<%BANNER%>

Pasteurization of Beer by a Continuous Dense-Phase CO2 System


PAGE 1

PASTEURIZATION OF BEER BY A CONTINUOUS DENSE-PHASE CO2 SYSTEM By GILLIAN FOLKES A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Gillian Folkes

PAGE 3

This document is dedicated to beer drinkers everywhere. Enjoy.

PAGE 4

ACKNOWLEDGMENTS I would like to Dr. Charles Sims, Dr. Marty Marshall, Dr. Al Wysocki, Dr. Andre Khuri, and especially Dr. Murat Balaban, my major advisor. He not only has supported me with encouragement and challenges, but has also been an exceptional mentor and friend. Secondly, I would like to thank my parents who were not only very good at raising a child, but were also exceptional at raising an adult. They gave me a solid foundation not only for academics, but also for life and love. Finally, I would like to thank my husband, Roi Dagan, who knew exactly what to say, even when he said, I love you. Now go do some work! He truly understands what research is all about. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT.......................................................................................................................xi CHAPTER 1 INTRODUCTION........................................................................................................1 2 LITERATURE REVIEW.............................................................................................5 Beer Production and Consumption in the United States..............................................5 Beer Composition.........................................................................................................6 Yeast Cultures in Beer..................................................................................................7 Beer Quality..................................................................................................................8 Beer Color.....................................................................................................................9 Beer Haze....................................................................................................................10 Composition of Beer Haze..................................................................................10 Formation of Haze...............................................................................................11 Haze Removal and Beer Stabilization Techniques.............................................13 Beer Foam...................................................................................................................15 Beer Foam Composition......................................................................................15 Beer Foam Formation..........................................................................................16 Beer Foam Stabilization......................................................................................16 Beer Flavor.................................................................................................................17 Processing of Beer......................................................................................................18 Pasteurization......................................................................................................18 Flash pasteurization......................................................................................19 Nonthermal methods....................................................................................19 Effects of Dense-Phase CO2 Pasteurization...............................................................20 Microbiology.......................................................................................................20 Theories of Cell Death by Dense-Phase CO2 Pasteurization..............................23 Effects of Cosolvents...........................................................................................24 Quality Attributes................................................................................................25 Dense-Phase CO2 Pasteurization of Beer...................................................................26 Objectives...................................................................................................................26 v

PAGE 6

3 MATERIALS AND METHODS...............................................................................28 The Dense-Phase CO2 System....................................................................................28 Beer Samples..............................................................................................................29 Experimental Design..................................................................................................30 Cleanability Study...............................................................................................30 Experimental Design...........................................................................................31 Procedures...........................................................................................................32 Analysis of Treated Samples......................................................................................33 Microbial Reduction Experiments.......................................................................33 Haze Measurement..............................................................................................33 Foam Capacity and Stability Measurements.......................................................33 Polyacrylamide Gel Electrophoresis of Beer Proteins........................................34 Flavor...................................................................................................................34 Sensory................................................................................................................36 Fresh beer vs. dense-phase CO2 processed beer...........................................37 Dense-phase CO2 processed beer vs. heat pasteurized beer.........................37 Storage studies..............................................................................................38 Statistical Analysis......................................................................................................38 Conjoint Analysis of Beer Purchase Decision............................................................38 4 RESULTS AND DISCUSSION.................................................................................39 Microbial Reduction Experiments..............................................................................39 Mode of Cell Death....................................................................................................44 Effect of Dense-phase CO2 Processing on Haze........................................................46 Effect of Dense-phase CO2 Processing on Foam Capacity and Stability...................48 Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Sensory Evaluation..............................................................................................................53 Beer Aroma.........................................................................................................53 Beer Flavor..........................................................................................................56 Beer Aroma After Storage...................................................................................59 Beer Flavor After Storage...................................................................................61 Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Gas Chromatography-Olfactometry and Mass Spectrometry.......................................65 Conjoint Analysis of Beer Purchase Decisions..........................................................72 5 CONCLUSION...........................................................................................................75 vi

PAGE 7

APPENDIX A RAW EXPERIMENT DATA.....................................................................................77 B SENSORY AND CONJOINT BALLOTS.................................................................88 C STATISTICAL MATERIAL.....................................................................................90 LIST OF REFERENCES...................................................................................................92 BIOGRAPHICAL SKETCH.............................................................................................98 vii

PAGE 8

LIST OF TABLES Table page 3-1 Experimental Design................................................................................................32 4-1 Microbial Reduction Results....................................................................................42 4-2 Beer Haze in NTU After Processing and After Storage at 1.67 C for 30 Days......47 4-4 Retention Times and Aroma Descriptors for Compounds Detected by Both Assessors on ZB-5 and Carbowax Columns............................................................65 4-5 Linear Retention Indices and Identification of Compounds using GC-O ...............67 4-6 Average Integration Areas of Identified Compounds in Fresh and CO2 Processed Beer.........................................................................................................69 4-7 Conjoint Analysis Transformation...........................................................................74 A-1 Yeast Counts for 27 Treatment Combinations Done in Duplicate...........................78 A-2 Foam and Liquid Volumes.......................................................................................87 viii

PAGE 9

LIST OF FIGURES Figure page 2-1 Yeast Growth Stages during Fermentation ...............................................................8 2-2 Molecular Mechanisms for Haze Formation in Beer ..............................................14 3-1 Schematic of the Continuous Dense-Phase CO2 Pasteurization System..................29 4-1 Scanning Electron Microscopy Picture of Yeast in Fresh Beer...............................45 4-2 Scanning Electron Microscopy Picture of Yeast in Beer Dense-Phase CO2 Processed at 27.6 MPa, 10% CO2, at 21C, With a Residence Time of 5 Minutes.....................................................................................................................45 4-3 Scanning Electron Microscopy Picture of Yeast in Beer Pasteurized at 74C for 30 Seconds................................................................................................................45 4-4 Beer Haze Following Processing and Following Storage at 1.67C for 30 Days....47 4-5 Foam Capacity and Stability of Beer Samples After Processing.............................50 4-6 Foam Capacity and Stability of Beer Samples After Storage at 1.67C for 30 Days.....................................................................................................................50 4-7 Aroma Evaluation of Fresh and CO2 Processed Beer Samples................................54 4-8 Evaluation of Aromas of Fresh, CO2 Processed and Heat Pasteurized Beer Samples....................................................................................................................56 4-9 Evaluation of Beer Flavor Between Fresh and CO2 Processed Samples.................57 4-10 Evaluation of Beer Flavors Between Fresh, CO2 Processed, and Heat Pasteurized Samples.................................................................................................58 4-11 Aroma Evaluation of Aroma Between Aged, CO2 Processed, Heat Pasteurized, and a Non-Stored, Fresh Hidden Control.................................................................60 4-12 Evaluation of Flavor Between Aged, CO2 Processed, Heat Pasteurized, and a Non-Stored, Fresh Hidden Control........................................................................61 ix

PAGE 10

4-13 Evaluation of Flavor Between Fresh and CO2 Processed Beer After Storage at 1.67C for 30 Days, and a Non-Stored, Fresh Hidden Control................................63 4-14 Evaluation of Flavor of Fresh, CO2 Processed, and Heat Pasteurized Samples After 30 Days at 1.67C, and a Non-Stored, Fresh, Hidden Control.......................64 4-15 Typical FID Chromatogram for Fresh Beer.............................................................70 4-16 Typical Aromagram for Fresh Beer.........................................................................70 A-1 Polyacrylamide Gels................................................................................................87 B-1 Sample Sensory Ballot.............................................................................................88 B-2 Sample Ballot for Conjoint Analysis.......................................................................89 x

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PASTEURIZATION OF BEER BY A CONTINUOUS DENSE-PHASE CO2 SYSTEM By Gillian Folkes August 2004 Chair: Murat Balaban Major Department: Food Science and Human Nutrition The world production of beer grew 26% between 1987 and 1997, despite the growing market of competing beverages. In 1999-2000, $7.7 billion worth of beer was produced in the United States. Bottled beer is currently flash-pasteurized. Because beer is a delicate beverage, off-flavors are easily formed during pasteurization. With freshness being top priority, it is evident a method of pasteurization using no heat would be of great help to the brewing industry. Currently there is great interest in dense-phase CO2 as an alternative processing method, and studies using the combination of carbon dioxide and pressure for pasteurization have been successful. Not only is microbial inactivation achieved, but also no taste or aroma changes are perceived, and vitamin quality is maintained. The purpose of this investigation was to evaluate the effectiveness of dense-phase CO2 pasteurization system with beer. A predicted maximum log reduction in yeast xi

PAGE 12

populations of 7.38 logs was seen at 26.5 MPa, 21C, 9.6% CO2, and 4.77 minutes residence time. Haze was slightly reduced by dense phase CO2 pasteurization from 146 NTU to 95 NTU. At this same treatment combination, aroma and flavor of beer sample means were not considered significantly different (p=.3415) from fresh beer sample means when evaluated in a difference from control test, using fresh beer as the reference. Foam capacity and stability were affected minimally by CO2 processing, however changes would most likely be unnoticed by consumers. In addition, a conjoint analysis was performed to examine motives during beer purchase decision. The attribute eliciting the larges influence on purchase decision was lower price, followed by a preference for draft beer taste and an extended shelf life. Furthermore, indications of the mode of cell death were absorption of CO2 into the cell membrane creating a physical disruption in membrane structure, visible as divots in scanning electron microscopy pictures. A continuous dense-phase CO2 system was effective in the pasteurization of beer. The ability to produce a clear, consistently fresh beer that forms a good head, with an extended shelf life is top-priority to brewers. Dense-phase CO2 pasteurization can make this possible. xii

PAGE 13

CHAPTER 1 INTRODUCTION Beer dates back to 4000 B.C. when the Babylonians described ale in some of the worlds oldest writings; however it is believed that beer existed far before these records (Toussaint-Samat et al., 1991). Egyptians believed that Osiris, the god of agriculture, made a decoction of barley that had germinated in the Nile river. Becoming distracted he left it in the sun and forgot it. Upon his return the liquid had fermented. He drank it and proclaimed mankind should profit from it (Toussaint-Samat, 1992). The world production of beer grew 26% between 1987 and 1997, despite the growing market of competing beverages (Gonzalez del Cueto and Miguel, 1999). In 1999-2000, $7.7 billion worth of beer was produced in the United States (Summerour, 2001). Currently bottled beer is flash-pasteurized. Because beer is a delicate and heat labile beverage, off-flavors are easily formed during pasteurization. In a study by Kaneda et al. (1994), non-pasteurized beer was compared to pasteurized bottled beer. Off-flavors of bottled beers pasteurized at 15-30 pasteurization units (P.U.) had a similar off-flavor profile to non-pasteurized beers stored at 20C for 6-10 days. With freshness being top priority, it is evident a method of pasteurization using no heat would be of great help to the brewing industry. Its use would lead to a beer with a longer shelf life, cheaper production and distribution cost because of the elimination of some refrigerated distribution centers and trucks, and a fresher taste and aroma. Freshness is of utmost importance in the brewing industry. Current evidence of this is the use of colored glass 1

PAGE 14

2 bottles, refrigerated distribution houses and trucks, and the popularization of the born-on date. One method of non-thermal pasteurization is high hydrostatic pressure processing. There are a variety of foods such as jams, jellies, guacamole, and pate where high isostatic pressure technology is applied to inactivate organisms at an efficacy equal to that of traditional thermal pasteurization methods. The use of high pressure insures no noticeable changes in taste or flavors of these foods and does not destroy vitamins (Hoover, 1998). The combination of high pressure pasteurization with other preservation methods could lead to a more desirable process or product. The effect of the combination of pressure, time, and temperature on food has been studied, showing that with the use of heat and pressure effective pasteurization can take place at low to moderate pressures (Kalchayanand et al., 1998). Since achieving high pressures can be expensive, the combined methods insure not only safe food products but also lower costs. However, the use of heat still may degrade vitamins or change the taste, aroma, or texture of the food. In addition, high hydrostatic pressure processing is currently a batch process. For the large volumes that would be processed in the case of beer, a continuous process is more desirable. Currently there is great interest in dense-phase CO2 as an alternative processing method, and studies using the combination of carbon dioxide and pressure for pasteurization have been successful. Not only is microbial inactivation achieved, but also no taste, or aroma changes are perceived and vitamin quality is maintained. The addition of carbon dioxide to the high pressure treatment allows pressures as low as 14 to 107

PAGE 15

3 MPa to be used for effective microbial inactivation (Kincal, 2000), instead of 400-900 MPa. These pressures are lower than those in the combined pressure/heat treatment, and this creates a cost effective alternative pasteurization method because of the small amount of energy used for pressurization and the use of an inexpensive gas. The need for inhibition of pathogen/spoilage organism growth and the characteristic carbonation of beer make it an excellent candidate for dense-phase CO2 pasteurization. The purpose of this investigation was to evaluate the effectiveness of dense-phase CO2 pasteurization system for beer. The success of this system will rely on its ability to inactivate microorganisms, preserve fresh beer taste and aroma, not exacerbate or possibly prevent beer haze, and insure proper foam formation and stabilization. Therefore a study examining these characteristics was conducted in order to compare it with current pasteurization methods and to make inferences regarding the intermolecular changes occurring in beer during the high pressure carbon dioxide pasteurization. Although there is promise in the use of dense-phase CO2 pasteurization technology with beer, research must be done to insure the technical and economic feasibility of the procedure. The objectives of this study were: 1. To quantify and elucidate mechanisms for inactivation of yeast by dense-phase CO2 pasteurization of beer 2. To prove there are no significant changes in flavor and aroma of beer after dense-phase CO2 pasteurization 3. To prove there are no significant changes in chill haze formation of beer after dense-phase CO2 pasteurization 4. To prove there are no significant changes in foam formation and stability after dense-phase CO2 pasteurization

PAGE 16

4 5. To evaluate consumer and industry acceptability and the economic impacts of dense-phase CO2 pasteurization

PAGE 17

CHAPTER 2 LITERATURE REVIEW Beer Production and Consumption in the United States The brewing industry in the United States is an active part of the national economy, paying billions of dollars annually in taxes and wages. One sign that beer is a significant force in todays economy is that it is included in the basket of goods used to calculate the Consumer Price Index. Currently the U.S. brewing industry employs approximately 1.66 million workers and pays over $47 billion in wages annually (The Beer Institute, 2003a, 2003b, 2003c). In 2002, the U.S. brewing industry recorded its seventh straight year of growth. Currently, there are over 3,500 brands of malt beverages on the market and in the year 2000 it was estimated that 199,650,000 barrels of beer were produced in the United States. For the same year, consumption of malt beverages reached 21.8 gallons per capita in the U.S. (The Beer Institute, 2003a, 2003b, 2003c). Beer consumption is mostly male-dominated, with men accounting for more than 80% of the volume consumed. A large number of these beer drinkers are white and prefer domestic light beer, followed by domestic draft beer. African American drinkers make up about 10% of the beer market. In general, they are the biggest consumers of malt liquors, followed by ice beer. Considering all beer styles, light beer has the strongest following among women consumers. Women beer drinkers are also more attracted to specialty micro-brewed beers because of their greater variety (Goldammer, 2000). 5

PAGE 18

6 Beer Composition Beer contains a variety of components, and although many do not regard beer as a health food, many components of beer are conducive to good health. Beer is mostly water and is made from malted grains such as barley or wheat, hops, yeast, and sometimes adjuncts such as corn or rice. Beer contains about 4-5% alcohol by volume, which if consumed moderately can help reduce the risk of cardiovascular disease (Baxter, 2000). On average, a beers two main components, excluding water, are carbohydrates (1 to 60 g/liter) and proteins (2 to 6 g/liter) usually in the form of peptides. The carbohydrates are found in the form of branched dextrans and not as free sugars, which would have been consumed by the yeast during fermentation. These dextrans not only have little immediate impact on blood sugar levels as free sugars but also are less cariogenic. Another health claim of beer is that it contains no fat (Baxter, 2000). Because beers are made from malted grains they are a good source of B vitamins. Beer is generally not considered as the main source of B vitamins; however it becomes an important source of these in malnourished societies. Yeast also contributes to the B vitamins in beer. For example, about 1 liter (2 pints) of beer may provide one third to one half of a consumers daily requirement of 5 B vitamins, including folate. From the use of malted barley also comes silicon, which maintains healthy bones, magnesium, and potassium (Baxter, 2000). Beverages such as fruit juices and wines have already been recognized for their readily available antioxidants. Beer contains two distinct sources of antioxidants: melanoidins and polyphenols. Melanoidins are formed during the roasting of malted barley from Maillard reactions. Polyphenols come from not only the malt in the form of

PAGE 19

7 ferulic acid, but also from hops in the form of catechin. Although antioxidant levels are comparatively low in beer, their high bioavailability makes beer as competitive as other foods rich in antioxidants (Baxter, 2000). Yeast Cultures in Beer In the production of beer, only 4 ingredients are necessary: water, malt, hops, and yeast. After germination of the barley to create malt, the grain is ground and water added. This solution is then boiled to create the wort. It is then cooled and the yeast is added to start the fermentation. In the brewing operation, yeast functions as the means of transforming the fermentable sugars in the malt into alcohol, CO2, and heat through fermentation, as shown in the Gay-Lussac equation: C6H12O52(C2H5OH) + 2(CO2) +heat There are two types of yeast used in brewing: top and bottom fermenters. Ales are made from top fermenting yeast, most commonly Sacchromyces cerevisiae, and lager beers are made from bottom fermenting yeasts such as Sacchromyces carlsbergensis. Although yeasts are usually referred to as facultative anaerobes, they are actually aerobes. During fermentation of beer, yeasts switch between oxidative and fermentative metabolisms, depending on the presence of oxygen; however they cannot grow anaerobically indefinitely. The cell membrane of all eukaryotes, yeasts included, contain unsaturated fatty acids and sterols. These compounds can only be produced by the yeast under aerobic conditions and although these compounds do exist in wort, the amounts are too low to sustain yeast growth. Therefore, when pitching or adding the initial yeast culture to the wort to start the fermentation, approximately 1 x 107 to 2 x 107 cells per ml are added. Because the wort is oxygenated prior to fermentation, yeast numbers may only increase by three cell divisions, or a factor of 8. Subsequent multiplication of the

PAGE 20

8 yeast is inhibited because of the lack of further aeration of the wort which would lead to off flavors in the beer. Furthermore, fermentation includes the lag, exponential growth, and stationary phases of yeast growth as seen in Figure 1. Figure 2-1: Yeast Growth Stages during Fermentation (modified from Campbell, 1997). In kegged or draft beer, because it is not pasteurized after packaging, yeast cultures are still viable. Without proper refrigeration or pasteurization the shelf life of this beer is only several hours, because the yeast cultures will continue to ferment the beer and will create off-flavors in the process. With proper refrigeration or pasteurization this is prevented. Bottled beer is flash pasteurized to make it a shelf stable product between 71.5 to 74 C and held for 15 to 30 seconds. The pasteurization step kills all yeast that remain in beer after fermentation and packaging are complete (Goldammer, 2000). Beer Quality The importance of quality in the brewing industry is evident in many of todays brewing practices. The popularization of the born-on date and extensive expenditures

PAGE 21

9 on advertising such characteristics as quality ingredients . time honored practices . finest heritage of certain beers, highlight the industrys pride in and strive towards quality. The visual quality of a beer is the first to be judged by the consumer. The package of beer, either the draught, bottle, or can, can easily help or hinder the sales of a product. Recent innovations in the area of packaging have been tamper proof kegs, oxygen scavenging crowns for bottles, narrowing the necks of cans to conserve metal, the use of a widget which creates better foam in beers of low carbonation, and the use of plastic bottles for sales of beer in recreational areas where glass bottles may cause danger (Bamforth, 2000). The next characteristics to be inspected are the beers color, clarity, and foaming capabilities, and then finally aroma and flavor. Beer Color Beer color is mostly dependent on the color of the malt after roasting and the other solid grist materials used to make the beer. The color forming molecules in malt and grist are primarily melanoidins, which were formed during the roasting of the malt by Maillard browning reactions. During the malt roast, the more intense the kilning the darker the malt and the resulting beer. Sugar content of the malt also determines the amount of browning that will occur, with higher modified grains having more sugars and darker color after roasting (Bamforth, 2003). The second source of color in brewing is the possible oxidation of polyphenols and tannins. These compounds originate in the malt and hops in beer and if large amounts of oxygen are allowed into the brewing process further darkening of the beer will occur. This is similar to polyphenol oxidase reactions in apples, potatoes, and mushrooms (Bamforth, 2003).

PAGE 22

10 Beer Haze Haze can be defined as the formation of a colloidal suspension that scatters light and makes a beverage appear cloudy. In beer there are two classes of hazes: biological and non-biological. Biological haze is irreversible and results from an infection of beer by wild yeasts or bacteria, resulting in spoilage. Non-biological haze is defined as being either chill or permanent haze resulting from native constituents of the beer. Chill haze is haze that forms upon cooling beer to 0C and redissolves upon warming to 20C or more. The term permanent haze should be used for haze which remains in beer at 20C or above. Composition of Beer Haze Haze in beer is formed by interactions between proteins and polyphenolic compounds. Most proteins that are haze-active contain large amounts of the amino acid proline and originate from the barley protein hordein (Asano et al., 1982). Hordein is a proline-rich prolamine, or alcohol soluble protein. Because proline has a cyclized side chain and can form cis bonds, the frequent inclusion of this amino acid in a proteins sequence gives the protein a more elongated, flexible structure. This allows greater interactions with polyphenolic compounds and in turn greater haze production. On the contrary, tighter coiled proteins have a significantly lower affinity for polyphenolic compounds. More specifically, a 19 kDa proline-rich protein has been found by immunoelectrophoretic analysis by Asano et al. (1982) to have significant positive effects on beer haze. Further, Hejgaard and Kaergaard (1983) found a 40 kDa beer protein involved in both foam and beer haze. In 1993, Kano and Kamimura indicated that both

PAGE 23

11 20 and 40 kDa protein fractions contributed to foam and colloid stability of beer with the 20 kDa fraction correlating more with haze formation than the 40 kDa protein. Polyphenols can be classified into 2 groups: haze-active and non-haze active polyphenols. They are classified by the number of binding sites per molecule, with haze-active polyphenols having two or more sites per molecule. Multiple sites allow the polyphenol not only to interact with one protein but also to cross-link with other proteins to create the colloid that results in haze (Seibert and Lynn, 1998). Characteristics of a haze-active polyphenol are not only multiple binding sites per molecule, but also the ability to form multiple hydrogen bonds with proteins through the many phenolic groups on each molecule. A phenol of approximately 1,000 Da usually has between 12 and 16 phenolic hydroxyl groups. This facilitates the interaction of multiple proteins with the polyphenol (Haslam, 1998). The polyphenols in beer that are naturally occurring haze-active compounds are proanthocyanidins. More specifically the most predominant proanthocyanidin dimers in beer are procyanidin B3 and prodelphinidin B3 (McMurrough et al., 1992). Formation of Haze Much work has been done on the dynamics of the polyphenol/protein interactions. Although the type of protein and polyphenol involved does effect the haze produced, when one discusses a fixed system such as a certain beer where the native proteins and polyphenols do not differ, pH exerts a significant effect on haze produced. For beer the most haze will be seen at pH 4.0-4.2, with less haze forming at lower and higher pHs. Haze formed at pH 3.0 was only 1/7 of the amount resulting at pH 4.0 (Siebert et al., 1996b).

PAGE 24

12 The effect of pH directly correlates to what many researchers have noted as the driving force for protein/polyphenol interactions: hydrophobic effects (Oh et al., 1980; Siebert et al., 1996b; Haslam, 1998). Hydrogen bond deployment by polyphenols is best seen as a secondary feature that follows hydrophobically driven associations (Jencks, 1969; Haslam, 1998). The importance of hydrophobic interactions has also been demonstrated by Oh et al. in 1980 by observing the interactions of tannin and gelatin or other polyproline proteins. Interactions increased with increases in temperature and ionic strength as one would expect for hydrophobic interactions. Seibert and Troukhanova (1996a) developed a model of haze formation in beverage corresponding to haze-active protein and polyphenol concentrations. At a constant protein concentration, haze increased to a maximum and then declined as polyphenol concentration increased. It was hypothesized that this occurred because a haze-active protein would have a fixed number of sites to which a polyphenol could bind, and a haze-active polyphenol has two or more sites to bind to a protein. When the concentrations of proteins and polyphenol binding sites is closest, this results in maximum utilization of binding sites and cross-linking, resulting in large colloidal particles and maximum light scattering. However in a matrix such as beer, there is a large excess of haze-active protein to polyphenols and each polyphenol should be able to find a binding site on each protein. These protein dimers created would therefore not be cross-linked together and result in a small colloidal particle size and in turn less haze. However in all instances Seibert et al. (1996b) affirm that pH, ethanol content, and temperature all affect the extent of haze formation.

PAGE 25

13 Combinations of proteins with polyphenols generally seem to result from three different mechanisms. As shown in Figure 2, protein polyphenol interactions can occur as hydrogen bonding between oxygen atoms of peptide bonds and hydroxyl groups of polyphenols, hydrophobic bonding between hydrophobic amino acids such as proline, tryptophan, phenylalanine, tyrosine, leucine, isoleucine, and valine and the hydrophobic ring structure of polyphenols, and ionic nodding between positively charged groups of proteins and negatively charged hydroxyl groups of polyphenols. However at the acidic pH of beer, the first two mechanisms would come into effect, while the third would not because the hydroxyl groups of polyphenols would have no charge and ionic bonding could not occur (Asano et al., 1982). Haze Removal and Beer Stabilization Techniques There are three types of fining agents used in brewing: those that remove high molecular weight polyphenols to reduce chill-haze, those that remove high molecular weight proteins to reduce chill-haze, and those that reduce yeast biomass but do not affect chill-haze. The most common polyphenol agent used is polyvinylpolypyrrolidine or PVPP. Many brewers prefer using a polyphenol scavenger because it does not remove protein, which could lead to a reduction of foamability. Users of PVPP also benefit from the fact that because complex polyphenols are being removed, as the beer ages and simple phenolic compounds complex to form polyphenols, the amount of polyphenol present in the aged product will remain low. The amount of polyphenol is directly proportional to the amount of haze present. Therefore even aged beers will have less haze and retain foam formation if PVPP is used (Fix, 1999).

PAGE 26

14 Figure 2-2: Molecular Mechanisms for Haze Formation in Beer (Asano et al., 1982) Although foam-forming ability can be disrupted by some fining agents, silica gels are the most widely used protein scavengers because they are highly specific for haze-active proteins. Silica gels work by complexing with haze-active proteins as a polyphenol would, therefore disrupting the formation of a haze (Fix, 1999). Yeast-active agents are less used and will have no effect on chill haze. The most common yeast-active agent is Isinglass, which comes from the swim bladder of tropical fish (Fix, 1999).

PAGE 27

15 Beer Foam Foam can be defined as a thermodynamically unstable colloidal system in which gas is momentarily entrapped in a liquid matrix. Foams can be made in two ways: by supersaturation or mechanically. When a gas is dissolved in a liquid under pressure, such as in a carbonated beverage, and pressure is released the gas becomes supersaturated and gas bubbles form creating a foam. These bubbles do not form spontaneously, but instead form from air pockets that are already present on the side of the container. Another example of supersaturation foam formation is the formation of the foam structure in bread where CO2 collects in air pockets in the dough. Foams that are formed mechanically could be formed by sparging, beating, or shearing. This would be the case in whipped products like meringues (Walstra, 1996). Beer Foam Composition Beer foam is created when supersaturated CO2 is released and forms gas bubbles in the continuous liquid matrix. This matrix contains amphiphilic proteins in the beer, which migrate to the surface of the air/liquid interface. This migration is motivated by the decrease in the proteins free energy as the protein migrates out of solution to the interface. The result of this migration is that interfacial tension is lowered, which makes it more favorable for the two thermodynamically incompatible phases (air and liquid) to co-exist. The air/liquid interface is stabilized because the proteins have formed a strong viscoelastic film. Although all proteins are amphiphilic, they can differ greatly in the their surface activity, making some proteins better emulsifiers/stabilizers than others. Two characteristics that govern a proteins ability to act as a surface-active protein are the

PAGE 28

16 properties and topology of the proteins surface and its conformational stability, flexibility, and adaptability (Damodaran, 1996). Although an increasing amount of hydrophobic amino acids does mean a protein could be more surface active, it is the distribution of these amino acids that is the most important. A protein with clumps of hydrophobic groups will be a better foam stabilizer than one with randomly dispersed hydrophobic groups. Also, a protein that is in a molten globule state, or can more easily unfold at the air/liquid surface will be more surface-active and therefore a better foam former/stabilizer (Damodaran, 1996). Beer Foam Formation For surface-active proteins, there is a sequence of events that leads to formation and stabilization of the foam. First the protein has to have the ability to rapidly diffuse and absorb to the interface. Secondly, the protein should be able to rapidly unfold and reorient its polypeptide segments at the interface, and thirdly, while the protein is at the interface it should be able to interact with neighboring proteins or molecules to form a strong, continuously cohesive viscoelastic film, able to tolerate mechanical or thermal forces. The first two steps above are critical for the formation of a foam and the third is imperative for foam stability. Beer Foam Stabilization Foam stability is dependent on the presence of amphipatic polypeptides from malt, alpha-acids from hops, and the absence of lipophilic materials. Brewers insure good foam stability by the addition of propylene glycol alginate and the use of nitrogen gas. Nitrogen works by providing bubbles of very small diameter causing a higher concentration of small bubbles, which results in a more stable foam. Foam formation and stability may also affect the flavor and aroma of beer (Bamforth, 2000).

PAGE 29

17 Beer Flavor Because beer is a delicate and heat labile beverage, off-flavors are easily formed during pasteurization. In a study by Kaneda et al. (1994), non-pasteurized beer was compared to pasteurized bottled beer. Off-flavors, volatile aldehydes of bottled beers pasteurized at 15-30 pasteurization units (P.U.) had a similar off-flavor profile to non-pasteurized beers stored at 20C for 6-10 days. With freshness being top priority to brewers, it is evident a method of pasteurization using no heat would be of great benefit to the brewing industry. When evaluating a new non-thermal pasteurization method, the effect on beer flavor would need to be evaluated as well. Although, much flavor analysis in the brewing industry is done by panelists, analysis of flavors by instrumentation is also noteworthy. One technique that has become increasingly popular is solid phase microextraction (SPME). This method uses a small volume of sorbent dispersed typically on the surface of a fiber to isolate and concentrate analytes from a sample matrix. After either exposure to the headspace or wicking into a liquid sample matrix, analytes will absorb to the fiber until an equilibrium is reached. Analytes are then thermally desorbed into an analytical instrument for separation and quantification. It is a solvent-free, non-destructive, and simple preparation technique (Pawliszyn, 2001). SPME can be used in conjunction with gas chromatography-olfactometry (GC-O) to determine what flavor compounds may be of interest in a particular sample, or the importance of specific compounds in the overall perception of a foods flavor or aroma. In traditional gas chromatography, a mixture of volatile compounds partition between the gas and stationary phases and travel at different velocities, resulting in different residence times within the column. Upon eluting compounds can be characterized using various

PAGE 30

18 detectors (Kamimura et al., 2003). In GC-O, the separation capabilities of gas chromatography are coupled with the sensitivity and precision of the human nose. This method of flavor analysis detects compounds that have an influence on flavor or aroma of the food (Naim et al., 1998). GC-O produces an aromagram which indicates an olfactory impression as a function of retention time plus an intensity descriptor (Kamimura et al., 2003). GC-O has been used with success to describe hop aromas, off-flavors, and aged beer flavor compounds, using both immersion and headspace SPME sampling (Kamimura et al., 2003). Processing of Beer Pasteurization Most beer, packaged in bottles or cans, is pasteurized after filling into containers by passing through a steam tunnel. During this process, the bottles are passed under a series of water sprays with the temperature of the water increasing as beer passes through the tunnel. After reaching the desired temperature, bottles are cooled by a water spray and then exit the tunnel to air dry. The heat treatment used is referred to in terms of pasteurization units, with 1 PU = 60 C for 1 minute, with 5 PU resulting in a sufficient kill of approximately 102/ml yeast cells that would remain in pre-filtered beer. The steam tunnel used for bottles or cans can be operated at various temperatures, usually 60C or 62C for a longer time, depending on operating needs. Up to 30 PU may be applied depending on pre-pasteurization procedures such as filtering. Because heat treatment could adversely affect flavor, the choice of pre-pasteurization treatments and PU value is always a compromise between extending shelf life and beer quality (Campbell, 1997).

PAGE 31

19 Flash pasteurization Although flash pasteurization is not common in North American breweries, it is very popular in Europe and Asia. During flash pasteurization, beer is heated to at least 71.5 to 74 C and held for 15 to 30 seconds, resulting in 13.26 PU. This is achieved by the use of a twoor three-stage plate heat exchanger with hot water as the heat source. The adjustment of the flow rate determines the PU value for the treatment. After flash pasteurization, beer is then aseptically packaged into sterilized bottles or cans. Nonthermal methods The most common non-thermal processing technique used by brewers to extend shelf life is sterile filtration. Sterile filtration has been used as an alternative to pasteurization for many years. It has the advantage over pasteurization in that the risk of flavor damage by heat is eliminated. The term "sterile filtration" refers to the reduction of yeast and bacteria to levels that do not result in spoilage of the beer over its planned shelf life by use of one or many filters before packaging the beer in sterile containers. This can be accomplished without the loss of color or flavor compounds (Goldammer, 2000). The brewer sets a specification for the maximum allowable concentration of yeast and bacteria for quality control purposes, which may have entered the brewing process inadvertently, in sterile-filtered beer since not all microorganisms will be removed during the process. There are no critical levels for allowable microorganisms, and therefore extensive monitoring of the process must occur. The process is also labor intensive because of the time needed to clean the filters to prevent fouling (Goldammer, 2000).

PAGE 32

20 Effects of Dense-Phase CO2 Pasteurization Microbiology Recently there has been great interest in inactivation of microorganisms in food or model food systems using high pressure carbon dioxide. This new technology has been shown to inactivate microorganisms as well as conventional heat pasteurization without the loss of nutrients or quality changes. The use of carbonation as a means of preserving food started as early as 1939 with the study by Brown et al.(1939) where apple cider was carbonated and microbial inactivation and flavor changes were recorded. The carbonation of the juice was shown to preserve the cider for up to 3 months at approximately 21oC with no change in flavor. The use of carbonation was also investigated for its use in soft drinks as a preservation agent. Even at the lowest amount of gas pressure (3 volumes of CO2 where 1 volume=1 L of CO2 per L of beer) sterility was achieved on approximately the 20th day depending on the Brix of the beverage (Insalata, 1952). Researchers have also evaluated the use of pressurized CO2 and decompression to reduce microbial loads. In 1951, Fraser showed that 99% of E. coli numbers were rendered non-viable by a decompression of CO2 from 500 psi to atmospheric pressure. Although carbonation with CO2 has been shown as an affective preservative some bacteria are not affected. Molin (1983) examined the growth inhibiting effect of carbon dioxide on a variety of food related bacteria by sparging spiked growth media with the gas, arriving at a variety of pressures. Only partial success was attained. Although 100% carbon dioxide did slow the growth of all organisms some were affected less than others. CO2 had approximately 75% inhibitory affect on Bacillus cereus, Brochothrix thermosphacta, and Aeromonas hydrophila, and a 53%-29% inhibitory effect on

PAGE 33

21 Escherichia coli and Streptococcus faecalis. Inhibitory rates for anaerobic bacteria were even lower. This discovery proved that carbonation of foods alone would not inactivate all food-related bacteria. Because of the need for a preservation method that is safe, inexpensive, and non-destructive to heat sensitive compounds, the use of supercritical carbon dioxide (SC-CO2) was tested as a food preservation method on cells in cultures or broths. SC-CO2 was chosen because of its safety, cost, and high purity. CO2 also has a low critical pressure and critical temperature which result in excellent solvent power of CO2 when used in HPCD pasteurization. Kamihira et al.(1987) used HPCD to sterilize cultures of bakers yeast, Escherichia coli, and Staphylococcus aureus, however they were only successful with cultures with moisture contents from 70-90%. This same success was echoed in 1991 and 1992 when Saccharomycces cerevisiae was inactivated using sub and supercritical CO2 and both cell inactivation and disintegration were studied. (Lin et al., 1991; Lin et al., 1992a, 1992b; Nakamura et al., 1994) A similar study was performed again using bakers yeast and Bacillus magetarium in a spore form (Enomoto et al., 1997a, 1997b) However in all studies, in striving for not only the inactivation of cells but also their disintegration, the vessel must undergo pressurization and depressurization, in some cases, several times creating an energy-intensive process. However, this is only true in batch systems. In a continuous system depressurization would happen naturally. Similar studies have been performed using Leuconostoc (Lin et al., 1993) and Kluyveromyces fragilis, Saccromyces cerevisia, and Candida utilis (Isenschmid et al., 1995) showing that inactivation can occur without disruption of the cell wall, but instead through the leaching of cellular contents by the CO2. The effect on HPCD on bacterial

PAGE 34

22 spores has also been examined with the most successful inactivation occurring in the subcritical region of CO2 (Enomoto et al., 1997a). It has been shown that the amount of inactivation is proportional to the amount of dissolved CO2 in the sample. Kumagai et al. (1997) measured dissolved CO2 gravimetrically. Results showed higher inactivation of Saccharomyces cerevisiae as CO2 levels increased. Shimoda et al. (2001) found similar results and were able to model death kinetics of Saccharomyces cerevisiae as first-order through the critical temperature and pressure of CO2. Additionally, higher water activities and higher pressures have also shown higher inactivation because of their effects on CO2 sorption of the yeast cells (Kumagai et al., 1997). To further enhance this effect not only have higher pressures been used but also modifications to the necessary machinery have taken place. In studies by Shimoda et al. (1998) and Ishikawa et al. (1995, 1997), a filter was placed in the process vessel to create microbubbles of CO2 entering the vessel. It was shown that these microbubbles were more effective in inactivating bacterial cultures than the process without the filter. Another innovation has been the development of a semi-continuous system for dense-phase CO2 pasteurization. One such system was shown to be more efficient on the inactivation of Saccharomyces cerevisiae (Spilimbergo et al., 2003b). Conditions such as temperature and pressure have also shown significant effects on the antimicrobial properties of CO2. Studies on E. coli in Ringers solutions showed a decrease in survival rate with increases in dissolved CO2, increases in pressure1.2-5 MPa, and increases in temperature from 25-45C (Ballestra et al., 1996). Studies on Listeria monocytogenes also showed similar results in reference to pressure and temperature

PAGE 35

23 variations with a complete inactivation at 6.08 MPa CO2 treatment at 115, 75, and 60 minutes at 25, 35, and 45 C, respectively (Erkman, 2000a, 2000b). Recently most studies have concentrated on the inactivation of microorganisms in food samples rather than cultures or broths. Studies by Wei et al.(1991) showed the effectiveness of using CO2 on cultures of Listeria or Salmonella suspended in water and spiked onto food samples. The study showed the treatment to be applicable in some food systems: chicken, shrimp, orange juice, and egg yolk, but ineffective in a whole eggsalmonella mix, proving this technology may be dependent on the foods characteristics and their ability to shelter bacteria from of carbon dioxide. Staphylococcus aureus suspended in broth and compared to a suspension in raw milk was shown to undergo inactivation at lower pressures and in a shorter length of time. Also compared was raw milk was compared to orange, peach and carrot juices and the milk showed protection of the bacteria (Erkman, 1997, 2000c). Kimchi has also been studied because of its unique properties as a fermented food which may spoil if fermentation is not stopped at the appropriate time. In two different studies by Hong et al. (1997, 1999), Lactobacillus was shown to undergo inactivation after 200 min at 6.86 MPa CO2 pressure in kimchi compared to 30 min at approximately 13.73 MPa in the broth suspension. This affirms the point that the success of HPCD should be evaluated on a case-by-case basis for different foods. Theories of Cell Death by Dense-Phase CO2 Pasteurization Researchers have reported that the inactivation rate of all microorganisms is sensitive to pressure, temperature, and exposure time to CO2. Furthermore Hong et al. (1999) found similar results, but also concluded microbial inactivation was mainly due to

PAGE 36

24 the transfer rate of CO2 into the cells which could lead to viability loss. To further elucidate the mode of cell death by dense-phase CO2 pasteurization, much research has been performed. Using scanning electron micrographs, Ballestra et al. (1996) noted cell deformation of Escherichia coli after processing at 35C and 5MPa of CO2 and Spilimbergo et al. (2003a) noted the same results on gram-positive and gram-negative bacteria. Shimoda et al. (2001) studied the effect of the concentration of CO2 on Saccharomyces cerevisiae and found that cell death during continuous versus batch treatments was due to the anesthesia effect of CO2. This effect can be defined as loss of cell viability because of the diffusion of molecular CO2 into the plasma membrane of the cell, compromising the construction of membrane domains (Isenschmid et al., 1995). More specifically, while coining the term anesthesia effect, Isenschmid et al. (1995) found that this effect occurred at temperatures higher than 18C. At these higher temperatures, dependency was also noted on the dissolved CO2 concentration, with increases in temperature and dissolved CO2 concentration resulting in increased cell death. However, below 18C a solvent effect was observed as the reason for viability loss in yeast cells (Isenschmid et al., 1995). Effects of Cosolvents Because the solvent effect is suspected as a mode of cell death during dense-phase CO2 processing, one must also examine the effects of co-solvents such as ethanol that may be present in the food matrix. Solvent characteristics of CO2 can be greatly modified by the addition or existence of certain cosolvents. CO2 is a non-polar solvent, and has limited affinity for polar solutes. It is often used to extract organic solute molecules. A polar modifier can be added to CO2 to improve the solubility and

PAGE 37

25 selectivity for polar molecules. Most often the cosolvents of choice are lower alcohols. Cosolvents are usually added in 5-10% amounts by volume and even in these small amounts, can have significant effects, especially in surface processes. For example, the addition of ethanol may have significant effects on extraction by allowing increased absorption on surface sites, preventing the re-adsorption of the compound of interest (Clifford and Williams, 2000). The addition of a cosolvent modifies the critical temperature and pressure of the original fluid (Taylor, 1996). Therefore the solubility of the materials of low volatility is enhanced and as a result, lower pressures can be used to achieve the same extraction yield (Brunner and Peter, 1982). This has been shown to increase the amount of oil extracted from sunflower seeds (Raghuram Rao et al., 1992), and rape seed, soy bean, cottonseed, flax seed, and peanuts, and to enhance the extraction of herbal components such as borage seed oil and hiprose fruit (Illes et al., 1994). Quality Attributes Although the use of HPCD as a pasteurization technique for foods has been well studied, its effect on foods quality characteristics needs more examination. During the injection of CO2 the pH of the foodstuff drops dramatically. Because pH plays such a central role in regulation of food systems, it is imperative to be able to predict what will happen to quality attributes of food after HPCD processing. Models for pH of food systems processed with HPCD have been developed to help forecast pH extremes which may occur during processing (Meyssami et al., 1992). Specific research has also been done with single strength orange juice to examine HPCD effect on enzymes, cloud, color, and Brix value, and total acidity. It was shown that although significant pH changes can occur during processing, the final pH of the product after de-pressurization was not

PAGE 38

26 significantly changed, cloud was enhanced, enzymes were inactivated, and flavor and aroma were unaffected (Arreola et al., 1991). Commercially, the addition of CO2 has also been used to preserve cottage cheese with no effect on the mouthfeel or flavor of the product (Mermelstein, 1997). Flavor changes and protein viability of model systems have also been examined after rupture of yeast cells (Lin et al., 1991). Dense-Phase CO2 Pasteurization of Beer The need for inhibition of pathogen/spoilage organism growth and the characteristic carbonation of beer lend it to be an excellent candidate for HPCD pasteurization. Freshness is of utmost importance in the brewing industry. Current evidence of this is the use of colored glass bottles, refrigerated distribution houses and trucks, and the popularization of the born-on date. Although there is promise in the use of HPCD pasteurization technology with beer, research must be done to insure the technical and economic feasibility of the procedure. Parameters that must be optimized are the process pressure, amount of CO2 to add, process temperature, and residence time of the product under pressure. These parameters will be evaluated first on the basis of microbial log reduction and then on their effects of the quality attributes of taste, aroma, foam formation and stability, and haze. Therefore, a study examining these characteristics will be conducted in order to compare the process with current pasteurization methods and to make inferences regarding the intermolecular changes occurring during the high pressure carbon dioxide pasteurization. Objectives 1. To quantify and elucidate mechanisms of inactivation of yeast by dense-phase CO2 pasteurization of beer as a function of process pressure, temperature, residence time, and CO2 percentage

PAGE 39

27 2. To prove there are no significant changes in flavor and aroma of beer after dense-phase CO2 pasteurization 3. To prove there are no significant changes in chill haze formation of beer after dense-phase CO2 pasteurization 4. To prove there are no significant changes in foam formation and stability after dense-phase CO2 pasteurization 5. To evaluate consumer and industry acceptability of dense-phase CO2 pasteurization

PAGE 40

CHAPTER 3 MATERIALS AND METHODS The Dense-Phase CO2 System The continuous dense-phase CO2 system was constructed by APV (Chicago, IL) for Praxair (Chicago, IL) and given to the University of Florida (Gainesville, FL). The system is housed in the pilot plant of the Food Science and Human Nutrition Department at the University of Florida. The system mixes cooled, pressurized liquid CO2 with a liquid feed pressurized by its own pump (Figure 3-1). The mixture then proceeds through a holding tube (79.2 m, 0.635 cm ID) for a specified residence time, which is modified by changing the flow rate of the mixture. In the holding tube, temperature can be controlled by electrical heating tape, insulation, and a controller system, and operating pressure is maintained. When exiting the holding tube the mixture is depressurized by passing through a back pressure valve and is then ready for collection. The machine works by pressurizing the liquid feed first, to 6.89 MPa using a reciprocating pump with a stroke length of 30 mm and a back pressure valve and then to the desired operating pressure using a second reciprocating pump and another back pressure valve. CO2 is then introduced by a reciprocating pump at approximately 6.2 MPa or higher. Pressurizing both the liquid feed and the CO2 insures both will remain liquid and mix with the desired proportions. The CO2/feed mixture then passes through the second reciprocating pump that maintains the operating pressure which is always 6.89 28

PAGE 41

29 MPa or more. This pressure is maintained throughout the holding tube. The mixture then exits the holding tube, is depressurized, and collected in sterile bottles as aseptically as possible, using alcohol to sterilize the end of the holding tube. P Main Pump Juice stream Vacuum Heating systemHold tube TreatedCO2juiceCO tank2 Expansion valve Pump Pump Chiller 4 1 2 3 5 6 7 8 9 Figure 3-1: Schematic of the Continuous Dense-Phase CO2 Pasteurization System Pressure, temperature of the holding tube, flow rate (which in turn controls residence time in the holding tube), and weight % of CO2 were all controllable independent variables. Thermocouples and pressure sensors were also located throughout the machine to monitor operating parameters. Beer Samples Fresh beer samples (less than one month old) were purchased from Market Street Pub in Gainesville, FL. The beer was purchased in 58.7 L (15.5 gallon) kegs and transported to the pilot plant in the Food Science and Human Nutrition Department at the University of Florida and stored at 1.67C until used. The beer is an ale brewed using an

PAGE 42

30 all-barley malt extract and whole hops which insured a high level of consistency from batch to batch. The clarity of the beer had not been improved by the use of brewing aids. Experimental Design Cleanability Study Experiments were conducted to evaluate the cleanability of the dense-phase CO2 system. Principal and Oxonia (27.5% hydrogen peroxide, 5.8% peroxy acetic acid, 66.7% inert ingredients) (Ecolab, St. Paul, MN) were used as sanitizing agents. A non-pathogenic spoilage organism Lactobacillus fermentum (ATCC, Manassas, VA) was used in cleanability experiments because of its frequent use as a test organism for thermal inactivation studies. A freeze dried culture of Lactobacillus fermentum was rehydrated in 25 mls Lactobacillus MRS broth (Difco Laboratories, Sparks MD) and incubated at 37C for 24 hours. Butterfields phosphate buffer was used to dilute this culture to 106 CFU/ml. The machine was sanitized by pumping Principal at 50 ml per 18.9 L hot tap water, followed by 18.9 L cold tap water rinse and then followed by Oxonia at 946 ml per18.9 L cold water. Some Oxonia solution was left in the system overnight. The next day, 6 L of sterile water was used to rinse the machine after the remaining Oxonia had been pumped through. Then 6 L of the cell suspension was pumped through the sanitized machine, collected, and plated to measure recovery of the microorganisms. The machine was then rinsed with water and re-sanitized. The machine was then allowed to sit for 2 hours. Then 6 L sterile water followed by 6 L of sterile Butterfields phosphate buffer were run and collected into sterile bottles. The collected buffer was filtered using a vacuum pump and .45 micron mixed cellulose water testing filters (Fischer Scientific, Pittsburg, PA). 1.2 L of buffer was filtered per filter and all filters were plated on Lactobacillus MRS

PAGE 43

31 agar (Difco Laboratories a subsidiary of Becton, Dickinson, and Company, Sparks MD). Plates were incubated overnight at 37C. The recovered cell suspension plates showed that 105/ml lactobacilli were recovered after pumping the suspension through the machine, i.e. approximately 900,000 CFU remained in the machine after the cell suspension exited. After the sanitizing procedure was completed, collection of buffer pumped through the system resulted in no CFUs on the plates for the filters. Results show that the Principal and Oxonia solutions were effective in cleaning the machine without pressure. Experimental Design A Central Composite Design (CCD) was used because it is an economical design, allowing a researcher to fit a second order prediction equation to a response surface. The independent variables were pressure, CO2 %, residence time, and temperature. Twenty-seven treatment combinations were selected by CCD. The aim was to establish an optimum set of operating conditions based on the dependent variable of yeast population reduction. More experiments were conducted after this one on a subset of the 27 points to evaluate the effect on the quality attributes of haze, foam, and aroma/flavor changes. The subset was selected from the 27 points by choosing the most effective combination treatment for log reduction of yeast, a more economical version of the previous, and adding a heated beer sample (74C, 30 seconds), and an untreated, fresh beer control. The experimental design is shown in Table 3-1.

PAGE 44

32 Table 3-1: Experimental Design PressureTempCO2TimePressure (MPa)0.000-1.4140.0000.000-1=20.7-1.000-1.000-1.000-1.0000=27.61.000-1.000-1.000-1.0001=34.5-1.000-1.000-1.0001.0001.414214=37.31.000-1.000-1.0001.000-1.414214=17.8-1.000-1.0001.000-1.0001.000-1.0001.000-1.000Co2 (%)-1.000-1.0001.0001.000-1=81.000-1.0001.0001.0000=100.0000.000-1.4140.0001=120.0000.0000.000-1.4141.414214=12.83-1.4140.0000.0000.000-1.414214=7.170.0000.0000.0000.0000.0000.0000.0000.000Temp (degrees C)0.0000.0000.0000.0001-=251.4140.0000.0000.0000=350.0000.0000.0001.4141=450.0000.0001.4140.0001.414214=49-1.0001.000-1.000-1.000-1.414214=211.0001.000-1.000-1.000-1.0001.000-1.0001.000Residence1.0001.000-1.0001.000Time (minutes)-1.0001.0001.000-1.000-1=41.0001.0001.000-1.0000=5-1.0001.0001.0001.0001=61.0001.0001.0001.0001.414214=6.410.0001.4140.0000.000-1.414214=3.56 Procedures For every experiment, the dense-phase CO2 system was sanitized one day prior to experiments by the procedure described in the cleanability experiment. Then the machine was rinsed with 6 L of sterile de-ionized water. As the feed tank emptied, beer was added and allowed to run in the machine without pressure for at least 2 hold-up volumes (5 L) to insure all water had exited the machine. Operating parameters were set for each treatment and 2 hold-up volumes (approximately 5 L) of beer were allowed to pass to insure steady state conditions before sample collection. Samples were collected in sterile bottles. The 27 treatment combinations were run in duplicate on separate days, using different batches of beer.

PAGE 45

33 For heat pasteurized beer, fresh beer was pumped by a peristaltic pump (Cole Parmer, Chicago, IL) into a temperature controlled water bath (Hart Scientific, American Fork, UT) through copper tubing. The beer was pasteurized at 74 C for 30 seconds and collected in a sterile bottle. Copper tubing was sanitized using pumping soapy hot water, then hot water, then alcohol through the system prior to use. Analysis of Treated Samples Microbial Reduction Experiments All beer samples were plated on potato dextrose agar for enumeration of yeast colonies. Beer was diluted serially using 90 ml dilution bottles containing Butterfields phosphate buffer (Hardy Diagnostics, Santa Maria, CA). Dilutions were done in duplicate and plated in duplicate. Plates were incubated at 37C for 5 days. Haze Measurement Beer samples were evaluated for haze using a Hach turbidimeter (Hach, Loveland, CO) and the AOAC procedure for beer haze measurement (AOAC 10.013). Readings were recorded as NTU, nephelometric turbidity units. Foam Capacity and Stability Measurements Beer samples (25 ml) were degassed at room temperature for 24 hours by using a stir plate and stir bar, and placed in graduated cylinders with their tops cut off to facilitate the addition and removal of the homogenizer blade. A Virtus 45 homogenizer (The Virtus Company, Gardiner, New York) was used to mechanically foam the beer at setting 20 for 60 seconds, keeping the blade at an equal depth for all beer samples. At the end of the 60 seconds the homogenizer was turned off and the volume of foam produced in ml was recorded, as an indication of foam capacity. Sixty seconds after that, the amount of

PAGE 46

34 beer liquid collected in the bottom of the cylinders was recorded as well, as an indication of foam stability (Odabasi, 2003). Polyacrylamide Gel Electrophoresis of Beer Proteins Polyacrylamide gel electrophoresis (PAGE) was run on beer proteins to determine if beer proteins were effected by dense-phase CO2 pasteurization. Beer samples (1ml) were centrifuged at 1.6C at 10,000g for 15 minutes in 30 kDa Millipore Centricon Centrifugal Filter Devices (Millipore, Bedford, MA) to concentrate beer proteins. The retentate was collected and mixed 1:1 with PAGE sample buffer containing dye (Biorad, Hercules, CA) and kept on ice. 30 uml of each sample was then loaded onto a pre-cast polyacrylamide gel (Biorad, Hercules, CA) and the gels were run at 100 V until dye bands had reached the end of the gel. A molecular weight marker was also loaded on each gel as a known standard (SigmaMarker, Sigma Aldrich, St. Louis, MO). Samples were run in duplicate on 15% and 18% polyacrylamide gels. Differences in proteins were evaluated by Rf values (mm traveled by band/mm traveled by dye front). Flavor Gas chromatography-olfactometry was performed on beer samples using solid phase microextraction (SPME) as the sample preparation method. The fiber used was 100um polydimethylsiloxane coating SPME fiber assembly (Supelco, Bellefonte, PA). Flavor profiles of fresh and processed beer were created by headspace SPME sampling. Fiber exposure time was optimized prior to sampling. Aliquots (7ml) of fresh beer were poured into 40 ml vials and each sealed with caps containing Teflon-coated septa. Volatiles were subsequently extracted using a pre-conditioned 100 m PDMS fiber (Supelco, Bellafonte, PA) under different extraction conditions, which varied in time (5,

PAGE 47

35 10, 15 min) and temperature of fiber exposure (30 and 40C). Equilibrium was reached at 40C for 15 minutes. Chromatographic data is available by contacting Dr. Murat Balaban at the Food Science and Human Nutrition Department, University of Florida. For sampling, the whole coated fiber was exposed to the headspace of the samples and after the extraction conditions were completed, the fiber was removed from the headspace and immediately inserted into a GC-splitless injector, where aroma compounds were allowed to be desorbed for 2 min. Volatile components were separated in a HP-5890 GC (Palo Alto, CA) equipped with a sniffing port (DATU, Geneva, NY), a flame ionization detector (FID), and a ZB-5 column (30 m x 0.32 mm i.d. x 0.5 mm film thickness) from J&W Scientific (Folsom, CA) and in the same model GC with a Carbowax column 30 m x 0.32 mm i.d x 0.5 um film thickness). For both columns the initial oven temperature was 40C which was then increased at 20C/min to a temperature of 120C. The temperature was then increased at 5C/min to a temperature of 160C, and then increased to 240C at 15C/min and held at this final condition for 5 min. Injection and detection port temperatures were 250 and 250C, respectively. A 0.2 L aliquot of alkane standard solution was also injected in the splitless mode. A GC splitter split the column effluent between the FID and the olfactometer in a 1:2 ratio, respectively. Two trained assessors (training based on aroma active compounds present in beer) were employed to evaluate each treatment in duplicate. Each assessor was asked to describe each odor detected in the GC-O effluent and to indicate the aroma intensity continuously during the chromatographic run using a linear potentiometer. This device has a pointer that can be moved across a 10-cm span to indicate aroma intensity. Aroma descriptors along with their respective retention times

PAGE 48

36 were recorded manually, and later transcribed into the chromatographic software for inclusion with the olfactometry time-intensity data. Time-intensity aromagrams were obtained for each treatment. Aroma-active compounds were defined as only those compounds producing an intensity response at the same retention time and similar descriptor from at least half of the panel responses. Mean aroma intensities of each aroma-active compound were calculated by averaging the peak height among each chromatographic run. Chromatograms were recorded, compounds tentatively identified by comparing LRI values and descriptors, and their correspondent peak areas integrated. Beer samples were also evaluated by GC-MS using the same chromatographic conditions as above to further aid in peak identification. A ZB-5 column (60 m x 0.25 mm i.d x 0.25 um film thickness) was used. Mass spectra were matched with several flavor libraries of mass spectra for identification using Xcaliber Software, Version 1.3. Sensory Sensory analysis was done to compare the aroma and flavor profiles of fresh beer to dense-phase CO2 processed beer and dense-phase CO2 processed beer to heat pasteurized beer, and a ranking test was used to rate the likability of different dense-phase CO2 processed beer treatments. A randomized complete block design was used (Ott, 1993) and difference from control measurements were recorded using Compusense on a line scale with anchors at 0 and 10 of no difference and extremely different (Compusense, Guelph, Ontario, Canada). All sensory tests were performed in the taste panel facility in Building 120, University of Florida, Gainesville, Florida which consisted of privacy booths for each panelists. The relative humidity was approximately 60% and the room temperature was between 23-25C. Forty-five to sixty untrained panelists were

PAGE 49

37 used in each test and selected on the basis of age (21+ years or older) and familiarity with beer and willingness to sign the informed consent document. No incentive was given to participate, however snacks were available afterwards to counteract any effects of the alcohol on panelists. Samples were labeled with random numbers and were presented on a white plate with the reference sample at the top center. In the next row the first 3 samples from left to right, and in the bottom row the last 2 samples from left to right. Samples were degassed overnight by placing on a hot plate with stirring to equalize carbonation levels and were served at room temperature. Data was assumed to be normal and an analysis of variance was performed using SAS Software (The SAS Institute, Cary, North Carolina). Tukeys mean separation (alpha=.1) tests were also performed on SAS if there was a significant difference in mean to examine mean separations. Fresh beer vs. dense-phase CO2 processed beer A difference from control test was used with the reference being fresh beer and the samples consisting of a hidden control of fresh beer, and three different dense-phase CO2 processed beer treatments. Panelists were asked to first rate the intensity of the differences in the aroma of the samples and then to rate the intensity of the differences in the flavor of the samples. Dense-phase CO2 processed beer vs. heat pasteurized beer A difference from control test was used with the reference being fresh beer and the samples consisting of a hidden control of fresh beer, and one dense-phase CO2 processed beer treatment (26.7MPa, 21C, 10% CO2, 5 minute residence time), and one heat pasteurized beer sample (74C, 30 seconds). Panelists were asked to first rate the

PAGE 50

38 intensity of the differences in the aroma of the samples and then to rate the intensity of the differences in the flavor of the samples. Storage studies All sensory tests were repeated after 30 days of storage at 1.67C. Statistical Analysis Sensory data were gathered by the Compusense system (Compusense, Guelph, Ontario, Canada), and data were pooled and analyzed using Excel spreadsheets for Windows and SAS software Version 9.0 (The SAS Institute, Cary, North Carolina). An analysis of variance was done on each set of data and a Tukeys mean separation was performed. A 90% confidence level was used. Conjoint Analysis of Beer Purchase Decision A conjoint analysis was performed to determine the part-worth values of price, flavor, and shelf stability of beer in a consumers purchase decision. This was done using a full factorial of combinations created from the characteristics of bottled vs. draft flavor, refrigerated vs. shelf stability, and $6.00/six 12 oz. bottles vs. $8.00/six 12 oz. bottles. These eight combinations were presented to consumers (21 and older) and consumer were asked to rank the combinations in order of purchase choices without ties. Part-worth values for each level of each attribute were then calculated. The higher the part-worth value the more influence the level of a characteristic has on purchase decision (Hair, et al., 1998).

PAGE 51

CHAPTER 4 RESULTS AND DISCUSSION This chapter contains experimental results obtained for the dense-phase CO2 pasteurization of beer, presented in five sections, namely yeast log reduction, foam, haze, sensory evaluation, and gas chromatography results. Microbial Reduction Experiments Yeast plate counts are listed in Table A-1 in Appendix A. Yeast log reductions for control and treated samples are presented in Table 4.1. A prediction equation was calculated using the log reduction responses and response surface methodology, where x1=pressure in MPa, x2=temperature in C, x3=CO2 %, and x4=residence time in minutes (Khuri and Cornell, 1996). SAS software was used using the code in Appendix C. The prediction equation in coded variables is: Predicted Log Reduction = 5.16 .135x1 .80x2 .043x3 + .05x4 .17x12 +.514(x2x1) +1.28(x22) +.012(x3x1) +.734(x3x2) -.971(x32) -.077(x4x1) + .852(x4x2) -.087(x4x3) -.756(x42) And in non-coded variables the prediction equation is: Predicted log reduction =.653 +.553x1 .016x2 .860x3 +2.623x4 -.012x12 +.004(x2x1) +.007(x22) +.0004(x3x1) +.019(x3x2) -.121(x32) -.006(x4x1) + .043(x4x2) -.022(x4x3) -.372 (x42) The ANOVA table used to generate these results is: The RSREG Procedure Coding Coefficients for the Independent Variables Factor Subtracted off Divided by X1 27.550000 9.750000 X2 35.000000 14.000000 39

PAGE 52

40 X3 10.002500 2.827500 X4 4.985000 1.425000 Response Surface for Variable Y Response Mean 4.572981 Root MSE 0.651957 R-Square 0.7307 Coefficient of Variation 14.2567 Type I Sum Regression DF of Squares R-Square F Value Pr > F Linear 4 13.142063 0.2135 7.73 0.0001 Quadratic 4 19.328234 0.3140 11.37 <.0001 Crossproduct 6 12.515434 0.2033 4.91 0.0008 Total Model 14 44.985731 0.7307 7.56 <.0001 Sum of Residual DF Squares Mean Square Total Error 39 16.576879 0.425048 Parameter Estimate Standard from Coded Parameter DF Estimate Error t Value Pr > |t| Data Intercept 1 0.653046 7.574578 0.09 0.9317 5.161059 X1 1 0.552928 0.223678 2.47 0.0179 -0.134616 X2 1 -1.015581 0.144890 -7.01 <.0001 -0.800175 X3 1 1.860443 0.884747 2.10 0.0420 -0.042847 X4 1 2.623033 1.754089 1.50 0.1429 0.050220 X1*X1 1 -0.012258 0.003265 -3.75 0.0006 -1.165235 X2*X1 1 0.003771 0.001670 2.26 0.0296 0.514805 X2*X2 1 0.006516 0.001573 4.14 0.0002 1.277046 X3*X1 1 0.000423 0.008352 0.05 0.9598 0.011674 X3*X2 1 0.018548 0.005763 3.22 0.0026 0.734228 X3*X3 1 -0.121398 0.038833 -3.13 0.0033 -0.970548 X4*X1 1 -0.005533 0.016703 -0.33 0.7422 -0.076869 X4*X2 1 0.042709 0.011525 3.71 0.0007 0.852052 X4*X3 1 -0.021641 0.057625 -0.38 0.7093 -0.087194 X4*X4 1 -0.372491 0.153716 -2.42 0.0201 -0.756389 Sum of Factor DF Squares Mean Square F Value Pr > F X1 5 8.623020 1.724604 4.06 0.0046 X2 5 32.363641 6.472728 15.23 <.0001 X3 5 8.654597 1.730919 4.07 0.0045 X4 5 8.468728 1.693746 3.98 0.0051 Canonical Analysis of Response Surface Based on Coded Data Critical Value Factor Coded Uncoded X1 -0.009502 27.457356 X2 0.241912 38.386764 X3 0.061900 10.177523 X4 0.166366 5.222071

PAGE 53

41 Predicted value at stationary point: 5.067764 Eigenvectors Eigenvalues X1 X2 X3 X4 1.437413 0.093337 0.967866 0.144464 0.183449 -0.776571 -0.159294 -0.097086 -0.391092 0.901247 -1.050380 -0.383442 -0.147191 0.863977 0.291290 -1.225588 0.904923 -0.179288 0.282347 0.263153 Stationary point is a saddle point. Estimated Ridge of Maximum Response for Variable Y Coded Estimated Standard Uncoded Factor Values Radius Response Error X1 X2 X3 X4 0.0 5.161059 0.210737 27.550000 35.000000 10.002500 4.985000 0.1 5.255745 0.210093 27.412807 33.620157 9.977017 4.981320 0.2 5.378149 0.208419 27.299691 32.245753 9.942388 4.965593 0.3 5.528925 0.206472 27.195846 30.877347 9.904840 4.945407 0.4 5.708271 0.205532 27.096467 29.512775 9.866025 4.923116 0.5 5.916267 0.207389 26.999570 28.150655 9.826552 4.899667 0.6 6.152952 0.214171 26.904196 26.790168 9.786696 4.875515 0.7 6.418347 0.227957 26.809822 25.430817 9.746596 4.850902 0.8 6.712462 0.250290 26.716139 24.072285 9.706333 4.825973 0.9 7.035307 0.281894 26.622955 22.714361 9.665955 4.800817 1.0 7.386885 0.322772 26.530142 21.356901 9.625492 4.775491

PAGE 54

Table 4-1: Microbial Reduction Results (Independent Variables are in Coded Units) 42 PressureTempCO2TimeLog Reduction Replicate Log Reduction0.000-1.4140.0000.0006.116.06-1.000-1.000-1.000-1.0006.116.061.000-1.000-1.000-1.0006.116.06-1.000-1.000-1.0001.0006.116.061.000-1.000-1.0001.0004.714.18-1.000-1.0001.000-1.0006.116.061.000-1.0001.000-1.0004.414.18-1.000-1.0001.0001.0004.243.961.000-1.0001.0001.0004.243.960.0000.000-1.4140.0003.193.280.0000.0000.000-1.4143.633.58-1.4140.0000.0000.0003.463.400.0000.0000.0000.0004.714.660.0000.0000.0000.0006.116.060.0000.0000.0000.0006.116.061.4140.0000.0000.0003.933.810.0000.0000.0001.4144.714.360.0000.0001.4140.0004.714.18-1.0001.000-1.000-1.0002.893.211.0001.000-1.000-1.0003.063.38-1.0001.000-1.0001.0004.013.761.0001.000-1.0001.0004.113.81-1.0001.0001.000-1.0003.373.301.0001.0001.000-1.0004.013.81-1.0001.0001.0001.0004.414.361.0001.0001.0001.0004.414.180.0001.4140.0000.0006.116.06

PAGE 55

43 All terms, regardless of significance we re kept as to mirror what independent variables would effect log reduction in a real system, as compared to this model. The overall shape of the response surface was a saddle point with occurred at x1=27.4 MPa, x2= 38 C, x3=10.1% CO2, x4=5.2 minutes residence time, and predicted log reduction of 5.1 logs. Using the prediction equation above, th e predicted maximum log reduction of 7.38 occurred at 26.5 MPa, 21 C, 9.6% CO2, and 4.77 minutes residence time. As stated in literature, log reduction was direc tly proportional to absorption of CO2 (Kumagai et al., 1997). The lowest experimental temperatur e would allow the greatest amount of CO2 to be dissolved in the beer. A lthough higher pressure and higher CO2 % would dissolve more CO2, and longer residence times would allow more equilibration, data shows that saturation is reached at 9.6% CO2, before the highest CO2 % of 12. Higher levels of CO2 only resulted in an excess of CO2. Also an increase of pressure from 26.62 to 37.33 MPa did not result in significant increase in CO2 uptake. Because beer was carbonated before processing the possible saturation of CO2 occurred at lower pressures and CO2 levels than expected. Overall, a predicted log reduc tion of 7.38 makes dense-phase CO2 pasteurization a formidable alternative to h eat pasteurization for the brew ing industry. Currently most brewers try to limit the amount of heat needed to pasteurize by pref iltering the beer to decrease the number of yeast that must be ki lled by heat. This allows a brewer use to a minimal amount of heat during pasteurizati on, thus limiting flavor damage. However the use of dense-phase CO2 not only would allow the elimination of the filtering step prior to pasteurization, but also would prevent flavor damage by heat because the pr ocess is

PAGE 56

44 optimized for 21C. This would simplify the extension of beer shelf life, while insuring no heat damage. Mode of Cell Death Yeast populations remained unchanged after one month of storage at 1.67C, indicating that there may not be an injury/repair mechanism due to the nature of the dense-phase CO2 pasteurization. To elucidate the mode of death, samples were examined using scanning electron microscopy. Images 1, 2, and 3 compare yeast from unprocessed (fresh) beer to that from beer after pasteurization at 27.6 MPa, 10% CO2, 21C, for 5 minutes, and heat pasteurization at 74C for 30 seconds, respectively. Fresh yeast are pert and round, with a smooth appearance. Heat pasteurized yeasts still appeared round and pert with slightly textured surfaces. After dense-phase CO2 pasteurization, some cells show explosive decompression but most have a shrunken appearance with divots in the surface. This illustrates the ability of dense-phase CO2 to affect cell meextraction of their components. mbranes by

PAGE 57

45 Figure 4-1: Scanning Electron Microscopy Picture of Yeast in Fresh Beer Processed at 27.6 MPa, 10% CO2, at 21C, With a Residence Time of 5 Figure 4-2: Scanning Electron Microscopy Picture of Yeast in Beer Dense-Phase CO2 Minutes Figure 4-3: Scanning Electron Microscopy Picture of Yeast in Beer Pasteurized at 74 C for 30 Seconds

PAGE 58

46 Alcohol in beer acts as a co-solvent, aiding in the extraction of hydrophobic and some semi-polar m aterials. It has been reported that the solvent characteristics of CO2 can be greatly modified by the addition cosolvents. For example, the additi 20.7 MPa CO2 processed samples at 101 NTU, heated mple 120.7 NTU, and fresh samples at 146 NTU. Turbidity after one month of storage at 1.67C, showed that all samples increased in haze significantly, however a mean separation was not performed due to the increase in haze was most likely due to microbial growth, not differences in beer protein (Table 4-2). The following ANOVA results were used in the analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 3 4720.916667 1573.638889 1110.80 <.0001 Error 8 11.333333 1.416667 Corrected Total 11 4732.250000 n or existence of certai on of ethanol may have significant effects on extraction by allowing increased absorption on surface sites, preventing the re-adsorption of the compound of interest (Clifford and Williams, 2000). Cosolvents have also been shown to increase the amount of oil extracted from sunflower seeds (Raghuram Rao et al., 1992), and rape seed, soy bean, cottonseed, flax seed, and peanuts, and to enhance the extraction of herbal components such as borage seed oil and hiprose fruit (Illes et al., 1994). Effect of Dense-phase CO2 Processing on Haze Turbidity results after processing showed significant differences between all sample means (p<.0001) on the ANOVA table and Tukeys mean separations (alpha=.1) are shown in Figure 4-4. The 27.6 MPa CO2 processed samples had the lowest turbidityaverage at 95.3 NTU, followed by sa

PAGE 59

47 R-Square Coeff Var Root MSE y Mean Source DF Type I SS Mean Square F Value Pr > F x 3 4720.916667 1573.638889 1110.80 <.0001 Source DF Type III SS Mean Square F Value Pr > F x 3 4720.916667 1573.638889 1110.80 <.0001 Table 4-2: Beer Haze in NTU After Processing and After Storage at 1.67 C for 30 Days 0.997605 1.028283 1.190238 115.7500 After ProcessingAfter StorageDifferenceSampleTurbidity (NTU)Trubidity (NTU)Over TimeFresh146425P-valueFresh1444403.702E-07Fresh14843127.6 MPa95424P-value27.6 MPa954284.198E-0727.6 MPa9644220.720.7100392Heated121440P-valueHeated1204448.506E-08Heated121452 20.7 MPa101404P-value MPa1024001.247E-07 MPa Figure 4-4: Beer Haze Following Processing and Following Storage at 1.67C for 30 Days D

PAGE 60

48 Changes seen in haze due to dense-phase CO2 processing may have been caused bthe drastic pH change associated with this process. During processing the pH drops to approximately 3, which would affect protein conformation and polyphenol conformation,interfering with protein-polyphenol complexes. To further examine the cause of the hdifferences seen between beer samples Polyacrylamide gel electrophoresis of beer proteins was performed. Gels showed no differences in protein bands between fresh, CO2 treated, and heat pasteurized samples. It must be concluded that differences in hazbetween samples after processing cannot be attributed to changes in beer proteinsdense-phase CO2 processing. Gel pictures are available in Appendix A. y aze e during Effect of Ded Stability seSum of Source DF Squares Mean Square F Value Pr > F 544.000000 22.67 0.0003 Error 8 192.000000 24.000000 Corrected Total 11 1824.000000 SE y Mean nse-phase CO2 Processing on Foam Capacity an Foam capacity results after processing showed differences between fresh, denphase CO2 processed, and heated beer sample means (p=.0003). Tukeys mean separation (alpha=.1) showed that the CO2 processed samples were not significantly different from each other, however they did differ from heated samples. Fresh and heated samples were also found to be significantly different. The following ANOVA results were used in the analysis for foam capacity and foam stability, respectfully: Foam Formation: Model 3 1632.000000 R-Square Coeff Var Root M

PAGE 61

49 0.894737 1.530931 4.898979 320.0000 Source DF Type I SS Mean Square F Value Pr > F Source DF Type III SS Mean Square F Value Pr > F oam Stability: Source DF Squares Mean Square F Value Pr > F Model 3 164.0000000 54.6666667 41.00 <.0001 Error 8 10.6666667 1.3333333 Corrected Total 11 174.6666667 R-Square Coeff Var Root MSE y Mean 0.938931 2.074312 1.154701 55.66667 Source DF Type I SS Mean Square F Value Pr > F x 3 164.0000000 54.6666667 41.00 <.0001 Source DF Type III SS Mean Square F Value Pr > F x 3 164.0000000 54.6666667 41.00 <.0001 Mean separations are shown on Figure 4-5. Heated samples had the highest foam capacity at 333%, followed by 27.6 MPa CO2 processed beer with 324%, 20.7 MPa CO2 processed beer with 321%, and fresh beer with 301% foam capacity (Figure 4-5). After 30 days of storage at 1.67C, sample means again showed significant differences (p=.0006). Heat and fresh samples were not significantly different and had the highest both with foam capacities of 307% (Figure4-6). x 3 1632.000000 544.000000 22.67 0.0003 x 3 1632.000000 544.000000 22.67 0.0003 F Sum of foam capacity with 332% and 327%, respecitively. CO2 processed samples followed,

PAGE 62

50 Figure 4-5: Foam Capacity and Stability of Beer Samples After Processing b b a a b a Figure 4-6: Foam Capacity and Stability of Beer Samples After Storage at 1. 67C for 30 Days Foam stability results after processing showed differences between fresh, dense-phase CO2 processed, and heated beer sample means (p<.0001) (Figure 4-5). Tukeys Foam Formation, Aged: Source DF Squares Mean Square F Value Pr > F mean separation (alpha=.1) showed that the CO2 processed samples were not significantly different from each other, however did differ from fresh and heated samples. The following ANOVA results were used in the analysis: Sum of

PAGE 63

51 Model 3 1584.000000 528.000000 18.86 0.0006 DF Type I SS Mean Square F Value Pr > F x 3 1584.000000 528.000000 18.86 0.0006 18.86 0.0006 Source DF Squares Mean Square F Value Pr > F Error 8 224.000000 28.000000 Corrected Total 11 1808.000000 R-Square Coeff Var Root MSE y Mean 0.876106 1.663995 5.291503 318.0000 Source Source DF Type III SS Mean Square F Value Pr > F x 3 1584.000000 528.000000 Foam Stability, Aged: Sum of Model 3 334.6666667 111.5555556 9.30 0.0055 Error 8 96.0000000 12.0000000 Corrected Total 11 430.6666667 0.777090 7.585624 3.464102 45.66667 Source DF Type I SS Mean Square F Value Pr > F x 3 334.6666667 111.5555556 9.30 0.0055 Source DF Type III SS Mean Square F Value Pr > F x 3 334.6666667 111.5555556 9.30 0.0055 2 R-Square Coeff Var Root MSE y Mean Fresh and heated sample means were not significantly different. Heated samples had the highest foam stability at 60%, followed by fresh beer with 59%, and CO

PAGE 64

52 processed samples with 52% foam stability for both CO2 treatments. After 30 days of storage at 1.67C, sample means again showed significant differences (p<.0055). Heated samples were significantly different from all other treatments with 55% foam stability. Fresh and CO2 processed samples were not significantly different with fresh beer foam stability at 44%, and 27.6 MPa and 20.7 MPa at 43% and 41%, respectively (Figure 4-6). Changes seen in foam characteristics due to dense-phase CO2 processing may have been caused by the extraction of cell membrane or cell wall parts that may have changed ing. In both foam capacity and foam stability, data showed that both CO2 processing and heat pasteurization may have significant effects, however, all beers formed a stable head at a level that would probably insure customer satisfaction. Original foam volumes and liquid volumes are listed in Table A-2. To further examine the cause of the foam differences seen between beer sample means Polyacrylamide gel electrophoresis of beer proteins was performed. Gels showed no differences in protein bands between fresh, CO2 treated, and heat pasteurized samples. attributed to changes in beer proteins during processing. Gel pictures are available in Appendix A. In the cases of both beer haze and foam, dense-phase CO2 pasteurization of beer in no way decreased the quality of the finished product. Consumers would be expected to see no differences between fresh and CO2 pasteurized beer. the amount of hydrophobic compounds in the beer, therefore, affecting foam It must be concluded that differences in foam between samples after processing cannot be

PAGE 65

53 Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Sensory Difference from control panels and gas chromatography-olfactometry were erformed to prove there were no aroma or flavor changes due to dense-phase CO2 asteurization. Panels were conducted within 1 day of processing and repeated after 30 ays of storage at 1.67C. All samples were decarbonated and served at room mperature. For sensory panels, analysis of variance (=.05) was used to compare eatment means and, when differences existed, a Tukeys mean separation test was Science and Human Nutrition Department, University of Florida. Beer % CO2, for 5 minutes, at 21C. Samples were evaluated on a line scale with 0=no ifference from the fresh beer reference and 10=extremely different from the fresh beer ference. The following ANOVA results were created and used for analysis: Sum of Model 63 643.348046 10.211874 1.85 0.0022 Error 116 641.861454 5.533288 Corrected Total 179 1285.209500 R-Square Coeff Var Root MSE y Mean 0.500578 67.56231 2.352294 3.481667 Evaluation p p d te tr employed to compare specific treatment means at =.10. A sample ballot is included in Appendix B and raw data is available by contacting Dr. Murat Balaban in the Food Aroma A difference from control test comparing aromas of fresh beer, beer processed at 27.6 MPa with 10% CO2, for 5 minutes, at 21C, and beer processed at 20.7 MPa with 10 d re Source DF Squares Mean Square F Value Pr > F

PAGE 66

54 row 59 573.3695000 9.7181271 1.76 0.0051 column 2 2.9962126 1.4981063 0.27 0.763 Source DF Type III SS Mean Square F Value Pr > row 59 573.5187454 9.7206567 1.76 0.0051 Source DF Type I SS Mean Square F Value Pr > F treatment 2 66.9823333 33.4911667 6.05 0.0032 3 F treatment 2 66.9877883 33.4938942 6.05 0.0032 Sample means were significantly different (p=.0032), with average scores of 2.72, 3.52, and 4.21, respectively on the ten point scale. All samples were rated as only slightly different from the reference and similarities were seen between fresh and the 27.6 MPa beer, and between the 2 COprocessed sample means. Mean separations are shown using letters on Figure 4-7. column 2 2.9962126 1.4981063 0.27 0.7633 2 Figure 4-7: Aroma Evaluation of Fresh and COProcessed Beer Samples (mean C, 30 seconds). Samples were again evaluated on a line scale with 0=no difference from the fresh beer reference 2 separations labeled) A difference from control test was then conducted to compare the aroma of fresh, 27.6 MPa CO2 processed beer, and heat pasteurized beer (74 A AB B a b

PAGE 67

55 andlts were created and used for analysis: Sum of r > F Error 86 484.8647953 5.6379627 Corrected Total 134 961.1760000 R-Square Coeff Var Root MSE y Mean Source DF Type I SS Mean Square F Value Pr > F .0108 treatment 2 25.2840000 12.6420000 2.24 0.1124 90 0.61 0.5463 F 5 treatment 2 26.4211450 13.2105725 2.34 0.1021 column 2 6.8645381 3.4322690 0.61 0.5463 2.75, 3.29, and 3.81 and are shown in Figure 4-8. 10=extremely different from the fresh beer reference. The following ANOVA resu Source DF Squares Mean Square F Value P Model 48 476.3112047 9.9231501 1.76 0.0113 0.495550 72.39145 2.374439 3.280000 row 44 444.1626667 10.0946061 1.79 0 column 2 6.8645381 3.43226 Source DF Type III SS Mean Square F Value Pr > row 44 445.3035982 10.1205363 1.80 0.010 There were no significant differences between sample means(p=.1021). Average scores were

PAGE 68

56 A Samples Beer Flavor A difference from contr Figure 4-8: Evaluation of Aromas of Fresh, CO2 Processed and Heat Pasteurized Beer ol test comparing flavor of fresh beer, beer processed at 27.6 MPa with 10% CO2, for 5 minutes, at 21C, and beer processed at 20.7 MPa with 10% CO2, for 5 minutes, at 21C. Samples were evaluated on a line scale with 0=no difference from the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 63 681.309167 10.814431 2.02 0.0005 Error 116 619.517278 5.340666 R-Square Coeff Var Root MSE y Mean 0.523751 60.09504 2.310988 3.845556 Source DF Type I SS Mean Square F Value Pr > F row 59 666.5731111 11.2978493 2.12 treatment 2 11.5821111 5.7910556 1.08 column 2 39444 1.5722 0.30 0.7449 Corrected Total 179 1300.826444 0.0003 0.3415 3.15 697

PAGE 69

57 row 59 Source DF Type III SS Mean Square F Value Pr > F 666.9011463 11.3034093 2.12 0.0003 treatment 2 11.5196545 5.7598272 1.08 0.3435 column 2 3.1539444 1.5769722 0.30 0.7449 Sample means were not significantly different (p=.3435), with average scores of 3.63, 3.71, and 4.2, respectively on the ten point scale. Scores are shown on Figure 4-9. Figure 4-9: Evaluation of Beer Flavor Between Fresh and CO2 Processed Samples A difference from control test was then conducted to compare the flavor of fresh, 27.6 M the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 48 658.515137 13.719065 2.51 <.0001 Error 86 469.693011 5.461547 Corrected Total 134 1128.208148 R-Square Coeff Var Root MSE y Mean Pa CO2 processed beer, and heat pasteurized beer (74C, 30 seconds). Samples were again evaluated on a line scale with 0=no difference from

PAGE 70

58 0.583682 55.69185 2.336995 4.196296 Source DF Type I SS Mean Square F Value Pr > F row 44 571.6481481 12.9920034 2.38 0.0003 treatment 2 76.3779259 38.1889630 6.99 0.0015 column 2 10.4890631 5.2445316 0.96 0.3869 Source DF Type III SS Mean Square F Value Pr > F row 574.7377992 13.0622227 2.39 0.0003 treatment 2 78.8786056 39.4393028 7.22 0.0013 column 2 10.4890631 5.2445316 0.96 0.3869 Sample means were significantly different (p=.0013), with average scores of 3.66, 3.67, different from the reference and similarities were seen between fresh and the 27.6 MPa beer, and the heated sample was seen as significantly different from the others. Mean separations are shown using letters on Figure 4-10. 44 and 5.26, respectively on the ten point scale. All samples were rated only slightly Figure 4-10: Evaluation of Beer Flavors Between Fresh, CO Processed, and Heat 2 Pasteurized Samples (mean separations labeled) A A B

PAGE 71

59 Beer Aroma After Storage A storage study was conducted at 1.67 C for 30 days and beer aroma and taste tests wproce7 2stored, fresh hidden control. Samples were evaluated on a line scale with 0=no difference from the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 56 961.287287 17.165844 3.77 <.0001 Error 155 706.211393 4.556203 Corrected Total 211 1667.498679 R-Square Coeff Var Root MSE Sample_Aroma Mean F 1 Samp_ 3 24.9533962 8.3177987 1.83 0.1448 1 n Square F Value Pr > F 3.95 <.0001 1.86 0.1383 Samp_Code 1 1.0002111 1.0002111 0.22 0.6401 ere repeated. A difference from control test comparing aromas of fresh beer, beer ssed at 27.6 MPa with 10% CO2, for 5 minutes, at 21C, and beer processed at 20.MPa with 10% CO, for 5 minutes, after 30 days of storage at 1.67C at 21C, and a non0.576485 74.35419 2.134526 2.870755 Source DF Type I SS Mean Square F Value Pr > Samp_Set 52 935.3336792 17.9871861 3.95 <.000 Samp_Code 1 1.0002111 1.0002111 0.22 0.640 Source DF Type III SS Mea Samp_Set 52 935.3336792 17.9871861 Samp_ 3 25.4545507 8.4848502 There were no significant differences seen between sample means and averages scores of 2.35, 3.3, 2.99, and 2.84 are shown on Figure 4-11.

PAGE 72

60 Figure 4-11: Aroma Evaluation of Aroma Between Aged, CO2 Processed, Heat Pasteurized, and a Non-Stored, Fresh Hidden Control 27.6 MPa CO2 processed beer, heat pasteurized beer (74C, 30 seconds) after 30 days of storage at 1.67C, and a non-aged fresh hidden control. Samples were again evaluated on a line scale with 0=no difference from the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 50 825.213462 16.504269 3.03 <.0001 Error 129 703.666538 5.454779 Corrected Total 179 1528.880000 R-Square Coeff Var Root MSE taste Mean 0.539750 58.38867 2.335547 4.000000 r > F set 44 533.1900000 12.1179545 2.22 0.0003 number 3 290.1186667 96.7062222 17.73 <.0001 code 3 1.9047954 0.6349318 0.12 0.9504 A difference from control test was then conducted to compare the aroma of aged, Source DF Type I SS Mean Square F Value P

PAGE 73

61 Source DF Type III SS Mean Square F Value Pr > F 003 number 3 290.0674621 96.6891540 17.73 <.0001 set 44 533.1900000 12.1179545 2.22 0.0 code 3 1.9047954 0.6349318 0.12 0.9504 Sample means were significantly different (p<.0001), with average scores of 2.56, 3.34, 4.87, and 1.99, respectively on the ten point scale. Mean separations are shown using letters on Figure 4-12. Figure 4-12: Evaluation of Flavor Between Aged, CO2 Processed, Heat Pasteurized, and a Non-Stored, Fresh Hidden Control (mean separations labeled) 27.6 MPa with 10% CO2, for 5 minutes, at 21C, and beer processed at 20.7 MPa with 10% Cith 0=no difference from Beer Flavor After Storage A difference from control test comparing the flavor of aged beer, beer processed at O2, for 5 minutes, after 30 days of storage at 1.67C at 21C, and a non-stored, fresh hidden control. Samples were evaluated on a line scale w A A A B A

PAGE 74

62 the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Model 56 798.921305 14.266452 2.92 <.0001 Error 155 756.016242 4.877524 Corrected Total 211 1554.937547 R-Square Coeff Var Root MSE Taste_Difference Mean 0.513796 61.58965 2.208512 3.585849 Source DF Type I SS Mean Square F Value Pr > F Samp_ 3 13.3967925 4.4655975 0.92 0.4349 <.0001 Samp_ 3 13.7483426 4.5827809 0.94 0.4231 0.8419653 0.8419653 0.17 0.6784 age scores of Samp_Set 52 784.6825472 15.0900490 3.09 <.0001 Samp_Code 1 0.8419653 0.8419653 0.17 0.6784 Source DF Type III SS Mean Square F Value Pr > F Samp_Set 52 784.6825472 15.0900490 3.09 Samp_Code 1 Sample means were not significantly different (p=.4231), with aver 3.58, 3.91, 3.65, and 3.20, respectively on the ten point scale. Scores are shown on Figure 4-13.

PAGE 75

63 Figure 4-13: Evaluation of Flavor Between Fresh and CO2 Processed Beer After Storage at 1.67C for 30 Days, and a Non-Stored, Fresh Hidden Control A difference from control test was then conducted to compare the flavor of aged, 27.6 MPa CO2 processed beer, heat pasteurized beer (74C, 30 seconds) after 30 days of storage at 1.67C, and a non-aged fresh hidden control. Samples were again evaluated on a line scale with 0=no difference from the fresh beer reference and 10=extremely different from the fresh beer reference. The following ANOVA results were created and used for analysis: Sum of Source DF Squares Mean Square F Value Pr > F Error 131 648.607913 4.951205 Corrected Total 179 1441.913111 R-Square Coeff Var Root MSE Sample_Aroma Mean 0.550175 69.80193 2.225130 3.187778 Source DF Type I SS Mean Square F Value Pr > F Samp_Set 44 581.5481111 13.2170025 2.67 <.0001 Samp_ 3 211.7388889 70.5796296 14.26 <.0001 Samp_Code 1 0.0181980 0.0181980 0.00 0.9517 Model 48 793.305198 16.527192 3.34 <.0001

PAGE 76

64 Samp_Set 44 581.5481111 13.21 Source DF Type III SS Mean Square F Value Pr > F 70025 2.67 <.0001 Samp_ 3 211.6859758 70.5619919 14.25 <.0001 0.0181980 0.00 0.9517 Sample means were significantly different (p<.0001), with average scores of 2.92, .71, 6.14, and 3.23, respectively on the ten point scale. Mean separations are shown sing letters on Figure 4-14. Samp_Code 1 0.0181980 3 u FigurSamples After 30 Days at 1.67C, and a Non-Stored, Fresh, Hidden Control ing all decarbonated beer samples, overall, panelists could easily distinguish heat pasteurized beer, but in only on one occasion was a difference noted between fresh and the 20.7 MPa CO2 processed beer. This difference in aroma may have been caused by the increased spillage of cell components, caused by CO2 processing at 20.7MPa. These cell components have been known to contribute to off-flavors in beer (Lin et al., 1991). No differences were noted in flavor between the same samples and in following less-volatile compounds, many of which could be hop constituents, could have masked any differences in aroma during beer tastings. On average, panelists could only e 4-14: Evaluation of Flavor of Fresh, CO2 Processed, and Heat Pasteurized (mean separations labeled) Consider A A A A B

PAGE 77

65 discern heat pasteurized beer during sensory tests. The flavor differences caused by dense-phase CO2 pasteurization were negligible. In an industry with its foundations in producing a consistently fresh product, dense-phase CO2 pasteurization is an alternative to heat pasteurization that results in the same extended shelf life, while preserving fresh beer taste. Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Gas Chromatography-Olfactometry and Mass Spectrometry Odor descriptors and retention times of aroma active compounds present in beer the ason the ZB-5 column were used to characterize beer samples. was ined from the ZB-5 and Carbowax columns and their spectroscopy. the sample, however was not etected previously by GC-O. Mass spectra library matches, chromatograms, and d Science and egration d using LRI alues and confirmed using mass spectra (Table 4-6). When comparing values of samples can be observed in Table 4-4. A variety of aroma compounds were detected by sessors, however only those detected more than 50% of the time by both assessors Linear retention indices (LRI) were calculated for aroma active compounds. Tentative identification of the aroma compounds present in fresh beer (Table 4-5) conducted based on their LRI values obta aroma descriptors. Compounds in red were then confirmed by mass Methyl nerolate was also identified by mass spectra in d aromagrams are available by contacting Dr. Murat Balaban in the Foo Human Nutrition Department, University of Florida. To evaluate if dense-phase CO2 processing did cause flavor changes, int areas (reported in mV), were compared for compounds that had been identifie v compounds before and after dense-phase CO2 pasteurization, many compounds had negligible differences, however in the case of ethyl hexanoate, the amount of the

PAGE 78

66 co mpound decreased on average by approximately 49%. Because of ethyl hexanoates processing. Table 4-4: Retention Tis and Aroma Dcriptors for Compnds Detected byoth Although stripping did occur, sensory panelists did not recognize differences between the 26.7MPa dense-phase CO2 and fresh beer. This is possible because even though ethyl hexanoate did decrease in CO2 processed beer, if it remained above its flavor threshold, no flavor change would be detectable to panelists. Overall, it can be concluded that although some stripping did occur, no appreciable changes were detected between fresh and CO2 processed beer samples. volatile nature, not only did this compound elute early in the chromatography run (4.26 minutes on Carbowax column), but was also easily stripped by the carbon dioxide during me es ou B Assessors on ZB-5 and Carbowax Columns (desciptors are listed for the app Retention Times (minutes)Descriptor on ZB-5Descriptor on Carbowaxropriate column only)2.82fruity3.21banana, bubble gum3.71fruit4.17Fruity4.26musty, fruity4.92Fruity/sweet5.58fruity, sour5.96fruity, wine6.18Fruity6.5Fruity7.3pepper8.7green apple/sweet9.6sweet, cooked/coconutty11.2green apple/sweet11.9apple pie13.1slight red burnt juice15.15burnt apple/sweet13.61apple pie14.81

PAGE 79

67Table 4-5: Linear Retention Iesd Itiftiof Cpounds using GC-O (Comnds in Red Wer Confired by Mass Stropy em)1051) ndic andenican oom Compounds on -5RI ofomundsn C108113119769125879141145020156128194269184365396 pouwaTenive eth2-mhylbhexenal Z (11hexanal E (12eth2-mhylpethhexoatehexenal Z (85hexol, 141ethl heptoatcare (eth3-hyoxyethoctteethl phenylacedamscdodena140dodano aci pecsco) eteion TeLRI 22313147464258561291111 RntimofZBL Cpo oarboxtatIdentificatioRI.81yl etutano (.25-345).77-228).1yl etop (751).28yl an (.9-33).56an2 (7).92yane 6.51en1011)7.30linalool (1552)8.71yl drhee (1134)9.61yl anoa (11.1yta11.9aenone (183.1ecl (7)3.61ecicd ( n (Lateanoate1242)(1346)xanoat195)te27)1568)

PAGE 80

68 Retention index values are useful in compound identification when LRI values arecompared across different columns and/or different chromatographic conditions. Severalfactors affect the retention time of an analyte in a chromatographic run such as column length, carrier gas flow, film thickness of the column, column packaging, and compound, however, do not affect their retention indices, because these are relative values, calculated from alkane standards run at the same chromatographic conditions. temperature program. All these factors significantly affect the retention time of a

PAGE 81

Table 4-6: Average Integration Areas of Identified Compounds in Fresh and CO2 Processed Beer 69Identified Compounds Average Fresh Integration Areas Average CO2 Integration Areas Fresh-CO2 Integration Areas % change ethyl hexanoate 150837744733948.66 ethyl heptanoate 53490353276821360.40 ethyl octanoate 274993254857201367.32 ethyl phenylacetate 1766819087-1419-8.03 dodecenal 28457305302073-7.28 dodecanoic acid 139467144960-5493-3.94

PAGE 82

70 When comparing the relationship between the GC-FID chromatogram (Figure 4-15) and the aromagram peak intensities for each compound (Figure 4-16), one can observe that aroma active components that may drive the aroma profile either did not have a response at all (peaks with RT= 4.17, 4.92, 6.18 min) or had a relativelyresponse (peaks with RT= 6.53, 9.62, 13.3, 13.61) in the FID detector. The latter observation likely occurred due to the presence of these compounds in lower quantitito their poor response and specific interactions with the FID detector, or have lower low es, threshold values. Moreover, compounds that have relatively high peak intensities to the FID detector can be odorless compounds that do not contribute to the typical key notes of the sample and thus they can not be detected by the assessor while conducting the GC-O sniffing. However, an assessor can also not identify a specific aroma compound due to selective anosmia or due to the low sensory threshold of the specific compound. Figure 4-15: Typical FID Chromatogram for Fresh Beer Figure 4-16: Typical Aromagram for Fresh Beer

PAGE 83

71 The data produced by GC-O has a qualita tive component in which the assessor describe perception. This usually involves association of the precept with a word or group of words in a le xicon. Qualitative GC-O data are either measureent of odor potency or perceived in tensity plotted against a retention index (Fried and Acree, 2000). The major advantag e of using GC-O is the ability to detect the prce of aroma impact co mponents. One does not need to know ahead of timh components to measure. Usi ng this technique many highly potent aroma etected that would ha ve been missed simply because they were ch minute concentrations. Howe ver, GC-O has limitati ons since it can not gonistic interactions from ot her aroma active components in 2000). reduce experimental error and varia tion, several factors can be monitored and samp le temperature, time of day, duration of repeated standardization of sniffers, and use of a standard at people can be trained to co nsistently iden tify smells if e ically and trained to sniff with standard chemicals and larieprecis of threshold detec tion of an individual makes it more difficult odor experience than the beginning of the e variation to intensity ratings (Rouseff and Naim, 2000; although some stripping did occur, no etween fresh and CO2 processed beer samples. This the sensor y panels, can be used to support the case s th e n atu re o f th eir m rich esen e whic ine synergistic or anta un exp ec ted impact compounds are d present at su determ the sam including sample preparation, room analysis, repetition of analysis, lexicon. It is well documented th they voc for the assessor to detect the end of an experience, lending som Friedrich and Acree, 2000). appreciable changes were detected b finding, along with the results from p To le (Rouseff and Naim ar abu stan da s. T rdiz he ed pe riod ion Overall, it can be concluded that

PAGE 84

72 for dening in is ion Department, University of Florida. The attribute eliciting the largest influence on purchase decision was price, which was 53.9% of a consumers purchase decision. Overall, consumers preferred paying $6.00/six 12 ounce bottles (average rank=3.705), rather than $8.00/6 12 ounce bottles (average rank=5.279). The next most influential attribute was beer flavor, being 34% of or with an average rank of 4.003, compared to bottled beer flavor with an average rank of 4.995. Finally, beer shelf life showed the least effect on beer purchase decision, being 12.1% of the purchase decision. Consumers preferred shelf stable beer with an average rank of 4.321 over beer that required refrigeration that had an average rank of 4.674. This conjoint analysis gives insight into characteristics of the beer that influence which beer a consumer will purchase. This information is useful when introducing a new se-phase CO2 as an alternative to heat pasteurization of beer. Furthermore, because of beers complex mix of flavors, the minimal stripping that does occur durdense-phase CO2 pasteurization is easily masked by other flavor compounds, resultinga final product no different from fresh beer. Conjoint Analysis of Beer Purchase Decisions A conjoint analysis is used to quantify how important a specific characteristic is ofa product in a consumers purchase decision (Table 4-7). A conjoint analysis of beerpurchase decisions was conducted to elucidate how flavor, shelf life requirements, and price affect purchase decision. Two levels of each attribute were tested, creating a full 23 factorial. Panelists were asked to rank 8 descriptions of beers in the order in which they would purchase the beers. A sample ballot is available in Appendix B and ranking dataavailable by contacting Dr. Murat Balaban in the Food Science and Human Nutrit the purchase decision. Consumers preferred draft beer flav

PAGE 85

73 producr its t to an existing market, as in the case of introducing dense-phase pasteurized beeto the current beer market. Using the data from the conjoint analysis, dense-phase CO2 pasteurized beer would successful because of its draft beer taste, coupled withextended shelf life. However, if the cost of using dense-phase CO2 increased the price ofthe beer significantly, consumer may opt not to purchase the product. Having seen that price is 53.9% of the purchase decision, it would be in the best interest to minimize or prevent a price increase by lowering the cost of production by recycling CO2 during processing and eliminating cold storage during transit for the product.

PAGE 86

74Table 4-7: Conjoint Analysis Transformation FaFlaDraBotdStogeRefgerSh StaPri$$ ctor LevAverage Rank of LeDeviation from Overall Average nkReversed iationSquared vianStandardized eioEstimated Part-rtRange of Part-WorthFactor porcevorft03-07972428005.0tle9505-9524-19-05raridate21-09790307075.1elfble7404-7403-01-08ce6.05-05956314149.98.7909-7960-31-15 sImtan1.8134%0.6412%2.8753% DetioDviatnWoh0.70.8.910.50.8.900.20.1.320.00.1.310.22.1.450.72.0.42 elvelRaDev4.0.490.44.9.490.4d4.3.170.14.6.170.1003.7.790.7005.2.770.7

PAGE 87

CHAPTER 5 g in a 5-log or higher decrease in yeast populations. 2or essing and after 30 days of storage at 1.67C. Beer haze was significantly reduced by dense-phase CO2 pasteurization. Beer foam capacity and stability were affected by dense-phase CO2 pasteurization, however not to the detriment of the finished products quality. Beer purchase decision is most affected by price, then beer flavor, and finally shelf life requirements. Consumers prefer shelf stable, draft beer and are adverse to an increase in price. In general, a continuous dense-phase CO2 system was effective in the pasteurization of beer. The success of this system in the brewing industry would rely on its predicted 7.38 log reduction in yeast populations, while preserving fresh beer aroma, flavor, foaming capacity and stability, and aiding in the reduction of beer haze. By resulting in an extended shelf life dense-phase CO pasteurization is a formidable alternative to heat pasteurization, and would be preferred to heat pasteurization because of its ability of maintain fresh beer characteristics. The ability to produce a clear, consistently fresh tasting beer that forms a good head, with an extended shelf life is top-priority to brewers. Dense-phase CO2 pasteurization can make this possible. In addition, it is predicted that consumer acceptability of this technology in the brewing industry would be high based on consumers priorities when it comes to beer CONCLUSION The conclusions of this study are the following: Dense-phase CO2 pasteurization is effective in the pasteurization of beer, resultinAlthough some stripping did occur during dense-phase CO pasteurization, flavand aroma changes detected by panelists were negligible after proc 75

PAGE 88

76 characteristics. Consumers prefer a beer that does not have to be refrigerated and that has draft beer taste; however they would be adverse to an increase in beer prices. If brewers concentrated on cost-effective processi ng techniques, such as recycling CO2 throughout the brewery and elimination of the co ld storage of product, dense-phase CO2 pasteurization could be used to create a pr oduct with extended shelf life and draft beer taste with no increase in price. Indications of the mode of ce ll death were absorption of CO2 into the cell membrane creating a physical disruption in membrane structure, visible as divots in scanning electron microscopy pictures. It is recommended that more research be conducted to elucid ate the mode of cell death and the role of alcohol as a co-solvent in yeast death by dense-phase CO2 pasteurization Further research in the area of beer flavor stripping would also lend to the application of this technology to beve rages in general. Recycling of the CO2 stream and recovery of stripped volatiles would al so be valuable research topics.

PAGE 89

APPENDIX A RAW ATA EXPERIMENT D

PAGE 90

78 Table A-1: Yeast Counts for 27 Treatment Combinationon Dic s De inuplate BABve ToCol000000000000000000000000593616.700.50.00.50.000000000.2422211.210.50.20000000000000000000000000000000000000000 TreatmentDilutionAAAve ABve AtalonimaLedu log)110000006.1200000300000400000500000600000210000006.12000003000004000005000006000003135227.56752.82100.5153100.515400050006010.5000541010.5251253.020000530cont04000500060005100006.120003000400050006000 ction(no/n106106866594106 es Reiningog R.00.001.001.00.00

PAGE 91

79Table A-1: Continued TreatmentDilutionAAAve ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)610000210.550.004.4116200000003000000040000000500000006000000071454.5586.55.5550.003.37022111100.50.75300000004000000050cont000006000000081111121.51.25125.004.01200000003000000040000000500000006010.50000.259100000000.006.112000000030000000400000005100.50000.2560000000101000010.50.2525.004.712000000030000000400000005000000060000000111502.50001.25125.004.013720000000300000004100.50000.255100.50000.256010.50000.251210002221100.004.110620000000300000004000010.50.255000000060000000 370626

PAGE 92

80Table A-1: Continued TreatmentDilutionotalColonies RemainingLog Reduction log(no/n)00.006.11062010.50000.25300000004000000050000000600000002118cont85128.58.25825.003.19412000100.50.253000010.50.25400000005010.50000.2560000000221000010.50.2525.004.71262010.50000.2530000000400000005000000060000000231142.5433.53300.003.63352201100.50.7530000000400000005000100.50.2560000000241010.50000.2525.004.71262000000030000000400000005000000060000000251100.50000.2525.004.712620000000300000004000000050002010.560000000261contcont#DIV/0!000#DIV/0!0.006.11062contcont#DIV/0!000#DIV/0!30cont00000400000005000000060000000 AAAve ABBAve BAve T201000000

PAGE 93

Table A-1: Continued TreatmentDilutionAAAve ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)20100000000.006.11062010.50000.25300000004000000050000000600000002118cont85128.58.25825.003.19412000100.50.253000010.50.25400000005010.50000.2560000000221000010.50.2525.004.71262010.50000.2530000000400000005000000060000000231142.5433.53300.003.63352201100.50.7530000000400000005000100.50.2560000000241010.50000.2525.004.71262000000030000000400000005000000060000000251100.50000.2525.004.712620000000300000004000000050002010.560000000261contcont#DIV/0!000#DIV/0!0.006.11062contcont#DIV/0!000#DIV/0!30cont00000400000005000000060000000 81

PAGE 94

Table A-1: Continued Treatmen t Dilution A A Ave ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)27100000000.006.11062000121. 5 0.7 5 3000100. 5 0.2 5 40000000 5 010. 5 100. 5 0. 5 60000000 82 TreatmentDilutionAAAve ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)1100000000.006.056920000000300000004000000050000000600000002100000000.006.05692000000030000000400000005000000060000000315769787700.003.211821110000.53010.5010.50.5400000005000000060000000411115128.54.75475.003.38022010.5100.50.530000000400000005000000060000000

PAGE 95

Table A-1: Continued 83 Treatmen t Dilution A A Ave ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)5100000000.006.0569200000003000000040000000500000006000000061021010.50.7575.004.1818200000003000000040000000500000006000000071676.54655.75575.003.297221111111300000004000000050con t 000006000000081211.52221.75175.003.8139200000003000000040000000500000006010.50000.259100000000.006.05692000000030000000400000005100.50000.2560000000

PAGE 96

Table A-1: Continued TreatmentDilutionAAAve ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)101111010.50.7575.004.18182000000030000000400000005000000060000000111343.5100.52200.003.7559200000003000000040210000.55100.50000.2560000000121121.52221.75175.003.81392000000030000000400000005000000060000000131111211.51.25125.003.96002000000030000000400000005000000060000000141111211.51.25125.003.96002000000030000000400000005000000060000000151100.5010.50.550.004.35792000000030000000400000005000000060000000 84

PAGE 97

Table A-1: Continued Treatmen t Dilution A A Ave A BBAve BAve TotalColonies RemainingLog Reduction log(no/n)161111010.50.7575.004.181820000000300000004000000050000000600000001712436664.5450.003.40372100.5010.50.530000000400000005000000060000000181021232.51.75175.003.81392000010.50.253000000040000000500000006000000019100000000.006.0569200000003000000040000000500000006000000020100000000.006.05692010.50000.2530000000400000005000000060000000211877.5454.56600.003.27882000100.50.2530000000400000005000000060000000 85

PAGE 98

Table A-1: Continued TreatmentDilutionAAAve ABBAve BAve TotalColonies RemainingLog Reduction log(no/n)221111010.50.7575.004.18182010.50000.25300000004000000050000000600000002312223543300.003.579820000000300000004000000050000000600000002411110000.550.004.35792000000030000000400000005000000060000000251010.50000.2525.004.659020000000300000004000000050002010.56000000026100000000.006.0569200000003000000040000000500000006000000027100000000.006.05692000000030000000400000005000000060000000 86

PAGE 99

87 Table A-2: Foam and Liquid Volumes After ProcessingAfter StorageTreatmentfoam volumeliquid volumefoam volumeliquid volumeFresh75108114Fresh76108114Fresh75118314Fresh Average75.3333333310.3333333381.666666671426.7 MPa8012781426.7 MPa8112761426.7 MPa8212761526.7 MPa Average811276.6666666714.3333333320.7 MPa7812781420.7 MPa8212761620.7 MPa8112761420.7 MPa Average80.333333331276.6666666714.66666667Heated84108112Heated83108410Heated83108412108311.33333333 Bottom Gel: 18% cross-linked Lanes from Left to Right: Same as Above Heated Average83.33333333 lamide Gels Figure A-1: Polyacry Top Gel: 15% cross-linked Lanes from Left to Right: 1. Heated Beer 2. 20.7 MPa 3. 27.6 MPa 4. Fresh Beer 5. Heated Beer 6. 20.7 MPa 7. 27.6 MPa 8. Fresh Beer 9. Molecular Weight Marker

PAGE 100

88 APPENDIX B SENSORY AND CONJOINT BALLOTS Figure B-1: Sa m ple Sensory Ballot

PAGE 101

89 Beer Pu rchase Survey Read each of the descriptions below, then rank them in the order in which you would purchase the beers. For instance, the beer you would most likely purchase would be ranked #1 and the beer you are least likely to purchase would be #8. There can be no ties. Beer Descriptions Rankings Bottled Beer Taste Must be refrigerated $8.00/ six 12 oz. bottles Draft Beer Taste Shelf Stable $6.00/ six 12 oz. bottles Bottled Beer Taste Shelf Stable $6.00/ six 12 oz. bottles Draft Beer Taste Must be refrigerated $6.00/ six 12 oz. bottles Draft Beer Taste Must be refrigerated $8.00/ six 12 oz. bottles Draft Beer Taste Shelf Stable $8.00/ six 12 oz. bottles Bottled Beer Taste Shelf Stable $8.00/ six 12 oz. bottles Bottled Beer Taste Must be refrigerated $6.00/ six 12 oz. bottles Figure B-2: Sample Ballot for Conjoint Analysis

PAGE 102

APPENDIX C STATISTICAL MATERIAL Response Surface SAS Code data firstmicro; input X1 X2 X3 X4 Y; u1=27.6; u2=35; u3=10; u4=5; v1=6.9; v2=10; v3=2; v4=1; z1=(X1-u1)/v1; z2=(X2-u2)/v2; z3=(X3-u3)/v3; z4=(X4-u4)/v4; cards; 27.6 21 10 5 6.1106 20.7 25 8 4 6.1106 34.5 25 8 4 6.1106 2 0.7 25 8 6 6.1106 34.5 25 8 6 4.7126 20.7 25 12 4 6.1106 34.5 25 12 4 4.4116 2 0.7 25 12 6 4.2355 34.5 25 12 6 4.2355 2 7.6 35 7.175 5 3.1941 2 7.6 35 10 3.56 3.6335 17.8 35 10 5 3.4574 27.6 35 10 5 4.7126 27.6 35 10 5 6.1106 27.6 35 10 5 6.1106 37.3 35 10 5 3.9345 27.6 35 10 6.41 4.7126 27.6 35 12.83 5 4.7126 20.7 45 8 4 2.8866 34.5 45 8 4 3.0594 20.7 45 8 6 4.0137 34.5 45 8 6 4.1106 20.7 45 12 4 3.3702 34.5 45 12 4 4.0137 20.7 45 12 6 4.4116 34.5 45 12 6 4.4116 27.6 49 10 5 6.1106 27.6 21 10 5 6.0569 20.7 25 8 4 6.0569 34.5 25 8 4 6.0569 90

PAGE 103

91 20.7 25 8 6 6.0569 34.5 25 8 6 4.1818 20.7 25 12 4 6.0569 34.5 25 12 4 4.1818 20.7 25 12 6 3.96 34.5 25 12 6 3.96 27.6 35 7.175 5 3.2788 27.6 35 10 3.56 3.5798 17.8 35 10 5 3.4037 27.6 35 10 5 4.6590 27.6 35 10 5 6.0569 27.6 35 10 5 6.0569 37.3 35 10 5 3.8139 27.6 35 10 6.41 4.3579 27.6 35 12.83 5 4.1818 20.7 45 8 4 3.2118 34.5 45 8 4 3.3802 20.7 45 8 6 3.7559 34.5 45 8 6 3.8139 20.7 45 12 4 3.2972 34.5 45 12 4 3.8139 20.7 45 12 6 4.3579 34.5 45 12 6 4.1818 27.6 49 10 5 6.0569 ; proc print data=firstmicro;run; proc sort data=firstmicro; by X1 X2 X3 X4; run; proc rsreg data=firstmicro; model y=x1 x2 x3 x4; ridge max; run; Example Code for ANOVA and Tukeys Test PROC IMPORT OUT= WORK.taste_031804 DATAFILE= "C:\Documents and Settings\GFolkes\Desktop\Sensory Results\031804 results SAS taste.xls" DBMS=EXCEL2000 REPLACE; SHEET="Sheet1$"; GETNAMES=YES; RUN; proc print; run; p roc sort; by Samp_Set Samp_ Samp_Code Taste_Differe nce; proc glm; class Samp_Set Samp_ Samp_Code; model Taste_Difference=Samp_Set Samp_ Samp_Code; means Samp_/tukey alpha=.01; run;

PAGE 104

LIST OF REFERENCES Arreola, A. G., Balaban M. O., Marshall M. R., Peplow A. J., Wei C. I., Cornell J. A. 1991. Supercritical carbon dioxide effects on some quality attributes of single juice. Journal of Food Science. 56: 1030-1033 hingawa, K., Hashimoto, N. 1982. Characterization of haze-forming teins of beer and their roles in chill haze formation. Journal of American iety of Brewing Chemists. 40: 147-154. P., Abreu da Silva, A., Cuq, J. L. 1996. Inactivation of Escherichia coli by rbon dioxide under pressure. Journal of Food Science. 61 (4): 829-831, 836. 000. Brewing and brewing research: past, present, and future. Science of Food and Agriculture. 80: 1371-1378 C. W. 2003. Tap Into the Art and Science of Brewing, Oxford University ed Kingdom. y ingredients in beer. Ferment. 13: 20-24 beerservesamerica.org/economic/index/phtml Last Accessed: institute.org/pdfs/PER_CAPITA-MALT_BEVERAGES.PDF, Last Accessed: 07/22/2003b rg/pdfs/PRODUCTION_OF_MALT_BEVERAGES_IN_US.P 07/22/2003c Neuman, F. 1939. Preserving Cider by Carbonation. Am. 371-373 982. On the solubility of glycerides and fatty acids in in the presence of an entrainer. Separation Science and 199-214. Microbiology: Fundamentals and Frontiers., American ology, Washington, DC. strength orange Asano, K ., S proc So Balstra, le ca Bamforth, C. W. 2 Journal of the Bamforth Press, Oxford, Un it Baxter, D. 2000. Health The Beer Institute, www. 07/22/2003a The Beer Institute, www.beer CONSUMPTION_OF_ The Beer Institute, www.beerinstitute.o DF, Last Accessed: Brown, H. D., Fitzgerald, C. i. 36: Soc. of Hortic. Sc Brunner, G., Peter, S. 1 compressed gases Technology. 17(1): Campbell, I. 1997. Food Society for Microbi 92

PAGE 105

93 Clifford, J. R., Williams A. A., 2000. Supercritical Fluid Methods and Protocols. wa, New Jersey. od Proteins and Their Applications. Marcel Decker, New akoda M., Amaya N. 1997a. Lethal effect of high-xide on a bacterial spore. Journal of Fermentation ad : 305-307 ., Nagai K., Hashimoto T., Hakoda M. 1997b. Inactivation of y high-pressure carbon dioxide treatment with or without n. Biosci. Biotech. Biochem. 61: 1133-1137 microbial effect of pressurized carbon dioxide on Staphylococcus d milk. Lebensm.-Wiss. u.-Technol. 30: 826-829 ictive Modeling of Listeria Monocytogenes inactivation under n dioxide. Lebensm.-Wiss. u.-Technol. 33: 514-519. ct of carbon dioxide pressure on Listeria Monocytogenes in od Microbiology. 17: 589-596. Antimicrobial effect of pressurized carbon dioxide on Enterococcus in physiological saline and foods. Journal of the Science of Food and f Brewing Science. Brewers Publications, Boulder, Colorado. raser, D. 1951. Bursting bacteria by release of gas pressure. Nature. 167 (4236): 33-34. Friedrich, J.E.; Acree, T.E. 2000. Flavor Chemistry: Industrial and Academic Research. n DC. ndbook. Apex Publishers, Colchester, Essex, a Miguel 1999. An overview and projection to the year American brewing industry. Brewers Digest. 74: 60-66 1998. Practical Phenolics: From Structure to Molecular Recognition and gical Action. Cambridge University Press, Cambridge, United Kingdom. cation and properties of the major antigenic beer ing. 89: 402-410. Humana Press, Toto Damodaran, S. 1996. Fo York. Enomoto, A., Nakamura K., H pressure carbon dio Bioengineering. 83 Enomoto, A., Nakamura K food microorganisms b explosive decompressio Erkman, O. 1997. Anti aureus in broth an Erkman, O. 2000a. Pred high pressure carbo Er kman, O. 2000b. Effe physiological saline and foods. Fo Erkman, O. 2000c. faecalis Agriculture. 80: 465-470 Fix, G. 1999. Principles o F American Chemical Society. Washingto Goldammer, Ted. 2000. The Brewers Ha United Kingdom. Gonzalez, del Cuerto A006 of the North Mor 2 Haslam, E. Physiolo Hejgaard, J., Kaergaard, P. 1983. Purifi protein of barley origin. Journal of the Institute of Brew

PAGE 106

94 Hair, J., Anderson, R., Tathom, R., Black, W. 1998. Multivariate Data Analysis. 5th ed. Pretence Hall, Upper Saddle River, NJ. Hong, S., Park W., Pyun Y. 1997. Inactivation of Lactobacillus sp. from Kimchi by high pressure carbon dioxide. Lebensm.-Wiss. u.-Technol. 30: 681-685 Hong, S., Park W., Pyun Y. 19on of Lactobacillus plantarum as influence by pressure and temperature of pressurized carbon dioxide. Hoov the R & D Associates. 49/50: 294-296 Illes, solvents, Proc. 3rd Symp. Supercritical Fluids, France. 2: 511. Illes, t by supercritical CO and propane. Journal of Supercritical Fluids. 10: 209. Insala 84-85, 190 n the survival of yeast cells. Journal of Biotechnology. 39: 229-237 Ishikawa, H., Shimoda, M., Shiratsuchi, H., Osajima, Y. 1995. Sterilization of hem. 59 (10): 1949-1950. T., Osajima, Y. 1997. method. Biosci. Biotech. Biochem. 61 (6): 1022-1023. Jenck hydrostatic pressure-pasteurization. Food Microbiology. 15: 207-214 Kamis with gric. Biol. Chem. 51: 407-412 ce. 99. Non-thermal inactivati International Journal of Food Science and Technology. 34: 125-130 er, D. G. 1998. High-Pressure Pasteurization. Activities report of V. Szalai, O., Szebenyi, N. I., Grosz, M., Hethelyi, I. 1994. Oil recovery fromborage seed with carbon dioxide and propane V. Szalai, O., Then, M., Daood, H., Perniczki, S. 1997. Extraction of hiprose frui 2ta, N. 1952. CO2 Versus Beverage Bacteria. Food Engineering. Isenschmid, A., Marison I. W., von Stockar, U. 1995. The influence of pressure and temperature of compressed CO2 o microorganis ms by the supercritical carbon dioxide micro-bubble method. Biosci. Biotech. Bioc Ishikawa, H., Shim oda, M., Tamaya, K., Yonekura, A., KawanoInactivation of Bacillus spores by the supercritical c arbon dioxide micro-bubble s, W. P. 1969. Catalysts in chemistry and enzymology. McGrw-Hill Pbulishers, New York. Kalchayanand, N., Sikes A., Dunne C. P., Ray B. 1998. Factors influencing death and injury of foodbourne pathogens by hira, M., Taniguchi M., Kobayashi T. 1987. Sterilization of microorganismsupercritical carbon dioxide. A Kamimura, M., Kaneda H. 1992. Off-Flavors in Beer Developments in Food Scien28: 433-472

PAGE 107

95 Kaneda, H., Kano Y., Osawa T., Kawakishi S., Koshino S. 1994. Free radical reactiin beer during pasteurization. Internati on onal Journal of Food Science and Technology. 29: 195-200 Kano, Y., Kamimura, M. 1993. Simple methods for determination of the molecular weight distribution of beer proteins and their application to foam and haze studies. Khurie Surfaces: Designs and Analyses. CRC Press, Inc. Boca Raton, FL. Kinca, is, University of Florida Kumalls and sterilization by high-pressure CO2. Biosci. Biotech. Biochem. 61 (6): 931-935. Lin, Hrized carbon dioxide. Biotechnol. Prog. 7: 201-204 Lin, Hae by supercritical and subcritical carbon dioxide. Biotechnol. Prog. 8: 458-461 Lin, Hn e Chemical Engineering Journal. 52: b29-b34 with pressurized carbon dioxide. Biotechnol. Prog. 8: 165-166 Maru, A. A., Goldstein, H., Navarro, A., Seabrooks, J. R., Ryder, D. S. 2003. Investigation of Beer Flavor by Gas Chromatography-Olfactometry. Journal of McMurrough, I. Kelly, R., Byrne, J. 1992. Effect of removal of sensitive protein and y ts. 50: 67-76. od Journal of the American Society of Brewing Chemists. 51: 21-28. A., Cornell, J. 1996. Respons l, D. 2000. A continuous high pressure carbon dioxide system for cloud retentionmicrobial reduction and quality change in orange juice. Masters Thes gai, H., Hata, C., Nakamura, K. 1997. CO2 Sorption by microbial ce ., Chan E., Chen C., Chen L. 1991. Disintegration of yeast cells by pressu ., Yang Z., Chen L. 1992a. Inactivation of Saccharomyces cerevisi ., Yang Z., Chen L. 1993. Inactivation of Leuconostoc dextranicum with carbodioxide under pressure. Th Lin, H., Yang Z., Chen L. 1992b. An improved method for disruption of microbial cells kami American Society of Brewing Chemists. 61(1): 23-32. proanthocyanidins on colloidal stability of lager beer. Journal of American Societof Brewing Chemis Mermelstein, N. H. 1997. Extending dairy product shelf life with carbon dioxide. FoTechnology. 51: 72 Meyssami, B., Balaban M., Teixeira A. 1992. Prediction of pH in Model Systems Pressurized with Carbon Dioxide. Biotechnology. 8: 149-154 Molin, G. 1983. The resistance top carbon dioxide of some food related bacteria. European Journal of Applied Microbiology Biotechnology. 18: 214-217

PAGE 108

96 Nakamura, K., Enomoto A., Fukishima K., Nagai K., Hakoda M. 1994. Disruption of microbial cells by the flash discharge of highpressure carbon dioxide. Biosci. Biotech. Biochem. 58: 1297-1301 Naim. Chemical and Sensory Analysis of Off-Flavors in Citrus Products. Chapter 25 in Flavor Odabasi, A. Z. 2003. Shell Color and Other Quality Attributes of Brown Eggs as 28: 394-of Foods and Flavors. Editors: Rouseff, R. L., Cadwallader, K. R, Kluwer Raghastry, S. V. G. K., Mukhopadyah, M. 1992. Modeling solute-cosolvent interactions for supercritical fluid extraction of RouseIndustrial and Academic Research. American Chemical Society. Washington DC. Siebe in 44: 1997-2005. rotein. Journal of the American Society of Brewing Chemists. 56: 24-31. Siebe. ood Chemistry. 44: 80-85. eath kinetics of Saccharomyces cerevisiae. Journal of Applied Microbiology. 91: 306-311. Shimmoto Y., Cocunubo-Castellanos H., Tonoike H., Kawano T., Ishikawa H., Osajima Y. 1998. Antimicrobial effects of pressured carbon dioxide in a continuous flow system. Journal of Food Science. 63: 709-712 M., Rouseff, R. L., Zehavi, U., Schutz, O., Halvera-Toledo, E. 1998 Analysis: Developments in Isolation and Characterization. Editors: Mussinan and Morello. American Chemical Society, Washington DC. Affected by the Hens Age and Vanadium in Their Diet. Ph.D. Dissertation. University of Florida. Oh, H. I., Hoff, J. E., Armstrong, G. S., Haff, L. A. 1980. Hydrophobic interaction intannin-protein complexes. Journal of Agricultural and Food Chemistry.398. Ott, R. L. 1993. An Introduction to Statistical Methods and Data Analysis. 4th ed. Wadsworth, Inc., Belmont, CA. Pawliszyn, J. 2001. Solid Phase Microextraction. Chapter 6 in Headspace Analysis Academic/Plenum Publishers, New York. arum Rao, G. V., Srinivas, P.,S fragrances. Journal of Supercritical Fluids. 5(1): 19. ff, R.; Naim, M. 2000. Flavor Chemistry: rt, K. J., Carrasco, A., Lynn, P. Y. 1996b. Formation of protein-polyphenol hazebeverages. Journal of Agricultural and Food Chemistry. Siebert, K. J., Lynn, P. Y. 1998. Comparison of polyphenol interactions with polyvinylpolypyrrolidine and haze-active p rt, K. J., Troukhanova, N. V., 1996a. Nature of polyphenol-protein interactionsJournal of Agricultural and F Shimoda, M., Cocunubo-Castellanos, J., Kago, H., Miyake, M., Osajima, Y., Hayakawa, I. 2001. The influence of dissolved CO2 concentration on the d oda, M., Yama

PAGE 109

97 Spilimbergo, S., Elvassore, N., Bertucco, A. 2003a. Microbial inactivation by high-pressure. The Journal of Supercritical Fluids. 22: 55-63. Spilimbergo, S., Elvassore, N., Bertucco, A. 2003b. Inactivation of microorganisms by e. 1 Summerour, J. 2001. Tasting success. Progressive Grocer. 80: 43-44 Tayloercritical Fluid Extraction. John Wiley and Sons, Inc., New York. Toussden, MA. Toussaint-Samat, M., Alberny R., Horman I. 1991. 2 Million Years of the Food Walstra, P. Chapter 3: Dispersed Systems: Basic Considerations in Food Chemistry. Wei, C. I., Balaban M. O., Fernando S. Y., Peplow A. J. 1991. Bacterial effect of high al of supercritical CO2 by a semi-continuous process. Italian Journal of Food Scienc(15): 115-124. r, L. T., 1996. Sup aint-Samat, M. 1992. A History of Food. Blackwell Publishers Ltd., Mal Industry. Nestle S. A., (Editor: Fennema, O.) Marcel Dekker, New York. 1996. Pgs. 143-147. pressure CO2 treatment on foods spiked with Listeria or Salmonella. JournFood Protection. 54: 189-193

PAGE 110

BIOGRAPHICAL SKETCH Gillian Folkes was born in 1978 in Tampa, FL, to Edward V. Folkes, Jr. and Gilda Folke of Florida as a National Merit Scholar and graduated in May of 2000 with a bachelors degrefollowing fall, Gillian started her Ph.D. program under the instruction of Dr. Murat Balabarrying Roi Dagan in May of 2004, Gillian will 2004,orking for ABC Research Corporation. s. After graduating high school as salutatorian, Gillian attended the University e in food science with honors and a minor in business administration. The an as an Alumni Fellow. After m receive her Ph.D. in food science and a minor in food resource economics in August and will continue to reside in Gainesville, w 98


Permanent Link: http://ufdc.ufl.edu/UFE0006549/00001

Material Information

Title: Pasteurization of Beer by a Continuous Dense-Phase CO2 System
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006549:00001

Permanent Link: http://ufdc.ufl.edu/UFE0006549/00001

Material Information

Title: Pasteurization of Beer by a Continuous Dense-Phase CO2 System
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006549:00001


This item has the following downloads:


Full Text












PASTEURIZATION OF BEER BY A CONTINUOUS DENSE-PHASE CO2 SYSTEM


By

GILLIAN FOLKES


















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Gillian Folkes

































This document is dedicated to beer drinkers everywhere. Enjoy.















ACKNOWLEDGMENTS

I would like to Dr. Charles Sims, Dr. Marty Marshall, Dr. Al Wysocki, Dr. Andre

Khuri, and especially Dr. Murat Balaban, my major advisor. He not only has supported

me with encouragement and challenges, but has also been an exceptional mentor and

friend.

Secondly, I would like to thank my parents who were not only very good at raising

a child, but were also exceptional at raising an adult. They gave me a solid foundation

not only for academics, but also for life and love.

Finally, I would like to thank my husband, Roi Dagan, who knew exactly what to

say, even when he said, "I love you. Now go do some work!" He truly understands what

research is all about.
















TABLE OF CONTENTS
Page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES .......................................... viii

LIST OF FIGURES .................................................. ............................ ix

ABSTRACT .................................................... ................. xi

CHAPTER

1 IN T R O D U C T IO N ................................................. .............................................. .

2 LITER A TU R E R EV IEW .................................................................... ...............5...

Beer Production and Consumption in the United States ........................................5...
B eer C om p o sition ................................................................................................. 6
Y east Cultures in B eer .... .. ................................. ............................ .......... .. 7
B e er Q u ality ......................................................................................................... 8
B e er C o lo r.......................................................... ................................................ 9
B e e r H a z e ................................................................................................................. ... 1 0
C om position of B eer H aze ..................................... ..................... ............... 10
Form ation of H aze ........................................................... .............................. .. 11
Haze Removal and Beer Stabilization Techniques ........................................13
B e e r F o a m ............... ............................................................................................... ... 1 5
B eer Foam C om position .................................... ....................... ............... 15
B eer F oam F orm ation ......................................... ......................... .............. 16
B eer F oam Stabilization ....................................... ....................... ............... 16
B e e r F lav o r ................................................................................................................ 1 7
Processing of Beer ....................................... ........ ................... 18
P a steu riz atio n ...................................................................................................... 18
Flash pasteurization ..................................... .. .......... .... ........ .. .......... .. 19
N ontherm al m ethods ............................................................... ............... 19
Effects of Dense-Phase CO2 Pasteurization .......................................... ................ 20
Microbiology .............. .. .. .................... ................... 20
Theories of Cell Death by Dense-Phase CO2 Pasteurization ..............................23
Effects of Cosolvents ......................................................................... ............ 24
Quality Attributes ............... ...... ............... 25
D ense-Phase CO2 Pasteurization of Beer .............................................. ................ 26
O b j e ctiv e s ............................................................................................................... .. 2 6


v









3 M A TERIALS AND M ETH OD S .......................................................... ................ 28

The D ense-Phase CO 2 System ................................... ...................... ................ 28
B eer Sam ples ............................................................................................. . 29
E xperim mental D design ............... ................ .............................................. 30
C leanability Study ............. .. ............... .............................................. 30
E xperim ental D esign ................................................................... .................. 3 1
P procedures ........................................................................... .......................... 32
A analysis of Treated Sam ples ....................................... ....................... ................ 33
M icrobial R education Experim ents.................................................. ................ 33
H aze M easurem ent ............ .... ............... ................................................ 33
Foam Capacity and Stability M easurements.................................. ................ 33
Polyacrylamide Gel Electrophoresis of Beer Proteins ..................................34
F lavor ................................................................................... ...................... 34
Sensory .............................................................................................36
Fresh beer vs. dense-phase CO2 processed beer.....................................37
Dense-phase CO2 processed beer vs. heat pasteurized beer......................37
Storage studies ...........................................................................................38
Statistical Analysis...............................................38
Conjoint Analysis of Beer Purchase Decision....................................... ................ 38

4 RE SU L TS A N D D ISCU SSION ............................................................ .................. 39

M icrobial R education Experim ents......................................................... ................ 39
M ode of C ell D death ................ ... .................. ....................... ......................... 44
Effect of Dense-phase CO2 Processing on Haze ...................................................46
Effect of Dense-phase CO2 Processing on Foam Capacity and Stability ................48
Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Sensory
E v a lu atio n ............................................................................................................. 5 3
B e er A ro m a ......................................................................................................... 5 3
B eer F lav or .......................................................................... . .. ............... 56
B eer A rom a A after Storage.............................................................. ............... 59
B eer F lavor A after Storage ...................................................... ..... ................... 6 1
Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Gas
Chromatography-Olfactometry and Mass Spectrometry ..................................65
Conjoint Analysis of Beer Purchase Decisions .............. ....................................72

5 C O N C L U S IO N ...........................................................................................................7 5









APPENDIX

A R A W EX PER IM EN T D A TA .......................................... ...................... ............... 77

B SENSORY AND CONJOINT BALLOTS................ ...................................88

C STA TISTICAL M A TERIAL ...................................... ....................... ................ 90

L IST O F R E FE R E N C E S .................................................. ........................................... 92

BIO GR APH ICAL SK ETCH .................................................................... ................ 98















LIST OF TABLES


Table page

3-1 E xperim ental D esign .. ...................................................................... ................ 32

4-1 M icrobial R education R results ............................................................... ................ 42

4-2 Beer Haze in NTU After Processing and After Storage at 1.670 C for 30 Days......47

4-4 Retention Times and Aroma Descriptors for Compounds Detected by Both
Assessors on ZB-5 and Carbowax Columns....................................... ................ 65

4-5 Linear Retention Indices and Identification of Compounds using GC-O ...............67

4-6 Average Integration Areas of Identified Compounds in Fresh and CO2
P ro cessed B eer ......................................................................................................... 6 9

4-7 Conjoint A analysis Transform ation...................................................... ................ 74

A-i Yeast Counts for 27 Treatment Combinations Done in Duplicate........................78

A -2 Foam and L iquid V olum es........................................ ....................... ............... 87















LIST OF FIGURES


Figure page

2-1 Yeast Growth Stages during Ferm entation ......................................... ...............8...

2-2 Molecular Mechanisms for Haze Formation in Beer ........... ..................14

3-1 Schematic of the Continuous Dense-Phase CO2 Pasteurization System...............29

4-1 Scanning Electron Microscopy Picture of Yeast in Fresh Beer...............................45

4-2 Scanning Electron Microscopy Picture of Yeast in Beer Dense-Phase CO2
Processed at 27.6 MPa, 10% CO2, at 210C, With a Residence Time of 5
M in u te s ..................................................................................................................... 4 5

4-3 Scanning Electron Microscopy Picture of Yeast in Beer Pasteurized at 740C for
3 0 S e c o n d s ............................................................................................................... 4 5

4-4 Beer Haze Following Processing and Following Storage at 1.670C for 30 Days....47

4-5 Foam Capacity and Stability of Beer Samples After Processing..........................50

4-6 Foam Capacity and Stability of Beer Samples After Storage at 1.670C for
3 0 D a y s .................................................................................................................. ... 5 0

4-7 Aroma Evaluation of Fresh and CO2 Processed Beer Samples.............................. 54

4-8 Evaluation of Aromas of Fresh, CO2 Processed and Heat Pasteurized Beer
S a m p le s ................................................................................................................. .. 5 6

4-9 Evaluation of Beer Flavor Between Fresh and CO2 Processed Samples..............57

4-10 Evaluation of Beer Flavors Between Fresh, CO2 Processed, and Heat
P asteurized Sam ples .... ................................................................... .............. 58

4-11 Aroma Evaluation of Aroma Between Aged, CO2 Processed, Heat Pasteurized,
and a N on-Stored, Fresh H idden Control............................................ ................ 60

4-12 Evaluation of Flavor Between Aged, CO2 Processed, Heat Pasteurized, and
a N on-Stored, Fresh H idden Control................................................... ................ 61









4-13 Evaluation of Flavor Between Fresh and CO2 Processed Beer After Storage at
1.670C for 30 Days, and a Non-Stored, Fresh Hidden Control...............................63

4-14 Evaluation of Flavor of Fresh, CO2 Processed, and Heat Pasteurized Samples
After 30 Days at 1.670C, and a Non-Stored, Fresh, Hidden Control....................64

4-15 Typical FID Chromatogram for Fresh Beer........................................................70

4-16 Typical A rom agram for Fresh B eer .................................................... ................ 70

A -i P olyacrylam ide G els .. ...................................................................... ................ 87

B -i Sam ple Sensory B allot .................................................................... ................ 88

B -2 Sam ple B allot for Conjoint A analysis .................................................. ................ 89















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

PASTEURIZATION OF BEER BY A CONTINUOUS DENSE-PHASE CO2 SYSTEM

By

Gillian Folkes

August 2004

Chair: Murat Balaban
Major Department: Food Science and Human Nutrition

The world production of beer grew 26% between 1987 and 1997, despite the

growing market of competing beverages. In 1999-2000, $7.7 billion worth of beer was

produced in the United States.

Bottled beer is currently flash-pasteurized. Because beer is a delicate beverage,

off-flavors are easily formed during pasteurization. With "freshness" being top priority,

it is evident a method of pasteurization using no heat would be of great help to the

brewing industry.

Currently there is great interest in dense-phase CO2 as an alternative processing

method, and studies using the combination of carbon dioxide and pressure for

pasteurization have been successful. Not only is microbial inactivation achieved, but also

no taste or aroma changes are perceived, and vitamin quality is maintained.

The purpose of this investigation was to evaluate the effectiveness of dense-phase

CO2 pasteurization system with beer. A predicted maximum log reduction in yeast









populations of 7.38 logs was seen at 26.5 MPa, 210C, 9.6% C02, and 4.77 minutes

residence time. Haze was slightly reduced by dense phase CO2 pasteurization from 146

NTU to 95 NTU. At this same treatment combination, aroma and flavor of beer sample

means were not considered significantly different (p=.3415) from fresh beer sample

means when evaluated in a difference from control test, using fresh beer as the reference.

Foam capacity and stability were affected minimally by CO2 processing, however

changes would most likely be unnoticed by consumers.

In addition, a conjoint analysis was performed to examine motives during beer

purchase decision. The attribute eliciting the large influence on purchase decision was

lower price, followed by a preference for draft beer taste and an extended shelf life.

Furthermore, indications of the mode of cell death were absorption of CO2 into the

cell membrane creating a physical disruption in membrane structure, visible as divots in

scanning electron microscopy pictures.

A continuous dense-phase CO2 system was effective in the pasteurization of beer.

The ability to produce a clear, consistently fresh beer that forms a good head, with an

extended shelf life is top-priority to brewers. Dense-phase CO2 pasteurization can make

this possible.














CHAPTER 1
INTRODUCTION

Beer dates back to 4000 B.C. when the Babylonians described ale in some of the

world's oldest writings; however it is believed that beer existed far before these records

(Toussaint-Samat et al., 1991). Egyptians believed that Osiris, the god of agriculture,

made a decoction of barley that had germinated in the Nile river. Becoming distracted he

left it in the sun and forgot it. Upon his return the liquid had fermented. He drank it and

proclaimed mankind should profit from it (Toussaint-Samat, 1992).

The world production of beer grew 26% between 1987 and 1997, despite the

growing market of competing beverages (Gonzalez del Cueto and Miguel, 1999). In

1999-2000, $7.7 billion worth of beer was produced in the United States (Summerour,

2001).

Currently bottled beer is flash-pasteurized. Because beer is a delicate and heat

labile beverage, off-flavors are easily formed during pasteurization. In a study by

Kaneda et al. (1994), non-pasteurized beer was compared to pasteurized bottled beer.

Off-flavors of bottled beers pasteurized at 15-30 pasteurization units (P.U.) had a similar

off-flavor profile to non-pasteurized beers stored at 20C for 6-10 days. With "freshness"

being top priority, it is evident a method of pasteurization using no heat would be of great

help to the brewing industry. Its use would lead to a beer with a longer shelf life, cheaper

production and distribution cost because of the elimination of some refrigerated

distribution centers and trucks, and a fresher taste and aroma. Freshness is of utmost

importance in the brewing industry. Current evidence of this is the use of colored glass









bottles, refrigerated distribution houses and trucks, and the popularization of the "born-

on" date.

One method of non-thermal pasteurization is high hydrostatic pressure processing.

There are a variety of foods such as jams, jellies, guacamole, and pate where high

isostatic pressure technology is applied to inactivate organisms at an efficacy equal to

that of traditional thermal pasteurization methods. The use of high pressure insures no

noticeable changes in taste or flavors of these foods and does not destroy vitamins

(Hoover, 1998).

The combination of high pressure pasteurization with other preservation methods

could lead to a more desirable process or product. The effect of the combination of

pressure, time, and temperature on food has been studied, showing that with the use of

heat and pressure effective pasteurization can take place at low to moderate pressures

(Kalchayanand et al., 1998). Since achieving high pressures can be expensive, the

combined methods insure not only safe food products but also lower costs. However, the

use of heat still may degrade vitamins or change the taste, aroma, or texture of the food.

In addition, high hydrostatic pressure processing is currently a batch process. For the

large volumes that would be processed in the case of beer, a continuous process is more

desirable.

Currently there is great interest in dense-phase CO2 as an alternative processing

method, and studies using the combination of carbon dioxide and pressure for

pasteurization have been successful. Not only is microbial inactivation achieved, but also

no taste, or aroma changes are perceived and vitamin quality is maintained. The addition

of carbon dioxide to the high pressure treatment allows pressures as low as 14 to 107









MPa to be used for effective microbial inactivation (Kincal, 2000), instead of 400-900

MPa. These pressures are lower than those in the combined pressure/heat treatment, and

this creates a cost effective alternative pasteurization method because of the small amount

of energy used for pressurization and the use of an inexpensive gas.

The need for inhibition of pathogen/spoilage organism growth and the

characteristic carbonation of beer make it an excellent candidate for dense-phase CO2

pasteurization.

The purpose of this investigation was to evaluate the effectiveness of dense-phase

CO2 pasteurization system for beer. The success of this system will rely on its ability to

inactivate microorganisms, preserve fresh beer taste and aroma, not exacerbate or

possibly prevent beer haze, and insure proper foam formation and stabilization.

Therefore a study examining these characteristics was conducted in order to compare it

with current pasteurization methods and to make inferences regarding the intermolecular

changes occurring in beer during the high pressure carbon dioxide pasteurization.

Although there is promise in the use of dense-phase CO2 pasteurization technology

with beer, research must be done to insure the technical and economic feasibility of the

procedure. The objectives of this study were:

1. To quantify and elucidate mechanisms for inactivation of yeast by dense-phase CO2
pasteurization of beer

2. To prove there are no significant changes in flavor and aroma of beer after dense-
phase CO2 pasteurization

3. To prove there are no significant changes in chill haze formation of beer after
dense-phase CO2 pasteurization

4. To prove there are no significant changes in foam formation and stability after
dense-phase CO2 pasteurization






4


5. To evaluate consumer and industry acceptability and the economic impacts of
dense-phase CO2 pasteurization














CHAPTER 2
LITERATURE REVIEW

Beer Production and Consumption in the United States

The brewing industry in the United States is an active part of the national economy,

paying billions of dollars annually in taxes and wages. One sign that beer is a significant

force in today's economy is that it is included in the basket of goods used to calculate the

Consumer Price Index. Currently the U.S. brewing industry employs approximately 1.66

million workers and pays over $47 billion in wages annually (The Beer Institute, 2003a,

2003b, 2003c).

In 2002, the U.S. brewing industry recorded its seventh straight year of growth.

Currently, there are over 3,500 brands of malt beverages on the market and in the year

2000 it was estimated that 199,650,000 barrels of beer were produced in the United

States. For the same year, consumption of malt beverages reached 21.8 gallons per

capital in the U.S. (The Beer Institute, 2003a, 2003b, 2003c).

Beer consumption is mostly male-dominated, with men accounting for more than

80% of the volume consumed. A large number of these beer drinkers are white and

prefer domestic light beer, followed by domestic draft beer. African American drinkers

make up about 10% of the beer market. In general, they are the biggest consumers of

malt liquors, followed by ice beer. Considering all beer styles, light beer has the

strongest following among women consumers. Women beer drinkers are also more

attracted to specialty micro-brewed beers because of their greater variety (Goldammer,

2000).









Beer Composition

Beer contains a variety of components, and although many do not regard beer as a

health food, many components of beer are conducive to good health. Beer is mostly

water and is made from malted grains such as barley or wheat, hops, yeast, and

sometimes adjuncts such as corn or rice. Beer contains about 4-5% alcohol by volume,

which if consumed moderately can help reduce the risk of cardiovascular disease

(Baxter, 2000).

On average, a beer's two main components, excluding water, are carbohydrates (1

to 60 g/liter) and proteins (2 to 6 g/liter) usually in the form of peptides. The

carbohydrates are found in the form of branched dextrans and not as free sugars, which

would have been consumed by the yeast during fermentation. These dextrans not only

have little immediate impact on blood sugar levels as free sugars but also are less

cariogenic. Another health claim of beer is that it contains no fat (Baxter, 2000).

Because beers are made from malted grains they are a good source of B vitamins.

Beer is generally not considered as the main source of B vitamins; however it becomes an

important source of these in malnourished societies. Yeast also contributes to the B

vitamins in beer. For example, about 1 liter (2 pints) of beer may provide one third to

one half of a consumer's daily requirement of 5 B vitamins, including folate. From the

use of malted barley also comes silicon, which maintains healthy bones, magnesium, and

potassium (Baxter, 2000).

Beverages such as fruit juices and wines have already been recognized for their

readily available antioxidants. Beer contains two distinct sources of antioxidants:

melanoidins and polyphenols. Melanoidins are formed during the roasting of malted

barley from Maillard reactions. Polyphenols come from not only the malt in the form of









ferulic acid, but also from hops in the form of catechin. Although antioxidant levels are

comparatively low in beer, their high bioavailability makes beer as competitive as other

foods rich in antioxidants (Baxter, 2000).

Yeast Cultures in Beer

In the production of beer, only 4 ingredients are necessary: water, malt, hops, and

yeast. After germination of the barley to create malt, the grain is ground and water

added. This solution is then boiled to create the wort. It is then cooled and the yeast is

added to start the fermentation. In the brewing operation, yeast functions as the means of

transforming the fermentable sugars in the malt into alcohol, C02, and heat through

fermentation, as shown in the Gay-Lussac equation:

C6H1205-2(C2H5OH) + 2(CO2) +heat

There are two types of yeast used in brewing: top and bottom fermenters. Ales are

made from top fermenting yeast, most commonly Sacchromyces cerevisiae, and lager

beers are made from bottom fermenting yeasts such as Sacchromyces carlsbergensis.

Although yeasts are usually referred to as facultative anaerobes, they are actually aerobes.

During fermentation of beer, yeasts switch between oxidative and fermentative

metabolisms, depending on the presence of oxygen; however they cannot grow

anaerobically indefinitely. The cell membrane of all eukaryotes, yeasts included, contain

unsaturated fatty acids and sterols. These compounds can only be produced by the yeast

under aerobic conditions and although these compounds do exist in wort, the amounts are

too low to sustain yeast growth. Therefore, when "pitching" or adding the initial yeast

culture to the wort to start the fermentation, approximately 1 x 107 to 2 x 107 cells per ml

are added. Because the wort is oxygenated prior to fermentation, yeast numbers may

only increase by three cell divisions, or a factor of 8. Subsequent multiplication of the









yeast is inhibited because of the lack of further aeration of the wort which would lead to

off flavors in the beer. Furthermore, fermentation includes the lag, exponential growth,

and stationary phases of yeast growth as seen in Figure 1.




S. cerevisiae
-population



1 Ethanol



Sa-Amino nitrogen


0
I,-


0 -Fermentable sugar




Time -*

Figure 2-1: Yeast Growth Stages during Fermentation (modified from Campbell, 1997).

In kegged or draft beer, because it is not pasteurized after packaging, yeast cultures

are still viable. Without proper refrigeration or pasteurization the shelf life of this beer is

only several hours, because the yeast cultures will continue to ferment the beer and will

create off-flavors in the process. With proper refrigeration or pasteurization this is

prevented. Bottled beer is flash pasteurized to make it a shelf stable product between

71.5 to 740 C and held for 15 to 30 seconds. The pasteurization step kills all yeast that

remain in beer after fermentation and packaging are complete (Goldammer, 2000).

Beer Quality

The importance of quality in the brewing industry is evident in many of today's

brewing practices. The popularization of the "born-on date" and extensive expenditures









on advertising such characteristics as "quality ingredients time honored practices ...

finest heritage" of certain beers, highlight the industry's pride in and strive towards

quality.

The visual quality of a beer is the first to be judged by the consumer. The package

of beer, either the draught, bottle, or can, can easily help or hinder the sales of a product.

Recent innovations in the area of packaging have been tamper proof kegs, oxygen

scavenging crowns for bottles, narrowing the necks of cans to conserve metal, the use of

a "widget" which creates better foam in beers of low carbonation, and the use of plastic

bottles for sales of beer in recreational areas where glass bottles may cause danger

(Bamforth, 2000). The next characteristics to be inspected are the beer's color, clarity,

and foaming capabilities, and then finally aroma and flavor.

Beer Color

Beer color is mostly dependent on the color of the malt after roasting and the other

solid grist materials used to make the beer. The color forming molecules in malt and

grist are primarily melanoidins, which were formed during the roasting of the malt by

Maillard browning reactions. During the malt roast, the more intense the killing the

darker the malt and the resulting beer. Sugar content of the malt also determines the

amount of browning that will occur, with higher modified grains having more sugars and

darker color after roasting (Bamforth, 2003).

The second source of color in brewing is the possible oxidation of polyphenols

and tannins. These compounds originate in the malt and hops in beer and if large

amounts of oxygen are allowed into the brewing process further darkening of the beer

will occur. This is similar to polyphenol oxidase reactions in apples, potatoes, and

mushrooms (Bamforth, 2003).









Beer Haze

Haze can be defined as the formation of a colloidal suspension that scatters light

and makes a beverage appear cloudy. In beer there are two classes of hazes: biological

and non-biological. Biological haze is irreversible and results from an infection of beer

by wild yeasts or bacteria, resulting in spoilage. Non-biological haze is defined as being

either "chill" or "permanent" haze resulting from native constituents of the beer. Chill

haze is haze that forms upon cooling beer to 0C and redissolves upon warming to 20C

or more. The term permanent haze should be used for haze which remains in beer at

200C or above.

Composition of Beer Haze

Haze in beer is formed by interactions between proteins and polyphenolic

compounds. Most proteins that are haze-active contain large amounts of the amino acid

proline and originate from the barley protein hordein (Asano et al., 1982). Hordein is a

proline-rich prolamine, or alcohol soluble protein. Because proline has a cyclized side

chain and can form cis bonds, the frequent inclusion of this amino acid in a protein's

sequence gives the protein a more elongated, flexible structure. This allows greater

interactions with polyphenolic compounds and in turn greater haze production. On the

contrary, tighter coiled proteins have a significantly lower affinity for polyphenolic

compounds. More specifically, a 19 kDa proline-rich protein has been found by

immunoelectrophoretic analysis by Asano et al. (1982) to have significant positive effects

on beer haze. Further, Hejgaard and Kaergaard (1983) found a 40 kDa beer protein

involved in both foam and beer haze. In 1993, Kano and Kamimura indicated that both









20 and 40 kDa protein fractions contributed to foam and colloid stability of beer with the

20 kDa fraction correlating more with haze formation than the 40 kDa protein.

Polyphenols can be classified into 2 groups: haze-active and non-haze active

polyphenols. They are classified by the number of binding sites per molecule, with haze-

active polyphenols having two or more sites per molecule. Multiple sites allow the

polyphenol not only to interact with one protein but also to cross-link with other proteins

to create the colloid that results in haze (Seibert and Lynn, 1998). Characteristics of a

haze-active polyphenol are not only multiple binding sites per molecule, but also the

ability to form multiple hydrogen bonds with proteins through the many phenolic groups

on each molecule. A phenol of approximately 1,000 Da usually has between 12 and 16

phenolic hydroxyl groups. This facilitates the interaction of multiple proteins with the

polyphenol (Haslam, 1998).

The polyphenols in beer that are naturally occurring haze-active compounds are

proanthocyanidins. More specifically the most predominant proanthocyanidin dimers in

beer are procyanidin B3 and prodelphinidin B3 (McMurrough et al., 1992).

Formation of Haze

Much work has been done on the dynamics of the polyphenol/protein interactions.

Although the type of protein and polyphenol involved does effect the haze produced,

when one discusses a fixed system such as a certain beer where the native proteins and

polyphenols do not differ, pH exerts a significant effect on haze produced. For beer the

most haze will be seen at pH 4.0-4.2, with less haze forming at lower and higher pHs.

Haze formed at pH 3.0 was only 1/7 of the amount resulting at pH 4.0 (Siebert et al.,

1996b).









The effect of pH directly correlates to what many researchers have noted as the

driving force for protein/polyphenol interactions: hydrophobic effects (Oh et al., 1980;

Siebert et al., 1996b; Haslam, 1998). Hydrogen bond deployment by polyphenols is best

seen as a secondary feature that follows hydrophobically driven associations (Jencks,

1969; Haslam, 1998). The importance of hydrophobic interactions has also been

demonstrated by Oh et al. in 1980 by observing the interactions of tannin and gelatin or

other polyproline proteins. Interactions increased with increases in temperature and ionic

strength as one would expect for hydrophobic interactions.

Seibert and Troukhanova (1996a) developed a model of haze formation in beverage

corresponding to haze-active protein and polyphenol concentrations. At a constant

protein concentration, haze increased to a maximum and then declined as polyphenol

concentration increased. It was hypothesized that this occurred because a haze-active

protein would have a fixed number of sites to which a polyphenol could bind, and a haze-

active polyphenol has two or more sites to bind to a protein. When the concentrations of

proteins and polyphenol binding sites is closest, this results in maximum utilization of

binding sites and cross-linking, resulting in large colloidal particles and maximum light

scattering. However in a matrix such as beer, there is a large excess of haze-active

protein to polyphenols and each polyphenol should be able to find a binding site on each

protein. These protein dimers created would therefore not be cross-linked together and

result in a small colloidal particle size and in turn less haze. However in all instances

Seibert et al. (1996b) affirm that pH, ethanol content, and temperature all affect the extent

of haze formation.









Combinations of proteins with polyphenols generally seem to result from three

different mechanisms. As shown in Figure 2, protein polyphenol interactions can occur

as hydrogen bonding between oxygen atoms of peptide bonds and hydroxyl groups of

polyphenols, hydrophobic bonding between hydrophobic amino acids such as proline,

tryptophan, phenylalanine, tyrosine, leucine, isoleucine, and valine and the hydrophobic

ring structure of polyphenols, and ionic nodding between positively charged groups of

proteins and negatively charged hydroxyl groups of polyphenols. However at the acidic

pH of beer, the first two mechanisms would come into effect, while the third would not

because the hydroxyl groups of polyphenols would have no charge and ionic bonding

could not occur (Asano et al., 1982).

Haze Removal and Beer Stabilization Techniques

There are three types of fining agents used in brewing: those that remove high

molecular weight polyphenols to reduce chill-haze, those that remove high molecular

weight proteins to reduce chill-haze, and those that reduce yeast biomass but do not affect

chill-haze.

The most common polyphenol agent used is polyvinylpolypyrrolidine or PVPP.

Many brewers prefer using a polyphenol scavenger because it does not remove protein,

which could lead to a reduction of foamability. Users of PVPP also benefit from the fact

that because complex polyphenols are being removed, as the beer ages and simple

phenolic compounds complex to form polyphenols, the amount of polyphenol present in

the aged product will remain low. The amount of polyphenol is directly proportional to

the amount of haze present. Therefore even aged beers will have less haze and retain

foam formation if PVPP is used (Fix, 1999).















O H 0 H
,-IN NZ N Protein
H O H 0
Hydrogen "
bonding H

(1) Hydrogen bonding HO

Intermolecular HO Poyphenol
hydrogen bonding CO
0 OH 0 H
H N O NH 0 Protein



Proline
0 o0 H

N N NH N Protein
Intermolecular H 0 '-f P t On
hydrogen bonding :
H
HOr60
(2) Hydrophobic bonding HO
:HO O Polyphenol
Hydrophobic bonding--- :
S OH
OH ,-

Lysine

6 0
'Y"N t Protein
Intermolecular 0 H: 0 H
hydrogen bonding

NH3 +
(3) Ionic bonding Ionic bonding-->
HO
HOtO Polyphenol

POH
OH


Figure 2-2: Molecular Mechanisms for Haze Formation in Beer (Asano et al., 1982)


Although foam-forming ability can be disrupted by some fining agents, silica gels


are the most widely used protein scavengers because they are highly specific for haze-


active proteins. Silica gels work by completing with haze-active proteins as a


polyphenol would, therefore disrupting the formation of a haze (Fix, 1999).


Yeast-active agents are less used and will have no effect on chill haze. The most


common yeast-active agent is Isinglass, which comes from the swim bladder of tropical


fish (Fix, 1999).









Beer Foam

Foam can be defined as a thermodynamically unstable colloidal system in which

gas is momentarily entrapped in a liquid matrix. Foams can be made in two ways: by

supersaturation or mechanically. When a gas is dissolved in a liquid under pressure, such

as in a carbonated beverage, and pressure is released the gas becomes supersaturated and

gas bubbles form creating a foam. These bubbles do not form spontaneously, but instead

form from air pockets that are already present on the side of the container. Another

example of supersaturation foam formation is the formation of the foam structure in

bread where CO2 collects in air pockets in the dough. Foams that are formed

mechanically could be formed by sparging, beating, or shearing. This would be the case

in whipped products like meringues (Walstra, 1996).

Beer Foam Composition

Beer foam is created when supersaturated CO2 is released and forms gas bubbles in

the continuous liquid matrix. This matrix contains amphiphilic proteins in the beer,

which migrate to the surface of the air/liquid interface. This migration is motivated by

the decrease in the proteins' free energy as the protein migrates out of solution to the

interface. The result of this migration is that interfacial tension is lowered, which makes

it more favorable for the two thermodynamically incompatible phases (air and liquid) to

co-exist. The air/liquid interface is stabilized because the proteins have formed a strong

viscoelastic film.

Although all proteins are amphiphilic, they can differ greatly in the their surface

activity, making some proteins better emulsifiers/stabilizers than others. Two

characteristics that govern a protein's ability to act as a surface-active protein are the









properties and topology of the protein's surface and its conformational stability,

flexibility, and adaptability (Damodaran, 1996).

Although an increasing amount of hydrophobic amino acids does mean a protein

could be more surface active, it is the distribution of these amino acids that is the most

important. A protein with clumps of hydrophobic groups will be a better foam stabilizer

than one with randomly dispersed hydrophobic groups. Also, a protein that is in a molten

globule state, or can more easily unfold at the air/liquid surface will be more surface-

active and therefore a better foam former/stabilizer (Damodaran, 1996).

Beer Foam Formation

For surface-active proteins, there is a sequence of events that leads to formation

and stabilization of the foam. First the protein has to have the ability to rapidly diffuse

and absorb to the interface. Secondly, the protein should be able to rapidly unfold and

reorient its polypeptide segments at the interface, and thirdly, while the protein is at the

interface it should be able to interact with neighboring proteins or molecules to form a

strong, continuously cohesive viscoelastic film, able to tolerate mechanical or thermal

forces. The first two steps above are critical for the formation of a foam and the third is

imperative for foam stability.

Beer Foam Stabilization

Foam stability is dependent on the presence of amphipatic polypeptides from

malt, alpha-acids from hops, and the absence of lipophilic materials. Brewers insure

good foam stability by the addition of propylene glycol alginate and the use of nitrogen

gas. Nitrogen works by providing bubbles of very small diameter causing a higher

concentration of small bubbles, which results in a more stable foam. Foam formation and

stability may also affect the flavor and aroma of beer (Bamforth, 2000).









Beer Flavor

Because beer is a delicate and heat labile beverage, off-flavors are easily formed

during pasteurization. In a study by Kaneda et al. (1994), non-pasteurized beer was

compared to pasteurized bottled beer. Off-flavors, volatile aldehydes of bottled beers

pasteurized at 15-30 pasteurization units (P.U.) had a similar off-flavor profile to non-

pasteurized beers stored at 20C for 6-10 days. With "freshness" being top priority to

brewers, it is evident a method of pasteurization using no heat would be of great benefit

to the brewing industry.

When evaluating a new non-thermal pasteurization method, the effect on beer

flavor would need to be evaluated as well. Although, much flavor analysis in the

brewing industry is done by panelists, analysis of flavors by instrumentation is also

noteworthy. One technique that has become increasingly popular is solid phase

microextraction (SPME). This method uses a small volume of sorbent dispersed typically

on the surface of a fiber to isolate and concentrate analytes from a sample matrix. After

either exposure to the headspace or wicking into a liquid sample matrix, analytes will

absorb to the fiber until an equilibrium is reached. Analytes are then thermally desorbed

into an analytical instrument for separation and quantification. It is a solvent-free, non-

destructive, and simple preparation technique (Pawliszyn, 2001).

SPME can be used in conjunction with gas chromatography-olfactometry (GC-O)

to determine what flavor compounds may be of interest in a particular sample, or the

importance of specific compounds in the overall perception of a food's flavor or aroma.

In traditional gas chromatography, a mixture of volatile compounds partition between the

gas and stationary phases and travel at different velocities, resulting in different residence

times within the column. Upon eluting compounds can be characterized using various









detectors (Kamimura et al., 2003). In GC-O, the separation capabilities of gas

chromatography are coupled with the sensitivity and precision of the human nose. This

method of flavor analysis detects compounds that have an influence on flavor or aroma of

the food (Naim et al., 1998). GC-O produces an aromagram which indicates an olfactory

impression as a function of retention time plus an intensity descriptor (Kamimura et al.,

2003). GC-O has been used with success to describe hop aromas, off-flavors, and aged

beer flavor compounds, using both immersion and headspace SPME sampling

(Kamimura et al., 2003).

Processing of Beer

Pasteurization

Most beer, packaged in bottles or cans, is pasteurized after filling into containers by

passing through a steam tunnel. During this process, the bottles are passed under a series

of water sprays with the temperature of the water increasing as beer passes through the

tunnel. After reaching the desired temperature, bottles are cooled by a water spray and

then exit the tunnel to air dry. The heat treatment used is referred to in terms of

pasteurization units, with 1 PU = 600 C for 1 minute, with 5 PU resulting in a sufficient

kill of approximately 102/ml yeast cells that would remain in pre-filtered beer. The

steam tunnel used for bottles or cans can be operated at various temperatures, usually

600C or 620C for a longer time, depending on operating needs. Up to 30 PU may be

applied depending on pre-pasteurization procedures such as filtering. Because heat

treatment could adversely affect flavor, the choice of pre-pasteurization treatments and

PU value is always a compromise between extending shelf life and beer quality

(Campbell, 1997).









Flash pasteurization

Although flash pasteurization is not common in North American breweries, it is

very popular in Europe and Asia. During flash pasteurization, beer is heated to at least

71.5 to 740 C and held for 15 to 30 seconds, resulting in 13.26 PU. This is achieved by

the use of a two- or three-stage plate heat exchanger with hot water as the heat source.

The adjustment of the flow rate determines the PU value for the treatment. After flash

pasteurization, beer is then aseptically packaged into sterilized bottles or cans.

Nonthermal methods

The most common non-thermal processing technique used by brewers to extend

shelf life is sterile filtration. Sterile filtration has been used as an alternative to

pasteurization for many years. It has the advantage over pasteurization in that the risk of

flavor damage by heat is eliminated.

The term "sterile filtration" refers to the reduction of yeast and bacteria to levels

that do not result in spoilage of the beer over its planned shelf life by use of one or many

filters before packaging the beer in sterile containers. This can be accomplished without

the loss of color or flavor compounds (Goldammer, 2000).

The brewer sets a specification for the maximum allowable concentration of yeast

and bacteria for quality control purposes, which may have entered the brewing process

inadvertently, in sterile-filtered beer since not all microorganisms will be removed during

the process. There are no critical levels for allowable microorganisms, and therefore

extensive monitoring of the process must occur. The process is also labor intensive

because of the time needed to clean the filters to prevent fouling (Goldammer, 2000).









Effects of Dense-Phase CO2 Pasteurization

Microbiology

Recently there has been great interest in inactivation of microorganisms in food or

model food systems using high pressure carbon dioxide. This new technology has been

shown to inactivate microorganisms as well as conventional heat pasteurization without

the loss of nutrients or quality changes.

The use of carbonation as a means of preserving food started as early as 1939

with the study by Brown et al.(1939) where apple cider was carbonated and microbial

inactivation and flavor changes were recorded. The carbonation of the juice was shown

to preserve the cider for up to 3 months at approximately 21C with no change in flavor.

The use of carbonation was also investigated for its use in soft drinks as a preservation

agent. Even at the lowest amount of gas pressure (3 volumes of CO2 where 1 volume=l1

L of CO2 per L of beer) sterility was achieved on approximately the 20th day depending

on the Brix of the beverage (Insalata, 1952). Researchers have also evaluated the use of

pressurized CO2 and decompression to reduce microbial loads. In 1951, Fraser showed

that 99% of E. coli numbers were rendered non-viable by a decompression of CO2 from

500 psi to atmospheric pressure.

Although carbonation with CO2 has been shown as an affective preservative some

bacteria are not affected. Molin (1983) examined the growth inhibiting effect of carbon

dioxide on a variety of food related bacteria by sparging spiked growth media with the

gas, arriving at a variety of pressures. Only partial success was attained. Although 100%

carbon dioxide did slow the growth of all organisms some were affected less than others.

CO2 had approximately 75% inhibitory affect on Bacillus cereus, Brochothrix

thermosphacta, and Aeromonas hydrophila, and a 53%-29% inhibitory effect on









Escherichia coli and Streptococcus faecalis. Inhibitory rates for anaerobic bacteria were

even lower. This discovery proved that carbonation of foods alone would not inactivate

all food-related bacteria.

Because of the need for a preservation method that is safe, inexpensive, and non-

destructive to heat sensitive compounds, the use of supercritical carbon dioxide (SC-C02)

was tested as a food preservation method on cells in cultures or broths. SC-CO2 was

chosen because of its safety, cost, and high purity. CO2 also has a low critical pressure

and critical temperature which result in excellent solvent power of CO2 when used in

HPCD pasteurization. Kamihira et al.(1987) used HPCD to sterilize cultures of baker's

yeast, Escherichia coli, and Staphylococcus aureus, however they were only successful

with cultures with moisture contents from 70-90%. This same success was echoed in

1991 and 1992 when Saccharomycces cerevisiae was inactivated using sub and

supercritical CO2 and both cell inactivation and disintegration were studied. (Lin et al.,

1991; Lin et al., 1992a, 1992b; Nakamura et al., 1994) A similar study was performed

again using baker's yeast and Bacillus magetarium in a spore form (Enomoto et al.,

1997a, 1997b) However in all studies, in striving for not only the inactivation of cells

but also their disintegration, the vessel must undergo pressurization and depressurization,

in some cases, several times creating an energy-intensive process. However, this is only

true in batch systems. In a continuous system depressurization would happen naturally.

Similar studies have been performed using Leuconostoc (Lin et al., 1993) and

Kluyveromycesfragilis, Saccromyces cerevisia, and Candida utilis (Isenschmid et al.,

1995) showing that inactivation can occur without disruption of the cell wall, but instead

through the leaching of cellular contents by the CO2. The effect on HPCD on bacterial









spores has also been examined with the most successful inactivation occurring in the

subcritical region of CO2 (Enomoto et al., 1997a).

It has been shown that the amount of inactivation is proportional to the amount of

dissolved CO2 in the sample. Kumagai et al. (1997) measured dissolved CO2

gravimetrically. Results showed higher inactivation of Saccharomyces cerevisiae as CO2

levels increased. Shimoda et al. (2001) found similar results and were able to model

death kinetics of Saccharomyces cerevisiae as first-order through the critical temperature

and pressure of CO2. Additionally, higher water activities and higher pressures have also

shown higher inactivation because of their effects on CO2 sorption of the yeast cells

(Kumagai et al., 1997). To further enhance this effect not only have higher pressures

been used but also modifications to the necessary machinery have taken place. In studies

by Shimoda et al. (1998) and Ishikawa et al. (1995, 1997), a filter was placed in the

process vessel to create microbubbles of CO2 entering the vessel. It was shown that these

microbubbles were more effective in inactivating bacterial cultures than the process

without the filter. Another innovation has been the development of a semi-continuous

system for dense-phase CO2 pasteurization. One such system was shown to be more

efficient on the inactivation of Saccharomyces cerevisiae (Spilimbergo et al., 2003b).

Conditions such as temperature and pressure have also shown significant effects

on the antimicrobial properties of CO2. Studies on E. coli in Ringers solutions showed a

decrease in survival rate with increases in dissolved C02, increases in pressurel.2-5 MPa,

and increases in temperature from 250-450C (Ballestra et al., 1996). Studies on Listeria

monocytogenes also showed similar results in reference to pressure and temperature









variations with a complete inactivation at 6.08 MPa CO2 treatment at 115, 75, and 60

minutes at 25, 35, and 450 C, respectively (Erkman, 2000a, 2000b).

Recently most studies have concentrated on the inactivation of microorganisms in

food samples rather than cultures or broths. Studies by Wei et al.(1991) showed the

effectiveness of using CO2 on cultures of Listeria or Salmonella suspended in water and

spiked onto food samples. The study showed the treatment to be applicable in some food

systems: chicken, shrimp, orange juice, and egg yolk, but ineffective in a whole egg-

salmonella mix, proving this technology may be dependent on the foods characteristics

and their ability to shelter bacteria from of carbon dioxide. Staphylococcus aureus

suspended in broth and compared to a suspension in raw milk was shown to undergo

inactivation at lower pressures and in a shorter length of time. Also compared was raw

milk was compared to orange, peach and carrot juices and the milk showed protection of

the bacteria (Erkman, 1997, 2000c).

Kimchi has also been studied because of its unique properties as a fermented food

which may spoil if fermentation is not stopped at the appropriate time. In two different

studies by Hong et al. (1997, 1999), Lactobacillus was shown to undergo inactivation

after 200 min at 6.86 MPa CO2 pressure in kimchi compared to 30 min at approximately

13.73 MPa in the broth suspension. This affirms the point that the success of HPCD

should be evaluated on a case-by-case basis for different foods.

Theories of Cell Death by Dense-Phase CO2 Pasteurization

Researchers have reported that the inactivation rate of all microorganisms is

sensitive to pressure, temperature, and exposure time to CO2. Furthermore Hong et al.

(1999) found similar results, but also concluded microbial inactivation was mainly due to









the transfer rate of CO2 into the cells which could lead to viability loss. To further

elucidate the mode of cell death by dense-phase CO2 pasteurization, much research has

been performed. Using scanning electron micrographs, Ballestra et al. (1996) noted cell

deformation of Escherichia coli after processing at 350C and 5MPa of CO2 and

Spilimbergo et al. (2003a) noted the same results on gram-positive and gram-negative

bacteria. Shimoda et al. (2001) studied the effect of the concentration of CO2 on

Saccharomyces cerevisiae and found that cell death during continuous versus batch

treatments was due to the "anesthesia effect" of CO2. This effect can be defined as loss

of cell viability because of the diffusion of molecular CO2 into the plasma membrane of

the cell, compromising the construction of membrane domains (Isenschmid et al., 1995).

More specifically, while coining the term "anesthesia effect", Isenschmid et al. (1995)

found that this effect occurred at temperatures higher than 180C. At these higher

temperatures, dependency was also noted on the dissolved CO2 concentration, with

increases in temperature and dissolved CO2 concentration resulting in increased cell

death. However, below 18C a solvent effect was observed as the reason for viability

loss in yeast cells (Isenschmid et al., 1995).

Effects of Cosolvents

Because the solvent effect is suspected as a mode of cell death during dense-phase

CO2 processing, one must also examine the effects of co-solvents such as ethanol that

may be present in the food matrix. Solvent characteristics of CO2 can be greatly

modified by the addition or existence of certain cosolvents. CO2 is a non-polar solvent,

and has limited affinity for polar solutes. It is often used to extract organic solute

molecules. A polar modifier can be added to CO2 to improve the solubility and









selectivity for polar molecules. Most often the cosolvents of choice are lower alcohols.

Cosolvents are usually added in 5-10% amounts by volume and even in these small

amounts, can have significant effects, especially in surface processes. For example, the

addition of ethanol may have significant effects on extraction by allowing increased

absorption on surface sites, preventing the re-adsorption of the compound of interest

(Clifford and Williams, 2000).

The addition of a cosolvent modifies the critical temperature and pressure of the

original fluid (Taylor, 1996). Therefore the solubility of the materials of low volatility is

enhanced and as a result, lower pressures can be used to achieve the same extraction yield

(Brunner and Peter, 1982). This has been shown to increase the amount of oil extracted

from sunflower seeds (Raghuram Rao et al., 1992), and rape seed, soy bean, cottonseed,

flax seed, and peanuts, and to enhance the extraction of herbal components such as

borage seed oil and hiprose fruit (Illes et al., 1994).

Quality Attributes

Although the use of HPCD as a pasteurization technique for foods has been well

studied, its effect on food's quality characteristics needs more examination. During the

injection of CO2 the pH of the foodstuff drops dramatically. Because pH plays such a

central role in regulation of food systems, it is imperative to be able to predict what will

happen to quality attributes of food after HPCD processing. Models for pH of food

systems processed with HPCD have been developed to help forecast pH extremes which

may occur during processing (Meyssami et al., 1992). Specific research has also been

done with single strength orange juice to examine HPCD effect on enzymes, cloud, color,

and Brix value, and total acidity. It was shown that although significant pH changes can

occur during processing, the final pH of the product after de-pressurization was not









significantly changed, cloud was enhanced, enzymes were inactivated, and flavor and

aroma were unaffected (Arreola et al., 1991). Commercially, the addition of CO2 has

also been used to preserve cottage cheese with no effect on the mouthfeel or flavor of the

product (Mermelstein, 1997). Flavor changes and protein viability of model systems

have also been examined after rupture of yeast cells (Lin et al., 1991).

Dense-Phase CO2 Pasteurization of Beer

The need for inhibition of pathogen/spoilage organism growth and the

characteristic carbonation of beer lend it to be an excellent candidate for HPCD

pasteurization. Freshness is of utmost importance in the brewing industry. Current

evidence of this is the use of colored glass bottles, refrigerated distribution houses and

trucks, and the popularization of the "born-on" date.

Although there is promise in the use of HPCD pasteurization technology with

beer, research must be done to insure the technical and economic feasibility of the

procedure. Parameters that must be optimized are the process pressure, amount of CO2

to add, process temperature, and residence time of the product under pressure. These

parameters will be evaluated first on the basis of microbial log reduction and then on

their effects of the quality attributes of taste, aroma, foam formation and stability, and

haze. Therefore, a study examining these characteristics will be conducted in order to

compare the process with current pasteurization methods and to make inferences

regarding the intermolecular changes occurring during the high pressure carbon dioxide

pasteurization.

Objectives

1. To quantify and elucidate mechanisms of inactivation of yeast by dense-phase CO2
pasteurization of beer as a function of process pressure, temperature, residence
time, and CO2 percentage






27


2. To prove there are no significant changes in flavor and aroma of beer after dense-
phase CO2 pasteurization

3. To prove there are no significant changes in chill haze formation of beer after
dense-phase CO2 pasteurization

4. To prove there are no significant changes in foam formation and stability after
dense-phase CO2 pasteurization

5. To evaluate consumer and industry acceptability of dense-phase CO2 pasteurization

















CHAPTER 3
MATERIALS AND METHODS

The Dense-Phase CO2 System

The continuous dense-phase CO2 system was constructed by APV (Chicago, IL)

for Praxair (Chicago, IL) and given to the University of Florida (Gainesville, FL). The

system is housed in the pilot plant of the Food Science and Human Nutrition Department

at the University of Florida. The system mixes cooled, pressurized liquid CO2 with a

liquid feed pressurized by its own pump (Figure 3-1). The mixture then proceeds through

a holding tube (79.2 m, 0.635 cm ID) for a specified residence time, which is modified by

changing the flow rate of the mixture. In the holding tube, temperature can be controlled

by electrical heating tape, insulation, and a controller system, and operating pressure is

maintained. When exiting the holding tube the mixture is depressurized by passing

through a back pressure valve and is then ready for collection.

The machine works by pressurizing the liquid feed first, to 6.89 MPa using a

reciprocating pump with a stroke length of 30 mm and a back pressure valve and then to

the desired operating pressure using a second reciprocating pump and another back

pressure valve. CO2 is then introduced by a reciprocating pump at approximately 6.2

MPa or higher. Pressurizing both the liquid feed and the CO2 insures both will remain

liquid and mix with the desired proportions. The C02/feed mixture then passes through

the second reciprocating pump that maintains the operating pressure which is always 6.89









MPa or more. This pressure is maintained throughout the holding tube. The mixture

then exits the holding tube, is depressurized, and collected in sterile bottles as aseptically

as possible, using alcohol to sterilize the end of the holding tube.




Pump

0 C02
(r C^ iChille-r 9
SHold tube





Heating / --
system '

Pump
Juice ( Treated
stream juice


Figure 3-1: Schematic of the Continuous Dense-Phase CO2 Pasteurization System


Pressure, temperature of the holding tube, flow rate (which in turn controls

residence time in the holding tube), and weight % of CO2 were all controllable

independent variables. Thermocouples and pressure sensors were also located

throughout the machine to monitor operating parameters.

Beer Samples

Fresh beer samples (less than one month old) were purchased from Market Street

Pub in Gainesville, FL. The beer was purchased in 58.7 L (15.5 gallon) kegs and

transported to the pilot plant in the Food Science and Human Nutrition Department at the

University of Florida and stored at 1.670C until used. The beer is an ale brewed using an









all-barley malt extract and whole hops which insured a high level of consistency from

batch to batch. The clarity of the beer had not been improved by the use of brewing aids.

Experimental Design

Cleanability Study

Experiments were conducted to evaluate the cleanability of the dense-phase CO2

system. Principal and Oxonia (27.5% hydrogen peroxide, 5.8% peroxy acetic acid,

66.7% inert ingredients) (Ecolab, St. Paul, MN) were used as sanitizing agents. A non-

pathogenic spoilage organism Lactobacillusfermentum (ATCC, Manassas, VA) was used

in cleanability experiments because of its frequent use as a test organism for thermal

inactivation studies. A freeze dried culture of Lactobacillusfermentum was rehydrated in

25 mls Lactobacillus MRS broth (Difco Laboratories, Sparks MD) and incubated at 37C

for 24 hours. Butterfield's phosphate buffer was used to dilute this culture to 106

CFU/ml.

The machine was sanitized by pumping Principal at 50 ml per 18.9 L hot tap water,

followed by 18.9 L cold tap water rinse and then followed by Oxonia at 946 ml perl8.9 L

cold water. Some Oxonia solution was left in the system overnight. The next day, 6 L of

sterile water was used to rinse the machine after the remaining Oxonia had been pumped

through. Then 6 L of the cell suspension was pumped through the sanitized machine,

collected, and plated to measure recovery of the microorganisms. The machine was then

rinsed with water and re-sanitized. The machine was then allowed to sit for 2 hours.

Then 6 L sterile water followed by 6 L of sterile Butterfield's phosphate buffer were run

and collected into sterile bottles. The collected buffer was filtered using a vacuum pump

and .45 micron mixed cellulose water testing filters (Fischer Scientific, Pittsburg, PA).

1.2 L of buffer was filtered per filter and all filters were plated on Lactobacillus MRS









agar (Difco Laboratories a subsidiary of Becton, Dickinson, and Company, Sparks MD).

Plates were incubated overnight at 37C.

The recovered cell suspension plates showed that 105/ml lactobacilli were

recovered after pumping the suspension through the machine, i.e. approximately 900,000

CFU remained in the machine after the cell suspension exited. After the sanitizing

procedure was completed, collection of buffer pumped through the system resulted in no

CFUs on the plates for the filters. Results show that the Principal and Oxonia solutions

were effective in cleaning the machine without pressure.

Experimental Design

A Central Composite Design (CCD) was used because it is an economical design,

allowing a researcher to fit a second order prediction equation to a response surface. The

independent variables were pressure, CO2 %, residence time, and temperature. Twenty-

seven treatment combinations were selected by CCD. The aim was to establish an

optimum set of operating conditions based on the dependent variable of yeast population

reduction.

More experiments were conducted after this one on a subset of the 27 points to

evaluate the effect on the quality attributes of haze, foam, and aroma/flavor changes. The

subset was selected from the 27 points by choosing the most effective combination

treatment for log reduction of yeast, a more economical version of the previous, and

adding a heated beer sample (740C, 30 seconds), and an untreated, fresh beer control.

The experimental design is shown in Table 3-1.










Table 3-1: Ex erimental Desi n_
Pressure Temp C02 Time
0.000 -1.414 0.000 0.000
-1.000 -1.000 -1.000 -1.000
1.000 -1.000 -1.000 -1.000
-1.000 -1.000 -1.000 1.000
1.000 -1.000 -1.000 1.000
-1.000 -1.000 1.000 -1.000
1.000 -1.000 1.000 -1.000
-1.000 -1.000 1.000 1.000
1.000 -1.000 1.000 1.000
0.000 0.000 -1.414 0.000
0.000 0.000 0.000 -1.414
-1.414 0.000 0.000 0.000
0.000 0.000 0.000 0.000
0.000 0.000 0.000 0.000
0.000 0.000 0.000 0.000
1.414 0.000 0.000 0.000
0.000 0.000 0.000 1.414
0.000 0.000 1.414 0.000
-1.000 1.000 -1.000 -1.000
1.000 1.000 -1.000 -1.000
-1.000 1.000 -1.000 1.000
1.000 1.000 -1.000 1.000
-1.000 1.000 1.000 -1.000
1.000 1.000 1.000 -1.000
-1.000 1.000 1.000 1.000
1.000 1.000 1.000 1.000
0.000 1.414 0.000 0.000


Pressure (MPa)
-1=20.7
0=27.6
1=34.5
1.414214=37.3
-1.414214=17.8

Co2 (%)
-1=8
0=10
1=12
1.414214=12.83
-1.414214=7.17

Temp (degrees C)
1-=25
0=35
1=45
1.414214=49
-1.414214=21

Residence
Time (minutes)
-1=4
0=5
1=6
1.414214=6.41
-1.414214=3.56


Procedures

For every experiment, the dense-phase CO2 system was sanitized one day prior to

experiments by the procedure described in the cleanability experiment. Then the

machine was rinsed with 6 L of sterile de-ionized water. As the feed tank emptied, beer

was added and allowed to run in the machine without pressure for at least 2 hold-up

volumes (5 L) to insure all water had exited the machine. Operating parameters were set

for each treatment and 2 hold-up volumes (approximately 5 L) of beer were allowed to

pass to insure steady state conditions before sample collection. Samples were collected

in sterile bottles. The 27 treatment combinations were run in duplicate on separate days,

using different batches of beer.









For heat pasteurized beer, fresh beer was pumped by a peristaltic pump (Cole

Parmer, Chicago, IL) into a temperature controlled water bath (Hart Scientific, American

Fork, UT) through copper tubing. The beer was pasteurized at 740 C for 30 seconds and

collected in a sterile bottle. Copper tubing was sanitized using pumping soapy hot water,

then hot water, then alcohol through the system prior to use.

Analysis of Treated Samples

Microbial Reduction Experiments

All beer samples were plated on potato dextrose agar for enumeration of yeast

colonies. Beer was diluted serially using 90 ml dilution bottles containing Butterfield's

phosphate buffer (Hardy Diagnostics, Santa Maria, CA). Dilutions were done in

duplicate and plated in duplicate. Plates were incubated at 37C for 5 days.

Haze Measurement

Beer samples were evaluated for haze using a Hach turbidimeter (Hach, Loveland,

CO) and the AOAC procedure for beer haze measurement (AOAC 10.013). Readings

were recorded as NTU, nephelometric turbidity units.

Foam Capacity and Stability Measurements

Beer samples (25 ml) were degassed at room temperature for 24 hours by using a

stir plate and stir bar, and placed in graduated cylinders with their tops cut off to facilitate

the addition and removal of the homogenizer blade. A Virtus 45 homogenizer (The

Virtus Company, Gardiner, New York) was used to mechanically foam the beer at setting

20 for 60 seconds, keeping the blade at an equal depth for all beer samples. At the end of

the 60 seconds the homogenizer was turned off and the volume of foam produced in ml

was recorded, as an indication of foam capacity. Sixty seconds after that, the amount of









beer liquid collected in the bottom of the cylinders was recorded as well, as an indication

of foam stability (Odabasi, 2003).

Polyacrylamide Gel Electrophoresis of Beer Proteins

Polyacrylamide gel electrophoresis (PAGE) was run on beer proteins to determine

if beer proteins were effected by dense-phase CO2 pasteurization. Beer samples (lml)

were centrifuged at 1.60C at 10,000g for 15 minutes in 30 kDa Millipore Centricon

Centrifugal Filter Devices (Millipore, Bedford, MA) to concentrate beer proteins. The

retentate was collected and mixed 1:1 with PAGE sample buffer containing dye (Biorad,

Hercules, CA) and kept on ice. 30 uml of each sample was then loaded onto a pre-cast

polyacrylamide gel (Biorad, Hercules, CA) and the gels were run at 100 V until dye

bands had reached the end of the gel. A molecular weight marker was also loaded on

each gel as a known standard (SigmaMarker, Sigma Aldrich, St. Louis, MO). Samples

were run in duplicate on 15% and 18% polyacrylamide gels. Differences in proteins

were evaluated by Rf values (mm traveled by band/mm traveled by dye front).

Flavor

Gas chromatography-olfactometry was performed on beer samples using solid

phase microextraction (SPME) as the sample preparation method. The fiber used was

100um polydimethylsiloxane coating SPME fiber assembly (Supelco, Bellefonte, PA).

Flavor profiles of fresh and processed beer were created by headspace SPME sampling.

Fiber exposure time was optimized prior to sampling. Aliquots (7ml) of fresh beer

were poured into 40 ml vials and each sealed with caps containing Teflon-coated septa.

Volatiles were subsequently extracted using a pre-conditioned 100 |tm PDMS fiber

(Supelco, Bellafonte, PA) under different extraction conditions, which varied in time (5,









10, 15 min) and temperature of fiber exposure (30 and 400C). Equilibrium was reached

at 400C for 15 minutes. Chromatographic data is available by contacting Dr. Murat

Balaban at the Food Science and Human Nutrition Department, University of Florida.

For sampling, the whole coated fiber was exposed to the headspace of the samples

and after the extraction conditions were completed, the fiber was removed from the

headspace and immediately inserted into a GC-splitless injector, where aroma

compounds were allowed to be desorbed for 2 min.

Volatile components were separated in a HP-5890 GC (Palo Alto, CA) equipped

with a sniffing port (DATU, Geneva, NY), a flame ionization detector (FID), and a ZB-5

column (30 m x 0.32 mm i.d. x 0.5 mm film thickness) from J&W Scientific (Folsom,

CA) and in the same model GC with a Carbowax column 30 m x 0.32 mm i.d x 0.5 um

film thickness). For both columns the initial oven temperature was 400C which was then

increased at 200C/min to a temperature of 1200C. The temperature was then increased at

5C/min to a temperature of 1600C, and then increased to 2400C at 150C/min and held at

this final condition for 5 min. Injection and detection port temperatures were 250 and

2500C, respectively. A 0.2 p.L aliquot of alkane standard solution was also injected in the

splitless mode. A GC splitter split the column effluent between the FID and the

olfactometer in a 1:2 ratio, respectively. Two trained assessors (training based on aroma

active compounds present in beer) were employed to evaluate each treatment in

duplicate. Each assessor was asked to describe each odor detected in the GC-O effluent

and to indicate the aroma intensity continuously during the chromatographic run using a

linear potentiometer. This device has a pointer that can be moved across a 10-cm span to

indicate aroma intensity. Aroma descriptors along with their respective retention times









were recorded manually, and later transcribed into the chromatographic software for

inclusion with the olfactometry time-intensity data. Time-intensity aromagrams were

obtained for each treatment. Aroma-active compounds were defined as only those

compounds producing an intensity response at the same retention time and similar

descriptor from at least half of the panel responses. Mean aroma intensities of each

aroma-active compound were calculated by averaging the peak height among each

chromatographic run. Chromatograms were recorded, compounds tentatively identified

by comparing LRI values and descriptors, and their correspondent peak areas integrated.

Beer samples were also evaluated by GC-MS using the same chromatographic

conditions as above to further aid in peak identification. A ZB-5 column (60 m x 0.25

mm i.d x 0.25 um film thickness) was used. Mass spectra were matched with several

flavor libraries of mass spectra for identification using Xcaliber Software, Version 1.3.

Sensory

Sensory analysis was done to compare the aroma and flavor profiles of fresh beer

to dense-phase CO2 processed beer and dense-phase CO2 processed beer to heat

pasteurized beer, and a ranking test was used to rate the likability of different dense-

phase CO2 processed beer treatments. A randomized complete block design was used

(Ott, 1993) and difference from control measurements were recorded using Compusense

on a line scale with anchors at 0 and 10 of "no difference" and "extremely different"

(Compusense, Guelph, Ontario, Canada). All sensory tests were performed in the taste

panel facility in Building 120, University of Florida, Gainesville, Florida which consisted

of privacy booths for each panelists. The relative humidity was approximately 60% and

the room temperature was between 23-250C. Forty-five to sixty untrained panelists were









used in each test and selected on the basis of age (21+ years or older) and familiarity with

beer and willingness to sign the informed consent document. No incentive was given to

participate, however snacks were available afterwards to counteract any effects of the

alcohol on panelists.

Samples were labeled with random numbers and were presented on a white plate

with the reference sample at the top center. In the next row the first 3 samples from left

to right, and in the bottom row the last 2 samples from left to right. Samples were

degassed overnight by placing on a hot plate with stirring to equalize carbonation levels

and were served at room temperature.

Data was assumed to be normal and an analysis of variance was performed using

SAS Software (The SAS Institute, Cary, North Carolina). Tukey's mean separation

(alpha=. 1) tests were also performed on SAS if there was a significant difference in mean

to examine mean separations.

Fresh beer vs. dense-phase CO2 processed beer

A difference from control test was used with the reference being fresh beer and the

samples consisting of a hidden control of fresh beer, and three different dense-phase CO2

processed beer treatments. Panelists were asked to first rate the intensity of the

differences in the aroma of the samples and then to rate the intensity of the differences in

the flavor of the samples.

Dense-phase CO2 processed beer vs. heat pasteurized beer

A difference from control test was used with the reference being fresh beer and the

samples consisting of a hidden control of fresh beer, and one dense-phase CO2 processed

beer treatment (26.7MPa, 21C, 10% C02, 5 minute residence time), and one heat

pasteurized beer sample (740C, 30 seconds). Panelists were asked to first rate the









intensity of the differences in the aroma of the samples and then to rate the intensity of

the differences in the flavor of the samples.

Storage studies

All sensory tests were repeated after 30 days of storage at 1.67C.

Statistical Analysis

Sensory data were gathered by the Compusense system (Compusense, Guelph,

Ontario, Canada), and data were pooled and analyzed using Excel spreadsheets for

Windows and SAS software Version 9.0 (The SAS Institute, Cary, North Carolina). An

analysis of variance was done on each set of data and a Tukey's mean separation was

performed. A 90% confidence level was used.

Conjoint Analysis of Beer Purchase Decision

A conjoint analysis was performed to determine the part-worth values of price,

flavor, and shelf stability of beer in a consumer's purchase decision. This was done using

a full factorial of combinations created from the characteristics of bottled vs. draft flavor,

refrigerated vs. shelf stability, and $6.00/six 12 oz. bottles vs. $8.00/six 12 oz. bottles.

These eight combinations were presented to consumers (21 and older) and consumer

were asked to rank the combinations in order of purchase choices without ties. Part-

worth values for each level of each attribute were then calculated. The higher the part-

worth value the more influence the level of a characteristic has on purchase decision

(Hair, et al., 1998).















CHAPTER 4
RESULTS AND DISCUSSION

This chapter contains experimental results obtained for the dense-phase CO2

pasteurization of beer, presented in five sections, namely yeast log reduction, foam, haze,

sensory evaluation, and gas chromatography results.

Microbial Reduction Experiments

Yeast plate counts are listed in Table A-i in Appendix A. Yeast log reductions for

control and treated samples are presented in Table 4.1. A prediction equation was

calculated using the log reduction responses and response surface methodology, where

xi=pressure in MPa, x2=temperature in C, X3=CO2 %, and x4=residence time in minutes

(Khuri and Cornell, 1996). SAS software was used using the code in Appendix C. The

prediction equation in coded variables is:

Predicted Log Reduction = 5.16 -.135xl -.80x2 -.043x3 +.05x4 -1.17xl2 +.514(x2xl)
+1.28(x22) +.012(x3xl) +.734(x3x2) -.971(x32) -.077(x4xl) + .852(x4x2) -.087(x4x3) -
.756(x42)


And in non-coded variables the prediction equation is:

Predicted log reduction =.653 +.553xl -1.016x2 -1.860x3 +2.623x4 -.012xl2
+.004(x2xl) +.007(x22) +.0004(x3xl) +.019(x3x2) -.121(x32) -.006(x4xl) + .043(x4x2)
-.022(x4x3) -.372 (x42)
The ANOVA table used to generate these results is:

The RSREG Procedure
Coding Coefficients for the Independent Variables
Factor Subtracted off Divided by
X1 27.550000 9.750000
X2 35.000000 14.000000













10.002500
4.985000


2.827500
1.425000


Response Surface for Variable Y

Response Mean 4.572981
Root MSE 0.651957
R-Square 0.7307
Coefficient of Variation 14.2567


Type I Sum
DF of Squares


13.142063
19.328234
12.515434
44.985731


R-Square F Value Pr > F


0.2135
0.3140
0.2033
0.7307


7.73
11.37
4.91
7.56


0.0001
<.0001
0.0008
<.0001


Sum of
DF Squares

39 16.576879


Parameter


Parameter DF Estimate


Intercept
X1
X2
X3
X4
X1*Xi
X2*X1
X2*X2
X3*X1
X3*X2
X3*X3
X4*X1
X4*X2
X4*X3
X4*X4


0.653046
0.552928
-1 .015581
1.860443
2.623033
-0.012258
0.003771
0.006516
0.000423
0.018548
-0.121398
-0.005533
0.042709
-0.021641
-0.372491


Standard
Error

7.574578
0.223678
0.144890
0.884747
1.754089
0.003265
0.001670
0.001573
0.008352
0.005763
0.038833
0.016703
0.011525
0.057625
0.153716


t Value Pr > |t|


0.09
2.47
-7.01
2.10
1 .50
-3.75
2.26
4.14
0.05
3.22
-3.13
-0.33
3.71
-0.38
-2.42


0.9317
0.0179
<.0001
0.0420
0.1429
0.0006
0.0296
0.0002
0.9598
0.0026
0.0033
0.7422
0.0007
0.7093
0.0201


Sum of
Factor DF Squares Mean Square F Value Pr > F


8.623020
32.363641
8.654597
8.468728


1.724604
6.472728
1.730919
1.693746


4.06 0.0046
15.23 <.0001
4.07 0.0045
3.98 0.0051


Canonical Analysis of Response Surface Based on Coded Data


Critical Value


Factor


Coded


-0.009502
0.241912
0.061900
0.166366


Uncoded


27.457356
38.386764
10.177523
5.222071


Regression

Linear
Quadratic
Crossproduct
Total Model


Residual

Total Error


Mean Square

0.425048


Estimate
from Coded
Data

5.161059
-0.134616
-0.800175
-0.042847
0.050220
-1.165235
0.514805
1.277046
0.011674
0.734228
-0.970548
-0.076869
0.852052
-0.087194
-0.756389









41





Predicted value at stationary point: 5.067764


X1

0.093337
-0.159294
0.383442
0.904923


Eigenvectors
X2


0.967866
-0.097086
-0.147191
-0.179288


0.144464
-0.391092
0.863977
0.282347


0.183449
0.901247
0.291290
0.263153


Stationary point is a saddle point.

Estimated Ridge of Maximum Response for Variable Y


Standard
Error

0.210737
0.210093
0.208419
0.206472
0.205532
0.207389
0.214171
0.227957
0.250290
0.281894
0.322772


Uncoded Factor Values
X2


27.550000
27.412807
27.299691
27.195846
27.096467
26.999570
26.904196
26.809822
26.716139
26.622955
26.530142


35.000000
33.620157
32.245753
30.877347
29.512775
28.150655
26.790168
25.430817
24.072285
22.714361
21.356901


10.002500
9.977017
9.942388
9.904840
9.866025
9.826552
9.786696
9.746596
9.706333
9.665955
9.625492


4.985000
4.981320
4.965593
4.945407
4.923116
4.899667
4.875515
4.850902
4.825973
4.800817
4.775491


Eigenvalues


1.437413
0.776571
1.050380
1.225588


Coded
Radius

0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0


Estimated
Response

5.161059
5.255745
5.378149
5.528925
5.708271
5.916267
6.152952
6.418347
6.712462
7.035307
7.386885















Table 4-1: Microbial Reduction Results
(Independent Variables are in Coded Units)


Pressure Temp C02 Time Log Reduction Replicate Log Reduction
0.000 -1.414 0.000 0.000 6.11 6.06
-1.000 -1.000 -1.000 -1.000 6.11 6.06
1.000 -1.000 -1.000 -1.000 6.11 6.06
-1.000 -1.000 -1.000 1.000 6.11 6.06
1.000 -1.000 -1.000 1.000 4.71 4.18
-1.000 -1.000 1.000 -1.000 6.11 6.06
1.000 -1.000 1.000 -1.000 4.41 4.18
-1.000 -1.000 1.000 1.000 4.24 3.96
1.000 -1.000 1.000 1.000 4.24 3.96
0.000 0.000 -1.414 0.000 3.19 3.28
0.000 0.000 0.000 -1.414 3.63 3.58
-1.414 0.000 0.000 0.000 3.46 3.40
0.000 0.000 0.000 0.000 4.71 4.66
0.000 0.000 0.000 0.000 6.11 6.06
0.000 0.000 0.000 0.000 6.11 6.06
1.414 0.000 0.000 0.000 3.93 3.81
0.000 0.000 0.000 1.414 4.71 4.36
0.000 0.000 1.414 0.000 4.71 4.18
-1.000 1.000 -1.000 -1.000 2.89 3.21
1.000 1.000 -1.000 -1.000 3.06 3.38
-1.000 1.000 -1.000 1.000 4.01 3.76
1.000 1.000 -1.000 1.000 4.11 3.81
-1.000 1.000 1.000 -1.000 3.37 3.30
1.000 1.000 1.000 -1.000 4.01 3.81
-1.000 1.000 1.000 1.000 4.41 4.36
1.000 1.000 1.000 1.000 4.41 4.18
0.000 1.414 0.000 0.000 6.11 6.06









All terms, regardless of significance were kept as to mirror what independent

variables would effect log reduction in a real system, as compared to this model. The

overall shape of the response surface was a saddle point with occurred at xl=27.4 MPa,

x2= 380C, x3=10.1% C02, x4=5.2 minutes residence time, and predicted log reduction of

5.1 logs.

Using the prediction equation above, the predicted maximum log reduction of 7.38

occurred at 26.5 MPa, 21C, 9.6% CO2, and 4.77 minutes residence time. As stated in

literature, log reduction was directly proportional to absorption of CO2 (Kumagai et al.,

1997). The lowest experimental temperature would allow the greatest amount of CO2 to

be dissolved in the beer. Although higher pressure and higher CO2 % would dissolve

more C02, and longer residence times would allow more equilibration, data shows that

saturation is reached at 9.6% C02, before the highest CO2 % of 12. Higher levels of CO2

only resulted in an excess of CO2. Also an increase of pressure from 26.62 to 37.33 MPa

did not result in significant increase in CO2 uptake. Because beer was carbonated before

processing the possible saturation of CO2 occurred at lower pressures and CO2 levels than

expected.

Overall, a predicted log reduction of 7.38 makes dense-phase CO2 pasteurization a

formidable alternative to heat pasteurization for the brewing industry. Currently most

brewers try to limit the amount of heat needed to pasteurize by prefiltering the beer to

decrease the number of yeast that must be killed by heat. This allows a brewer use to a

minimal amount of heat during pasteurization, thus limiting flavor damage. However the

use of dense-phase CO2 not only would allow the elimination of the filtering step prior to

pasteurization, but also would prevent flavor damage by heat because the process is









optimized for 210C. This would simplify the extension of beer shelf life, while insuring

no heat damage.

Mode of Cell Death

Yeast populations remained unchanged after one month of storage at 1.67C,

indicating that there may not be an injury/repair mechanism due to the nature of the

dense-phase CO2 pasteurization. To elucidate the mode of death, samples were examined

using scanning electron microscopy. Images 1, 2, and 3 compare yeast from unprocessed

(fresh) beer to that from beer after pasteurization at 27.6 MPa, 10% CO2, 210C, for 5

minutes, and heat pasteurization at 740C for 30 seconds, respectively. Fresh yeast are

pert and round, with a smooth appearance. Heat pasteurized yeasts still appeared round

and pert with slightly textured surfaces. After dense-phase CO2 pasteurization, some

cells show explosive decompression but most have a shrunken appearance with divots in

the surface. This illustrates the ability of dense-phase CO2 to affect cell membranes by

extraction of their components.










Figure 4-1: Scanning Electron Microscopy Picture of Yeast in Fresh Beer


Figure 4-2: Scanning Electron Microscopy Picture of Yeast in Beer Dense-Phase CO2
Processed at 27.6 MPa, 10% C02, at 210C, With a Residence Time of 5
Minutes


Figure 4-3: Scanning Electron Microscopy Picture of Yeast in Beer Pasteurized at 74C
for 30 Seconds










Alcohol in beer acts as a co-solvent, aiding in the extraction of hydrophobic and

some semi-polar materials. It has been reported that the solvent characteristics of CO2 can

be greatly modified by the addition or existence of certain cosolvents. For example, the

addition of ethanol may have significant effects on extraction by allowing increased

absorption on surface sites, preventing the re-adsorption of the compound of interest

(Clifford and Williams, 2000). Cosolvents have also been shown to increase the amount

of oil extracted from sunflower seeds (Raghuram Rao et al., 1992), and rape seed, soy

bean, cottonseed, flax seed, and peanuts, and to enhance the extraction of herbal

components such as borage seed oil and hiprose fruit (Illes et al., 1994).

Effect of Dense-phase CO2 Processing on Haze

Turbidity results after processing showed significant differences between all

sample means (p<.0001) on the ANOVA table and Tukey's mean separations (alpha=.1)

are shown in Figure 4-4. The 27.6 MPa CO2 processed samples had the lowest turbidity

average at 95.3 NTU, followed by 20.7 MPa CO2 processed samples at 101 NTU, heated

sample 120.7 NTU, and fresh samples at 146 NTU. Turbidity after one month of storage

at 1.670C, showed that all samples increased in haze significantly, however a mean

separation was not performed due to the increase in haze was most likely due to

microbial growth, not differences in beer protein (Table 4-2). The following ANOVA

results were used in the analysis:


Sum of
Source DF Squares Mean Square F Value Pr > F

Model 3 4720.916667 1573.638889 1110.80 <.0001

Error 8 11.333333 1.416667

Corrected Total 11 4732.250000







47


R-Square Coeff Var


0.997605


1.028283


DF Type I SS Mean Square


3 4720.916667


DF Type III SS

3 4720.916667


1573.638889


Mean Square

1573.638889


F Value Pr > F

1110.80 <.0001


F Value Pr > F

1110.80 <.0001


Table 4-2: Beer Haze in NTU After Processing and After Storage at 1.670 C for 30 Days
After Processing After Storage Difference
Sample Turbidity (NTU) Trubidity (NTU) Over Time
Fresh 146 425 P-value
Fresh 144 440 3.702E-07
Fresh 148 431
27.6 MPa 95 424 P-value
27.6 MPa 95 428 4.198E-07
27.6 MPa 96 442
20.7 MPa 101 404 P-value
20.7 MPa 102 400 1.247E-07
20.7 MPa 100 392
Heated 121 440 P-value
Heated 120 444 8.506E-08
Heated 121 452_


Beer Haze


600

400

300

200

100

0


Following Processing and
Storage


Following






IAfter Processing
mAfter Storage


Fresh 27.6 MPa 20.7 MPa Heated
Processing Conditions

Figure 4-4: Beer Haze Following Processing and Following Storage at 1.670C for 30
Days


Source


Root MSE

1.190238


y Mean

115.7500


Source










Changes seen in haze due to dense-phase CO2 processing may have been caused by

the drastic pH change associated with this process. During processing the pH drops to

approximately 3, which would affect protein conformation and polyphenol conformation,

interfering with protein-polyphenol complexes. To further examine the cause of the haze

differences seen between beer samples Polyacrylamide gel electrophoresis of beer

proteins was performed. Gels showed no differences in protein bands between fresh,

CO2 treated, and heat pasteurized samples. It must be concluded that differences in haze

between samples after processing cannot be attributed to changes in beer proteins during

dense-phase CO2 processing. Gel pictures are available in Appendix A.

Effect of Dense-phase CO2 Processing on Foam Capacity and Stability

Foam capacity results after processing showed differences between fresh, dense-

phase CO2 processed, and heated beer sample means (p=.0003). Tukey's mean

separation (alpha=. 1) showed that the CO2 processed samples were not significantly

different from each other, however they did differ from heated samples. Fresh and heated

samples were also found to be significantly different. The following ANOVA results

were used in the analysis for foam capacity and foam stability, respectfully:




Foam Formation:


Sum of
Source DF Squares Mean Square F Value Pr > F

Model 3 1632.000000 544.000000 22.67 0.0003

Error 8 192.000000 24.000000

Corrected Total 11 1824.000000


R-Square Coeff Var Root MSE


y Mean












4.898979


Type I SS

1632.000000



Type III SS

1632.000000


Mean Square

544.000000



Mean Square

544.000000


F Value

22.67



F Value

22.67


Foam Stability:



Source

Model

Error

Corrected Total


R-Square

0.938931


Source

x



Source

x


Sum of
Squares

164.0000000

10.6666667

174.6666667


Coeff Var

2.074312


Mean Square

54.6666667

1.3333333


Root MSE

1.154701


Type I SS

164.0000000



Type III SS

164.0000000


F Value Pr > F

41.00 <.0001


y Mean

55.66667


Mean Square

54.6666667



Mean Square

54.6666667


F Value

41 .00



F Value

41 .00


Pr > F

<.0001



Pr > F

<.0001


Mean separations are shown on Figure 4-5. Heated samples had the highest foam


capacity at 333%, followed by 27.6 MPa CO2 processed beer with 324%, 20.7 MPa CO2


processed beer with 321%, and fresh beer with 301% foam capacity (Figure 4-5). After


30 days of storage at 1.67C, sample means again showed significant differences


(p=.0006). Heat and fresh samples were not significantly different and had the highest


foam capacity with 332% and 327%, respectively. CO2 processed samples followed,


both with foam capacities of 307% (Figure4-6).


Source


Source


Pr > F

0.0003



Pr > F

0.0003


0.894737


1.530931


320.0000










Foam Capacity and Foam Stability Following
Processing


500
400
300 C AB B A
200
100
n


D Foam Capacity
* Foam Stability


Fresh 27.6 MPa 207 MPa Heated
Beer Processing Conditions
Figure 4-5: Foam Capacity and Stability of Beer Samples After Processing

Foam Capacity and Foam Stability Following
Storage


450
400
350
300 A B B
250
200
150
100
50
0


[] Foam Capacity
* Foam Stability


Fresh 27.6 MPa 20.7 MPa Heated
Beer Processing Conditions

Figure 4-6: Foam Capacity and Stability of Beer Samples After Storage at 1.67C for 30
Days

Foam stability results after processing showed differences between fresh, dense-

phase CO2 processed, and heated beer sample means (p<.0001) (Figure 4-5). Tukey's

mean separation (alpha=.1) showed that the CO2 processed samples were not

significantly different from each other, however did differ from fresh and heated samples.

The following ANOVA results were used in the analysis:

Foam Formation, Aged:

Sum of


Source


DF Squares Mean Square


F Value Pr > F













Model

Error

Corrected Total


R-Square

0.876106


Source

x


1584.000000

224.000000

1808.000000


Coeff Var

1.663995


528.000000

28.000000


Root MSE

5.291503


Type I SS

1584.000000




Type III SS

1584.000000


Source

x


18.86 0.0006


y Mean

318.0000


Mean Square

528.000000




Mean Square

528.000000


F Value

18.86




F Value

18.86


Pr > F

0.0006




Pr > F

0.0006


Foam Stability, Aged:


Source

Model


Error

Corrected Total


R-Square

0.777090


Sum of
Squares

334.6666667

96.0000000

430.6666667


Coeff Var

7.585624


DF Type I SS

3 334.6666667



DF Type III SS

3 334.6666667


Fresh and heated sample means were not significantly different. Heated samples


had the highest foam stability at 60%, followed by fresh beer with 59%, and CO2


Mean Square

111.5555556

12.0000000


F Value

9.30


Pr > F

0.0055


Source


Root MSE

3.464102


y Mean

45.66667


Source

x


Mean Square

111.5555556



Mean Square

111.5555556


F Value

9.30



F Value

9.30


Pr > F

0.0055



Pr > F

0.0055









processed samples with 52% foam stability for both CO2 treatments. After 30 days of

storage at 1.670C, sample means again showed significant differences (p<.0055). Heated

samples were significantly different from all other treatments with 55% foam stability.

Fresh and CO2 processed samples were not significantly different with fresh beer foam

stability at 44%, and 27.6 MPa and 20.7 MPa at 43% and 41%, respectively (Figure 4-6).

Changes seen in foam characteristics due to dense-phase CO2 processing may have

been caused by the extraction of cell membrane or cell wall parts that may have changed

the amount of hydrophobic compounds in the beer, therefore, affecting foaming.

In both foam capacity and foam stability, data showed that both CO2 processing

and heat pasteurization may have significant effects, however, all beers formed a stable

head at a level that would probably insure customer satisfaction. Original foam volumes

and liquid volumes are listed in Table A-2.

To further examine the cause of the foam differences seen between beer sample

means Polyacrylamide gel electrophoresis of beer proteins was performed. Gels showed

no differences in protein bands between fresh, CO2 treated, and heat pasteurized samples.

It must be concluded that differences in foam between samples after processing cannot be

attributed to changes in beer proteins during processing. Gel pictures are available in

Appendix A.

In the cases of both beer haze and foam, dense-phase CO2 pasteurization of beer in

no way decreased the quality of the finished product. Consumers would be expected to

see no differences between fresh and CO2 pasteurized beer.










Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Sensory
Evaluation

Difference from control panels and gas chromatography-olfactometry were

performed to prove there were no aroma or flavor changes due to dense-phase CO2

pasteurization. Panels were conducted within 1 day of processing and repeated after 30

days of storage at 1.67C. All samples were decarbonated and served at room

temperature. For sensory panels, analysis of variance (V=.05) was used to compare

treatment means and, when differences existed, a Tukey's mean separation test was

employed to compare specific treatment means at V=.10. A sample ballot is included in

Appendix B and raw data is available by contacting Dr. Murat Balaban in the Food

Science and Human Nutrition Department, University of Florida.

Beer Aroma

A difference from control test comparing aromas of fresh beer, beer processed at

27.6 MPa with 10% C02, for 5 minutes, at 21C, and beer processed at 20.7 MPa with

10% CO2, for 5 minutes, at 210C. Samples were evaluated on a line scale with 0=no

difference from the fresh beer reference and 10=extremely different from the fresh beer

reference.

The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 63 643.348046 10.211874 1.85 0.0022

Error 116 641.861454 5.533288

Corrected Total 179 1285.209500


R-Square Coeff Var Root MSE y Mean

0.500578 67.56231 2.352294 3.481667






54


Source DF Type I SS Mean Square F Value Pr > F

row 59 573.3695000 9.7181271 1.76 0.0051
treatment 2 66.9823333 33.4911667 6.05 0.0032
column 2 2.9962126 1.4981063 0.27 0.7633


Source DF Type III SS Mean Square F Value Pr > F

row 59 573.5187454 9.7206567 1.76 0.0051
treatment 2 66.9877883 33.4938942 6.05 0.0032
column 2 2.9962126 1.4981063 0.27 0.7633

Sample means were significantly different (p=.0032), with average scores of 2.72,

3.52, and 4.21, respectively on the ten point scale. All samples were rated as only

slightly different from the reference and similarities were seen between fresh and the 27.6

MPa beer, and between the 2 CO2 processed sample means. Mean separations are shown

using letters on Figure 4-7.

Difference From Control Aroma
Evaluation
0
a 10
5-

0


5 Hidden 27.6 MPa 20.7 MPa
Control

Processing Conditions
Figure 4-7: Aroma Evaluation of Fresh and CO2 Processed Beer Samples (mean
separations labeled)

A difference from control test was then conducted to compare the aroma of fresh,

27.6 MPa CO2 processed beer, and heat pasteurized beer (740C, 30 seconds). Samples

were again evaluated on a line scale with 0=no difference from the fresh beer reference












and 10=extremely different from the fresh beer reference. The following ANOVA results


were created and used for analysis:


Source

Model


Error

Corrected Total


R-Square

0.495550


DF Type I SS

44 444.1626667
2 25.2840000
2 6.8645381





DF Type III SS

44 445.3035982
2 26.4211450
2 6.8645381


There were no significant differences between sample means(p:


scores were 2.75, 3.29, and 3.81 and are shown in Figure 4-8.


1021). Average


Sum of
Squares

476.3112047

484.8647953

961.1760000


Mean Square

9.9231501

5.6379627


F Value

1 .76


Pr > F

0.0113


Coeff Var

72.39145


Root MSE

2.374439


Source

row
treatment
column


y Mean

3.280000


Source

row
treatment
column


Mean Square

10.0946061
12.6420000
3.4322690





Mean Square

10.1205363
13.2105725
3.4322690


F Value

1 .79
2.24
0.61





F Value

1 .80
2.34
0.61


Pr > F

0.0108
0.1124
0.5463





Pr > F

0.0105
0.1021
0.5463










Difference From Control Aroma
Evaluation





0


5 Hidden 27.6 MPa Heated
Control

Processing Conditions

Figure 4-8: Evaluation of Aromas of Fresh, CO2 Processed and Heat Pasteurized Beer
Samples

Beer Flavor

A difference from control test comparing flavor of fresh beer, beer processed at

27.6 MPa with 10% C02, for 5 minutes, at 21C, and beer processed at 20.7 MPa with

10% CO2, for 5 minutes, at 210C. Samples were evaluated on a line scale with 0=no

difference from the fresh beer reference and 10=extremely different from the fresh beer

reference. The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 63 681.309167 10.814431 2.02 0.0005

Error 116 619.517278 5.340666

Corrected Total 179 1300.826444


R-Square Coeff Var Root MSE y Mean

0.523751 60.09504 2.310988 3.845556


Source DF Type I SS Mean Square F Value Pr > F

row 59 666.5731111 11.2978493 2.12 0.0003
treatment 2 11.5821111 5.7910556 1.08 0.3415
column 2 3.1539444 1.5769722 0.30 0.7449










Source DF Type III SS

row 59 666.9011463
treatment 2 11.5196545
column 2 3.1539444

Sample means were not significantly different (p=

3.63, 3.71, and 4.2, respectively on the ten point scale.


Difference From Control Taste
Evaluation


10-

5

0-


Hidden
Control


Mean Square F Value Pr > F

11.3034093 2.12 0.0003
5.7598272 1.08 0.3435
1.5769722 0.30 0.7449

=.3435), with average scores of

Scores are shown on Figure 4-9.


Processing Conditions
Figure 4-9: Evaluation of Beer Flavor Between Fresh and CO2 Processed Samples

A difference from control test was then conducted to compare the flavor of fresh,

27.6 MPa CO2 processed beer, and heat pasteurized beer (740C, 30 seconds). Samples

were again evaluated on a line scale with 0=no difference from the fresh beer reference

and 10=extremely different from the fresh beer reference.

The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 48 658.515137 13.719065 2.51 <.0001

Error 86 469.693011 5.461547

Corrected Total 134 1128.208148


R-Square Coeff Var Root MSE


27.6 MPa 20.7 MPa


y Mean












55.69185


Source

row
treatment
column


Type I SS

571.6481481
76.3779259
10.4890631




Type III SS

574.7377992
78.8786056
10.4890631


Source

row
treatment
column


Mean Square

12.9920034
38.1889630
5.2445316




Mean Square

13.0622227
39.4393028
5.2445316


F Value

2.38
6.99
0.96




F Value

2.39
7.22
0.96


Pr > F

0.0003
0.0015
0.3869




Pr > F

0.0003
0.0013
0.3869


Sample means were significantly different (p=.0013), with average scores of 3.66,


3.67, and 5.26, respectively on the ten point scale. All samples were rated only slightly


different from the reference and similarities were seen between fresh and the 27.6 MPa


beer, and the heated sample was seen as significantly different from the others. Mean


separations are shown using letters on Figure 4-10.


Difference From Control Taste

Evaluation


10

5

0


Hidden

Control


27.6 MPa Heated


Processing Conditions

Figure 4-10: Evaluation of Beer Flavors Between Fresh, CO2 Processed, and Heat
Pasteurized Samples (mean separations labeled)


0.583682


2.336995


4.196296







59


Beer Aroma After Storage

A storage study was conducted at 1.670 C for 30 days and beer aroma and taste


tests were repeated. A difference from control test comparing aromas of fresh beer, beer


processed at 27.6 MPa with 10% C02, for 5 minutes, at 21C, and beer processed at 20.7


MPa with 10% CO2, for 5 minutes, after 30 days of storage at 1.670C at 21C, and a non-


stored, fresh hidden control. Samples were evaluated on a line scale with 0=no difference

from the fresh beer reference and 10=extremely different from the fresh beer reference.

The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 56 961.287287 17.165844 3.77 <.0001

Error 155 706.211393 4.556203

Corrected Total 211 1667.498679


R-Square Coeff Var Root MSE Sample_Aroma Mean

0.576485 74.35419 2.134526 2.870755


Source DF Type I SS Mean Square F Value Pr > F

Samp_Set 52 935.3336792 17.9871861 3.95 <.0001
Samp_ 3 24.9533962 8.3177987 1.83 0.1448
Samp_Code 1 1.0002111 1.0002111 0.22 0.6401


Source DF Type III SS Mean Square F Value Pr > F

Samp_Set 52 935.3336792 17.9871861 3.95 <.0001
Samp_ 3 25.4545507 8.4848502 1.86 0.1383
Samp_Code 1 1.0002111 1.0002111 0.22 0.6401



There were no significant differences seen between sample means and averages

scores of 2.35, 3.3, 2.99, and 2.84 are shown on Figure 4-11.










Difference From Control Aroma
Evaluation-One Month Later

0
o10





C Hidden 27.6 MPa 20.7 MPa Aged Keg
Control

Processing Conditions

Figure 4-11: Aroma Evaluation of Aroma Between Aged, CO2 Processed, Heat
Pasteurized, and a Non-Stored, Fresh Hidden Control

A difference from control test was then conducted to compare the aroma of aged,

27.6 MPa CO2 processed beer, heat pasteurized beer (740C, 30 seconds) after 30 days of

storage at 1.670C, and a non-aged fresh hidden control. Samples were again evaluated on

a line scale with 0=no difference from the fresh beer reference and 10=extremely

different from the fresh beer reference.

The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 50 825.213462 16.504269 3.03 <.0001

Error 129 703.666538 5.454779

Corrected Total 179 1528.880000


R-Square Coeff Var Root MSE taste Mean

0.539750 58.38867 2.335547 4.000000


Source DF Type I SS Mean Square F Value Pr > F

set 44 533.1900000 12.1179545 2.22 0.0003
number 3 290.1186667 96.7062222 17.73 <.0001
code 3 1.9047954 0.6349318 0.12 0.9504









Source DF Type III SS Mean Square F Value Pr > F
set 44 533.1900000 12.1179545 2.22 0.0003
number 3 290.0674621 96.6891540 17.73 <.0001
code 3 1.9047954 0.6349318 0.12 0.9504



Sample means were significantly different (p<.0001), with average scores of 2.56,

3.34, 4.87, and 1.99, respectively on the ten point scale. Mean separations are shown

using letters on Figure 4-12.





Difference From Control Aroma
Evaluation-One Month Later
0)
I-
Co
0


4 Hidden 27.6 MPa Heated Aged Keg
Control

Processing Conditions
Figure 4-12: Evaluation of Flavor Between Aged, CO2 Processed, Heat Pasteurized, and
a Non-Stored, Fresh Hidden Control (mean separations labeled)

Beer Flavor After Storage

A difference from control test comparing the flavor of aged beer, beer processed at

27.6 MPa with 10% C02, for 5 minutes, at 21C, and beer processed at 20.7 MPa with

10% CO2, for 5 minutes, after 30 days of storage at 1.670C at 210C, and a non-stored,

fresh hidden control. Samples were evaluated on a line scale with 0=no difference from








62



the fresh beer reference and 10=extremely different from the fresh beer reference. The


following ANOVA results were created and used for analysis:


Source

Model


Error

Corrected Total


R-Square

0.513796


Sum of
Squares

798.921305

756.016242

1554.937547


Coeff Var

61.58965


Source

Samp_Set
Samp_
Samp_Code



Source

Samp_Set
Samp_
Samp_Code


Root MSE

2.208512


Type I SS

784.6825472
13.3967925
0.8419653



Type III SS

784.6825472
13.7483426
0.8419653


Mean Square

14.266452

4.877524


F Value

2.92


Pr > F

<.0001


Taste Difference Mean

3.585849


Mean Square

15.0900490
4.4655975
0.8419653



Mean Square

15.0900490
4.5827809
0.8419653


F Value

3.09
0.92
0.17



F Value

3.09
0.94
0.17


Pr > F

<.0001
0.4349
0.6784



Pr > F

<.0001
0.4231
0.6784


Sample means were not significantly different (p=.4231), with average scores of


3.58, 3.91, 3.65, and 3.20, respectively on the ten point scale. Scores are shown on


Figure 4-13.










Difference From Control Taste

Evaluation-One Month Later
I-
0






< Hidden 27.6 MPa 20.7 MPa Aged Keg
Control

Processing Conditions

Figure 4-13: Evaluation of Flavor Between Fresh and CO2 Processed Beer After Storage
at 1.670C for 30 Days, and a Non-Stored, Fresh Hidden Control

A difference from control test was then conducted to compare the flavor of aged,

27.6 MPa CO2 processed beer, heat pasteurized beer (740C, 30 seconds) after 30 days of

storage at 1.670C, and a non-aged fresh hidden control. Samples were again evaluated on

a line scale with 0=no difference from the fresh beer reference and 10=extremely

different from the fresh beer reference.

The following ANOVA results were created and used for analysis:

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 48 793.305198 16.527192 3.34 <.0001

Error 131 648.607913 4.951205

Corrected Total 179 1441.913111


R-Square Coeff Var Root MSE Sample_Aroma Mean

0.550175 69.80193 2.225130 3.187778


Source DF Type I SS Mean Square F Value Pr > F

Samp_Set 44 581.5481111 13.2170025 2.67 <.0001
Samp_ 3 211.7388889 70.5796296 14.26 <.0001
Samp_Code 1 0.0181980 0.0181980 0.00 0.9517









Source DF Type III SS Mean Square F Value Pr > F
Samp_Set 44 581.5481111 13.2170025 2.67 <.0001
Samp_ 3 211.6859758 70.5619919 14.25 <.0001
Samp_Code 1 0.0181980 0.0181980 0.00 0.9517


Sample means were significantly different (p<.0001), with average scores of 2.92,

3.71, 6.14, and 3.23, respectively on the ten point scale. Mean separations are shown

using letters on Figure 4-14.


Difference From Control Taste
Evaluation-One Month Later

10
lO5




S< Hidden 27.6 MPa Heated Aged Keg
Control

Processing Conditions
Figure 4-14: Evaluation of Flavor of Fresh, CO2 Processed, and Heat Pasteurized
Samples After 30 Days at 1.670C, and a Non-Stored, Fresh, Hidden Control
(mean separations labeled)

Considering all decarbonated beer samples, overall, panelists could easily

distinguish heat pasteurized beer, but in only on one occasion was a difference noted

between fresh and the 20.7 MPa CO2 processed beer. This difference in aroma may have

been caused by the increased spillage of cell components, caused by CO2 processing at

20.7MPa. These cell components have been known to contribute to off-flavors in beer

(Lin et al., 1991). No differences were noted in flavor between the same samples and in

following less-volatile compounds, many of which could be hop constituents, could have

masked any differences in aroma during beer tasting. On average, panelists could only









discern heat pasteurized beer during sensory tests. The flavor differences caused by

dense-phase CO2 pasteurization were negligible. In an industry with its foundations in

producing a consistently fresh product, dense-phase CO2 pasteurization is an alternative

to heat pasteurization that results in the same extended shelf life, while preserving fresh

beer taste.

Effect of Dense-phase CO2 Processing on Beer Aroma and Flavor by Gas
Chromatography-Olfactometry and Mass Spectrometry

Odor descriptors and retention times of aroma active compounds present in beer

samples can be observed in Table 4-4. A variety of aroma compounds were detected by

the assessors, however only those detected more than 50% of the time by both assessors

on the ZB-5 column were used to characterize beer samples.

Linear retention indices (LRI) were calculated for aroma active compounds.

Tentative identification of the aroma compounds present in fresh beer (Table 4-5) was

conducted based on their LRI values obtained from the ZB-5 and Carbowax columns and

their aroma descriptors. Compounds in red were then confirmed by mass spectroscopy.

Methyl nerolate was also identified by mass spectra in the sample, however was not

detected previously by GC-O. Mass spectra library matches, chromatograms, and

aromagrams are available by contacting Dr. Murat Balaban in the Food Science and

Human Nutrition Department, University of Florida.

To evaluate if dense-phase CO2 processing did cause flavor changes, integration

areas (reported in mV), were compared for compounds that had been identified using LRI

values and confirmed using mass spectra (Table 4-6). When comparing values of

compounds before and after dense-phase CO2 pasteurization, many compounds had

negligible differences, however in the case of ethyl hexanoate, the amount of the










compound decreased on average by approximately 49%. Because of ethyl hexanoate's

volatile nature, not only did this compound elute early in the chromatography run (4.26

minutes on Carbowax column), but was also easily stripped by the carbon dioxide during

processing.

Table 4-4: Retention Times and Aroma Descriptors for Compounds Detected by Both
Assessors on ZB-5 and Carbowax Columns
Retention Times (minutes) Descriptor on ZB-5 Descriptor on Carbowax
(desciptors are listed for the appropriate column only)
2.82 fruity
3.21 banana, bubble gum
3.71 fruit
4.17 Fruity
4.26 musty, fruity
4.92 Fruity/sweet
5.58 fruity, sour
5.96 fruity, wine
6.18 Fruity
6.5 Fruity
7.3 pepper
8.7 green apple/sweet
9.6 sweet,
cooked/coconutty
11.2 green apple/sweet
11.9 apple pie
13.1 slight red
apple/sweet
13.61 apple pie
14.81 burnt juice
15.15 burnt

Although stripping did occur, sensory panelists did not recognize differences

between the 26.7MPa dense-phase CO2 and fresh beer. This is possible because even

though ethyl hexanoate did decrease in CO2 processed beer, if it remained above its

flavor threshold, no flavor change would be detectable to panelists. Overall, it can be

concluded that although some stripping did occur, no appreciable changes were detected

between fresh and CO2 processed beer samples.














Table 4-5: Linear Retention Indices and Identification of Compounds using GC-O (Compounds in Red Were Confirmed by Mass
Spectroscopy)
Retention Time LRI of Compounds on ZB-5 LRI of Compounds on Carbowax Tentative Identification (LRI)
2.82 1081 ethyl 2-methylbutanoate (1051)
3.21 1135 hexenal Z-3 (1145)
3.71 1197 hexanal E-2 (1228)
4.17 769 ethyl 2-methylpopanoate (751)
4.26 1258 ethyl hexanoate (1242)
4.92 879 hexenal Z-3 (853)
5.58 1416 hexanol, 2 (1417)
5.96 1452 ethyl heptanoate (1346)
6.5 1020 carene (1011)
7.3 1560 linalool (1552)
8.7 1128 ethyl 3-hydroxyhexanoate (1134)
9.6 1194 ethyl octanoate (1195)
11.2 1269 ethyl phenylacetate
11.9 1849 damascenone (1827)
13.1 1365 dodecenal (1407)
13.61 1396 dodecanoic acid (1568)






68


Retention index values are useful in compound identification when LRI values are

compared across different columns and/or different chromatographic conditions. Several

factors affect the retention time of an analyte in a chromatographic run such as column

length, carrier gas flow, film thickness of the column, column packaging, and

temperature program. All these factors significantly affect the retention time of a

compound, however, do not affect their retention indices, because these are relative

values, calculated from alkane standards run at the same chromatographic conditions.













Table 4-6: Average Integration Areas of Identified Compounds in Fresh and CO2 Processed Beer
Average Fresh Integration
Identified Compounds Areas Average CO2 Integration Areas Fresh-CO2 Integration Areas % change
ethyl hexanoate 15083 7744 7339 48.66
ethyl heptanoate 534903 532768 2136 0.40
ethyl octanoate 274993 254857 20136 7.32
ethyl phenylacetate 17668 19087 -1419 -8.03
dodecenal 28457 30530 -2073 -7.28
dodecanoic acid 139467 144960 -5493 -3.94







70


When comparing the relationship between the GC-FID chromatogram (Figure 4-

15) and the aromagram peak intensities for each compound (Figure 4-16), one can


observe that aroma active components that may drive the aroma profile either did not


have a response at all (peaks with RT= 4.17, 4.92, 6.18 min) or had a relatively low


response (peaks with RT= 6.53, 9.62, 13.3, 13.61) in the FID detector. The latter


observation likely occurred due to the presence of these compounds in lower quantities,


to their poor response and specific interactions with the FID detector, or have lower


threshold values. Moreover, compounds that have relatively high peak intensities to the


FID detector can be odorless compounds that do not contribute to the typical key notes of


the sample and thus they can not be detected by the assessor while conducting the GC-O


sniffing. However, an assessor can also not identify a specific aroma compound due to


selective anosmia or due to the low sensory threshold of the specific compound.

1500-


o IT
N "
.'- -' "
500- co o 1O CO N M L N 0G 0
4 1-N m mm o)
0L, ) 1' o[ AII I I I


0 2 4 6 8 10 12
Time Minutes


Figure 4-15: Typical FID Chromatogram for Fresh

1500 -

1000- CO LO r" -
0 O N|C
i500- 0 N
o -.i -

0- -

0 2 4 6 8 10 12
Time Minutes

Figure 4-16: Typical Aromagram for Fresh Beer


14 16 18 20 22


Beer


10 1








14 16 18 20 22









The data produced by GC-O has a qualitative component in which the assessor

describes the nature of their perception. This usually involves association of the precept

with a word or group of words in a lexicon. Qualitative GC-O data are either

measurement of odor potency or perceived intensity plotted against a retention index

(Friedrich and Acree, 2000). The major advantage of using GC-O is the ability to detect

the presence of unexpected aroma impact components. One does not need to know ahead

of time which components to measure. Using this technique many highly potent aroma

impact compounds are detected that would have been missed simply because they were

present at such minute concentrations. However, GC-O has limitations since it can not

determine synergistic or antagonistic interactions from other aroma active components in

the sample (Rouseff and Naim, 2000).

To reduce experimental error and variation, several factors can be monitored

including sample preparation, room and sample temperature, time of day, duration of

analysis, repetition of analysis, repeated standardization of sniffers, and use of a standard

lexicon. It is well documented that people can be trained to consistently identify smells if

they are standardized periodically and trained to sniff with standard chemicals and

vocabularies. The precision of threshold detection of an individual makes it more difficult

for the assessor to detect the end of an odor experience than the beginning of the

experience, lending some variation to intensity ratings (Rouseff and Naim, 2000;

Friedrich and Acree, 2000).

Overall, it can be concluded that although some stripping did occur, no

appreciable changes were detected between fresh and CO2 processed beer samples. This

finding, along with the results from the sensory panels, can be used to support the case









for dense-phase CO2 as an alternative to heat pasteurization of beer. Furthermore,

because of beer's complex mix of flavors, the minimal stripping that does occur during

dense-phase CO2 pasteurization is easily masked by other flavor compounds, resulting in

a final product no different from fresh beer.

Conjoint Analysis of Beer Purchase Decisions

A conjoint analysis is used to quantify how important a specific characteristic is of

a product in a consumer's purchase decision (Table 4-7). A conjoint analysis of beer

purchase decisions was conducted to elucidate how flavor, shelf life requirements, and

price affect purchase decision. Two levels of each attribute were tested, creating a full 23

factorial. Panelists were asked to rank 8 descriptions of beers in the order in which they

would purchase the beers. A sample ballot is available in Appendix B and ranking data is

available by contacting Dr. Murat Balaban in the Food Science and Human Nutrition

Department, University of Florida.

The attribute eliciting the largest influence on purchase decision was price, which

was 53.9% of a consumer's purchase decision. Overall, consumers preferred paying

$6.00/six 12 ounce bottles (average rank=3.705), rather than $8.00/6 12 ounce bottles

(average rank=5.279). The next most influential attribute was beer flavor, being 34% of

the purchase decision. Consumers preferred draft beer flavor with an average rank of

4.003, compared to bottled beer flavor with an average rank of 4.995. Finally, beer shelf

life showed the least effect on beer purchase decision, being 12.1% of the purchase

decision. Consumers preferred shelf stable beer with an average rank of 4.321 over beer

that required refrigeration that had an average rank of 4.674.

This conjoint analysis gives insight into characteristics of the beer that influence

which beer a consumer will purchase. This information is useful when introducing a new









product to an existing market, as in the case of introducing dense-phase pasteurized beer

to the current beer market. Using the data from the conjoint analysis, dense-phase CO2

pasteurized beer would successful because of its draft beer taste, coupled with it's

extended shelf life. However, if the cost of using dense-phase CO2 increased the price of

the beer significantly, consumer may opt not to purchase the product. Having seen that

price is 53.9% of the purchase decision, it would be in the best interest to minimize or

prevent a price increase by lowering the cost of production by recycling CO2 during

processing and eliminating cold storage during transit for the product.










Table 4-7: Conjoint Analysis Transformation
Average Deviation from Range of
Rank of Overall Average Reversed Squared Standardized Estimated Part- Part- Factor
Factor Level Level Rank Deviation Deviation Deviation Worth Worths Importance
Flavor
Draft 4.003 -0.497 0.497 0.247 0.828 0.910 1.815 34.0%
Bottled 4.995 0.495 -0.495 0.245 -0.819 -0.905
Storage
Refridgerated 4.321 -0.179 0.179 0.032 0.107 0.327 0.645 12.1%
Shelf Stable 4.674 0.174 -0.174 0.030 -0.101 -0.318
Price
$6.00 3.705 -0.795 0.795 0.632 2.114 1.454 2.879 53.9%
$8.00 5.279 0.779 -0.779 0.607 -2.031 -1.425 __














CHAPTER 5
CONCLUSION

The conclusions of this study are the following:

* Dense-phase CO2 pasteurization is effective in the pasteurization of beer, resulting
in a 5-log or higher decrease in yeast populations.

* Although some stripping did occur during dense-phase CO2 pasteurization, flavor
and aroma changes detected by panelists were negligible after processing and after
30 days of storage at 1.67C.

* Beer haze was significantly reduced by dense-phase CO2 pasteurization.

* Beer foam capacity and stability were affected by dense-phase CO2 pasteurization,
however not to the detriment of the finished product's quality.

* Beer purchase decision is most affected by price, then beer flavor, and finally shelf
life requirements. Consumers prefer shelf stable, draft beer and are adverse to an
increase in price.

In general, a continuous dense-phase CO2 system was effective in the

pasteurization of beer. The success of this system in the brewing industry would rely on

its predicted 7.38 log reduction in yeast populations, while preserving fresh beer aroma,

flavor, foaming capacity and stability, and aiding in the reduction of beer haze. By

resulting in an extended shelf life dense-phase CO pasteurization is a formidable

alternative to heat pasteurization, and would be preferred to heat pasteurization because

of its ability of maintain fresh beer characteristics. The ability to produce a clear,

consistently fresh tasting beer that forms a good head, with an extended shelf life is top-

priority to brewers. Dense-phase CO2 pasteurization can make this possible.

In addition, it is predicted that consumer acceptability of this technology in the

brewing industry would be high based on consumers' priorities when it comes to beer









characteristics. Consumers prefer a beer that does not have to be refrigerated and that has

draft beer taste; however they would be adverse to an increase in beer prices. If brewers

concentrated on cost-effective processing techniques, such as recycling CO2 throughout

the brewery and elimination of the cold storage of product, dense-phase CO2

pasteurization could be used to create a product with extended shelf life and draft beer

taste with no increase in price.

Indications of the mode of cell death were absorption of CO2 into the cell

membrane creating a physical disruption in membrane structure, visible as divots in

scanning electron microscopy pictures.

It is recommended that more research be conducted to elucidate the mode of cell

death and the role of alcohol as a co-solvent in yeast death by dense-phase CO2

pasteurization Further research in the area of beer flavor stripping would also lend to

the application of this technology to beverages in general. Recycling of the CO2 stream

and recovery of stripped volatiles would also be valuable research topics.















APPENDIX A
RAW EXPERIMENT DATA





















Table A-1: Yeast Counts for 27 Treatment Combinations Done in Duplicate

Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
1 1 0 0 0 0 0 0 0 0.00 6.1106
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
2 1 0 0 0 0 0 0 0 0.00 6.1106
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
3 1 3 52 27.5 9 3 6 16.75 1675.00 2.8866
2 1 0 0.5 0 1 0.5 0.5
3 1 0 0.5 0 1 0.5 0.5
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 1 0.5 0 0 0 0.25
4 1 0 1 0.5 42 2 22 11.25 1125.00 3.0594
2 0 0 0 1 0 0.5 0.25
3 0 cont 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
5 1 0 0 0 0 0 0 0 0.00 6.1106
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0 _



















Table A-1: Continued

Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
6 1 0 0 0 0 2 1 0 5 50 00 4 4116
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
7 1 4 5 4 5 5 8 6 5 5 5 550 00 3 3702
2 1 1 1 1 0 05 075
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 cont 0 0 0 0 0
6 0 0 0 0 0 0 0
8 1 1 1 1 1 2 1 5 1 25 125 00 4 0137
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 1 05 0 0 0 025
9 1 0 C0 0 0 0 0 0 0 00 6 1106
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 1 0 05 0 0 0 025
6 0 0 0 0 0 0 0
10 1 0 0 0 0 1 0 5 0 25 25 00 4 7126
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
11 1 5 0 2 5 0 0 0 1 25 125 00 4 0137
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 1 0 05 0 0 0 025
5 1 0 05 0 0 0 025
6 0 1 05 0 0 0 025
12 1 0 0 0 2 2 2 1 100 00 4 1106
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 1 05 025
5 0 U 0 0 05 0 0 0
6 0 0 0 0 0 0 0



















Table A-1: Continued

Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
20 1 0 0 0 0 0 0 0 000 6 1106
2 0 1 05 0 0 0 025
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
21 1 8 cont 8 5 12 85 825 82500 3 1941
2 0 0 0 1 0 05 025
3 0 0 0 0 1 05 025
4 0 0 0 0 0 0 0
5 0 1 05 0 0 0 025
6 0 0 0 0 0 0 0
22 1 0 0 0 0 1 0 5 0 25 25 00 47126
2 0 1 05 0 0 0 025
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
23 1 1 4 2 5 4 3 3 5 3 300 00 3 6335
2 2 0 1 1 0 05 075
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 1 0 05 025
6 0 0 0 0 0 0 0
24 1 0 1 0 5 0 0 0 0 25 25 00 47126
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
25 1 1 0 0 5 0 0 0 0 25 25 00 47126
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 2 0 1 05
6 0 0 0 0 0 0 0
26 1 cont cont #DIV/01 0 0 0 #DIV/01 0 00 6 1106
2 cont cont #DIV/01 0 0 0 #DIV/01
3 0 cont 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0

















Table A-1: Continued
Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
20 1 0 0 0 0 0 0 0 0.00 6.1106
2 0 1 0.5 0 0 0 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
21 1 8 cont 8 5 12 8.5 8.25 825.00 3.1941
2 0 0 0 1 0 0.5 0.25
3 0 0 0 0 1 0.5 0.25
4 0 0 0 0 0 0 0
5 0 1 0.5 0 0 0 0.25
6 0 0 0 0 0 0 0
22 1 0 0 0 0 1 0.5 0.25 25.00 4.7126
2 0 1 0.5 0 0 0 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
23 1 1 4 2.5 4 3 3.5 3 300.00 3.6335
2 2 0 1 1 0 0.5 0.75
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 1 0 0.5 0.25
6 0 0 0 0 0 0 0
24 1 0 1 0.5 0 0 0 0.25 25.00 4.7126
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
25 1 1 0 0.5 0 0 0 0.25 25.00 4.7126
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 2 0 1 0.5
6 0 0 0 0 0 0 0
26 1 cont cont #DIV/0! 0 0 0 #DIV/0! 0.00 6.1106
2 cont cont #DIV/0! 0 0 0 #DIV/0!
3 0 cont 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0 _
6 0 0 _0 0 _0 0 0 _
















Table A-1: Continued
Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
27 1 0 0 0 0 0 0 0 0.00 6.1106
2 0 0 0 1 2 1.5 0.75
3 0 0 0 1 0 0.5 0.25
4 0 0 0 0 0 0 0 ______
5 0 1 0.5 1 0 0.5 05 __
6___ 0 0 0 0 0 0 0 __


Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
1 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
2 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
3 1 5 7 6 9 7 8 7 700.00 3.2118
2 1 1 1 0 0 0 05
3 0 1 0.5 0 1 0.5 0.5
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
4 1 1 1 1 5 12 8.5 4.75 475.00 3.3802
2 0 1 0.5 1 0 0.5 0.5
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 _
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
















Table A-1: Continued
Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
5 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
6 1 0 2 1 0 1 0.5 0.75 75.00 4.1818
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
7 1 6 7 6.5 4 6 5 5.75 575.00 3.2972
2 1 1 1 1 1 1 1
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 cont 0 0 0 0 0
6 0 0 0 0 0 0 0
8 1 2 1 1.5 2 2 2 1.75 175.00 3.8139
2 0 0 0 0 0 0 0_
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 1 0.5 0 0 0 0.25
9 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 1 0 05 0 0 0 0.25
6 0 0 0 0 0 0 0

















Table A-1: Continued

Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
10 1 1 1 1 0 1 0.5 0.75 75.00 4.1818
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
11 1 3 4 3.5 1 0 0.5 2 200.00 3.7559
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 2 1 0 0 0 0.5
5 1 0 0.5 0 0 0 0.25
6 0 0 0 0 0 0 0
12 1 1 2 1.5 2 2 2 1.75 175.00 3.8139
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
13 1 1 1 1 2 1 1.5 1.25 125.00 3.9600
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
14 1 1 1 1 2 1 1.5 1.25 125.00 3.9600
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
15 1 1 0 0.5 0 1 0.5 0.5 50.00 4.3579
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 01 0
____ 6 0 0 0 0 0 0 0
















Table A-1: Continued
Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
16 1 1 1 1 0 1 0.5 0.75 75.00 4.1818
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
17 1 2 4 3 6 6 6 4.5 450.00 3.4037
2 1 0 0.5 0 1 0.5 0.5
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
18 1 0 2 1 2 3 2.5 1.75 175.00 3.8139
2 0 0 0 0 1 0.5 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
19 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
20 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 1 0.5 0 0 0 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
21 1 8 7 7.5 4 5 4.5 6 600.00 3.2788
2 0 0 0 1 0 0.5 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0 _
__ 5 0 0 0 0 0 0 0__
___ 6 0 0 0 0 0 0 0
















Table A-1: Continued

Treatment Dilution
A A Ave A B B Ave B Ave Total Colonies Remaining Log Reduction log(no/n)
22 1 1 1 1 0 1 0.5 0.75 75.00 4.1818
2 0 1 0.5 0 0 0 0.25
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
23 1 2 2 2 3 5 4 3 300.00 3.5798
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
24 1 1 1 1 0 0 0 0.5 50.00 4.3579
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
25 1 0 1 0.5 0 0 0 0.25 25.00 4.6590
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 2 0 1 0.5
6 0 0 0 0 0 0 0
26 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0
6 0 0 0 0 0 0 0
27 1 0 0 0 0 0 0 0 0.00 6.0569
2 0 0 0 0 0 0 0
3 0 0 0 0 0 0 0
4 0 0 0 0 0 0 0
5 0 0 0 0 0 0 0 __
__6 0 0 0 0 0 0 0











After Processing After Storage
Treatment foam volume liquid volume foam volume liquid volume
Fresh 75 10 81 14
Fresh 76 10 81 14
Fresh 75 11 83 14
Fresh Average 75.33333333 10.33333333 81.66666667 14
26.7 MPa 80 12 78 14
26.7 MPa 81 12 76 14
26.7 MPa 82 12 76 15
26.7 MPa Average 81 12 76.66666667 14.33333333
20.7 MPa 78 12 78 14
20.7 MPa 82 12 76 16
20.7 MPa 81 12 76 14
20.7 MPa Average 80.33333333 12 76.66666667 14.66666667
Heated 84 10 81 12
Heated 83 10 84 10
Heated 83 10 84 12
Heated Average 83.33333333 10 83 11.33333333


figure A-I: roiyacryiamiae Jeis


Top Gel: 15%
cross-linked
Lanes from Left to
Right:
1. Heated Beer
2. 20.7 MPa
3. 27.6 MPa
4. Fresh Beer
5. Heated Beer
6. 20.7 MPa
7. 27.6 MPa
8. Fresh Beer
9. Molecular
Weight
Marker
Bottom Gel: 18%
cross-linked
Lanes from Left to
Right:
Same as Above


Table A-2:


Foam and Liquid Volumes














APPENDIX B
SENSORY AND CONJOINT BALLOTS


You will be given a reference sample marked 000. Please use the following line

scales to say how different each sample's aroma is from the reference's aroma.

Sample 352

1 --------------------------------------------]

ro difference at all extremely different

Sample 428


Io difference at all extremely different
rio difference at all extremely different


Figure B-1: Sample Sensory Ballot