the color of the dried matooke from the color of the fresh matooke (Baini and Langrish, 2007). The total color difference (ΔE*) was calculated form equation 2.4 using the Visual-Basic-based program for analysis of images (Sturm and Hofacker, 2009). It was observed for the Table 5.7, that there was an inverse relationship between the effects of drying parameters on the total color difference (ΔE*). The results of total color difference (ΔE*), was lowest for effect of thickness was 7 mm, but it took the longest time to dry. This was followed by the effect of air velocity which was 6 m/s, and lastly for the effect of drying air temperature was 70˚C. Drying Characteristics of Matooke 74 Table 5.9 Effect of processing parameters on total color difference (ΔE*) Drying parameter Effect of air temperature on total color change Processing Parameters Total color Dew point Air Thickness Air difference (mm) temperature temperature velocity(m/s) (ΔE*) (˚C) (˚C) 50 15 4.5 3 7.751 60 15 4.5 3 7.103 70 15 4.5 3 3.948 Effect of dew point temperature on total color change 60 15 4.5 3 7.103 60 25 4.5 3 6.869 60 35 4.5 3 5.701 Effect of air velocity on the total color change Effect of thickness on the total color change 50 50 50 15 15 15 3.0 4.5 6.0 3 3 3 8.879 7.751 2.897 60 60 60 15 15 15 4.5 4.5 4.5 3 5 7 7.103 4.968 1.968 5.3.6 Summary of effects of processing parameters Table 5.10 show the summary of effects of processing parameters: drying air temperature, dew point temperature, air velocity and thickness on drying time, drying rate constant (k), effective diffusivity, Deff, total color difference, ΔE*. The desirable responses are: shortest drying time, highest drying rate constant, highest effective diffusivity and smallest color change. The processing parameter which fulfills all the four mentioned criteria would be considered as the best, thus the combination of those processing parameters would give the optimum processing parameters of matooke. From Table 5.10, it was observed that drying air temperature of 70˚C and air velocity of 6 m/s fulfill all the four criteria, thickness of 3 mm and dew point temperature fulfill three of the criteria. This implies that the of combination of: 70˚C, 6 m/s, 3mm and 15˚C would be the optimum processing parameters for drying matooke. Investigation Effect of air temperature (˚C) Effect of dew point temperature (˚C) Effect of air velocity (m/s) Effect of size (slice thickness) (mm) Processing Parameters (Independent Variables) Dependent Variables (Responses) Air Temperature (˚C) Dew Point Temperature (˚C) Air velocity (m/s) Thickness (mm) Time (min) Drying rate Constant, k (/h) Effective Diffusivity, Deff (m2/s) Total Color Difference, ΔE* 50 15 4.5 3 200 1.34 3.40E-10 7.75 60 15 4.5 3 115 2.10 5.32E-10 7.10 70 15 4.5 3 90 2.49 6.31E-10 3.95 60 15 4.5 3 115 2.10 5.32E-10 7.10 60 25 4.5 3 135 1.76 4.45E-10 6.87 60 35 4.5 3 145 1.72 4.36E-10 5.70 50 15 3.0 3 420 0.59 1.50E-10 8.90 50 15 4.5 3 200 1.34 3.40E -10 7.75 50 15 6.0 3 92 1.61 4.09E-10 2.90 60 15 4.5 3 115 2.10 5.32E-10 7.10 60 15 4.5 5 250 0.96 6.75E-10 4.97 60 15 4.5 7 370 0.60 8.27E-10 1.97 Drying Characteristics of Matooke Table 5.10 Summary of effects of process parameters 75 Drying Characteristics of Matooke 76 5.3.7 Effect of drying air temperature on surface product temperature of matooke Knowledge of the product temperature during drying is important to evaluate the influence of the drying process on physicochemical and quality attributes. This is because the reactions between food components are often accelerated during drying leading to significant reduction in quality and nutritional value. Drying characteristics for matooke 86 The reaction rates are strongly affected by the temperature and moisture content5.3.7 of Effect foodofduring (Labuza, 1972). 5.21 shows the effect of drying airdrying temperature on surface productFigure temperature of matooke air temperature on surface product temperature at air velocity of 4.5 m/s dew Knowledge of the product temperature during drying is important to evaluate the influence of the point temperature 15˚C andand thickness of 3mm. drying process on of physicochemical quality attributes. This is because the reactions between food components are often accelerated during drying leading to significant reduction in quality nutritional value. The reaction rates are strongly affected by the temperature and moisture Figure and 5.21 Development of infrared surface product temperature with drying of food during (Labuza, 1972). the effect of temperature on point time atcontent a given set airdrying temperature of Figure 70˚C5.21 at shows air velocity ofair4.5 m/s dew surface product temperature at air velocity of 4.5 m/s dew point temperature of 15˚C and temperature of 15˚C and thickness of 3mm. thickness of 3mm. 8 Temperature (˚C) 60 6 Air Temp Product Temp 40 4 MC 20 2 0 Moisture Content (kg water/kg dry matter) 80 0 - 50 Time (min) 100 150 Figure 5.21 Development of infrared surface product temperature with drying time at a given set It was observed 5.21, that was a direct relationship air temperaturefrom of 70˚CFigure at air velocity of 4.5 m/sthere dew point temperature of 15˚C and thicknessbetween of surface3mm. product temperature and air temperature at any given drying time, in that the surface product temperature increased with increase in drying air temperature. At the onset of the drying process, the surface product It was observed from Figure 5.21, that there was a direct relationship between surface product temperature of and matooke was atmuch lower the air temperature due to the temperature air temperature any given dryingthan time, in thatset the surface product temperature increased with increase in drying airAs temperature. At the onset of the drying process, the effect of evaporative cooling. drying continued, the moisture at surface the surface product temperature of matooke was much lower than the set air temperature due to the effect of decreased and the internal resistance to moisture transport increased, thus evaporative cooling. As drying continued, the moisture at the surface decreased and the internal the evaporation zone transport movedincreased, from the intozone themoved material. resistance to moisture thus surface the evaporation from theThis surfaceimplied the material. This implied that the heat necessary had to be further that theintoheat necessary for evaporation had to for beevaporation further transferred from the transferred from the surface into the material to evaporate the moisture, in order to accomplish Drying Characteristics of Matooke 77 surface into the material to evaporate the moisture, in order to accomplish the phenomena of drying as being a simultaneously heat and mass transfer process. This temperature gradient was required for heat and mass transfer to take place. As the moisture content of product continued to decrease, the product temperature also increased until the set air temperature was reached. This implied that when the product temperature became relatively constant, the samples were relatively dry suggesting that it could have attained the equilibrium moisture content at the set air temperature. It was noted that, the gelatinization temperature of matooke is 72˚C, glass transition temperature is approximately less than67˚C, (Muranga 1998), however, from the results it was observed that the surface product temperature at the end of drying, on average was approximately less than 67˚C, implying that the glass transition temperature was not exceeded. 5.3.8 Drying Effect of surface product temperature on the drying characteristics for matooke 88 time of matooke Effect of surface temperature on the product drying time temperature of matooke Figure 5.3.8 5.22 shows theproduct effect of surface on the drying temperature. Figure 5.22 shows the effect of surface product temperature on the drying temperature. 8 Air Temp Temperature (˚C) 80 Product Temp 6 MC (db) 60 4 40 2 20 0 Moisture Content (kg water/kg dry matter) 100 0 - 50 100 Time (min) 150 200 Figure 5.22 Development of air temperature with drying time for a given set infrared surface product temperature of 50˚C at air velocity of 4.5 m/s dew point temperature of 15˚C and Figure 5.22 Development of air temperature with drying time for a given set thickness of 3 mm. infrared surface product temperature of 50˚C at air velocity of 4.5 m/s dew point temperature of 15˚C and thickness of 3 mm. It was observed from Figure 5.22, that the air temperature shot up during the first region of drying, It was when observed from Figurereached 5.22, the that air temperature shotstarted up during the product temperature set the temperature, the air temperature decreasing tending to the set product temperature. The drying air temperature becomes the first region of drying, when the product temperature reached the set relatively constant but slightly above the set product temperature, because the temperature gradient required for continuation of the drying process. The first region when the air temperature shot could be attributed to the high moisture content of 78 Drying Characteristics of Matooke temperature, the air temperature started decreasing tending to the set product temperature. The drying air temperature becomes relatively constant but slightly above the set product temperature, because the temperature gradient required for continuation of the drying process. The first region when the air temperature shot could be attributed to the high moisture content of the product that it required high energy to evaporate the moisture from the product and also to overcome the effect of evaporative cooling effect. The effect of setting the product temperature, reduced the drying time by half. Comparing the results of drying when air temperature was set at 50˚C, and when the product temperature was set at 50˚C, with other processing parameters kept constant at air velocity 4.5 m/s, dew point temperature 15˚C and thickness 3mm. The drying time for set air temperature was approximately 200 min, while that for the set product temperature was approximately 100 min. This indicated that by setting the product temperature, it reduced the drying time by 50%; in other words, setting the product temperature the drying process automatically turned itself to stepwise (stepdown) drying process. This was similar to step-wise drying reported by Chua et al., 2001, which significantly reduced the drying time to reach the desired moisture content with improved product color. Drying Characteristics of Matooke 5.4 79 Conclusions and Recommendations. Conclusions This study has bridged the gap of lack of information on drying behavior of matooke. Both thickness and dew point temperature showed a direct relationship with drying time, while the air temperature and air velocity showed an inverse relationship with drying time. The effect of size (thickness) on drying time showed that the drying time required reaching the recommended microbiologically shelf-stable product from the moisture sorption isotherm studies of 0.1 (kg water/kg dry matter) at air temperature of 60˚C, dew point temperature of 18˚C and air velocity of 4.5 m/s were: 115, 250 and 375 min for 3, 5 and 7 mm respectively. It was recommended that for single-layer drying, the slice thickness = 3 mm was suitable for drying of matooke. However, from practical point of view, by considering the difference in loading capacities between single layer drying (2.33 kg/m2 taking 1.92h to dry) and multi-layer drying (6.78 kg/m² taking 5.07 h to dry), and the number of batches per day (12 working hours). It is more practical to run 2 batches (multi-layer drying) than 6 batches (single-layer drying) in a 12h. The effect of drying air temperature on drying time at dew point temperature of 18˚C, thickness of 3 mm and air velocity of 4.5 m/s were, the drying time was: 200, 115 and 90 min for 50, 60 and 70oC respectively. The effect of dew point temperature on drying time at air temperature of 60˚C, thickness of 3mm and air velocity of 4.5 m/s, the drying time was: 115, 135 and 140 min for 15, 25 and 35˚C respectively. The effect of air velocity on drying time at air temperature of 50˚C, thickness of 3mm and dew point temperature of 18˚C, the drying time was: 420, 200 and 92 min for 3, 4.5 and 6 m/s respectively. In contrast to Krokida et al., (2000), air velocity also has a significant effect on drying characteristics of matooke. The drying of matooke takes place in three steps: the warm-up and the two falling rate periods; the first falling rate and second falling rate. The drying rate constant was inversely related to thickness and dew point temperature, while it was directly related to air velocity and air temperature. The drying rate constant for all processing parameters ranged from 0.6 – 2.1/h. The effective diffusivity for all processing parameters ranged from were in the range of 80 Drying Characteristics of Matooke directly related to thickness, air velocity and air temperature while it was inversely related to dew point temperature. The effective diffusivity ranged from 1.43E-10 - 3.86E-10 m2/s. The activation energy (Ea) has been reported to be the minimum energy that must be supplied to break water-solid and/or water-water interactions and to move the water molecule from one point to another in solid (Demirel and Turhan, 2003). The results show that activation energy (Ea) for matooke was 16.3kJ/mol. (1,605kJ/kg). Comparing the activation energy (Ea) with the net isosteric heat of desorption (qst), Ea (1,605kJ/kg) was higher than qst (1,298) at 0.1 (kg water/kg dry matter) (the moisture content used in the drying analyses). This suggests that moisture molecules travels effectively in liquid form in matooke slice during the drying process. The total color difference (ΔE*) between the fresh and dry samples, showed an inverse relationship between the effects of processing parameters on the total color difference (ΔE*), lowest was for the effect of thickness at 7 mm, but it took the longest time to dry. This was followed by effect of air velocity at 6 m/s, then effect of drying air temperature at 70˚C. In the drying system controlled by setting the drying air temperature, product temperature increased with drying time, tending to the set drying air temperature. However, the product temperature did not exactly equal to the drying air temperature because of temperature gradient was required for heat and mass transfer to take place. Whereas, in the drying system controlled by setting product temperature, the drying air temperature shot up to high as possible (Figure 5.23). The drying system controlled by setting product temperature, the drying time was reduced by 50% of that obtained for setting air drying temperature. This implied that the drying system controlled by product temperature, automatically turned itself into stepwise (step-down) drying system. Recommendations Considering all the effect of processing parameters on drying characteristics of matooke showed that the of combination of drying air temperature 70˚C, air velocity 6 m/s, thickness 3mm and dew point temperature 15˚C would be the optimum processing parameters for drying matooke. Drying Characteristics of Matooke 81 For commercial processing of matooke to employ multi-layer drying of loading capacity equal or less than 7 kg/m², drying air temperature 70˚C, air velocity 6 m/s, thickness 3mm and dew point temperature 15˚C More investigation is required to develop a drying system which is controlled by product temperature instead of drying air temperature. Further investigation on the effect of surface product temperature in association with case hardening should be carried out. Optimization of processing parameters for Matooke 82 Chapter Six Objective: Optimization of processing parameters for Matooke 6.1 Background Drying is a major food processing operation and is classified as simultaneous heat and mass transfer process (Madamba et al., 1994). Drying comprises a series of processes which contribute to the final product and have significant impact on the physicochemical and quality attributes (characteristics) of the dried product (Krokida et al., 2000). During drying, in addition to moisture removal, there are many physicochemical reactions which occur and can lead to various types of quality deterioration (Krokida et al., 2000). The physicochemical and quality attributes considered in this investigation which could be affected during the drying process and are critical to the functionality of flour were: starch content and pasting properties (final and peak viscosity). The final viscosity and peak viscosity are correlated to final sample quality (Kolawole et al., 2012; Iglett et al. 2012; Osungbaro et al., 2010; Ikegwu et al., 2009; Niba et al. 2002). In the production of dried foodstuffs, minimizing drying cost is the most critical factor. However, the conditions which produce minimum costs are likely not to give the desired quality characteristics. Sturm et al., (2012) reported that many industrial drying processes were found experimentally, but their validity has not been evaluated since. Optimization of the drying operation leads to an improvement in the quality of the dried product, a reduction in the cost of processing as well as optimizing the throughput (Mudamba et al., 1994). This is because the design and optimization of dryers used for food crops is dependent on the thermal and physical properties of the specific crops (Rasouli et al., 2011). Therefore, there was need for optimizing the process parameters with reference to the physicochemical and quality characteristics of matooke. It has been reported by previous researchers that drying temperature, relative humidity and velocity of drying air have significant effect on the drying kinetics and quality of the dried product (Baini & Langrish, 2007; Sturm et al., 2012; Islam et al., 2012). However, their effect Optimization of processing parameters for Matooke 83 on physicochemical and quality characteristics of dried matooke needed to be established. Banana flour in general has been reported to have good prospects and health benefits, but has not been fully utilized, could be due to high costs associated with its conversion to flour and limited research highlighting the potential and beneficial properties (Islam et al., 2012). Therefore, the objective of this study was to optimize the processing parameters in relation to the physicochemical properties and quality of the matooke and the following were the specific objectives: • To establish the effect of the process parameters on individual responses (drying time, final viscosity and starch content). • To optimize the process parameters with respect to drying time final viscosity and starch content • To validate the optimum process parameters for drying. 6.2 Materials and Methods The materials and methods employed for this study were as explained in section 5.2, and the samples used for the determination of starch content and pasting properties were kept at 4˚C. The dry matooke was milled to flour using the starch mill as explained in chapter 3. The final viscosity, peak viscosity and starch content were used for the optimization process. Response Surface Methodology (RSM) techniques were employed for analysis using Design-Expert version 8, software (Stat-Ease, Inc., Minneapolis, USA). All experiments were carried out in triplicates. Independent Variables (Factors) Drying Temperature (50, 60 and 70˚C) Dew-point Temperature (15, 25 and 35˚C) Air velocity (3, 4.5, 6 m/s) Dependent Variables (Responses) Drying time to reach moisture content of 10% (wb) Final Viscosity (RVU) Peak Viscosity (RVU) Starch Content (%) 84 Optimization of processing parameters for Matooke The data analysis employed the Response Surface Methodology (RSM) which is one of the experimental designing methods which can surmount the limitations of conventional methods collectively and has an advantage that it reduces the number of experimental trials needed to evaluate multiple parameters and their interactions (Noordin et al., 2004; Sturm et al., 2012; Ramkrishna & Susmita, 2012).The Design-Expert version 8 software was used and employed Response Surface Methodology (RSM), to generate the design and optimize the drying conditions for matooke. The generated experimental design was considered to be efficiently applicable when the G Efficiency was greater than 60% (Myers, 2009). Analysis of variance (ANOVA) was employed to determine the significance of the relationship between the independent and dependent variables, for which was judged by the coefficient of determination (R-squared and adjusted R-squared) and p-value (p<0.05). The model and its coefficients were considered significant when the R-squared and adjusted R-squared were greater than 0.900 and p-value (p<0.05) which indicated that the model was statistically significant at α-level of 0.05. The model was valid when the lack of fit was not significant (p>0.05) which indicated a good fit, a low coefficient of variance (C.V.% <10) which indicated that the experiments performed were highly reliable, and adequate precision (Adeq Precision > 4) which indicated that the model as fitted was adequate for predicting (Noordin et al., 2004; Ramkrishna & Susmita, 2012). 6.3 Results and Discussion 6.3.1 Establishing the effect of the process parameters on individual responses The design of the experiment for establishing the effect of the process parameters on individual responses (Noordin et al., 2004) was generated by D-Optimality shown in Table 6.1. Optimization of processing parameters for Matooke 85 Table 6.1 Experimental design Factors Responses Run Air Temperature (˚C) Dew Point Temperature (˚C) Time (min) Final Peak Starch Viscosity Viscosity Content (RVU) (RVU) (% db) 1 60 25 134 250.13 388.83 82.92 2 60 35 136 264.08 394.81 82.81 3 50 15 200 258.71 394.82 83.29 4 50 35 158 258.67 393.31 83.30 5 60 35 130 261.88 392.57 82.21 6 70 35 128 278.08 411.42 84.68 7 70 15 90 266.83 408.92 83.73 8 50 35 157 258.67 410.17 82.34 9 60 25 138 263.22 408.92 83.81 10 50 15 172 242.30 409.54 81.13 11 70 15 84 273.71 409.23 83.22 12 60 15 104 265.08 409.39 83.23 13 70 25 125 264.13 409.31 83.62 14 50 15 175 257.58 407.42 82.46 15 70 35 127 269.42 412.55 82.01 16 50 25 210 303.25 408.36 82.12 The G Efficiency = 86% indicated that the experimental design was efficient for analysis. The experiments for establishing the effect of processing parameters on the individual responses were carried out at the same air velocity of 4.5 m/s, thickness of 3mm as recommended from the drying characteristics study (Chapter 3) and using matooke at harvest maturity age of 17 weeks as recommended optimum maturity window (Chapter 3). Table 6.1 shows the results of the dependent variable (drying time to reach moisture content of 0.10 (kg water/kg dry matter) which were obtained for the above set conditions. From the ANOVA analysis (Table 6.2), it was observed that the model and model terms were significant implying that the process parameters had a significant effect on the drying time. Optimization of processing parameters for Matooke 86 Table 6.2 ANOVA for Response Surface Quadratic Model analysis of variance table for drying time Source Model x1-Air Temp x2-Dew P.Temp x1x2 x12 x22 Residual Lack of Fit Pure Error Cor Total R-Squared Adj R-Squared C.V. % Sum of Squares 17863.00 12085.12 548.12 2325.38 1516.79 1489.31 1076.00 558.33 517.67 18939.00 df 5 1 1 1 1 1 10 3 7 15 Mean Square F Value p-value Prob > F 3572.60 12085.12 548.12 2325.38 1516.79 1489.31 107.60 186.11 73.95 33.20 112.32 5.09 21.61 14.10 13.84 < 0.0001 Significant < 0.0001 0.0476 0.0009 0.0038 0.0040 not significant 2.52 0.1419 0.9432 0.9148 7.32 The model for drying time was significant (R2 = 0.943 and p<0.0001). Lack of fit was not significant (p>0.1) indicated a good fit that the model was valid. A low coefficient of variance (C.V.% = 7.32) indicated that the experiments performed were highly reliable. Adequate precision (Adeq Precision = 17.82 > 4) indicated that the model as fitted was adequate for predicting. The results of the dependent variables (final viscosity) were obtained for the above set of processing parameters given in Table 6.1. From ANOVA analysis (Table 6.3), it was observed that the model and model terms were not significant implying that the processing parameters had no effect on the final viscosity. Optimization of processing parameters for Matooke 87 Table 6.3: ANOVA for Response Surface Quadratic Model analysis of variance table for final viscosiy (pasting properties) Source Model x1-Air Temp x2-Dew P.Temp x1x2 Residual Lack of Fit Pure Error Cor Total R-Squared Adeq Precision C.V. % Sum of Squares df Mean Square 190.17 108.18 40.09 15.07 2467.81 2150.41 317.40 2657.98 3 1 1 1 12 5 7 15 63.39 108.18 40.09 15.07 205.65 430.08 F Value p-value Prob > F 0.31 0.53 0.19 0.073 0.8190 0.4822 0.6667 0.7912 not significant 9.49 0.0051 significant 45.34 0.0715 1.398 5.42 The model for final viscosity was not significant (R2 = 0.0715 and p>0.05). The lack of fit was significant (p<0.01), implying that the model was not valid. Adequate precision (Adeq Precision = 1.4 less than 4) indicated that the model as fitted was not adequate for predicting. However, a low coefficient of variance (C.V.% = 5.42) indicated that the experiments were highly reliable. It also showed that all model terms: air temperature and dew point temperature, were not significant since p-value were greater than 0.05 (p<0.05). This could be attributed to the fact that the set temperature were below the gelatinization temperature of 72˚C (Muranga, 1998). This was an indication that the processing parameters did not have a significant effect on final viscosity (pasting properties). The results of the dependent variables (peak viscosity) were obtained for the above set of processing parameters given in Table 6.1. From ANOVA analysis (Table 6.4), it was observed that the model and model terms were not significant implying that the processing parameters had no effect on the peak viscosity. Optimization of processing parameters for Matooke 88 Table 6.4: ANOVA for Response Surface Quadratic Model analysis of variance table for peak viscosity (pasting properties) Source Sum of Squares df Mean Square F Value p-value Prob > F Model 363.20 5 72.64 1.03 0.4502 x1-Air Temp 137.85 1 137.85 1.96 0.1922 x2-Dew P. Temp 21.86 1 21.86 0.31 0.5899 x1 5.12 1 5.12 0.07 0.7930 x22 191.66 1 191.66 2.72 0.1302 0.07 0.7993 0.74 0.5589 2 x1x2 4.81 1 4.81 Residual 704.89 10 70.49 Lack of Fit 170.47 3 56.82 Pure Error 53.41 7 76.43 1068.08 15 Cor Total R-Squared Adj R-Squared CV % not significant not significant 0.340 0.010 2.08 The model for final viscosity was not significant (R2 = 0.34 and p>0.05). The lack of fit was significant (p<0.01), implying that the model was not valid. The low coefficient of variance (C.V.% = 2.08) indicated that the experiments were highly reliable. It also showed that all model terms: air temperature and dew point temperature, were not significant since p-value were greater than 0.05 (p<0.05). This was also an indication that the processing parameters did not have a significant effect on peak viscosity (pasting properties). Table 6.1 shows the results of the starch content which were obtained for the set of processing parameters. From ANOVA analysis (Table 6.5), it was observed that the model and model terms were not significant implying that the process parameters had no effect on the starch content, since the set temperatures were below the gelatinization temperature of 72˚C. Optimization of processing parameters for Matooke 89 Table 6.5 ANOVA for Response Surface Quadratic Model analysis of variance table for starch content. Source Model x1-Air Temp x2-Dew P. Temp Residual Lack of Fit Pure Error Cor Total Sum of Squares 2.82 2.81 3.358E-3 8.42 1.31 7.10 11.23 R-Squared C.V. % Adeq Precision 0.2507 0.97 3.016 df 2 1 1 13 6 7 15 Mean F p-value Square Value Prob > F 1.41 2.17 0.1532 not significant 2.81 4.34 0.0576 3.358E-3 5.186E-3 0.9437 0.65 0.22 1.01 0.22 0.9596 not significant The model for starch content was not significant (R2 = 0.251 and p>0.05). Also the model terms were not significant, which indicated that the processing parameters had no significant impact on starch content. Adequate precision (Adeq Precision = 1.4 less than 4) indicated that the model as fitted was not adequate for predicting. However, the lack of fit was significant (Prob>F) implying that the model was valid although it was not significant. A low coefficient of variance (C.V.% = 0.97) indicated that the experiments were highly reliable. Considering Tables 6.3, 6.4 and 6.5 showing the effect of processing parameters on physicochemical properties of matooke flour, pasting properties (final and peak viscosities) and starch content. It was observed that the processing parameters had no significant effect on the physicochemical properties of matooke flour. This could be attributed to the fact that the drying air temperatures were below the gelatinization temperature of 72˚C (Muranga 1998). 6.3.2 Optimization of process Parameters The D-Optimal experimental design used in drying process optimization is shown in Table 6.6.The independent variables were: air temperature, dew point temperature and airflow. The dependent variables (responses) were Optimization of processing parameters for Matooke 90 drying time to reach moisture content of 0.1 (kg water/kg dry matter). The objective for the optimization was therefore to minimize the drying time and maximize starch content and peak viscosity. The G Efficiency = 62% indicated that the experimental design was efficient for analysis. Table 6.6 Drying time for different set drying conditions of matooke Run Air Temperature (˚C) Dew Point Temperature (˚C) Air velocity (m/s) Time (min) 1 50 25 6 168.75 2 70 35 3 250.50 3 50 35 6 212.00 4 60 35 3 315.00 5 70 35 6 103.50 6 50 35 3 401.00 7 50 15 6 117.00 8 70 15 3 205.00 9 60 25 6 117.30 10 50 35 6 260.00 11 50 25 3 365.00 12 50 15 4.5 204.67 13 70 35 6 129.67 14 70 15 6 83.30 15 70 15 3 205.00 16 50 15 3 328.00 17 70 15 6 80.00 18 60 15 4.5 154.00 19 70 25 4.5 125.00 20 70 25 4.5 125.00 The criteria for optimization was to minimize drying time and keeping in range the drying air temperature and air velocity was used to determine the suitable processing parameters. The ANOVA analysis (Table 6.7), showed that the model for drying time is significant (R2 = 0.9876 and p<0.0001). The lack of fit was not significant Optimization of processing parameters for Matooke 91 (p>0.1) indicated a good fit and that the model was valid. A low coefficient of variance (C.V.% = 7.43) indicated that the experiments were highly reliable. Adequate precision (Adeq Precision = 32.44 greater than 4) indicated that the model as fitted was adequate for predicting. It also showed that all model terms were significant since p-value was less than 0.05 (p<0.05). Table 6.7. ANOVA for Response Surface Quadratic Model analysis of variance table Source Sum of Squares Model x1-Air Temp x2-Dew P. Temp x3-Air Flow x1x2 x1x3 x2x3 x12 x22 x32 Residual Lack of Fit Pure Error Cor Total 1.717E+005 51221.41 14786.18 96196.24 2616.68 3272.26 150.50 227.44 359.88 1359.28 2153.95 654.07 1499.88 1.738E+005 C.V. % R-Squared Adj R-Squared Pred R-Squared Adeq Precision 7.43 0.988 0.977 0.952 32.436 Mean Square df 9 1 1 1 1 1 1 1 1 1 10 5 5 19 F Value 19074.32 88.55 51221.41 237.80 14786.18 68.65 96196.24 446.60 2616.68 12.15 3272.26 15.19 150.50 0.70 227.44 1.06 359.88 1.67 1359.28 6.31 215.40 130.81 0.44 299.98 p-value Prob > F < 0.0001 Significant < 0.0001 < 0.0001 < 0.0001 0.0059 0.0030 0.4227 0.3284 0.2252 0.0308 0.8082 Not significant Final Model Equation in terms of actual factors: Y=1381.36-18.4*x1+5.99*x2-1.62*x3-0.15*x1*x2+7.9E-3*x1*x3+1.7e-32 2 2 *x3+0.1*x1 +0.1* x2 +5.9E-4*x3 *x 2 Where:Y= Drying time to reach moisture content of 10% (db). x1 = Drying Air Temperature (˚C) x2 = Dew Point Temperature (˚C) x3 = Air velocity(m/s) 6.1 92 Optimization of processing parameters for Matooke Figure 6.1B – 6.3B (appendix B) show the response surfaces of the effect of air velocity, drying air temperature and dew point temperature on drying time indicating the predicted time. Figure 6.4B – 6.6B (appendix B) show the response surfaces of the effect of drying air temperature and dew point temperature on final viscosity, peak viscosity and starch content indicating the predicted values respectively. 6.3.3 Optimum processing parameters for drying matooke Therefore, optimization showed that optimum processing parameters for matooke were: Drying Air Temperature = 69.58˚C Dew Point Temperature = 18˚C Air velocity = 6 m/s (overflow mode) Thickness = 3 mm (as recommended in Chapter 3) The drying air temperature can be rounded to 70˚C and air velocity to 6m/s. Considering the quality attribute (total color change), there was minimum color change with air velocity of 6m/s. The product temperature development during drying for both drying processes: controlled by drying air temperature and another one controlled by product temperature does not reach the drying air temperature because of the temperature gradient required for continued drying (Chapter 5). It was concluded that, 70˚C was the optimum drying air temperature for drying matooke as given in Table 6.8. Table 6.8: Optimum drying conditions/parameters for matooke (Musa sp, AAA-AEHB) Parameter Optimum Conditions Drying air temperature (˚C) 70 Dew point temperature (˚C) 18 Air velocity (m/s) 6 Optimization of processing parameters for Matooke 93 6.3.4 Validation of the optimum processing parameters for matooke Table 6.9 shows the comparison between the predicted and actual experimental values. The validation of the optimum drying conditions investigated by applying the following optimum drying conditions: thickness 3 mm, drying air temperature 70˚C, dew point temperature 18˚C and air velocity 6 m/s. This confirms that those optimum conditions predicted by the model are the best for the drying of matooke. Table 6.9. Comparison between the predicted values and actual values. Response Actual Predicted Time (min) 84.08 77.80 Final viscosity (RVU) 243.21 272.34 Peak viscosity (RVU) 396.17 411.12 Starch content (% db) 83.46 83.65 The optimum drying air temperature of 70˚C is greater than the recommended drying air temperature by Muranga, (1998) of 50˚C for drying matooke and yet the results of starch content (%db) are within the same range. Also the results for pasting properties are within the same range as reported by Muranga et al., 2010 (Table 10). In addition, the results were in line with Islam et al., (2012), who reported that appropriate conditions of dehydration enhance the drying performance and drying air temperature of 65˚C gave faster drying of green bananas without any irregular burning or stake burning. Also Leite et al., (2007) dried bananas at 60 and 70˚C, reported that the chemical composition was not affected by drying conditions. In addition, considering the results of product temperature development during the drying controlled product temperature, the drying temperature went up to 90˚C without affecting the physicochemical properties (Chapter 5). This implies that air drying temperature has no effect on the physicochemical properties as long as the product temperature is below the gelatinization temperature of matooke starch (72˚C). Optimization of processing parameters for Matooke 94 Table 6.10 Comparison of current results of this study with previous research Model Prediction Validation Exp. results Source (Muranga, et al.,2010) Drying air temperature (˚C) 70 70 50 Final viscosity (RVU) 272.34 243.21 244.7 Peak viscosity (RVU) 411.12 396.17 375.92 Starch content (%db) 83.65 83.46 ND 6.4 Conclusions and Recommendations The optimum processing parameters for single-layer drying of matooke were determined to obtain the minimum drying time, maximum starch content, final viscosity and peak viscosity. The drying air temperature and air velocity were kept within the range, the thickness was maintained at 3 mm (as recommended in Chapter 5, and the average dew point temperature of ambient at Bushenyi was used. The optimum processing parameters for single-layer drying of matooke are the same as those obtained in Chapter 5 for the analysis of drying behavior. For which the desirable responses were: shortest drying time, highest drying rate constant, highest effective diffusivity and smallest color change. The processing parameter which fulfills all the four mentioned criteria would be considered as the best, thus the combination of those processing parameters would give the optimum processing parameters of matooke. The processing parameters do not have a significant effect on the physicochemical and functional properties, and quality attribute as long as the product temperature does not exceed 72˚C (gelatinization temperature of matooke). Therefore, the optimum processing parameters for single-layer drying of matooke are: Thickness 3mm, air temperature 70˚C, dew point temperature 18˚C and air velocity 6m/s overflow mode. General Conclusions and Recommendations 95 Chapter Seven General Conclusions and Recommendations Banana is an important crop in Uganda; it is a staple food for more than 70% of Uganda’s population and contributes to about 42% of household income in rural areas. The most common type of banana is locally called matooke, Musa sp triploid acuminate genome group (AAA-EAHB). The study was carried out using a research dryer in the laboratory employing the overflow mode. The results can be applied directly for commercial processing of bananas (matooke), nevertheless, verification for online processing plant is recommended. This study has bridged the gap of lack of accurate information on the harvest maturity window for commercial processing of matooke into flour. The physiological maturity age of matooke (mbwazirume cv. Musa sp. AAA EAHB) at Bushenyi located at the Latitude: 0.59°S, Longitude: 30.21° E and Altitude 1570 m is 21 weeks. The optimum harvest maturity window for commercial processing of matooke at Bushenyi is between 15-21 weeks. Matooke harvested within the above optimum harvest maturity window can be used for processing raw matooke flour referred to as Raw Tooke Flour (RTF) of a consistent quality on the market for use in food and non food industries. In case of natural disaster like storm, matooke from the age of 12-15 weeks can be processed into matooke flour since at that age it already has starch content above 80% (db). The matooke from maturity age between 3 – 12 weeks can be processed for flour for people who require less starch in their diet. In addition, matooke harvested beyond the above maturity window after 21 weeks is recommended for processing Instant Tooke Flour (ITF). This is because beyond the recommended harvest maturity window, it would be in the initial stages of starch loss which plays a major role in textural changes in ripening which is good for ITF. 96 General Conclusions and Recommendations The finger weight model is recommended for farmers to estimate harvest maturity age for matooke and the combined model of finger weight and pulp peel ratio is recommended for commercial processors. Matooke cv mbwazirume (Musa sp. AAA-EAHB) at Bushenyi. Theoretically the results of the study can be used elsewhere. Nevertheless, it is recommended that similar work be carried out in the matooke growing areas to determine the applicable models for those particular regions since development of matooke and bananas in general is affected by location (Robinson & Saúco, 2010) This study has provided information for the first time on the moisture sorption isotherm of matooke. The results of this study on moisture sorption isotherm of matooke can be used for prediction of quality, design, modeling and optimization of many processes such as drying, aeration and storage. In addition, it is also important for predicting quality changes during packaging and storage of dried matooke. Matooke isotherms exhibited type II curves behavior which is characteristic of foodstuffs as reported in literature (Johnson & Brennan, 2000; Yan et al., 2008). The equilibrium moisture content decreases with increase in temperature at a given water activity (aw), indicating that matooke becomes less hygroscopic with increase in temperatures giving a clear stability domain of matooke after drying leading to extensive shelf life. This complements Muranga, (1998), who attributed the extensive shelf life to low sugar and fat. The GAB model best described all the adsorption and desorption moisture isotherms for matooke, this was consistent with literature on bananas and plantain. The GAB monolayer has been reported to be indicative for optimum moisture content for storage conditions for hydrated foods. For the case of matooke, the GAB monolayer for adsorption isotherm at 30˚C was 0.087 (kg water/kg dry matter), which was below the recommended safe water activity level of 0.6. The water activity level of 0.6 corresponds to equilibrium moisture content of 0.11 (kg water/kg dry matter) on the adsorption isotherm at 30˚C. Therefore, for commercial processing of matooke in order to obtain a microbiologically shelf-stable dry matooke, it is recommended to dry matooke to moisture content below or equal to 11 (kg water/kg dry matter) equivalent to 10% (wb). General Conclusions and Recommendations 97 The hysteresis phenomenon was exhibited by the moisture sorption isotherms for matooke. It was reported that in food preservation and storage technology, the microbiological stability of stored food item depend more on water activity than total moisture content (Caurie, 2007). Therefore, hysteresis could be used as a food quality index. It is therefore recommended that further investigations on hysteresis of matooke at different temperatures be carried out to establish the optimum range of temperatures safe storage, since increased hysteresis being indicative of reduced stability and reduced hysteresis being indicative of improved stability of stored dry matooke. The isoteric heat of sorption for both adsorptions increased with decreased moisture content, however, desorption was higher than the adsorption and this was in line with literature on bananas and plantains (Ciro et al., 2008; Johnson & Brennan, 2000). It was reported that isosetric heat of sorption for adsorption isotherm is a measure for the energy released during adsorption and that for desorption as the energy required for breaking the intermolecular forces between the molecules of water vapor and the surface of adsorbent. Thus the ioseteric heat of sorption is considered as indicative of the intermolecular attractive forces between the sorption sites and the water vapor. The total isosteric heat of sorption for adsorption isotherm for matooke ranged from 4,586 – 2,386.39 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (db) respectively. While the total isosteric heat of sorption for desorption isotherm for matooke ranged from 18,194 – 2,391 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (db) respectively. This study has revealed that the minimum energy required for drying matooke from 3.5 (kg water/kg dry matter) (80%wb) to 0.11 (kg water/kg dry matter) (10%wb) is equal to 8,124 kJ/kg of water removed. Implying that the minimum energy required for drying of 1 kg of fresh matooke from 80 - 10(% wb) is 5,793 kJ. This study has bridged the gap of lack of information on drying behavior of matooke. Both thickness and dew point temperature showed a direct relationship with drying time, while the air temperature and air velocity showed an inverse relationship with drying time to reach the desired moisture content. The effect of processing parameters on the drying characteristics showed that air velocity rate, drying air temperature and thickness had a significant effect, 98 General Conclusions and Recommendations while dew point did not have significant effect on the drying characteristics of matooke. The drying of matooke takes place in three steps: the warm-up and the two falling rate periods; the first falling rate and second falling rate. This was in line with the literature on bananas and plantains (Prachayawarakon et al., 2008). For all processing parameters, the drying rate constant ranged from 0.59 – 2.49 /h, and effective diffusivity ranged from 1.5E-10 – 8.27E-10 m2/s. The activation energy (Ea) has been reported to be the minimum energy that must be supplied to break water-solid and/or water-water interactions and to move the water molecule from one point to another in solid. The results show that activation energy (Ea) for matooke was 16.3kJ/mol (1,605kJ/kg). Comparing the activation energy (Ea) with the net isosteric heat of desorption (qst), Ea (1,605kJ/kg) was higher than qst (1,297.62) at 0.1 (kg water/kg dry matter) (the moisture content used in the drying analyses). This suggests that moisture molecules travel in liquid form in matooke slices as reported in literature (Demirel & Turhan, 2003). The total color difference (ΔE*) between the fresh and dry samples, showed an inverse relationship between the effects of processing parameters on the total color difference (ΔE*). The total color difference (ΔE*), was lowest for effect of thickness was 7 mm, but it took the longest time to dry. It was followed by the effect of air velocity which was 6 m/s, and lastly for the effect of drying air temperature which was 70˚C. In the drying system controlled by setting the drying air temperature, product temperature increased with drying time, tending to the set drying air temperature. However, the product temperature did not exactly equal to the drying air temperature. Whereas, in the drying system controlled by setting product temperature, the drying air temperature shot up to high as possible, for example when the product temperature was set at 50˚C, the air drying temperature shot up to 90˚C and decreased tending to the set product temperature with increase in drying time. In this case also, product temperature became constant and nearly equal to air temperature which was an indication that the samples had dried and attained the equilibrium moisture content at the set air temperature. The drying system controlled by setting product temperature, the drying time was reduced by 50% of setting air General Conclusions and Recommendations 99 drying temperature. This implied that the drying system controlled by product temperature, automatically turned itself into stepwise (step-down) drying system. It was therefore recommended to carry out further investigation to develop a drying system which is controlled by product temperature instead of drying air temperature. Further investigation on the effect of surface product temperature in association with case hardening should be carried out. The processing parameters did not have a significant effect on the physicochemical (starch content) and functional properties (peak and final viscosity) of matooke. The optimum drying air temperature of 70˚C is greater than the recommended drying air temperature by Muranga, (1998) of 50˚C for drying matooke and yet the results of starch content (%db) are within the same range. This implies that drying matooke using air drying temperature of 70˚C gives the same product at a reduced drying time and hence the most appropriate. Considering the development of product temperature during drying, (drying systems, one controlled by setting drying air temperature and one controlled setting the product temperature) and the fact that the processing parameters do not have influence on the physicochemical and functional properties of matooke. This suggests that any drying air temperature can be used for drying as long as the product temperature does not exceed 68˚C (the onset of gelatinization for matooke). Therefore, the optimum processing parameters for single-layer drying of matooke are: thickness = 3mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6 m/s overflow mode. However, from practical point of view, by considering the difference in loading capacities between single layer drying (2.33 kg/m2 taking 1.92h to dry) and multi-layer drying (6.78 kg/m² taking 5.07 h to dry), and the number of batches per day (12 working hours). It is more practical to run 2 batches (multi-layer drying) than 6 batches (single-layer drying) in a 12h. It is therefore recommended for commercial processing of matooke to employ multi-layer drying of loading capacity equal or less than 7 kg/m², thickness 3 mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6m/s overflow mode. General Conclusions and Recommendations 100 7.1 Recommended Procedure for Commercial Processing Matooke Flour • Matooke located at the Latitude: 0.59°S, Longitude: 30.21° E and Altitude: 1570 m. should be harvested at maturity age between 15 – 21 weeks. • The finger weight model was recommended for farmers to estimate harvest maturity for matooke and the combined model of finger weight and pulp peel ratio is recommended for industrial use. • Recommended Standard Procedure for Commercial Processing Matooke Flour (Raw Tooke Flour) (Modification of the Patent No. AP/P/2005/003308 as follows: Multi-layer drying: 7.2 – De-finger the fingers from the bunch – Peel the fingers – Pre-treat with 1% Sodium Meta-bisulphate solution for 3 – 5 minutes – Slice thickness should be equal to 3 mm – Pre-treatment in 1% Sodium Meta-bisulphate for 3- 5 minutes – Drain the slices – Load the trays and load the trays in the dryer – Drying air temperature between 65 - 70˚C – Dew point temperature of ambient conditions between 16 – 22˚C – Air velocity rate equal to 6 m/s – Loading capacity should not exceed 7 kg/m2. – Drying time equal to 6 h ± 20 min Recommendations for future work • The maturity experiments can be carried out for different cultivars and for other climatic zones to determine the respective optimum maturity window. • Similar studies are carried out for different cultivars and in other matooke growing areas to validate the developed models and/or improve them for those particular regions since development of matooke and bananas in general is affected by location. General Conclusions and Recommendations 101 • There is need for research to develop electronic or magnetic sensors/units which can be used to estimate maturity age by employing the developed models for estimating the harvest maturity window. • The hysteresis phenomenon was exhibited and since it could be used as a food quality index, it is recommended that further investigations of hysteresis phenomenon at different temperatures to establish the ranges of increased hysteresis being indicative of reduced stability and reduced hysteresis being indicative of improved stability of stored product. • There is need for more work to be carried out to optimize energy requirement for the above recommended parameters. • Research on developing a drying system which can be operated by setting the product temperature instead of the conventional method of stepwise drying. • Research on developing equipment for non destructive determination of maturity age of matooke using the combined maturity model (finger weight combined with pulp/peel ratio. • Further investigation on the effect of surface product temperature in association with case hardening should be carried out. 102 Summary Summary Cooking banana is one of the most important crops in Uganda; it is a staple food and source of household income in rural areas. The most common cooking banana is locally called matooke, a Musa sp triploid acuminate genome group (AAA-EAHB). It is perishable and traded in fresh form leading to very high postharvest losses (22-45%). This is attributed to: non-uniform level of harvest maturity, poor handling, bulk transportation and lack of value addition/processing technologies, which are currently the main challenges for trade and export, and diversified utilization of matooke. Drying is one of the oldest technologies employed in processing of agricultural produce. A lot of research has been carried out on drying of fruits and vegetables, but little information is available on matooke. Drying of matooke and milling it to flour extends its shelf-life is an important means to overcome the above challenges. Raw matooke flour is a generic flour developed to improve shelf stability of the fruit and to find alternative uses. It is rich in starch (80 - 85%db) and subsequently has a high potential as a calorie resource base. It possesses good properties for both food and non-food industrial use. Some effort has been done to commercialize the processing of matooke but there is still limited information on its processing into flour. It was imperative to carry out an indepth study to bridge the following gaps: lack of accurate information on the maturity window within which matooke for processing into flour can be harvested leading to non-uniform quality of matooke flour; there is no information on moisture sorption isotherm for matooke from which the minimum equilibrium moisture content in relation to temperature and relative humidity is obtainable, below which the dry matooke would be microbiologically shelf-stable; and lack of information on drying behavior of matooke and standardized processing parameters for matooke in relation to physicochemical properties of the flour. The main objective of the study was to establish the optimum harvest maturity window and optimize the processing parameters for obtaining standardized microbiologically shelf-stable matooke flour with good starch quality attributes. This research was designed to: i) establish the optimum maturity harvest window within which matooke can be harvested to produce a consistent quality of matooke flour, ii) establish the sorption isotherms for matooke, iii) establish Summary 103 the effect of process parameters on drying characteristics of matooke, iv) optimize the drying process parameters for matooke, v) validate the models of maturity and optimum process parameters and vi) standardize process parameters for commercial processing of matooke. Samples were obtained from a banana plantation at Presidential Initiative on Banana Industrial Development (PIBID), Technology Business Incubation Center (TBI) at Nyaruzunga – Bushenyi in Western Uganda. A completely randomized design (CRD) was employed in selecting the banana stools from which samples for the experiments were picked. The cultivar Mbwazirume which is soft cooking and commonly grown in Bushenyi was selected for the study. The static gravitation method recommended by COST 90 Project (Wolf et al., 1985), was used for determination of moisture sorption isotherms. A research dryer developed for this research. All experiments were carried out in laboratories at TBI. The physiological maturity of matooke cv. mbwazirume at Bushenyi is 21 weeks. The optimum harvest maturity window for commercial processing of matooke flour (Raw Tooke Flour - RTF) at Bushenyi is between 15-21 weeks. The finger weight model is recommended for farmers to estimate harvest maturity for matooke and the combined model of finger weight and pulp peel ratio is recommended for commercial processors. Matooke isotherms exhibited type II curve behavior which is characteristic of foodstuffs. The GAB model best described all the adsorption and desorption moisture isotherms. For commercial processing of matooke, in order to obtain a microbiologically shelf-stable dry product. It is recommended to dry it to moisture content below or equal to 10% (wb). The hysteresis phenomenon was exhibited by the moisture sorption isotherms for matooke. The isoteric heat of sorption for both adsorptions and desorption isotherms increased with decreased moisture content. The total isosteric heat of sorption for matooke: adsorption isotherm ranged from 4,586 – 2,386 kJ/kg and desorption isotherm from 18,194– 2,391 kJ/kg for equilibrium moisture content from 0.3 – 0.01 (db) respectively. The minimum energy required for drying matooke from 80 – 10% (wb) is 8,124 kJ/kg of water removed. Implying that the minimum energy required for drying of 1 kg of fresh matooke from 80 - 10% (wb) is 5,793 kJ. 104 Summary The drying of matooke takes place in three steps: the warm-up and the two falling rate periods. The drying rate constant for all processing parameters ranged from 5,793 kJ and effective diffusivity ranged from 1.5E-10 - 8.27E-10 m2/s. The activation energy (Ea) for matooke was 16.3kJ/mol (1,605 kJ/kg). Comparing the activation energy (Ea) with the net isosteric heat of sorption for desorption isotherm (qst) (1,297.62) at 0.1 (kg water/kg dry matter), indicated that Ea was higher than qst suggesting that moisture molecules travel in liquid form in matooke slices. The total color difference (ΔE*) between the fresh and dry samples, was lowest for effect of thickness of 7 mm, followed by air velocity of 6 m/s, and then drying air temperature at 70˚C. The drying system controlled by set surface product temperature, reduced the drying time by 50% compared to that of a drying system controlled by set air drying temperature. The processing parameters did not have a significant effect on physicochemical and quality attributes, suggesting that any drying air temperature can be used in the initial stages of drying as long as the product temperature does not exceed gelatinization temperature of matooke (72˚C). The optimum processing parameters for single-layer drying of matooke are: thickness = 3 mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6 m/s overflow mode. From practical point of view it is recommended that for commercial processing of matooke, to employ multi-layer drying of loading capacity equal or less than 7 kg/m², thickness 3 mm, air temperatures 70˚C, dew point temperature 18˚C and air velocity 6 m/s overflow mode. Zusammenfassung 105 Zusammenfassung Kochbananen sind eine der wichtigsten Anbauprodukte in Uganda, sie sind Grundnahrungsmittel und Einkommensquelle in ländlichen Gebieten. Die bekannteste Kochbanane, umgangssprachlich matooke genannt, ist die Musa sp triploid acuminate genome group (AAA-EAHB). Sie ist verderblich und wird im frischen Zustand gehandelt, so dass sehr hohe Nachernteverluste auftreten (22-45%). Gründe dafür sind: Ernte bei nicht einheitlichem Reifegrad, wenig schonende Aufbewahrung, Transport in großen Einheiten und vor allem findet keine Weiterverabeitung bzw. Konservierung/Trocknung statt, was die Grundlage für Handel, Export und vielfältige Nutzung von matooke wäre. Trocknung ist eine der ältesten Technologien in der landwirtschaftlichen Produktionskette. Es gibt zahlreiche Forschungsarbeiten zur Trocknung von Obst und Gemüse, aber wenig ist über die Trocknung von matooke bekannt. Trocknung von matooke und anschließendes Vermahlen diente der Wertsteigerung und –erhaltung und würde die zuvor genannten Nachteile kompensieren. Rohes matooke -Mehl wurde ursprünglich entwickelt, um die Lagerstabilität der verarbeiteten Früchte zu verbessern und alternative Nutzungen zu finden, es hat einen hohen Stärkeanteil (80 - 85% db) und ist somit eine kalorienreiche Nahrungsquelle. Es hat gute Eigenschaften sowohl für den Nahrungsmittelbereich, als auch für den non-food Bereich. Einige Bemühungen wurden unternommen, um die Weiteverarbeitung von matooke auszuweiten, aber Wenig ist über die Verarbeitung zu Mehl bekannt. Es war somit notwendig detailliertere Studien in die Wege zu leiten, um folgende Lücken zu füllen: genaue Kenntnisse über das optimale Reifefenster zur Ernte von matooke, um gleichmäßige und qualitativ hochwertiges Mehl zu erhalten; es gibt keine Informationen über des Sorptionsverhalten von matooke, aus dem das Feuchtegleichgewicht in Abhängigkeit von Temperatur und relativer Luftfeuchte zu ermitteln ist, bei dem getrocknete matooke mikrobiologisch stabil gelagert werden kann; außerdem gibt es keine Informationen über das Trocknungsverhalten von matooke sowie standarisierte Prozessparameter von matooke in Bezug auf physiochemisch Eigenschaften des Mehl. Das Hauptanliegen dieser Forschungsarbeit besteht darin, ein optimales Reifeund Erntefenster zu definieren, so wie optimale Weiterverarbeitungsparameter zu bestimmen, die ein standarisiertes mikrobiologisch einwandfreies Mehl 106 Zusammenfassung mit guten Stärke Eigenschaften garantieren. Diese Arbeit hat folgende Ziele : i) ein optimales Reifefenster zu definieren, innerhalb dessen matooke geerntet werden kann um gleichbleibendes matooke Mehl zu erhalten, ii) die Sorptionsisothermen für matooke aufzuzeigen iii) die Auswirkungen verschiedener Trocknungsvarianten von matooke aufzeigen, vi) die Trocknungsparameter optimieren v) Wertungen der verschiedenen Modelle zu Reifegrad und Verarbeitungsschritten erstellen vi) Standardkriterien für eine kommerzielle Weiterverarbeitung von matooke zu definieren. Das Probenmaterial stammt ausschließlich von der Bananen Plantage der “Presidential Initiative on Banana Industrial Development (PIBID), Technology Business Incubation Center (TBI) at Nyaruzunga – Bushenyi” in West Uganda. Nach einem vollständig randomisierter Entwurf (completely randomized design (CRD) wurden die Bananen Stauden für die Versuche beerntet. Für die Untersuchungen wurde der weichkochende Kultivar Mbwazirume, der in Bushenyi weitverbreitet ist, ausgewählt. Zur Bestimmung der Sorptionsisothermen wurde die Methode COST 90 Project (Wolf et al., 1985) eingesetzt. Es wurde ein eigens entwickelter Versuchstrockner verwendet. Alle Versuche wurden in den Laboren des TBI durchgeführt. Die physiologische Reife von matooke cv. mbwazirume aus Bushenyi ist nach 21 Wochen erreicht. Das optimale Erntefenster zur Weiterverarbeitung als matooke Mehl (Raw Tooke Flour - RTF) in Bushenyi liegt zwischen 15 – 21 Wochen. Zur Ermittlung der Erntereife von matooke wird den Landwirten die Ermittlung des Fingergewichts empfohlen, während den kommerziellen Weiterverarbeitern das kombinierte Modell Fingergewicht und Verhältnis Fruchtfleisch zu Schale empfohlen wird. Matooke Isothermen deuten auf den Typ II hin, welcher charakteristisch für Nahrungsmittel ist. Am besten beschreibt das GAP Modell die Sorption und Desorption der Isothermen. Um ein mikrobiologisch stabiles und kommerziell handelsfähiges Lagerprodukt zu erhalten wird empfohlen unter einen Restwassergehalt von 10 % (wb) zu trocknen. Die erhaltenen Sorptionsisothermen für matooke weisen auf das Hysterese Phänomen hin. Umso geringer die Restfeuchte, umso höher ist die isoterische Sorptionswärme sowohl für Adsorptions- als auch für Desorptionsisothermen. Die gesamte isoterische Sorptionswärme für matooke setzt sich wie folgt zusammen: die Adsorptionsisotherme schwanken zwischen 4,586 – 2,386 Zusammenfassung 107 kJ/kg und die Desorptionsisothermen zwischen 18,194– 2,391 kJ/kg für ein Feuchtigkeitsgleichgewicht von 0.3 – 0.01 (db). Um matooke von 80 auf 10% (wb) runter zu trocknen ist die Energie von mindestens 8,124 kJ/kg zu entfernendes Wasser notwendig. Was wiederum bedeutet, dass die Energie von mindestens 5,793 kJ notwendig ist um 1 kg frisches matooke von 80 auf 10 % (wb) zu trocknen. Die Trocknung von matooke erfolgt in drei Schritten: das Aufwärmen, gefolgt von zwei abfallende Temperaturstufen. Die konstante Trocknungsrate für alle Prozessparameter variiert zwischen 5,793 kJ und der tatsächliche Diffusionskoeffizient liegt zwischen 1.5E-10 - 8.27E-10 m2/s. Die Aktivierungsenergie (Ea) für matooke ist 16.3kJ/mol (1,605 kJ/kg). Vergleicht man die Aktivierungsenergie (Ea) mit der isoterische Sorptionswärme von Adsorptions- als auch von Desorptionsisothermen (qst) (1,297.62) bei 0.1 (kg Wasser/kg TS), so wird klar, dass Ea größer als qst ist, was vermuten lässt, dass sich die Feuchtigkeitsmoleküle in matooke Scheiben im flüssigen Zustand befinden. Die geringsten Farbänderungen (ΔE*) treten bei einer Schichtdicke der Scheiben von 7 mm, einer Umluftgeschwindigkeit im Trockner von 6 m/s, und einer Trocknungstemperatur von 70 °C auf. Eine Reduzierung der Trocknungszeit um 50 % kann erreicht werden, wenn die Oberflächentemperatur des Produktes anstatt die Lufttemperatur im Trockner als Grundlage für die Steuerung des Trockners genutzt wird. Es sind weder physiochemische noch qualitative Einbußen zu erwarten, wenn die Temperatur in den ersten Stufen der Trocknung unter 72 °C liegt und somit keine Verkleisterung des Produktes stattfindet. Die optimalen Prozessparameter bei der einlagigen Trocknung von matooke sind : Scheibendicke = 3 mm, Lufttemperatur 70˚C, Taupunkt Temperatur 18˚C und Luftströmung über dem Produkt von 6 m/s. Bei der kommerziellen Trocknung von matooke wird eine mehrschichtige Trocknung mit einer Beladung des Trockners von weniger oder gleich 7 kg/m² empfohlen und die o.g. Prozessparameter beizubehalten, Scheibendicke = 3 mm, Lufttemperatur 70˚C, Taupunkt Temperatur 18˚C und Luftströmung über dem Produkt von 6 m/s. 108 References References Adeniji T. A., Hart A. D., Tenkouano A., Barimalaa I. S., & Sanni L. O. (2010). Comparative study of pasting properties of improved plantain, banana and cassava varieties with emphasis on industrial application. African Journal of Food Agriculture Nutrition and Development. 10 (5) 2601– 2614. Adewole E., Oso O. A., Talabi J. Y. Ogunmodede O.T., Ajiboye B. O., Ojo O. A., Onikanni S. A., & Adewumi F. D. (2012). Pasting characteristics of plantain (Balbisiana hybrids) and banana (Musa acuminate) starches. Aguirre-Cruz Andres, Roselis Carmona-Garcia & Luis A. Bello-Perez (2010). Moisture adsorption of banana flours (Musa paradisiaca) unmodified and modified by acid-treatment. Paper presented at the XII Congreso Nacional de Ciencia Y Technologia de Alimentos, Guanajuato, Gto. Ajibola O.O. (1986). Desorption isotherms for plantains at several temperatures. Journal of Food Science, 169-171. Akcaoz Handan (2011). Analysis of energy use for banana production. A case study from Turkey. African Journal of Agricultural Research, 6 (25), 5618-5624 Al-Muhtaseb A.H., McMinn W.A.M., & Magee T.R.A (2004). Water sorption isotherms of starch powders. Part 1. mathematical description of experimental data. Journal of Food Engineering, 61, 297-307. Al-Muhtaseb A.H., McMinn W.A.M., & Magee T.R.A (2004). Water sorption isotherms of starch powders. Part 2. thermodynamic characteristics. Journal of Food Engineering, 62, 135-142. Andrade P. Ricardo D., Lemus M. Roberto. & Pérez C. Carmen E. (2011). Models of sorption isotherms for food. Uses and limitations. Vitae, Revista de Facultad de Quimica Farma Ceutica, 18 (3), 325-334. Argyropoulos Dimitrios, Rainer Alex, Robert Kohler & Joachim Müller (2012). Moisture sorption isotherms and isosteric heat of sorption of leaves and stems of lemon balm (Melissa officinalis L.) established by dynamic vapor sorption. LWT – Food Science and Technology, 47, 324-331. References 109 Aviara N.A., Ajibola O.O., Aregbesola O.A. & Adedeji M.A. (2006). Moisture sorption isotherms of sorghum mat at 40 and 50˚C. Journal of Stored Products Research, 42, 290-301. Ayala Aponte Alfredo, Lilian Serma Cock & Gloria Rodriguez Delapav (2011). Moisture adsorptions in yellow pitahaya (Selenicereusmegalanthus). DYNA year 78,.170, 7-14. Babalis Stamatios J. & Belessiotis Vassilios G. (2004). Influence of the drying conditions on the drying constants and moisture diffusivity during the thin-layer drying of figs. Journal of Food Engineering 65, 449-458. Baini, R., & Langrish, T.A.G., (2007). Choosing an appropriate drying model for intermittent and continuous drying of bananas. Journal of Food Engineering. 79, 330 - 343. Bell Leonard N. & Labuza Theodore P. (2000). Moisture Sorption. Practial Aspects of Isotherm Measurement and Use, Second edition, ISBN. 00100426. Bello-Pérez L.A., A. De Francisco, E. Agama-Acevedo, F. Gutierrez-Meraz & F.J.L. García-Suarez, (2005). Morphological and molecular studies of banana starch. Food Science and Technology. International Journal. 11 (5), 367-372. Bello-Perez Luis A., Edith Agama-Acecedo, Paula Osotio-Diaz, Rub G. Utrilla-Coello & Fancisco J. Garcia-Suarez (2011). Banana and Mango Flours in Four and Brads and their Fortification in Health and Disease Prevention. Edited by Victor Ronald WaAson and Vinood Patel ISBN. 978-0-12-380886-8. Bezerra Carolina Vieirra, Edna Regina Amante, Diana Cardoso de Oliveria, Antonio M.C. Rodrigues, Luiza Helena Meller da Silva, (2013). Green banana (Musa Cavendish) flour obtained in spouted bed – Effect of drying on physical-chemical, functional and morphological characteristics of the starch. Industrial Crops and Production, 41, 241-249. Caurie Mathew (2007). Hystteresis phenomenon in foods. International Journal of Food Science & Technology, 42 (1) 45-49. 110 References Chen Chiachung (2006). Obtaining the isosteric sorption heat directly by sorption isotherm equations. Journal of Food Engineering, 74, 178185. Chillet M., O. Huhert and L. De Lapeyre De Bellaire (2006). Effect of the physiological age of bananas (musa spp.) on their susceptibility to wound anthracnose due to Colletotrichum musae. Plant Disease, 1181 – 1185. Chillet M., O. Huhert & L. De Lapeyre De Bellaire (2010). Postharvest disease. effect of the physiological age of bananas (musa sp.) on their susceptibility to wound anthracnose due to Colletotrichum musae. Proceedings for International on Banana & Plantain in Africa. Eds T. Dubois et al. Acta Hort. 879, ISHS 2010. Chong Chien Hwa, Chung Lim Law, Micheal Cloke, Ching Lik Hii & Luqman Chuah Abdullah (2008). Drying kinetics and product quality of dried Chempedak. Journal of Food Engineering, 88, 522–527. Chua K.J., Mujumdar A.S., Halader M.N.A., Chou S.K. & Ho J.C. (2001). Batch drying of banana pieces – effect of stepwise change in air drying air temperature on drying kinetics and product color. Food Research International, 34, 721-731. Ciro Hector, Jairo Alexander Osorio & Elkin Alonso Cortes (2008). Determination of the isosteric heat of plantain pulp (musa paradisiacal) by sorption isotherms. DYNA 156 127-134 Corrêa Paulo C., André L.D. Goneli, Paulo C.A. Júnior, Gabriel H.H. de Oliviera & Domingos S.M. Valente (2010). Moisture sorption isotherms and isosteric heat of sorption of coffee in different processing levels. International Journal of Food Science & Technology, 45, 2016-2022. Corzo Otoniel, Nelson Bracho, Alberto Vasquez & Angle Pereira (2008). Optimization of a thin layer drying of coroba slices. Journal of Food Engineering, 85, 372–380. Dadzie B. K. (1998). Post-harvest characteristics of black sigatoka resistant banana, cooking and plantain hybrids. INIBAP Technical Guidelines INIBAP-ISBN. 2-910810 -24-0 References 111 Dadzie B. K. & J. E. Orchard (1997). Routine postharvest screening of bnana/ plantain hybrids. criteria and methods. INIBAP Technical Guidelines INIBAP-ISBN. 2-910810-22-4 Daramola B. & Osanyinlusi S.A. (2006). Production, characterization and application of banana (Musa spp) flour in whole maize. African Journal of Biotechnology, 5 (10), 992-995. Demirel Devlet & Mahir Turhan, (2003). Air-drying behavior of dwarf cavendish and gros michel banana slices. Journal of Food Engineering, 59, 1-11. Dhatt A.S. & B.V.C. Mahajan (2007). Horticulture postharvest technology harvesting, handling and storage of horticultural crops. Punjab Horticultural Postharvest Technology Centre, Punjab Agricultural University Campus, Ludhiana (16-07-2007) Doymaz Ibrahim (2006). Thin-layer drying behavior of mint leaves. Journal of Food Engineering, 74, 370-375. Doymaz Ibrahim (2007). The kinetics of forced convection air-drying of pumpkin slices. Journal of Food Engineering, 79, 243-248. Erbay Zafer & Filiz Icier (2009). Optimization of hot air drying of olive leaves using random surface method. Journal of Food Engineering, 91, 533– 541. Falade K.O. and O.O. Awole (2005). Adsorption isotherms and heat of sorption of fresh preosmosed oven-dried bananas. Journal of Food Agriculture & Environment, 3 (1), 97-102. Falade O. Kolawole & Ayojesutomi O. Olugbuyi (2010). Effect of maturity and drying method on the physicochemical and reconstitution properties of plantain flour. International Journal of Food Science and Technology, 45, 170–178. FAOSTAT 2011. FAO Agricultural Production Statistics Database. Hassian M.D., Bala B.K., Hossian M.A. & Mondolo M.R.A. (2001). Sorption isotherm and isosteric heat of sorption of pineapple. Journal of Food Engineering, 48, 103-107. 112 References Hofsetz Kelly, Celso Costa Lopes, Miriam Dupas Hubiger, Luis Mayor & Alberto M. Sereno, (2007). Changes in the physical properties of banans on applying HTST pulse during air-drying. Journal of Food Engineering. 83, 531–540. Hunter Lab (2008). Application Notes Vol.8 No. 7 Iglesias, H.A., & J.Chirife (1982). Handbook of food isotherms. Water Sorption Parameters for Food and Food Components. Academic Press, New York, London. IITA (2009). International Institute of Tropical Agriculture (IITA) Ikegwu O.J., Nwobas N.V., Odoh M.O. & Oledinma N.U. (2009). Evaluation of pasting and some functional properties of stach isolated from some improved cassava varieties in Nigeria. Electronic Journal of Environmental, Agricultural and Food Chemistry, 8 (8), 647-665. Inglett George E., Diejun Chen, Jingyuan Xu and Suyong Lee (2012). Pasting and rheological properties of β-glucan-enriched hydroclloids from oat bran concentrate. Journal of Food Processing and Preservation. 1-7 Islam M.S., Haque M.A. and Islam M.N. (2012). Effect of drying parameters on dehydration of green banana (Musa sepientum) and its use in potato (Solanum tuberosum) chips formulation. The Agriculturists 10 (1), 8797. Jamali A., M. Kouhila, L. Ait Mohamed, J.T. Jaouhari, A Idlimam & N. Abdenouri (2006). Sorption isotherms of Chenopodium ambrosiodes leaves at three temperatures. Journal of Food Engineering, 72, 77–84. Jiang Hao, Min Zhanga & Arun S. Mujumdar (2010). Physicochemical changes during different stages of MFD/FD banana chips. Journal of Food Engineering, 101, 140–145. Johnson P.N.T. & Brennan J.G. (2000). Moisture sorption isotherm characteristics of plantains (Musa, AAB). Journal of Food Engineering, 44, 79-84. References 113 Johnson P.N.T., Brennan J.G. & Addo-Yobo F.Y. (1998). Air-drying characteristics of plantain (Musa AAB). Journal of Food Engineering, 37, 233-242. Jose A. Larrauri Garcia & Fulgencio Saura Calixto (2000). Evaluation of CIELab colour parameters during clarification of a syrup from Mesquite pods (Prosopis Oallida L.). International Journal of Food Science and Technology 2000, 35, 385–389. Kaddumukas P., Kyamuhangire W., Muonga J. & Muranga F.I. (2005). The effect of drying methods on the quality of green banana flour. African Crop Science Conference Proceedings, 7, 1267-1271. Kader A.A. (1999). Fruit maturity, ripening and quality relationships. Proceeding of International Symposium on effect of Pre- and Post Harvest Factors on Storage of Fruit. Ed. L. Mitchalczuk, Act Hortculture, 485, ISHS, 203208. Kalawole O. Falade & Titilayo A. Kolawole (2012). Physical, functional and pasting properties of different maize (Zea mays) cultivars as modified by an increase in γ-irradiation doses. The International Journal of Food Science & Technology, 47, 801–807. Karamura Eldad, E. Frison, D.A. Karamura & S. Sharrock (1998). banana production systems in Eastern and Southern Africa. bananas and food security - ed. by C. Picq, E. Fouré & E.A. Frison, INIBAP, Montpellier, 401-412, ISBN 2-910810-36-4 Karim Md. Azharrul & M.N.A. Hawlader, (2005). Drying characteristics of banana. theoretical moldelling and experimental validation. Journal of Food Engineering, 70 (1), 35–45. Karugaba & Kimaru 1999, Banana production in Uganda, an essential food and cash crop. RELEMA –Techinical Handbook No. 18, ISBN 9966896-39-2. Karugaba Aloysius & Gathiru Kimaru (1999). Banana production in Uganda. ISBN 9966-896-39-2 114 References Kaymak-Ertekin Figen & Atil Gedik (2004). Sorption isotherms and isosteric heat of sorption of grapes, apricots, apples and potatoes. Lebensm.Wiss. U-Technol. 37, 429-438. Kechaou N. & Maalej M. (1999). Desorption isotherms for imported banana application of the GAB theory, Drying Technology, 17 (6), 1201-1213. Kiranoudis C.T., Maroulis Z.B., Tsami E. & Marinos-Kouris D. (1993). Equilibrium moisture content and heat of desorption of some vegetables. Journal of Food Engineerin, 20, 55-74. Kirk Ronald, Ronald Sawyer (1991). Pearson’s Composition and Analysis of Foods. 9th Edition, Longman Scientific and Technical. ISBN No. 0-58240910-1 Krokida M.K. & Marinos-Kouris D. (2004). Rehydration kinetics of dehydrated products. Journal of Food Engineering, 57, 1-7. Krokida M.K., Kiranoudis C.T. & Maroulis Z.B. (1999). Viscoelastic behavior of dehydrated products during rehydration. Journal of Food Engineering, 40, 269-277. Krokida M.K., Kiranoudis C.T., Maroulis Z.B. & Marinos-Kouris D. (2000). Drying related properties of apple. Drying Technology, 18 (6), 12511267. Labuza T. P., L. Mcnally, Denise Gallagher, J. Hawkes and F. Hurtado (1972). Stability of Intermediate Moisture Foods. 1. Lipid Oxidation. Journal of Food Science, 37, 154-159. Lai His-Mei & Hsiao-Hsien Cheng (2004). Properties of pregelatinized rice flour made by hot air or gum puffing. International Journal of Food Science and Technology, 39, 201-212. Larrauri Garcia J. A. & F. Saura Clixto (2000). Evaluation of CIE-lab colour parameters during the clarification of a sugar syrup from Mesquite pods (Prosopis Pallida .L) International Journal of Science and Technology, 35, 385-389. References 115 Laylieam, Somchai, & Kosittrakun Mani, (1998). A research note. Effects of harvest maturity on banana quality. Journal of Food Quality, 22, 539544. Lehmann Undine, Gisela Jacobasch, and Detlef Scmiedl, (2002). Characterization of Resistant Starch Type III from Banana (Musa acuminata). Journal of Agricultural and Food Chemistry, 50, 52365240. Leite J.B., M.C. Mancini, & S.V. Borges, (2007). Effect of drying temperature on the quality of dried bananas cv. prata and d’άgua. LWT, 40, 319323. Ling L.-H., E.M. Osman, J.B. Fernandes & P.J. Relly, Ames (1992). Physical properties of starch from cavendish banana fruit. Starch/Stärke, 24 (6), 184-188. Madamba P.S., Driscoll R.H. & Buckle K.A. (1994). Shrinkage, drying and porosity of garlic during drying. Journal of Food Engineering, 23, 309319. Maskan Medeni (2000). Microwave/air and microwave finish drying of banana. Jounal of Food Engineering, 44, 71-78. Mathlouthi Mohamed (2001). Water content, water activity, water structure and the stability of foodstuffs. Food Control, 12, 409-417. McLaughlin C.P. & Magee T.R.A. (1998). Determination of sorption isotherm and the isosteric heats of sorption for potatoes. Journal of Food Engineering, 35, 267-280. McMinn W.A.M., & T.R.A Magee (2003). Thermodynamic properties of moisture sorption of potato. Journal of Food Engineering, 60, 157-165. Mehta Sanjeev & Singh Amarjit (2006). Adsorption isotherms for red chilli (Capsicum annum L.). European Food Research and Technology, 233. 849-852. Menkov N.D. & Dinkov K.T. (1999). Moisture sorption isotherm of obacco seeds at three temperatures. Journal of Agricultural Engineering Research, 74, 261-266. 116 References Menkov N.D., A.G. Durakowa, & A. Krasteva (2005). Moisture sorption isotherms of common bean flour at several temperatures. Electronoc Journal of Environmental, Agricultural and Food Chemistry, 4 (2), 892– 898. Millan-Testa C.E., M.G. Mendez-Montealvo, M.A. Ottenhof, I.A. Farhat & L.A. Bello-Pérez (2005). Determination of the molecular and structural charateristics of okenia, mango and banana starch. Journal of Agricultural and Food Chemistry, 53, 495-501. Moreira R. Chenlo F., Torres M.D. & Vallejo N. (2008). Thermodynamic analysis of experimental sorption isotherms of loquat and quince fruits. Journal of Food Engineering, 88, 514-521. Muchui M.N., Njorege C.K., Kahangi E,M. & Onyango C.A. (2010). Determination of maturity indices of tissue cultured bananas (Musa spp.) ‘Williams’ and ‘Grade Naine’. (2010). Proceedings of International Conference on Banana & Plantain in Africa, Eds,. T.Dubois et al., Act Hortculture 879, ISHS, 425-430. Mujumadar, A.S. (2007). Handbook of Industrial Drying. CRC Press. Boca Raton, FL, USA. Mujumdar, A.S. (2000). Drying Technology in Agriculture and Food Sciences. ISBN 1-57808-148-3, Science Publishers Inc. Enfield, NH, USA. Muranga F. I. 1998 Composition and Physiochemical Characteristics of Starches of different Banana Varieties. PhD Thesis submitted at Makerere University. Muranga F. I., Mutambuka M., Nabugoomu F., and Lindhauer M. G. (2010). Optimization of Raw Tooke Flour, vital gluten and water absorption in Tooke/wheat composite bread: Effect of raw Tooke flour and vital gluten on wheat flour physicochemical and dough rheological properties (Part I). African Journal of Food Science, 4 (5), 223–230. Muranga, F.I., Sampath, H., Marlett, J.A., & Ntambi, J.M., (2007). Impact of processing technique on the apparent bioavailability of cooking banana (matooke) starch. African Journal of Biochemistry Research, 1(5), 72 – 77. References 117 Myers Rymond, H., Douglass C. Montyomery & Christine M. AdersonCook, (2009). Response Surface Methodology. Process and product optimization using designed experiments. Third Edition, ISBN 978-0470-17446-3. John Wiley & Sons Inc. Publishers. Hoboken, NJ, USA. Nelson C. Scot, Randy C. Ploetz & Angela Kay Kepler (2006). Musa species (banana and plantain) Species profiles for pacific island agroforestry www.traditionaltree.org) Newport Scientific, (2006). Rapid Visco Analyzer Series S4A (RVA-Super4) Installation and Operation Manual Nguyen Minh-Hue, & William E. Price, (2007). Air-drying of banana. Influence of experimental parameters, slab thickness, banana maturity and harvesting season. Journal of Food Engineering, 79, 200-207. Niba L. L., M. M. Bokanga, F.L. Jackson, D.S. Schlimme & B.W. Li (2002). Physiochemical properties of starch granular characteristics of flour from various Manihot Esculenta (cassava) genotypes. Journal of Food Science, 67 (5), 1701-1704 Noordin M.Y., Venkatesh V.C., Sharif S. , Elting S. & Abdullah A. (2004). Applicaton of response surface methodology in describing the performance of coated carbide tools when turning AISI 1045 steel. Journal of Materials Processing Technology, 145, 46-58. Nyombi K., P.J.A. van Asten, P.A. Leffelaar, M. Corbeels, C.K. Kaizzi, & K.E. Giller, 2009. Allometric growth relationships of East Africa highland bananas (Musa spp., AAA-EAHB) cv. Kisansa and Mbwazirume. Annals of Applied Biology 155, 403−418. Nyombi Kenneth (2010). Understanding growth of Eat African Highland Banana. Experiment and simulation. PhD thesis submitted at WageningenUniversity. ISBN. 978-90-8585-550-7. Oluwalana I.B., Oluwamukomi M.O., Fagbemi T.N. & Oluwafemi G.I., (2011). Effect of temperature and period of blanching on the pasting and functional properties of plantain (Musa parasidiaca) flour. Journal of Stored Products and Postharvest Research, 2 (8), 164-169. 118 References Osungbaro, Taiwo. O., Jimoh D., & Osundeyi E., (2010). Functional and pasting properties of composite cassava-sorgum flour meals. Agriculture and Biology Journal of North America, 1 (4), 715-720. Pacheco-Delahaye Empratríz, Ronald Maldonado, Elevina Pérez & Mily Schroeder (2008). Production and characterization of unripe plantain (Musa paradisiacal L.) flours. Interciencia, 33 (4), 291-296. Pahlevanzadeh H. & M. Yazdani (2005). Moisture adsorption isotherms and isosteric energy for almond. Journal of Food Processing Engineering, 28, 331–345. Perera O. Conrad (2005). Selected quality attributes of dried foods. Drying Technology, 23, 717–730. Phoungchandang S. & Woods J.L. (2000). Moisture diffusion and desorption isotherms for banana. Journal of Food Science, Food Engineering and Physical Properties, 65 (4), 651-657. Robinson John C. and Victor Galan Saúco (2010): Bananas and Plantains. 2nd Edition ISBN 13: 978 1 84593 658 7. Selvamani, P., Manivannan K., & Jagan Mona, (2009). Proximate composition and pasting behavior of starch from Indian bananas (Musa sp). Botany Research International, 2 (2) 103-106. Shewfelt L. Robert (2009). Measuring quality and maturity in postharvest handling (Second Edition) A System Approach. Editors. Wojciech J. Florkowski, Robert L. Shewfelt, Bernhard Brueckner & Stanley E. Prussie. ISBN. 978-0-12-374112-7 Siddiq M., A. Ravi, J.B. Harte & K.D. Dolan (2010). Physical and functional characteristics of selected dry bean (Phaseolus vulgaris L.) flours. LWT – Food Science and Technology, 43, 232–237. Singh S. P., Bhatt L., & Prasad M. (2004). Pyhsico-quality changes associated with growth and development of banana (Musa sp.) fruit CV dwarf Cavendish. Agricultural Science Digest, 24 (3), 197–199. References 119 Siripatrawan U. & Jantawat P. (2006). Determination of moisture sorption isotherms for jasmine rice crackers using BET and GAB models. Food Science and Technology International, 12 (6), 459-465. Stover Robert H. & Simmonds Norman W. (1987). Bananas, Longman Scientific & Technical, ISBN 0582463572 Sturm B., & Hofacker W., (2009). Optical monitoring and control of drying process. In DAAAM International Scientific Book 2009; B. Katalinic Ed; DAAAM International. Vienna. Austria, 591-212. Sturm Barbara, Werner C. Hofacker & Oliver Hensel, (2012). Optimizing the drying parameters for hot-air-dried apples. Drying Technology. An International Journal, 30, (14), 1570-1582. Sun Da-Wen & Byrne C. 1(998). Selection of EMC/ERH isotherm equations for rapeseed. Journal of Agricultural Engineering Research, 69, 307315. Tapre, A.R., & Jain, R.K., (2012). Study of advanced maturity stages of banana. Technical Journals online.com, IJERS, I (III) 272-274. Thornton P. and Cramer L. (2012). Impacts of climate change on the agricultural and aquatic systems and natural resources within the CGIAR’s mandate. CCAFS Working Paper 23. Thuwapanichayanan Ratiya, Somkiat Prachayawarakorn, Jaruwan Kunwisawa & Somchart Soponronnarit, (2011). Determination of effective moisture diffusivity and assessment of quality attributes of banana slices during drying. LWT – Journal of Food Science and Technology, 44, 15021510. Thys R.S.C., Caciano P.Z. Norena, Ligia D.F. Marczak, Andereia G. Aires, & F. Cladera-Olivera (2010). Adroption isotherms of pinhao (Aracria angstifolia seeds) starch and thermodynamic analysis. Journal of Food Engineering 100, 468 – 473. Timmermann E.O., Chirife J. & Iglesias H.A. (2001). Water sorption isotherms of food and foodstuffs. EAT or GAB parameters. Journal of Food Engineering, 48,19-31. 120 References Tribess T.B., Hernandez-Uribe J.P., Mendez-Montealvo M.G.C., Menezes E.W., Bello-Perez L.A. & Tadini C.C. 2009. Thermal properties and resistant starch content of green banana flour (Musa cavendishii) produced at different drying conditions. Food Science and Technology, 42, 1022-1025. Tsami E., (1991). Research notes. Net isoseteric heat of sorption in dried fruits. Journal of Food Engineering 14, 327-335. Turner W.David, Jeanie A. Fortescue and Dane S. Thomas (2007). Environmental physiology of bananas (Musa sp). Brazil Journal of Plant Physiology, 19 (4) 463–484. UBOS, (2010). Uganda Bureau of Statistics Vatanasuchart Nednapis, Boonma Niyomwit & Karuna Wongkrajang (2012). Resistant starch content, in vitro starch digestibility and physic-chemical properties of flour and starch from Thai bananas. Maejo International Journal of Science and Technology 6 (02), 259 – 271. Veg-Galvez A., R. Lemus-Mondaca, P. Fito and A. Andres (2007). Note. Moisture Isotherms and Isosteric Heat of Red Bell Pepper (var. Lumuyo). Food Science and Technology International, 13 (4), 309–316. Veg-Galvez Antonio, Marlene Palacios & Roberto Lemus-Mondaca (2008). Moisture sorption isotherms and isosteric heat determination in Chilean papaya (Vasconcllea pubescens). Quim. Nova., 31 (6), 1417-1421. Vinzenz B. M. Bauer, (2011). Banana (Musa spp. AAA-EA) marketing in Uganda. Should bananas be weighed in future? Conference paper presented at the Conference on International Research on Food Security, Natural Resources Management and Rural Development. Trpentag 2011. Universty of Bonn. Wang N. & Brennan J.G. (1991). Moisture sorption isotherm characteristics of potatoes at four temperatures. Journal of Food Engineering, 14, 269287. Wolf, W., Spiess, W.E.L., & Jung, G., (1985). Standardization of isotherm measurement (COST-Project 90 and 90 BIS). In Properties of Water in References 121 Foods in Relation to Quality and Stability, by D. Simatos & J.L. Multon, 661-679. Yan Zhengyong, Maria J. Sousa-Gallagher and Fernanda A.R. Oliveria (2008). Sorption isotherms and moisture sorption hysteresis of intermediate moisture content banana. Journal of Food Engineering, 86, 342 – 348. Yang Zhao, Zhu Elong & Zhu Zongsheng (2012). Moisture sorption isotherm and net isosteric heats of sorption of green soybean. International Journal of Food Engineering, 8 (3) Article15, 1-16. Yazdani M., Suzandehchi P., Azizi M. & Ghobadi P. (2006). Moisture sorption isotherms and isosteric heat for pistachio. European Food Research and Technology, 223, 577-584. Yi Yuan, Liming Zhang, Yujie Dai & Jiugao Yu (2007). Physicochemical properties of starch obtained from Dioscorea nipponica Makino comparison with other tuber starches. Zhang Pingyi, Roy L. Whistler, James N. BeMiller, & Bruce R. Hamaker (2005). Banana starch. Production, Physicochemical Properties, and Digestibility. A Review. Carbohydrate Polymers. 59, 443-458. Zogaz, N.P., Maroulis, Z.B., & Marinos-Kouris, D. (1996). Moisture diffusivity data compilation in foodstuffs. Drying Technology, 14(10), 2225-2253. Appendix A 122Appendix A 130 Appendix A 130 Appendix Appendix A A Appendix A 1.6 1.6 3 mm Moisture RatioRatio (x/xo(x/x ) Moisture o) 1.2 3 mm 1.2 5 mm 5 mm 0.8 7 mm 0.8 7 mm 0.4 y = 1.4156e-0.035x R² = 0.9844 y = 1.4156e-0.035x R² = 0.9844 y = 1.3592e-0.016x R² = 0.9873 y = 1.3592e-0.016x R² = 0.9873 y = 1.228e-0.01x R² = 0.9902 y = 1.228e-0.01x R² = 0.9902 0.4 0 0 0 100 200 300 400 0 100 Time 200(min) 300 400 Time (min) Figure 5.1A Plot of moisture ratio (experimental and predicted) of different Figure 5.1A Plot of moisture ratio (experimental predicted) of thickness versus the thickness versus the drying time of and matooke atdifferent air temperature 60odrying C, dew o o time of5.1A matooke temperature 60 C, dew point temperature C and airthickness velocity 4.5m/s. Figure Plot at of air moisture (experimental and predicted) 15 of different versus the drying point temperature 15oCratio and air velocity 4.5m/s. o o time of matooke at air temperature 60 C, dew point temperature 15 C and air velocity 4.5m/s. Time (min) 0 100 Time (min) 200 300 400 0 100 200 300 400 0 0 -2 Ln(Moisture Ratio) Ln(Moisture Ratio) -2 -4 -4 -6 -6 -8 -8 -10 -10 3 mm 3 mm 5 mm 5 mm 7 mm 7 mm y = -0.0347x + 0.3476 R² = 0.9844 y = -0.0347x + 0.3476 R² = 0.9844 y = -0.0164x + 0.3069 R² = 0.9873 y = -0.0164x + 0.3069 = 0.9873 y = R² -0.01x + 0.2054 R² = 0.9902 y = -0.01x + 0.2054 R² = 0.9902 Figure 5.2A Plot of Logarithm of moisture ratio at different thickness versus the drying time of matooke at air temperature 60oC, dew point temperature 15oC and air velocity 4.5m/s. Appendix A 131 Figure 5.2A Plot of Logarithm of moisture ratio at different thickness versus the drying time of matooke at Appendix AppendixAA 131123 o o air temperature 60 C, dew point temperature 15 C and air velocity 4.5m/s. Moisture Ratio (x/xo)Ratio (x/xo) Moisture Figure 5.2A Plot of Logarithm of moisture ratio at different thickness versus the drying time of matooke at 1.6 o o air temperature 60 C, dew point temperature 15 C and air velocity 4.5m/s. y = 1.1057e-0.022x 50˚C R² = 0.9882 1.6 1.2 y = 1.3527e-0.035x 60˚C -0.022x y = 1.1057e 50˚C R² = 0.9838 R² = 0.9882 1.2 0.8 70˚C 60˚C 0.8 0.4 -0.035x 1.3527e-0.04x yy==1.3318e 0.9838 R²R²==0.9852 70˚C y = 1.3318e-0.04x R² = 0.9852 0.40 - 100 Time (min) 200 300 0 Figure 5.3A Plot of moisture ratio (experimental and predicted) of different drying air temperature versus 100 200 300 o the drying timePlot of matooke at dew point temperature 15 C, air velocity 4.5m/s thickness 3mm Figure 5.3A of moisture ratio (experimental and and predicted) of different Time (min) drying air temperature versus the drying time of matooke at dew point Figure 5.3A Plot of moisture ratio (experimentalTime and(min) predicted) of different drying air temperature versus temperature 15oC, air velocity 4.5m/s and thickness 3mm o the drying time of matooke at dew point temperature 15 C, air velocity 4.5m/s and thickness 3mm 0 0 100 200 Time (min) Ln(Moisture Ln(Moisture Ratio) Ratio) 0 -1 0 100 200 70˚C 50˚C -1 -2 60˚C + 0.1717 y = -0.0224x R² = 0.9884 70˚C -2 -3 y = -0.0415x + 0.3355 R² = 0.9841 -4 -3 y = -0.0415x + 0.3355 R² = 0.9841 -4 50˚C 60˚C y = -0.0224x + 0.1717 R² = 0.9884 y = -0.035x + 0.3021 R² = 0.9838 y = -0.035x + 0.3021 R² = 0.9838 Figure 5.4A Plot of Logarithm of moisture ratio at different drying air temperature versus the drying time of matooke at dew point temperature 15oC, air velocity 4.5m/s and thickness 3mm Figure 5.4A Plot of Logarithm of moisture ratio at different drying air temperature versus the drying time of o matooke at dew point temperature 15 C, air velocity 4.5m/s and thickness 3mm Appendix A 124 1.6 Moisture Ratio (x/xo) 1.2 15˚C y = 1.4169e-0.035x R² = 0.984 25˚C y = 1.3612e-0.029x R² = 0.9884 35˚C y = 1.3729e-0.029x R² = 0.9724 0.8 0.4 0 - 50 100 150 200 Time (min) Figure 5.5A Plot of moisture ratio (experimental and predicted) of different Figure 5.5A temperature Plot of moisture ratio andofpredicted) of different point temperature dew point on (experimental drying time matooke at airdew temperature 60oon C, air o drying time of matooke at air temperature 60 C, air velocity 4.5m/s and thickness 3mm. velocity Appendix A4.5m/s and thickness 3mm. 133 Time (min) 0 Ln(Moisture Ratio) - 50 100 150 200 -2 15˚C -4 25˚C 35˚C -6 y = -0.035x + 0.3484 R² = 0.984 y = -0.0293x + 0.3084 R² = 0.9882 y = -0.0287x + 0.3169 R² = 0.9724 Figure 5.6A Plot of Logarithm of moisture ratio at different dew point o Figure 5.6A Plot of moisture at different temperature on time velocity of temperature onLogarithm dryingoftime of ratio matooke at dew air point temperature 60drying C, air o matooke at air temperature 60 C, air velocity 4.5m/s and thickness 3mm. 4.5m/s and thickness 3mm. Appendix Appendix AA 134 125 134 Appendix A 2.4 2.4 Moisture Moisture Ratio Ratio (x/x(x/x o) o) y = 1.0073e-0.009x -0.009x = 0.9816 y =R² 1.0073e R² = 0.9816 3 m/s 3 m/s 2 2 1.6 1.6 4.5 m/s 4.5 m/s 1.2 1.2 6m/s 6m/s 0.8 0.8 y = 1.2212e-0.02x -0.02x = 0.9879 y =R²1.2212e R² = 0.9879 y = 2.1568e-0.05x -0.05x = 0.7941 y =R²2.1568e R² = 0.7941 0.4 0.4 0 0 0 0 100 100 200 300 200Time (min)300 400 400 500 500 Time (min) Figure Plot of moisture ratio (experimental and predicted) of drying different Figure5.7A 5.7A Plot of moisture ratio (experimental and predicted) of different air velocity on the time air o drying time o o Figure 5.7A Plot moisture time ratio predicted) of air thickness velocity velocity on atthe drying ofdew matooke at airoftemperature of on 50 C, dew point C, point and temperature 15different C and the of the 3mm. of matooke air of temperature of 50(experimental o o of matooke at air temperature of 50 C, dew point temperature of 15 C and the thickness of 3mm. temperature of 15oC and the thickness of 3mm. Ln(Moisture Ln(Moisture Ratio) Ratio) 0 0 0 0 -1 -1 -2 -2 Time (min) Time (min) 100 100 200 200 3 m/s 3 m/s 4.5 m/s 4.5 m/s 6 m/s 6 m/s -3 -3 300 300 y = -0.0099x + 0.133 R² = 0.9946 y = -0.0099x + 0.133 R² = 0.9946 y = -0.0224x + 0.0848 R² = 0.9879 y = -0.0224x + 0.0848 R² = 0.9879 y = -0.0269x + 0.0831 R² = 0.9908 y = -0.0269x + 0.0831 R² = 0.9908 Figure 5.8A Plot of Logarithm of moisture ratio at air velocity on the drying time of matooke at air temperature of 50oC, dew point temperature of 15oC and the thickness of 3mm. 126 Appendix B Appendix B Figure 6.1B Response surface showing the effect of air temperature and dew point temperature on drying time Figure 6.2B Response surface showing the effect of air temperature and air velocity on drying time. (Air velocity (m/s): 3, 4.5 & 6 are represented by air flow (m3/h): 400, 600 & 800 respectively.) Appendix B 127 Figure 6.3B Response surface showing the effect of air velocity and dew point temperature on drying time. (Air velocity (m/s): 3, 4.5 & 6 are represented by air flow (m3/h): 400, 600 & 800 respectively). Figure 6.4B Response surface showing the effect of air temperature and dew point temperature on pasting properties (final viscosity). 128 Appendix B Figure 6.5B Response surface showing the effect of air temperature and dew point temperature on pasting properties (peak viscosity). Figure 6.6B Response surface showing the effect of air temperature and dew point temperature on starch content