Effect of Pasteurization of Bottled Unfiltered Beer to Beer Volatile Substances Bernardo João Marques Saraiva Biological Engine

Effect of Pasteurization of Bottled Unfiltered Beer to Beer Volatile Substances

Bernardo João Marques Saraiva

Thesis to obtain the Master of Science Degree in Biological Engineering

Supervisors: Prof. Dr. Ana Cristina Anjinho Madeira Viegas

Prof. Eng. Pavel Dostálek

Examination Committee

Chairperson: Prof. Dr. Helena Maria Rodrigues Vasconcelos Pinheiro Supervisor: Prof. Dr. Ana Cristina Anjinho Madeira Viegas Member of the Committee: Prof. Dr. Marília Clemente Velez Mateus

November 2019

Preface

The work presented in this thesis was performed at Department of Biotechnology of University of Chemistry and Technology (Prague, Czech Republic), during the period from March to July 2019, under the supervision of Prof. Eng. Pavel Dostálek, and within the frame of the Erasmus+ program. The thesis was co-supervised at Instituto Superior Técnico by Prof. Dr. Ana Cristina Anjinho Madeira Viegas.

Declaration

I declare that this document is an original work of my own authorship and that it fulfills all the requirements of the Code of Conduct and Good Practices of the Universidade de Lisboa.

Acknowledgements

First of all, I would like to thank my supervisors who accompanied me throughout this extensive process which culminated with the achievement of this document, to whom I am extremely grateful for all the advices and recommendations given in order to facilitate my learning process and to help me handing the best piece of work possible. Thank you to all the crew at Brevnov Monastery Brewery of St. Adalbert, along with University of Chemistry and Technology of Prague and Instituto Superior Técnico for giving me this chance of performing this intellectually challenging and amusing project. Furthermore, I would like to thank all staff at Instituto Superior Técnico that contributed to my academic success throughout this past 5 years, which had a remarkable influence on my personal and intellectual growth. I learned at least a little bit from every professor that came across my academic path, and looking back, I honestly all those struggles I had along the path made me readier to overcome any type of adversity I may face in life and to embrace any professional challenge I may come across. The learning process is never complete, but I doubtlessly consider I have made the best choice I could have ever made to achieve great goals in the engineering field. At a personal level, I have to praise my parents, Maria and Jorge for everything they have done for me, for an unconditional support, for providing me everything I ever needed to be as successful as I could and for all the sacrifices they have made to ensure I would have a brilliant academic path ever since the day I was born. I cannot put into words how thankful I am to my progenitors, who were prouder than anyone else of my personal accomplishments, since they were the ones making everything so it could become possible and supported me in every single moment. Secondly, I would like to mention the influence of Rita Marçal on my academic route and the help I got from her along the way. A remarkable, caring and brilliant person and student who displayed academic excellence at the same time she encouraged everyone around her to improve and provided needful help for them to achieve their own goals. I must proceed to thank my four best friends Casimiro, Diogo, Jorge and Miguel for the crucial role they have in my life and whose endless and unconditional support made me look at the brighter side of adversity and focus on the right topics at the right time. Few others encourage me better than this group of outstanding human beings, with who I spent some of my most joyful times. I would also like to show my appreciation for all my other family members, namely my godmother Ana, for all the faith and pride they have put in me and for contributing significantly for my education. Since I have spent a number of tough times at university, it is also logical to acknowledge those who brought me the happiest moments I have spent in here and with who I have gladly shared my everyday life throughout these last 5 years. Therefore, thank you João, Henrique, José, Rafael, Xavier, Vivar, Gamelas, Guilherme, Catarina, Beatriz, Joana, Miguel, Ricardo, Mestre, Lucas, Sebastião, Diogo, Cristina, Marta, Pedro, Inês, Mariana, Francisco, Francisca and many others who provided me some memorable moments and were a big part of my university life, which special highlight to my beloved godchildren girls Joana, Inês, Débora, Estela and Marta. I never thought I would

say this but I am currently very sure that I will miss these days I have spent at Instituto Superior Técnico because, as this stage of my life is coming to a wonderful disclosure after years of striving, I can only look at the positive side of everything and those moments were innumerable. Lastly, I would like to thank the city of Prague for welcome me so well and be a part of this thesis to obtain master degree in biological engineering. I have spent some of the best moments I have ever experienced over there. Grateful for the contributions of Gabrielle, Margaux, Alix, Lucas, Maxence, Simone, François throughout this process and for the much-needed moral support they gave me, with a honorable mention to my dear Nadja, who put all the effort she got to bring me up and doubtlessly believed I would be successful even when every single aspect pointed to the opposite direction. I do not know what is coming ahead of me, but all I can say is that I will, as always, face it with a smile on my face resultant from all these wonderful people I have in my life and for the role of Instituto Superior Técnico on it.

Abstract

Pasteurization is a heat-treatment process that has been used since the 19th century to extend the shelf-life of food and beverages by eliminating most pathogens and spoiling microorganisms that may harm the quality and safeness of the product. It is a commonly used method in the food industry ever since the market became more demanding and more awareness regarding the influence of diet on personal health has been raised. However, in the case of beer, a number of contradictory reports has emerged concerning the influence of pasteurization in its flavor attributes. Some of them mention pasteurization enhances beer staling and causes a more prominent loss of fresh flavors, whereas others report this heat-treatment process favors the stability of beer organoleptic properties in addition to preventing its spoilage. Therefore, this study was performed in order to evaluate the impact of pasteurization on the sensory quality of beer, as well as on its more influent aroma-active compounds. For this, triangle and preference tests were performed using tunnel and flash pasteurized 12% lagers and 20% imperial lagers to assess which pasteurization method produced better results regarding sensory quality maintenance. In addition, IPA (Indian Pale Ale) samples were used to thoroughly evaluate and compare the sensory profile of pasteurized and non-pasteurized bottled aged beers, which were furtherly submitted to analysis of hop-derived compounds, to check the influence of pasteurization on these flavoring agents through storage time (up to 4 months). Furthermore, sets of pasteurized and non-pasteurized 12% lager and 20% imperial lager matured samples were used to perform analysis with regard to their carbonyl profile, for testing the effect of pasteurization on the formation of staling indicators. Tunnel pasteurization proved to be a more efficient method to ensure a long-term quality product. Use of pasteurization appears to have a positive effect on perseverance of overall flavor attributes and, in contrast, dismissing this treatment caused spoilage notes on the tested beers within a period of 4 months. Hop-derived and carbonyl compounds analysis on their respective beer types did not exhibit any significant consequence of pasteurization usage, apart from slight evidences of a higher degree of oxidation for pasteurized samples in comparison to non-pasteurized ones, although further studies with beers stored for longer periods are needed, in both cases, to reach more accurate conclusions.

Keywords: Beer, Pasteurization, Spoilage, Carbonyls, Staling, Hop oils

Resumo

A pasteurização é um processo de tratamento térmico usado desde o século 19 para prolongar a vida útil de alimentos e bebidas, eliminando grande parte dos microrganismos patogénicos e outros potencialmente nocivos que possam prejudicar a qualidade e a segurança do produto. É um método usado frequentemente na indústria alimentar desde que o mercado se tornou mais exigente e aumentou a conscientização sobre a influência da dieta na saúde humana. No entanto, no caso da cerveja, surgiram vários relatos contraditórios sobre a influência da pasteurização nos seus atributos de sabor. Alguns deles mencionam que a pasteurização promove o staling da cerveja e causa uma perda mais notória de sabores frescos, enquanto outros relatam que este processo de tratamento térmico favorece a estabilidade das propriedades organoléticas da cerveja, além de evitar a sua deterioração. Assim, este estudo foi realizado com o objetivo de avaliar o impacto da pasteurização na qualidade sensorial da cerveja e também nos seus compostos químicos aromáticos mais influentes. Para isso, foram realizados testes de triângulo e de preferência, utilizando amostras de lager 12% e de lager imperial 20% submetidas a pasteurização em garrafa ou em flash, para avaliar qual o método de pasteurização que produz melhores resultados em relação à manutenção da qualidade sensorial das cervejas. Além disso, amostras de IPA foram usadas para avaliar e comparar minuciosamente o perfil sensorial de cervejas envelhecidas em garrafa submetidas ou não a pasteurização, que foram posteriormente submetidas a análise de compostos derivados dos lúpulos, para verificar a influência da pasteurização na concentração desses agentes aromatizantes ao longo do tempo de armazenamento. Adicionalmente, conjuntos de amostras maturadas de lager 12% e de lager imperial 20% submetidas ou não a pasteurização, foram usadas para realizar análises em relação ao seu perfil de compostos carbonilo, para testar o efeito da pasteurização na formação destes indicadores de staling. A pasteurização em garrafa provou ser um método mais eficiente para garantir um produto de qualidade a longo prazo. O uso da pasteurização aparenta ter um efeito positivo na perseverança dos atributos gerais do sabor da cerveja e, por outro lado, descartar esse tratamento causou sinais de deterioração nas cervejas testadas após um período de 4 meses. A análise de compostos derivados de lúpulo e de compostos carbonílicos nos seus respetivos tipos de cerveja analisados, não indicou qualquer correlação com o uso ou não de pasteurização, além de pequenas evidências que apontam para um maior grau de oxidação em amostras pasteurizadas em comparação com amostras não pasteurizadas, embora estudos adicionais com cervejas armazenadas por períodos mais longos são necessários, em ambos os casos, para chegar a conclusões mais precisas.

Palavras-chave: Cerveja, Pasteurização, Deterioração, Carbonilos, Staling, Óleos de lúpulo

List of Contents

1. INTRODUCTION – LITERATURE REVIEW ...... 1

1.1 BRIEF HISTORY OF BEER ...... 1

1.2 BEER AGING ...... 2

1.2.1 CHEMICAL CHANGES IN BEER DURING STORAGE ...... 4

1.2.1.1 VOLATILE COMPOUNDS ...... 5

1.2.1.1.1 CARBONYL COMPOUNDS ...... 5

1.2.1.1.1.1 ALKANALS AND ALKENALS ...... 9

1.2.1.1.1.2 STRECKER ALDEHYDES...... 9

1.2.1.1.1.3 KETONES AND VICINAL DIKETONES ...... 10

1.2.1.1.1.4 CYCLIC ACETALS ...... 12

1.2.1.1.1.5 HETEROCYCLIC COMPOUNDS AND MAILLARD REACTION ...... 13

1.2.1.1.1.6 VOLATILE ESTERS ...... 16

1.2.1.1.1.7 SULFUR COMPOUNDS ...... 17

1.2.1.1.1.8 VARIATION OF CARBONYLS CONCENTRATION DURING BEER AGING ...... 18

1.2.1.1.1.9 SUMMARY ...... 20

1.2.1.2 REACTIVE OXYGEN MECHANISMS OF AGING PROCESS IN BEER ...... 21

1.2.1.2.1 OXIDATION OF UNSATURATED FATTY ACIDS...... 23

1.2.2 INHIBITION OF FLAVOR DEGRADATION ...... 25

1.2.3 HOP-DERIVED COMPOUNDS ...... 27

1.2.3.1 RESINS ...... 28

1.2.3.2 ESSENTIAL OILS ...... 29

1.3 PASTEURIZATION IN BEER PRODUCTION ...... 31

1.3.1 TUNNEL PASTEURIZATION ...... 34

1.3.2 FLASH PASTEURIZATION ...... 35

1.4 THESIS SCOPE AND GOALS ...... 36

2. MATERIALS AND METHODS ...... 36

2.1 RELEVANT FEATURES OF BEER PRODUCTION PROCESS ...... 36

2.2 PASTEURIZATION OF BEER AND STORAGE CONDITIONS ...... 37

2.3 SENSORY ANALYSIS METHODS: ...... 37

2.3.1 TRIANGLE TEST ...... 37

2.3.2 PREFERENCE TEST ...... 38

2.3.3 SENSORY TEST ...... 38

2.4 CARBONYL COMPOUNDS ANALYSIS ...... 39

2.4.1 FILTRATION ...... 39

2.4.2 DERIVATIZATION METHOD ...... 39

2.4.3 HS-SPME GC-MS PROCEDURE AND CONDITIONS ...... 42

2.4.4 EVALUATION OF HS-SPME GC-MS RESULTS ...... 42

2.4.5 OTHER DEVICES AND EQUIPMENT ...... 43

2.4.6 CALIBRATION LINES ...... 43

2.5 ESSENTIAL HOP OILS ANALYSIS ...... 45

2.5.1 FILTRATION ...... 45

2.5.2 ISOLATION AND PURIFICATION OF HOP ESSENTIAL OILS COMPOUNDS ...... 45

2.5.3 GC-MS CONDITIONS AND RESULTS TREATMENT ...... 49

3. RESULTS AND DISCUSSION ...... 50

3.1 SENSORY ANALYSIS OF TUNNEL OR FLASH PASTEURIZED 12% LAGERS AND 20% IMPERIAL

LAGERS: ...... 50

3.1.1 TRIANGLE TESTS ...... 50

3.1.2 PREFERENCE TEST ...... 54

3.2 CARBONYL AGING MARKERS ANALYSIS IN LAGERS ...... 56

3.3 SENSORY TESTS PERFORMED FOR MONASTERY IPA SAMPLES ...... 62

3.4 ESSENTIAL HOP OILS ANALYSIS IN MONASTERY IPA’S ...... 71

4. CONCLUSIONS AND FUTURE PROSPECTS ...... 79

5. REFERENCES ...... 81

6. APPENDICES ...... 92

APPENDIX I – CALIBRATION LINES DESIGNED FOR CARBONYL ANALYSIS ...... 92

APPENDIX II – SUPPORTING DATA FOR ANALYSIS OF TRIANGLE TESTS RESULTS ...... 93

APPENDIX III – SUPPORTING DATA FOR ANALYSIS OF CARBONYL AGING MARKERS ...... 95

APPENDIX IV – MONASTERY IPA’S SENSORY TESTS RESULTS SHEETS ...... 95

APPENDIX V – IUPAC NAMES OF HOP ESSENTIAL OILS ANALYZED COMPOUNDS...... 99

List of Tables

TABLE 1- VARIETY OF CARBONYL COMPOUNDS AND RESPECTIVE CHEMICAL CLASSES MOSTLY REPRESENTED IN BEER...... 6 TABLE 2- ALCOHOL PERCENTAGE, COLOR, PH AND BITTERNESS OF 8 COMMERCIAL BEERS TESTED: 3 LAGER BEERS, L-A, L-B, L-C AND 5 SPECIALTY BEERS, S-A, S-B, S-C, S-D, S-E...... 20 TABLE 3- CARBONYL AGING MARKERS IN BEER AND THE TYPE OF REACTION INVOLVED IN THEIR

FORMATION DURING STORAGE...... 20

TABLE 4- VOLUME, WEIGHT AND DEGREE OF PURITY OF EVERY CARBONYL COMPOUND ADDED TO THE MIXED STANDARDS SOLUTION...... 40 TABLE 5- LIST OF HS-SPME GC-MS PARAMETERS AND CONDITIONS FOR ANALYSIS OF CARBONYL COMPOUNDS...... 42

TABLE 6- LIST OF GC-MS PARAMETERS AND CONDITIONS FOR ANALYSIS OF HOP ESSENTIAL OILS...... 49 TABLE 7- VOLUME OF FILTRATE (L) COLLECTED FOR EACH BEER SAMPLE...... 50 TABLE 8- TRIANGLE TESTS RESULTS REGARDING TASTE DIFFERENCES BETWEEN TUNNEL AND FLASH

PASTEURIZED LAGER BEERS (12% LAGER AND 20% IMPERIAL LAGER (IL)) AND AVERAGE

DIFFICULTY LEVEL OF THE ASSESSORS POPULATION CHOICE. CORRECT ANSWERS ARE REFERRED

TO THE NUMBER OF ASSESSORS WHO WERE ABLE TO CORRECTLY IDENTIFY THE DIFFERENTLY

PASTEURIZED SAMPLE FOR EACH TYPE OF LAGER AND PD IS A STATISTICAL PARAMETER WHICH

REPRESENTS THE PROPORTION OF CORRECT ASSESSMENTS. THE NUMBER OF TRAINED ASSESSORS IS ALSO DISCRIMINATED FOR EACH CASE...... 51 TABLE 9- RESULTS OF PREFERENCE TEST CARRIED OUT AFTER THE THIRD SET OF TRIANGLE TESTS

PERFORMED FOR 20% IMPERIAL LAGERS...... 54 TABLE 10- CARBONYL COMPOUNDS CONCENTRATION (MG/L), AND RESPECTIVE CHARACTERISTIC

RETENTION TIME (MIN), DETERMINED FOR NON-PASTEURIZED (NP) AND PASTEURIZED (P) 12%

LAGER AND 20% IMPERIAL LAGER (IL) SAMPLES STORED FOR 0 MONTHS, THOROUGH DERIVATIZATION METHOD FOLLOWED BY GC-MS ANALYSIS...... 56

TABLE 11- CARBONYL COMPOUNDS CONCENTRATION (MG/L), AND RESPECTIVE CHARACTERISTIC

RETENTION TIME (MIN), DETERMINED FOR NON-PASTEURIZED (NP) AND PASTEURIZED (P) 12%

LAGER AND 20% IMPERIAL LAGER (IL) SAMPLES STORED FOR 2 MONTHS, THOROUGH DERIVATIZATION METHOD FOLLOWED BY GC-MS ANALYSIS...... 57 TABLE 12- CARBONYL COMPOUNDS CONCENTRATION (MG/L), AND RESPECTIVE CHARACTERISTIC

RETENTION TIME (MIN), DETERMINED FOR NON-PASTEURIZED (NP) AND PASTEURIZED (P) 12%

LAGER AND 20% IMPERIAL LAGER (IL) SAMPLES STORED FOR 4 MONTHS, THOROUGH DERIVATIZATION METHOD FOLLOWED BY GC-MS ANALYSIS...... 57 TABLE 13- THE MOST ACTIVE HOP OILS-DERIVED ODORANTS COMMONLY FOUND IN BEER AND THEIR

RESPECTIVE RETENTION TIME ON THE GC COLUMN USED, FLAVOR DESCRIPTIONS AND FLAVOR

THRESHOLDS ...... 72

TABLE 14- RESULTS OBTAINED FOR HOP ESSENTIAL OILS ANALYSIS OF NON-PASTEURIZED (NP) AND

PASTEURIZED (P) STANDARD MONASTERY IPA AND MONASTERY IPA + ESSENTIAL OILS SAMPLES STORED FOR 0 MONTHS...... 73

TABLE 15- RESULTS OBTAINED FOR HOP ESSENTIAL OILS ANALYSIS OF NON-PASTEURIZED (NP) AND

PASTEURIZED (P) STANDARD MONASTERY IPA AND MONASTERY IPA + ESSENTIAL OILS SAMPLES STORED FOR 2 MONTHS...... 74 TABLE 16-RESULTS OBTAINED FOR HOP ESSENTIAL OILS ANALYSIS OF NON-PASTEURIZED (NP) AND

PASTEURIZED (P) STANDARD MONASTERY IPA AND MONASTERY IPA + ESSENTIAL OILS SAMPLES

STORED FOR 4 MONTHS...... 75 TABLE 17- CALIBRATION EQUATIONS OBTAINED FOR EACH ANALYZED CARBONYL AND RESPECTIVE

COEFFICIENT OF DETERMINATION (R2)...... 93 TABLE 18- MINIMUM NUMBER OF CORRECT RESPONSES REQUIRED FOR SAMPLES TO BE CONSIDERED SIGNIFICANTLY DIFFERENT AT THE STATED ALPHA-LEVEL...... 94

TABLE 19- DESCRIPTION OF STUDIED CARBONYL COMPOUNDS, REGARDING THE CHEMICAL GROUP THEY

BELONG TO, AS WELL AS THEIR RESPECTIVE FLAVOR THRESHOLDS, FORMATION MECHANISMS AND FLAVOR DESCRIPTORS ...... 95 TABLE 20- SENSORY TESTS RESULTS FOR MONASTERY IPA SAMPLES STORED FOR 0 MONTHS...... 96 TABLE 21- SENSORY TESTS RESULTS FOR MONASTERY IPA SAMPLES STORED FOR 2 MONTHS...... 97 TABLE 22- SENSORY TESTS RESULTS FOR MONASTERY IPA SAMPLES STORED FOR 4 MONTHS...... 98

List of Figures

FIGURE 1- SENSORY CHANGES DURING BEER AGING ACCORDING TO DALGLIESH (1977)...... 3

FIGURE 2- CHEMICAL STRUCTURE OF A FEW CARBONYL COMPOUNDS CONNECTED TO BEER STALING

MENTIONED ON TABLE 2...... 6

FIGURE 3- FORMATION OF (E)-2-NONENAL IN BEER BY ALDOL CONDENSATION OF ACETALDEHYDE AND

HEPTANAL...... 7

FIGURE 4- REPRESENTATION OF THE DERIVATIZATION REACTION SCHEME INVOLVING CARBONYL

COMPOUNDS AND PFBOA AGENT ...... 8

FIGURE 5- SCHEMATIZATION OF THE STRECKER REACTION MECHANISM WHICH GIVES RISE TO IMPORTANT

FLAVOR-ACTIVE ALDEHYDES ...... 10

FIGURE 6- SCHEMATIZATION OF THE PATHWAYS FOR DIACETYL AND 2,3-PENTANEDIONE FORMATION

DURING YEAST METABOLISM AND SUBSEQUENT REDUCTION, AS WELL AS VALINE AND ISOLEUCINE

SYNTHESIS ...... 11

FIGURE 7- MAILLARD REACTION SCHEME DEVELOPED BY HODGE ...... 14

FIGURE 8- TWO DIFFERENT PATHWAYS FOR FURFURAL AND HMF PRODUCTION IN BEER, RESPECTIVELY

FROM PENTOSES AND HEXOSES...... 15

FIGURE 9- ISOAMYL ACETATE FORMATION AND DEGRADATION REACTIONS ...... 17

FIGURE 10- CONCENTRATIONS OF 3-METHYLBUTANAL (A), 2-ISOBUTYL-4,5-DIMETHYL-1,3-DIOXOLANE

(B), FURFURAL (C), FURFURYL ETHYL ETHER (D) AND DIACETYL (E) AND THE EVOLUTION OF BEER

COLOR (F) DURING STORAGE OF EIGHT BEERS FOR ONE YEAR AT 20°C ...... 19

FIGURE 11- REACTIONS PRODUCING REACTIVE OXYGEN SPECIES IN BEER ...... 22

FIGURE 12- REACTION BETWEEN ETHANOL AND HYDROXYL RADICALS IN BEER ACCORDING TO ANDERSEN

AND SKIBSTED (1998)...... 22

FIGURE 13- FORMATION MECHANISM OF 2,4,5-TRIMETHYL-1,3-DIOXOLANE IN BEER...... 23

FIGURE 14- FORMATION OF HYDROPEROXY FATTY ACIDS 9-LOOH AND 13-LOOH BY AUTO-OXIDATION

OF LINOLEIC ACID ...... 25

FIGURE 15- PROTON-CATALYZED CLEAVAGE OF 9-LOOH AND 13-LOOH HYDROPEROXY ACIDS OF

LINOLEIC ACID, RESULTING IN (E)-2-NONENAL FORMATION...... 25

FIGURE 16- STABILIZATION OF PHENOXY RADICALS BY DELOCALIZATION...... 26

FIGURE 17- DEPICTION OF DRIED HOPE CONE CHEMICAL COMPOSITION...... 27

FIGURE 18- ALPHA-ACIDS AND THEIR RESPECTIVE ISOMERIZATION PRODUCTS…...... 28

FIGURE 19- CLASSIFICATION OF HOP ESSENTIAL OILS ACCORDING TO THEIR CHEMICAL COMPOSITION 29

FIGURE 20- CHEMICAL STRUCTURES OF SOME ODORANTS IN HOP ESSENTIAL OILS.(117) ...... 31

FIGURE 21- GC-MS APPARATUS USED FOR CARBONYLS ANALYSIS IN BEER SAMPLES...... 43

FIGURE 22- STEAM DISTILLATION APPARATUS...... 47

FIGURE 23- SOLID-PHASE EXTRACTION EQUIPMENT...... 47

FIGURE 24- ROTARY EVAPORATOR...... 48

FIGURE 25- COLUMN BAR CHART CONTAINING INFORMATION ABOUT THE DETERMINED CONCENTRATION

(MG/L) FOR EACH CARBONYL COMPOUND TESTED ON EVERY NON-PASTEURIZED (NP) AND

PASTEURIZED (P) 12% LAGER AND 20% IMPERIAL LAGER (IL) SAMPLE FROM EVERY SET OF

ANALYZED BOTTLES (1ST SET – BOTTLES STORED FOR 0 MONTHS; 2ND SET - BOTTLES STORED FOR

2 MONTHS; 3RD SET - BOTTLES STORED FOR 4 MONTHS)…...... 58

FIGURE 26- COLUMN BAR CHART CONTAINING INFORMATION ABOUT THE DETERMINED CONCENTRATION

(MG/L) FOR EACH CARBONYL COMPOUND TESTED ON EVERY NON-PASTEURIZED (NP) AND

PASTEURIZED (P) 12% LAGER AND 20% IMPERIAL LAGER (IL) SAMPLE FROM EVERY SET OF

ANALYZED BOTTLES (1ST SET – BOTTLES STORED FOR 0 MONTHS; 2ND SET - BOTTLES STORED FOR

2 MONTHS; 3RD SET - BOTTLES STORED FOR 4 MONTHS)...... 59

FIGURE 27- SPIDER WEB PLOTS REPRESENTING THE SENSORY PROFILE OF MONASTERY IPA SAMPLES

STORED FOR 0 MONTHS...... 63

FIGURE 28- SPIDER WEB PLOT CONTAINING COMBINED FLAVOR PROFILES OF MONASTERY IPA SAMPLES

STORED FOR 0 MONTHS……… ...... 64

FIGURE 29- SPIDER WEB PLOTS REPRESENTING THE SENSORY PROFILE OF MONASTERY IPA SAMPLES

STORED FOR 2 MONTHS...... 64

FIGURE 30- SPIDER WEB PLOT CONTAINING COMBINED FLAVOR PROFILES OF MONASTERY IPA SAMPLES

STORED FOR 2 MONTHS...... 65

FIGURE 31- SPIDER WEB PLOTS REPRESENTING THE SENSORY PROFILE OF MONASTERY IPA SAMPLES

STORED FOR 4 MONTHS...... 65

FIGURE 32- SPIDER WEB PLOT CONTAINING COMBINED FLAVOR PROFILES OF MONASTERY IPA SAMPLES

STORED FOR 4 MONTHS……… ...... 66

FIGURE 33- SPIDER WEB PLOTS CONTAINING COMBINED FLAVOR PROFILES OF EACH OF THE 4 DIFFERENT

MONASTERY IPA SAMPLES STORED FOR PERIODS OF 0, 2 AND 4 MONTHS……… ...... 66

FIGURE 34- CALIBRATION LINES GRAPHICS DESIGNED FOR 2-METHYLPROPANAL, 2-METHYLBUTANAL,

3-METHYLBUTANAL AND 2,3-PENTANEDIONE ……… ...... 92

FIGURE 35- CALIBRATION LINES GRAPHICS DESIGNED FOR HEPTANAL, FURFURAL, DIACETYL AND

OCTANAL ...... 48

FIGURE 36- CALIBRATION LINES GRAPHICS DESIGNED FOR BENZALDEHYDE AND (E)-2-NONENAL...... 93

List of Abbreviations

ABV – Alcohol by volume AEDA – Aroma extraction dilution analysis B.C. – Before Christ DNPH – 2,4-Dinitrophenylhydrazine EBC – European Brewery Convention Units EBU – European Bitterness Units PFBOA – o-(2,3,4,5,6-Pentafluorobenzyl)hydroxylamine GC – Gas chromatography GC-MS – Gas chromatography/Mass spectrometry engaged system HMF – 5-Hydroxymethyl furfural HPLC-UV – High-performance liquid chromatography/Ultraviolet mass detection engaged system HS-SPME GC-MS - Headspace solid-phase micro-extraction method in combination with gas chromatography/Mass spectrometry engaged system IL – Imperial lager IPA – Indian Pale Ale IS – Internal standard LOX – Lipoxygenases SPME- Solid-phase micro-extraction ppb – Parts per billion PU – Pasteurization Unit

1. Introduction – Literature Review

1.1 Brief history of beer

Beer is one of the oldest beverages to ever been produced by humans. The first written reports about its production are dated back from the fifth millennium B.C. coming from the ancient region of Babylonia, part of an area which belongs to the nowadays called Iraq.(1)(2) It was recorded as part of the history of ancient Egyptian and Mesopotamian civilizations, although it is believed that beer existed far before these written records.(2)

Mythologically speaking, the Egyptians accredited beer existence to Osiris, the god of agriculture. Legend says that Osiris seeded a plantation of barley which later germinated along the banks of the Nile river, which he later forgot and left it in the sun for a long period of time. Upon his return he found a fermented liquid. After trying it he proclaimed Mankind should benefit from his gift.(3) However, there are a lot of historical or mythical figures from numerous civilizations who credit themselves or are credited for the invention of beer, as it is the case of Flemish king

Gambrinus.(4)

It is most likely that the first beer that has ever been produced resulted from spontaneous natural fermentation of barley due to the presence of wild yeasts.(5) This is a common characteristic to almost any cereal containing sugar, meaning that beer-like beverages are directly associated to domesticated cereal cultivation cultures.(5)(6)

Eventually, beer made its way from the Middle East across the Mediterranean Sea to Europe, where it became part of current life. Beer as we know it nowadays, so called modern beer, is assumed to have been invented by German monks during the early Middle Ages when hops were firstly introduced as a bittering and flavoring agent and it quickly became a trend. Before that time, many different herbs and spices were used as an attempt to balance the sweet malted barley flavors in beer.(5)

Monks became the most remarkable brewers of the Middle Ages, as it was very frequent for monasteries to have a brewery on site, especially in Central and Northern Europe, as some of them still remain active. Furthermore, apart from the introduction of hops on the brewing process, monks were also credited for many brewing innovations such as the idea of lagering or cold storing for flavoring improvement, which are still currently used techniques.(5)

Ever since then, the brewing industry has become a huge global business whether through big multinational companies or smaller producers ranging from regional breweries to local brewpubs.(7) The widespread increasing business results on a more demanding and competitive

market, in which the search for quality and affordable products is constant.(8) Therefore, quality maintenance and reproducibility are significant distinguishing factors and a main concern for the leading exporting beer brands and are mostly related to storage conditions as beer aging appears to be the main issue affecting the quality of the product.(9)

1.2 Beer aging

The chemical composition of beer changes during its storage, which can alter the sensory properties such as aroma, color and taste. Unlike some wines, beer aging is usually considered to have a negative impact for its sensory quality as a variety of new different flavors may arise, depending on the beer type and storage conditions.(9)

The shelf-life of food products is an important feature for both manufacturers and consumers. Its evaluation is mostly aimed at assuring food safety, although it additionally concerns quality aspects including physical, chemical, and sensorial properties. Therefore, shelf-life studies can provide important information to ensure a high-quality product during its storage period.(10)

By definition, the shelf-life of food is the period during which the food retains an acceptable quality from a safety and organoleptic point of view. It usually relies on attributes related to formulation, processing, packaging and storage conditions. Being perishable products, foods can be deteriorated through a number of factors which can compromise their quality and safety. These can be categorized into chemical and physical factors.(10)

In order to minimize the degradation of food during processing or storage, kinetic models which describe degradation rates can be determined. These kinetic models are elaborated to relate the influence of external factors, such as temperature, on the intrinsic properties of each food. Their purpose is to be used for prediction, control, optimization and simulation of various food processing operations, allowing maximization of shelf-life.(10)

Alike other food products, beer is also vulnerable to alterations during storage periods. Its shelf-life is especially related to its microbiological, colloidal, foam, color and flavor stabilities. In the past, the appearance and growth of beer spoiling microorganisms was treated initially as the primary trouble causing phenomenon. However, the main concern of investigation works is currently focused on the factors affecting beer aroma and taste, being these considered as the most important parameters regarding the quality of the product, since technology development applied to brewing industry allowed assuring control over microbiological issues.(9)

Consumers expect the flavor of a certain beer brand to be constant every time. Nonetheless, if not consumed fresh, beer aging may be responsible for noticeable flavor changes or even flavor loss. Although several efforts have been made through numerous studies and

experiences, beer aging phenomenon remains difficult to control since multiple chemical changes and reactions may occur throughout storage time leading to beer staling and spoilage.(9)

In 1977, Dalgliesh developed a more detailed study regarding these changes which occur during storage, describing the variation of the intensity of different flavors and aromas present in beer in function of storage time.(11)

Figure 1- Sensory changes during beer aging according to Dalgliesh (1977).(11)

A few conclusions were reached through the graphic plot displayed on Fig. 1. During beer aging a constant decrease on bitterness intensity through time was detected throughout storage days, alongside with an equivalent constant increase in sweet taste, namely caramel and leathery aromas. Furthermore, a rapid emergence of what is designated as ribes flavor can also be observed, which is referred to a characteristic blackcurrant leaves odor. As time goes by, the intensity of ribes flavor decreases followed by a development of cardboard flavor, which increases as ribes aroma vanishes.(11) However, another study performed by Meilgaard pointed to a parabola type of development, constantly increasing its intensity until reaching a maximum, followed by a decrease, similarly to what occurs with ribes flavor.(12)

As a major point, a relation between the intensity decrease of positive flavor attributes of beer, such as fruity and floral aromas, and the formation and further increment of stale flavors can be established.(13)(14)

The term beer staling is generally used to describe solely the appearance and development of cardboard flavor. Despite being the main manifestation of staling in some cases,

especially in lager beers, this cannot be generalized as cardboard flavor is not the only indicator of beer staling. (9) Regarding other types of beers, there is a big variety of stale flavors which can be more or less prominent depending on the characteristics of each beer and respective aging process. In 1981, Whitear noted some burnt, alcoholic, astringent, liquorice and caramel flavors resultants from the aging of a strong ale, whereas cardboard and metallic flavors were not detected.(15)

Flavor deterioration is caused mainly due to contact of beer with oxygen and can happen as rapidly as higher it is the oxygen content of bottled beer. In fact, a close correlation seems to exist between the formation and growing intensity of ribes taste and headspace air, meaning that this flavor can be avoided in the absence of excessive contact with oxygen.(16) However, further studies proved that beer staling can still occur at very low oxygen levels, suggesting that beer staling phenomenon is in part a non-oxidative process.(17)

Another aspect to take into account which also plays a big role concerning beer aging characteristics are storage temperatures, as they directly affect the chemical reactions that occur within the bottle.(9) Nonetheless, the reaction rate does not increase constantly with temperature, as it depends on the activation energy of each reaction, which differs according to the reaction type. Therefore, storing beer at different temperatures does not necessarily generate the same relative levels of staling compounds.(9)

For example, a study driven by a Japanese team in 1999, showed that cardboard flavor intensity presents different time courses during lager beer storage at 20℃ and 30℃.(18) The results were indicative of a higher prominence of cardboard flavor at 30℃, which matched some previous findings from 1995, that suggested lager beer stored at 25℃ tends to predominantly develop caramel taste, whereas cardboard notes are more dominant when stored at temperatures between 30℃ and 37℃.(19)

1.2.1 Chemical changes in beer during storage

Deterioration phenomenon is the result of both formation and degradation reactions, which lead to the development and dissipation of certain flavors. In order to have noticeable effects, the formation of new compounds must reach concentrations above their respective flavor threshold, while in contrast, degradation of compounds must reach concentrations below their flavor threshold to cause losses of initial fresh beer flavors.(9) Nevertheless, interactions that may occur amongst different molecules, whether they are initially present in beer or not, can also have an effect on the resultant aroma and taste by suppressing or enhancing the impact of certain flavors, which frequently happens amidst volatile compounds.(20)

1.2.1.1 Volatile Compounds

Volatile compounds represent a big part of the composition of every beer. The introduction of gas chromatography (GC) techniques allowed a more detailed outlook on the changes in beer volatiles during storage, which were assumed to have a significant relation with the formation of staling compounds.(21)(22) More recently, development of new techniques such as aroma extraction dilution analysis (AEDA) permitted an improvement in order to detect volatiles in food products and also to evaluate their relevance for odor perception.(23) This method allowed the identification of several staling compounds in beer.(24)

1.2.1.1.1 Carbonyl Compounds

From the beginning, carbonyl compounds have drawn the most interest amongst volatile substances possibly responsible for beer staling, as they were previously known to cause flavor changes in other food products such as milk, oils, vegetables and butter.(9)

The majority of carbonyl compounds in beer have their origin on raw materials used for its production or emerge through chemical reactions, particularly in the pre and post fermentation stages of the production process. More precisely, carbonyls are mostly produced in heat-enhanced reactions that occur during kilning of malt, mashing, wort processing and pasteurization of the final product, although most of the ones formed during pre-fermentation stages are reduced by yeasts during fermentation, generating the correspondent alcohols and acetate esters.(25)

On Table 1 are presented a number of carbonyl compounds most frequently found in beer, as well as the chemical class they belong to, and Fig. 2 illustrates the chemical structure of some of the most relevant ones mentioned, relatively to beer aging phenomenon.

Table 1- Variety of carbonyl compounds and respective chemical classes mostly represented in beer.(26)

Chemical carbonyl class Compounds Propanal (I), Butanal, Pentanal, Hexanal, Saturated linear aldehydes Heptanal (VI), Octanal (VII), Nonanal, Decanal 2-Propenal, (E)-2-Butenal, (E)-2-Pentenal, (E)-2-Hexenal, Unsaturated linear aldehydes (E)-2-Heptenal, (E)-2-Octenal, (E)-2-Nonenal (VIII), (E,E)-2,4-Decadienal 2-Butanone, 2-Pentanone, 2-Hexanone, Linear methylketones 2-Octanone, 2-Nonanone, 2-Decanone

2-Methylpropanal (II), 2-Methylbutanal (III), Branched chain aldehydes 3-Methylbutanal (IV)

Phenyl aldehydes Benzaldehyde (XI), Phenyl acetaldehyde

Furfural (V), 5-Methylfurfural, Furan derivatives 5-Hydroxymethyl furfural, 2-Acetylfuran Hydroxycarbonyls Hydroxyacetone Glyoxal, Methylglyoxal, Diacetyl (IX), Dicarbonyls 2,3-Pentanedione (X) Methional, Ethylpiruvate, Methyl isobutyl Others ketone

Figure 2- Chemical structure of a few carbonyl compounds connected to beer staling mentioned on Table 2.

Concerning beer storage, in 1966 Hashimoto became the first ever researcher to report a remarkable increase in concentration of volatile carbonyls through time, in association with the development of the so called stale flavors.(27) Alkanals and alkenals were the most predominant formed compounds, being also those which revealed the highest influence in beer taste.(28) In particular, acetaldehyde was one of the first compounds for which a concentration increment was observed in aged beer (29), as a result of ethanol oxidation, and (E)-2-nonenal was initially described as the molecule inductive of cardboard flavor in aged beer.(30)(31) Since then, a big part of the following studies were focused on confirming the role of (E)-2-nonenal as a staling indicator, although there were used extreme storage conditions, as they are mostly referred to heated and very acidified beer (pH 2), which were used to facilitate the obtainment of detectable levels.(32)(33)

Under regular storage conditions, standard methods hitherto developed such as direct analysis by gas chromatography posterior to either headspace or standard solvent extraction were not able to be used, as a clear separation of (E)-2-nonenal from other more abundant volatile peaks was impossible to achieve.(9) In 1974, a group of researchers developed a method which enabled to follow the variation of (E)-2-nonenal under normal storage conditions, which consisted on extracting beer with dichloromethane, followed by derivatization of (E)-2-nonenal with 2,4-dinitrophenylhydrazine (DNPH) under acidic conditions before the final chromatographic steps.(34) This method provided some interesting results as (E)-2-nonenal levels increased above the flavor threshold.(34) Further research suggested that the formation of (E)-2-nonenal in aging beer was due to aldol condensation of acetaldehyde with heptanal, in a reaction catalyzed by amino acids (Fig. 3).(35)

Figure 3- Formation of (E)-2-nonenal in beer by aldol condensation of acetaldehyde and heptanal.(35)

Various similar techniques were since developed aimed at verifying the increasing concentration of other linear C4-C10 alkanals and alkenals during beer storage, although there were used different approaches for isolation of compounds.(36)(37) Many of these methods remained based on derivatization of carbonyl compounds, in order to surpass the interference caused by the beer matrix.(9) The only variable feature between them was the use of different derivatization agents on beer extracts, being o-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine (PFBOA) and hydroxylamine hydrochloride the most frequently used ones, preceding GC-MS

(gas chromatography/mass spectrometry engaged system) analysis.(38)(39)

PFBOA is currently the most widely used derivatization agent, partly because the natural pH of beer (≈ 4.5) is favorable for the generation and further determination of carbonyl compounds as PFBOA derivatives.(37) Another advantage of PFBOA usage is that the derivatization reaction with this agent occurs in aqueous phase, dismissing the need of any previous separation technique for carbonyl compounds.(38) Furthermore, detection limits in the range of few ppb can be easily obtained using detection modes such as electron capture detection and mass spectrometry with electric impact ionization or negative chemical ionization.(40)(41)(42)

The derivatization reaction that occurs between carbonyl compounds and PFBOA in aqueous solution is represented below on Fig. 4.

Figure 4- Representation of the derivatization reaction scheme involving carbonyl compounds and PFBOA agent.(38)

PFBOA reacts with the carbonyl group of each compound originating the corresponding oximes. For the entirety of carbonyl compounds, except for symmetrical ketones and formaldehyde, two geometrical isomers are formed, as it is displayed on Fig. 4. However, diones can generate several isomers depending on the spatial arrangement of the resulting derivatives on the multiple reactive carbonyl groups, which is the case of 2,3-pentanedione that originates eight isomers with PFBOA and 2,3-butanedione which forms five, being these two of the most frequently found diones in aged beer.(37) Regarding analysis, in gas chromatography two isomers

of linear alkanals can be separated and further quantification can be made by integrating the area underneath both peaks and adding them one to another, which corresponds to the total quantity of reaction products for each carbonyl.(38)

More recently, other methods using solid-phase micro-extraction (SPME) or stir bar and absorptive extraction with on-site PFBOA derivatization of carbonyls have been developed, allowing higher yields and a more efficient process.(43)(44) Alternative techniques include extraction of carbonyls by vacuum or steam distillation followed by GC-MS analysis for the first case or solid-phase extraction with HPLC-UV detection for the second presented alternative.(45)(46)

The results of various research works aimed at figuring out which carbonyls have the most influence regarding beer staling and emergence during aging process are not consistent, and in some cases are rather contradictory.(9)

A number of publications mention no significant variation in (E)-2-nonenal during beer aging, which was previously assumed to be responsible for cardboard flavor development.(47)(48)(49) In contrast, other authors observed its continuous formation during storage, however stating that it occurs independently of the oxygen content in bottled beers.(45)(46) Notwithstanding, Hashimoto provided evidences that carbonyl compounds are in fact important for flavor staling by conducting an experimental work in which carbonyl scavengers were used, such as hydroxylamine, and the results demonstrated a considerable diminish in certain aspects of aging flavor in beer.(50)

1.2.1.1.1.1 Alkanals and alkenals

Despite most of the attention being drawn towards (E)-2-nonenal and its impact on flavor stability, other studies were performed having as main goal evidencing the concentration increase of other alkanals and alkenals upon beer storage. Greenhoff was able to demonstrate that the levels of all linear C4-C10 2-alkenals increased above their flavor thresholds, particularly long-chained ones starting from 2-heptenal.(36) Another report claimed that alkadienals, namely

(E,Z)-2,6-nonadienal and (E,E)-2,4-decadienal, also take part in flavor staling.(51)

1.2.1.1.1.2 Strecker aldehydes

Strecker aldehydes are also commonly formed during beer storage.(49)(52) This class of carbonyls includes compounds such as 2-methylpropanal, 2-methylbutanal, 3-methylbutanal, benzaldehyde, phenylacetaldehyde and methional. They are known to have aroma-active properties, like malty aroma for the case of 3-methylbutanal and flowery and sweet honey-like flavors provided by benzaldehyde, which are relevant contributors for the sensory profile of aged beer.(53) As Strecker aldehydes concentration is directly related to oxygen contact, those which

do not hold any relevant aroma characteristics regarding stale flavor formation can be used as suitable markers for beer oxidation.(54)(55)

The formation of Strecker aldehydes in beer is catalyzed by the reaction between 훼-dicarbonyls, which are generated through sugar degradation reactions that occur in pre-fermentation stages, and 훼-amino acids which are mainly derived from barley and malt.(53)(56) Each Strecker aldehyde has a different 훼-amino acid precursor and its formation involves deamination and decarboxylation of the given 훼-amino acid in the presence of 훼-dicarbonyl compounds.(57) Cysteine, valine, leucine, isoleucine, methionine and phenylalanine originate acetaldehyde, 2-methylpronanal, 3-methylbutanal, 2-methylbutanal, methional and phenylacetaldehyde/benzaldehyde, respectively.(57)

The products of Strecker reaction are thus an 훼-aminoketone and a Strecker aldehyde which contains one carbon atom less than the corresponding amino acid.(57) Its mechanism is showed below on Fig. 5.

Figure 5- Schematization of the Strecker reaction mechanism which gives rise to important flavor-active aldehydes.(57)

1.2.1.1.1.3 Ketones and vicinal diketones

As far as ketones are concerned, the main focus was given to carotenoid-derived compounds such as 훽-damascenone. Although a significant increase was not detected, carotenoid-derived flavors were already suspected to be staling components.(58) As for 훽-damascenone, its development brings up sweet red fruit aromas which are related to the decrease of beer bitterness intensity.(59)

In the presence of high oxygen levels, the formation of vicinal diketones, predominantly 2,3-butanedione (also known as diacetyl) and 2,3-pentanedione, has proven to be an issue, as

they can easily surpass their flavor thresholds in oxidized beer.(60) Diacetyl is responsible for the emergence of an intense buttery taste, as well as caramel flavor, while 2,3-pentanedione contributes more honey-like notes to beer.(61)

Monitoring and controlling vicinal diketones levels plays an important part in flavor changes during beer maturation because these compounds can be detected by humans at fairly low levels, given their low flavor thresholds, and additionally, an excessive concentration of vicinal diketones in beer can be a sign of improper fermentation or infection caused by bacteria or wild yeasts.(61)

Vicinal diketones are produced during fermentation and are a result of a long-chain reaction. The precursors of these compounds are released by yeast cells during amino acid synthesis. Valine synthesis produces 훼-acetolactate, while isoleucine synthesis forms 훼-ketobutyrate. These precursors subsequently break down into vicinal diketones (Fig. 6). However, during beer maturation diacetyl and 2,3-pentanedione levels may decrease as yeasts are able to metabolize them as an energy source, reducing them into the respective alcohols. Nonetheless, yeast removal or inactivation before maturation stage may result in a significant presence of vicinal diketones in beer before consumption.(61)

Figure 6- Schematization of the pathways for diacetyl and 2,3-pentanedione formation during yeast metabolism and subsequent reduction, as well as valine and isoleucine synthesis.(62)

Reduction of vicinal diketones levels is one of the main goals of traditional lagering, as they are most easily detectable in lighter beers where the flavor is not masked by stronger malt and hop flavors.(62)

Brewers have developed several tools in order to minimize their effect, as it is the case of kräusening method. This method consists on adding actively fermenting beer to maturing beer, allowing fresh yeast to consume vicinal diketones in addition to contributing to natural carbonation.(61) Alternative biological techniques have also been created, more substantially in order to avoid extensive maturation times, such as the addition of an enzyme isolated from Acetobacter bacteria to wort, which is able to entirely prevent diacetyl formation simply by converting 훼-acetolactate directly into flavorless acetoin.(61)

However, the most commonly used method is related to yeast strain selection. Some strains produce only small amounts of vicinal diketones and are capable of metabolizing them regardless of their concentration, whereas others are not. Hence, genetically modifying yeast strains are a frequently used solution to overcome this issue.(62)

Other classes of volatile carbonyl compounds subject to concentration increases during beer storage include cyclic acetals, heterocyclic compounds, esters and sulfur compounds. These components are particularly indicative of beer oxidation, although their influence regarding sensory changes is variable, depending on their organoleptic properties and subsequent flavor thresholds in beer.(9)

1.2.1.1.1.4 Cyclic acetals

Dioxolanes are the main subclass of cyclic acetals which can be found in beer, with particular focus to 2,4,5-trimethyl-1,3-dioxolane and 2-isobutyl-4,5-dimethyl-1,3-dioxolane isomers.(63) Despite being already present in unboiled wort, these compounds are lost by evaporation during wort boiling. The formation of this type of dioxolanes is directly related to enzymatic reduction of vicinal diketones as they are a result of condensation reactions between alcohols and aldehydes, enhanced by acidic conditions.(64) More precisely, 2,3-butadienol generated from diacetyl reduction may react with 3-methylbutanal, giving rise to 2-isobutyl-4,5- dimethyl-1,3-dioxolane, or with acetaldehyde, originating 2,4,5-trimethyl-1,3-dioxolane, which are compounds commonly found in matured beer.(64)

These compounds have a flavor threshold of around 0.9 ppm, at which concentration they can produce phenolic and astringent notes to beer. However, they are never present in sufficiently high concentrations (maximum concentration reported was 0.1 ppm) to make a significant 64) contribution to the flavor of beer.(

1.2.1.1.1.5 Heterocyclic compounds and Maillard reaction

Regarding heterocyclic compounds with carbonyl functions, a role as sensitive indicators of beer flavor deterioration has been assigned, despite being generally found in concentrations far below their flavor thresholds.(65)(66) Within this class of chemical compounds, furfural and 5-hydroxymethyl furfural (HMF) have a particular relevance, as their levels reportedly increase in aged beer with time at an approximately linear rate, which varies logarithmically with storage temperatures.(67)

The increase of furfural and HMF levels in aged beer seems to occur independently from oxygen concentration, although they appear to be directly related to sensory changes, namely flavor staling, meaning that furfural and HMF cannot be used as suitable markers for oxidation, but they can serve as indicators of heat-induced flavor damages to beer.(68)

The emergence of furfural and HMF in beer is clearly correlated with alcohol content and color (which is derived from the types and content of malt used for beer production), as furfural is present in higher concentrations in high alcoholic and dark beers, whilst its levels in lager blonde beers are substantially lower.(69)

Furfural and HMF are produced through several routes of the Maillard reaction. The chemistry underlying the Maillard reaction is rather complex, as it encompasses multiple reactions in chain.(70) The starting point of this reactional mechanism consists on a condensation reaction between a reducing sugar and a compound containing a free amino group. The condensation product is a N-substituted glycosilamine, which rearranges to form an Amadori intermediate product. Further degradation of Amadori products can go several pathways, depending on the pH of the system. Under acidic conditions, it undergoes mainly 1,2-enolisation with formation of furfural, if the reducing sugar involved in the first step is a pentose such as xylose or arabinose, or HMF when hexoses such as glucose or mannose are involved. Under alkaline conditions, the degradation of Amadori intermediates is thought to involve 2,3-enolisation, which gives rise to reductones and a variety of fission products that include dicarbonyl compounds which undergo further Strecker degradation.(70) Other pathways for furfural and HMF formation do not involve amino compounds nor aminocarbony reactions.(71)

Fig. 7 displays a comprehensive scheme of Maillard reaction and its multiple routes developed by Hodge(72) and Fig. 8 represents the reactional mechanisms of two different pathways for the formation of furfural and HMF.

Figure 7- Maillard reaction scheme developed by Hodge.(72)

Figure 8- Two different pathways for furfural and HMF production in beer, respectively from pentoses and hexoses.(71)

Similarly to Strecker degradation, amino acid synthesis also has an important role for Maillard reaction occurrence in beer. In this case, amino acids come up as the main providers of amino groups which react with reducing sugars in aminocarbony pathway (Fig. 8). Thus, one of the most obvious negative consequences of Maillard reaction in beer is related to the loss of nutritive value, as it causes decrease of protein levels.(70)

Moreover, another clear effect of Maillard reaction is color intensification. The degree of browning and darkening of a beer increases with the extent of Maillard reaction. In the final stage of the reaction, colored intermediates and other reactive precursors condense and polymerize, in a reaction catalyzed by amine presence, generating brown polymers known as melanoidins

(Fig. 7).(70) Hence, it can be concluded that browning of beer is directly proportional to the amount and reducing power of the sugars involved, which is also connected to alcohol levels in beer.(73)

Sugar degradation occurs during mashing stage of beer production. Thus, temperature and heating time play a big role in mashing efficiency and therefore in the formation of Maillard products.(70) Maillard himself, reported that the rate of the reaction increases with temperature.(74)

An increase in temperature leads to an increase of the reactivity between the sugar and the amino group.(74)

Regarding beer storage, the principle is the same. The occurrence of Maillard reaction during storage is related to the fact that yeasts are mostly able to reduce furfural and HMF during fermentation. Thus, the concentration of the precursor remains nearly unchanged before and after fermentation, leading to the formation of furfurals during maturation stage, especially when beer is stored at high temperatures (≤ 27℃).(71)

The pH factor also has a crucial influence concerning the balance of the various chemical pathways of the Maillard reaction.(70) Similarly to temperature, the reactivity of the sugar and amino group is also highly influenced by pH. Amino groups are more reactive at their unprotoned form, while reactant sugars have a higher reactivity at their open-chained form. Both of these forms are favored at higher pH. However, the natural pH of beer is acidic, thus promoting furfural and HMF pathways.(70) The pH of beer is closely related to the malt ratio in the wort, being a higher pH level resultant from a higher malt content.(71) Therefore, upon maturation stage, lower pH values enhance beer staling.(71)

Methods to avoid furfural and HMF formation and subsequent flavor damage, include anti-oxidative mashing and yeast strain selection.(71) Both of these regulating methods have the goal of increasing the pH of the system and therefore, flavor stability. Anti-oxidative mashing consists on using carbon dioxide during mashing stage to insulate the mash from air, preventing oxygen contact which is responsible for oxidative reactions that promote a lower pH.(71)(75) In contrast, yeast strain selection can be used to promote higher pH levels in beer by selecting strains with higher cellular volume, as cell size was shown to have an effect on this parameter.(71) Therefore, beer produced by yeast strains with bigger cell sizes tends to have higher pH values upon fermentation, hence inhibiting furfural and HMF pathways of Maillard reaction, lowering their levels during beer maturation and contributing for flavor stability.(71)

Other furan-derived compounds can also be formed during aging of beer, such as furfuryl ethyl ether and furanones.(68)(76) Identically to furfural and HMF, furfuryl ethyl ether can also be used as a heat-induced flavor damage indicator in beer and may increase to levels above its flavor threshold during storage.(68) Regarding furanones, there is no data available on their importance for beer staling.(76)

1.2.1.1.1.6 Volatile esters

Volatile esters are another class of chemical compounds with a significant level of importance during beer aging process, as they are responsible for fruity notes and other positive flavor attributes of fresh beer.(9) In opposition to every other class of components previously

mentioned, some esters decrease their concentration during beer storage, to levels that go below their flavor threshold, which results in losses of positive taste notes.(76) This is the case of isoamyl acetate, produced by yeasts during fermentation, which gives a banana-like flavor to beer.(77)

The decrease of isoamyl acetate concentration during beer aging is due to hydrolysis. During storage, the equilibrium is shifted to the formation of acetic acid and isoamyl alcohol as it is demonstrated on Fig. 9.(69)

Figure 9- Isoamyl acetate formation and degradation reactions.(78)

As a consequence of isoamyl acetate hydrolysis during beer aging, fruity flavors initially present in beer (more characteristic of specialty ales) may disappear, which decreases the intensity of background flavors, increasing the sensory perception of eventual stale flavors.(69)

In contrast to hydrolysis, esterification reactions may also occur between carboxylic acids and alcohols in aging beer, increasing the levels of the respectively formed esters. This is the case of ethyl cinnamate, which adds a cinnamon-like taste to beer, 2-methyl-butyrate, which is associated to the development of winy flavors, and also some lactones with fruity aromas.(79)(80)(81) These reactions take place in acidic environments, thus relying on the pH level of the finished product.(69)

1.2.1.1.1.7 Sulfur compounds

Sulfur compounds are not usually present in high concentrations, however they generally have an extremely low flavor threshold in beer and small concentration changes may have noticeable effects on its flavor.(9) Dimethyl trisulfide has reportedly reached levels above its flavor threshold in aged beer.(82) This compound is associated to onion-like and ribes flavor emergence in beer.(82)

1.2.1.1.1.8 Variation of carbonyls concentration during beer aging

Fig. 10 disposes several graphic representations (A-E) of the variation of certain carbonyl compounds concentration (in 휇g/L) during a period of one year of beer aging.(69) This experience was conducted in 8 different types of Belgian commercial beers, including 3 pilsner beers (L-A, L-B and L-C) and 5 specialty ales (S-A, S-B, S-C, S-D and S-E). The most relevant properties of each beer are presented on Table 2. All beers were stored under the same conditions at a temperature of 20℃. The evolution of the color intensity of each beer within this period, measured in EBC’s, is represented in graphic F.

Figure 10- Concentrations of 3-methylbutanal (A), 2-isobutyl-4,5-dimethyl-1,3-dioxolane (B), furfural (C), furfuryl ethyl ether (D) and diacetyl (E) and the evolution of beer color (F) during storage of eight beers for one year at 20°C.(69)

Table 2- Alcohol percentage, color, pH and bitterness of 8 commercial beers tested: 3 lager beers, L-A, L-B, L-C and 5 specialty beers, S-A, S-B, S-C, S-D, S-E.(69)

Beer Type Alcohol (%v/v) Color (EBC) pH Bitterness (EBU) L-A 5.34 10.3 4.37 25.0 L-B 5.09 9.2 4.61 23.2 L-C 5.24 10.9 4.56 20.0 S-A 5.44 30.5 4.29 20.2 S-B 7.82 18.3 4.39 16.4 S-C 8.84 47.4 4.44 18.6 S-D 9.77 16.7 4.44 32.6 S-E 10.60 16.4 4.30 20.6

1.2.1.1.1.9 Summary

A summary of some the most important carbonyl aging markers in beer and the respective type of aging reaction involved in their formation during storage is presented on Table 3. Most of these compounds were analyzed in beer samples used for this experimental work.

Table 3- Carbonyl aging markers in beer and the type of reaction involved in their formation during storage.

Carbonyl aging marker Aging reaction Acetaldehyde Oxidation of ethanol (E)-2-Nonenal Aldol condensation 2-Methylpropanal Strecker degradation of amino acids 2-Methylbutanal Strecker degradation of amino acids 3-Methylbutanal Strecker degradation of amino acids 2,4,5-Trimethyl-1,3-dioxolane Aldehyde acetalization

Oxidation of excretion products from yeast Diacetyl metabolism

Oxidation of excretion products from yeast 2,3-Pentanedione metabolism

5-Hydroxymethyl Furfural Maillard reaction Furfural Maillard reaction Etherification of ethanol and Maillard Furfuryl Ethyl Ether compounds

Isoamyl Acetate Hydrolysis of esters produced by yeasts Benzaldehyde Strecker degradation of amino acids

1.2.1.2 Reactive oxygen mechanisms of aging process in beer

In terms of chemistry, beer can be considered as a water-ethanol solution with a pH value of around 4.2, in which several different molecules are dissolved. These molecules are originated from the raw materials used for beer production (water, malt, hops, adjuncts) and the wort production, fermentation and maturation processes, which comprehend various chemical reactions.(9)

Nonetheless, the components of freshly bottled beer are not in a chemical equilibrium, thus being constantly subject to several chemical conversions during storage resultant from its thermodynamic activity. These reactions eventually determine the type of the aging characteristics of beer, being as relevant accordingly to the rates at which they occur under certain storage conditions. Reaction rates are dependent of the substrate concentrations and rate constants, which differ between reaction types and level of temperature influence.(9)

Some of the most important aging reactions are initiated by oxygen, as it causes a rapid deterioration of beer flavor and it is inevitably present in bottled beer.(83) In the ground state (3O2), oxygen is very stable and will not easily react with organic compounds, but in the presence of iron/copper ions in beer, oxygen can capture an electron and form highly reactive superoxide anion (O2-), while iron/copper is oxidized.(84)

However, the superoxide anion can be protonated and form the perhydroxyl radical

(OOH.). This reaction has a pKa of 4.8, meaning that the majority of superoxide will be in perhydroxyl form at the normal pH of beer. In contrast, the presence of iron/copper ions can induce the formation of peroxide anion (O22-) by reducing superoxide. Peroxide anion is immediately protonated in beer to hydrogen peroxide (H2O2), which can lead to the formation of hydroxyl radicals (OH.) by specific metal-induced reactions such as Fenton and Haber-Weiss reactions.(84)

Oxygen species have increasing reactivity according to their reduction status (superoxide anion < perhydroxyl radical < hydroxyl radical) and the concentration of free radicals in aging beer relies on iron/copper and oxygen concentrations, and also increases with higher storage temperatures.(84) Furthermore, the generation of free radicals in beer may not occur from the beginning of the aging process, starting only after a definite period of time, called lag time, which is related to the anti-oxidant activity of beer.(85)

The reaction chain mechanism responsible for the production of reactive oxygen species in beer is represented on Fig. 11. These reactive oxygen species are capable of further reacting with all types of organic compounds present in beer, resulting in various alterations in its sensory profile.(9)

Figure 11- Reactions producing reactive oxygen species in beer.(84)

Reactive oxygen species are in the basis of all oxidation reactions that occur during beer storage. The findings of Andersen and Skibsted in 1998, supported the importance of the formation of hydroxyl radicals for flavor damaging, as one of the most reactive species that has been identified.(86) In beer, these radicals can react with ethanol, which is the second most abundant compound and a good radical scavenger, and form acetaldehyde alongside with other minor products (Fig. 12).(86)

Figure 12- Reaction between ethanol and hydroxyl radicals in beer according to Andersen and Skibsted (1998).(86)

As mentioned above, acetaldehyde is also relevant for the formation of staling indicator (E)-2-nonenal, since the pathway involving aldol condensation with heptanal (Fig. 3) was shown to be much more plausible for the formation (E)-2-nonenal than oxidation of 2-nonenol, due to decreasing reactivity of alcohols with their molecular weight, resulting in a very low conversion rate for this reaction.(87)

Also directly related to acetaldehyde formation from ethanol oxidation are aldehyde acetalization reactions. As previously mentioned, cyclic acetals like dioxolanes are originated during beer storage from condensation between aldehydes and alcohols, as it is the case of 2,4,5-trimethyl-1,3-dioxolane, which results from a condensation reaction between acetaldehyde and 2,3-butanediol (Fig. 13).(64)

Figure 13- Formation mechanism of 2,4,5-trimethyl-1,3-dioxolane in beer.(64)

1.2.1.2.1 Oxidation of unsaturated fatty acids

Lipid oxidation was also deeply studied as it was assumed to be linked to beer staling. Oxidative breakdown of lipids was suggested to lead to the formation of saturated and unsaturated aldehydes, namely (E)-2-nonenal, and other carbonyl compounds.(88)(89)

In beer and wort, the most significant lipid substrates arise from malted barley, as it is the case of linoleic acid (C18:2) and linolenic acid (C18:3), which are mainly released from

triacylglycerols by the activity of lipases.(90) Hydrolysis of triacylglycerols to fatty acids occurs during mashing, as malt lipases remain active throughout most of this process.(91) There is no clear evidence that lipid oxidation does not occur after beer bottling, although it is mostly linked to earlier production processes.

However, some oxidation intermediates of linoleic acid have been pointed as

(E)-2-nonenal precursors.(92) In fact, degradation of linoleic acid can result in (E)-2-nonenal formation but only in very acidified beer (pH≈2).(93) This does not exclude the possibility of the potential created by enzymatic and non-enzymatic oxidation of linoleic acid to enable the generation of (E)-2-nonenal during beer storage, since evidences were found that in wort, (E)-2-nonenal forms Schiff bases (imines) with amino acids or other proteins, which pass into beer and is then released during storage in a reaction enhanced at low pH.(93)(94)

Nonetheless, this is not the main source for (E)-2-nonenal emergence during beer maturation, being a non-oxidative aldol condensation pathway the most significant for the formation of this compound.(95) Oxidation of linoleic acid comes up as a secondary pathway for (E)-2-nonenal formation, which can occur through two separate oxidation routes: an enzymatic oxidation with lipoxygenases (LOX), which occurs predominantly during malting and mashing stages, or an auto-oxidation pathway (Fig. 14).(96)

Auto-oxidation of linoleic acid is initiated by the removal of a hydrogen atom from oxygen reactive species. In this case, it is more likely to be initiated by peroxy radicals (ROO.) as hydroxyl radicals are more likely to react with more abundant components in beer, such as sugars.(9) The methylene group of linoleic acid in position 11 is then activated due to its neighboring double bonds, leading to the formation of a pentadienyl radical after proton removal, which is further stabilized by the formation of two hydroperoxy acids at positions 9 and 13 (9-LOOH and 13-LOOH). These hydroperoxy acids may be then subject to degradation processes leading to a variety of volatile compounds relevant for flavor staling, such as (E)-2-nonenal (Fig. 15).(9)

Figure 14- Formation of hydroperoxy fatty acids 9-LOOH and 13-LOOH by auto-oxidation of linoleic acid.(97)

Figure 15- Proton-catalyzed cleavage of 9-LOOH and 13-LOOH hydroperoxy acids of linoleic acid, resulting in (E)-2-nonenal formation.(98)

1.2.2 Inhibition of flavor degradation

Beer aging is a result of oxidative and non-oxidative processes, which involve formation and degradation reactions, responsible for generating flavor damaging products. Beer staling is often regarded as solely the result of oxidation, although non-oxidative processes may be just as important and more difficult to control.(9) Thus, beer staling can occur either in the presence or

absence of oxygen. However, the most challenging part of comprehending beer aging phenomenon remains evaluating the relevance and extent of the specific reported reactions that take part in this overall process and their respective degree of significance upon certain storage conditions.(9)

Notwithstanding, several studies have been performed in this direction and some mechanisms that occur were already decoded, allowing the development of prevention techniques for flavor degradation and improvement of flavor stability.

Methods to control flavor deterioration include usage of chelating agents for capturing metal ions responsible for the activation of molecular oxygen and subsequent transformation into reactive oxygen species.(9) Some of the most frequently used chelating agents are polyphenols, amino acids and melanoidins.(99) Polyphenols, originated mostly from malted barley, are important anti-oxidants as they are also able to react with free radicals to produce phenoxy radicals which are stable due to delocalization of the free radical over the aromatic ring, therefore preventing further oxidation reactions (Fig. 16).(9) In particular, lower molecular weight polyphenols show more effectiveness as anti-oxidants, because they have a higher reducing power.(100) However, some polyphenols can have a contrary effect and function as pro-oxidants due to their ability to transfer electrons to transition metal ions.(17)

Figure 16- Stabilization of phenoxy radicals by delocalization.(9)

Another commonly used anti-oxidant for beer production is sulfite, as it postpones the formation of free radicals with effectiveness.(101) Sulfite is naturally generated by yeast metabolism of sulfate provided by raw materials and can also inhibit Maillard reactions, and therefore furfural formation, meaning that sulfite provides a very suitable option as an inhibitor of beer staling.(102)

Nevertheless, yeast straining selection and temperature control at any stage of beer production process play a very big role in decreasing the amount of staling components and their precursors in the final product.(9) While higher temperatures can promote staling reactions at any point of the process, yeast size has shown to be correlated to the pH level of the final product and yeast reducing power has proven to be an important variable for staling compounds degradation, as it can generate the reduction of aldehydes to alcohols during fermentation and the conversion of other flavor-active compounds originated during earlier stages of the brewing process.(103)

1.2.3 Hop-derived compounds

Hops (Humulus lupulus) belong to the Humulus genus and Cannabaceae family and are one of the four essential ingredients in beer along with barley, water and yeast.(104) Hop plants are cultivated in moderate climate regions with specific requirements related to temperature, daylight exposure, rainfall and soil fertility, existing over one hundred different varieties of hops grown all around the world.(104)(105)

Hops have a significant degree of importance regarding beer flavor stability, preservation and organoleptic properties.(105) Originally, hops were introduced in beer production process to inhibit the growth of microorganisms that could spoil beer.(106) Nowadays, hops are also added to impart bitterness and pleasant aromas to beer.(107)

Generally, hops are classified into two broad categories based on the main purpose of usage: bittering hops and aroma hops. Further division of bittering hops is also commonly done, with respect to their chemical composition, which is based on the content of 훼-acids, responsible for bitterness.(105)(107)

Hops are characterized by a rather complex chemical profile, containing a wide variety of compounds with different properties (Fig. 17). However, for brewing purposes the main focus is driven towards hop resins, essential oils and polyphenols, which contain the precursors of bitter and other flavor compounds.(105)

Figure 17- Depiction of dried hope cone chemical composition.(108)

1.2.3.1 Resins

Resins contain 훼-acids, thus being directly linked to the intensity of bitter taste of the final product. Moreover, hop bitter acids influence beer stability by dissipating the transmembrane pH gradient, resulting in growth inhibition of some microorganisms affecting beer quality, such as lactic acid bacteria.(106) The main species of 훼-acids are humulone, cohumulone and adhumulone. However, during wort boiling they are converted into their respective iso-훼-acids (Fig. 18) which are far more soluble in wort and possess more bitterness compared to their non-isomerized precursors.(108)

Nonetheless, 훼-acids and iso-훼-acids can undergo oxidation during beer maturation.(109) The degradation of these hop bitter acids not only decreases sensory bitterness but also results in the formation of products involved in the appearance of aging flavors.(110) Iso-훼-acids are particularly sensitive to oxidation during storage, as they rapidly degrade in the presence of oxygen reactive species.(111) However, not all isomers are equally subject to oxidation, being trans isomers much more sensitive than cis isomers.(112) Hence, the concentration ratio of trans/cis isomers was proposed as a good marker for flavor deterioration of beer.(113)

Figure 18- α-acids and their respective isomerization products.(105)

The mechanisms for hop acids degradation have not been completely elucidated. However, the oxidation products are known to be humulinones, which are more polar and less bitter than the corresponding iso-훼-acids, and volatile carbonyl compounds with various chain lengths.(114)(110) Moreover, other research evidenced that electrons are released from iso-훼-acids in the presence of suitable electron acceptors, which do not necessarily require involvement of oxygen species, meaning that iso-훼-acids can be subject to oxidative-type degradation in the absence of molecular oxygen.(115)

1.2.3.2 Essential oils

The hop components that are volatile in steam are part of the hop essential oils, which have significant aroma and flavor-active properties and are very prone to changes in their composition during storage.(105) There are multiple classification systems regarding the composition of hop essential oils, which comprise solely around 3% of the gross composition of the dried hop cone.(116) Fig. 19 illustrates one possible approach for essential oils sub-categorization.

Figure 19- Classification of hop essential oils according to their chemical composition.(117)

The most abundant and important components from the aroma point of view are terpenes, which are divided into two groups: monoterpenes and sesquiterpenes.(117) Considering monoterpenes, the main compounds that belong to this sub-class of hop essential oils are β-myrcene, limonene, α-phellandrene, β-phellandrene, p-cymene, sabinene, α-pinene and

β-pinene.(117)(118) Concerning sesquiterpenes, the most abundant ones are the isomers

α-humulene (or α -caryophyllene) and β-caryophyllene, and β-farnesene.(117)

Hop oils also contain a variety of oxygenated compounds including aldehydes, ketones, alcohols, esters, acids and epoxides.(117) Odor-active terpene ketones such as β-damascenone occur in hop essential oils although in very small amounts, while 2-undecanone is often reported as the most abundant ketone in hop oils.(119) The presence of aldehydes in essential oils is also regarded as very limited and dependent of the ripeness of the hop cone, as compounds like furfural are only detected in fully ripe cones.(120) Other species of aldehydes commonly found in hop essential oils are hexanal, heptanal and octanal.(121)

Linalool, geraniol, myrcenol, nerol, humulenol-I, humulenol-II, and humulol are the terpene alcohols frequently detected in hop oils and matured beer, although from these, linalool and geraniol are original components of the essential oils, providing potent fruity and flowery aromas.(117)(118) Many esters of terpene alcohols are found in hop oils, such as methyl geranate, methyl nerolate and geranyl acetate, as well as methyl esters of various acids, originated from fatty acids biosynthesis in hop cones.(123)(124) Notwithstanding, similar compounds are also produced during fermentation, making it difficult to determine whether these hop-derived esters survive every stage of the brewing process and are the ones detected in the final product or not.(124)

As previously mentioned, hop essential oils are prone to many changes in their composition and are relatively easily degradable, which can cause significant flavor loss.(117) Oxidation and polymerization reactions often occur during the aging of hop extracts, leading to deterioration of hop oils.(117) These reactions take place mainly during beer storage, although they may also happen in other stages of the brewing process, as it is the case of myrcene, which undergoes auto-oxidation and polymerization throughout the whole process.(125)

Caryophyllene oxide and humulene epoxides I, II and III are generated by oxidation of the respective sesquiterpenes (β-caryophyllene and α-humulene isomers).(120) In the case of α-humulene, its degradation reaction can even occur at room temperature, being highly susceptible to oxidation under regular storage conditions.(126)

However, the presence of some particular hop-derived aroma-active compounds in beer depends not only on the storage conditions and subsequent reactions that occur in this stage, but also on the hop addition method and fermentation, since yeasts are capable of promoting biotransformation reactions of hop compounds (essentially monoterpenoids), such as conversion of geraniol into linalool and β-citronellol.(105) Many hop essential oils components survive into beer whereas others are generated through degradation reactions.(116) Nonetheless, the hoppy character of the beer is not attributed to a single compound and it is instead a complex mixture of aromas caused by the presence of several volatile substances, each with different properties, which define its sensory profile.(105) In some cases, these volatile substances act in synergy, enhancing the perception of a certain flavor. For instance, compounds such as geraniol and

β-citronellol, which consists on a reduced form of nerol, while coexisting with excess linalool, can produce a more intense citrus flavor in beer, affecting it positively.(127)

Amidst all these compounds, Fig. 20 represents the chemical structures of a few aroma-active compounds founds in hop oils.

Figure 20- Chemical structures of some odorants in hop essential oils.(117)

1.3 Pasteurization in beer production

Since the 18th century that methods based on using heat to increase the shelf-life of food products and beverages have been used.(128) The first scientist to reportedly use such methods for this purpose was Swedish chemist Carl Wilhelm Scheele and later on, in the 19th century, French scientist Louis Pasteur invented this industrial process, at that time aimed for preventing beer and wine spoilage.(128)

At the beginning Pasteur used this treatment solely for wine closed in wooden containers and did not recommended its usage for beer, as pasteurization seemed to negatively affect the organoleptic properties of beer.(129) The subsequent alterations on its most relevant chemical compounds caused by heat treatment were shown to display significant effects in beer aroma and flavor.(129)

Further studies focused on the preservation and shelf-life extension of edible products led to a more accurate understanding of the laws of thermal inactivation of biological processes in microorganisms.(128) This allowed the development of improvements on the pasteurization technique for food product conservation, especially useful for discovering the more suitable target temperature for each, as well as a pasteurization dose, which led to the definition of the concept of pasteurization unit.(128) A pasteurization unit (PU) is defined as the decimal reduction of microorganism number that occurs in a product held at 60°C for a period of time of 60 seconds.(130) The total number of PU's for a particular pasteurization process for beer can be estimated from the following equation:

푃푈 = 푡 × 1.393 (푇−60) (1)

where T is the temperature in degrees Celsius and t is the time in minutes at which the beer is held at that temperature. Although in reality, the beverage does not instantaneously reach the desired temperature, nor does it cool down instantaneously.

A key concept crucial for the application of pasteurization process to extend the shelf-life of products is the thermal resistance of the different microorganisms.(128) As pasteurization consists of microbial death subject to heating, the rate at which this event occurs is an important measurement to define the process and it is given by a differential equation translated by a first order kinetic reaction.

푑푁 = −푘푁 (2) 푑휏

푁 represents the number of surviving microorganisms of a certain population at heating time 휏. The constant 푘 stands for a kinetic constant which can also be called inactivation constant. This is only applicable if considering a one of a kind population of microorganisms or a homogeneous mixture with identical physical properties and thermal resistance.(128)

There are a number of factors which influence the heat-induced kill rate of microorganisms. Parameters such as:

- species and physiological state of the microorganisms - pH of the surrounding environment

- ethanol concentration - carbon dioxide concentration - concentration of other compounds in beer such as metal cations, proteins, etc. - homogeneity of the suspension, presence of aggregates and degree of binding to solid surfaces

affect the time required to induce microorganism inactivation, which is majorly dependent on their initial concentration. In order to take into account the influence of all the above factors a study was developed, resulting in the creation of specific tables for each type of beer (and other beverages) with different pasteurization levels.(131)

Beer has been recognized for a long time as a safe beverage with a remarkable microbiological stability. The reason why, is that beer is an unfavorable medium for many microorganisms due to the presence of ethanol, hop bitter compounds, the high content of carbon dioxide, the low pH and the presence of only traces of nutritive substances such as glucose and amino acids.(132) However, in spite of these unfavorable features, a few microorganisms still manage to grow in beer and cause an increase of turbidity and unpleasant sensory changes. This phenomenon is called beer spoilage.(133) The reported beer spoilage microorganisms include mainly Gram-positive lactic acid bacteria belonging to the genera Lactobacillus and Pediococcus, although some Gram-negative species and wild yeasts can also be responsible for beer spoilage.(134) Hop compounds, namely 훼-acids, have demonstrated to protect beer from microbial infection, however, it was found that these compounds only inhibit growth of Gram-positive bacteria and not of Gram-negative species, despite lactic acid bacteria have reportedly developed resistance to hop-derived anti-bacterial mechanisms.(133)(135) Therefore, a more effective technique to prevent the growth of beer spoilage microorganisms must be applied, as it is the case of pasteurization.

Beer is pasteurized after fermentation in order to keep it microbiologically stable. Even though it is a commonly used method in most breweries, a number of them prefer not to pasteurize their beer. Those which do not pasteurize their beer, usually do not filter it as well, selling an unfiltered product.(136) Therefore, research on pasteurized unfiltered beer remains important, as the influence of pasteurization on the chemical storage stability is not very well described, and only few studies have been carried out on this subject. Additionally, it is still not clear if the effects of pasteurizing beer are mostly positive or negative, since rather contradictory results have been obtained.(136)

Some experiments pointed that pasteurization has a negative influence on the chemical storage stability, as decreasing levels in polyphenols and bitterness were observed, as well as an increase in staling aldehydes and a decrease in the concentration of volatile esters associated with beer freshness.(137) In contrast, others reached the conclusion that pasteurization of beer

improves its oxidative stability during storage compared with non‐pasteurized beer, based on electron spin resonance (ESR) spectroscopy assays.(138)

Heat enhances reactions between beer compounds and oxygen. Thus, when pasteurization temperature is reached, dissolved oxygen content can decrease to about 30% of the original concentration.(136) As previously acknowledged, oxidation process results in considerable damage to beer sensorial properties, so lengthy pasteurization is not recommended, in order to avoid undesirable effects due to the promotion of this kind of reactions. Another particular aspect of beer pasteurization are eventual changes in color. Usually it results in an increased intensity in the color of beer which is also related to the concentration of dissolved oxygen. It acquires a reddish tone, similar to what can be verified in rusty iron due to oxygen effect.(139) Despite all these changes, no significant alterations in pH have been observed during pasteurization.(136)

Because of the resulting changes in flavor stability in some cases, pasteurization has been pointed as a generator of beer staling. However, the discussion that still remains has to do with the spoilage of unpasteurized beer after long periods of storage and the even more prominent negative effects of dismissing pasteurization due to the lack of heat treatment necessary for inactivation of enzymes and other deteriorating microorganisms.(128) There are two types of pasteurization process used in beer production: tunnel pasteurization and flash pasteurization.

1.3.1 Tunnel Pasteurization

Tunnel pasteurization is usually employed for beer which is packed in cans or bottles. It is a longer process than flash pasteurization but the temperature the beer is exposed to is lower. During the process the beer passes through a tunnel which is broken up into a number of chambers that are at different temperatures, usually between 20°C and 60°C.(140)

In tunnel pasteurization bottles or cans are filled and closed, then funneled into the pasteurizer tunnel before any labelling is added. The tunnel has a low ceiling with spray heads at regular intervals. Temperature controlled water is sprayed down on to the packages. The bottles or cans slowly move through the pasteurizer on either a walking beam or conveyor belt. The tunnel is divided into many temperature zones to slowly heat the product up to pasteurization temperature, keep them at this specified holding temperature and then bring them back down to room temperature.(141)

Modern tunnel pasteurizers contain sophisticated control systems to manage the temperatures, deal with line hold-ups and slow-downs in a way to prevent over or under pasteurization of the product. Water is normally recirculated to improve energy efficiency.(140)

1.3.2 Flash Pasteurization

Regarding flash pasteurization, it is generally used when beer is to be filled into kegs and exposes beer to higher temperatures but for a much shorter period of time. This is possible because the higher the temperature the more rapidly microorganisms and enzymes are degraded. In this process the product is handled in a controlled, continuous flow and subjected to a heating temperature, normally in the range of 71.5°C to 74°C, for a period of time usually between 15 to 30 seconds. The flash pasteurization temperature and hold-time are dependent upon beer types and characteristics as well as the total number of viable microorganism cells.(141)

In terms of operation, once the keg is connected to the machine, the flow goes through the opening valves and the liquid reaches the pump. Afterwards it is pumped towards a heat exchanger, which comprises two separate steps of heating. At the first stage, the incoming beer exchanges heat with the out coming one, which is already leaving the second step of the pasteurization process. This, of course, can only happen once there is beer already flowing inside the equipment. At the very first operation, water is used to prepare the system as it is heated to remain inside the pasteurizer for further usage as a warm fluid that will exchange heat with the incoming cold beer. By running water inside the equipment, it is also aimed to calibrate the flash pasteurizer by setting the intended levels of flow-rate, temperature and pasteurization units. The second step of heat exchange occurs between pre-heated beer from the first stage and hot water kept inside the pasteurizer constantly heated by a steam stream. The already pasteurized beer will then leave the flash pasteurizer after exchanging heat with the incoming cold beer initiating a new cycle. It is also common to add a cooling step before reintroducing pasteurized beer into the keg, using a cold fluid for heat exchange.

There are some major differences between tunnel and flash pasteurization. Whereas some time-saving and economic advantages are brought by the use of flash pasteurizer, since there are lower energetic resource requirements, other aspects are not so beneficial. Since flash pasteurization is done prior to filling, it does not have any effect on organisms inadvertently introduced during filling.(140) Therefore, a well-controlled, aseptic filling operation to prevent reintroduction of microorganisms is essential. Other issues concerning flash pasteurization are related to flavor stability, as changes in flavor caused by the application of heat are more prone to occur in flash pasteurized beers than in beers which are pasteurized after bottling due to higher contact with oxygen when heat is applied.(142)

1.4 Thesis scope and goals

The present study was conducted in the Laboratory of Brewing and Malting of the Department of Biotechnology of University of Chemistry and Technology Prague, in association with Brevnov Monastery Brewery of St. Adalbert from March to July 2019. The main goal was to investigate the effect of pasteurization in unfiltered beer, regarding its chemical stability and sensory properties, after storage periods of 0, 2 and 4 months. For this purpose, samples of Czech typical light lager (12% ABV - alcohol by volume) and imperial lager (20 % ABV), as well as samples of Monastery IPA (Indian Pale Ale), were used, all produced at the brewery. This investigation work was split into four different branches. Samples of the two types of unfiltered lager beer were submitted to both tunnel and flash pasteurization and triangle and preference test trials were performed after each of the mentioned storage time intervals, in order to detect if there were significant sensorial changes caused by each type of pasteurization method and also to evaluate if these changes were more easily distinguishable and have more prominent effects in higher alcoholic beers. In parallel, another group of bottled 12% lager and 20% imperial lager samples were stored for periods of 0, 2 and 4 months at 21°C, while a number of each was previously tunnel pasteurized, whereas others were not. These samples were taken into the laboratory at the end of each storage period where determination and comparison of carbonyl compounds concentration was performed, to figure out the effect of pasteurization and storage time in these aging markers. Regarding IPA samples, the procedure included the addition of a solution containing essential hop oils to some bottles and similar to what was done with lagers, only a number of those bottles was tunnel pasteurized before being stored for 0, 2 and 4 months at 21°C. After each storage period, a deep sensory analysis was performed to evaluate the effects of pasteurization on beer essential hop oils and the resultant flavors and also to verify if sensory profile changes were more prominent in samples with higher concentration of essential oils. Each sensory analysis was followed by isolation and purification of essential oils extract from all IPA samples at the laboratory and further analytical methods were used to determine the content of essential oils compounds.

2. Materials and Methods

2.1 Relevant features of beer production process

Some relevant features of 12% light lager, 20% imperial lager and Monastery IPA production are worth to mention. For lager 12% lager, yeast strain RIBM 7 (Collection of Research Institute of Brewing and Malting, Prague) was used and for 20% imperial lager, strain RIBM 95. For hopping Monastery IPA, varieties of Saphir, Summit and Chinook were added during hop boiling and varieties of Cascade and Citra were added during whirlpool. Essential hop oils concentrated solution added to some IPA bottles was made by Barth® company and was prepared by steam distillation of hops variety Citra.

2.2 Pasteurization of beer and storage conditions

The pasteurization of beer took place at Brevnov Monastery Brewery of St. Adalbert on the 4th of March 2019. The flash pasteurizer was used to sterilize 3 kegs containing 15 L of 20% imperial lager and 3 kegs containing 15 L of 12% light lager. All these kegs were pasteurized at 40 PU’s. The maximum temperature at which beer remained in the stand was 71-72°C and the residence time was approximately 45 seconds. Pressure was regulated to 10 bar, as this is an important parameter regarding beer carbonation, according to the brew master. All these parameters, including flow-rate, were previously set-up and monitorized during the process. Tunnel pasteurization was applied to 4 sets of 12 bottles of 0.75 L. Each different set of bottles corresponded to 12% light lager, 20% imperial lager, Monastery IPA and Monastery IPA with addition of hop essential oils solution samples. All these bottles were pasteurized in a water bath at 38 PU’s and the maximum temperature reached inside the pasteurizer was 62°C. The entire process of pasteurization (heating, pasteurization, cooling) lasted about 50 minutes. In addition to the above-mentioned tunnel pasteurized beers, an equal number of bottles for each type of beer was also used for storage, in the unpasteurized version (for comparison). All bottled beers were stored in a room at 21°C for periods of 0, 2 and 4 months. After these storage periods, 2 bottles of each different set of pasteurized and unpasteurized beers were taken into the Laboratory of Brewing and Malting of the Department of Biotechnology of University of Chemistry and Technology Prague for further chemical analysis.

2.3 Sensory analysis methods:

2.3.1 Triangle test

The triangle test is a commonly used method in the field of brewing to detect significant differences among samples. A panel of assessors is asked to determine any differences or similarities between 3 samples of beer, whereas only 2 of them are equal while the other one is taken from a difference source or has a different production method. The assessors must not only try to identify the different sample, but also how significant was the sensorial difference felt in comparison with the other 2 samples. The samples are coded in a way that leaves no hints about the non-identical sample and the number of assessors must be enough to allow the obtainment of statistical relevance for the results.(143)

In case no difference is easily identified by the assessors, they are still forced to make a guess due to statistical significance purposes. However, this test is not only applicable for a single attribute as it can be applied to determine differences on several attributes at the same time. Having a large number of panelists for this type of test is strongly recommended in order to cover the whole spectrum of sensorial diversity and sensitivity and therefore to avoid the risk of having many false positive or false negative results, assuring the highest possible level of reliability.(143)

A usual appliance for triangle test happens for example, when a brewery develops a new process to improve the flavor balance of a beer or a process that may cause an impact on its sensorial stability and aims to find out whether the new beer is preferred over the currently produced beer before an expensive and large-scale consumer test is planned. It initially must be proven that the new beer is different from the current beer regarding its flavor, based on a statistical parameter.(143)

In this case, triangle test was used in order to find differences between tunnel pasteurized and flash pasteurized beer, using coded samples of 12% light lager and 20% imperial lager. The aim of this test was also to verify if the differences were more easily detected in higher alcoholic beers or not. A triangle test was performed after periods of 0, 2 and 4 months of pasteurized beer storage and the results were compared, in order to check if taste differences between tunnel pasteurized and flash pasteurized beer were more effortlessly detected after longer periods of storage. Any difference in sample appearance was avoided by using opaque tasting cups. The order and type of sample was random, that is, the taster could receive 2 flash pasteurized beer samples and 1 tunnel pasteurized beer sample or vice-versa. In a statistical level, significant differences were only considered when correctly noticed by more than 50% of the panel of assessors.

2.3.2 Preference test

When statically significant differences are noticed by the panel of assessors in triangle tests, they are usually followed by preference tests where assessors have the chance to try again both different samples correctly identified during the triangle test, this time coded with letters A and B, and choose which one they think it has the best palate. The test must be performed by all assessors even if they previously were not able to correctly identify the different sample on the triangle test. Preference test is done for marketing purposes, so producers can gain some information about which is the best product to launch in the market or if the newly developed products do not show any improvements in comparison to the ones previously developed. For this event, preference tests were done so the assessors could select their preference between tunnel and flash pasteurized lager beers.

2.3.3 Sensory test

Sensory analysis is used to obtain a systematic description of the flavor of samples under test. In a practical level, this method can be used to profile market competitors, by evaluating the sensory properties of the product they develop, before or after the development of a new product. In this test, the assessors either evaluate the intensity of a certain attribute within a defined scale or give an overall appreciation. Intensity scales must be chosen in order to produce the best discrimination among samples and the most reproducible results. Results are regarded as the

final average of all assessors rates and are used to form a sensory profile of the sample.(144) For the purpose of this work, sensory analysis was done to create the sensory profile of all Monastery IPA samples, with and without the addition of essential oils solution (unpasteurized samples and samples pasteurized before storage periods of 0, 2 and 4 months). All results were posteriorly compared, with the aim of verifying the differences concerning the sensorial profile of pasteurized and unpasteurized beer, the effect of time on the sensory properties of the ale and if the sensory differences were more prominent when increasing the concentration of essential oils.

2.4 Carbonyl compounds analysis

The headspace solid-phase micro-extraction method in combination with gas chromatography and mass spectrometry (HS-SPME GC-MS) was used to determine carbonyl compounds. This method comprises a few sample preparation steps described below.

2.4.1 Filtration

After storage periods of 0, 2 and 4 months, bottles of unfiltered pasteurized and non-pasteurized 12% light lager and 20% imperial lager were taken into University of Chemistry and Technology Prague facilities, where they were initially stored in a cold room in the dark at 15°C. The first step of sample processing for carbonyl compounds analysis was filtration of the beers samples. For this purpose, diatomaceous earth powder (LABICOM s.r.o, Czech Republic), also known as kieselguhr, was used to form a thick layer within filter papers to retain the yeast, filtrating the liquid through the filter funnel. Samples of 100 mL were collected in falcon tubes, to which parafilm was added to enclose the caps, and were stored in the freezer.

2.4.2 Derivatization method

All samples were taken out of the freezer at the same time, prior to chemical analysis. Derivatization method with PFBOA was used for the reasons presented in subchapter 1.2.1.1.1.

The method used was an optimized version of what is described in literature.(38) Derivatization solution was initially prepared by dissolving 0.150 g of PFBOA (≥ 99%, Sigma-Aldrich, Switzerland) in demineralized water in a 25 mL flask. The solution was stored in glass vials at 4°C. A solution of internal standard (IS) was achieved by measuring 30 휇L of 3-fluorobenzaldehyde (≥ 97%, Sigma-Aldrich, Switzerland) in a 50 mL flask with addition of ethanol (gradient grade for liquid chromatography, ≥ 99.9%, Merck, Germany) until the volume line. 4 mL of this solution were then pipetted into another 50 mL flask and ethanol was added again. This solution was stored in glass vials at -18°C. On the day of analysis, one vial was collected and stored at 4°C. A mixed solution of carbonyl compounds standards was prepared in a 200 mL flask half filled with ethanol, due to their high volatility. To this flask was pipetted:

• 400 휇L of 2-methylpropanal (99%, Sigma-Aldrich, Switzerland) • 400 휇L of 2-methylbutanal (97%, Fluka, Switzerland) • 400 휇L of 3-methylpropanal (97%, Sigma-Aldrich, Switzerland) • 200 휇L of heptanal (97%, Merck, Germany) • 200 휇L of octanal (98%, Merck, Germany) • 1200 휇L of furfural (98%, Alfa Aesar, Germany) • 300 휇L of benzaldehyde (98%, Alfa Aesar, Germany) • 300 휇L of 2,3-pentanedione (97%, Sigma-Aldrich, Switzerland) • 300 휇L of diacetyl (97%, Sigma-Aldrich, Switzerland) • 20 휇L of (E)-2-nonenal (97%, Sigma-Aldrich, Switzerland) and completed with ethanol until the 200 mL volume line. All these compounds were weighed in a scale after being respectively added to the solution, for further calibration calculations. Thereby, Table 4 indicates the weight measured for each carbonyl compound in the mixed standard solution.

Table 4- Volume, weight and degree of purity of every carbonyl compound added to the mixed standards solution.

Mixed Solution of Carbonyl Compounds Standards Compound Volume (µL) Weight (g) Degree of Purity (%) 2-Methylpropanal 400 0.3920 99 2-Methylbutanal 400 0.4612 97 3-Methylbutanal 400 0.4538 97 Heptanal 200 0.3245 97 Octanal 200 0.3048 98 Furfural 1200 1.7525 98 Benzaldehyde 300 0.4888 98 2,3-Pentandione 300 0.4474 97 Diacetyl 300 0.3661 97 (E)-2-Nonenal 20 0.0253 97

Three different volumes of this solution (0.02, 0.1 and 0.2 mL) were transferred to three different 100 mL flasks to which ethanol was added until the volume line. The purpose of this, was to achieve three different sets of derivatization reagent solutions with three different concentrations (low, medium and high) of carbonyl standards for a more accurate calibration process. Three replicates were prepared for each concentration. The thus prepared solutions with diluted carbonyl standards were stored at 4°C.

In addition to mixed carbonyl standard solutions, another sample was prepared with standard commercial beer for calibration purposes. For the preparation of this sample, a bottle of Gambrinus Originál 10° (4.3% ABV) was used. Approximately 150 mL of cold beer was degassed by shaking on a shaker for 5 minutes at a frequency of 175 min-1. To prepare the samples for monitoring free carbonyl compounds 2.5 g of NaCl (p.a., Lach-Ner, s.r.o., Czech Republic), 10 mL of shake beer (free from carbon dioxide) and 100 휇L of 3-fluorobenzaldehyde were introduced in a dark glass vial. Same approach was made for the remaining calibration solutions preparation, to which were added 100 휇L of the respective mixed standard solution in addition to the previously mentioned components. In total, 13 calibration solutions were prepared, as follows: 4 containing solely degassed commercial beer (Beer 1, Beer 2, Beer 3 and Beer 4), one of them to be used as a “blank” for GC-MS apparatus calibration (Beer 1), 3 to which were added a low concentration mixed carbonyl standards solution (Low 1 , Low 2 and Low 3), 3 to which were added a medium concentration mixed carbonyl standards solution (Medium 1, Medium 2 and Medium 3) and 3 to which were added a high concentration mixed carbonyl standards solution (High 1, High 2 and High 3). After closure, all calibration solutions contents were mixed on a Vortex mixer for about 1 minute.

Derivatization reagent solutions were prepared in dark glass vials with 3 g of NaCl, 10 mL of demineralized water and 20 µL of PFBOA solution. The contents of the vials were then mixed on a Vortex mixer for about 1 minute. The amount of derivatization reagent solutions prepared was equal to the number of prepared calibration solutions, so each can singularly undergo through carbonyl derivatization reaction process.

The addition of an internal standard solution is an important feature of this method. By standardizing all calibration solutions with a chemical substance added in the same concentration to all of them, a relation between the detected carbonyl compounds peaks (and respective concentrations) and the internal standard peak in the chromatogram can be established. This is done to correct for the loss of analyte during sample preparation or sample inlet, as the measurement procedure using HS-SPME may not be homogeneous in all samples. This ratio for the samples is then used to obtain the analyte concentrations from a calibration curve. The internal standard is a compound that must have similar but not identical properties to the chemical species of interest in the samples, as the effects of sample preparation should, relative to the amount of each species, be the same for the signal from the internal standard as for the signals from the species of interest in the ideal case. Logically, it also must not react with the chemical species of interest nor create any sort of interferences for the analysis.

2.4.3 HS-SPME GC-MS procedure and conditions

The vials were heated at 50°C for 5 minutes upon extraction. Sampling involves a fiber coated with a sorbent phase placed in the sample headspace. The analytes were extracted onto the coated SPME fiber and injected into an injection port of GC-MS analysis system. On Table 5 are listed the conditions for GS-MS analysis of carbonyls content.

Table 5- List of HS-SPME GC-MS parameters and conditions for analysis of carbonyl compounds.

Parameter Conditions Syringe type Fiber Incubation temperature 50°C Extraction time 3600 s GC apparatus Agilent HP 6890N GC column HP-5 MS 30 m × 0.25 mm × 0.25 μm Mode Constant flow Inert mass selective detector Agilent MS 5975 Carrier gas He 6.0 Injector settings Splitless mode Pressure 17.58 psi Inlet temperature 250°C Purge flow 10.0 mL/min Purge time 0.10 min Total flow 14.3 mL/min Initial temperature 40°C (0 min) 10°C/min to 140 ° C (0 min) Temperature program 7°C/min to 250°C (14 min) 20°C/min to 300°C (2 min) Run time 44.21 min

2.4.4 Evaluation of HS-SPME GC-MS results

Identification of carbonyl compounds was performed on the basis of obtained mass spectra using electronic impact ionization mode. The analytes were identified based on their mass spectra and on a characteristic mass of m/z 181. The peak at m/z 181 corresponds to the partition compound with the structure C6F5CH2+ of derivatization reaction products mentioned in section 1.2.1.1.1 (Fig. 4). This partition is common to all carbonyl derivatives. As mentioned in section 1.2.1.1.1 the derivatization reactions for each carbonyl compound provide two geometric isomers of oximes, so that each analyte is represented in the chromatogram by a pair of elution bands.

The results of the individual analysis were standardized using an internal standard. Quantification was performed using manual integration tool.

2.4.5 Other devices and equipment

In this section are listed the devices and equipment used in this analytical method that have not been descriptively mentioned before.

• 20 mL dark vial with silicone septum – Supelco, USA

• Milli-Q® Reference Water Purification System – Merck, Germany • Automatic dosing system COMBI PAL CTC Analytics – CTC PAL AG, Switzerland • SPME holder – Supelco, USA • SPME coated fiber assembly 50/30 μm DVB/CAR/PDMS, Stableflex – Supelco, USA • Agilent GC 6890N gas chromatograph – Agilent Technologies, USA • HP-5MS 30m x 0.25mm x 0.25µm column – Agilent Technologies, USA • Agilent 5975B Inert MSD Mass Detector – Agilent Technologies, USA

Figure 21- GC-MS apparatus used for carbonyls analysis in beer samples.

2.4.6 Calibration Lines

As previously mentioned, quantification of double band carbonyl peaks for the 13 calibration solutions was made using manual integration to obtain the area underneath those peaks. Internal standard method was used for sample normalization. The highest area of internal standard measured in all samples (which corresponds to Beer 3 sample) was selected and

defined as multiplicative factor for all carbonyl areas. This means that the area of each carbonyl compound in every calibration sample was multiplied by internal standard area of Beer 3 sample and divided by the respective internal standard area measured for each sample. This was the first step for calibration lines obtainment. Following this step, an average of each carbonyl area for every set of samples (depending on the concentration of mixed carbonyl standards solution added) was determined. Meaning that in the end of this step, the normalized areas measured for each carbonyl on Beer 2, Beer 3 and Beer 4 samples were combined into one, designated as Beer, which represents the average peak area for each compound in this set of samples. Same logic was applied for the remaining sets of samples, producing Low, Medium and High average combinations. Since the goal of adding mixed carbonyl standards solutions was to detect increasing concentration of carbonyls in beer samples, the resulting areas for each carbonyl in Beer set was individually subtracted to Low, Medium and High sets (Low-Beer, Medium-Beer and High-Beer). In sum, the concentration of each substance added in the mixed standard solutions was isolated from the typical concentrations already found in standard commercial beer.

In order to covert peak areas into concentrations, the following calculation process was applied: the weighted mass for each carbonyl (Table 4) was multiplied by the respective degree of purity so final pure mass could be obtained. This value was divided by the product between the initial volume of the flask to which carbonyl standards were pipetted (200 mL) and the volume of the diluted solutions (100 mL) and multiplied by the volume taken from the initial solution to achieve Low (0.02 mL), Medium (0.1 mL) and High (0.2 mL) calibration solutions. As a result of this operation, the concentration of every carbonyl compound on all mixed standard solutions was known. The final step consisted on establishing a correlation between these concentrations determined for Low, Medium and High solutions and the quantified peaks for Low-Beer, Medium-Beer and High-Beer. Linear regressions for each carbonyl compound were drawn, which also included the origin point for accuracy purposes. During this process, some results were looked down, as they were senseless and occurred probably due to sample preparation errors. In the end, a y-x equation translating peak area in function of concentration was obtained for each staling carbonyl studied.

These equations were used for determination of carbonyl content in 12% light lager and 20% imperial lager pasteurized and unpasteurized samples. The processing of these samples was similar to the preparation of calibration solutions (2.5 g of NaCl, 10 mL of filtered beer and 100 휇L of internal standard). Two replicates for each sample were prepared, achieving a total of 24 samples (12% Lager NP 1, 12% Lager NP 2, 12% Lager P 1, 12% Lager P 2, 20% IP NP 1, 20% IP NP 2, 20% IP P 1, 20% IP P 2 for each storage period – 0, 2 and 4 months). Same number of derivatization solutions was prepared for these measurements. The individual carbonyl content of these samples was also quantified using manual integration tool and peak areas were normalized with the previously selected internal standard area (belonging to Beer 3 sample). An average between the values obtained for the two replicates of each sample was also performed.

By substituting the resulting average peak areas on the y parameter of the respective calibration equations determined for each carbonyl compound, their concentrations can be obtained in every beer sample. The calibration lines graphics designed for each analyzed carbonyl and respective equations, can be found in Appendix I.

2.5 Essential hop oils analysis

Similar to 12% lager and 20% imperial lager bottles, unfiltered Monastery IPA pasteurized and non-pasteurized samples were shifted to the cold chamber within the facilities of University of Chemistry and Technology Prague after 0, 2 and 4 months of storage at the brewery. Each sample was filtrated before isolation and purification process of hop essential oils, prior to GC-MS analysis for identification and quantification. Internal standard method was also used for homogenization of the results.

2.5.1 Filtration

The filtration procedure for Monastery IPA’s was identical to what was used for non-pasteurized and pasteurized lagers. A teaspoonful amount of kieselguhr (LABICOM s.r.o, Czech Republic) was used inside filter papers to retain the yeast, filtrating the liquid through the filter funnel. However, in this case, the whole content inside the bottles was filtrated and collected inside plastic bottles and the filtrate volume was measured for each sample, unlike the 100 mL samples harvested for carbonyl analysis in lager beers. Plastic bottles were enclosed with parafilm and kept in the freezer at -4°C.

2.5.2 Isolation and purification of hop essential oils compounds

Isolation of hop essential oils compounds from beer samples comprised three different parts: steam distillation, solid-phase extraction and evaporation. All chemicals, apparatus and devices used for each part of the process are described below:

Chemicals:

- For steam distillation:

• 30% silicone antifoam in H2O – Sigma-Aldrich, Switzerland • Denatured ethanol – VWR Chemicals, France • Distilled water • Ice • Sodium chloride, p.a. – Penta s.r.o., Czech Republic

- For solid-phase extraction:

• n-Hexane > 97 % – VWR Chemicals, France • Ethanol absolute, gradient grade for liquid chromatography, ≥ 99.9% –Merck, Germany • Distilled water • Borneol (0.751 g/L) – Merck, Germany

- For evaporation:

• Anhydrous sodium sulfate, p.a. – Penta s.r.o., Czech Republic • n-Hexane > 97 % – VWR Chemicals, France • Ice • n-Hexane ≥ 99 %, PESTINORM® – VWR Chemicals, France

Apparatus and devices:

- For steam distillation:

• Steam distillation apparatus (Bunsen burner, heating mantle, thermos, Allihn condenser, steam generator, 4000 mL round-bottom boiling flask, 1000 mL beaker for cooling, three receivers of respective volumes 250 mL, 200 mL and 50 mL) (Fig. 22)

- For solid-phase extraction:

• Solid-phase extraction equipment (suction pump, water pump and glass box) – Agilent Technologies, USA (Fig. 23) • Bond elut SI-1 column (500 μg /3 mL) – Agilent Technologies, USA • Bond elut C18 column (500 μg /3 mL) – Agilent Technologies, USA • 500 mL flasks

- For evaporation:

• Teflon (PTFE) filters, 0.45 μm porosity • 100 mL round-bottom flasks • IKA Werke RV 06-ML Rotary evaporator – Artisan Technology Group, USA (Fig. 24)

Figure 22- Steam distillation apparatus.

Figure 23- Solid-phase extraction equipment.

Figure 24- Rotary evaporator.

The measured volume for each filtrated sample was added to the boiling flask together with a pipetted amount of silicone antifoam, where it was steam-distilled for approximately 1.5 hours. The aim of this separation technique was to separate hop essential oils compounds from water and ethanol in solution due to their higher volatility. The distillates were collected in three different receivers. In the first receiver, 5 mL of denatured ethanol was prior added, in the second 50 mL and in the third one 10 mL, to prevent volatiles evaporation. The second receiver was placed in a beaker filled with ice and the third one in a thermos filled with ice and sodium chloride. The distillates of a volume of approximately 350 mL for every sample were collected from all three receivers into 500 mL Erlenmeyer flasks and transferred into 600 mL plastic bottles. Secondly, the obtained distillates from the beer matrices were extracted using solid-phase extraction under vacuum (20 bar). A SI-1 polar column was used for pre-cleaning the distillates and a C18 single-use column was used to trap the essential oils in the sorbent, as essential oils compounds are long-chained non-polar molecules. The columns were prior conditioned, respectively with 5 mL of n-hexane and 5 mL of distilled water for SI-1 and 5 mL of ethanol absolute and 5 mL of distilled water for C18. The columns were connected together, with SI-1 on top, and then with the distillate bottles through rubber tubes. Eventually, this step lasted 3-4 hours. The next task was to isolate the entrapped essential oils fractions in the sorbent of the C18 column. This was achieved by eluting the columns with 6 mL of n-hexane under vacuum. In the end, 100 μL of borneol (0.751 g/L) were added as internal standard to all samples. The eluates were then

transferred to 100 mL Falcon tubes and dried overnight with anhydrous sodium sulfate. The next day, the eluates were filtered from the anhydrous salt in PTFE disc filters with n-hexane, followed by evaporation under vacuum in the rotary evaporator to remove solvent content. The remaining fraction was then transferred to a vial and added up to 1.5 mL with GC-grade n-hexane, ready for GC-MS analysis.

2.5.3 GC-MS conditions and results treatment

The conditions set for GC-MS apparatus for hop essential oils analysis and equipment description are presented on Table 6.

Table 6- List of GC-MS parameters and conditions for analysis of hop essential oils.

Parameter Conditions Agilent GC 6890N gas chromatograph GC apparatus – Agilent Technologies, USA HP-5MS 30m x 0.25mm x 0.25µm column GC column – Agilent Technologies, USA Agilent 5975B Inert MSD Mass Detector Inert mass selective detector – Agilent Technologies, USA Carrier gas He 6.0 Injection source CTC PAL ALS Mode Constant flow Initial flow 1.0 mL/min Injector settings Splitless mode Pressure 16.06 psi Inlet temperature 280°C Purge flow 5.0 mL/min Purge time 0.30 min Total flow 9.2 mL/min Initial temperature 40°C (7 min) 9°C/min to 120 ° C (3 min) 1.5°C/min to 135°C (2 min) Temperature program 2.5°C/min to 200°C (1 min) 20°C/min to 300°C (5 min) Run time 67.89 min

Individual components were identified in each sample according to the retention times and m/z ratio, using the internal standard calibration method. Afterwards, peaks of the components of interest in the chromatograms were manually integrated, followed by calculation of their amounts. Since 100 μL of borneol (0.751 g/L) were added to all essential oils fractions,

the correspondent mass of internal standard added was 75.1 μg. This value was multiplied by the area of each peak of interest in every sample and divided by the product between borneol peak area of the respective sample chromatogram and volume of filtrate collected. Table 7 indicates the collected volume after filtration of each beer sample.

Table 7- Volume of filtrate (L) collected for each beer sample.

Volume of Filtrate Collected (L)

1st Set 2nd Set 3rd Set

IPA Non-Pasteurized 0.64 0.72 0.71 IPA Pasteurized 0.68 0.72 0.70 IPA + Essential Oils Non-Pasteurized 0.66 0.70 0.70 IPA + Essential Oils Pasteurized 0.64 0.72 0.71

1st set- Beers stored for 0 months

2nd set- Beers stored for 2 months

3rd set- Beers stored for 4 months

As a result of this operation, the concentration (μg/L) of every relevant hop essential oils compound found in each beer sample was obtained.

3. Results and Discussion

In this chapter, results obtained from the experiments performed are presented and discussed. They will be divided into four parts: firstly, results obtained from the sensory analysis of pasteurized 12% light lager and 20% imperial lager are presented, followed by carbonyl compounds analysis results in non-pasteurized and pasteurized 12% light lager and 20% imperial lager samples, which precede results of pasteurized and non-pasteurized Monastery IPA samples sensory evaluations, and lastly, results from essential hop oils analysis from non-pasteurized and pasteurized Monastery IPA samples are presented.

3.1 Sensory analysis of tunnel or flash pasteurized 12% lagers and 20% imperial lagers:

3.1.1 Triangle tests

The results of the three sets of triangle tests performed using pasteurized lager beers are shown in Table 8. It is important to refer that the number of assessors was different for each set of tests and fairly scarce due to limited staff availability. In addition, not all assessors had the same level of training, although all of them were familiar with the test and its objectives.

Table 8- Triangle tests results regarding taste differences between tunnel and flash pasteurized lager beers (12% Lager and 20% Imperial Lager (IL)) and average difficulty level of the assessors population choice. Correct answers are referred to the number of assessors who were able to correctly identify the differently pasteurized sample for each type of lager and Pd is a statistical parameter which represents the proportion of correct assessments. The number of trained assessors is also discriminated for each case.

Triangle tests

1st Set 2nd Set 3rd Set

12% 12% 12% 20% IL 20% IL 20% IL Lager Lager Lager

Correct Answers 3 4 5 5 3 5

Wrong Answers 6 5 5 5 5 3

Trained Assessors 5 5 6 6 3 3

푷풅 33% 44% 50% 50% 38% 63%

Difficulty level of choice Difficult Difficult No easy No easy No easy No easy (Average)

1st set- Beers stored for 0 months

2nd set- Beers stored for 2 months

3rd set- Beers stored for 4 months

The analysis and interpretation of results is based on the value of 훼, or 훼-risk parameter which is part of the specific terminology to this method. The parameter represents the probability of concluding that a perceptible difference exists when, in reality, there is not one. The value must be chosen according to the level of reliability desired to achieve and considered as acceptable by the entity in charge of the test. It measures the sensitivity of the test along with 훽-risk (risk of concluding that no significant difference exists when, in reality, there is one) and 푃푑 (allowable proportion of discriminators in the population, i.e., those who can detect the difference between the products). In this case, the selected value for 훼-risk was 5% (0.05).(143) The number of correct assessments needed for statistical significance of the results was defined in accordance to the information provided on Table 18 of Appendix II. Since the number of assessments performed in all sets was inferior to the minimum number considered in this source, the results were extrapolated, in proportion with the available information. Thus, it was concluded that in order to achieve the desired significance level of 훼-risk, more than 50% of the panel of assessors must correctly identify the different sample (푃푑>50%) so it can be stated that differences between tunnel pasteurization and flash pasteurization are apparent in the tested sample.

According to this evaluation method, no apparent differences were found between tunnel and flash pasteurized samples for both lagers stored for few days between the pasteurization process and the tasting trial. The same can be declared about 12% light lager and 20% imperial lager bottles stored for 2 months after pasteurization, whereas the trial performed for the third set of beers exhibited comparatively different results (Table 8).

As expected, a short period of storage time was not enough to detect significant palate differences between tunnel and flash pasteurized samples in any type of lagers. A period of days is not sufficient to produce relevant aging reactions affecting beer flavor, regardless of the pasteurization technique used.(69) Therefore, the assessors displayed difficulties identifying the different sample present in each triangle test, since no distinguishing factors could be felt while tasting the samples. This resulted in a ratio of correct answers given below 50% for both 12% light lagers and 20% imperial lagers. Nonetheless, the number of assessors who were able to correctly identify the different pasteurized sample was slightly higher on 20% imperial lager trials (Table 8).

Regarding the second set of samples, an equal amount of correct and wrong guesses was obtained for both pasteurized lagers in their respective triangle tests, resulting in a 푃푑 value of 50% in both cases (Table 8). However, this value still was not enough to consider the significance of the results as the selected 훼-risk value implied that the majority of assessors were able to correctly identify the different sample. Yet, in comparison to the first set of trials, a higher

푃푑 value was obtained in both groups of samples, meaning that a larger proportion of assessors were able to notice taste differences and distinguish tunnel pasteurized lagers from flash pasteurized lagers after ones being stored for 2 months.

After 4 months of storage, triangle tests results continued to point to no significant taste differences between tunnel and flash pasteurized 12% light lagers, as a 푃푑 of 38% was obtained for this set samples. Independently of storage time, the majority of the panel of assessors was never able to correctly distinguish the effects of tunnel pasteurization from the sensory changes caused by flash pasteurization in this lager type. Although it can be the case of both pasteurization methods affecting the aging of this beer in the same way and thus producing similar results, the scarce number of assessments performed and the low level of experience of some assessors does not allow to reach doubtless conclusions.

However, the same did not happen with 20% imperial lager trials. For 20% imperial lagers stored for 4 months, 63% of the assessors population successfully discriminated those which went through tunnel pasteurization from those which underwent flash pasteurization prior to storage. These results matched the statistical significance levels required, evidencing apparent differences between tunnel and flash pasteurized imperial lagers concerning their flavor. Therefore, these samples were further submitted to preference test.

Comparing the results obtained for each set of beers, some aspects can be discussed. Storage time seems to have an impact on distinguishing more clearly the effect of each pasteurization method on the flavor of 20% imperial lagers, as 푃푑 value constantly increased from the first to the last set of samples. The prominence of pasteurization effects on aging reactions and resultant flavors was more noticeable on further aged beers. In contrast, a relation between storage time and effects of each kind of pasteurization on 12% light lager aging cannot be established, since the proportion of assessors who could successfully identify the different sample displayed an irregular variation for this type of tested lager. This means that contrary to 20% imperial lagers, storage time did not allow an easier distinction between tunnel and flash pasteurized 12% lagers. However, it must be pointed that 푃푑 value increased from first to the second set of 12% lager samples, while it decreased from the second to the third one, meaning that the latter set of results prevents this correlation with storage time. Moreover, when comparing both sorts of lagers, 푃푑 values obtained for 20% imperial lagers were superior to the ones obtained for 12% light lagers in every set of samples, except for the second set where an equal value for this parameter was achieved in both cases. So, it can be presumed that alterations on beer sensory profile caused by the two different pasteurization methods can be more easily distinguished in higher alcoholic beers, setting apart these pasteurization techniques regarding prominence of induced staling reactions through storage time.

Furthermore, the assessors were asked to select the degree of difficulty (between very easy, easy, no easy, difficult and very difficult) of their choice on each triangle test performed. The average level of difficulty selected by the assessors panel was difficult, no easy and no easy, for each pair of lager trials, respectively for the first, second and third sets of samples (Table 8). In accordance to the results obtained and to what was mentioned above, the assessors displayed more difficulties choosing the different sample on triangle tests performed with beers stored for shorter periods, namely the first set of samples which contained lagers stored for a period inferior to one month. However, although tasters made their selections with more ease within the second and third sets of lagers in comparison to the first one, they did not exhibit a clear distinction in regard to their choices on the second and third sets, assigning a similar difficulty level of choice for both cases.

Nonetheless, although some conclusions and remarks can be deducted from these triangle tests results, the fact that a limited population of assessors was used, in addition to the inclusion of a low number of specialized members amongst the population, must be accentuated. It also must be considered that the panel changed throughout the sets of trials, meaning that not all members of the initial panel were present on the second and third sets of triangle tests and were switched for different evaluators. So, the development of this experiment was not followed and accompanied by the same group of people, as tasters were constantly replaced by others with different levels of experience and backgrounds. In order to enable achievement of more supported conclusions, these tests should have been performed with a higher number of assessors, preferably with a majority of trained members for purpose of the test, and the panel

should be preserved as much as possible from the beginning to the end of the successive tasting trials, keeping a homogenous spectrum of sensorial abilities and acquaintance with the process. In fact, the recommended number of assessors for this type of trials should be between 18 to 36 assessors.(139) The available population was far below this interval in all trials, which compromises the accuracy of these results. Notwithstanding, their level of significance is somewhat relevant and partly match the expectances, as differences between the effects of both pasteurization methods applied on these sets of lagers were noticed.

3.1.2 Preference test

Since the results obtained on triangle tests performed for 20% imperial lager bottles stored for 4 months after pasteurization matched the significance requirements to consider the existence of an apparent difference between the outcome of application of flash pasteurization and tunnel pasteurization to these samples, a preference test was performed so tasters could reevaluate both differently pasteurized samples and choose the one they consider to have a better taste. The results of the carried out preference test are demonstrated on Table 9.

Table 9- Results of preference test carried out after the third set of triangle tests performed for 20% imperial lagers.

Assessor Answer 1 A 2 A 3 A 4 A 5 A 6 A 7 A 8 A

A- Tunnel pasteurized 20% imperial lager samples stored for 4 months B- Flash pasteurized 20% imperial lager samples stored for 4 months

After identifying perceptible differences between the two differently pasteurized 20% imperial lager samples, the panel of assessors had a chance to try again both samples, this time coded with the letters A and B, and unanimously showed preference for sample A, regarding its taste. Sample A corresponded to imperial lagers submitted to tunnel pasteurization. The choice was clear regardless of the training level of the assessors, since preference was shown by both specialized and unspecialized tasters, which favors the accuracy and reliability of preference test results. Furthermore, oxidized flavor as well as some astringent and acidic notes were remarked by the majority of the assessors while tasting sample B.

A plausible explanation for this evident preference for tunnel pasteurized imperial lagers over flash pasteurized ones, is related to higher oxygen exposure of beer during flash pasteurization, resulting in more prominent oxidation effects during storage, especially after longer periods.(142) As mentioned in Introduction chapter (see sections 1.3.1 and 1.3.2), changes in flavor stability are more prone to occur in flash pasteurized beers in comparison to tunnel pasteurized beers.(142) As heat enhances several aging reactions related to reactive oxygen mechanisms which severely affect beer flavor stability, heating beer up to high temperatures in an open system, as it happens during flash pasteurizer operation, may cause noticeable flavor damages. In comparison to tunnel pasteurization, where beers are pasteurized after bottle filling, the oxygen contact is much higher during flash pasteurization, since the oxygen content inside the bottles is much lower when compared to oxygen exposure in an open-air heating operation.

However, staling effects were only significantly distinguishable by the panel of assessors after 4 months of beer storage and only for 20% imperial lager samples (Table 8), which means that periods of 0 or 2 months of storage time were not enough to make a clear distinction between the staling degree of tunnel and flash pasteurized imperial lagers.

Although it must always be reminded that the risk of inaccurate triangle test results is high due to the low number of available assessors and lack of specialization level of some members, the collected data from these trials permits reaching some interesting conclusions. Firstly, few storage days are not sufficient to produce any sort of significant staling indicators (69) neither in tunnel or flash pasteurized lagers, regardless of their alcoholic levels, which makes it hard to separate tunnel pasteurized beers from flash pasteurized ones solely by performing tasting trials. Two months of storage generated a higher proportion of correct responses, probably due to a more advanced level of staling of flash pasteurized samples either in 12% lagers as in 20% imperial lagers. A significant portion of assessors was able to successfully distinguish flavor differences between flash and tunnel pasteurized 20% imperial lager samples stored for 4 months after pasteurization, assuming preference for tunnel pasteurized samples. Therefore, a longer aging process causes more substantial staling notes in flash pasteurized imperial lagers in comparison to tunnel pasteurized samples. The same could not be concluded for 12% light lagers.

As the purpose of these tests was to figure out differences between two different pasteurization methods and elect the more suitable one for potential commercialization, results obtained indicate that tunnel pasteurization should be applied for lagers stored for long periods (equal or superior to 3 or 4 months) instead of flash pasteurization, due to the greater negative impact of the latter on beer flavor. However, since no significant noticeable flavor differences were found between both pasteurization methods for lagers stored for periods equal or inferior to 2 months, it may be more suitable to use flash pasteurization for pasteurizing lagers instead of tunnel, considering that tunnel pasteurization represents a more expensive alternative due to higher spending of energetic resources, as it involves a much longer operation time. The chambers within the tunnel pasteurizer are filled with water which is heated up to a specific

designated temperature for each chamber. The energy required for singularly heating multiple volumes of water in different compartments and holding their temperature for a period of minutes, represents much higher utility expenses than those required for flash pasteurizer operation. Thus, in the case of both pasteurization techniques producing identical flavor impacts on lager beers stored up to 2 months, the most advantageous option from the ecological and economical point

of view should be selected, which is flash pasteurization (see section 1.3.2).(142) Ultimately, if trying to extent the shelf-life of lager beers, breweries should opt for tunnel pasteurization if beers are about to be stored for a period equal or superior to 3 or 4 months, which can be the case of exported products, which include local storage, transportation, customhouse (in some cases), distribution and shelf time before consumption, since tunnel pasteurization appears to be a less flavor damaging alternative. If lager bottles are just meant to be sold to local brewpubs, brewhouses, restaurants or any other selling points which beforehand do not require storage periods superior to 2 months, then flash pasteurization of lagers appears to be the most rentable option, assuring the most achievable quality of the product within this shelf-life period.

3.2 Carbonyl aging markers analysis in lagers

After identification and quantification of carbonyls peaks on the chromatograms generated by HS-SPME GC-MS system and subsequent calculation procedure explained in section 2.4.6, carbonyls concentration was individually determined for tunnel pasteurized and unpasteurized 12% light lager and 20% imperial lager samples stored for 0, 2 and 4 months. The results of these analysis can be observed on Table 10, 11 and 12.

Table 10- Carbonyl compounds concentration (mg/L), and respective characteristic retention time (min), determined for non-pasteurized (NP) and pasteurized (P) 12% Lager and 20% Imperial Lager (IL) samples stored for 0 months, thorough derivatization method followed by GC-MS analysis.

Concentration (mg/L)

1st set Retention Time Compound 12% Lager NP 12% Lager P 20% IL NP 20% IL P (min) 13.843 2-Methylpropanal 1.79E+00 1.80E+00 1.79E+00 4.91E+00 15.328 2-Methylbutanal 9.05E-01 8.56E-01 8.48E-01 1.57E+00 15.592 3-Methylbutanal 2.30E+00 2.12E+00 2.01E+00 4.10E+00 17.772 2,3-Pentanedione 3.12E+01 3.43E+01 2.28E+01 5.39E+01 18.659 Furfural 6.95E+01 5.43E+01 5.75E+01 4.84E+01 19.295 Heptanal 9.11E-02 6.45E-02 6.52E-02 1.34E-01 19.845 Diacetyl 3.35E+01 4.63E+01 4.47E+01 5.80E+01 20.836 Octanal 2.35E-01 1.33E-01 2.31E-01 3.95E-01 21.386 Benzaldehyde 1.94E+00 1.06E+00 4.06E+00 1.47E+00 22.324 (E)-2-Nonenal 5.29E+00 4.26E+00 7.40E+00 1.01E+01

Table 11- Carbonyl compounds concentration (mg/L), and respective characteristic retention time (min), determined for non-pasteurized (NP) and pasteurized (P) 12% Lager and 20% Imperial Lager (IL) samples stored for 2 months, thorough derivatization method followed by GC-MS analysis.

Concentration (mg/L)

2nd set Retention Time Compound 12% Lager NP 12% Lager P 20% IL NP 20% IL P (min) 13.843 2-Methylpropanal 1.93E+00 2.22E+00 8.93E+00 1.03E+01 15.328 2-Methylbutanal 8.22E-01 7.03E-01 2.48E+00 2.48E+00 15.592 3-Methylbutanal 1.97E+00 1.71E+00 6.32E+00 6.40E+00 17.772 2,3-Pentanedione 1.99E+01 2.60E+01 6.41E+01 9.64E+01 18.659 Furfural 3.74E+01 2.54E+01 6.44E+01 6.16E+01 19.295 Heptanal 4.30E-02 2.54E-02 2.21E-01 6.32E-01 19.845 Diacetyl 5.05E+01 4.80E+01 8.02E+01 6.16E+01 20.836 Octanal 1.54E-01 8.57E-02 5.54E-01 6.89E-01 21.386 Benzaldehyde 2.51E+00 2.96E+00 1.60E+00 1.97E+00 22.324 (E)-2-Nonenal 5.83E+00 4.17E+00 1.16E+01 1.22E+01

Table 12- Carbonyl compounds concentration (mg/L), and respective characteristic retention time (min), determined for non-pasteurized (NP) and pasteurized (P) 12% Lager and 20% Imperial Lager (IL) samples stored for 4 months, thorough derivatization method followed by GC-MS analysis.

Concentration (mg/L)

3rd set Retention Time Compound 12% Lager NP 12% Lager P 20% IL NP 20% IL P (min) 13.843 2-Methylpropanal 2.24E+00 3.97E+00 1.03E+01 6.47E+00 15.328 2-Methylbutanal 3.51E-01 1.29E+00 1.28E+00 1.42E+00 15.592 3-Methylbutanal 5.43E-01 4.19E+00 3.36E+00 3.83E+00 17.772 2,3-Pentanedione 2.25E+01 1.87E+02 6.63E+01 3.60E+01 18.659 Furfural 1.76E+01 3.85E+02 3.50E+01 7.18E+01 19.295 Heptanal 1.31E-01 9.39E-01 3.66E-01 3.81E-01 19.845 Diacetyl 1.90E+01 4.83E+02 2.93E+01 8.16E+01 20.836 Octanal 4.00E-02 8.30E-01 2.88E-01 2.62E-01 21.386 Benzaldehyde 1.69E+00 9.58E+00 1.82E+00 6.69E+00 22.324 (E)-2-Nonenal 1.99E+00 8.70E+00 5.11E+00 6.31E+00

In order to facilitate comparison of the results obtained for the two beers exposed to the conditions under study, the carbonyl concentration values were displayed in a column bar chart shown in Fig. 25.

Figure 25- Column bar chart containing information about the determined concentration (mg/L) for each carbonyl compound tested on every non-pasteurized (NP) and pasteurized (P) 12% lager and 20% imperial lager (IL) sample from every set of analyzed bottles (1st set – bottles stored for 0 months; 2nd set - bottles stored for 2 months; 3rd set - bottles stored for 4 months).

As detected carbonyls concentration was much higher for pasteurized 12% lager stored for 4 months than for any other sample, overshadowing the remaining results and complicating a thorough comparative analysis, the column bar chart was redesigned on a more suitable axis (Fig. 26).

Figure 26- Column bar chart containing information about the determined concentration (mg/L) for each carbonyl compound tested on every non-pasteurized (NP) and pasteurized (P) 12% lager and 20% imperial lager (IL) sample from every set of analyzed bottles (1st set - bottles stored for 0 months; 2nd set - bottles stored for 2 months; 3rd set - bottles stored for 4 months).

All these results obtained were compared to referenced flavor threshold values for the respective carbonyl species, indicated in the consulted literature (see Appendix III).(145) The only compound found below its flavor threshold interval was heptanal, in pasteurized 12% lager and non-pasteurized 20% imperial lager from the first set, and on both 12% lager samples from the second set of analyzed bottles. Which means that all tested carbonyls had an influence on the flavor profile of every lager bottle, apart from the referred exceptions. However, the gap between the flavor threshold levels and the concentrations measured is different for each carbonyl in every case, meaning that they have different contributions for beer staling.

Attending to the first set of samples, it can be noticed that, in general, carbonyls concentration is slightly higher on pasteurized 20 % imperial lager, in comparison to 12% lagers and to its non-pasteurized version (Table 10), suggesting that reactive oxygen mechanisms are activated during pasteurization, promoting oxidation reactions which lead to staling carbonyls formation in a short term, and that this effect is more prominent in higher alcoholic beers, since pasteurization did not produce the same effect in 12% lagers (Fig. 26). In fact, pasteurized and non-pasteurized 12% lager samples featured similar carbonyl profiles. However, furfural was detected in higher concentrations in non-pasteurized lagers, although the difference to pasteurized samples is not significant enough to look for further explanations, as furfural concentrations measured in this first set of samples remain within the same range in all four analyzed lagers.

The second set of samples suggests existence of a possible correlation between the degree of staling and alcoholic content, as imperial lagers display far superior concentrations of staling carbonyls compounds than 12% light lagers (Table 11). The only compound which was quantified in higher concentrations in 12% lagers was benzaldehyde. A reasonable explanation for this, is a higher availability of phenylalanine substrate, which promotes formation of benzaldehyde through Strecker degradation mechanism (see section 1.2.1.1.1.2), on 12% lagers, although the difference from the results obtained for imperial lagers is not very substantial. However, this is related to production method issues. After 2 months of storage, furfural concentration values were substantially higher in 20% imperial lagers than in 12% light lager samples (Fig. 26). These results match the previously mentioned premise that furfural is present in higher concentrations in high alcoholic beers (see section 1.2.1.1.1.5). Nonetheless, furfural concentrations did not vary as expected among pasteurized and non-pasteurized samples. Heat treatment supposedly enhances the formation of this Maillard reaction intermediate (67), although furfural was not detected in higher concentrations in pasteurized lagers during the first and second sets of analyzed samples. Moreover, vicinal diketones and (E)-2-nonenal concentration was also far superior in imperial lagers (Table 11). However, no correlation between (E)-2-nonenal formation and alcohol levels has been reported yet. In contrast, this correlation has been verified and reported for vicinal diketones in an experimental work that involved carbonyls determination in Czech lagers.(61) Which is expected, as being vicinal diketones yeast metabolism sub-products, it may be assumed that their formation is somehow related to fermentation yield and thus to alcohol content.(145) Regarding the effect of pasteurization process in these unfiltered lagers, only a slight improvement for 12% lager samples can be mentioned, as non-pasteurized version produced a carbonyl profile slightly more indicative of a higher staling degree (Fig. 26). The same was not verified for 20% imperial lagers, as no regular variation for each carbonyl concentration was observed.

Comparing the evolution of carbonyl profiles between identical samples from the first and second sets, the aging effect can be confirmed, particularly for 20% imperial lagers (Fig. 26). Although no regular variation could be observed for 12% lagers, since an increase on concentration was verified for some carbonyl compounds while for some others it decreased, the mean results show that 2 months of storage had a negative impact on these lager samples. Notwithstanding, aging effect was significantly more prominent on 20% imperial lager samples. The concentration increase was more constantly verified from first to second set of samples regarding staling carbonyls analyzed (Table 10 and 11).

The analysis of samples stored for 4 months did not allow establishment of expected or logically coherent conclusions. Firstly, it must be pointed that results obtained for pasteurized 12% light lager are unreasonable and out of proportion in contrast to the remaining tested samples, so they must not be considered or attributed to any significance (Fig. 26). This discrepancy occurred probably due to sample preparation mistakes, which can be either related

to addition of incorrect volume of internal standard (it cannot be related to derivatization samples preparation, otherwise it would equally affect the results obtained for every sample), or to issues connected with HS-SPME sampling system in this particular case. Furthermore, the results obtained for 12% light lager non-pasteurized and for 20% imperial lager non-pasteurized are rather inconsistent, since it appears to have occurred an aging process setback in both cases, as carbonyls concentration significantly diminished in practically all tested compounds in these samples from the second to the third set (Table 11 and 12). These results do not have any logical meaning, since aging process in lagers must have produced higher concentrations of staling carbonyls, contrary to what was determined. Identical incoherent results were obtained for pasteurized 20% imperial lager sample aged for 4 months, except for a noticeable increase on furfural and diacetyl concentrations. However, these are not enough to validate the results. Since unfiltered beers were tested, the possibility of sulfite adducts with carbonyls being formed (due to higher sulfite production resultant from yeast metabolism), disabling carbonyl detection through

HS-SPME GC-MS, could exist.(146) Nonetheless, these adducts have only reportedly been formed with (E)-2-nonenal and sulfite concentrations never reach such high concentrations in beer to cause such drastic reduction of carbonyl content.(146) In addition, this possibility fails to explain why pasteurized 12% light lager displayed such abnormal carbonyl profile. Hence, evidences point to issues occurred on the HS-SPME GC-MS operative system during this set of analysis, which led to obtainment of inaccurate results. Therefore, no further conclusions could be reach nor a valid comparative analysis involving the third set of samples could be performed.

Ultimately, one more issue related to the accuracy of these results can be signalized. This issue, has to do with a low accuracy level of the results obtained for 2,3-pentanedione concentration. The origin of this issue is related to calibration process, since the coefficient of determination value (R2) is far inferior for 2,3-pentanedione calibration line when compared to the obtained values for this coefficient for any other carbonyl calibration line drawn (see Appendix I). This compromises a significant accuracy level for the determined concentrations of 2,3-pentanedione in all samples and it occurred most probably due to quantification errors on 2,3-pentanedione peaks displayed on calibration samples.

3.3 Sensory tests performed for Monastery IPA samples

Sensory tests were performed aimed at designing the sensory profile of Monastery IPA samples produced at Brevnov Monastery Brewery of St. Adalbert. For this purpose, standard Monastery IPA bottles and Monastery IPA bottles to which were added a concentrated solution of essential hop oils were used. These bottles were pasteurized and stored for periods of 0, 2 and 4 months before trials (achieving three different sets of descriptive analysis trials) and the unpasteurized version of each sample was used for comparison. In this case, assessors were asked to evaluate eight different flavor attributes on an intensity scale from 1 to 5 (1 - very weak, 2 - weak, 3 - moderate, 4 - strong, 5 - very strong), as well as to assess the overall beer rating on a scale from 1 to 9 (1 - unusable, 2 - very bad, 3 - bad, 4 - under average, 5 - average, 6 - above average, 7 - good, 8 - very good, 9 - excellent) and also to give a qualitative remark about the character of bitterness. The 8 different appraised flavor attributes were the following:

- Overall intensity of flavor - Intensity of hop aroma - Intensity of citrus aroma - Intensity of spicy aroma - Intensity of fruity aroma - Intensity of resin aroma - Intensity of floral aroma - Intensity of bitterness

The results sheets obtained from the three different sets of descriptive analysis trials are represented on Table 20, 21 and 22 from Appendix IV.

Similar to what occurred with triangle tests, it was not possible to achieve an equal number of assessors for all sets of sensory tests. The first set of samples was evaluated by 12 assessors, of which 8 could be considered as trained assessors. Solely 8 assessors were present for the second and third sets of trials, while only 4 of them were specialized for this purpose in both situations. The procedure of these tests, as well as intensity scales chosen, followed the guidelines and recommendation provided by consulted literature.(144) No reference substance for comparison was used, as the goal was to perform a comparative analysis between the tested samples. While performing the evaluation of a given beer, assessors were provided as much samples as they required to make a confident analysis.

In order to facilitate the understanding of the results, spider web plots were designed (Fig. 27-33) with the average ratings scored for each attribute. These plots were designed for each beer of the different sets of samples, so the prominence or insignificance of a certain attribute could be verified on each of those samples, and in the end, the plots obtained were overlaid (using a common axis scale) amongst the respective sets, so differences regarding the sensory profile of the 4 samples belonging to each set (referred to samples stored for the same period) could be observed.

Figure 27- Spider web plots representing the sensory profile of Monastery IPA samples stored for 0 months.

Sensory Profile of IPA's

IPA + Essential Oils Non-Pasteurized IPA + Essential Oils Pasteurized IPA Non-Pasteurized IPA Pasteurized

Overall intensity of flavor 4.00 3.50 Intensity of bitterness 3.00 Intensity of hop aroma 2.50 2.00 1.50 1.00 0.50 Intensity of floral aroma 0.00 Intensity of citrus aroma

Intensity of resin aroma Intensity of spicy aroma

Intensity of fruity aroma

Figure 28- Spider web plot containing combined flavor profiles of Monastery IPA samples stored for 0 months.

Figure 29- Spider web plots representing the sensory profile of Monastery IPA samples stored for 2 months.

Sensory Profile of IPA's

IPA + Essential Oils Non-Pasteurized IPA + Essential Oils Pasteurized IPA Non-Pasteurized IPA Pasteurized

Overall intensity of flavor 4.00 3.50 Intensity of bitterness 3.00 Intensity of hop aroma 2.50 2.00 1.50 1.00 0.50 Intensity of floral aroma 0.00 Intensity of citrus aroma

Intensity of resin aroma Intensity of spicy aroma

Intensity of fruity aroma

Figure 30- Spider web plot containing combined flavor profiles of Monastery IPA samples stored for 2 months.

Figure 31 -Spider web plots representing the sensory profile of Monastery IPA samples stored for 4 months.

Sensory Profile of IPA's

IPA + Essential Oils Non-Pasteurized IPA + Essential Oils Pasteurized IPA Non-Pasteurized IPA Pasteurized

Overall intensity of flavor 4.00 3.50 Intensity of bitterness 3.00 Intensity of hop aroma 2.50 2.00 1.50 1.00 0.50 Intensity of floral aroma 0.00 Intensity of citrus aroma

Intensity of resin aroma Intensity of spicy aroma

Intensity of fruity aroma

Figure 32- Spider web plot containing combined flavor profiles of Monastery IPA samples stored for 4 months.

Figure 33- Spider web plots containing combined flavor profiles of each of the 4 different Monastery IPA samples stored for periods of 0, 2 and 4 months.

Some conclusions can be taken from direct analysis of the results presented on the tables displayed on Appendix IV and plots displayed above. Primarily, among the three different sets of beers, pasteurized beers scored better results on overall beer ratings in comparison to their non-pasteurized versions (Table 20-22). The assessors showed preference for pasteurized IPA’s regardless of the storage time they were subjected to. Non-pasteurized samples were only assessed either as under average or average products, while the overall assessment results mean always exhibited better ratings for the correspondent pasteurized versions on a range from average to very good products (Table 20-22). Notwithstanding the fact that the panel of assessors was not consistent and was submitted to several alterations throughout the three sets of trials, which is relevant because these ratings rely on the specific personal preferences of the tasters and on their sensitivity and thresholds by which they sense individual attributes, these results corroborate the theory that pasteurization has a positive effect on flavor stability of beers, as documented by Lund.(147) Particularly due to the preference for pasteurized beers occurring independently from the panel of evaluators.

Furthermore, the worst scores attributed to non-pasteurized IPA’s were given to samples stored for 4 months (Table 22), the longest storage period tested, and assessors made reference to their acidic taste and unpleasant aroma, including some remarks from specialized members who recognized both of these samples as oxidized spoiled products. In contrast, standard Monastery IPA and Monastery IPA + essential oils pasteurized bottles from the third set of samples were not as badly assessed and no remarks concerning oxidation-induced off-flavors were made. Hence, it can be assumed that pasteurization process produced a better effect on the oxidative stability of tested beers. Lund assigned the increased anti-oxidant capacity of pasteurized beer to enhancement of Maillard reaction products (namely melanoidins), sulfite formation and increased polyphenolic concentration, as effects of pasteurization treatment.(147) His experiments also reported a lower concentration of oxygen radicals in pasteurized beers in comparison to non-pasteurized ones, which was suggested to be caused by uptake of residual oxygen leading to accelerated radical reactions during the pasteurization followed by a subsequent reduction in the radical concentration after pasteurization.(147) In order words, reactive oxygen mechanisms are triggered during pasteurization but the remaining oxygen content inside the bottles after pasteurization is substantially lower, preventing a higher rate of oxidative processes during aging period, contributing for a higher chemical stability which is translated into a higher flavor stability from the period subsequent to pasteurization until consumption. Lund’s experiments results are coherent between them and support the thesis that pasteurization not only prevents microbial growth and spoilage, as it also boosts anti-oxidant power of beer, which can explain the preference for pasteurized IPA’s displayed by assessors during sensory tests.

Concerning non-pasteurized samples, 4 months seemed to be enough for assessors to remark spoilage and oxidized notes. Attending the fact that these samples were not submitted to pasteurization and thus could have been subjected to microbial growth, namely lactic acid

bacteria, it can be concluded that the shelf-life of non-pasteurized Monastery IPA samples is inferior to 4 months, as the quality of the product appeared to be compromised.

Overall, although the scores were not very dissimilar, it can also be pointed that a slight preference for samples to which were added a hop essential oils solution was shown, as the overall beer ratings were scarcely better in comparison to standard Monastery IPA’s amongst each set of trials (comparing non-pasteurized versions one to another and the same between pasteurized IPA’s). Hop essential oils solutions added flavor-active and aroma-active components or increased the concentration of the ones present in the standard ale, and despite the low significance margin of differences sensed, these samples were assessed as preferable (Table 20-22). However, no different effects on the character of bitterness were observed, as tasters assessed this parameter in a similar way, without regard to the addition of hop essential solutions, use of pasteurization or storage time.

Comparing the sensory profiles provided by the spider web plots developed for each tested beer, some aspects can be pointed. Overall intensity of flavor was more accentuated on pasteurized beers when contrasting the results obtained against non-pasteurized beers in all sets of tested samples (Fig. 27-32). Besides, the ratings scored for pasteurized ales were more consistent throughout time than those scored for non-pasteurized ales. These results also validate Lund’s postulates on the positive impact pasteurization has on beer flavor stability. Oxidative aging reactions are responsible for the formation and degradation of flavor-active compounds which alter the overall flavor intensity of a certain beer over time in an irregular way and, as mentioned above, are more prominent in the absence of pasteurization. Addition of essential oils had an influence in this attribute but only noticeable on the results obtained for the first two sets of trials (Fig. 27-32), since a correlation exists between higher scores for overall flavor intensity and addition of essential oils solution (between samples with the same level of pasteurization). The sensory tests performed with samples stored for 4 months did not allow the establishment of this correlation, as amongst non-pasteurized bottles, standard IPA scored a much higher rate on overall flavor intensity than its correspondent version to which were added essential oils (Fig. 32).

Intensity of hop aroma, which is directly connected to overall flavor intensity since hops are responsible for imparting the wide majority of aromas in beer and affect its organoleptic properties more than any other constituent, was characterized in a similar manner to the previously assessed attribute. More noticeable on pasteurized IPA’s in every trial and higher consistency on pasteurized samples test results. So ultimately, it can be stated that pasteurization slows down oxidation of hop-derived compounds, among other chemical species. However, in this case it seems to exist a more distinguishable impact on the addition of hop essential oils solution on this parameter. When compared the mean scores obtained for pasteurized and non-pasteurized standard Monastery IPA’s and their respective versions which include the addition of this solution, an increment regarding intensity of hop aroma can be verified in every

set of trials for Monastery IPA’s to which an essential oils solution was added (Fig. 27-33). Which is expected, considering that the purpose of adding this solution to IPA bottles was to increase hop essential oils components concentration to levels even more superior to their respective flavor thresholds, which is the same as increasing the intensity of hop aroma. Going back on overall intensity of flavor attribute analysis, it was mentioned that an increase on the attributed scores caused by addition of essential oils solution solely was not verified on the third set of samples when comparing pasteurized and non-pasteurized IPA’s amid them. Yet, the correlation between intensity of hop aroma and addition of essential oils was constant throughout every set of trials. Focusing on the results achieved on the third set of trials (Fig. 32), it can be concluded that the discrepancy demonstrated between standard non-pasteurized IPA and non-pasteurized IPA with addition of essential oils overall flavor intensity results is not due to an increase of intensity of hop aroma on standard non-pasteurized IPA. The highest intensity of flavor noticed by the assessors has a different source, is not related to a diminished concentration of hop essential oils, which would be contradictory, but probably occurs due to higher degree of staling caused by other compounds not related to hops addition, namely carbonyls. Thus, the higher intensity of flavor felt for this sample has to do with chemical changes that occur in malt-derived components resultant from a more advanced storage time, which generate easily detectable off-flavors. Even though both samples underwent the same storage time, sensory test results point to a higher degree of staling of standard non-pasteurized Monastery IPA stored for 4 months in comparison to its version with a higher concentration of hop oils.

Focusing on intensity of bitterness parameter, the analysis of the results leads to the following illations: the first set of tests produced expected results (Fig. 28). Bitterness is more well-preserved in pasteurized samples and essential oils solution addition produces a more recognizable bitterness taste among pairs of non-pasteurized and pasteurized samples. The second set of trials produced nearly identical conclusions (Fig. 30), as the only noticeable difference stands on the fact that similar scores were assessed for this parameter for pasteurized standard Monastery IPA and for its respective version to which essential oils were added, cancelling the influence of the addition of this concentrated solution. The third set of trials produced very distinct results (Fig. 32). Intensity of bitterness was far more felt on pasteurized Monastery IPA than on every other sample, non-pasteurized IPA and pasteurized IPA with addition of essential oils achieved identical scores and bitterness was least noticeable on non- pasteurized Monastery IPA with addition of hop oils bottles.

However, it does not mean that these results are inconsistent with intensity of hop aroma results, since it is known that hops are responsible for imparting bitterness on beer. The point is, hops are also used for adding other aromas to the sensory profile of beer, as it is the case of citrus, fruity, spicy, floral and resin aromas, which were analyzed in these trials. Degradation of hop compounds does not occur in a uniform way in every beer, hence the intensity level of these aromas is subject to non-equitable alterations. This results on emphasis and loss of certain flavors throughout the aging process. Furthermore, there is a case of high emphasis of a given attribute

resulting in an effect called “flavor masking”, which consists on diminishing the intensity felt for a certain flavor even if its concentration was not subject to any alterations (or subject to minor alterations). Nonetheless, aging process is not exclusive for hop-derived compounds, as carbonyls are also very prone to formation and degradation reactions which have a considerable impact on the flavor stability of a beer. Therefore, each beer has its own particular aging process with singular consequences both at a chemical and at an organoleptic level.

Contextualizing this information with the results displayed throughout the four series of spider web plots containing the combined results of every set for each IPA sample (Fig. 33), the subsequent deductions were achieved: the effects of aging process on pasteurized Monastery IPA bottles with addition of hop oils, resulted mainly in an increased intensity of practically all specific flavor attributes (floral, resin, fruity, spicy and citrus aromas), although apart from citrus aroma none of the remaining attributes kept their consistency after 2 more storage months, since a clear intensity decrease was observed during the third set of trials in comparison to the second one. However, intensity of bitterness, along with intensity of hop aroma and overall flavor intensity parameters kept their consistency throughout time on these IPA samples, which means that alterations on aroma-active hop-derived compounds did not affect bitterness. Standard Monastery IPA pasteurized samples kept a rather consistent sensory profile between 0 and 2 months of storage, apart from a considerable development of resin aroma intensity. The assessments performed after 4 months of storage indicated a slight diminish of resin aroma intensity and development of floral aromas in comparison to equivalent beers stored for 2 months, while bitterness was slightly more noticeable than in every other previous tasting trials. In this case, bitterness was also not loss or masked during aging process, which is probably due to an overall consistency of the remaining flavor attributes, which were not subject to severe alterations during storage. From Fig. 33, it can be easily comprehended that non-pasteurized Monastery IPA’s exhibit the most irregular sensory profiles. In both cases, the difference between the sensory profiles designed for the first and second sets of tests is notable. Despite a certain consistency level achieved for overall flavor intensity, intensity of hop aroma and intensity of bitterness on non-pasteurized standard Monastery IPA and non-pasteurized Monastery IPA with addition of essential oils, the differences regarding intensity of floral, resin, fruity, spicy and citrus notes are remarkable in both cases, as it can be seen from the verifiable distance between blue and orange lines in these spider web plots. The plots also displayed significant differences between the second and third sets of sensory tests performed for both non-pasteurized ales. However, these sensory changes produced a higher impact on overall flavor intensity and bitterness intensity assessments for standard Monastery IPA’s, as the essential oils concentrated version was able to demonstrate some consistency in these parameters. The remarkable differences between the first and second sets of assessments for both of these non-pasteurized IPA’s can be explained due to highly noticeable effects of aging process in these samples organoleptic properties. In contrast, the dissimilarities between the second and third sets of plots can be justified with spoilage effects noted on these ales during the third set of trials, resultant from the lack of heat treatment for preventing microbiological growth, which disabled a clear

identification of certain aromas and hindered the assessments. Notwithstanding, it must be stated the fact that no constant loss of positive flavor notes (namely fruity, floral and citrus aromas) throughout storage time was noticed neither in pasteurized or non-pasteurized IPA’s, which indicates that deterioration of compounds responsible for the different flavor notes occurred at a constant rate, although pasteurization enabled a better preservation of IPA’s sensory profile, since identification of flavor notes on trials performed for non-pasteurized ales after 4 months of storage was highly conditioned by their notable spoilage degree.

3.4 Essential hop oils analysis in Monastery IPA’s

Regarding essential hop oils analysis performed for Monastery IPA samples, the identification process was broader, as hop oils compounds were singularly identified in the obtained chromatograms. However, certain compounds were found in some IPA samples, whereas in others were not. Multiple peaks were eluted in every case, as not only hop compounds were present in the analyzed vials, but also solvents used for isolation and purification process and a fraction of non-vaporized carbonyls, which made identification of hop essential oils task more difficult. So, the first step of this procedure was seeking for a group of reference hop-derived compounds with strong odor-active properties commonly present in beers of the same style, according to a previous report on the topic.(148) Table 13 represents these referenced compounds, as well as their characteristic retention time on the GC column used (see section 2.5.3), and also their respective flavor description and reported flavor threshold levels.(149)

Table 13- The most active hop oils-derived odorants commonly found in beer and their respective retention time on the GC column used, flavor description and reported flavor thresholds.(148)(149)

Retention Time Flavor threshold Compound Flavor description (min) (흁g/L) 휷-Myrcene 17.550 Herbs, spicy 13 2-Methylbutyl-2- 17.808 Fruity 50 methylpropanoate D-Limonene 18.710 Citrus, floral 10 Linalool 20.784 Floral, citrus, rose-like 1 2-Nonanone 20.897 Varnish, stale, ketone 100-200 Camphor 22.765 Acrid, spicy N.D. 휷-Citronellol 24.961 Floral, citrus 4-5 2-Undecanone 28.961 Varnish, bitter, ketone 400 Methylgeranate 30.644 Fruity, waxy N.D. Geraniol 34.149 Floral, citrus, rose-like 5-6 2-Dodecanol 35.630 Sweet 500 휷-Caryophyllene 36.906 Floral, spicy 100-120 휶-Humulene 38.841 Woody, spicy 60-80 Nerolidol 44.380 Floral, citrus N.D. Humulol 46.721 Hay, grassy N.D. Caryophyllene oxide 46.891 Musty, spicy N.D. Humulene epoxide I 47.181 Hay, grassy 10 흉-Cadinol 48.586 Herbs, floral N.D. 휶-Cadinol 49.286 Herbs, floral N.D. Farnesol 52.252 Floral N.D.

N.D. – Not detected

After seeking for these compounds on IPA hop oils extract chromatograms, some other chemical species also considered to be aroma-relevant or that may indicate oxidation of essential oils components were searched and quantified posteriorly to detection. The results obtained for each IPA sample essential oils analysis are displayed below on Table 14-16.

Table 14- Results obtained for hop essential oils analysis of non-pasteurized (NP) and pasteurized (P) standard Monastery IPA and Monastery IPA + essential oils samples stored for 0 months.

1st Set

Concentration (흁g/L)

Retention time IPA + Essential IPA + Essential Compound IPA NP IPA P (min) Oils NP Oils P 11.937 3-Hexanone 2.10E+00 N.D. N.D. 3.87E+00 12.097 2-Hexanone 2.10E+00 N.D. N.D. N.D. 12.236 3-Hexanol 2.25E+00 N.D. 3.04E+00 N.D. 12.388 2-Hexanol 1.54E+00 N.D. N.D. N.D. 17.550 β-Myrcene 3.75E+00 1.33E+03 2.74E+00 N.D. 17.808 2-Methylbutyl-2- N.D. N.D. N.D. N.D. methylpropanoate 18.710 D-Limonene 1.12E+00 2.83E+02 1.54E+00 1.69E+00 20.784 Linalool 2.12E+00 N.D. 1.80E+00 1.21E+00 20.897 2-Nonanone N.D. N.D. N.D. N.D. 22.765 Camphor 2.57E+00 N.D. 3.27E+00 3.34E+00 24.754 Decanal N.D. N.D. N.D. N.D. 28.961 2-Undecanone 9.79E-01 N.D. N.D. N.D. 29.299 2-Undecanol 6.99E-01 N.D. N.D. N.D. 29.582 Cyanamide, dibutyl N.D. N.D. N.D. 7.19E+00 30.644 Methyl geranate 1.26E+00 N.D. N.D. N.D. 34.149 Geraniol 3.93E-01 N.D. N.D. N.D. 35.630 2-Dodecanol N.D. N.D. N.D. N.D. 36.906 β-Caryophyllene 2.16E+00 1.12E+02 1.62E+00 1.03E+00 38.841 α-Humulene 4.82E+00 2.19E+02 3.71E+00 2.09E+00 43.633 α-Calacorene N.D. N.D. 3.06E+00 N.D. 44.380 (E)-Nerolidol 2.77E+00 N.D. N.D. N.D. 46.721 Humulol 1.69E+00 N.D. N.D. N.D. 47.181 Humulene epoxide I 1.06E+01 2.44E+02 6.02E+00 5.19E+00 48.254 (Z,Z)-α-Farnesene N.D. N.D. N.D. N.D. 48.586 τ-Cadinol 3.14E+00 N.D. N.D. 1.24E+00 49.286 α-Cadinol 3.70E+00 N.D. 3.57E+00 N.D. 52.252 Farnesol 5.79E+00 N.D. 3.10E+00 N.D.

Sum 5.56E+01 2.19E+03 3.35E+01 2.68E+01

N.D. – Not detected

Table 15- Results obtained for hop essential oils analysis of non-pasteurized (NP) and pasteurized (P) standard Monastery IPA and Monastery IPA + essential oils samples stored for 2 months.

2nd Set

Concentration (흁g/L)

Retention time IPA + Essential IPA + Essential Compound IPA NP IPA P (min) Oils NP Oils P 11.937 3-Hexanone N.D. N.D. N.D. N.D. 12.097 2-Hexanone N.D. N.D. N.D. N.D. 12.236 3-Hexanol N.D. N.D. N.D. N.D. 12.388 2-Hexanol N.D. N.D. N.D. N.D. 17.550 β-Myrcene 1.52E+01 6.06E+01 1.69E+01 8.68E+01 2-Methylbutyl-2- 17.808 1.39E+01 N.D. N.D. N.D. methylpropanoate 18.710 D-Limonene 3.19E+01 1.06E+02 1.68E+01 1.81E+01 20.784 Linalool 1.77E+01 5.42E+01 1.56E+01 N.D. 20.897 2-Nonanone N.D. N.D. N.D. N.D. 22.765 Camphor N.D. N.D. 5.69E+00 N.D. 24.754 Decanal N.D. N.D. N.D. N.D. 28.961 2-Undecanone N.D. N.D. N.D. N.D. 29.299 2-Undecanol N.D. N.D. 1.16E+01 N.D. 29.582 Cyanamide, dibutyl 1.91E+02 7.04E+02 1.03E+02 N.D. 30.644 Methyl geranate 6.70E+00 N.D. 7.48E+00 N.D. 34.149 Geraniol N.D. N.D. N.D. N.D. 35.630 2-Dodecanol N.D. N.D. N.D. N.D. 36.906 β-Caryophyllene N.D. N.D. N.D. N.D. 38.841 α-Humulene 9.55E+00 6.44E+01 6.49E+00 N.D. 43.633 α-Calacorene N.D. N.D. N.D. N.D. 44.380 (E)-Nerolidol 1.39E+01 N.D. 1.43E+01 N.D. 46.721 Humulol 1.25E+01 N.D. 1.67E+01 N.D. 47.181 Humulene epoxide I 8.63E+01 2.58E+02 6.33E+01 1.14E+02 48.254 (Z,Z)-α-Farnesene 3.09E+01 N.D. N.D. N.D. 48.586 τ-Cadinol 2.58E+01 1.12E+02 1.77E+01 N.D. 49.286 α-Cadinol 2.75E+01 1.21E+02 2.03E+01 N.D. 52.252 Farnesol 4.54E+01 N.D. 2.97E+01 N.D. Sum 5.28E+02 1.48E+03 3.45E+02 2.19E+02

N.D. – Not detected

Table 16- Results obtained for hop essential oils analysis of non-pasteurized (NP) and pasteurized (P) standard Monastery IPA and Monastery IPA + essential oils samples stored for 4 months. 3rd Set

Concentration (흁g/L)

Retention time IPA + Essential IPA + Essential Compound IPA NP IPA P (min) Oils NP Oils P 11.937 3-Hexanone N.D. N.D. N.D. N.D. 12.097 2-Hexanone N.D. N.D. N.D. N.D. 12.236 3-Hexanol N.D. N.D. N.D. N.D. 12.388 2-Hexanol N.D. N.D. N.D. N.D. 17.550 β-Myrcene 1.71E+01 2.76E+01 2.78E+01 3.08E+01 2-Methylbutyl-2- 17.808 N.D. N.D. N.D. N.D. methylpropanoate 18.710 D-Limonene 2.52E+01 2.26E+01 3.07E+01 2.85E+01 20.784 Linalool 4.02E+01 4.94E+01 5.61E+01 2.71E+01 20.897 2-Nonanone N.D. 1.37E+01 N.D. N.D. 22.765 Camphor N.D. N.D. N.D. N.D. 24.754 Decanal N.D. N.D. N.D. N.D. 28.961 2-Undecanone N.D. N.D. N.D. 1.11E+01 29.299 2-Undecanol 1.22E+01 N.D. 4.11E+01 N.D. 29.582 Cyanamide, dibutyl N.D. N.D. N.D. N.D. 30.644 Methyl geranate 1.70E+01 1.76E+01 3.30E+01 2.52E+01 34.149 Geraniol N.D. N.D. N.D. N.D. 35.630 2-Dodecanol N.D. 6.37E+01 N.D. N.D. 36.906 β-Caryophyllene N.D. N.D. 2.11E+01 8.36E+00 38.841 α-Humulene 1.04E+01 1.33E+01 3.64E+01 4.27E+01 43.633 α-Calacorene N.D. N.D. N.D. N.D. 44.380 (E)-Nerolidol 4.29E+01 3.63E+01 3.10E+01 2.93E+01 46.721 Humulol N.D. 8.42E+01 1.28E+02 1.05E+02 47.181 Humulene epoxide I 1.68E+02 1.58E+02 2.13E+02 2.54E+02 48.254 (Z,Z)-α-Farnesene N.D. N.D. N.D. N.D. 48.586 τ-Cadinol 5.48E+01 5.86E+01 5.44E+01 6.38E+01 49.286 α-Cadinol 7.33E+01 8.98E+01 8.74E+01 8.56E+01 52.252 Farnesol 1.26E+02 1.06E+02 1.26E+02 7.47E+01 Sum 5.87E+02 7.41E+02 8.86E+02 7.86E+02

N.D. – Not detected

The IUPAC name of the compounds which are not designated in such way on the tables exhibited above can be found in Appendix V. Caryophyllene oxide and 훽-citronellol are not mentioned in any results sheet because they were not detected in any IPA sample chromatogram.

To begin with this analysis, the most active odorants (mentioned on Table 13) found and quantified in the tested IPA samples were compared to their respective flavor thresholds (for those to which flavor thresholds have been documented). β-caryophyllene was only found above its flavor threshold in pasteurized standard IPA stored for 0 months. In parallel, α-humulene was solely detected above its threshold in the same sample as β-caryophyllene was and also in pasteurized standard IPA stored for 2 months. Compounds such as 2-methylbutyl-2- methylpropanoate, 2-nonanone, geraniol and 2-dodecanol were only detected once amid all 12 analyzed samples and in concentrations far below the respective flavor thresholds in every case. The same occurred for 2-undecanone, although it was detected twice. In contrast, linalool, which was detected in most samples, was always quantified in levels above its threshold. β-myrcene and D-limonene were also quantified above the respective threshold except for both non-pasteurized IPA’s and pasteurized IPA with essential oils solution belonging to the first set of analyzed samples. Lastly, humulene epoxide I levels were also determined above its threshold in every sample with exception for essential oils concentrated samples from the first set. Therefore, overall, it can be concluded that β-myrcene, D-limonene, linalool and humulene epoxide I are the components that play a bigger role on the sensory profile perceived for these studied Monastery IPA samples.

Regarding the first set of analysis performed, pasteurized standard IPA produced abnormal results. It is not logical that the sum of essential oils components is far superior in this sample than in any other 11 tested samples, in addition to quantified concentrations of β-myrcene and D-limonene being unrealistically superior to their respective flavor thresholds, especially given that this sample was not concentrated in essential hop oils. The cause of this abnormality must be related to sampling issues for the GC-MS apparatus. Thus, these results were excluded from further analysis. Furthermore, contrary to expected, Monastery IPA samples concentrated in essential oils did not produced higher total concentrations of these chemical species. The exact same scenario was verified for the second set of analysis where the total sum of quantified hop-derived components was far superior for pasteurized standard Monastery IPA and no increase was verified for essential oils concentrated samples. However, the third set of results displayed lower variance amid the four analyzed samples regarding the total concentration of essential oils detected and also displayed higher total concentrations for non-pasteurized and pasteurized Monastery IPA’s with addition of essential oils in comparison to the standard ale versions.

Examining the variation of linalool throughout the three sets of samples, a somewhat constant increase in concentration with storage time can be observed in all four IPA samples,

however namely in non-pasteurized samples. This increasing concentration of linalool in beer over time is not surprising, there are some reported mechanisms that justify this occurrence, namely release of linalool from linalool glycosides or isomerization of nerol and geraniol to linalool, being both of these biotransformation processes induced by yeast metabolism.(150)(151) Regarding the first presented alternative, it is relevant to introduce the definition of glycoside. A glycoside is a molecule that contains a sugar moiety and an aglycone.(152) In beer, glycosides of aliphatic or monoterpene alcohols (such as linalool) are the most frequently ones formed, as well as glycosides of phenols. For the case of linalool, its correspondent glycoside is formed through a bounding to of β-D-glucose.(152) An increase in linalool concentration has reportedly been observed during fermentation and maturation of unfiltered beer, thanks to the activity of

β-glucosidase, the yeast enzyme which hydrolyses the linalool glycoside and releases linalool.(152) Thus, this mechanism supports the results obtained in this experimental work. Furthermore, studies performed with Saccharomyces cerevisiae strains, demonstrated the ability of these yeasts to isomerize geraniol and nerol to linalool.(151) These mechanisms, not only explain increasing levels of linalool upon fermentation and maturation of unfiltered beer, but also justify vanishing of nerol, geraniol and 훽-citronellol (since 훽-citronellol is a reaction product of geraniol reduction), as it was observed in analyzed IPA’s.

Yeast metabolism has a significant effect on hop oils compounds and subsequent analysis. Yeast glycosidic enzymes are responsible for the release of many monoterpene alcohols from their bounded glycosides. Moreover, the presence of 훽-citronellol resulting from geraniol reduction is exclusively attributed to yeast metabolism, although this compound was not found in any sample.(152) However, some biotransformations caused by yeast metabolism are disadvantageous and undesirable. Which is the case of isomerization of linalool to α-terpineol, as the latter has a lower odor threshold and produces an unpleasant aroma.(151) Notwithstanding, given that α-terpineol was not identified in any sample, this conversion has no relevance for results analysis. In addition, the isolated and analyzed hop oils fractions from Monastery IPA’s did not contain only the components of interest, but also many ester and higher alcohols produced by yeast, such as ethyl hexanoate, ethyl heptanoate, ethyl octanoate, 2-phenylethyl acetate, octanoic acid, n-decanoic acid, dodecanoic acid and phenylethyl alcohol. These compounds coeluted with some hop essential oils compounds in some cases, making the quantification of the compounds of interest difficult. This is one of the limitations of the method used for analysis. Moreover, the presence of these sub-products of yeast metabolism possibly lowers the transfer rates of hop oils during beer fermentation and maturation, affecting the flavor properties of the final product.(116)

Despite its lower solubility in beer, which is due to its non-polarity, β-myrcene was detected in rather equivalent amounts to linalool, which has a far superior rate of dissolving in beer due its polarity, in practically every IPA sample stored for 0 and 2 months. The difference between the concentrations of these components is slightly more remarkable in samples stored for 4 months, as linalool levels become higher due to the above-mentioned reasons. Furthermore,

β-myrcene can also undergo auto-oxidation, yielding compounds such as α-pinene and β-pinene, besides terpenoids such as linalool, nerol or geraniol.(125) However, in this case, such mechanisms do not seem to have occurred in any analyzed sample, since α-pinene, β-pinene and nerol were not detected at all and geraniol was only detected in a very small peak in one of them. Therefore, correlation between β-myrcene concentration decrease and linalool concentration increase cannot be established.

Following caryophyllene oxide and humulene epoxide I levels is a good way to evaluate beer oxidation degree, in particular oxidation of hop-derived components. As mentioned in Introduction (see section 1.2.3.2), these compounds are originated through oxidation of the respective sesquiterpenes. The fact that there were no traces of caryophyllene oxide detected on samples in which β-caryophyllene was, is a positive sign of oxidative stability of the beer. In contrast, humulene epoxide I was found in every sample. Which is not surprising since α-humulene oxidation can occur even at room temperature (see section 1.2.3.2). Thus, it is also not surprising that a constant increase in humulene epoxide I concentration in function of storage time was verified for every sample, except for pasteurized standard Monastery IPA, as a result of the previously referred incongruous results obtained for this sample. This increase was more accentuated on Monastery IPA samples to which were added essential oils, which is related to higher availability of oxidizable substrate. Addition of essential oils, at an overall level, did not produce a much significant difference amongst the results obtained for this variety of Monastery IPA samples, apart from a slight increase on β-myrcene and D-limonene levels and on methyl geranate concentration quantified on the third set of samples. As no effects of pasteurization on standard Monastery IPA’s can be discussed, the focus is driven towards the effects of pasteurization process on IPA’s to which were added an essential oils solution. Apparently, attending to humulene epoxide l quantified levels, it can be concluded that the pasteurized sample was more oxidized than its correspondent non-pasteurized version, indicating that pasteurization induces oxidative reactions in these ales.

Contextualizing essential hop oils analysis results with sensory tests results for these Monastery IPA samples, a few aspects can be commented. Although essential oils analysis suggest a higher oxidation level for pasteurized Monastery IPA with addition of essential oils concentrated solution in comparison to its non-pasteurized version, more noticeable in bottles stored for 2 and 4 months, it is relevant not to forget the influence of microbiological stability on beer organoleptic and sensorial properties, which is more responsible for emergence of unpleasant flavors when not achieved, as it was verified for non-pasteurized samples during tasting trials. Therefore, it must be remarked that beer oxidation and beer spoilage are not equivalent concepts. Proceeding with the analysis, by visualizing spider web plots designed on Fig. 33 (see section 3.3) and attending to Table 14-16 displayed results, an overall match between increased intensity of floral and citrus aroma perceived and D-limonene, linalool and nerolidol concentrations appears to exist for both Monastery IPA’s concentrated in essential oils and also for non-pasteurized standard Monastery IPA. This correlation was expected, given that these

compounds found in these samples essential oils analysis are in fact responsible for imparting flowery and citrus flavors. Perception of floral aroma also seems to be enhanced by 휏-cadinol, 훼-cadinol and farnesol concentrations. A correlation between intensity of spicy aroma described and 훽-caryophyllene and 훼-humulene levels can also be pointed.

Regarding factors that may affect the accuracy of these results, besides difficulties on quantification of some peaks of interest due to coelution with other compounds resultant from yeast metabolism, some other points can be mentioned. Analysis of unfiltered beer has some inherent conditions such as absorption of certain hop-derived compounds by yeast, namely

훼-humulene and its oxidation derivatives.(154) This can affect the accuracy of quantification of these compounds, considering that yeasts were removed in the first stage of the analysis process. Moreover, use of steam distillation technique for isolation of essential hop oils has some associated disadvantages, such as the risk of formation of oxidized, hydrolyzed, isomerized or polymerized artifacts of the volatiles of interest.(155) In addition, it requires an attentive control of the process as over-heating and disjunction of connected pieces can lead to leakages and loss of relevant content.

4. Conclusions and Future Prospects

Regarding tunnel or flash pasteurization of two types of Czech unfiltered lager beers (12% light lager and 20% imperial lager), results obtained from triangle and preference tests allow to suggest that tunnel method is better suited for preserving the organoleptic properties and flavor quality of lager beers in a long-term. The preference found for tunnel pasteurized beers over flash pasteurized ones is possibly due to higher oxygen contact and subjection to higher temperatures during flash pasteurization, which cause further damage on the sensory profile of beer. However, statistically significant results regarding flavor differences between tunnel and flash pasteurized lagers were only noticed on samples stored for 4 months. Thus, tunnel pasteurization is a more appropriate method for lagers to be exported or beforehand require longer storage periods before consumption, while flash pasteurization is more suitable for lagers about to be shifted from the brewery to a local market selling point or to any other commercialization route which merely implies a storage period equal or inferior to 2 months, since flash method has less associated expenses and apparently equally affects the flavor properties of lagers, in comparison to tunnel method, within this period. However, it would be relevant to collect samples and preform triangle trials within shorter storage periods, in order to have a more precise notion on what is the aging period required to detect significant flavor differences between flash and tunnel pasteurized lagers, since only periods of 2 and 4 months were tested. Furthermore, these trials should be attended by a higher number of assessors than those who attended the trials performed for this experimental work, so a higher accuracy level of the results can be achieved.

Analysis of carbonyl aging markers performed in unfiltered 12% light lagers and 20% imperial lagers did not allow to reach major conclusions about the effect of pasteurization on beer staling. Especially because the third set of analyzed bottles, which contained samples stored for 4 months, generated absurd results which were not taken into account. It is only mentionable that a higher concentration of staling carbonyls was detected in samples aged for 2 months, in comparison to their respective versions stored solely for a period of days prior to analysis, and that staling effect seemed to be more prominent in higher alcoholic beers and slightly more evident in pasteurized samples. Therefore, these analyses should be repeated before further conclusions can be reached. It would have also been positive to test these lagers after longer storage periods in order to investigate about the aging time from which carbonyls concentration stabilizes and aging reactions cease.

Pasteurization of unfiltered IPA samples produced a positive effect on flavor stability and consistency over time. In contrast, absence of pasteurization in tested IPA’s resulted in a more irregular sensory profile over time and spoilage notes perceived after 4 months of aging, meaning that the shelf-life of non-pasteurized samples is inferior to 4 months due to product quality issues. Furthermore, assessments performed suggested that pasteurization preserves better the overall quality of the product, along with some specific flavor attributes such as bitterness and intensity of hop aroma, in any stage of maturation process. Addition of essential hop oils did not produce significant differences, apart from a higher perception of certain flavor attributes. In order to complement these results, it would be interesting to perform a microbiological test to identify which microorganisms may be possibly responsible for IPA samples spoilage after 4 months of storage. Although it was assumed beer spoilage is most likely due to lactic acid bacteria growth, as acidic notes were remarked during sensory tests, it could also be relevant to identify and quantify all microorganisms present in spoiled beer, so a microbiological profile of these analyzed IPA samples can be developed and used for future comparative analysis.

Essential hop oils analysis of unfiltered IPA samples revealed that pasteurization enhances oxidation of hop-derived aroma-active compounds, although in not such a superior rate in comparison to non-pasteurized samples. Moreover, β-myrcene, D-limonene, linalool and humulene epoxide I were identified as the components most responsible for the aromas perceived in the sensory analysis of these samples, with particular relevance to linalool, to which increasing concentrations through time were assumed to be related with yeast metabolism. Notwithstanding, more studies aimed at determining the effect of pasteurization on hop-derived compounds in unfiltered beer should be carried out, since it was not possible to perform a comparative analysis due to lack of published reports on the topic.

5. References

1- Folkes, G. (2004). Pasteurization of beer by a continuous dense-phase CO2 system (Doctoral dissertation). University of Florida, Gainesville, USA. Retrieved from http://etd.fcla.edu/UF/UFE0006549/folkes_g.pdf 2- Toussaint-Samat, M., Alberny R., Horman I. (1991). 2 Million Years of the Food Industry.

1st ed. Nestle S. A.

3- Toussaint-Samat, M. (1992). A History of Food. 1st ed. Malden, MA, USA. Blackwell Publishers Ltd. 4- BeerHistory.com. (2016, September 07). King Gambrinus. Who was he?. Retrieved from http://www.beerhistory.com/library/holdings/gambrinus.shtml (Consulted in 11/09/2019) 5- Heartland Brewery. (2014, December 02). History of Beer. Retrieved from http://www.heartlandbrewery.com/history-of-beer/ (Consulted in 11/09/2019) 6- Beer100. (n.d.). The History of Beer. Retrieved from https://www.beer100.com/beerhistory/ (Consulted in 11/09/2019) 7- Pavsler, A., & Buiatti, S. (2009). Lager beer. In V. R. Preedy (Ed.), Beer in Health and Disease Prevention (pp. 31–45). Amsterdam: Elsevier Academic Press. 8- Losada, M. (2018). Effect of pasteurization on the sensory quality of craft beers (Master thesis). Universidade do Porto, Portugal. Retrieved from https://repositorio- aberto.up.pt/bitstream/10216/113876/2/277162.pdf 9- Vanderhaegen, B., Neven, H., Verachtert, H., & Derdelinckx, G. (2006). The chemistry of beer aging – A critical review. Food Chemistry, 95, 357–381.

10- Kilcast, D., and Subramaniam, P. (2011). Food and Beverage Stability and Shelf Life. 1st ed. Cambridge, UK. Woodhead Publishing 11- Dalgliesh, C. E. (1977). Flavour stability. Proceedings of the European Brewery Convention Congress, 623–659. 12- Meilgaard, M. (1972). Stale flavor carbonyls in brewing. Brewers Digest, 47, 48–57. 13- Bamforth, C. W. (1999). The science and understanding of the flavour stability of beer: a critical assessment. Brauwelt International, 98–110. 14- Whitear, A. L., Carr, B. L., Crabb, D., & Jacques, D. (1979). The challenge of flavour stability. Proceedings of the European Brewery Convention Congress, 13–25. 15- Whitear, A. L. (1981). Factors affecting beer stability. EBC-Flavour symposium. Monograph VII (pp. 203–210). 16- Clapperton, J. F. (1976). Ribes flavour in beer. Journal of the Institute of Brewing, 82, 175–176. 17- Bamforth, C. W. (2000). Beer quality: oxidation. Brewers’ Guardian, 129, 31–34. 18- Kaneda, H., Kobayashi, M., Takashio, M., Tamaki, T., & Shinotsuka, K. (1999). Beer staling mechanism. MBAA Technical Quarterly, 36, 41–47.

19- Kaneda, H., Kobayashi, N., Furusho, S., Sahara, H., & Koshino, S. (1995). Chemical evaluation of beer flavor stability. MBAA Technical Quarterly, 32, 76–80. 20- Meilgaard, M. (1975). Flavor chemistry of beer; Part I: flavor interaction between principal volatiles. MBAA Technical Quarterly, 12, 107–117. 21- Ahrenst-Larsen, B., & Levin Hansen, H. (1963). Gaschromatographische Untersuchungen über die Geschmaksstabilität von Bier. Brauwissenschaft, 16, 393–397. 22- Jamieson, A. M., Chen, E. C., & Van Gheluwe, J. E. A. (1969). A study of the cardboard flavour in beer by gas chromatography. Proceedings of the American Society of Brewing Chemists, 123–126.

23- Belitz, H. D., & Grosch, W. (1999). Food chemistry. 1st ed. Berlin Heidelberg: Springer Verlag. 24- Gijs, L., Chevance, F., Jerkovic, V., & Collin, S. (2002). How low pH can intensify beta-damascenone and dimethyl trisulfide production through beer aging. Journal of Agricultural and Food Chemistry, 50, 5612–5616. 25- Angelino, F., Kolkman, J. R., van Gemert, L. J., Vogels, J. T., van Lonkhuijsen, H. J., & Douma, A. C. (1999). Flavour stability of pilsener beer. Proceedings of the European Brewery Convention Congress, 103–112. 26- Saison, D., De Schutter, D.P., Delvaux, F.R.F. (2008). Optimisation of a complete method for the analysis of volatiles involved in the flavour stability of beer by solid-phase microextraction in combination with gas chromatography and mass spectrometry. Journal of Chromatography, 1190, 342–349. 27- Hashimoto, N. (1966). Rep. Res. Lab. Kirin Brewery Co., 9, 1. 28- Meilgaard, M., & Moya, E. (1970). A study of carbonyl compounds in beer – 1. Background and literature review. MBAA Technical Quarterly, 7, 135–142. 29- Engan, S. (1969). Some changes in beer flavour during aging. Journal of the Institute of Brewing, 75, 371–376. 30- Jamieson, A. M., & Van Gheluwe, J. E. A. (1970). Identification of a compound responsible for cardboard flavor in beer. Proceedings of the American Society of Brewing Chemists, 192–197. 31- Palamand, S. R., & Hardwick, W. A. (1969). Studies on the relative flavor importance of some beer constituents. MBAA Technical Quarterly, 6, 117–128. 32- Drost, B. W., Van Eerde, P., Hoekstra, S. F., & Strating, J. (1971). Fatty acids and staling of beer. Proceedings of the European Brewery Convention Congress, 451–458. 33- Wohleb, R., Jennings, W. G., & Lewis, M. J. (1972). Some observations on (E)-nonenal in beer as determined by a headspace sampling technique for gas–liquid chromatography. Proceedings of the American Society of Brewing Chemists, 1–3. 34- Wang, P. S., & Siebert, K. J. (1974). Determination of trans-2-nonenal in beer. MBAA Technical Quarterly, 11, 110–117.

35- Hashimoto, N., & Kuroiwa, Y. (1975). Proposed pathways for the formation of volatile aldehydes during storage of bottled beer. Proceedings of the American Society of Brewing Chemists, 33, 104–111. 36- Greenhoff, K., & Wheeler, R. E. (1981). Analysis of beer carbonyls at the part per billion level by combined liquid chromatography and high-pressure liquid chromatography. Journal of the Institute of Brewing, 86, 35–41. 37- Hashimoto, N., & Eshima, T. (1977). Composition and pathway of formation of stale aldehydes in bottled beer. Journal of the American Society of Brewing Chemists, 35, 145-150. 38- Ojala, M., Kotiaho, T., Siirila, J., & Sihvonen, M. (1994). Analysis of aldehydes and ketones from beer as O-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine derivates. Talanta, 41, 1297–1309. 39- Barker, R. L., Pipasts, P., & Gracey, D. E. F. (1989). Examination of beer carbonyls as their oximes by gas chromatography–mass spectrometry. Journal of the American Society of Brewing Chemists, 47, 9–14. 40- Yamada, H., & Somiya, I. (1989). Ozone Science Engineering. The Journal of the International Ozone Association 11, 127. 41- Glaze W. H., Koga M. & Cancilla D. (1989). Ozonation byproducts. 2. Improvement of an aqueous-phase derivatization method for the detection of formaldehyde and other carbonyl compounds formed by the ozonation of drinking water. Environ. Sci. Technol. 23, 838-847. 42- Choudhury, T.K., Kotiaho, T., Cooks, R.G. (1992). Detection of low molecular weight aldehydes in aqueous solution by membrane introduction mass spectrometry. Talanta; 39, 573. 43- Vesely, P., Lusk, L., Basarova, G., Seabrooks, J., & Ryder, D. (2003). Analysis of aldehydes in beer using solid-phase microextraction with on-fiber derivatization and gas chromatography/mass spectrometry. Journal of Agricultural and Food Chemistry, 51, 6941–6944. 44- Ochiai, N., Sasamoto, K., Daishima, S., Heiden, A. C., & Hoffmann, A. (2003). Determination of stale-flavor carbonyl compounds in beer by stir bar sorptive extraction with in situ derivatization and thermal desorption–gas chromatography–mass spectrometry. Journal of Chromatography, 986, 101–110. 45- Lermusieau, G., Noel, S., Liegeois, C., & Collin, S. (1999). Nonoxidative mechanism for development of trans-2-nonenal in beer. Journal of the American Society of Brewing Chemists, 57, 29–33. 46- Santos, J. R., Carneiro, J. R., Guido, L. F., Almeida, P. J., Rodrigues, J. A., & Barros, A. A. (2003). Determination of E-2-nonenal by high-performance liquid chromatography with UV detection assay for the evaluation of beer ageing. Journal of Chromatography, 985, 395–402.

47- Foster, R. T., Samp, E. J., & Patino, H. (2001). Multivariate modeling of sensory and chemical data to understand staling in light beer. Journal of the American Society of Brewing Chemists, 59, 201–210. 48- Narziss, L., Miedaner, H., & Lustig, S. (1999). The behaviour of volatile aromatic substances as beer ages. Monatsschrift Fur Brauwissenschaft, 52, 164–175. 49- Schieberle, P., & Komarek, D. (2002). Changes in key aroma compounds during natural beer aging. In K. R. Cadwallader & H. Weenen (Eds.), Freshness and Shelf Life of Foods (pp. 70–79). Washington DC, USA: ACS. 50- Hashimoto, N. (1981). Flavour stability of packaged beers. In J. R. A. Pollock (Ed.), Brewing Science (pp. 347–405). London: Academic Press. 51- Harayama, K., Hayase, F., & Kato, H. (1994). Evaluation by a multivariate-analysis of the stale flavor formed while storing beer. Bioscience Biotechnology and Biochemistry, 58, 1595–1598. 52- Miedaner, H., Narziss, L., & Eichhorn, P. (1991). Einige faktoren der geschmacksstabilität – sensorische und analytische bewertung. Proceedings of the European Brewery Convention Congress, 401–408. 53- Hofmann, T. & Schieberle, P. (2000). Formation of aroma-active Strecker-aldehydes by a direct oxidative degradation of Amadori compounds. Journal of Agricultural and Food Chemistry. 48, 4301– 4305. 54- Bohmann, J. J. (1985). Determination of the aging behavior of beer. 2. The behavior of some volatile compounds under heat-treatment. Monatsschrift Fur Brauwissenschaft, 38, 79–85. 55- Narziss, L., Miedaner, H., & Eichhorn, P. (1999). Studies on taste stability of beer. Monatsschrift Fur Brauwissenschaft, 52, 80–85. 56- Fontana, M., & Buiatti, S. (2009). Amino acids in beer. In V. R. Preedy (Ed.), Beer in Health and Disease Prevention (pp. 273–284), London: Academic Press. 57- Resconi, V.C., Escudero, A., Campo, M.M., (2013). The development of aromas in ruminant meat. Molecules 18, 6748-6781. 58- Strating, J., & Van Eerde, P. (1973). The staling of beer. Journal of the Institute of Brewing, 79, 414–415. 59- Chevance, F., Guyot-Declerck, C., Dupont, J., & Collin, S. (2002). Investigation of the beta-damascenone level in fresh and aged commercial beers. Journal of Agricultural and Food Chemistry, 50, 3818–3821. 60- Wheeler, R. E., Pragnell, M. J., & Pierce, J. S. (1971). The identification of factors affecting flavour stability in beer. Proceedings of the European Brewery Convention Congress, 423–436. 61- Craft Beer & Brewing. (2014, November 19). Vicinal Diketones. Retrieved from https://beerandbrewing.com/dictionary/OFH8CHBicP/ (Consulted in 26/09/2019) 62- Krogerus, K., & Gibson, B. R. (2013). 125th Anniversary Review: Diacetyl and its control during brewery fermentation. Journal of the Institute of Brewing, 119, 86–97.

63- Vanderhaegen, B., Neven, H., Coghe, S., Verstrepen, K. J., Verachtert, H., & Derdelinckx, G. (2003). Evolution of chemical and sensory properties during aging of top-fermented beer. Journal of Agricultural and Food Chemistry, 51, 6782–6790. 64- Peppard, T. L., & Halsey, S. A. (1982). The occurrence of 2 geometrical-isomers of 2,4,5- trimethyl-1,3-dioxolane in beer. Journal of the Institute of Brewing, 88, 309–312. 65- Bernstein, L., & Laufer, L. (1977). Further studies on furfural: the influence of raw materials, processing conditions and pasteurization temperatures. Journal of the American Society of Brewing Chemists, 35, 21–35. 66- Shimizu, C., Nakamura, Y., Miyai, K., Araki, S., Takashio, M., & Shinotsuka, K. (2001). Factors affecting 5-hydroxymethyl furfural formation and stale flavor formation in beer. Journal of the American Society of Brewing Chemists, 59, 51–58. 67- Madigan, D., Perez, A., & Clements, M. (1998). Furanic aldehyde analysis by HPLC as a method to determine heat-induced flavor damage to beer. Journal of the American Society of Brewing Chemists, 56, 146–151. 68- Vanderhaegen, B., Neven, H., Daenen, L., Verstrepen, K. J., Verachtert, H., & Derdelinckx, G. (2004). Furfuryl ethyl ether: Important aging flavor and a new marker for the storage conditions of beer. Journal of Agricultural and Food Chemistry, 52, 1661–1668. 69- Vanderhaegen, B., Delvaux, F., Daenen, L., Verachtert, H. and Delvaux, F. R. (2007). Aging characteristics of different beer types. Journal of Agricultural and Food Chemistry, 103(2), 404-412. 70- Martins, S. I. F. S., Jongen, W. M. F., & Van Boekel, M. A. J. S (2001). A review of Maillard reaction in food and implications to kinetic modelling. Trends in Food Science and Technology, 11, 364– 373. 71- Dunlop, A.P. (1948). Furfural formation and behaviour. Ind. Eng. Chem. 40 (2), 204-209. 72- Hodge, J. E. (1953). Chemistry of Browning Reactions in Model Systems. Journal of Agricultural and Food Chemistry, 1, 928–943. 73- Hashiba, H. (1982). The Browning Reaction of Amadori Compounds Derived from Various Sugars. Agricultural and Biological Chemistry 46, 547–548. 74- Maillard, L.C. (1912). Action Des Acides Amines Sur Les Sucres. Formation Des Melanoidins Par Voie Methodique. Compt. Rend. 154, 66–68. 75- Grigsby, J. H., Palamand, S. R., Davis, D. P., & Hardwick, W. A. (1972). Studies on the reactions involved in the oxidation of beer. Proceedings of the American Society of Brewing Chemists, 87-92. 76- Lustig, S., Miedaner, H., Narziss, L. (1993). Untersuchungen flu chtiger̈ aromastoffe bei der bieralterung mittels multidimensionaler gaschromatographie. Proceedings of the European Brewery Convention Congress, 445–452. 77- Neven, H., Delvaux, F., & Derdelinckx, G. (1997). Flavor evolution of top fermented beers. MBAA Technical Quarterly, 34, 115–118.

78- Alchetron. (n.d). Isoamyl Acetate. Retrieved from https://alchetron.com/Isoamyl-acetate (Consulted in 23/09/2019) 79- Williams, R. S., & Wagner, H. P. (1978). The isolation and identification of new staling related compounds form beer. Journal of the American Society of Brewing Chemists, 36, 27–31. 80- Gijs, L., & Collin, S. (2002). Effect of the reducing power of a beer on dimethyltrisulfide production during aging. Journal of the American Society of Brewing Chemists, 60, 68–70. 81- Eichhorn, P., Komori, T., Miedaner, H., & Narziss, L. (1989). Alterungscarbonyle im sub-ppb bereich. Proceedings of the European Brewery Convention Congress, 717–724. 82- Gijs, L., Perpete, P., Timmermans, A., & Collin, S. (2000). 3- Methylthiopropionaldehyde as precursor of dimethyl trisulfide in aged beers. Journal of Agricultural and Food Chemistry, 48, 6196–6199. 83- Bamforth, C. W., & Parsons, R. (1985). New procedures to improve the flavor stability of beer. Journal of the American Society of Brewing Chemists, 43, 197–202. 84- Kaneda, H., Kobayashi, M., Takashio, M., Tamaki, T., & Shinotsuka, K. (1999). Beer staling mechanism. MBAA Technical Quarterly, 36, 41–47. 85- Uchida, M., & Ono, M. (1996). Improvement for oxidative flavor stability of beer – role of OH-radical in beer oxidation. Journal of the American Society of Brewing Chemists, 54, 198–204. 86- Andersen, M. L., & Skibsted, L. H. (1998). Electron spin resonance spin trapping identification of radicals formed during aerobic forced aging of beer. Journal of Agricultural and Food Chemistry, 46, 1272–1275. 87- Irwin, A. J., Barker, R. L., & Pipasts, P. (1991). The role of copper, oxygen, and polyphenols in beer flavor instability. Journal of the American Society of Brewing Chemists, 49, 140–148. 88- Dale, R. S., & Pollock, J. R. A. (1977). A means to reduce formation of precursors of 2-trans-nonenal in brewing. Journal of the Institute of Brewing, 83, 88–91. 89- Tressl, R., Bahri, D., & Silwar, R. (1979). Bildung von aldehyden durch lipidoxidation und deren bedeutung als off-flavor-komponenten in bier. Proceedings of the European Brewery Convention Congress, 27–41. 90- Baxter, E. D. (1984). Recognition of 2 lipases from barley and green malt. Journal of the Institute of Brewing, 90, 277–281. 91- Schwarz, P., Stanley, P., & Solberg, S. (2002). Activity of lipase during mashing. Journal of the American Society of Brewing Chemists, 60, 107–109. 92- Stenroos, L., Wang, P., Siebert, K., & Meilgaard, M. (1976). Origin and formation of 2-nonenal in heated beer. MBAA Technical Quarterly, 13, 227–232. 93- Graveland, A., Pesman, L., & Van Eerde, P. (1972). Enzymatic oxidation of linoleic acid in barley suspensions. MBAA Technical Quarterly, 9, 98–104.

94- Noel, S., & Collin, S. (1995). Trans-2-Nonenal degradation products during mashing. Proceedings of the European Brewery Convention Congress, 483–490. 95- Noel, S., Metais, N., Bonte, S., Bodart, E., Peladan, F., Dupire, S., & Collin, S. (1999). The use of Oxygen 18 in appraising the impact of oxidation process during beer storage. Journal of the Institute of Brewing, 105, 269–274. 96- Stephenson, W. H., Biawa, J. P., Miracle, R. E., & Bamforth, C. W. (2003). Laboratory-scale studies of the impact of oxygen on mashing. Journal of the Institute of Brewing, 109, 273–283.

97- Belitz, H. D., & Grosch, W. (1999). Food chemistry. 1st ed. Berlin Heidelberg: Springer Verlag. 98- Ohloff, G. (1978). Recent developments in the field of naturally- occurring aroma components. Fortschritte der Chemie Organischer Naturstoffe, 35, 431–527. 99- Wijewickreme, A. N., & Kitts, D. D. (1998a). Metal chelating and antioxidant activity of model Maillard reaction products. In F. Shahidi, C. T. Ho, & N. van Chuyen (Eds.), Process induced chemical changes in food (pp. 245–254). New York, USA. Plenum 100- Buggey, L. A. (2001). A review of polyphenolic antioxidants in hops, brewing and beer. The brewer international, 4, 21–25. 101- Andersen, M. L., Outtrup, H., & Skibsted, L. H. (2000). Potential antioxidants in beer assessed by ESR spin trapping. Journal of Agricultural and Food Chemistry, 48, 3106–3111. 102- Wedzicha, B. L., & Kedward, C. (1995). Kinetics of the oligosaccharide-glycine-sulfite reaction - relationship to the browning of oligosaccharide mixtures. Food Chemistry, 54, 397–402. 103- Sanchez, B., Reverol, L., Galindo-Castro, I., Bravo, A., Ramirez, J. L., & Rangel-Aldao, R. (2001). Brewers yeast oxidoreductase with activity on Maillard reaction intermediates of beer. Proceedings of the European Brewery Convention Congress, 456–464.

104- Roberts, T. R., Wilson, R. J. H. (2006). Hops, in Handbook of Brewing, 2nd Edition. (B. Raton Ed.) (pp. 232-234) Florida, USA. Taylor & Francis. 105- Almaguer, C., Schönberger, C., Gastl, M., Arendt, E. K., Becker, T. (2014). Humulus lupulus – a story that begs to be told. A review. Journal of the Institute of Brewing, 120(4), 289-314. 106- Simpson, W. J. (1993). Studies on the sensitivty of lactic acid bacteria to hop bitter acids. Journal of the Institute of Brewing, 99(5), 405-411. 107- Aberl, A., Coelhan, M. (2012). Determination of Volatile Compounds in Different Hop Varieties by Headspace-Trap GC/MS - In Comparison with Conventional Hop Essential Oil Analysis. Journal of Agricultural and Food Chemistry, 60(11), 2785-2792. 108- Palamand, S. R., Aldenhoff, J. M. (1973). Bitter tasting compounds of beer. Chemistry and taste properties of some hop resin compounds. Journal of Agricultural and Food Chemistry, 21(4), 535-543.

109- Cook, A. H., Harris, G. (1950). The chemistry of hop constituents. Part I. Humulinone, a new constituent of hops. Journal of the Chemical Society (1), 1873- 1876. 110- Hashimoto, N., & Eshima, T. (1979). Oxidative degradation of isohumulones in relation to flavour stability of beer. Journal of the Institute of Brewing, 85, 136–140. 111- Kaneda, H., Kano, Y., Osawa, T., Kawakishi, S., & Kamada, K. (1989). The role of free radicals in beer oxidation. Journal of the American Society of Brewing Chemists, 47, 49–53. 112- De Cooman, L., Aerts, G., Overmeire, H., & De Keukeleire, D. (2000). Alterations of the profiles of iso-alpha-acids during beer ageing, marked instability of trans-iso-alpha-acids and implications for beer bitterness consistency in relation to tetrahydroiso-alpha-acids. Journal of the Institute of Brewing, 106, 169–178. 113- Araki, S., Tsuchiya, Y., Takashio, M., Tamaki, T., Shinotsuka, K. (1998). Identification of Hop Cultivars by DNA Marker Analysis. Journal of the American Society of Brewing Chemists, 56(3), 93-98. 114- Algazzali, V., Shellhammer, T. (2016). Bitterness Intensity of Oxidized Hop Acids: Humulinones and Hulupones. Journal of the American Society of Brewing Chemists, 74(1), 36-43. 115- Huvaere, K., Andersen, M. L., Olsen, K., Skibsted, L. H., Heyerick, A., & De Keukeleire, D. (2003). Radicaloid-type oxidative decomposition of beer bittering agents revealed. Chemistry-a European Journal, 9, 4693–4699. 116- Briggs, D. E., Boulton, C. A., Brookes, P. A., Stevens, R. (2004). Brewing: science and

practice: in Handbook of Brewing, 2nd Edition. (B. Raton Ed.) (pp. 882). Florida, USA. Taylor & Francis. 117- Sharpe, F. R., Laws, D. R. J. (1981). The essential oil of hops: a review. Journal of the Institute of Brewing, 87(2), 96-107. 118- Buttery, R. G., Mc Fadeen, W. H., Lundin, R. E., Kealy, M. P. (1964). Volatile hop constituents: conventional and capillary gas chromatographic separation with characterization by physical methods. Journal of the Institute of Brewing, 70(5), 396-401. 119- Sell, C. S. (2003). A Fragrant Introduction to Terpenoid Chemistry. Royal Society of Chemistry, UK, 229-268. 120- Naya, Y., Kotake, M. (1972). The Constituents of Hops (Humulus lupulus L.). VII. The Rapid Analysis of Volatile Components. Bulletin of the Chemical Society of Japan, 45(9), 2887-2891. 121- Jahnsen, V. J. (1963). Composition of hop oil. Journal of the Institute of Brewing, 69(6), 460-466. 122- Roberts, J. B. (1963). Hop oil II. Major constituents of the non-saponifiable fraction. Journal of the Institute of Brewing, 69(4), 343-346. 123- Eyres, G., Dufour, J. P. (2009). Hop Essential Oil: Analysis, Chemical Composition and Odor Characteristics. In V. R. Preedy (Ed.), Beer in Health and Disease Prevention (pp. 239–254), London: Academic Press.

124- Buttery, R. G., Ling, L. C. (1966). The chemical composition of the volatile oil of hops. Brew. Dig, 41, 71-77. 125- Dieckmann, R. H., Palamand, S. R. (1974). Autoxidation of some constituents of hops. I. Monoterpene hydrocarbon, myrcene. Journal of Agricultural and Food Chemistry, 22(3), 498-503. 126- Pickett, J. A., Sharpe, F. R., Peppard, T. L. (1977). Aerial oxidation of humulene. Chemistry & Industry(1), 30-31. 127- Takoi, K., Itoga, Y., Koie, K., Kosugi, T., Shimase, M., Katayama, Y., Nakayama, Y., Watari, J. (2010). The Contribution of Geraniol Metabolism to the Citrus Flavour of Beer: Synergy of Geraniol and β‐Citronellol Under Coexistence with Excess Linalool. Journal of the Institute of Brewing, 116(3), 251-260. 128- Basařová, G., Šavel, J., Basař, P., Basařová, P., Lejsek, T. (2017). The Comprehensive Guide to Brewing. From raw material to packaging. Nürnberg, Germany. Fachverlag. 129- Anderson, R. G. (1995). Louis Pasteur (1822-1895): An assessment of his impact on the brewing industry. Proceedings of the European Brewery Convention Congress, 13-23. 130- Del Vecchio, H. W., Dayharsch, A., & Baselt, F. C. (1951). Thermal death time studies on beer spoilage organisms. American Society of Brewing Chemists Proceedings 45-50. 131- O’Connor-Cox, E.S.C., Yiu, P.M., & Ingledew, W.M. (1991). Pasteurization: Thermal death of microbes in brewing. Technical Quarterly of Master Brewers Association of the Americas, 28, 67-77. 132- Bunker, H.J. (1955). The survival of pathogenic bacteria in beer. Proceedings of the European Brewery Convention Congress, 330–341. 133- Sakamoto, K., and Konings, W. N. (2003). Beer spoilage bacteria and hop resistance. International Journal of Food Microbiology, 89, 105–124. 134- Back, W. (1994). Secondary contamination in the filling area. Brauwelt Int. 4, 326–328. 135- Shimwell, J.L. (1937). On the relation between the staining properties of bacteria and their reaction towards hop antiseptic. Part I, II. Journal of the Institute of Brewing 43, 111–118. 136- Hoff S., Lund M.N., Petersen M.A., Frank W., Andersen M.L. (2013). Storage stability of pasteurized non-filtered beer. Journal of the Institute of Brewing, 119, 172–181. 137- Cao, L., Zhou, G. Q., Guo, P., and Li, Y. C. (2011). Influence of pasteurising intensity on beer flavour stability. Journal of the Institute of Brewing, 117(4), 587–592 138- Kaneda, H., Kano, Y., Osawa, T., Kawakishi, S., and Koshino, S. (1994). Free‐radical reactions in beer during pasteurization. International Journal of Food Science and Technology, 29(2), 195–200 139- Savel, J. (2004). Specific process units for temperature-dependent process evaluation in brewing. Technical Quarterly of Master Brewers Association of the Americas, 28, 43-50. 140- Portno, A.D. (1968). Pasteurization and sterilization of beer. Journal of the Institute of Brewing, 74, 291–300.

141- AEE - Institut für Nachhaltige Technologien. (2017, December 17). Pasteurization in beer production. Retrieved from http://wiki.zero- emissions.at/index.php?title=Pasteurization_in_beer_production (Consulted in 30/09/2019) 142- Horn, C. S., Franke, M., Blakemore, F. B., & Stannek, W. (1997). Modelling and simulation of pasteurization and staling effects during tunnel pasteurization of bottled beer. Food and Bioproducts Processing, 75(C1), 23–33. 143- ASBC. (2010). Triangle Test. Methods of Analysis. Retrieved from http://methods.asbcnet.org/login.aspx?ReturnUrl=%2fmethods%2fsensoryanalysis- 7.pdf (Consulted in 27/09/2019) 144- ASBC. (2010). Descriptive Analysis. Methods of Analysis. Retrieved from http://methods.asbcnet.org/login.aspx?ReturnUrl=%2fmethods%2fsensoryanalysis- 10.pdf (Consulted in 27/09/2019) 145- Andrés-Iglesias, C., Nešpor, J., Karabín, M., Montero, O., Blanco, C. A., Dostálek, P. (2016). Comparison of carbonyl profiles from Czech and Spanish lagers: traditional and modern technology. Food Science Technology, 66, 390–397. 146- Nyborg, M., Outtrup, H., and Dreyer, T. (1999). Investigations of the protective mechanism of sulfite against beer staling and formation of adducts with trans-2-nonenal. Journal of the American Society of Brewing Chemists, 57, 24-28. 147- Lund, M. N., Hoff, S., Berner, T. S., Lametsch, R., Andersen, M. L., (2012). Effect of pasteurization on the protein composition and oxidative stability of beer during storage. Journal of Agricultural and Food Chemistry, 60, 12362− 12370. 148- Takoi, K., Koie, K., Itoga, Y., Katayama, Y., Shimase, M., Nakayama, Y., Watari, J. (2010b). Biotransformation of Hop-Derived Monoterpene Alcohols by Lager Yeast and Their Contribution to the Flavor of Hopped Beer. Journal of Agricultural and Food Chemistry, 58(8), 5050-5058. 149- ASBC. (2010). Beer Flavor Database. Methods of Analysis. Retrieved from http://methods.asbcnet.org/Flavors_Database.aspx (Consulted in 21/10/2019) 150- Forster, A., Gahr, A. (2013). On the fate of certain hop substances during dry hopping. Brewing science, 66(7), 93-103. 151- King, A., Dickinson, R. J. (2000). Biotransformation of monoterpene alcohols by Saccharomyces cerevisiae, Torulaspora delbrueckii and Kluyveromyces lactis. Yeast, 16(6), 499-506. 152- Kollmannsberger, H., Biendl, M., Nitz, S. (2006). Occurence of glycosidically bound flavour compounds in hops, hop products and beer. Monatsschrift für brauwissenschaft, 5(6), 83-89. 153- Lermusieau, G., Bulens, M., Collin, S. (2001). Use of GC−Olfactometry to Identify the Hop Aromatic Compounds in Beer. Journal of Agricultural and Food Chemistry, 49(8), 3867-3874.

154- Yang, X., Lederer, C., McDaniel, M., Deinzer, M. (1993). Chemical analysis and sensory evaluation of hydrolysis products of humulene epoxides II and III. Journal of Agricultural and Food Chemistry, 41(8), 1300-1304. 155- Pickett, J. A., Coates, J., Sharpe, F. R. (1975). Distortion of essential oil composition during isolation by steam distillation. Journal of Agricultural and Food Chemistry 13, 571-572.

6. Appendices

Appendix I – Calibration lines designed for carbonyl analysis

Figure 34- Calibration lines graphics designed for 2-methylpropanal, 2-methylbutanal, 3-mehtylbutanal and 2,3-pentanedione.

Figure 35- Calibration lines graphics designed for furfural, heptanal, diacetyl and octanal.

Figure 36- Calibration lines graphics designed for benzaldehyde and (E)-2-nonenal.

Table 17- Calibration equations obtained for each analyzed carbonyl and respective coefficient of determination (R2).

Compound Equation R² 2-Methylpropanal y = 16 812 515,888x - 1 644 927,218 0,9766 2-Methylbutanal y = 24 705 334,768x - 3 651 804,540 0,9774 3-Methylbutanal y = 42 626 211,252x + 1 289 268,103 0,9806 2,3-Pentanedione y = 763 372,749x + 280 375,232 0,8816 Furfural y = 58 410,799x - 53 594,612 0,9669 Heptanal y = 49 644 793,789x + 4 624 984,391 0,9949 Diacetyl y = 1 073 929,045x - 154 192,227 0,9808 Octanal y = 45 115 523,509x + 6 038 560,954 0,9922 Benzaldehyde y = 1 017 954,242x + 177 790,814 0,9955 (E)-2-Nonenal y = 8 959 075,211x + 82 705,333 0,9833

Appendix II – Supporting data for analysis of triangle tests results

훼-risk: probability of concluding that a perceptible difference exists when one does not, also known as type I error, significance level or false positive rate.

• 0.10 – 0.05: slight evidence that a difference is apparent;

• 0.05 – 0.01: moderate evidence that a difference is apparent;

• 0.01 – 0.001: strong evidence that a difference is apparent;

• < 0.001: very strong evidence that a difference is apparent.

Most often an 훼-risk of 0.05 is used. On Table 18, n is referred to the number of assessments (tasters) and “훼” is the probability of concluding that a perceptible difference exists when one does not.

Table 18- Minimum number of correct responses required for samples to be considered significantly different at the stated α-level.(143)

Appendix III – Supporting data for analysis of carbonyl aging markers

Table 19- Description of studied carbonyl compounds, regarding the chemical group they belong to, as well as their respective flavor thresholds (μg/L), formation mechanisms and flavor descriptors.(145)

Appendix IV – Monastery IPA’s sensory tests results sheets

The results sheets from sensory tests performed with Monastery IPA samples stored for periods of 0, 2 and 4 months, mentioned in section 3.3, are presented below on Table 20, 21 and 22.

Table 20- Sensory tests results for Monastery IPA samples stored for 0 months.

Table 21- Sensory tests results for Monastery IPA samples stored for 2 months.

Table 22- Sensory tests results for Monastery IPA samples stored for 4 months.

Appendix V – IUPAC names of hop essential oils analyzed compounds

β-Myrcene – 7-Methyl-3-methylene-1,6-octadiene D-Limonene – 1-Methyl-4-prop-1-en-2-ylcyclohexene Linalool – 3,7-Dimethylocta-1,6-dien-3-ol Camphor – 1,7,7-Trimethylbicyclo[2.2.1]heptan-2-one Methyl Geranate – Methyl (2E)-3,7-dimethylocta-2,6-dienoate Geraniol – (2E)-3,7-Dimethylocta-2,6-dien-1-ol β-Caryophyllene – (1R,9S)-4,11,11-Trimethyl-8-methylenebicyclo[7.2.0]undec-4-ene 훼-Humulene – (1Z,4Z,8Z)-2,6,6,9-Tetramethylcycloundeca-1,4,8-triene 훼-Calacorene – (1S)-4,7-Dimethyl-1-propan-2-yl-1,2-dihydronaphthalene (E)-Nerolidol – 3,7,11-Trimethyldodeca-1,6,10-trien-3-ol Humulol – (3E,7E)-1,5,5,8-Tetramethylcycloundeca-3,7-dien-1-ol Humulene epoxide I – (1R,3E,7E,11R)-1,5,5,8-Tetramethyl-12-oxabicyclo[9.1.0]dodeca-3,7- diene (Z,Z)-α-Farnesene – (3Z,6Z)-3,7,11-Trimethyldodeca-1,3,6,10-tetraene τ-Cadinol – (1R,4S,4aS,8aS)-1,6-Dimethyl-4-propan-2-yl-3,4,4a,7,8,8a-hexahydro-2H- naphthalen-1-ol α-Cadinol – (1S,4R)-1,6-Dimethyl-4-propan-2-yl-3,4,4a,7,8,8a-hexahydro-2H-naphthalen-1-ol Farnesol – 3,7,11-Trimethyldodeca-2,6,10-trien-1-ol