Effects of Coffee Roasting Technologies on Bioactive Compounds and Cup Quality of Specialty Coffee Beans Grown in Ethiopia

ADDIS ABABA UNIVERSITY ADDIS ABABA INSTITUTE OF TECHNOLOGY SCHOOL OF CHEMICAL AND BIO-ENGINEERING

FOOD ENGINEERING PROGRAMME

A Thesis Submitted to the School of Chemical & Bio-Engineering in Partial Fulfillment of the Requirements for the Degree of Master of Science (M.Sc.) in Chemical Engineering (Food Engineering)

Muluken Zenebe Bolka

November 7, 2019 Addis Ababa, Ethiopia

Effects of Coffee Roasting Technologies on Bioactive Compounds and Cup Quality of Specialty Coffee Beans Grown in Ethiopia

By Muluken Zenebe Bolka

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Science (M.Sc.) in Chemical Engineering (Food Engineering) School of Chemical and Bio-Engineering Addis Ababa Institute of Technology Addis Ababa University

Advisor: Dr. Eng. Shimelis Admassu (Associate Professor)

Declaration

I, the undersigned, declare that this thesis presented for the attainment of the degree of Master of Science in Chemical Engineering (Food Engineering) has been composed solely by myself and has not been presented or submitted, in part or as a whole, for a degree in any other institution or university. The thesis presented is my original work and all sources of information or materials used for the thesis have been duly acknowledged.

Mr. Muluken Zenebe Bolka Signature: ______Date: ______

M.Sc. Candidate

This thesis has been submitted for examination with my approval and done under my supervision as University advisor. Dr. Eng. Shimelis Admassu Emire Signature: ______Date: ______Advisor

The undersigned members of the thesis examining board appointed to examine the thesis of Mr. Muluken Zenebe which is submitted for the degree of Master of Science in Chemical Engineering (Food Engineering) confirmed that the thesis fulfills the requirements of the graduate program and approved it to be accepted.

Approved by the Examining Board: Signature Date

______Chairman, School’s Head

Dr. Ing. Zebene Kifle ______Internal Examiner

Dr. Beteley Tekola ______External Examiner

Dedication

I dedicate this thesis to my father for inspiring me as a young boy to have ambitious dreams. Thank you Ane! I have made it here with your inspiration.

I also dedicate this thesis to my mother for teaching me what kind patience looks like in the face of tremendous obstacles. By your relentless prayers, I am here. I love you, Bayu!

Additionally, I dedicate this work to my brothers, Tedla (Teddy), Zelalem (Oche), Tibebu (Abuye), and sister, Medhanit (Mimi). I am fortunate to be your brother.

Finally, I dedicate this work to Bonna Coffee Processing for introducing me to my new found passion for “Coffee”.

iii

Acknowledgements

First of all, I am so thankful to God for giving me the wisdom and strength to undertake and complete this study. Next, I am very grateful to my advisor, Dr. Eng. Shimelis Admassu for the confidence he had when he inspired me to undertake such an interesting and bold thesis dissertation title. His guidance and comments have were key to accomplish the research objectives. I am indebted to my sponsor Hawassa University and the School of Chemical and Bio Engineering, Addis Ababa University (AAU), for giving me the opportunity to pursue my graduate study.

I would like to express my thanks to the managing director of Ethiopian Coffee and Tea Authority Coffee Quality Inspection and Certification Center (ECTACQICC), Mr. Berhanu Gezahegn for his kind permission to freely conduct the study at the center. His staffs at the coffee quality laboratory were very helpful, especially Mrs. Segenet, Mr. Seifedin and Mr. Tefera who shared their vast experience in coffee grading and sensory evaluation (cupping). I would like to also thank Mr. Taye and Mr. Agegnehu from Moyee Coffee PLC for allowing me to visit their company and use their roast analyzer device during their busy work time.

My gratitude also goes to the staffs of Agricultural Quality Research Laboratory at the Ethiopian Institute of Agricultural Research (EIAR), especially Dr. Kassaye who granted me the permission to carry out part of my research in the institute’s laboratory. I would like to extend my special thanks from the bottom of my heart to Mr. Dinka Mulugeta for his priceless technical support when conducting the laboratory experiments at EIAR. Without him, this work may not have been possible. Last but not least, I would like to thank the staffs of EIAR at Wondogenet branch, especially Mr. Eyasu, Mr. Mulugeta and Mr. Bekri for their warm hospitality and technical support in finalizing my research.

It remains for me to thank my family for their encouragements and love that in diverse ways made this study possible. I owe a special debt of gratitude to my uncle Mr. Asnake Bekele from whom I kindly obtained the samples for the study. Through him, I was able to access and get a hands on experience on the latest IKAWA® Pro roaster. A special thanks to the Aboneh’s family, Wondimagen’s family and Bonna Coffee Processing for making my stay in Addis enjoyable as always. I’m also grateful to Mr. Muluken Kebede and Mr. Woldeab Alemayehu for cheering me up throughout the entire journey. May God bless everyone who supported me during the study!

Abstract

The objective of this research was to study the effects of different type of coffee roasting technologies, namely Drum, Fluidized Bed and Traditional roasters on the cup quality and bioactive compounds of specialty coffee beans grown in Ethiopia. A total of 3 unwashed raw coffee varieties were selected for the study and collected from Yirgacheffe, Harar and Sidama coffee regions in Ethiopia. The sample coffee beans were specialty graded and roasted at light, medium & dark degree of roasts from 150°C - 200°C for 7 to 15 minutes. The roasted coffee beans were evaluated for cup quality at medium degree of roast by three professional cuppers. Next, the amount of caffeine, trigonelline and total chlorogenic acids present in each coffee samples were simultaneously analyzed using a high-performance liquid chromatography (HPLC). The amount of acrylamide formed in Yirgacheffe coffee during roasting was determined using UV-Visible Spectrophotometer. Based on cupping evaluation, traditional roaster resulted at the least total cup score for all the coffee varieties. Among other factors, only varietal difference and degree of roast were found to have significant (P < 0.05) effect on caffeine content of the sample coffee beans. Generally, a significant reduction in trigonelline and total chlorogenic acids content of the coffee beans was observed during roasting process, with darker roasts attaining the least values. A single analysis of variance on experimental results of acrylamide formation in roasted Yirgacheffe coffee showed significant (p < 0.05) differences for type of roaster and degree of roast. The highest acrylamide content, i.e. 2.303ppm was exhibited at light degree of roast using traditional roaster. Finally, drum roaster was found to be the best type of coffee roaster at medium degree of roast since it produced specialty coffee beans with the highest cup quality, optimum bioactive compounds content and minimum acrylamide formation at medium degree of roast.

Keywords: Acrylamide, Bioactive Compounds, Cup Quality, Roaster, Specialty Coffee

T Table of Contents a Chapter Title Page № bl Title Page ...... i e Declaration ...... ii of Dedication ...... iii C Acknowledgements ...... iv o Abstract ...... v n Table of Contents ...... vi te List of Tables ...... viii n List of Figures ...... ix ts List of Abbreviations ...... x

List of Appendices ...... xi 1. Introduction ...... 1 1.1. Background ...... 1 1.2. Statement of the problem ...... 2 1.3. Research Question ...... 3 1.4. Objectives ...... 3 1.4.1. General Objective ...... 3 1.4.2. Specific Objectives ...... 3 1.5. Significance of the study ...... 3 2. Literature Review ...... 4 2.1. Overview of Coffee Production and Marketing ...... 4 2.1.1. Global Overview of Coffee Production and Marketing ...... 4 2.1.2. National Overview of Coffee Production and Marketing ...... 5 2.2. Levels of Coffee Processing ...... 6 2.2.1. Primary Coffee Processing ...... 6 2.2.2. Secondary Coffee Processing ...... 8 2.2.3. Tertiary Coffee Processing ...... 9 2.3. Coffee Roasting Process Technology ...... 11 2.3.1. Main Aspects of Coffee Roasting Process ...... 11 2.3.2. Types of Coffee Roasting Technologies ...... 13

2.4. Evaluation of Roasted Coffee Quality ...... 15 2.4.1. Quality of Raw Coffee Beans ...... 16 2.4.2. Specific Roasting Conditions ...... 24 3. Materials and Methods ...... 27 3.1. Raw Material Collection, Transportation and Preparation...... 27 3.1.1. Raw Material Collection ...... 27 3.1.2. Raw Material Transportation and Storage ...... 27 3.1.3. Raw Material Preparation ...... 28 3.2. Experimental Framework of the Research ...... 34 3.3. Methods of Analysis ...... 35 3.3.1. Physical Examination of Coffee Beans...... 35 3.3.2. Cupping of Roasted Coffee Beans ...... 39 3.3.3. Determination of Bioactive Compounds ...... 40 3.3.4. Determination of Acrylamide Concentration...... 42 3.3.5. Experimental Design and Statistical Analysis ...... 44 4. Results and Discussion ...... 45 4.1. Physical Examination of Coffee Beans ...... 45 4.1.1. Sieve Analysis ...... 45 4.1.2. Color Analysis ...... 45 4.1.3. Moisture Analysis ...... 47 4.2. Cupping of Roasted Coffee Beans...... 48 4.3. Composition of Bioactive Compounds ...... 50 4.4. Acrylamide Concentration ...... 57 5. Conclusions and Recommendations...... 59 5.1. Conclusions ...... 59 5.2. Recommendations ...... 60 References ...... 61 Appendices ...... 67

vii

List of Tables

Table № Title Page № Table 2.1. Top ten green coffee producing countries in 2014 ...... 5 Table 2.2. Requirement for Roasted Coffee Beans and Roasted Ground Coffee ...... 15 Table 2.3. Chemical composition of dry roasted Coffea arabica and Coffea robusta beans ...... 24 Table 3.1. Agronomical Practices of the Obtained Coffee Samples ...... 27 Table 3.2. Description of the Sample Coffee Codes ...... 33 Table 3.3. SCAA Color Value of Roasted Coffee Beans ...... 38 Table 3.4 HPLC Equipment Operating Condition ...... 41 Table 4.1. SCAA color value of Sample coffee beans roasted at different conditions ...... 46 Table 4.2. Moisture Content (%) of Sample Raw Coffee Beans ...... 47 Table 4.3. Moisture Content of Yirgacheffe Coffee Roasted at Different Conditions...... 47 Table 4.4. Cupping Scores of Roasted Coffee Beans ...... 49 Table 4.5. Correlation coefficients among drum roasted coffees presented in matrix form ...... 58 Table 4.6. Correlation coefficients among fluidized bed roasted coffees presented in matrix form ...58 Table 4.7. Correlation coefficients among traditionally roasted coffees presented in matrix form .....58

viii

List of Figures

Figure № Title Page № Figure 2.1. Different parts of coffee cherry (fruit) ...... 7 Figure 2.2 Primary coffee processing ...... 7 Figure 2.3. Integrated Flow Sheet of Brewed, Instant, Decaffeinated and Roasted Coffee ...... 10 Figure 2.4. Main aspects of roasting coffee beans ...... 12 Figure 2.5. Drum roaster with low speed hot air blown through its bottom ...... 13 Figure 2.6. Fluidized bed roasting of coffee ...... 14 Figure 2.7. Traditional Ethiopian Oven Top Coffee Roaster...... 15 Figure 2.8. Chemical structure of Caffeine (1,3,7 trimethylaxanthine) ...... 19 Figure 2.9. Coffee Flavor Precursors ...... 23 Figure 2.10. Effect of roasting time on acrylamide concentration in coffee ...... 25 Figure 2.11. Roast Profile of Coffee (for Tanzanian Coffee Bean) ...... 26 Figure 2.12. Typical composition per 100 ml coffee brew from medium roasted coffee ...... 26 Figure 3.1. Drum Type Sample Roaster (BRZ 4, Probat®, USA, 2009) ...... 29 Figure 3.2. Fluidized Bed Sample Roaster (V3, IKAWA® Pro, UK, 2018)...... 30 Figure 3.3. Coffee Roast Profile Displayed on a Smartphone ...... 30 Figure 3.4. A Traditional Coffee Roaster Manually Wired in with Thermocouple ...... 31 Figure 3.5. Laboratory Coffee Grinder (VTA6S, MAHLKÖNIG GmbH & Co., Germany, 2009) ....32 Figure 3.6. Experimental Framework of the Research Work ...... 34 Figure 3.7. Primary and Secondary Coffee Bean Defects ...... 35 Figure 3.8. Sieve Analysis of Coffee Beans ...... 36 Figure 3.9. Measuring Color Value of Roasted Coffee Beans using Roast Analyzer...... 37 Figure 4.1. Caffeine Content of Drum Roasted Coffee Beans ...... 51 Figure 4.2. Caffeine Content of Coffee Beans Roasted Using Fluidized Bed Roaster ...... 51 Figure 4.3. Caffeine Content of Coffee Beans Roasted Using Traditional Roaster ...... 51 Figure 4.4. Trigonelline Content of Sample Coffee Beans Roasted Using Drum Roaster...... 52 Figure 4.5. Trigonelline Content of Coffee Beans Roasted Using Fluidized Bed Roaster ...... 53 Figure 4.6. Trigonelline Content of Coffee Beans Roasted Using Traditional Roaster ...... 54 Figure 4.7. Total CGAs Content of Sample Coffees Roasted by Drum Roaster ...... 55 Figure 4.8. Total CGAs content of Sample Coffees Roasted Using Fluidized Bed Roaster ...... 56 Figure 4.9. Total CGAs Content of Sample Coffees Roasted by Traditional Roaster ...... 56 Figure 4.10. Acrylamide Content of Roasted Yirgacheffe Coffee ...... 57

List of Abbreviations

AOAC Association of Official Analytical Chemists AQRL Agricultural Quality Research Laboratory asl Above sea level CGAs Chlorogenic Acids CQA Caffeoylquinic acid ECTACQICC Ethiopian Coffee and Tea Authority Coffee Quality Inspection and Certification Center ECX Ethiopian Commodities Exchange EIAR Ethiopian Institute of Agricultural Research EIPO Ethiopian Intellectual Property Organization ES Ethiopian Standard ESA Ethiopian Standard Agency HPLC High Performance Liquid Chromatography ISO International Standard Organization LPG Liquefied Petroleum Gas PTFE Polytetrafluoroethylene SCAA Specialty Coffee Association of America SNNPR Southern Nations, Nationalities and Peoples Region UCDA Uganda Coffee Development Association USDA United States Department of Agriculture

List of Appendices

Appendix № Title Page № Appendix 1. Preliminary Unwashed Coffee Quality Assessment Sheet ...... 67 Appendix 2. Cupping Protocol used at ECTACQICC...... 70 Appendix 3. Specialty Coffee Cupping or Grading Sheet ...... 71 Appendix 4. HPLC Chromatograms of Sample Raw Coffee Beans ...... 74 Appendix 5. Calibration Curves for the HPLC Analysis ...... 75 Appendix 6. HPLC Chromatograms of Roasted Coffee Beans ...... 76 Appendix 7. Photos of Sample Coffee Beans ...... 82 Appendix 8. Absorbance Reading of Standard Acrylamide Samples at 220nm ...... 84 Appendix 9. Calibration curve for Acrylamide Analysis ...... 84 Appendix 10. Absorbance Reading of Roasted Yirgacheffe Samples at 220nm ...... 85 Appendix 11. Photos of the Research Laboratory Works ...... 86

Chapter One 1. Introduction 1.1. Background Globally, coffee is the second major traded commodity next to oil (UCDA, 2007). And it is the most commercialized food product and most widely consumed beverage in the world (Farah, 2012). Coffee cultivation and processing is an important source of income for many developing countries including Ethiopia (Poltronieri & Rossi, 2016). It is backbone to Ethiopian economy with about 15 million people, mostly women, directly or indirectly deriving their livelihoods from and generating about 50% of the country’s agricultural export earnings (USDA, 2017).

Coffee is a tropical bushy plant that is highly cultivated for its cherry fruits which are processed in various ways to produce green (raw), roasted and grounded, instant/ soluble and brewed beverage according to market requirements (Nicolas, 2004). However, a good quality coffee could be rejected unless good processing practices are strictly followed (Sualeh & Mekonnen, 2015).

Roasting is the most important coffee value addition process that develops the characteristic flavor, aroma and color of coffee beans (Mussatto et al., 2011). However, achieving an ideal roast is a goal that is complicated since coffee beans behave differently and produce different results in physical properties, chemical composition and biological activities when roasted at different conditions (Kelly, 2018).

There are many types of coffee roasting technologies. Traditionally, coffee beans are roasted in a saucepan by continually stirring them to provide a more even heat (Kelly, 2018). The most common type of coffee roaster available for home and industrial use is drum roaster. It is a rotating horizontal cylinder that roasts coffee beans placed inside it by continually rotating and heating them by hot air pumped through either the center of the cylinder or through its perforated sides to ensure an even roast. One of the recent developments in coffee roasting technology is fluidized bed roasting in which high velocity hot gas directed towards the beans, usually from the bottom of the roasting machine, so that the gases heat and move the floating beans simultaneously (Clarke, 2001). This research studies the effect of these three types of roasting technologies on the quality of specialty coffee beans grown in Ethiopia (Corrêa et al., 2016).

1.2. Statement of the Problem

Roasting is an important process that creates the characteristic flavor and aroma of coffee (Marilisa & Anese, 2018). There are different types of coffee roasting technologies based on the type of heat source, mode of heat transfer, degree of automation, mode of operation, availability of auxiliary units, operating conditions, etc. (Clarke & Vitzthum, 2001). Modern types of coffee roasting machines are either gas fired or electrically heated with certain time-temperature controlling units, sensors and cooling units that enable better control of the degree of roast (Sinnott, 2010). On other hand, traditional coffee roasting machines are wood fired and completely manual which may impose difficulty to control the roasting process without burning the beans and compromising the flavor of the final beverage (Kelly, 2018). The mode of heat transfer in some coffee roasters is convective while it’s conductive in others (Koskei et al., 2015). Certain types of roasters enable full visual sight of the beans during roasting whereas other types are closed systems that provide sampler spoon and/ or sight glass to check the doneness of the beans (Sinnott, 2010).

In Ethiopia, there is an emerging coffee roasting industry that employs different kinds of roasting technologies to produce roasted coffee beans for both local and foreign markets (Nicolas, 2004). However, the use of different types of coffee roasting technologies may cause quality variations among coffee beans of the same origin, especially specialty coffee beans which have unique flavor and high market value (Clark, 2003). In addition, improperly roasted beans are known to be carcinogenic due to high level of phenolic compounds like acrylamide (Bae et al., 2014).

Although the quality of roasted coffee beans is mainly determined by the inherent chemical composition of the raw coffee beans (Summaa et al., 2007; Mastronardi, 2012; Shan et al., 2016; Uman et al., 2016) and the roasting conditions, i.e. the temperature-time profile and the skill set of an individual roaster (Sinnott, 2010; Kelly, 2018), certain type of coffee roasters may be better for achieving high cup quality of roasted coffee beans (Yeretzian et al., 2002; Poltronieri & Rossi, 2016; Diallo, 2019). Therefore, it is important to study the effect of different type of coffee roasting technologies on the quality of selected specialty coffee beans grown in Ethiopia.

1.3. Research Question

- Does roasting of coffee beans using different type of coffee roasting technologies cause variation in their cup quality? - Can type of coffee roaster affect the bioactive content of coffee bean? - Which type of coffee roasting technology is the most suitable for maintaining the unique cup characteristics of specialty coffee beans grown in Ethiopia? 1.4. Objectives 1.4.1. General Objective

The general objective of this research was to study the effect of type of coffee roasting technologies on the cup quality and bioactive compounds content of specialty coffee beans grown in Ethiopia. 1.4.2. Specific Objectives The specific objectives of this research were to: - Investigate the changes in the physical properties of the sample coffee beans roasted at different degree of roast using different type of roasting technologies. - Compare the cup quality of specialty coffee beans roasted at medium degree of roast using selected type of coffee roasting technologies. - Determine the content of bioactive compounds, i.e. caffeine, trigonelline and total chlorogenic acids found in sample coffee beans roasted using different type of roasting technologies. - Determine the concentration of acrylamide formed in a selected coffee variety roasted using different type of roasting technologies. 1.5. Significance of the study The outcome of this study would have significant contribution to filling the current research gap on the effects of coffee roasting technologies on the cup quality and bioactive compounds content of specialty coffee beans. In addition, the results obtained from this study can be used by coffee traders, coffee processors and researchers as useful inputs when purchasing, using and studying coffee roasting technologies, respectively.

Chapter Two 2. Literature Review 2.1. Overview of Coffee Production and Marketing Coffee is a tropical bushy plant which belongs to the Rubiaceae family, genus Coffea (Farah, 2012), and highly cultivated for its cherry fruits (Nicolas, 2004). The cherries of the plant are processed into seeds or beans used for making different coffee drinks. 2.1.1. Global Overview of Coffee Production and Marketing

Globally, coffee is the second major traded commodity next to oil and thus plays a vital role in the balance of trade between developed and developing countries, providing the latter with an important source of export earnings to pay for imports of capital and consumer goods. (UCDA, 2007). It is traded in various forms such as raw (green), roasted, grounded, instant/ soluble and brewed coffee. And coffee drink is one of the most consumed beverages in the world (UCDA, 2007).

Although more than 80 coffee species have been identified worldwide (Clark, 2003), there are only two commercially important species, namely Coffea arabica and Coffea robusta (Mastronardi, 2012). Among the two commercially important coffee species, Coffea arabica (Arabica) accounts for 64% while Robusta accounts for about 35% of the world’s coffee production. Compared to Arabica, the price of Robusta is much lower; therefore it is much more popular among industrial clients (Tin, 2006).

Coffee is grown and exported by more than 60 countries. Brazil, Vietnam, Columbia, Indonesia & Ethiopia are among the major coffee producer/ exporting countries in the world (Nicolas, 2004). And the United States, Germany, Japan, & Italy are among the major coffee consumer/ importing countries in the world (USDA, 2017). The following table summarizes the top ten coffee producers with their annual coffee production in 2014.

Table 2.1. Top ten green coffee producing countries in 2014

Rank Country Production (Million Metric Tons) 1 Brazil 2.8 2 Vietnam 1 3 Colombia 4 4 Indonesia 0.7 5 Ethiopia 0.6 6 India 0.4 7 Honduras 0.3 8 Guatemala 0.3 9 Peru 0.2 10 Uganda 0.2 Total World Production 8.8

Source: USDA (2017) In spite of the global high consumption amount, today’s coffee market is characterized by huge over supplies and declining quality. The continual splitting of large coffee estates, the excessive expansion of new plantations and the growing number of intermediaries in the marketing chain has resulted in a deterioration of coffee quality and its price (Nicolas, 2004).

2.1.2. National Overview of Coffee Production and Marketing

Coffee production is important to the Ethiopian economy with about 15 million people, mostly women, directly or indirectly deriving their livelihoods from it. It is the country’s major export item that the generates nearly 50% of its agricultural export earnings (USDA, 2017). Coffee is primarily grown in three regions of the country namely Oromia, Harari, and Southern Nations, Nationalities and Peoples Region (SNNPR). And it is a vital cash crop or primary source of income for many coffee growing farmers, processors & traders in the top coffee producing areas in the country mainly Sidama, Gedeo, Jima, Kaffa, Harar, & Neqemte (Nicolas, 2004).

Ethiopia is the largest producer of coffee in Sub-Saharan Africa and is the fifth largest coffee producer in the world next to Brazil, Vietnam, Colombia, and Indonesia, contributing about 7 to 10% of total world coffee production (Tefera & Tefera, 2013). In the year 2011/12, the country exported around 200,000 tons of green (unroasted) coffee to many countries out of which Germany, Saudi Arabia, Belgium, France, and USA were the top 5 export destinations(USDA, 2017).

The Ethiopian coffee production is mainly destined for export through the Ethiopian Commodities Exchange (ECX). Sidama, Harar and Yirgacheffee are internationally registered trademarks of Ethiopian fine coffees that are distributed and sold in many countries to foreign coffee companies, roasters and distributors from Ethiopian coffee exporters, coffee farmers and cooperative unions (EIPO, 2008). Sidama coffee takes the country’s largest coffee export share up to 45% in terms of volume and value during 2006 to 2013 fiscal year (Minten et al., 2014). Despite their smaller production size, Harar and Yirgacheffe coffees are the most expensive Ethiopian coffees in the international market (Dejene, 2011; Amamo, 2014) and among the renowned coffees in the global specialty coffee market ((Workneh, 2015)). However, Ethiopian coffee production and total export earning is unsatisfactory due to the country’s poor agricultural productivity and fluctuating global market price of coffee (USDA, 2017).

2.2. Levels of Coffee Processing

Processing plays a crucial role in determining the quality of coffee (Sualeh & Mekonnen, 2015). The process of converting harvested coffee cherries to coffee beverage (for humans) with characteristic flavor, aroma, color & nutritional properties using commercially feasible methods is called coffee processing (Clarke, 2001). And the application of technological principles to coffee processing is generally understood as coffee process technology (Nicolas, 2004). There are many types of coffee process technologies that enable the production of coffee in the form of green (raw), roasted & ground, instant/ soluble and brewed beverage according to market requirements. Generally, coffee processing can be classified in to three levels: primary, secondary, and tertiary coffee processing (UCDA, 2007).

2.2.1. Primary Coffee Processing

After harvesting, coffee fruits or cherries are processed in various ways to produce green coffee beans (Clarke, 2001). Arabica and Robusta coffees are different in their ideal growing climates, physical aspects, chemical composition, and characteristics of brew made with their ground roasted seeds (Farah, 2012). However, the cherry or fruit of any coffee species have the following common parts: - An external skin (exocarp), red or yellow when the fruit is ripe. - A mucilaginous flesh (mesocarp), known as pulp and mucilage.

- Generally, two beans (each one called an endosperm) which contain a germ (embryo). - Each bean is covered by silverskin (spermoderm) and surrounded by parchment (endocarp). - If one bean aborts, its place remains empty and the other one grows into a more rounded shape (a peaberry). The following figure shows the different parts of a coffee cherry or fruit.

Figure 2.1. Different parts of coffee cherry (fruit) Source: Nicolas (2004) Generally, primary coffee processing aims to lower the water content of fresh cherries to a level which allows the preservation of beans (about 11-12 %) by removing all the covering which surround the beans (i.e. the skin, pulp, and parchment (husk)) to expose two (sometimes one) coffee beans from each cherry as shown in the figure below (Nicolas, 2004).

WET/ DRIED COFFEE CHERRY PARCHMENT COFFEE GREEN COFFEE

Figure 2.2 Primary coffee processing Source: Nicolas (2004) 7

Generally, there are two methods of primary coffee processing; they are known as dry processing and wet processing (Clarke & Vitzthum, 2001).

A. Dry Processing

In dry processing, whole coffee cherries (i.e. pulp, parchment and bean) are dried in the sun for a period of up to three weeks until their moisture content is approximately 10%–12% (Nicolas, 2004). After drying, the dry cherries are mechanically dehulled to remove the whole hull (i.e. dried pulp and parchment) and expose the coffee beans (“green coffee”) which can be further cleaned or polished to remove a mucilaginous material (silver skin) adhering to their surface. To obtain a good-quality beverage, the beans are mechanically and electronically sorted to separate defective beans from the high quality beans before bagging or packaging (Clarke, 2001). This method is usually applied to coffea arabicas from Ethiopia, as well as to all coffea robustas, of which Brazil, Indonesia and Vietnam are the main producers. In addition, this method is commonly used in places where sun and space are abundant (Ullman, 2011). B. Wet Processing The wet-processing technique is more sophisticated and generally produces higher quality green coffee beans (Nicolas, 2004). In this method, fresh cherries are first dumped in flotation tanks to separate the unripe cherries from the ripe ones. Then the outer skin of the cherries is removed by squeezing them through slots. Next, the pulp surrounding the beans is extensively washed away after being fermented in slatted tanks where microbial and naturally occurring pectic enzymes solubilize the mucilage. The Fermentation times varies from 6 to 72 h, depending on the temperature, the amount of mucilage, and the concentration of peptic enzymes (Clarke & Vitzthum, 2001). The beans, still covered by the inner shell, the parchment, are dried either for a period of up to 1 week in the sun. The main producer of wet-processed coffee is Colombia, followed by the Central American and Eastern African countries (Ullman, 2011). 2.2.2. Secondary Coffee Processing Secondary coffee processing, also known as export grading, is the final post-harvest process that transforms clean coffee beans (i.e. after mechanical removal of parchment) into various coffee grades that meet the international standards (Anonymous, 2007). This stage involves: cleaning (polishing), drying, de-stoning, size grading, gravimetric sorting done manually or using electronic sensors and finally bagging coffee in jute bags of 60 Kg for export (Nicolas, 2004).

2.2.3. Tertiary Coffee Processing

Apart from exporting green coffee beans, coffee can also be processed to make higher value-added coffee products. This stage is known as tertiary coffee processing. It involves roasting, grinding, making of instant coffee, brewing, decaffeination and other processes (UCDA, 2007).

Whether done using simple saucepan or modern drum roaster, roasting is the most important coffee value addition process that develops the characteristic flavor, aroma and color of coffee. It eventually determines the quality of final coffee beverage (Mussatto et al., 2011). Relative to the raw material price, an added value of 100–300% is achieved through roasting (Yeretzian et al., 2002). Grinding is also a key coffee value addition process that underpins the preparation of beverage coffee and other coffee by-products. It is done using manual small scale grinders or motorized mills involve the use of a hammer plate, vertical plate or hammer mills for large scale production (Clarke & Vitzthum, 2001).

Brewing is another way of coffee value addition. There are various kinds of coffee brewing techniques or methods such as espresso, filter and French press (Ullman, 2011). Espresso is a method of forcing a near boiling water through a compressed pack of ground-roasted coffee. In French Press, ground coffee is mixed with hot water and a filtered press is passed through the brew separating the grinds from it. Whereas Filter method passes hot water through ground coffee sitting on filter paper (Kelly, 2018).

Decaffeination is the process of removing caffeine from coffee bean using solvent extraction method to minimize its “negative” physiological effects (Ullman, 2011). It is performed before roasting in order to minimize flavor and aroma losses that may occur during roasting processes. “Decaffeinated coffee” means a coffee with a maximum caffeine concentration of 0.1% related to the dry mass (Kirk-Othmer, 1985).

Instant or soluble coffee production typically involves treating ground-roast coffee with softened water at high temperature and pressure to extract the water-soluble compounds (Clarke, 2001). This soluble material is then cooled, centrifuged, concentrated by heating, and dried by freeze- drying or spray drying method to reduce moisture to approximately 5%. The yield and quality of final product can be improved by blending ground-roast coffee of Arabica species with ground- roast Robusta coffee which contains higher amounts of soluble solids (Kirk-Othmer, 1985).

Green (Raw) Coffee Beans

Swelling Roasting - Swelling the raw beans with water in order to be soluble & - Coffee beans are put in available for caffeine extraction. contact with hot gas or surface (2000C - 4500C ) for 1 to 12 min until the desired degree of Extracting roast (color, aroma, flavor) is Roasted & Ground reached. - Extracting the caffeine from the Coffee beans with a solvent such as dichloromethane, ethyl acetate, or Cooling supercritical carbon dioxide. - The beans are discharged from the roaster and cooled rapidly by water quenching and air Stripping cooling. - Steam stripping to remove all Boiling/ Leaching solvent residues from the beans. Grounding - Certain amount of ground coffee is added in boiling water, infusing it for several minutes - Roasted beans are grounded (up to 30 min). to coarse, medium, or fine Adsorbent Regeneration - Or by leaching a bed of ground coffee kept on particle size according to type a filter at slightly above atmospheric pressure - Regenerating adsorbents of coffee drink & consumers' for few minutes using boiling water or steam. such as activated carbon (if choices. applied). Extracting - Water-soluble compounds Filtering are extracted from the ground Drying - Desired concentration of soluble coffee with soft water at high solids is obtained using filter paper, - Drying the decaffeinated coffee temperatures. cloth or mesh screen. beans to moisture content of green (raw) coffee bean. Centrifuging

- To separate the solids from Simmering/ Calming Weighing & Packaging the liquids in the extract. - The boiled coffee is left to - To be roasted, ground & simmer (i.e. to become calmer & brewed like ordinary coffee and Concentrating reach temperature below boiling consumed by people with health - Further removal of water point) before being served in a problems related to caffeine. through evaporation. cup.

Decaffeinated Coffee Drying - To reduce the moisture to Coffee Drink approximately 5% using freez-drying or spray drying.

- Packed in paper bags or plastic containers to be quickly brewed and served to busy customers.

Instant Coffee

Figure 2.3. Integrated Flow Sheet of Brewed, Instant, Decaffeinated and Roasted Coffee Adopted from Clarke & Vitzthum (2001), Nicolas (2004) and Ullman (2011)

2.3. Coffee Roasting Process Technology

2.3.1. Main Aspects of Coffee Roasting Process

The aroma of green coffee beans is very weak and it is only through roasting that the seeds gain the characteristic aroma and flavor of coffee (Nicolas, 2004). During roasting, heat energy is induced from the roaster into the green bean through hot gases or hot metallic surfaces to raise their temperature in the range of 170–230 °C for 10 – 20 min (Yeretzian et al., 2002; Sinnott, 2010) and thereby start complex chemical and physical changes (Kelly, 2018). Generally, coffee roasting process can be divided into three stages or phases: a drying phase, during which moisture is eliminated, a roasting phase, where a number of complex pyrolytic reactions take place, and finally a cooling phase, where the freshly roasted coffee is quickly cooled in order to stop over roasting (Yeretzian et al., 2002).

During the initial stage, green coffee beans are brought to around 100 °C and undergo dehydration process where most of the free water is driven out and evaporate in an endothermic process (Wang, 2012). The 10 – 12% moisture content of green coffee beans is reduced to just 5 – 6% (Yeretzian et al., 2002). In addition, the green color of the bean changes to yellowish (Kelly, 2018), the smell of the beans changes to bread-like or freshly mowed grass and the silver skin comes off, as chaff, leaving the yellowing grassy smelling bean (Yeretzian et al., 2002).

During the second stage, the actual roasting process starts once the beans’ moisture content reaches about 6%. Hence, the temperature of the coffee beans rises again and reaches around 170oC, where exothermic and pyrolytic reactions begin generating heat, building high pressure inside closed voids within the porous coffee bean microstructure (Ivon, 2002) and releasing large quantities of gas (5–12 l/kg of mainly CO2) (Marilisa & Anese, 2018). During the roast phase over two hundred different reactions occur causing various volatiles to be produced and released from the coffee beans. The release of water vapor and CO2 account for the majority of the mass loss of the bean during roasting (Bustos-Vanegas et al., 2017). As a result, the coffee beans swell by 50 – 100% leading to violent release of built up water when cells explode under the high internal pressure (Sinnott, 2010). At this stage it is possible to hear the coffee beans pop. This sound is called the “first crack” or “first pop” (Yeretzian et al., 2002). At approximately 190oC, the coffee beans start developing a full body with the nice acidity, flavors, rich aroma and color for which coffee is known (Marilisa & Anese, 2018). Early roasting is characterized by the Maillard reaction followed

by caramelization. As this reaction occurs, the color of the bean changes from yellowy brown to the darker brown that is characteristically recognized as coffee (Kelly, 2018). This phase ends with more rapid popping sound called “second crack” or “second popping” which is caused by a pyrolytic reaction. This crack also causes the coffee bean’s cell walls to rupture leading to the oils trapped in the coffee to seep out (Sinnott, 2010).

As the roast continues into its final stage, charring and burning cause the coffee beans to gradually lose their positive acidity, rich aroma and develop a pronounced burnt taste (Yeretzian et al., 2002). The coffee beans go from dark brown to black (Kelly, 2018). When the desired degree of roast is reached, the beans must be discharged from the roaster and cooled rapidly by water quenching and/ or air cooling in order to avoid over-roasting due to subsequent exothermic reaction. Air cooling is preferred than water quenching because further water increases the risk of microbial growth (Nicolas, 2004). However, the cooling process must be fast, so it is better to do it using fan or forced air convection. The efficiency of this process depends on the properties of the coffee bean (Bustos-Vanegas et al., 2017).

The main aspects of the roasting process are illustrated in the figure below.

Figure 2.4. Main aspects of roasting coffee beans Source: Geiger et al. (2001)

2.3.2. Types of Coffee Roasting Technologies

Different types of roasters are designed with different mechanical principles, so that beans and hot gas come into contact in different ways to get properly roasted coffee bean (Kelly, 2018). The mode of heat transfers during roasting can involve conduction, convection, and radiation (Baggenstoss et al., 2008). In all type of the roasting machines, coffee is usually agitated to ensure that it is evenly roasted (Kelly, 2018).

I. Drum Roasters

The most common roasters available for home and industrial use are traditional drum roasters or rotating cylinder which involves placing the coffee in a rotating drum that tumbles the beans above a powerful gas burner, heating the coffee by hot air pumped through either the center of the cylinder or through its perforated sides and ensuring an even roast (Nicolas, 2004). Usually the roasting process takes 8 to 12 minutes. This roasting system is used in both batch and continuous roasting systems. And the process is ended by rapid cooling of the coffee (Marilisa & Anese, 2018). The mass loss rate of the beans required to achieve a commercial degree of roast is constant for cylindrical type of roasters (Guillermo et al., 2016). A typical drum roaster involves three forms of heat transfer to roast coffee (convection, conduction & radiation) (Diallo, 2019).

Figure 2.5. Drum roaster with low speed hot air blown through its bottom Source: Kelly (2018) Drum roasters require an operator with sensory expertise who must react to the audible, visual and aromatic changes that occur during the roasting process by adjusting the heat and airflow settings (Diallo, 2019).

II. Fluidized Bed Roasters One of the recent developments in coffee roasting technology is fluidized bed roasting in which high velocity hot gas directed towards the beans, usually from the bottom of the roasting machine, so that the gases heat and move the floating beans simultaneously (Clarke & Vitzthum, 2001). The maximum temperatures used in industrial fluidized bed roasters generally vary from 210oC to 240oC (Nicolas, 2004). Fluidized bed roasters are preferred for industrial use because they are faster with typical roasting times of 1 to 4 minutes (Kelly, 2018), allow better control of air temperature and speed inside the roasting chamber, and produce a more homogeneous color than other roasters. Fast roasted coffee beans are different from traditionally roasted beans: they are larger, have a reduced density and greater porosity. Because of these structural differences the fast roasted beans allow better water penetration and extraction of the soluble compounds (Kelly, 2018). According to Eggers & Pietsch (2001) and Diallo (2019), coffees roasted in fluidized bed roaster is almost exclusively based on convective heat transfer that results in low density and high yield coffee. Whereas, Nagaraju et al. (1997) stated that drum roasters mainly involve conductive heat transfer producing roasted coffees with less soluble solids, more degradation of chlorogenic acids, more burnt flavor, and higher loss of volatiles than those roasted using fluidized bed roasters.

Figure 2.6. Fluidized bed roasting of coffee Source: Kelly (2018)

III. Traditional Roasters Coffee beans are traditionally roasted in a saucepan or a tin can with a handle that is slowly and continually stirred over a fire to provide a more even heat (UCDA, 2007).

Figure 2.7. Traditional Ethiopian Oven Top Coffee Roaster 2.4. Evaluation of Roasted Coffee Quality

Roasted coffee beans are coffee beans that have been properly processed and roasted. They must be free from any artificial coloring, extraneous matter and undesirable flavor (ESA, 2002a). The quality of roasted coffee depends on many factors, such as the quality of raw coffee beans, the specific roasting conditions, the time since the beans are roasted, and the type of packaging material used for storing the coffee (Wang, 2012). According to the Ethiopian Standard Agency, roasted coffee beans shall comply with the following requirements shown in Table 2.2. Table 2.2. Requirement for Roasted Coffee Beans and Roasted Ground Coffee

# Characteristics Requirement Roasted Coffee Beans Roasted Ground Coffee 1 Moisture, % (w/w), max 3.0 5.4 2 Total ash, % (w/w) 2 – 5 2 - 5 3 Acid insoluble ash, % (w/w) 1 1 4 Water soluble matter, % (w/w) 25 – 32 25 - 32 5 Alkalinity of water soluble ash, ml 0.1N 3.5 – 5.0 3.5 – 5.0 HCl per g of material 6 Caffeine, % (w/w), min 0.8 0.8 7 Petroleum ether extract, % (w/w) 8.5 8.5

Source: ESA (2002a)

2.4.1. Quality of Raw Coffee Beans

Quality of raw or green coffee beans is a function of physical appearances, organoleptic or cup- quality, and inherent chemical constituents. It is a complex characteristic, which depends mainly on genetic factors, environment, agronomy, processing, storage condition, brewing process and taste of consumers (Sualeh & Mekonnen, 2015). The criteria commonly used to evaluate the quality of raw coffee beans include bean size, color, shape, roast potential, processing method, crop year, species, variety, flavor or cup quality and presence of defects (Banks et al., 1999). Among these, the last two, i.e. cup quality and presence of defects, are the most important criteria that are employed in the global coffee trade (Franca & Oliveira, 2008). The Specialty Coffee Association of America (SCAA) defines specialty coffee in its green stage as coffee that is free of primary defects, is properly sized and dried, presents no faults or taints in the cup and has distinctive attributes. In practical terms, this means that the coffee must be able to pass aspect or preliminary grading and cupping tests (Poltronieri & Rossi, 2016). 2.4.1.1. Origin, Species and Variety

Among the two main coffee species, Arabica coffee is considered to have better taste than Robusta (Ivon, 2002). All coffees of Ethiopian origin belong to the Coffea Arabica species. The country largely produces and exports different coffee varieties such as Harar, Yirgachefe, Sidama, Jima (Djimma), Limu, Teppi and Bebeka (Nicolas, 2004). Harar coffee fetches some of the highest price for unwashed coffee in the world market. It is mostly exported as unwashed grade 4 & 5. Most of its export goes to the Middle East and Japan. It grows on an altitude ranging from 1510 – 2120m above sea level (CLU, 2017). Yirgacheffe is widely recognized as specialty coffee district in Ethiopia. The coffee grows at an altitude ranging from 1770 – 2200m above sea level. It is exported as washed grade 2 (CLU, 2017). Sidama mostly produces washed coffee for export but also produces exceptionally highest quality unwashed (sun-dried) coffee too. The coffee grows at an elevation of 1400 to 2200m above sea level on the slope of the mountain shoulder of the rift valley (CLU, 2017). Limu mostly produces coffee which grows at an elevation of 1400 – 2100m above sea level (CLU, 2017). Jimma (Djimmah) coffee is usually a natural or sundried coffee. It grows at elevation of 1400 – 1800m asl. And Lekempte coffee is produced in the western part of the country. The bulk of the coffee comes from west Wellega zone, Oromia region. It grows at an altitude of 1700 – 2200m above sea level (CLU, 2017). Teppi coffee grows in south west Ethiopia

at an altitude of 1100 – 1900m asl. Bebeka coffee also grows in the south western part of the country at an altitude of 950 – 1285m asl (CLU, 2017).

2.4.1.2. Cup Quality of Coffee

Cup testing (cupping) is a standard approach used to evaluate raw coffee quality. It is performed by professional cuppers who judge the coffee for its flavor, mouth feel and after taste (Oliveira et al., 2018). However, this kind of assessment is inevitably subjective since it’s based on personal experience and memory of the professional cupper (Nicolas, 2004). Therefore, it is essential to explore alternative methods to accurately assess the chemical characteristics and the quality of coffee beverage (Clemente et al., 2018). The basic coffee attributes evaluated during cup tasting (cupping) are: 1. Aroma - the fragrance or odor perceived by the nose. 2. Taste- which is perceived by the tongue. 3. Flavor - which is the combination of aroma and taste. The flavor which contributes to the quality of the coffee is described in terms of winey, spicy and fragrant. Off-flavors such as grassy, onion, musty, earthy, etc., reduce coffee quality. 4. Body is a feeling of the heaviness or richness on the tongue. 5. Acidity is a sharp and pleasing taste. It can range from sweet to fruity/ citrus and is considered as a favorable attribute. Ethiopian coffees have excellent taste. Harrar coffee is soft, mild and fragrant with a strong winey flavor, but is more acid and has less body. It is famed in the world for its Mocha flavor. When cupped, the coffee has medium to light acidity, full body and strong mocha flavor with blueberry notes (CLU, 2017). Yirgacheffe coffee is widely recognized as specialty coffee in Ethiopia. It is set apart for its exceptional citrus and floral flavor. It has fine acidity and body. Sidama coffee has distinct inherent quality of balanced acidity and body. It has a cup characteristics of fine acidity, medium body with spicy and citrus flavor. Sidama coffee is usually blended with other coffees for gourmet (CLU, 2017). Limu coffee is known for its distinctive winey flavor. It has a well-balanced cup quality of medium acidity and body. Usually, Yirgacheffe, Sidama, & Limu Coffee have similar remarkable soft flavor and fragrant aroma with low acidy (Nicolas, 2004). Jimma coffee is known for its fair average cup quality with lower or light medium acidity, full or good heavy body, hard balanced flavor and singular taste (Nicolas, 2004; CLU, 2017). Lekempte coffee has a cup

characteristics of medium to pointed acidity and medium body with a wild fruity finish. Teppi coffee has a cup characteristics of low to medium acidity and smooth after taste. Bebeka coffee has light acidity and medium to good body cup characteristics (CLU, 2017). However, coffee supplied to local market has lower quality and even subjected to adulteration such as mixing with roasted barley (USDA, 2017). 2.4.1.3. Physical Characteristics of Raw Coffee Beans

Physical characteristics of raw coffee beans such as size, shape, mass, density, presence or absence of defects and color present significant variations during roasting process (Franca & Oliveira, 2008). Raw or green Harar coffee bean has a characteristics thick and oval shape, medium size and yellow (amber) color. Raw Yirgacheffe coffee beans are greenish to bluish in color, uniform, and thick, medium to bold, oval in shape and compacted. Green Sidama coffee beans are small to medium in size, and greenish to grayish in color. Raw Limmu coffee beans are flatter than other washed coffees, medium sized, greenish to bluish in color and mostly round in shape. Raw Lekempte coffee bean is generally medium in size and green to bluish in color. Raw Teppi coffee beans are large and thick with white center cut, pointed tips, and low density. Green Bebeka coffee beans are medium in size, greenish to bluish in color, and mostly round in shape (CLU, 2017).

Non-uniformity of coffee beans in a sample affects the heat transfer during roasting process and subsequently reduces cup quality (Sualeh & Mekonnen, 2015). Coffee bean size is usually assessed in order to ensure uniformity where over 85% of the beans should be in the same size range and remain on top of 14 size screen after sieving (Sivetz & Desrosier, 1979; ECX, 2015). Due to the non-enzymatic browning and pyrolysis reactions, the color of coffee beans changes during roasting process; and it must be monitored to control the effects of heating and the required degree of roast. Another important physical aspect of coffee beans that is commonly monitored during coffee roasting is moisture content. According to the Ethiopian Commodity Exchange’s ECX Coffee Contacts, the moisture content of an export grade coffee shall not be more than 11.5% by weight (ECX, 2015). It is highly related to weight loss behavior of the coffee beans and the reaction pathways for the production of volatile components at different processing temperatures (Clark, 2003).

2.4.1.4. Chemical Composition of Raw Coffee Bean

Coffees of superior quality are those that have chemical compounds responsible for the flavor and aroma such as caffeine, trigonelline, aldehydes, furans, ketones, sugars, proteins, amino acids, pyrroles, pyridines, pyrazines, oxazoles, carboxylic acids, fatty acids, and phenolic compounds in an equilibrated proportion to obtain good body, acidity, and smoothness of the beverage (Clemente et al., 2018).

I. Bioactive Compounds

Bioactive compounds are minor food constituents that have biological functions other than nutritional functions. Coffee is a functional food due to its high content of bioactive compounds that exert stimulating, antioxidant, antimicrobial, anticarcinogenic, anti-inflammatory, antihypertensive, and other beneficial biological properties (Farah, 2012).

Caffeine is the most important bioactive compound found in coffee. It is primarily responsible for the physiologically active or stimulating effect of coffee drink & coffee-based beverages (Ivon, 2002). It also contributes 10 to 30% of the often desirable bitterness of the coffee drink (Clemente et al., 2018). A dry raw coffee bean contains between 0.8% and 2.8% caffeine, depending on its species, origin, and degree of maturation (e.g. 1 % in Arabicas and 2 % in Robustas) (Nicolas, 2004). Despite its low content, caffeine is the single most frequently determined compound in coffee. Caffeine is an alkaloid with the chemical formula C8H10O2N4·H2O and chemical name 1,3,7 trimethylaxanthine (Ivon, 2002). It consists of needle-shaped crystals with a melting point of 2360C (see the figure below).

Figure 2.8. Chemical structure of Caffeine (1,3,7 trimethylaxanthine) Source: Ivon (2002)

Roasting is reported to cause approximately 30% reduction in caffeine content or level of raw coffee beans (Franca et al., 2005). However, some studies suggest that the caffeine content of a

coffee bean remain unaffected (Cruz et al., 2012) and only a small percentage reduction occurs due to the organic weight loss during roasting process (Farah, 2012). Trigonelline (N-methylpyridinium-3-carboxylate) is the second most important alkaloid in coffee after caffeine (Rodrigues & Bragagnolo, 2013). The concentrations of trigonelline in green Arabica coffee beans ranges from 0.74 - 1.54% (w/w) (Duarte, 2010; Gichimu et al., 2014). Recently, it has received considerable attention in the field of coffee research (Clarke & Vitzthum, 2001) because it is one of the precursors of aroma compounds in coffee as shown in Figure 2.9 and also contributes to the bitterness of coffee brew (Ky et al., 2001; Farah, 2012; Rodrigues & Bragagnolo, 2013; Kalaska et al., 2014). The sensory importance of trigonelline is due to the fact that it undergoes extensive thermal degradation of 50% to 90% (Kalaska et al., 2014) and generating pyridines & pyrroles that make up approximately half of the volatile substances released during coffee roasting process (Rodrigues & Bragagnolo, 2013). Therefore, the lower the trigonelline content in coffee beans, the lower the quality of the coffee beverage (Farah et al., 2006).

Chlorogenic acids (CGAs) are the most abundant phenolic compounds found in coffee beans (Clarke & Vitzthum, 2001). They are esters of quinic acid and trans-hydroxycinnamic acid; and the most abundant chlorogenic acid is 5-O-caffeoylquinic acid or 5-CQA which represents approximately 75% of the total chlorogenic acids (Farah et al., 2006; Workneh, 2015). Chlorgenic acids are mainly responsible for the antioxidant activity of coffee (Bicchi et al., 1995). They also play an important role in the formation of typical roasted coffee flavor (Perrone et al., 2008). Their composition in green coffee beans depends mainly on genetic properties, agricultural practices, harvesting and postharvest processing methods and environmental growing conditions (Alonso- Soales et al., 2009). An evaluation of the published works on CGAs in relation to coffee quality indicates that high CGAs levels, mainly 5-CQA, seem to be associated to the presence of immature beans (Workneh, 2015).

During roasting, Chlorogenic acids are strongly degraded and decomposed in to acids such as quinic and caffeic acids by hydrolysis, lactones, and volatile compounds such as the phenil, guaiacol, and 4-vinyl guaiacol by pyrolysis (Bicchi et al., 1995), which contribute to the 66% of acidity or astringency of coffee beverage (Clemente et al., 2018). The products are formed through low temperature hydrolysis, and their formation depends on the water content within the coffee

bean (Bicchi et al., 1995). According to (Clifford, 2000), a 1% weight loss of coffee (dry matter basis) during roasting may bring up to 10% decrease in the original CGAs content. Progressive reduction in the antioxidant activity of coffee has been observed with roasting (Workneh, 2015), showing medium roasted coffee the highest activity, due to the balance between the degradation of phenolic compounds and the generation of melanoidins during the process (Castilho et al., 2002). II. Maillard reaction products (MRPs) Maillard reaction products (MRPs) such as acrylamide, melanoidins which are high molecular weight nitrogenous and brown-colored compounds with antioxidant activities, and phenolic compounds like pyrazines are formed by complex interactions between amino acids and carbohydrates during the roasting of coffee at high temperature. At higher temperatures cellulose undergoes hydrolysis to smaller sugar derivatives including glucose and levoclucosan (Ivon, 2002). Acrylamide (acrylic amide) is a synthetic monomer that is metabolized to the expoxide, glycidamide (2,3-expoxypropanmide) which has neurotoxic potential. The mechanism of acrylamide neutrotoxicity is due to glycidamide conversion at processing temperature of 120 °C or higher (Tin, 2006). Phenolic compounds have been described as contributors to the quality of coffee flavor with a smoky-like odour. Guaiacol is formed by a thermal degradation of ferulic acid, and the level increases with roasting. Alkyl pyrazines, such as 2-ethyl 5-methylpyrazine, have a key role, with nutty and coffee-like odours. Pyrroles, originating from the degradation of amino acids and Amadori intermediates, have floreal, rose-like, smoky flavours (Ivon, 2002).

III. Proximate Composition Another important components of coffee bean are carbohydrates that make up about half of its dry weight (Marilisa & Anese, 2018). They participate in extensive chemical changes associated with coffee roasting and brewing process. Carbohydrates contribute to the formation of the typical coffee aroma during roasting (Ivon, 2002). Sucrose is the main carbohydrate found in green coffee. The reducing sugars reacts with Asparagine during Maillard reaction to form coffee flavor, color and acrylamide as the intermediate product (Marilisa & Anese, 2018).

Protein and fat are also major components of coffee beans that influence the characteristic aroma of roasted coffee (Koskei et al., 2015). A dry coffee bean contains 8–17 % lipids, up to 11 %

proteins and amino acids, and ca. 5% minerals (Nicolas, 2004). The common acids found in coffee beans are citric, acetic, malic, chlorogenic and quinic acids (Butt et al., 2011). Water or moisture content is also a critical parameter for evaluation of green coffee quality, in the sense that it affects mold growth, mycotoxin production, fermentation, physical, chemical and sensory parameters (Clarke & Vitzthum, 2001). Its amount within the beans is important from a commercial point of view as water is quite cheap compared to coffee (Mendonça et al., 2007). Its amount is usually controlled by processing conditions, and, for green coffee, it is kept at about 12% to allow for safe transportation and storage (Clark, 2003). Immediately after roasting, coffee moisture content can be close to zero percent (Yeretzian et al., 2002), especially for dark roasts and when water quenching is not employed. However, actual roasted coffee moisture values can reach up to 3% since coffee beans tend to absorb water from the surrounding air (Baggenstoss et al., 2008).

IV. Volatile Organic Compounds

Almost a thousand (app. 800) volatile compounds are identified in roasted coffee (Yeretzian et al., 2003). And around 40 of them are responsible for the main impression of coffee aroma (Baggenstoss et al., 2008; Belitz et al., 2009). The presence and intensity of such volatile compounds in the final product depends on the time-temperature profile during roasting and is an index of quality (Bustos-Vanegas et al., 2017). The most common aroma compounds are acetaldehyde, acetone, 2-methylfuran, 2-methylbutanol, 2-methylpyrazine, furfural, 2-propanone, furfuryl alcohol, 2,5-dimethylpyrazine and furfuryl acetate (Moon et al., 2013). Roasting leads to several changes in the chemical composition and biological activities of raw coffee bean due to the transformation of naturally occurring polyphenolic constituents into a complex mixture of Maillard reaction products, as well as the formation of organic compounds.

Figure 2.9. Coffee Flavor Precursors Source: Yeretzian (2002) Sulfur compounds are also changed by oxidation, thermal degradation, and/or hydrolysis, and the vanillin content increases considerably during the roasting process (Nicolas, 2004). During roasting, the chemical components of raw coffee bean form the typical coffee flavor as shown in the table below.

Table 2.3. Chemical composition of dry roasted Coffea arabica and Coffea robusta beans

Concentration Concentration Component (g/100g of dry coffee bean) (g/100g of roasted coffee bean) Coffea Arabica Coffea robusta Coffea arabica Coffea robusta Carbohydrates Sucrose 6.0 – 9.0 0.9 – 4.0 4.2 1.6 Reducing Sugars 0.1 0.4 0.3 0.3 Polysaccharides 34 – 44 48 – 55 31-33 37 Lignin 3.0 3.0 3.0 3.0 Pectin 2.0 2.0 2.0 2.0 Nitrogenous Compounds Protein 10.0 – 11.0 11.0 – 15.0 7.5 – 10 7.5 – 10 Free amino acids 0.5 0.8 – 1.0 ND ND Caffeine 0.9 – 1.3 1.5 – 2.5 1.1 – 1.3 2.4 – 2.5 Trigonelline 0.6 – 2.0 0.6 – 0.7 0.2 – 1.2 0.3 – 0.7 Nicotinic Acid NP NP 0.016 – 0.026 0.014 – 0.025 Lipids Coffee oil 15.0 – 17.0 7.0 – 10.0 17.0 11.0 Diterpenes 0.5 – 1.2 0.2 – 0.8 0.9 0.2 Minerals 3.0 – 4.2 4.4 – 4.5 4.5 4.7 Acids & Esters Chlorogenic acids 4.1 – 7.9 6.1 – 11.3 1.9 – 2.5 3.3 – 3.8 Aliphatic acids 1.0 1.0 1.6 1.6 Quinic acid 0.4 0.4 0.8 1.0 Melanoidins NP NP 25 25 Note that NP means not present while ND means not detected. Source: Clarke & Vitzthum (2001) 2.4.2. Specific Roasting Conditions

The final flavor of coffee is heavily dependent on how coffee beans are roasted, i.e. the specific roasting conditions such as roasting time, temperature profile, speed of air, design of roasting technology which includes the type of heat source, mode of heat transfer and degree of automation, and the skill of individual operating the roasting machine. Above all, roasting time and temperature strongly influence the physicochemical transformation of the coffee beans (Wang, 2012) and consequently affect the bioactivity and flavor of the beverage (Marilisa & Anese, 2018). When a coffee spends too long in the roaster without enough heat, its sweetness and fragrance ebb away (Nagaraju et al., 1997). Thus, it is important to control the roasting time and temperature so that

they are sufficient for the required chemical reactions to occur, without burning the beans and compromising the flavor of the final beverage (Kelly, 2018).

In conventional roasting process, the temperature ranges from 170 to 230oC while the time ranges from 10 to 20 minutes (Yeretzian et al., 2002; Sinnott, 2010). However, these values can vary depending on the degree of roast required (i.e. light, medium or dark), on the type of roaster used, and also on the variety, age, moisture content, etc. of the coffee beans (Jokanović et al., 2012). The control of temperatures and duration of roasting, in industry, are only effective if the quality of the raw coffee does not vary (Jokanović et al., 2012).

Light-roasted coffees may contain relatively higher amounts of acrylamide than very dark roasted beans due to the fact that acrylamide is formed at the beginning of the roasting step, declining then steeply towards the end of the roasting process as shown in the figure below (Summaa et al., 2007). The maximum antioxidant activity of coffee is obtained for medium roasted coffee (Bae et al., 2014). When coffee beans are roasted at higher temperature, the moisture content and total acidity of roasted coffee decrease whereas the pH and weight loss (%) increase (Moon et al., 2013).

Figure 2.10. Effect of roasting time on acrylamide concentration in coffee Source: Kocadağlı et al. (2012)

The final quality of roasted coffee is influenced by the design of the roasters (Wang, 2012). However, whatever type of coffee roasting technology and individual skill cannot make up for poor quality coffee beans (Kelly, 2018).

Figure 2.11. Roast Profile of Coffee (for Tanzanian Coffee Bean)

Source: Uman et al. (2016) Although the cup quality of coffee brew depends on many factors, mainly the quality of the roasted coffee beans, the grind size (degree of fineness or coarseness) of coffee powder, the extent of heat or boiling and the type of water used for brewing (Wang, 2012), a typical chemical composition of 100 ml coffee brew from medium roasted coffee is shown in the figure below.

Figure 2.12. Typical composition per 100 ml coffee brew from medium roasted coffee

Source: Clarke & Vitzthum (2001)

Chapter Three 3. Materials and Methods

3.1. Raw Material Collection, Transportation and Preparation

3.1.1. Raw Material Collection Three samples of unwashed (sun dried) Ethiopian coffee beans of Harar, Sidama, and Yirgacheffe variety were selected for the study. Each sample weighs about 1 kg and was freshly harvested from the 2017/18 crop year. The unwashed Harar coffee beans sample was collected from coffee farmers in Eastern Hararghe district of Harar Region. Whereas samples of the unwashed Sidama and Yirgacheffe coffee beans were collected from coffee farmers in Sidama zone and Gedeo zone of Southern Nations Nationalities & Peoples Region (SNNPR), respectively. The three coffee varieties were selected for the study due to their specialty nature (i.e. unique sensory characteristics and high cup quality). In addition, the samples are well known for their high market value or contribution to the export market and availability in the country (Nicolas, 2004). In addition, all the samples were chosen to be unwashed or dry processed for wholesome coffee quality. Table 3.1. Agronomical Practices of the Obtained Coffee Samples

Sample Code 01 02 03 Coffee Variety Yirgacheffe/ Yirgacheffee Harar Sidama/ Sidamo Coffee Species Arabica Arabica Arabica Region, Country SNNPR (South), Ethiopia Harar, Ethiopia SNNPR (South), Ethiopia Zone Gedeo Hararge Sidama District (Woreda) Kochere Eastern Hararghe Bonna Zuria Altitude (asl) 1770 – 2200m 1510 – 2120m 1400 to 2200m Annual Temperature 15 – 35oC 15 – 35oC 15 - 30oC Annual Rainfall 1000 – 2000mm 1000 – 2000mm 1200 – 2000mm Harvesting Season September to January September to January September to January Growing Practice Organic, Tree Shed-Garden Organic, Tree Shed-Garden Organic, Tree Shed-Garden

Processing Type Unwashed (Sun dried) Unwashed (Sun dried) Unwashed (Sun dried)

Source: CLU (2017)

3.1.2. Raw Material Transportation and Storage

The collected sample of raw coffee beans were aseptically transported in food grade polyethylene bags and stored at 20°C and 70% RH in air conditioned coffee storage facility found in the Ethiopian Coffee and Tea Authority Coffee Quality Inspection and Certification Center (ECTACQICC) in Addis Ababa, Ethiopia.

3.1.3. Raw Material Preparation

3.1.3.1. Roasting of Sample Coffee Beans

A. Drum Roaster

A four barrel drum roaster (BRZ 4, Probat®, USA, 2009) found at the coffee laboratory of ECTACQICC was used to roast the sample raw coffee beans selected for the study. Each barrel has a capacity of roasting about 100g of raw coffee beans. The machine uses LPG gas as a fuel and electricity to operate controlling devices, motors, temperature sensors and display devices. It was operated and controlled by a professional coffee roast master at the coffee laboratory found in ECTACQICC.

First, the minimum temperature of roast was set at 150oC for light degree of roast. And the maximum roasting temperature was limited to 200oC in order to ensure a dark roast at the end of each roasting process. The machine was heated by supplying an LPG gas from a cylinder through a hose and igniting it by electric switch. Then, the flames from the LPG gas directly heated a group of rotating metal rods which are found under each drums of the machine. The heated rods in turn rotated and transferred the heat to the horizontal cylinders (drums) by conduction. The drum roaster took about 3 minutes to reach around 150oC at which the raw coffee samples were separately fed in to the four drums. Immediately following this, the temperature inside the drum roaster begun to drop and reached to about 90oC after a minute because the coffee beans absorbed the heat from the drum through conduction. But after 2 minutes, the temperature of the drum roaster started to rise again and the coffee beans smelled like freshly mowed grass since they lost much of their moisture content.

Figure 3.1. Drum Type Sample Roaster (BRZ 4, Probat®, USA, 2009) (Seen in the Photo: Mr. Seifedin, a Professional Cupper at ECTACQICC Addis Ababa) Shortly after a minute, the temperature of the coffee beans reached around 150oC and the color of the beans changed to light brown, i.e. light degree of roasted. After another minute, the temperature reached around 170oC where the characteristic coffee aroma begun to develop and the color of the coffee beans became brown (i.e. light medium degree of roast). At this point, the first crack sound was heard. Immediately after few seconds, the roasting process became very dynamic and the color of the coffee beans changed quickly from brown to dark brown (i.e. medium degree of roast).

At around 200oC & 10 minutes from the start of the roasting process, the coffee beans started to become very dark brown in color (i.e. dark degree of roast) and smoky burnt aroma was sensed. A second crack sound was heard after few seconds and the coffee beans begun to look shiny and burnt in color (i.e. very dark degree of roast). To guarantee the required degree of roast, certain number (e.g. 10 or 20 beans) of sample coffee were manually removed every 20 seconds using a special tool called “Tryer” or “Sampler Spoon” until the roasting process is complete (Bustos- Vanegas et al., 2017). After 2 minutes, the roasting machine’s heat was turned off by the operator and the roasted coffee beans were quickly cooled using a fan that surges surrounding air down through perforated metal trays located under the machine to halt the exothermic process. Finally, the roasted coffee beans were transferred to coffee trays to allow sufficient time for CO2 degassing before grinding operation.

B. Fluidized Bed Roaster

About 50g capacity fluidized bed roaster (V3, IKAWA® Pro, UK, 2018) found at Vidaya’s Coffee PLC was used to study the effect of fluidized bed roaster on the sensory quality and bioactive content of the sample specialty coffee beans. The roaster works by blowing hot air through a chamber of beans from below, levitating and roasting them.

Figure 3.2. Fluidized Bed Sample Roaster (V3, IKAWA® Pro, UK, 2018) First, the coffee roast profile was created on a smartphone using an IKAWA® Pro application by setting the drop or minimum temperature to 150oC, maximum temperature to 200oC and time of roast to 7, 8, & 9 minutes for light, medium, and dark degree of roast, respectively. Then the roast profiles was easily transferred from the smartphone to the roaster by Bluetooth connection.

Figure 3.3. Coffee Roast Profile Displayed on a Smartphone

Then the sample raw coffee beans were filled on about 50g capacity cylindrical compartment found on top of the machine. In just 2 minutes, the machine heated up and reached 150oC. This was indicated by a green signal from the roaster. Hence, the coffee beans were fed in to the machine by turning cylindrical compartment counter clockwise. Following this, the coffee roasting process was started. Since the machine is automatic, it makes the necessary adjustments to the roasting process by lowering the airflow when the beans get less dense and increasing the heat gently. The thermometer readings from within the roaster were translated into real-time curves that clearly illustrated the course of the roast by a built-in software and displayed on the smartphone.

When the maximum roasting temperature of 200oC was reached, the machine stopped automatically and triggered the cooling fan that ended the roast with cool air blowing through the browned beans. The flaky chaff that blew off the coffee beans was caught by a glass cup below the machine and disposed manually. Then the roasted coffee beans were collected in a clean glass cup and ready to be brewed. The entire roast profile was saved on the smartphone in order to easily repeat the above roasting conditions for each sample that was subjected to the same degree of roast (Diallo, 2019).

C. Traditional Oven Top Roaster

About 50g capacity electric pizza pan (Maxima®, China, 2016) was used to roast the collected sample raw coffee beans. It was manually modified and wired in with a digital high temperature thermocouple to monitor and display the second-by-second temperature of the coffee beans during the roasting process.

Figure 3.4. A Traditional Coffee Roaster Manually Wired in with Thermocouple 31

3.1.3.2. Cooling of the Roasted Coffee Beans

Finally, the beans were air cooled after roasting in order to halt exothermic reaction. And the roasted coffee beans were allowed to stay overnight (for at least 8 hours after roasting) in order to allow sufficient time for full flavor development and CO2 degassing.

3.1.3.3. Grinding of the Roasted Coffee Beans

The roasted coffee beans were grinded at medium grind size of about 850µm using electronic laboratory scale coffee grinder (VTA6S, MAHLKÖNIG GmbH & Co., Hamburg, Germany, 2009) in a coffee laboratory found at ECTACQICC. The grinder was cleaned well after grinding each to protect mixing up one form another (Sualeh & Mekonnen, 2015). The roasted and ground coffee samples were separately labelled, packed in air tight odor free food grade polyethylene bags and stored in air conditioned raw material storage room found in the Agricultural Quality Research Laboratory (AQRL) at the head quarter of Ethiopian Institute of Agricultural Research (EIAR).

Figure 3.5. Laboratory Coffee Grinder (VTA6S, MAHLKÖNIG GmbH & Co., Germany, 2009) 3.1.3.4. Brewing of Sample Coffees for Cupping

Coffee brew was prepared from each sample of roasted & ground coffee according to CLU (2007) and ECX (2009) standard cupping protocol (Sualeh & Mekonnen, 2015). In order to create a truly blind cupping conditions and minimize bias, each of the three coffee varieties were assigned a unique sample code that represents the coffee varieties, type of roasters and degree of roasts as shown in the table below.

Table 3.2. Description of the Sample Coffee Codes

Sample Code Description 01 Raw or Green Yirgacheffe Coffee 02 Raw or Green Harar Coffee 03 Raw or Green Sidama Coffee 01DL Yirgacheffe Coffee Light Roasted on Drum Roaster 01FL Yirgacheffe Coffee Light Roasted on Fluidized Bed Roaster 01TL Yirgacheffe Coffee Light Roasted on Traditional Oven Top Roaster 01DM Yirgacheffe Coffee Medium Roasted on Drum Roaster 01FM Yirgacheffe Coffee Medium Roasted on Fluidized Bed Roaster 01TM Yirgacheffe Coffee Medium Roasted on Traditional Oven Top Roaster 01DD Yirgacheffe Coffee Dark Roasted on Drum Roaster 01FD Yirgacheffe Coffee Dark Roasted on Fluidized Bed Roaster 01TD Yirgacheffe Coffee Dark Roasted on Traditional Oven Top Roaster 02DL Harar Coffee Light Roasted on Drum Roaster 02FL Harar Coffee Light Roasted on Fluidized Bed Roaster 02TL Harar Coffee Light Roasted on Traditional Oven Top Roaster 02DM Harar Coffee Medium Roasted on Drum Roaster 02FM Harar Coffee Medium Roasted on Fluidized Bed Roaster 02TM Harar Coffee Medium Roasted on Traditional Oven Top Roaster 02DD Harar Coffee Dark Roasted on Drum Roaster 02FD Harar Coffee Dark Roasted on Fluidized Bed Roaster 02TD Harar Coffee Dark Roasted on Traditional Oven Top Roaster 03DL Sidama Coffee Light Roasted on Drum Roaster 03FL Sidama Coffee Light Roasted on Fluidized Bed Roaster 03DM Sidama Coffee Medium Roasted on Drum Roaster 03FM Sidama Coffee Medium Roasted on Fluidized Bed Roaster 03DD Sidama Coffee Dark Roasted on Drum Roaster 03FD Sidama Coffee Dark Roasted on Fluidized Bed Roaster

3.2. Experimental Framework of the Research

Figure 3.6. Experimental Framework of the Research Work

3.3. Methods of Analysis 3.3.1. Physical Examination of Coffee Beans

The raw sample coffee beans were physically examined for defect, odor quality, moisture content, size and shape. This examination was the first part of the preliminary test performed to grade the samples coffee beans for specialty coffee cupping. It represented 40% of the preliminary test conducted in the coffee laboratory found at ECTACQICC. 3.3.1.1. Coffee Bean Defect Count

Primary defects count was made for full black, full sour, fungus attacked beans, foreign matter, severe insect damage and coffee pod/ husk. And secondary defect observations were made for partial black, partial sour, floater, immature, withered, shell, slightly insect damaged, broken and soiled coffee beans.

3 4 1 2

5 6 7 8 9

10 11 12 13 14

Figure 3.7. Primary and Secondary Coffee Bean Defects Where: 1. Black; 2. Sour; 3. Fungus Attacked; 4. Insect Damaged; 5. Foreign Matters (Stones & Sticks) 6. Dried Cherries; 7. Husk; 8. Hull; 9. Parchment Coffee; 10. Floaters; 11. Stinkers; 12. Immature; 13. Malformed; 14. Shells

Source: Franca & Oliveira (2008)

3.3.1.2. Coffee Bean Odor Quality Evaluation

Odor quality assessment was conducted for clean, fairly clean, trace, light, moderate, and strong coffee odor by trained professionals using their sense of smell.

3.3.1.3. Sieve Analysis

The coffee beans were size graded using special set of screens (Pinhalense®, Brazil) numbered from 8 to 20 according to the diameter of the holes in millimeter. Sieves with rectangular holes were used for regularly shaped beans while sieves with circular holes were used for round peaberries. The diameters of the sieve holes ranges from 8/64 to 20/64 inches (Franca & Oliveira, 2008). The sieves were arranged in 1/64 inch increment as shown in the figure below. The percent of coffee beans retained on screen size of 14 (i.e.14/64 of an inch diameter) were recorded. And the coffee beans were repacked in food grade plastic bags and stored at 20oC and 70% relative humidity for later analysis (Sualeh & Mekonnen, 2015).

Figure 3.8. Sieve Analysis of Coffee Beans The shape of each sample coffee beans was visualized manually by trained coffee professionals at the coffee laboratory.

3.3.1.4. Color Analysis of Roasted Coffee Beans

An adequate degree of roast is required for the sample coffee beans to be proper for grinding and making coffee with pleasant sensory quality (Beattie, 2012). It was analyzed based on the color of different roast level (i.e. light, medium and dark) for roasted coffee beans (whole or ground) using a handheld Coffee Roast Analyzer (JAV-RDA-H, JAVALYTICSTM, Madison Instruments Inc., Australia, 2012) according to the Specialty Coffee Association of America (SCAA) Agtron roasting degree or color reference scale with the assistance of a trained professional (Sualeh & Mekonnen, 2015). 36

JAV-RDA-H degree of roast analyzer is a hand-held device weighing just 250g which has a user friendly interface embedded with Linux operating software and equipped with 1.8” Color LCD screen display, Infra-red sensor, and Infra-red LED light powered by Li-polymer 1200mAh battery. It is designed to measure the degree of roast of both whole and ground coffee (Paul, 2012).

Figure 3.9. Measuring Color Value of Roasted Coffee Beans using Roast Analyzer First, calibration of the equipment was performed before each measurement by placing the bottom part of surface of the equipment’s measurement hood over the top surface of the reference tile. Then, the whole roasted coffee beans sample was poured into the sample dish. A straight edge was used to form a flat surface that is level with edge of the dish. When the sample preparation is complete, the degree of roast was measured by placing the bottom of the equipment’s measuring hood on top of the sample dish. The result was instantly displayed. Three readings were performed to average the variations often found in whole bean samples. Calibration of the equipment was performed before each measurement of proceeding samples in order to avoid equipment reading errors.

Table 3.3. SCAA Color Value of Roasted Coffee Beans

SCAA Gourmet Result Common Names SCAA Names 97 ≤ 퐺표푢푟푚푒푡# < 100 Light- Extremely Light- 93 ≤ 퐺표푢푟푚푒푡# < 97 Light Extremely Light 90 ≤ 퐺표푢푟푚푒푡# < 93 Light+ Extremely Light+ 87 ≤ 퐺표푢푟푚푒푡# < 90 Cinnamon- Very Light- 83 ≤ 퐺표푢푟푚푒푡# < 87 Cinnamon Very Light 80 ≤ 퐺표푢푟푚푒푡# < 83 Cinnamon+ Very Light+ 77 ≤ 퐺표푢푟푚푒푡# < 80 Medium- (Light Medium) Light- 73 ≤ 퐺표푢푟푚푒푡# < 77 Medium Light 70 ≤ 퐺표푢푟푚푒푡# < 73 Medium+ (Medium Dark) Light+ 67 ≤ 퐺표푢푟푚푒푡# < 70 Dark- (Slightly Dark) Medium Light- 63 ≤ 퐺표푢푟푚푒푡# < 67 Dark Medium Light 60 ≤ 퐺표푢푟푚푒푡# < 63 Dark+ (Very Dark) Medium Light+ 57 ≤ 퐺표푢푟푚푒푡# < 60 City- Medium- 53 ≤ 퐺표푢푟푚푒푡# < 57 City Medium 50 ≤ 퐺표푢푟푚푒푡# < 53 City+ Medium+ 47 ≤ 퐺표푢푟푚푒푡# < 50 Full City- Moderately Dark- 43 ≤ 퐺표푢푟푚푒푡# < 47 Full City Moderately Dark 40 ≤ 퐺표푢푟푚푒푡# < 43 Full City+ Moderately Dark+ 37 ≤ 퐺표푢푟푚푒푡# < 40 French- Dark- 33 ≤ 퐺표푢푟푚푒푡# < 37 French Dark 30 ≤ 퐺표푢푟푚푒푡# < 33 French+ Dark+ 27 ≤ 퐺표푢푟푚푒푡# < 30 Italian- Very Dark- 23 ≤ 퐺표푢푟푚푒푡# < 27 Italian Very Dark 0 ≤ 퐺표푢푟푚푒푡# < 23 Italian+ Very Dark Where: “ – ” & “ + ” indicate lower and higher degree of roast, respectively. Gourmet# means color value according to Specialty Coffee Association of America (SCAA). Source: Paul (2012)

3.3.1.5. Determination of Moisture Content

The moisture content of raw coffee beans was measured using Grain Moisture Analyzer (GMM, Draminski®, Ukraine, 2014). The final moisture content of roasted coffee beans was measured by oven drying at 103oC for 12 hours according to the AOAC method 925.10 (AOAC, 1990). 3.3.2. Cupping of Roasted Coffee Beans

I. Preliminary Coffee Cupping

The second part of the preliminary coffee assessment was conducted out of 60% for the cup quality of the sample raw coffee beans which were roasted at medium degree of roast using 100g capacity LPG gas burning drum roaster (BRZ 4, Probat®, USA, 2009) found at the coffee laboratory in ECTACQICC. The cup quality evaluation or cupping of the sample coffees was performed by three professional coffee cuppers, namely Mrs. Segenet, Mr. Seifedin and Mr. Tefera, at the coffee laboratory. The evaluation of the sample coffees includes assessments for cup cleanness quality such as clean, fairly clean, and defective cups, assessment for acidity intensity such as pointed, moderately pointed, medium, light and lacking acidity, assessment for the coffee bean body quality like full, medium full, medium, light and thin body, assessment for flavor quality such as good, fairly good, average, fair and common flavor. Finally, the raw and cup value were summed up to obtain the total value out of 100% to determine whether the sample coffees qualify for specialty cupping.

II. Specialty Coffee Cupping

A sensory panel or team comprised of three experienced and certified professional cuppers carried out a blind cupping or assessment of basic cup quality attributes of coffee brew from each sample using a cupping protocol shown in Appendix 3. These include aromatic intensity, aromatic quality, color, acidity, astringency, bitterness, body, flavor and overall acceptability. Samples were separately coded to avoid individual biasness of the panel. Standard 180ml coffee testing cups were used. And each panelist was provided with a standard cupping form to give his/her independent judgment for each sample according to CLU (2007) and ECX (2009) method (Sualeh & Mekonnen, 2015). The effects of other variables that could have affected cup quality results were reduced by: - Subjecting the roasting of each sample to the skill set of a professional coffee roast master.

- Making every situations to be as consistent as possible across roasts by using “equivalent” profiles on all the roasting machines or technologies. The drop temperatures, maximum roasting temperature and development time ratios were consistently 150oC, 200oC and 15 minutes, respectively. - Roasting all samples 24 hours prior to cupping or sensory evaluation. - Conducting the coffee sensory evaluation by a cupping panel comprised of three professionals who have extensive (i.e. more than 2 years) experience of cupping specialty coffee beans and practice cupping at least 3-5 days per week. A mean of the cupping scores given for each coffee samples by the three professionals was approved by the quality supervisor and head of the coffee laboratory facility. 3.3.3. Determination of Bioactive Compounds

The caffeine, trigonelline and total chlorogenic acids (5-CQA) content of the roasted and ground coffee samples were simultaneously determined using a high-performance liquid chromatography method adopted from (Vignoli et al., 2014) at the Agricultural Quality Research Laboratory of the Ethiopian Institute of Agricultural Research (EIAR) according to the test methods ES ISO 10095 (ESA, 2001) and ES ISO 4052 (ESA, 2002b) described by the Ethiopian Standard Agency.

Sample Extraction Method

Around 0.5g of grinded sample coffee was weighed using precision scale and transferred into 50ml conical flask labelled according to the sample code. Then 40ml hot distilled water at 95oC was poured into the flask and the resulting coffee solution was boiled for 20 minutes on a heater. The solution was stirred using magnetic stirrer dropped in to the flask. After 20 minutes, the flask was carefully removed from the heater and boiling distilled water was poured up to the 50ml mark to make up for the evaporated water. Then the coffee extract was separately filtered using a 0.45µm Whatman filter paper and the supernatant was collected in 50ml beaker labelled according to its code. All the raw and roasted sample coffees were extracted according to the above method and the extracts were carefully transferred into clean HPLC vials through a 0.22µm PTFE syringe filter (Waters, Milford, Massachusetts, USA) to avoid injection of air bubbles into the HPLC system.

Sample Injection

The vials were transferred in to the auto sampler chamber of the HPLC system and the corresponding sample code was correctly fed into the HPLC computer prior to injection and running the system.

Equipment Setting and Operating Condition

Chromatographic analysis of alkaloids or bioactive compounds in the coffee extracts was performed using a High Performance Liquid Chromatography system (Agilent 1260 Infinity II, USA) with a diode array detector (DAD). Chromatographic separation was made on a normal phase C18 column (4.6 x 100mm, 5μm), with the column temperature held at 25oC. The mobile phase consisted of 5% aqueous acetic acid as solvent A and 100% acetonitrile as solvent B. The chromatography was run under isocratic condition by using 4:96 percent acetic acid (A) to acetonitrile (B) solvent ratio at a flow rate of 0.7 mL min-1. Table 3.4 HPLC Equipment Operating Condition

Item Experimental Condition Mobile phase composition 96% Acetonitile, 4% Acetic Acid Flow rate 0.7ml/min Injection Volume 10µl Column Temperature 25oC Wave length 272nm

The column was washed and reconditioned for 3 min in between each analysis using isopropanol. Chromatographic data for caffeine, trigonelline and chlorogenic acids was collected at 272 nm and converted to chromatograms using the instrument built-in software (Masslynx 4.1 SCN 714; Waters, USA). Peak identification was performed by comparing the retention times and peak areas of the analytes with that of pure standard solutions. The water used for all chromatographic analyses was distilled and all solvents were filtered by using 0.22 μm syringe filter before starting the chromatographic analysis. Quantitative analysis of the alkaloids was performed by using calibration curves constructed from series of standard solutions of the alkaloids and chromatographic peak areas.

3.3.4. Determination of Acrylamide Concentration

Several methods used to determine the acrylamide concentration in food are usually based on either GC-MS or LC-MS and are not suitable for coffee products due to their inability to avoid the presence of interferences that co-elute with the analyte and prevent its correct quantification (Delatour et al., 2004). The acrylamide concentration in the roasted coffee samples was determined using UV-Vis spectrophotometer following a validated analytical method according to (Soares et al., 2006). The method was based an improved sample preparation with two main purification. First, the samples were treated with ethanol and Carrez reagents in order to precipitate polysaccharides and proteins, respectively, and second with a layered solid extraction phase (SPE) to eliminate the main interferences (Soares et al., 2006).

Sample Preparation (Extraction) Method About 2g of the fine ground coffee sample was analytically weighed using precision scale and transferred into clean dry 50ml conical flask labelled corresponding to the sample. Then, the sample was dissolved by adding 10ml distilled water and 15ml absolute ethanol in to the flask. The resulting solution was thoroughly mixed/ shaken for about 1 minute. The extract was filtered through a 0.45µm Whatman filter paper and collected in 50ml beaker labelled corresponding to the sample. Then, the extract was acidified with acetic acid until pH 4 – 5. Carrez reagent I was prepared by analytically weighing 15g of potassium hexacyanoferrate (II) trihydrate, transferring it in to 100ml beaker, diluting it in water and adjusting the volume to the mark. The beaker was then labelled as Carrez I. And Carrez reagent II was prepared by analytically weighing 30g of zinc sulphate heptahydrate, transferring it in to 100ml beaker, diluting it in water and adjusting the volume to the mark. The beaker was then labelled as Carrez II. Then 1 ml of Carrez I and 1 ml of Carrez II reagent were added into to the beaker containing the acidified collected supernatant. The solution was transferred into 15 ml centrifugal flask and stored at 4oC for 15 minutes. Then the solution was centrifuged at 15000 g (15 minutes, 4oC). Next, the centrifuged solution was filtered through 0.22µm polyethylene filters and the filtrate was collected in 40ml vial labelled according to the sample code. The solution was reduced in volume to 2-3ml in rotary evaporator at 55oC and transferred to 20ml vial. The rotary vessel was rinsed with 2ml of water that was added to the solution in the vial.

SPE Clean-up Method The SPE columns were prepared by removing plungers from syringes and connecting the bottom (luer tip) with a polyethylene frit that has 10 µm pore size (Sep-Pak® Classic C18, Milford, Massachusetts, USA). The columns were first conditioned with 20 mL of methanol and then with 20 mL of water. The sample solution (~5 mL) was slowly loaded on the column cartridge using a pipet at a rate of 1 – 3 drops/sec and allowed to pass through tube with only gravity flow. The first eluate (~2 mL) was discarded and the remaining solution (~3 mL) was collected in 15 mL vial. The column was washed with 10ml of water and collected in the above 20ml vial. The sample analyte was eluted from the column with just 1 mL of methanol and collected in the vial to be kept in refrigerator at 4oC for later UV-Vis analysis. The column was once again conditioned with 20 mL of methanol and washed with 20 mL of water for SPE clean-up of the next samples.

Standard Preparation Method A 2gL-1 (2000ppm) stock solution of acrylamide (AA) was prepared by weighing 0.2g of Acrylamide (AA) and dissolving it using 100mL 1:1 mixture of distilled water and absolute ethanol in a beaker. The prepared stock solution was serially diluted with the 1:1 mixture of distilled water and absolute ethanol to produce working standard solutions of 0.4ppm, 1ppm, 2ppm, and 4ppm acrylamide concentration. The prepared stock and working standard solutions were stored in refrigerator at 4oC for later analysis.

UV-Vis Equipment Setting and Operation Method The instrument used for acrylamide analysis was UV-Vis Spectrophotometer (Agilent, Cary 300, Version 12.00, USA). The Wavelength range was set to 700 - 200nm and the Baseline was set to zero. Then two clean quartz cuvettes were filled with a 1:1 ratio mixture of absolute ethanol and distilled water. And they were placed in the reference and sample slot. Then the sample beam was blocked by closing the cover on top of the unit. Next, the blank sample in the sample slot was removed and replaced with the standard sample. A background scan was performed and the maximum absorbance was recorded at a wavelength of 220nm. Then the coffee sample solutions were scanned one after the other by removing previously scanned samples and changing the sample name. And the results were recorded in a computer.

3.3.5. Experimental Design and Statistical Analysis

A three factorial experimental design at three levels of treatments was used for the study. The selected factors were (A) Coffee Varieties, (B) Types of Coffee Roasters and (C) Degree of Roast. Three coffee varieties, namely Yirgacheffe, Harar and Sidama coffee, were roasted using three types of coffee roasters, i.e. “Drum Roaster”, “Fluidized Bed Roaster” and “Traditional Oven Top Roaster”, at three degree of roasts (i.e. Light, Medium and Dark Roast).

The experimental design was completely randomized, with a maximum of 3 observations or replication. The obtained experimental results or data were analyzed and interpreted by using analysis of variance (ANOVA) at a level of 5% significance (p < 0.05) using Design Expert software version 6.0.8.

Chapter Four 4. Results and Discussion

4.1. Physical Examination of Coffee Beans

4.1.1. Sieve Analysis

The size of the raw sample coffee beans were evaluated by three coffee professionals using sieve screens and weighted according to (Sualeh & Mekonnen, 2015). The amount of sample coffee beans retained in screen size 14 & above were weighed using electronic balance. And the results were reported in the preliminary unwashed coffee quality assessment sheet shown in Appendix 1. All the coffee samples obtained greater than 85% result, with 93% of Yirgacheffe, 95% of Harar, and 95% of Sidama coffee beans retained on screen size 14 & above. This indicated excellent uniformity in bean size of the sample coffees. And this uniformity reduced uneven heat transfer during roasting process. 4.1.2. Color Analysis

The color of each roasted sample coffee bean was reported using the Agtron color scale of the Specialty Coffee Association of America (SCAA). According to the equipment readings, the color value or SCAA Gourmet of the roasted sample coffee beans ranged from 97.8 to 25.1, i.e. “Extremely Light” to “Very Dark”. However, the SCAA names were converted to common names according to the conversion scale shown in Table 3.3 in order to incorporate them with the conventional naming of degree of coffee roast in many countries. From the obtained result, it was concluded that the coffee beans showed lighter degree of roast based on SCAA Agtron color scale and should be roasted more slowly to obtain darker degree of roast, especially on drum type of roaster.

Table 4.1. SCAA color value of Sample coffee beans roasted at different conditions

No. Sample Instrument Reading Common Name Code Color Value Color Name 1 01DL 90.3 Extremely Light Light 2 01DM 62.7 Medium Light Medium 3 01DD 56.6 Medium Medium 4 01FL 81.8 Very Light Light 5 01FM 68.2 Medium Light Medium 6 01FD 33.5 Dark Dark 7 01TL 82.9 Very Light Light 8 01TM 70.1 Light Medium 9 01TD 46.8 Moderately Dark Medium Dark 10 02DL 89.1 Very Light Light 11 02DM 91.6 Extremely Light Light 12 02DD 37.7 Dark Dark 13 02FL 97.8 Extremely Light Light 14 02FM 66.1 Medium Light Medium 15 02FD 38.9 Dark Dark 16 02TL 81.7 Very Light Light 17 02TM 75.3 Light Medium 18 02TD 25.1 Very Dark Dark 19 03DL 90.8 Extremely Light Light 20 03DM 87.4 Very Light Light 21 03DD 43.0 Moderately Dark Dark 22 03FL 94.6 Extremely Light Light 23 03FM 65.0 Medium Light Medium 24 03FD 38.2 Dark Dark 1. 01DL, 01DM, 01DD, 02DL, 02DM, 02DD, 03DL, 03DM, and 03DD are drum roasted sample coffees. 2. 01FL, 01FM, 01FD, 02FL, 02FM, 02FD, 03FL, 03FM and 03FD are fluidized bed roasted coffees. 3. 01TL, 01TM, 01TD, 02TL, 02TM and 02TD are sample coffees roasted by traditional roaster. 4. The results are expressed as the mean value of three measurements.

4.1.3. Moisture Analysis

According to obtained results shown in the table below, the moisture content of the raw sample coffee beans were in the optimum range of 8 – 12% (ECX, 2015) and can be stored at room temperature without growing fungus or mold. Although, there is small difference among the moisture contents of the three coffee varieties, climate and weather conditions of the sampling sites, the drying time during processing, varying storage conditions and way of handling during transportation might be one of the reasons for the moisture variation of the study samples.

Table 4.2. Moisture Content (%) of Sample Raw Coffee Beans

Green (Raw) Coffee Beans Moisture Content (%) Yirgacheffe 9.8 ± 0.1 Harar 10.0 ± 0.1 Sidama 10.3 ± 0.1 1. Results are expressed as the average of triplicate samples with mean ± 0.1 SD.

Further, the changes in the moisture content of the coffee beans were noticeable during the roasting process. The moisture content of Yirgacheffe coffee beans roasted at different degree of roast using the three type of coffee roasters is presented in Table 4.3 below.

Table 4.3. Moisture Content of Yirgacheffe Coffee Roasted at Different Conditions

Type of Roaster Degree of Roast Light Roast Medium Roast Dark Roast Drum 3.410 3.204 1.339 Fluidized Bed 3.477 1.884 1.311 Traditional 3.414 1.971 1.323 1. Results are expressed as the average of duplicate samples with mean ± 0.02 SD. 2. Values in the same column or raw are significantly (p<0.05) different by single Analysis of Variance (ANOVA).

According to analysis of variance, the moisture content of the roasted sample coffee beans showed significant (P<0.05) difference among the three degree of roasts and types of roasting technologies. The results ranged from 1.31% to 3.48%, being higher for lighter roast degrees and lower for darker roast. The least moisture content obtained at dark degree of roast using fluidized bed roaster may be due to a more moisture loss enhanced by hot air circulation.

4.2. Cupping of Roasted Coffee Beans

The three sample coffees were cupped at only medium degree of roast according to ECX (2015). The results were expressed as a mean value of three cupping scores given by three professional coffee cuppers at ECTACQICC coffee laboratory. Based on the cup scores given for coffee beans roasted by drum roaster, Sidama coffee obtained the highest 85.25% score followed by Yirgacheffe coffee with 83.17% score and Harar coffee with 82.33% score. This is because Sidama has got better sensory score in terms of fragrance, after taste, acidity and body. For coffee beans roasted using fluidized bed roaster, Yirgacheffe coffee obtained the highest 86.17% score followed by Sidama coffee with 84.5% score and Harar coffee with 81.25% score. This is because Yirgacheffe coffee showed remarkable sensory score in terms of flavor, after taste, balance and overall taste.

Only Harar and Yirgacheffe coffee beans were roasted on the traditional oven top roaster. According to the cupping test performed by three professional coffee cuppers at ECTACQICC, Yirgacheffe coffee obtained a higher average sensory score of 82.5% and Harar coffee with 80% score. Although Harar scored better for fragrance and flavor, Yirgacheffe showed remarkable sensory score in terms of after taste, acidity, body, balance and overall cup quality.

Among all the coffee varieties, Yirgacheffe coffee showed the highest cupping score using fluidized bed roaster. Whereas the highest cupping scores for Harar and Sidama Coffees were obtained by roasting them using drum roaster. However, the cup characteristics values of fragrance, flavor, after taste, acidity, body, balance, uniformity, cup cleanliness, sweetness and overall acceptability showed no significant (P > 0.05) difference for both coffee variety and type of roaster. Despite this, the top scores for fragrance and body were obtained from Sidama coffee roasted using drum roaster. And the top scores for flavor, after taste, and balance were recorded from Yirgacheffe coffee roasted using fluidized bed roaster.

Table 4.4. Cupping Scores of Roasted Coffee Beans

Cup Characteristics Value (10%)

Total After Clean Overall Sample Fragrance Flavor Acidity Body Uniformity Balance Sweetness Score Taste Cup Acceptability Code (100%) 01DM 8* 7.67* 7.08* 7.08* 7.42* 10* 8.08* 10* 10* 7.83* 83.17* 8.25a 7.75a 7.75a 7.5a 8.25a 10a 8.25a 10a 10a 8.25a 86a 7.75b 7.5b 6.25b 6.25b 6.5b 10b 8b 10b 10b 7.5b 79.75b 8c 7.75c 7.25c 7.5c 7.5c 10c 8c 10c 10c 7.75c 83.75c 01FM 7.83* 8.08* 8* 7.83* 7.83* 10* 8.25* 10* 10* 8.58* 86.42* 7.5a 8.25a 7.75a 8a 8a 10a 8a 10a 10a 8.75a 86.25a 8.5b 8.5b 8.5b 8.75b 9b 10b 8.75b 10b 10b 9b 91b 7.5c 7.5c 7.75c 6.75c 6.5c 10c 8c 10c 10c 8c 82c 01TM 7* 7* 7.5* 7.42* 7.67* 10* 7.92* 10* 10* 8* 82.5* 6.75a 6.5a 7.75a 7a 6.5 a 10a 8a 10a 10a 8a 80.5a 6.75b 6.5b 6.5b 6.75b 7.5 b 10b 6.75b 10b 10b 7b 77.75b 7.5c 8c 8.25c 8.5c 9c 10c 9c 10c 10c 9c 89.25c 02DM 7.58* 7.42* 7.33* 7.25* 7.33* 10* 7.75* 10* 10* 7.67* 82.33* 7a 7.25a 7.25a 6.5a 6.5a 10a 8a 10a 10a 8a 80.5a 7.5b 6.5b 6.25b 6.75b 6.5b 10b 6.75b 10b 10b 6.5b 76.75b 8.25c 8.5c 8.5c 8.5c 9c 10c 8.5c 10c 10c 8.5c 89.75c 02FM 7* 7.17* 6.92* 7.25* 7.75* 10* 7.67* 10* 10* 7.5* 81.25* 7.75a 8a 8a 8.25a 8.75a 10a 8.25a 10a 10a 8.25a 87.25a 6.5b 6.75b 6.5b 6.75b 7b 10b 8b 10b 10b 7.5b 79b 6.75c 6.75c 6.25c 6.75c 7.5c 10c 6.75c 10c 10c 6.75c 77.5c 02TM 7.42* 7.08* 7.08* 6.58* 7* 10* 7.5* 10* 10* 7.33* 80* 7.5a 7.25a 7.25a 7.25a 7.75a 10a 7.75a 10a 10a 7.5a 82.25a 8b 8b 7.75b 6b 6b 10b 8b 10b 10b 8b 81.75b 6.75c 6c 6.25c 6.5c 7.25c 10c 6.75c 10c 10c 6.5c 76c 03DM 8.25* 7.58* 7.58* 7.67* 8.08* 10* 8.08* 10* 10* 8* 85.25* 7.75a 6.75a 7a 6.75a 7.5a 10a 8a 10a 10a 7.5a 81.25a 8.5b 7.75b 7.5b 8b 7.75b 10b 7.75b 10b 10b 8b 85.25b 8.5c 8.25c 8.25c 8.25c 9c 10c 8.5c 10c 10c 8.5c 89.25c 03FM 7.92* 7.42* 7.58* 7.75* 7.83* 10* 8.08* 10* 10* 7.83* 84.42* 8a 7.75a 7.25a 7.75a 7.25a 10a 7.75a 10a 10a 7.5a 83.25a 7.5b 6.5b 7.25b 7.25b 7.25b 10b 8b 10b 10b 7.5b 81.25b 8.25c 8c 8.25c 8.25c 9c 10c 8.5c 10c 10c 8.5c 88.75c * Average cupping score of three professional cuppers. Data in same row sharing same letters represent cupping scores given by a single professional cupper. 01DM: Yirgacheffe coffee medium roasted by drum roaster, 02DM: Harar coffee medium roasted by drum roaster, 03DM: Sidama coffee medium roasted by drum roaster, 01FM: Yirgacheffe coffee medium roasted by fluidized bed roaster, 02FM: Harar coffee medium roasted by fluidized bed roaster, 03FM: Sidama coffee medium roasted by fluidized bed roaster, 01TM: Yirgacheffe coffee medium roasted by traditional roaster, 02TM: Harar coffee medium roasted by traditional roaster.

4.3. Composition of Bioactive Compounds

Although no extensive data was found on the caffeine content of the various coffee cultivars in Ethiopia (Workneh, 2015), the level of caffeine in all the three samples of specialty coffee beans was found to be very high and above 0.8%. The highest caffeine level was recoded for Sidama variety with 1.634%, whereas Yirgacheffe and Harar coffee varieties were measured to have caffeine level of 1.567% and 1.498%, respectively. This result is in agreement with the works of Farah et al. (2006) and Franca et al. (2005), who compared caffeine levels for samples of different sensory quality coffee beans and found that the highest caffeine levels corresponded to the highest (specialty) quality coffee samples. They suggested that the highest caffeine levels in good quality coffees may be associated to variations in growth and processing conditions.

Among the three experimental factors selected for the study, only coffee variety and degree of roast were found to have significant (푃 < 0.05) effect on caffeine level of the roasted coffee beans. The type of roasters showed no significant (푃 > 0.05) effect on the level of caffeine in the roasted coffee beans. This is in agreement with the results obtained by (Alonso-Soales et al., 2009) who concluded that the caffeine content of green coffee beans mainly depends on the coffee variety and location of origin. Generally, a significant reduction in caffeine content of the sample coffee beans was observed during the roasting process. Sidama coffee incurred up to 60% reduction in its caffeine level while Harar and Yirgacheffe coffees experienced an average of 53% and 40%, respectively. This is due to the increase in the solubility of caffeine in water as the roasting temperature increases. In addition, the caffeine loss was attributed to a drag by bounded water vapor released from the coffee matrix during roasting (Baggenstoss et al., 2008). The apparent higher caffeine percentages at darker degree of roast (i.e. medium or dark roast) in some of the roasted coffees shown in Figure 4.1 to 4.3 might be due to the decrease in moisture contents as roasting time increases (Vignoli et al., 2014).

1.8 1.634 1.567 1.6 1.499 1.4 1.2 1.062 1 0.873 0.884 0.793 0.8 0.686 0.72 0.5650.565 0.6 0.45

0.4 Caffeine Caffeine (%)Content 0.2 0 Harar Sidama Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.1. Caffeine Content of Drum Roasted Coffee Beans

1.8 1.634 1.567 1.6 1.498 1.4 1.2 0.9750.932 1 0.8860.882 0.84 0.8 0.651 0.635 0.6 0.471 0.469

0.4 Caffeine Caffeine (%)Content 0.2 0 Harar Sidama Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.2. Caffeine Content of Coffee Beans Roasted Using Fluidized Bed Roaster

1.8 1.567 1.6 1.498 1.473 1.4 1.2 0.994 1 0.857 0.81 0.8 0.685 0.6 0.463 0.4 Caffeine Caffeine (%)Content 0.2 0 Harar Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.3. Caffeine Content of Coffee Beans Roasted Using Traditional Roaster

Despite the availability of various scientific papers dealing with trigonelline, quantitative information on its distribution among Ethiopian coffee verities is scarce (Workneh, 2015). Among the three raw coffee samples, Harar coffee recorded the highest trigonelline content of 0.965%, followed by Sidama and Yirgacheffe coffees with nearly equal trigonelline content of 0.716% and 0.712%, respectively. This is well in the range of 0.6 to 2.0% trigonelline content of dry Arabica coffee beans (see Table 2.3 in chapter 2).

Generally, a significant reduction in trigonelline content of the coffee beans was observed during the roasting process, with darker roasts attaining the least values. During the early stage inside the drum roaster, i.e. when the coffee beans turn from raw or green to light roasted, Harar, Sidama and Yirgacheffe coffee showed 22%, 24% & 15% reduction in their trigonelline content, respectively. However, as the roasting processes progresses from light to medium degree of roast, a sharp increase in the trigonelline content was observed, especially for Harar and Yirgacheffe coffees. The highest trigonelline level was recorded at medium degree of roast for all the three coffee varieties, with Harar obtaining 0.895% w/w followed by Yirgacheffe and Sidama coffees attaining 0.868% & 0.584% w/w, respectively. As a result, the aroma of the final cup of coffee becomes higher (Farah et al., 2006). Hence, medium roasted coffees are usually preferred by professional coffee cuppers as the best degree of roast for cupping test. At dark degree of roast, all the coffees showed a dramatic decrease in their trigonelline content which results in low cup quality.

1 0.965 0.897 0.9 0.868 0.753 0.8 0.716 0.712 0.7 0.584 0.605 0.6 0.546 0.546 0.5 0.378 0.4 0.294 0.3

Trigonelline Trigonelline (%)Content 0.2 0.1 0 Harar Sidama Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.4. Trigonelline Content of Sample Coffee Beans Roasted Using Drum Roaster

Fluidized bed roaster brought about less negative impact on the trigonelline content of all the coffee samples during the early stage of roasting process. At light degree of roast, Harar Coffee showed only 5% reduction from its initial trigonelline content. Whereas, Yirgacheffe coffee showed 16% rise in its trigonelline content while Sidama coffee showed nearly no change. However, a steady decrease in trigonelline content was observed at medium degree of roast. Therefore, the highest trigonelline level of all the three coffee varieties was recorded at light degree of roast, with Harar obtaining 0.919% w/w while Yirgacheffee and Sidama coffee attaining 0.829% & 0.717% w/w, respectively. As a result, light degree of roast may be recommended for professional coffee roasters who wish to obtain the best cup qualities of their coffee using fluidized bed roasters.

0.965 1 0.92 0.9 0.855 0.829 0.756 0.8 0.716 0.716 0.712 0.7 0.577 0.6 0.5 0.453 0.392 0.412 0.4 0.3

0.2 Trigonelline (%) Content Trigonelline 0.1 0 Harar Sidama Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.5. Trigonelline Content of Coffee Beans Roasted Using Fluidized Bed Roaster

At light degree of roast, traditional oven top roaster also brought about less degradation effect on the initial trigonelline content of the two coffee samples. Yirgacheffe coffee obtained a 9% rise while Harar coffee showed nearly 11% reduction in their initial trigonelline content. However, a continuous decrease in their trigonelline content was observed after the beans pass light degree of roast and progress to darker roast. Thus, the maximum trigonelline content was achieved at light degree of roasted where Harar coffee recorded 0.860% w/w and Yirgacheffee attained 0.778% w/w. This high trigonelline content at light degree of roast contributes to greater coffee aroma (Farah et al., 2006) and eventually better cup quality.

1 0.965 0.9 0.861 0.778 0.8 0.724 0.712 0.7 0.639 0.6 0.481 0.5 0.392 0.4 0.3

0.2 Trigonelline Trigonelline (%)Content 0.1 0 Harar Yirgacheffe Coffee Variety Raw Light Medium Dark

Figure 4.6. Trigonelline Content of Coffee Beans Roasted Using Traditional Roaster According to the obtained results, medium roasted Harar coffee obtained the highest trigonelline content of 0.895% w/w using drum roaster, followed by 0.855% using fluidized bed roaster, and 0.724% using traditional oven top roaster. Similarly, medium roasted Yirgacheffee coffee showed the highest trigonelline content of 0.868% w/w using drum roaster, followed by 0.756% w/w using fluidized bed roaster 0.639%. And medium roasted Sidama coffee also showed the highest trigonelline content of 0.584% w/w using drum roaster and 0.577% using fluidized bed roaster. This indicates that drum roaster maintains better trigonelline level in the coffees than both fluidized bed roaster and traditional roaster. Therefore, drum type of roasters may be the best choices for roasting specialty coffee beans without jeopardizing their distinctive flavor and aroma.

There is only a limited information in order to compare the chlorogenic acids contents of the collected coffee samples with the chlorogenic acids present in other Ethiopian coffee beans (Alonso-Soales et al., 2009; Moon et al., 2013). An articles by (Farah et al., 2006) reported that raw Harar coffee beans contain about 5.6% (w/w) of total chlorogenic acids. And (Moon et al., 2013), while studying the role of coffee roasting conditions on the level of chlorogenic acids, reported total chlorogenic acid contents of 6.91%, 3.21% and 1.97% (w/w) in the green, light roasted and medium roasted coffee beans of Ethiopian origin, respectively. In a review by Farah & Donangelo (2006), the total chlorogenic acids produced during medium roasting of Harar has been reported as 1.93% (w/w) (Workneh, 2015).

Among the three raw coffee samples, Yirgacheffe coffee obtained the highest total chlorogenic acids content of 3.89% followed by 3.73% and 3.58% for Sidama and Harar coffees, respectively. During the roasting of the sample coffee beans using drum roaster, a significant reduction was observed in their chlorogenic acids level. At light degree of roast, Harar coffee showed 61% reduction followed by Sidama and Yirgacheffe with relatively close 47% & 46% reduction in their chlorogenic acids content, respectively. As the roasting progresses from light to medium degree of roast, Harar and Yirgacheffe coffees showed 21% and 13% increase in their chlorogenic acids level, respectively. However, Sidama coffee incurred 38% decrease from its chlorogenic acids level at light degree of roast. When the beans become dark roasted, they showed tremendous decrease in their original chlorgenic acids content, with 93% loss incurred by Sidama coffee followed by 91% for Harar, , and 81% for Yirgacheffe coffee.

4 3.728 3.893 3.583 3.5

3 2.436 2.5 2.154 1.98 2 1.705 1.408 1.5 1.225

Total CGAs CGAs Total(%) 1 0.723 0.5 0.312 0.253

0 Harar Sidama Yirgacheffe Coffee Variety Raw Light Medium Dark

Figure 4.7. Total CGAs Content of Sample Coffees Roasted by Drum Roaster

As the sample coffee beans progress from raw to dark degree of roast using fluidized bed roaster, a continuous reduction was observed in their chlorogenic acids level. At light degree of roast, both Harar and Sidama coffee showed around 55% reduction in their initial chlorogenic acids content. However, Yirgacheffe showed just 30% reduction from its original chlorogenic acid level. When the sample coffee beans become dark roasted, all of them incurred an average 90% reduction of their original chlorgenic acids level.

4 3.728 3.893 3.583 3.5

3 2.717 2.5

2 1.644 1.691 1.444 1.5 1.264

1.061 Total CGAs CGAs Total(%) 1 0.378 0.5 0.351 0.346

0 Harar Sidama Yirgacheffe Coffee Variety Raw Light Medium Dark

Figure 4.8. Total CGAs content of Sample Coffees Roasted Using Fluidized Bed Roaster When coffee beans roasted using traditional oven top roaster, i.e. Harar and Yirgacheffe coffee beans, incurred an average of 90% reduction in their original chlorgenic acids level. However, Yirgacheffe coffee showed better chlorgenic acid level at light degree of roast, whereas Harar coffee obtained higher level of chlorogenic acids at medium degree of roast.

4 3.893 3.583 3.5

2.5 1.844 2 1.697 1.385 1.5 1.295

Total CGAs CGAs Total(%) 1 0.488 0.5 0.295

0 Harar Yirgacheffe Coffee Variety

Raw Light Medium Dark

Figure 4.9. Total CGAs Content of Sample Coffees Roasted by Traditional Roaster

4.4. Acrylamide Concentration

The results obtained from the UV-Vis Spectrometry analysis were statistically analyzed using an analysis for variance for two factorial model of type of roasters and degree of roasting at three levels of treatment using Design Expert software version 6.0.8. The experiment showed significant (P < 0.05) differences in the roast results from each type of roaster and degree of roast. The highest acrylamide content of Yirgacheffe coffee was obtained at light degree of roast for both drum and traditional type of roasters while it was obtained at medium degree of roast for fluidized bed roaster. On the other hand, the lowest content of acrylamide was exhibited at light degree of roast for fluidized bed roaster, at medium degree of roast for drum roaster, and at dark degree of roast for traditional roaster. Among all the type of coffee roasters, fluidized bed roaster resulted at the least acrylamide content in the roasted sample at light degree of roast.

2.5 2.251 2.306 2.024 2

1.5 1.304 1.216 1.052 1 0.685

Acrylamide (ppm) Acrylamide Conc. 0.46 0.5 0.093 0 Drum Fluid Bed Traditional Type of Roaster Light Medium Dark

Figure 4.10. Acrylamide Content of Roasted Yirgacheffe Coffee The correlation coefficients of acrylamide formation with the degree of roast and bioactive compounds content of roasted Yirgacheffe coffee beans is presented in Table 4.5 to Table 4.7. The study indicated that acrylamide formation had a highly significant (P < 0.001) positive correlation with degree of roast for coffee beans roasted using drum roaster (Table 4.5). Whereas, the results revealed that acrylamide formation had a significant (P<0.05) negative correlation with caffeine and trigonelline content of the drum roasted coffee beans.

Table 4.5. Correlation coefficients among drum roasted coffees presented in matrix form

Roast Total Degree Caffeine Trigonelline CGAs Acrylamide Roast Degree 1 Caffeine -0.75** 1 Trigonelline - 0.78** 1 Total CGAs 0.52* - 0.75** 1 Acrylamide 0.96** -0.90** -0.43* - 1 **Highly significant (P<0.001), *Significant (0.010.05) effect A significant (P<0.05) positive correlation was observed between acrylamide formation and caffeine content of coffee beans roasted using fluid bed roaster (Table 4.6). However, the degree of roast showed no significant (P>0.05) effect on the acrylamide formation.

Table 4.6. Correlation coefficients among fluidized bed roasted coffees presented in matrix form

Roast Total Degree Caffeine Trigonelline CGAs Acrylamide Roast Degree 1 Caffeine -0.63* 1 Trigonelline 0.99** -0.54* 1 Total CGAs 0.96** -0.83** 0.92** 1 Acrylamide - 0.72* - - 1 **Highly significant (P<0.001), *Significant (0.010.05) effect For the traditionally roasted coffee beans, acrylamide formation had a highly significant (P<0.001) positive correlation with degree of roast and all the bioactive compounds content as presented in Table 4.7.

Table 4.7. Correlation coefficients among traditionally roasted coffees presented in matrix form

Total Roast Degree Caffeine Trigonelline CGAs Acrylamide Roast Degree 1 Caffeine 0.57* 1 Trigonelline 0.99** 0.67* 1 Total CGAs 0.99** 0.62* 0.99** 1 Acrylamide 0.89** 0.88** 0.94** 0.92** 1 **Highly significant (P<0.001) and *Significant (0.01

Chapter 5 5. Conclusions and Recommendations

5.1. Conclusions

During the study, the effect of different types of coffee roasting technologies, namely drum, fluidized bed and traditional coffee roasters was investigated on the cup quality and bioactive compounds of specialty graded Yirgacheffe, Harar and Sidama coffee varieties in Ethiopia. According to the experimental results of the study, Sidama and Harar coffees resulted at the best total cup score using drum roaster. Whereas, Yirgacheffe coffee achieved the highest score using fluidized bed roaster. Traditional oven top roaster resulted at the least total cup score for all the coffee varieties. However, no significant (P>0.05) differences in cup characteristics values were detected among the roasted sample coffee varieties disregard of the types of roasters.

Among the factors, only coffee variety and degree of roast were found to have significant (P<0.05) effect on caffeine contents of the coffee bean samples. Generally, a significant reduction in trigonelline and total chlorogenic acids content of the coffee beans was observed during the roasting process, with darker roasts attaining the least values. For drum roaster, the highest trigonelline level was recorded at medium degree of roast for Harar coffee obtaining 0.895% w/w followed by Yirgacheffee coffee with 0.868% w/w and Sidama coffee with the least 0.584% w/w. This may explain why medium degree of roast is usually preferred by professional coffee cuppers for cupping test. For fluidized bed roaster, the highest trigonelline level was recorded at light degree of roast for Harar coffee obtaining 0.919% w/w followed by Yirgacheffe with 0.829% and Sidama coffee with the least 0.717% w/w.

The results also show that drum roaster maintains better trigonelline level in the coffees than fluidized bed roaster and traditional roaster. Drum roasters might be the best choices for roasting specialty coffee beans without jeopardizing their distinctive flavor and aroma. And light degree of roast may be recommended for professional coffee roasters who wish to obtain the best cup qualities of their coffee using fluidized bed roasters.

A single analysis of variance on experiment results of a UV-Vis acrylamide analysis showed significant (P < 0.05) differences for type of roaster and degree of roast. The lowest content of

acrylamide was exhibited at light degree of roast for fluidized bed roaster, at medium degree of roast for drum roaster, and at dark degree of roast for traditional roaster. Generally, based on the obtained results, drum roaster is highly desirable to minimizing acrylamide formation while maximizing the cup quality and maintaining optimum bioactive compounds content at medium degree of roast. Finally, it was concluded that coffee roasting is both an art and a science. Although it is a complex process that depends on the inherent physicochemical composition of the raw coffee beans and the skill set of a person roasting the coffee beans, the type of coffee roasting technologies should be scientifically chosen and adapted to the different coffee varieties. This helps to develop the characteristic flavor and aroma of specialty coffee beans without compromising the quality of the coffee drink and causing any health risks to coffee drinkers.

5.2. Recommendations

 It is very important to extend this study to other varieties of coffees, types of coffee roasting technologies, and roasting conditions in order to really pinpoint how different types of roasting machines or technologies affect the quality of roasted specialty coffee beans in Ethiopia.  In addition, it is important to study the heat and mass transfer efficiencies of different brands and models of coffee roasting machines including their material of construction to further understand their effects on the quality of roasted coffee beans.  It is also recommended to encourage similar researches in the field of coffee processing technology for successful industrial transformation and economic development of Ethiopia.

References

Alonso-Soales, R. M., Serra, F., & Reniero, F. (2009). Botanical and geographical characterization of green coffee (Coffea arabica and Coffea canephora): chemometric evaluation of phenolic and methylxanthine contents. Journal of Agricultural and Food Chemistry, 57, 4224 - 4235. Amamo, A. A. (2014). Coffee production and marketing in Ethiopia. European Journal of Business and Management, 6, 109-121. Anonymous. (2007). Coffee Processing. Uganda Coffee Development Authority (UCDA). Retrieved from https://ugandacoffee.go.ug/primary-processing. AOAC. (1990). Official Methods of Analysis. Volume II. Association of Official Analytical Chemists, Inc. Washington, DC, USA Bae, J. H., Park, J. H., Im, S. S., & Song, D. K. (2014). Mini Review: Coffee and health. Integrative Medicine Research, 3, 189-191. Baggenstoss, J., Poisson, L., Kaegi, R., Perren, R., & Escher, F. (2008). Coffee Roasting and Aroma Formation: Application of Different Time temperature Conditions. Journal of Agricultural and Food Chemistry, 56(14), 5836-5846. Banks, M., McFadden, C., & Atkinson, C. (1999). The World Encyclopaedia of coffee. London: Anness Publishing Limited, pp. 256. Belitz, H. D., Grosch, W., & Schieberle, P. (2009). Coffee, Tea, Cocoa. Berlin: Vatican Springer, 340 – 400. Bicchi, C. P., Panero, O. M., Pellegrino, G. M., & Vanni, A. C. (1995). Characterization of green and roasted coffees through the chlorogenic acid fraction by HPLC-UV and principal component analysis. Journal of Agricultural and Food Chemistry, 43(6), 1549–1555. Bustos-Vanegas, J. D., Corrêa, P. C., Martins, M. A. ,et al. (2017). Developing predictive models for determining physical properties of coffee beans during the roasting process. Industrial Crops & Products, 112(2018), 839–845. Butt, M. S., Ahmed, M. T., Sultan, A. M., Yasin, M., & Imran, M. (2011). Evaluating the Effects of Decaffeination on Nutritional and Antioxidant Status of Different Coffee Brands. Internet Journal of Food Safety, 13, 198 - 207.

Clark, R. J. (2003). Coffee: green coffee/roast and ground. In Caballero, B., Trugo, L. C., & Finglas, P. (eds.), Encyclopedia of Food Science and Nutrition (2nd ed., Vol. 3). Oxford, UK: Academic Press, 584 – 601. Clarke, R. J., & Vitzthum, O. G. (2001). Coffee Recent Developments (Vol. 1). Oxford, England: Blackwell Science Ltd, pp. 272. Clemente, J. M., Martinez, H. E. P., Pedrosa, A. W., Neves, Y. P., Cecon, P. R., & Jifon, J. L. (2018). Boron, Copper, and Zinc Affect the Productivity, Cup Quality, and Chemical Compounds in Coffee Beans: Found in Coffee Quality: Cultivars, Blends, Processing, and Storage Impact. Journal of Food Quality, 2018(2), pp. 14. Clifford, M. N. (2000). Chlorogenic acids and other cinnamates-nature, occurrence, dietary burden, absorption and metabolism. Journal of the Science of Food and Agriculture, 80, 1033 -1043. doi: https://doi.org/10.1002/(sici)1097 CLU. (2017). Main Ethiopian Export Coffee, Broschure. Ethiopian Coffee & Tea Development and Marketing Authority Coffee Quality and Inspection Center. Addis Ababa, pp. 6. Corrêa, P. C., Vargas-Elías, G. A., Oliveira, G. H., Santos, F. L., Vasconcelos, W. L., & Nunes, E. H. M. (2016). Preservation of roasted and ground coffee during storage. Part 2: Bulk density and intergranular porosity. Agriambi: Official Journal of the Latin American Association, 20(7), pp. 6. Cruz, R., Cardoso, M. M., Fernandes, L. ,et al. (2012). Espresso coffee residues: A valuable source of unextracted compounds. Journal of Agriculture and Food Chemistry, 60, 7777 - 7784. Dejene, T. (2011). Quality and value chain analysis of Ethiopian coffee. (MSc thesis), Addis Ababa University, Addis Ababa, Ethiopia. Delatour, T., Perisset, A., Goldmann, T., Riediker, S., & Stadler, R. H. (2004). Improved Sample Preparation to Determine Acrylamide in Difficult Matrixes Such as Chocolate Powder, Cocoa, and Coffee by Liquid Chromatography Tandem Mass Spectrometry. Journal of Agricultural and Food Chemistry, 52, 4625-4631. Diallo, A. (2019). A Heated Comparison: Ikawa vs Probat Caravela Coffee: Caravela Ltd, pp. 11. Duarte, G. S., Pereira, A. A., Farah, A. (2010). Chlorogenic acids and other relevant compounds in Brazilian coffees processed by semi-dry and wet post-harvesting methods. Food Chemistry, 118, 851-855. ECX. (2015). Coffee Contracts (pp. 14): Ethiopian Commodity Exchange, pp. 14.

Eggers, R., & Pietsch, A. (2001). Coffee: Recent Developments. Oxford: Blackwell Science Ltd, pp. 257. EIPO. (2008). Ethiopia: A Portrait Country at the SCAA 20th Annual Conference and Exhibition, Ethiopian Intellectual Property Gazette, 11 – 13. ESA. (2001). Ethiopian Standard ES ISO 10095:2001 Coffee - Determination of caffeine content - Method by using High Performance Liquid Chromatography. Addis Ababa, pp. 4. ESA. (2002a). Ethiopian Standard ES 788:2002 Roasted Coffee Beans and Roasted Ground Coffee - Specification. Addis Ababa, pp. 9. ESA. (2002b). Ethiopian Standard ES ISO 4052:2002 Coffee -Determination of Caffeine content (Reference Method). Addis Ababa, Ethiopia, pp. 7. Farah, A. (2012). Coffee Constituents. In Y.-F. Chu (Ed.), Coffee: Emerging Health Effects and Disease Prevention (1st ed.): Blackwell Publishing Ltd, 21-58. Farah, A., Monteiro, M. C., Calado, V., Franca, A. S., & Trugo, E. L. C. (2006). Correlation between cup quality and chemical attributes of Brazilian coffee. Food Chemistry, 98(2), 373-380. Franca, A. S., Mendonca, J. C. F., & Oliveira, S. D. (2005). Composition of green and roasted coffees of different cup qualities. LWT, 38, 709 – 715. Franca, A. S., & Oliveira, L. S. (2008). Chemistry of Defective Coffee Beans. In E. N. Koeffer (Ed.), Food Chemistry Research Developments. Nova Science Publishers, Inc, 105-135 Gichimu, B. M., Gichuru, E. K., Mamati, G. E., & Nyende, A. B. (2014). Biochemical composition within coffea arabica and its relationship with cup quality. Journal of Food Research, 3, 31 - 44. Guillermo, A., Elías, V.-., Corrêa, P. C., Souza, N. R. D., Baptestini, F. M., & Melo, E. D. C. (2016). Kinetics of Mass Loss of Arabica Coffee During Roasting Process. Journal of the Brazilian Association of Agricultural Engineering, 36(2), 9. Ivon, F. (2002). Coffee Flavor Chemistry. Chichester, England: John Wiley & Sons, Ltd, pp. 410. Jokanović, M. R., Džinić, N. R., Cvetković, B. R., Grujić, S., & Odžaković, B. (2012). Changes of Physical Properties of Coffee Beans during Roasting. Acta Periodica Technologica, 43, 1-342.

Kalaska, B., Piotrowski, L., Leszczynska, A. ,et al. (2014). Antithrombotic effects of pyridinium compounds formed from trigonelline upon coffee roasting. Journal of Agriculture and Food Chemistry, 62, 2853-2860. Kelly, C. (2018). The Art of Coffee Roasting: Investigations into Sensor Development for the Application of Controlling Coffee Roasting (Doctorate of Philosophy), The University of Waikato, New Zealand, pp. 167. Kirk-Othmer. (1985). Encyclopedia of Chemical Technology (Vol. 7). Russia, 250 – 270. Koskei, K. R., Patrik, M., & Simon, M. (2015). Effects of Coffee Processing Technologies on Physicochemical Properties and Sensory Qualities of Coffee. African Journal of Food Science, 9(4), 7. Ky, C. L., Dussert, S., Guyot, B., Hamon, S., & Noirot, M. (2001). Caffeine, trigonelline, chlorogenic acids and sucrose diversity in wild Coffea arabica L. and C. canephora P. accessions. . Food Chemistry, 75, 223-230. Marilisa, A., & Anese, M. (2018). Effect of coffee roasting on in vitro α-glucosidase activity: Inhibition and mechanism of action. Food Research International, 111, 8. doi: 10.1016/j.foodres.2018.05.061 Mastronardi, G. (2012). Effects of Origin and Treatment of the Roasting Process on the Aromatic and Sensorial Composition of Coffee. (PhD. Thesis), University of Udine, Udine, Italy. Mendonça, J. C. F., Franca, A. S., & Oliveira, S. D. (2007). A comparative evaluation of methodologies for water content determination of in green coffee. LWT-Food Science and Technology, 40(7), 1300 - 1303. Minten, B., Tamru, S., Kuma, T., & Nyarko, Y. (2014). Structure and performance of Ethiopia’s coffee export sector. Washington, DC, USA: International Food Policy Research Institute. Moon, J. L., Sang, E. K., Jong, H. K., Sang, W. L., & Dong, M. Y. (2013). A Study of Coffee Bean Characteristics and Coffee Flavors in Relation to Roasting. Journal of the Korean Society of Food Science and Nutrition, 42(2), 255-261. Mussatto, S. I., Machado, E. M. S., Martins, S., & Teixeira, J. A. (2011). Production, Composition, and Application of Coffee and Its Industrial Residues. Food Bioprocess Technol, 4, 12. Nagaraju, V. D., Murthy, C. T., Ramalakshmi, K., & Rao, P. N. S. (1997). Studies on roasting of coffee beans in a spouted bed. Journal of Food Engineering, 31(2), 263-270.

Nicolas, W. J. (2004). Coffee: Growing, Processing, Sustainable Production, A Guidebook for Growers, Processors, Traders, and Researchers: Wiley-VCH, pp. 1021. Oliveira, G., Oliveira, A., Botelho, F. M., Treto, P. C., & Botelho, S. (2018). Coffee Quality: Cultivars, Blends, Processing, and Storage Impact. Journal of Food Quality, 2018(1), pp.7. Paul, H. (2012). User's Manual: Javalytics Degree of Roast Analyzer: Madison Instruments Inc. Retrieved from http://www.javalytics.com/wpcontent/uploads/2012/06/javalytics-jav-rda- h-handheld.pdf. Poltronieri, P., & Rossi, F. (2016). Review: Challenges in Specialty Coffee Processing and Quality Assurance. Challenges, 7(19), pp. 22. Rodrigues, N. P., & Bragagnolo, N. (2013). Identification and quantification of bioactive compounds in coffee brews by HPLC–DAD–MSn. Journal of Food Composition and Analysis, 32 105-115. Shan, O. E., Zzaman, W., & Yang, T. A. (2016). Impact of Different Temperature-Time Profiles during Superheated Steam Roasting on Some Physical Changes of Robusta Coffee. Pertanika Journal of Tropical Agricultural Science, 39(3), 311-320. Sinnott, K. (2010). The Art and Craft of Coffee: An Enthusiastic Guide to Selecting, Roasting, and Brewing Exquisite Coffee. Beverly, Massachusetts: Quarry Books, pp. 125. Sivetz, M., & Desrosier, N. W. (1979). Coffee Technology. Westport, Colombia: Avi Publishing Co, pp. 716. Soares, C. D., Cunha, S., & Fernandes, J. (2006). Determination of acrylamide in coffee and coffee products by GC-MS using an improved SPE clean-up. Food Additives and Contaminants, 23(12), 1276-1282. Sualeh, A., & Mekonnen, N. (2015). Manual for Coffee Quality Laboratory (Kirub, A., ed.). Addis Ababa, Ethiopia: Ethiopian Institute of Agricultural Research, pp. 24. Summaa, C. A., Callea, B. d. l., Broheea, M., Stadlerb, R. H., & Anklama, E. (2007). Impact of the roasting degree of coffee on the in vitro radical scavenging capacity and content of acrylamide. LWT, 40, 6. Tefera, A., & Tefera, T. (2013). Coffee Annual Report. In Q. Gray (ed.), Global Agricultural Information Network Report. Addis Ababa, Ethiopia: USDA Foreign Agricultural Service, pp. 9.

Tin, W. Y. (2006). Study of Acrylamide Formation in Coffee During Roasting. Malasia Science & Technology University College, pp. 82. UCDA. (2007). Coffee Sector Profile. Entebbe, Uganda, pp. 13. Ullman. (2011). Ulmans Encyclopedia of Industrial Engineering (7th ed.): Wiley-VCH, pp. 29456. Uman, E., Dashwood, M. C., & Dashwood, L. C. (2016). The effect of bean origin and temperature on grinding roasted coffee, Scientific Reports, pp. 8. USDA. (2017). Coffee: World Markets and Trade, USDA Foreign Agricultural Service, pp. 8. Vignoli, J., Viegas, M., Bassoli, D., & Benassi, M. (2014). Roasting process affects differently the bioactive compounds and the antioxidant activity of arabica and robusta coffees. Food Research International, 61, 279–285. Wang, N. (2012). Physicochemical Changes of Coffee Beans During Roasting. (Masters Thesis Dessertation), The University of Guelph, Ontario, Canada, pp. 100. Workneh, B. M. (2015). Studies on the Chemical Composition and Marker Compounds of Green Coffee Beans Grown in Different Parts of Ethiopia. (Degree of Doctor of Philosophy in Chemistry (Analytical Chemistry)), Addis Ababa University, Addis Ababa, pp. 183. Yeretzian, C., Jordan, A., Badoud, R., & Lindinger, W. (2002). From the green bean to the cup of coffee: investigating coffee roasting by on-line monitoring of volatiles. European Food Research and Technology, 214, 92-104. Yeretzian, C., Jordan, A., & Lindinger, W. (2003). Analysing the headspace of coffee by proton- transfer-reaction mass-spectrometry. International Journal of Mass Spectrometry, vol. 224, 115-139.

Appendices

Appendix 1. Preliminary Unwashed Coffee Quality Assessment Sheet

Appendix 2. Cupping Protocol used at ECTACQICC

1. Weighing 2. Grinding 3. Fragrance 4. Boiling Water Testing

8. Cup Tasting 5. "Brewing" 6. Aroma Testing 7. Skimming

Appendix 3. Specialty Coffee Cupping or Grading Sheet

Appendix 4. HPLC Chromatograms of Sample Raw Coffee Beans

Yirgacheffe

Harar

Sidama

Appendix 5. Calibration Curves for the HPLC Analysis

250 Caffeine y = 0.0265x + 0.4357 200 R² = 0.9999

150

Series1 100 Linear (Series1)

Conc. (mg/g) Conc. 50

0 0 2000 4000 6000 8000 Peak Area

250 Trigonelline y = 0.0602x - 0.2759 200 R² = 0.9999

150

Series1 100 Linear (Series1)

50 Conc. (mg/g) Conc.

0 Peak Area 0 1000 2000 3000 4000

250 CGAs y = 0.0267x - 1.218 200 R² = 1

150

Series1

100 Linear (Series1) (mg/g)

50 Conc. Conc.

0 Peak Area 0 2000 4000 6000 8000

Appendix 6. HPLC Chromatograms of Roasted Coffee Beans

01DL 01DM

01FL 01DD

1. 01DL, 01DM, and 01DD are sample Yirgacheffe coffees roasted using drum roaster at light, medium, and dark degree of roast, respectively. 2. 01FL is sample Yirgacheffe coffee roasted using fluidized bed roaster at light degree of roast.

01FM 01FD

01TL 01TM

1. 01FM and 01FD are sample Yirgacheffe coffees roasted using fluidized bed roaster at medium and dark degree of roast, respectively. 2. 01TL and 01TM are sample Yirgacheffe coffees roasted using traditional roaster at light and medium degree of roast, respectively.

01TD 02DL

02DM 02DD

1. 01TD is sample Yirgacheffe coffee roasted using traditional roaster at dark degree of roast. 2. 02DL, 02DM and 02DD are sample Harar coffees roasted using drum roaster at light, medium, and dark degree of roast, respectively.

02FL 02FM

02FD 02TL

1. 02FL, 02FM and 02FD are sample Harar coffees roasted using fluidized bed roaster at light, medium, and dark degree of roast, respectively. 2. 02TL is sample Harar coffee roasted using traditional roaster at light degree of roast.

02TM 02TD

03DL 03DM

1. 02TM and 02TD are sample Harar coffees roasted using traditional roaster at medium and dark degree of roast, respectively. 2. 03DL and 03DM are sample Sidama coffees roasted using drum roaster at light and medium degree of roast, respectively.

03FL 03DD

03FM 03FD

1. 03DD is sample Sidama coffee roasted using drum roaster at dark degree of roast. 2. 03FL, 03FM and 03FD are sample Sidama coffees roasted using fluidized bed roaster at light, medium, and dark degree of roast, respectively.

Appendix 7. Photos of Sample Coffee Beans

I. Raw Sample Coffee Beans

Sidama Yirgacheffe Harar II. Roasted Sample Coffee Beans

01DL 01DM 01DD

01FL 01FM 01FD

01TL 01TM 01TD

Appendix 8. Absorbance Reading of Standard Acrylamide Samples at 220nm

Sample Concentration F Mean SD % RSD Readings No. (ppm) 1 0.4 0.0528 0.0041 7.77 0.0481 R 0.0544 R 0.0558 2 1 0.2249 0.0085 3.78 0.2335 R 0.2165 R 0.2248 3 2 0.4579 0.0040 0.87 0.4552 R 0.4625 R 0.4560 4 4 0.6097 0.0055 0.90 0.6138 R 0.6118 R 0.6034 R = Repeated reading

Appendix 9. Calibration curve for Acrylamide Analysis

0.7 y = 0.1488x + 0.061 0.6 R² = 0.9077

0.5

0.4

0.3 Absorbance 0.2

0.1

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Concentration (ppm)

Appendix 10. Absorbance Reading of Roasted Yirgacheffe Samples at 220nm

Sample F Mean SD % RSD Readings Code 01DL 0.3621 0.0180 4.96 0.3502 R 0.3534 R 0.3828 01DM 0.2420 0.0047 1.94 0.2406 R 0.2473 R 0.2382 01DD 0.2550 0.0061 2.40 0.2479 R 0.2581 R 0.2589 01FL 0.0749 0.0025 3.40 0.0746 R 0.0724 R 0.0775 01FM 0.3960 0.0046 1.17 0.4000 R 0.3971 R 0.3909 01FD 0.1295 0.0041 3.18 0.1341 R 0.1282 R 0.1262 01TL 0.4041 0.0017 0.42 0.4021 R 0.4049 R 0.4053 01TM 0.2175 0.0028 1.31 0.2142 R 0.2191 R 0.2191 01TD 0.1629 0.0068 4.19 0.1682 R 0.1552 R 0.1652 R = Repeated reading

Appendix 11. Photos of the Research Laboratory Works

HPLC Analysis of Bioactive Compounds in Sample Coffees

Addition of Carrez Reagents in a Fume Hood

Precipitation of polysaccharides and large protein molecules due to addition of Carrez reagents

Centrifugation of Precipitated Samples

Rotary Evaporation of Centrifuged Samples SPE Clean-up of Sample Solutions

UV-Vis Spectrophotometer Used for Analysis of Acrylamide in the Sample Coffees