<%BANNER%>

Thermal Inactivation Kinetics of Escherichia coli and Alicyclobacillus acidoterrestris in Orange Juice


PAGE 1

THERMAL INACTIVATI ON KINETICS OF Escherichia coli AND Alicyclobacillus acidoterrestris IN ORANGE JUICE By VERTIGO MOODY 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 2003

PAGE 2

ACKNOWLEDGMENTS The author would like to express sincere gratitude to his major advisor, Dr. Arthur A. Teixeira for his confidence and enthusiasm throughout this research project. His guidance and support were essential for successful completion of this body of work. The author would also like to express gratitude and appreciation to his supervisory committee (Dr. Glen H. Smerage, Dr. Mickey Parish, Dr. Robert Braddock, and Dr. Spiros Svorounous) for their guidance and suggestions related to the research and the completion of this publication. Special thanks go to the faculty and staff of the Agricultural and Biological Engineering Department, especially Dr. David Chynoweth and Dr. Roger Nordstedt for the use of their lab space and equipment as well as Ms. Veronica Campbell for her guidance and technical skills in assisting with the laboratory aspect of this research project. Special thanks go to Dr. Braddock, Rockey Bryan and the staff at the Citrus Research and Education Center for assisting the author in coordinating visits to the center to conduct research and for troubleshooting problems with equipment. The author wishes to thank Dr. Parish and Lorrie Friedrich for their assistance with the microbiological aspect of this research project. Their help facilitated the completion of this project and enhanced the skills of the author for handling microorganisms in a laboratory setting. Finally, the author would like to thank his family and friends for their continued support and patience throughout this milestone in life. ii

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGMENTS..................................................................................................ii TABLE OF CONTENTS...................................................................................................iii LIST OF TABLES.............................................................................................................vi LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................xi CHAPTER 1 ESTIMATING THERMAL KINETIC PARAMETERS FOR Escherichia coli IN SINGLE-STRENGTH ORANGE JUICE USING TRADITIONAL ANALYSIS OF ISOTHERMAL BATH EXPERIMENTAL DATA.....................................................1 Introduction...................................................................................................................1 Literature Review.........................................................................................................2 Microbiology of Fruit Juices.................................................................................2 Mechanism of Acid Tolerance..............................................................................5 Spoilage.................................................................................................................6 Objectives.....................................................................................................................7 Methods and Materials.................................................................................................8 Scope of Work.......................................................................................................8 Preliminary Experiments.......................................................................................9 Preparation of Cultures..........................................................................................9 Source of strains.............................................................................................9 Acid adaptation preparation.........................................................................11 Experimental Apparatus......................................................................................12 Isothermal Inactivation Experiments...................................................................12 Estimating Dand z-values.................................................................................13 Results and Discussion...............................................................................................14 Preliminary Experiments.....................................................................................14 Saccharomyces cerevisiae............................................................................14 Escherichia coli cultured at neutral pH........................................................14 Acid-tolerant Escherichia coli cultures........................................................16 Thermal Inactivation of Escherichia coli............................................................17 iii

PAGE 4

2 ESTIMATING KINETIC PARAMETERS FOR THERMAL INACTIVATION OF Escherichia coli IN ORANGE JUICE USING THE PAIRED EQUIVALENT ISOTHERMAL EXPOSURES (PEIE) METHOD WITH A CONTINUOUS HIGH TEMPERATURE SHORT TIME (HTST) PROCESS TREATMENT.....................47 Introduction.................................................................................................................47 Literature Review.......................................................................................................48 First-order kinetics...............................................................................................49 The PEIE Method................................................................................................51 Objectives...................................................................................................................52 Methods and Materials...............................................................................................53 Preparation of Cultures........................................................................................53 Experimental Apparatus......................................................................................53 Calibration of Thermocouples.............................................................................54 Continuous Dynamic Thermal Treatments.........................................................55 Temperature Profiles...........................................................................................56 Estimating Dand z-Values with the PEIE Method............................................57 Validation Experiments.......................................................................................59 Results and Discussion...............................................................................................61 Continuous Dynamic Thermal Experiments Parameter Estimation.................61 Comparing PEIE and 3-Neck Flask Isothermal Methods...................................62 Validation Experiments.......................................................................................64 3 ESTIMATION OF KINETIC PARAMETERS FOR THERMAL INACTIVATION OF Alicyclobacillus acidoterrestris IN ORANGE JUICE.........................................85 Introduction.................................................................................................................85 Literature Review.......................................................................................................86 Occurrences of Alicyclobacillus acidoterrestris in Juice Products.....................86 The PEIE Method in Arrhenius Kinetics.............................................................87 The PEIE Method and TDT Kinetics..................................................................90 Objectives...................................................................................................................92 Methods and Materials...............................................................................................92 Preparation of Cultures........................................................................................92 Experimental Apparatus......................................................................................93 Continuous Dynamic Thermal Treatments.........................................................93 Temperature Profiles...........................................................................................94 Results and Discussion...............................................................................................96 Parameter Estimation by PEIE............................................................................96 Parameter Estimation using F value and TDT kinetics.......................................96 iv

PAGE 5

APPENDIX A MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli USING ARRHENIUS KINETICS...........................................................................109 B MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli USING THERMAL DEATH TIME (TDT) KINETICS..........................................118 C MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus acidoterrestris USING ARRHENIUS KINETICS..................................................126 D MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus acidoterrestris USING THERMAL DEATH TIME (TDT) KINETICS.................134 LIST OF REFERENCES.................................................................................................141 BIOGRAPHICAL SKETCH...........................................................................................144 v

PAGE 6

LIST OF TABLES Table page 1-1. Plate counts of survivors grown in standard nutrient broth and pH-modified nutrient broth for inducing acid tolerance.............................................................................23 1-2. D-values (seconds) for Escherichia coli in orange juice cultured at neutral pH (standard culture) in preliminary experiments.........................................................31 1-3. D-values (seconds) for Escherichia coli in orange juice cultured at low pH (acid adapted culture) in preliminary experiments............................................................37 1-4. D-values (seconds) from thermal inactivation experiments for Escherichia coli cultured at low pH....................................................................................................44 1-5. Comparison of TDT kinetic parameters with published data from Mazzotta (2001) and Splittstoesser et. al. (1996) using acid adapted and non-acid adapted Escherichia coli in orange juice...............................................................................46 2-1. Calibration of thermocouples....................................................................................69 2-2. Reynolds numbers for each flow rate for the continuous system..............................70 2-3 Rate constants used in Equation 2-1 for the heater section temperature profile.......74 2-4. Rate constants used in Equation 2-2 for the chiller section temperature profile.......74 2-5. Population survivor data for continuous experiments...............................................75 2-6. Estimation of Dand z-values from each iteration of the PEIE method...................76 2-7. Comparison of Dand z-values estimated by traditional method using isothermal treatments and PEIE method using continuous dynamic treatments.......................78 2-8. Kinetic parameters of thermal inactivation of Alicyclobacillus acidoterrestris spores in Cupuacu nector using the PEIE method and Isothermal method *......................78 2-9. Results of validation experiments, comparison of predicted number of survivors for PEIE analysis and Traditional isothermal batch analysis with experimental number of survivors...............................................................................................................84 vi

PAGE 7

3-1. Rate constants used in Equation 2-1 and 2-2 for the heater and chiller sections temperature profile for experimental set 1.............................................................101 3-2. Rate constants used in Equation 2-1 and 2-2 for the heater chiller sections temperature profile for experimental set 2.............................................................101 3-3. Population survivor data from Ultra High Temperature (UHT) heat treatments with Alicyclobacillus acidoterrestris in orange juice.....................................................102 3-4. Estimation of k and Ea values from each iteration of the PEIE method using Arrhenius kinetics..................................................................................................103 3-5. Estimation of Dand z-values from each iteration of the PEIE method using TDT kinetics...................................................................................................................105 3-6. Comparison of TDT kinetic parameters with published data from various sources using Alicyclobacillus acidoterrestris....................................................................107 vii

PAGE 8

LIST OF FIGURES Figure Page 1-1. Growth curve showing light absorbance at a wavelength of 600 nanometer vs time for Saccharomyces cerevisiae in yeast extract peptone dextrose (YEPD) broth. Sets are runs conducted on separate days.................................................................21 1-2. Growth curve showing absorbance of light at wavelength of 600 nanometer vs time for Escherichia coli ATCC #9637 in nutrient broth. Sets are experiments conducted on separate days......................................................................................22 1-3. Experimental apparatus (photograph)......................................................................24 1-4. Experimental apparatus (diagram)............................................................................25 1-5. Survivor curves from preliminary experiments at 50 o C, 54 o C and 56 o C for Saccharomyces cerevisiae in orange juice cultured at neutral Ph (standard culture)26 1-6. Preliminary experiments survivor curve at 59 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture)..................................................................27 1-7. Preliminary experiments survivor curves at 62 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture).........................................................28 1-8. Preliminary experiments survivor curves at 64 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture).........................................................29 1-9. TDT curve from preliminary experiments with Escherichia coli in orange juice cultured at neutral pH (standard culture). R 2 value of 0.90......................................30 1-10. Survivor curves from preliminary experiments at 52 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)...........................................32 1-11. Survivor curves from preliminary experiments at 55 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)...........................................33 1-12. Survivor curves from preliminary experiments at 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)...........................................34 1-13. Family of survivor curves from preliminary experiments at 52 o C, 55 o C, and 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).....35 viii

PAGE 9

1-14. TDT curve from preliminary experiments with Escherichia coli in orange juice cultured at low pH (acid adapted culture). R 2 value of 0.99...................................36 1-15. pH of broth vs. pH of orange juice product for Saccharomyces cerevisiae preliminary experiments...........................................................................................38 1-16. Survivor curve from thermal inactivation experiments at 52 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).......................................39 1-17. Survivor curve from thermal inactivation experiments at 55 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).......................................40 1-18. Survivor curve from thermal experiments at 58 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).......................................................41 1-19. Survivor curve from thermal inactivation experiments at 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).......................................42 1-20. Family of survivor curves at 52 o C, 55 o C, 58 o C and 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)...........................................43 1-21. TDT curve from the thermal inactivation experiment with Escherichia coli in orange juice cultured at low ph (acid adapted culture). R 2 value of 0.98...............45 2-1. Photo of the Microthermics HTST Lab 25 Labscale Pasteurizer.............................67 2-2. Schematic Diagram of the flow of the Microthermics pasteurizer...........................68 2-3. Thermal profile of product at a hold tube nominal temperature of 58 o C and residence times of 60 and 90 seconds......................................................................71 2-4. Thermal profile of product at a hold tube nominal temperature of 60 o C and residence times of 30 and 60 seconds......................................................................72 2-5. Thermal profile of product at a hold tube nominal temperature of 62 o C and residence times of 15 and 30 seconds......................................................................73 2-6. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic parameters from the PEIE method...........................................................................77 2-7. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic parameters from the PEIE method...........................................................................79 2-8. Comparison of TDT curves based upon data from the traditional and PEIE methods80 2-9. Comparison of TDT curves based upon data from the traditional and PEIE methods for Alicyclobacillus acidoterrestris spores in Cupuacu nectar (Vieira et. al. 2002) (Estimated curve based upon reference D-value and z-value).................................81 ix

PAGE 10

2-10. Temperature history and measured and predicted survivor responses for validation experiment I (10 second hold tube)..........................................................................82 2-11. Temperature history and measured and predicted survivor responses for validation experiment II (15 second hold tube)........................................................................83 3-1. Thermal profile of product at a hold tube nominal temperature of 95 o C, 100 o C, and 104 o C for experimental set one................................................................................99 3-2. Thermal profile of product at a hold tube nominal temperature of 95 o C, 100 o C, and 104 o C for experimental set two..............................................................................100 3-3. Arrhenius curve for Alicyclobacillus acidoterrestris in orange juice using kinetic parameters from the PEIE method.........................................................................104 3-4. TDT curve for Alicyclobacillus acidoterrestris in orange juice using kinetic parameters from the PEIE method.........................................................................106 3-5. Comparison of TDT curves based upon data from the PEIE method using TDT kinetics and Arrhenius kinetics..............................................................................108 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 THERMAL INACTIVATION KINETICS OF Escherichia coli AND Alicyclobacillus acidoterrestris IN ORANGE JUICE By Vertigo Moody December 2003 Chair: Arthur A. Teixeira Co-chair: Glen H. Smerage Major Department: Agricultural and Biological Engineering Growing concern about the safety of unpasteurized low-pH foods has changed the view of the microbial loads supported by these products. Recent outbreaks of Salmonella in single-strength unpasteurized orange juice and Escherichia coli O157:H7in apple juice have prompted food processors to seek ways of ensuring the safety of their products without compromising consumer acceptance. Spoilage is also a concern as it relates to the shelf life of fruit juice products. In order to achieve an optimum balance between safety, shelf life, and quality, good estimation of thermal inactivation parameters is essential for designing pasteurization processes that achieve all three goals. The purpose of this study was to validate a method for estimating thermal inactivation kinetic parameters of specific microorganisms. The method, called the Paired Equivalent Isothermal Exposures (PEIE) method, may be applied to products that are heated under non-isothermal conditions. This method simplifies the estimation of xi

PAGE 12

parameters by eliminating the need to perform tedious isothermal bath experiments, while still obtaining accurate estimations. The study was performed in three phases: 1) Estimating thermal kinetic parameters for Escherichia coli in single strength orange juice using traditional analysis of isothermal bath experimental data; 2) Estimating kinetic parameters for thermal inactivation of Escherichia coli in orange juice using the PEIE method with end-point data from continuous high-temperature short-time (HTST) process treatments and validation for each set of kinetic parameters, and 3) Estimating kinetic parameters for thermal inactivation of Alicyclobacillus acidoterrestris using the PEIE method. Estimating kinetic parameters from isothermal bath and continuous dynamic thermal treatment data gave parameters that were different. To confirm which parameters were more accurate, validation experiments were conducted at higher temperatures. Using the parameters from both methods the number of survivors from each experiment were compared with those predicted by each set of kinetics parameters. Results from validation experiments with Escherichia coli showed that model predictions agreed more closely with experimental data when kinetic parameters used were estimated by the PEIE method rather than the traditional isothermal bath method. The process conditions determined from the kinetic parameters estimated by the PEIE method yielded a 39.7% shorter time than that determined by the isothermal bath method. The PEIE method was used as the preferred method for estimating the kinetic parameters for Alicyclobacillus acidoterrestris in single-strength orange juice. xii

PAGE 13

CHAPTER 1 ESTIMATING THERMAL KINETIC PARAMETERS FOR Escherichia coli IN SINGLE-STRENGTH ORANGE JUICE USING TRADITIONAL ANALYSIS OF ISOTHERMAL BATH EXPERIMENTAL DATA Introduction Recent outbreaks of Escherichia coli and Samonella in low-pH fruit juices (including apple and orange) have prompted reevaluation of the ability of pathogenic microorganisms to survive in these high-acid food products. Unpasteurized fruit juices have become popular consumer products because flavor and texture quality are better than in pasteurized juices. Escherichia coli O157:H7 and Salmonella contaminated orange and apple juice and apple cider have raised the attention of the Food and Drug Administration, which previously considered high-acid foods with pH below 4.6 not to be potentially hazardous to consumers. These outbreaks provide a compelling reason to study these organisms tolerance to low pH and to study their effect on the safety and shelf life of these products. The design of pasteurization processes depends on estimating the thermal inactivation kinetic parameters. Performing thermal inactivation experiments on the acid-tolerant bacteria allows engineers to design thermal processes that more completely reduce the number of pathogenic microorganisms in the product to more safe levels. Accurate estimation of kinetic parameters is essential to food engineers. The purpose of this study is to characterize the thermal inactivation behavior of potentially pathogenic bacteria in orange juice. 1

PAGE 14

2 Literature Review Microbiology of Fruit Juices Up to the latter part of the 20 th century it was widely assumed that pathogenic microorganisms could not survive in low-pH, high-acid foods because of the belief that organic acids had an inhibitory and sometimes microbicidal effect (Parish 1997). The Food and Drug Administration generally considers foods with a pH greater than 4.6 to be potentially hazardous to consumers. Unpasteurized fruit juices have become a popular consumer food product because their flavor retention is better than that of pasteurized fruit juices. However, recent outbreaks of foodborne illness stemming from unpasteurized fruit juices have brought to the forefront the need for pasteurization of all processed fruit juices. Outbreaks involving Escherichia coli O157:H7 and Salmonella enterica in orange and in apple juices and apple cider have changed long held views on the safety of fruit juices and other low-pH products. Escherichia coli O157:H7 was first confirmed as a health concern in juices after an apple cider related outbreak in 1991 (Besser et. al. 1993). An outbreak of diarrhea and Hemolytic Uremic Syndrome (HUS) in southern Massachusetts was traced back to contamination of fresh-pressed apple cider (Besser et. al. 1993). Twenty-three persons were identified with Escherichia coli O157:H7 infections between October 23 and November 24 of 1991. An epidemiological study based on this case showed that when apple cider, with a pH ranging between 3.7 and 3.9, was inoculated with Escherichia coli O157:H7, bacteria survived for 20 days at refrigerated conditions (8 o C) (Besser et. al. 1993). Another outbreak of Hemolytic Uremic Syndrome (HUS) caused by the consumption of unpasteurized apple juice that was contaminated with Escherichia coli O157:H7 was documented in 1996 (Parish 1997). In this outbreak a large producer of

PAGE 15

3 fresh unpasteurized fruit products was implicated in the distribution of contaminated product. Salmonella has been isolated from apple cider samples (pH from 3.7 to 4.0) associated with an outbreak of gastroenteritis (Besser et. al. 1993). In 1989 an incident of typhoid fever caused by consuming orange juice contaminated with Salmonella typhi was documented in a New York hotel restaurant in which there were 45 confirmed and 24 probable cases of typhoid fever with 21 hospitalizations (Parish 1997). In 1996 (on June 19, in the state of Washington and on June 23, in the state of Oregon) health officials investigated clusters of outbreaks of diarrhea attributed to Salmonella and associated with a commercially distributed unpasteurized orange juice (CDC 1999). Samples of the unpasteurized orange juice yielded cultures of Salmonella when analyzed by the Food and Drug Administration (FDA). There were approximately 300 confirmed cases associated with this outbreak (CDC 1999). These recent outbreaks of food poisoning from Salmonella and Escherichia coli O157:H7 have called into question the safety of unpasteurized fruit juices and other low-pH, high-acid food products. Pasteurization is the traditional method of inactivating pathogenic and some spoilage-causing microorganisms in citrus products. The inhibitory effect of acid concentration and low pH toward the growth of most pathogenic bacteria alone does not ensure product safety (Parish 1997). Pathogens such as Salmonella, Escherichia coli O157:H7, Shigella, Vibrio, and Staphyloccocus have been shown to survive from hours to days and even weeks in various fruit juice products (Parish 1997).

PAGE 16

4 Miller and Kaspar (1994) showed the acid tolerance and survival of Escherichia coli O157:H7 in apple cider by testing two different strains. In their study they inoculated Trypticase soy broth (TSB) adjusted to various pHs, and commercial apple cider with those strains and observed the survival at each pH. Viable cells of Escherichia coli O157:H7 were still detectable in TSB at pH 2 after 24 hours of storage at refrigerated conditions. In apple cider cells were still detectable after 14 days of storage at 4 o C. Leyer et al. (1995) showed that acid-adapted Escherichia coli O157:H7 survived for 81 hours in apple cider with a pH of 3.42 stored at 6 o C, whereas the non adapted cells survived for only 28 hours. Semanchek and Golden (1996) showed that pathogenic Escherichia coli O157:H7 is capable of survival in apple cider for at least 10 days at a storage temperature of 20 o C with a minimal decrease in population of viable cells. In a study by Zhao et al. (1993) Salmonella survived in apple juice stored at 4 o C for more than 30 days at pH 3.6. These studies revealed that storage conditions affect the resistance to acid of these pathogens. Storage at refrigerated temperatures increases the time at which cells remain viable in the product. Zhao et al. (1993) showed that Escherichia coli O157:H7 was more rapidly inactivated in apple cider stored at 25 o C than at 4 o C. Ingham and Uljas (1998) reported that 84% to 91% of their inoculum of Escherichia coli cells was still viable in apple cider, without preservatives, after 21 days when stored at 4 o C. Similar studies conducted in different low-pH products showed an increase in the thermotolerance of Escherichia coli O157:H7. Leyer et al. (1995) reported that acid-adapted Escherichia coli O157:H7 in fermented meats showed a higher thermotolerance.

PAGE 17

5 Mechanism of Acid Tolerance The mechanism of acid tolerance of bacteria is not completely understood. Several theories have been proposed in an attempt to explain how bacteria are able to adjust and maintain their internal pH within homeostatic limits. These theories include the buffering capacity of cytoplasm, the low proton permeability of cells, and the extrusion of protons from the cytoplasm by a membrane-bound proton pump (Benjamin and Datta 1995). The antimicrobial effect of acids has been explained by the ability of undisassociated molecules to enter the cell membrane and release protons. This release of protons disrupts the electron transport system of the bacterial cell draining cellular energy resources (Diez-Gonzalex and Russell 1997). The electron transport system is highly dependent on the maintenance of a constant chemiosmotic potential across the inner mitochondrial membrane to ensure steady production of adenosine triphosphate (ATP) in the cellular environment. Bacteria capable of surviving in low pH (such as lactic acid bacteria) are able to decrease intracellular pH when extracellular pH decreases to maintain a low transmembrane pH gradient (Diez-Gonzalez and Russell 1997), thus decreasing the dissipation of the proton-motive force. Diez-Gonzalez and Russell (1997) studied the ability of Escherichia coli O157:H7 to change its intracellular pH in response to a change in the extracellular pH as a mechanism of acid tolerance and the ability to survive in low-pH products. They showed that Escherichia coli O157:H7 had a greater ability to control the level of acetate concentration within its internal environment than a non-pathogenic Escherichia coli strain. The O157:H7 strain maintained a maximum internal concentration of acetate less than 300 mM while the non-pathogenic strain accumulated as much as 500 mM of acetate internally when the external pH dropped to 5.9. The significance of the concentration of

PAGE 18

6 acetate in the cytoplasm gives insight into the ability of the bacteria to regulate the ions and thus reduce the impact of dramatic changes in external pH. Protein synthesis appears to be an essential aspect of the acid tolerance response of cells. OHara and Glenn (1994) showed inhibiting protein synthesis with compounds such as chloramphenicol prevented the development of acid tolerance in the cells. The nature of these proteins and their role in the acid tolerance response are not known. They also reported that the capacity to maintain alkaline intracellular pH is essential for the survival of root nodule bacteria in acidic environments. Spoilage In addition to product safety, the population size of viable microorganisms that remain in the product also affects the shelf life of the product with significant economic implications. Pasteurized single-strength juices and frozen juice concentrates are the predominant types of processed fruit juices commercially available. Yeasts, molds, and lactic acid bacteria have been implicated in the spoilage of fruit juices (Deak et al. 1993). Yeasts are the most problematic because of their ability to tolerate low-pH environment. In particular, Saccharomyces cerevisiae is the most commonly isolated species of yeasts from fruit juices that is responsible for spoilage. Twenty-five percent of yeast isolates from frozen concentrate were identified as Saccharomyces cerevisiae in a survey conducted in 1993 (Deak and Beuchat 1993). Yeasts lead to formation of films, alteration of color, and change in viscosity. The fermentation caused by yeasts produce products such as ethanol, carbon dioxide, and ethyl acetate, which alter the flavor of the products. The production of gases may also compromise the integrity of product packaging. The aim of pasteurization has been to eliminate the pathogenic

PAGE 19

7 microorganisms, reduce the population of spoilage-causing microorganisms and to inactivate enzymes for product safety and extended shelf life. Recent outbreaks of Salmonella and Escherichia coli O157:H7 in orange and apple juice and in apple cider provide a compelling reason to understand these microorganisms tolerance to low pH in relation to their ability to cause disease and how that tolerance affects thermal inactivation characteristics in those products for the purpose of food safety. Estimating the thermal inactivation characteristics of these pathogenic organisms in low-pH environments has both a food safety and economic impact on the design and processing of fruit juice products. Because the assumption (that inactivation caused by acid is sufficient) may no longer be valid, performing isothermal inactivation experiments on the acid tolerant strains of pathogenic microorganisms such as Escherichia coli allows engineers to design thermal processes that more completely reduce the number of viable microorganisms to levels that ensure the safety of the product. Economic impacts of microorganisms are also important in the food industry from a safety viewpoint and also from a shelf-life viewpoint. Yeasts such as Saccharomyces cerevisiae are implicated as the primary microorganisms responsible for spoilage of fruit juices and their limited shelf life at refrigerated conditions. Objectives Because of these impacts on the fruit juice processing industry, the objectives of this study were the following: To characterize the thermal inactivation kinetics of Saccharomyces cerevisiae and Escherichia coli in orange juice To estimate thermal-death-time parameters (Dand z-value) for Escherichia coli subjected to an acid adaptation procedure vs. standard cultures in orange juice

PAGE 20

8 To compare the estimated parameters for Escherichia coli and Saccharomyces cerevisiae with published data. Methods and Materials Scope of Work The scope of work undertaken in this study has been divided into two parts to determine the thermal inactivation kinetics of Escherichia coli and Saccharomyces cerevisiae in single-strength orange juice. The Saccharomyces cerevisiae strain was a wild type isolated from orange juice, and the Escherichia coli strain was obtained from the American Type Culture Collection. Growth curves were created for each microorganism to determine logarithmic and stationary phases of growth. Preliminary experiments were used to help determine the temperature range in which thermal inactivation of both microorganisms would yield measurable numbers of survivors in order to plot survivor curves. After the appropriate temperatures were selected, microorganisms were subjected to different time-temperature combinations in order to estimate the thermal-death-time (TDT) kinetic parameters. These kinetic parameters were estimated by traditional methods of analyzing the survivor curves at each constant temperature. This method entailed estimating the decimal reduction times (D-values) using linear regression to construct the straight line of best fit on a semilog plot of survivors vs time (survivor curve). The D-value is the reciprocal slope of this curve expressed as time required for the curve to cross one log cycle, or time for one log cycle reduction of the population. A semilog plot of D-values vs. temperatures allows estimation of the z-value, by taking the reciprocal of the slope of the curve. The z-value is expressed as the number of degrees of temperature change required for one log cycle change in D-value.

PAGE 21

9 Preliminary Experiments Analysis of the survivor curves generated from the preliminary experiments helped determine at which temperatures to conduct the thermal inactivation experiments. For the Escherichia coli, a procedure was developed and implemented to adapt the cells to survival in a low-pH medium similar to the pH of single-strength orange juice. This procedure more closely modeled the conditions experienced by Escherichia coli that survive in contaminated orange juice. Two sets of preliminary experiments were conducted. The first set involved the thermal inactivation of both microorganisms grown in neutral-pH broth. The second set involved acid-adapted Escherichia coli grown in low-pH broth. Preparation of Cultures Source of strains The strain of Saccharomyces cerevisiae chosen for this study was obtained from the yeast culture collection maintained in the microbiology laboratory at the University of Floridas Citrus Research and Education Center, Lake Alfred, FL (Zook 1997). Stock cultures were streaked onto potato dextrose agar (PDA) and incubated at 30 o C for 72 hours. A loop full of cells was aseptically transfered to 200 mL screw-cap flasks of yeast extract peptone dextrose (YEPD) broth and incubated for 48 hours at 30 o C while continuously shaken at 120 rpm on a junior orbit table shaker. Small aliquots of this broth were then put into 1 mL vials placed into a o C freezer and maintained as a stock culture. A small loop full of broth was streaked onto slants of PDA refrigerated at 10 o C and used as a working culture for a period of 3 weeks. After 3 weeks a new working culture was created from the stock culture using the above procedure.

PAGE 22

10 Growth curves for this particular strain of Saccharomyces cerevisiae were documented by Zook (1997). A new set of growth curves was created to verify those results. A small aliquot of working culture was inoculated into a flask of 200 mL of YEPD broth and incubated at 30 o C. One-millimeter samples were withdrawn at predetermined timed intervals for 30 hours. Turbidity of the samples was measured optically using a Spectronic 40 spectrophotometer (Figure 1-1). As documented by Zook (1997), the yeast completed their logarithmic phase after approximately 17 hours of incubation. The strain of Escherichia coli (preceptol culture ATCC #9637) used in this study was obtained from the American Type Culture Collection (ATCC). Working and stock cultures of this strain were made from the original freeze-dried culture obtained from ATCC. The reconstituted cultures were inoculated into 200 mL of nutrient broth and incubated at 37 o C while shaken at 120 rpm for 48 hours. Small aliquots of broth were placed in 1 mL vials placed in a o C freezer and maintained as a stock culture. A small loop full of broth was streaked onto slants of nutrient agar, incubated for growth and refrigerated at 10 o C. These slants were used as the working culture and maintained for a period of 3 weeks. Thereafter new slants were prepared from stock cultures. Growth curves for Escherichia coli were created in the same manner as those for the Saccharomyces cerevisiae. In addition to measuring turbidity, the culture was plated out after reaching logarithmic phase to estimate the concentration of cells. The average concentration was 7.6 x 10 7 colony forming units (cfu)/mL after 25 hours and 13 x 10 8 cfu/mL after 36 hours. These numbers were used to estimate the proportion of inoculum to medium in order to maintain a high initial concentration during the thermal

PAGE 23

11 inactivation experiments (Figure 1-2). For the Saccharomyces cerevisiae it was desirable to use the cells while in the logarightmic phase (Zook 1997); whereas, for the Escherichia coli cells in the stationary phase were used (Buchanan and Edelson 1996, OHara and Glenn 1994, Parish 1999). Acid adaptation preparation During the first set of preliminary experiments with Escherichia coli, thermal inactivation was conducted by inoculating the medium with standard cultures (strains grown at approximately neutral-pH conditions). Results showed that these cultures had no resistance at all to the low-pH conditions of the orange juice at any lethal temperature. It was reasoned that the cells should be subjected to an acid adaptation procedure in order to increase their thermal resistance at low pH. This procedure would provide a closer approximation of the growth environment the microorganisms would experience if growing in contaminated orange juice. For the second set of preliminary experiments, the Escherichia coli cells were subjected to an acid adaptation procedure before thermal inactivation. In this procedure, 200mL of nutrient broth was inoculated with 1 mL of stock culture and incubated at 37 o C for 24 hours. After 24 hours 6 mL of sterile 5% citric acid solution was injected into the broth to lower the pH to approximately 4.5. The broth was then incubated for an additional 24 hours. Then another 6mL of sterile 5% citric acid was injected into the broth to lower the pH to approximately 3.5. The broth was then incubated for an additional 48 hours. After 48 hours of incubation, the cells were ready to be used in the thermal inactivation experiments. The final pH of the broth was at approximately 3.4. A sample of broth was extracted and plated out for enumeration and to measure final pH at each incubation interval. Below pH 3.7 there was a one or two log cycle

PAGE 24

12 reduction in viable cells between the standard culture grown in neutral-pH broth and those grown in low-pH broth (Table 1-1). Experimental Apparatus Heating at constant temperature was accomplished by using a three-neck flask apparatus to reduce the thermal lags associated with glass or stainless steel tubes submersed in a constant temperature bath. The flask was equipped with a mercury-in-glass thermometer, rubber stoppers, a reflux condenser, a set of 9 needles, a 10 mL syringe, eight 3 mL syringes, and a heating plate (Figure 1-3 and 1-4). The inoculated orange juice was continuously mixed with a magnetic stirrer. A condenser placed in the middle neck of the flask recovered evaporated water vapor from the orange juice to assure a constant volume of inoculum. Isothermal Inactivation Experiments The flask, magnet, needles, rubber stoppers, condenser, and syringes were sterilized before each experimental run. The thermometer was submerged in 10% ethanol alcohol for 30 min to sanitize. The orange juice was reconstituted using sterile filtered deionized water. The orange juice concentrate was a commercial brand at 44 o Brix. Reconstitution was performed under aseptic conditions using the recipe shown on the label (1 part concentrate to 3 parts water). A 100mL sample of reconstituted orange juice was aseptically poured into the flask. The flask was resealed using the rubber stopper, placed on a heating plate, and allowed to reach equilibrium at the desired treatment temperature. Then 7 mLs of inoculum was suctioned into one 10 mL sterile syringe (under aseptic conditions) and injected into the flask. The effect of injecting the inoculum, which was at incubation temperature, on the equilibrium temperature of the flask was determined by allowing a 100mL sample of orange juice to equilibrate at each experimental

PAGE 25

13 temperature. A thermocouple probe was used to measure the temperature drop of the heated sample as the inoculum was injected into the three-neck flask apparatus. While maintaining equilibrium conditions the temperature was observed over a period of 30 minutes for any significant change. The results indicated that for each 7 mL of inoculum injected into the flask the temperature of the orange juice was lowered by precisely 1 o C. This lowered temperature was held constant throughout the experiment, and recorded as the lethal temperature of exposure for the survivor curve resulting from that experiment. A sample of inoculum was plated out before thermal inactivation to determine the dilution of cells to be injected into the 100 mL of orange juice in the flask. After injecting the inoculum into the flask, the timer was started, and eight successive 1 mL samples were taken from each run at predetermined time intervals. The extracted 1 mL samples were quickly transferred by injection into 9 mL of sterile peptone water maintained in an ice water bath to immediately quench further thermal inactivation. After the last sample was taken, three dilutions at each time interval were prepared and plated in duplicate. Isothermal experiments were performed at 52 o C, 55 o C, 58 o C, and 62 o C. Estimating Dand z-values Four replicate experiments were conducted at each temperature. The D-values obtained from each replicate at the same temperature were averaged for a single representative D-value at each temperature. Statistical analysis was performed on these values to determine the standard deviation. The z-value ( o C) was estimated from the negative inverse slope of the linear regression line of the log D-value vs temperature. Statistical analysis was performed using Microsoft Excel spreadsheet program using the 4 replicates at each testing temperature.

PAGE 26

14 Results and Discussion Preliminary Experiments Saccharomyces cerevisiae Survivor curves for preliminary experiments conducted at 50 o C, 54 o C, and 56 o C for Saccharomyces cerevisiae are shown in Figure 1-5. Note that tailing was observed in all of the survivor curves. This tailing phenomenon can probably be attributed to the presence of two variant populations in the inoculum. For Saccharomyces cerevisiae the two populations consist of spores and vegetative cells. Saccharomyces cerevisiae is known to produce spores under normal growth patterns (Zook 1997). At the relatively low temperature used in the preliminary experiments the more heat-resistant spores remained viable to germinate in the media on enumeration of the survivors while the heat quickly inactivated the vegetative population of cells. To assure a more uniform population of yeasts, it would be necessary to separate the spores from the vegetative cells. This separation requires growing the yeast on media that encourages sporulation, separating the spores by centrifugation, and verifying uniformity of population by microscopy. Our laboratory was not equipped for this purpose, so further work on Saccharomyces cerevisiae was set aside for future study. Escherichia coli cultured at neutral pH Temperatures chosen for the preliminary experiments were based upon work by Line et al. (1991) and Blackburn et al. (1997). Line et al. (1991) estimated the Dand z-values of Escherichia coli O157:H7 in ground beef subjected to various temperatures. Although the heating characteristics for ground beef are different than those of orange juice, it was useful to know the expected Dand z-values for nonpathogenic Escherichia coli. Line et al. (1997) estimated D-values of 78.2 min at 51.6 o C, 4.1 min at 57 o C, and 18

PAGE 27

15 sec at 62.7 o C in fatty ground beef. Blackburn et al. (1997) performed experiments with E coli O157:H7 in solutions that varied with pH and NaCl concentration. At 0.5% w/w concentration of NaCl and pH of 4.3 (closest to pH of the orange juice at 3.8) the D-values at 62.5 o C were 19 seconds, 34 seconds, 15 seconds, and 33 seconds for each specific strain of O157:H7. Using the results from both of these studies, the temperatures chosen for the preliminary experiments were 59 o C, 62 o C, and 64 o C in an attempt to show a significant difference between the D-values at each respective temperature. Survivor curves obtained from preliminary experiments conducted at 59 o C, 62 o C, and 64 o C with Escherichia coli cultured at neutral pH are shown in Figures 1-6 to 1-8. The TDT curve resulting from these experiments is shown in Figure 1-9, with a z-value of 6.4 o C. As shown in Figures 1-6 and 1-7 nearly all survivor curves showed tails at 59 o C and 62 o C. Therefore, D-values were obtained from the initial linear portion of the curves. Results of these replicates at each temperature are shown in Table 1-2. It should be noted that at the highest temperature (64 o C) the effective D-value was 1.2 seconds. With such a rapid decrease in the population of survivors over a 10 second interval, a sample extraction interval time of less than 5 seconds was needed to get countable plates which yielded at least 4 data points for each survivor curve. With the current technique for conducting isothermal bath experiments, this sample extraction interval was too short for one individual to perform accurately. The tailing phenomenon was observed only at the lower temperatures of 59 o C and 62 o C. The presence of tails suggested that a small fraction of the population was more tolerant of these conditions. It was postulated that the two populations likely differed in their tolerance to the acidic conditions of the orange juice. During this first set of

PAGE 28

16 preliminary experiments the Escherichia coli cells were cultured in neutral-pH broth and inactivated in low-pH orange juice. Existence of an acid-tolerant culture within the inoculum was suspected to account for the appearance of tailing. Since acid will inactivate vegetative cells the combination of it and the heat quickly kills the population that is relatively susceptible to acid, whereas the more resistant population persists. The lower temperatures used during the preliminary experiments were not high enough to inactivate the remaining resistant population of Escherichia coli, yet this was the population of greatest concern. Therefore, it became necessary to achieve a more heat-resistant acid-tolerant population. Acid-tolerant Escherichia coli cultures To test this hypothesis of the existence of acid tolerant subpopulations in the inoculum, a second set of preliminary experiments for the Escherichia coli was conducted using acid-tolerant cultures. Figures 1-10 to 1-12 show survivor curves obtained from these preliminary experiments for the acid-tolerant cultures at 52 o C, 55 o C, and 60 o C (Figure 1-13 shows the family of curves). Figure 1-14 shows the TDT curve resulting from these experiments at low pH. Table 1-3 lists the D-values obtained from analysis of the survivor curves at each temperature. The acid-adapted cultures displayed more resistance to heat than the non-acid-adapted Escherichia coli cultures. A comparison of the Escherichia coli grown in nutrient broth where the pH had not been adjusted vs adjusted pH nutrient broth showed a clear distinction between the thermal resistances of the cultures. The tailing observed in the survivor curves of the Escherichia coli grown in neutral broth did not show up in the survivor curves of the Escherichia coli grown in low-pH broth. At each replicate a sample was taken at a sufficiently long interval and plated out. The plates showed no growth at any of the temperatures for the

PAGE 29

17 isothermal experiments conducted with the acid adapted cultures. At 52 o C, 55 o C, 58 o C, and 60 o C the extended interval where no growth appeared on the plates was 56 min,15 min, 3 min, and 1.5 min, respectively. These results show that a more uniform population existed among the cells of the acid-adapted Escherichia coli. The acid adaptation procedure was successful in achieving its goals (elimination of the tailing phenomenon and higher thermal resistance). The difference in the thermal resistance between the two cultures along with the elimination of the tailing phenomenon demonstrated the importance of acid adaptation of the inoculum when working with low-pH fruit juices such as orange juice. Thermal Inactivation of Escherichia coli Based on results from the acid-tolerant preliminary experiments the best temperatures selected to give a significant difference between D-values were 52 o C, 55 o C, 58 o C, and 60 o C. At these temperatures the extraction intervals ranged from 7 minutes to 10 seconds. These times were appropriate to allow a sample to be taken at precise time intervals. Since pH was a major factor contributing to thermal inactivation of Escherichia coli, it was important to measure the pH for consistency during each experimental run. The pH of the orange juice used in the isothermal inactivation experiments vs the pH of the growth broth before inoculation of the Escherichia coli into the orange juice is shown in Figure 1-15. The pH of the orange juice ranged from 3.74 to 4.11 (a difference of 0.36) whereas the pH of the broth ranged from 3.29 to 4.09 (a difference of 0.8). For the orange juice the difference between the minimum and the maximum pH yielded no change in the number of survivors. To account for the difference in pH ranges, dilutions

PAGE 30

18 were plated out at one above and one below the target dilution. This method would also account for any variation in the initial concentration of cells. The isothermal survivor curves for Escherichia coli at 52 o C, 55 o C, 58 o C, and 60 o C are shown in Figures 1-16 through 1-19, respectively (Figure 1-20 shows the family of curves). Table 1-4 shows the results of the thermal inactivation experiments for Escherichia coli. The D-values were determined by taking an average of all the D-values for all the replications at each temperature. The standard deviation for D-values at each temperature was within 10% of the average value, thus the variation in the D-values among replications was not a significant source of experimental error. The TDT curve for the z-value of Escherichia coli in orange juice is shown in Figure 1-21. The z-value for this microorganism in orange juice was found to be 6.0 o C. This value agrees closely with the z-value from the preliminary experiments with the acid tolerant cultures. The R 2 -value from regression analysis was 0.98. These results were compared with those reported in the literature for the thermal inactivation of Escherichia coli in orange juice (Table 1-5). The cultures in this study were subjected to an acid-adaptation laboratory procedure before inoculation using a non-acid-resistant, low-heat-resistant strain of generic Escherichia coli, whereas Mazotta (2001) and Splittstoesser et al. (1996) used a naturally-occuring, acid-tolerant, pathogenic strain isolated from patients who had consumed contaminated product and showed clinical symptoms of Eshcherichia coli infection. Because of the natural genetic differences between generic and pathogenic strains of Escherichia coli, difference in heat resistance results among the three studies were expected. More importantly the Mazotta (2001) and Splittstoesser et al. (1996) study was expected to produce TDT kinetics

PAGE 31

19 different than those estimated in this study. Mazotta used single-strength orange juice adjusted to a pH of 3.9 with 1 N NaOH while Splittstoesser and colleagues used freshly prepared apple cider and commercial brand apple juice concentrates. Similar to this study, Mazotta conducted two sets of experiments using acid adapted and non-acid adapted cultures. Both this study and Mazottas showed a significant difference in the heat resistance between acid adapted and non-acid adapted cultures. This difference has a significant impact on the kinetic parameters estimated by thermal inactivation experiments with orange juice. Table 1-5 shows the D-values for Escherichia coli from all three studies. For both our study and Mazottas study, thermal inactivation kinetic parameters differ significantly between cultures grown in standard broth and those grown in pH-adjusted broth. In both studies acid-adapted cultures were at least twice as resistant as the non-acid-adapted cultures to thermal inactivation. The acid tolerance of Escherichia coli is important to their survival in low-pH products and may prove to be an important component of virulence for this species of bacteria (as it is able to survive the acidic conditions of the stomach, which relates to the infective dose). The acid tolerance of Escherichia coli significantly affects its thermal inactivation characteristics. Our study shows the value of acid adaptation before performing thermal inactivation experiments in low-pH products. The traditional recommended pasteurization treatment for orange juice (98 o C for 10 seconds) significantly affects the flavor of orange juice when compared with fresh untreated orange juice (Parish 1998). Parish (1998) showed that a 23 degree decrease in the temperature with the same treatment time had an impact on the sensory characteristics of orange juice.

PAGE 32

20 Most consumers prefer unpasteurized orange juice products to pasteurized products. However the recent outbreaks of disease associated with unpasteurized fruit juices has magnified the risk to consumer of these products. Data in this study suggest that a minimal treatment process can achieve the necessary reduction in population of pathogenic Escherichia coli in orange juice to a level that is safe for the consumer. With parameters estimated in this study the calculated thermal process time that will reduce the population of the acid-adapted Escherichia coli by 6 log cycles at a hold tube temperature of 67 o C is 11 seconds; whereas for the non-acid-adapted culture it would be 3.2 seconds, and could result in an unsafe product. The same difference in process time between acid-adapted and non-acid-adapted cultures was shown for the strain used in Mazottas study. The thermal process time for a 6.0 log cycle reduction of the acid-adapted culture at 67 o C is 22.81 seconds; whereas for the non-acid-adapted culture the thermal process time at the same hold tube temperature is 13.74 seconds. These process times differ by 39.7%. Results of both studies emphasize the importance of conducting experiments with cultures that are similar to those found in the product. Using the thermal inactivation kinetics from the non acid-adapted cultures from both studies leads to a significant difference in the final population of microorganisms present in the product.

PAGE 33

21 00.511.522.5051015202530Time (Hrs)ABS@600 nm Set One Rep 1 Set One Rep 2 Set One Rep 3 Set Two Rep 1 Set Two Rep 2 Figure 1-1. Growth curves showing light absorbance at a wavelength of 600 nanometer vs time for Saccharomyces cerevisiae in yeast extract peptone dextrose (YEPD) broth. Sets are runs conducted on separate days

PAGE 34

22 00.10.20.30.40.50.60102030405060708Time (Hrs)ABS@600 nm 0 Set two Rep 1 Set two Rep 1 Set three Set two Rep 2 Set two Rep 1 Figure 1-2. Growth curves showing absorbance of light at wavelength of 600 nanometer vs time for Escherichia coli ATCC #9637 in nutrient broth. Sets are experiments conducted on separate days

PAGE 35

23 Table 1-1. Plate counts of survivors grown in standard nutrient broth and pH-modified nutrient broth for inducing acid tolerance Acid-adapted Culture Non-acid-adapted Culture Incubation Hours Total Amount of Acid added (mL) pH of broth Plate Count (cfu) pH of broth Plate Count (cfu) 48 3 6.729 2.8 x 10 9 2.5 x 109 8.1 4.3 x 10 9 3.2 x 109 72 6 4.760 2.2 x 10 9 2.4 x 109 8.2 3.6 x 10 9 1.7 x 109 96 10 3.694 1.6 x 10 8 1.4 x 108 8.4 3.1 x 10 9 1.2 x 109 120 12 3.360 1.2 x 10 7 1.5 x 107 8.4 1.4 x 10 9 7.6 x 108

PAGE 36

24 Figure 1-3. Experimental apparatus (photograph)

PAGE 37

25 Figure 1-4. Experimental apparatus (diagram)

PAGE 38

26 01234567890100200300400500600700Time (sec)Log[survivors(cfu/ml)] 50 C 54 C 56 C Figure 1-5. Survivor curves from preliminary experiments at 50 o C, 54 o C and 56 o C for Saccharomyces cerevisiae in orange juice cultured at neutral Ph (standard culture)

PAGE 39

27 0246810050100150200250300Time (sec)Log[survivors(cfu/ml)] Run 1 Run 2 Run 3 Figure 1-6. Survivor curves from preliminary experiments at 59 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture)

PAGE 40

28 0246810050100150200250300Time (sec)Log[survivors(cfu/ml)] Run 1 Run 2 Run 3 Figure 1-7. Survivor curves from preliminary experiments at 62 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture)

PAGE 41

29 024681005101520Time (sec)Log[survivors(cfu/ml)] Run 1 Run 2 Figure 1-8. Survivor curves from preliminary experiments 64 o C for Escherichia coli in orange juice cultured at neutral pH (standard culture)

PAGE 42

30 00.10.20.30.40.50.60.70.80.915859606162636465Temperature (oC)Log[D-value(min)] Figure 1-9. TDT curve from preliminary experiments with Escherichia coli in orange juice cultured at neutral pH (standard culture). R 2 value of 0.90

PAGE 43

31 Table 1-2. D-values (seconds) for Escherichia coli in orange juice cultured at neutral pH (standard culture) in preliminary experiments Temperature Replicate 59 o C 62 o C 64 o C 1 6.25 4.81 1.1 2 7.14 3.57 1.3 3 6.55 2.95 NA Average 6.64 3.77 1.2 Std Deviation 0.45 0.94 0.14 z-value = 7.0 o C

PAGE 44

32 02468101201000200030004000Time (sec)Log[survivors (cfu/ml)] Run 1 Run 2 Figure 1-10. Survivor curves from preliminary experiments at 52 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture).

PAGE 45

33 0123456789100200400600800Time (sec)Log[survivors (cfu/ml)] Rep 1 Rep 2 Rep 3 Figure 1-11. Survivor curves from preliminary experiments at 55 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 46

34 012345678910050100150Time (sec)Log[survivors (cfu/ml)] Rep 1 Rep 2 Figure 1-12. Survivor curves from preliminary experiments at 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 47

35 012345678910050010001500200025003000Time (sec)Log [survivors(cfu/ml)] 52 C 55 C 60 C Figure 1-13. Family of survivor curves from preliminary experiments at 52 o C, 55 o C, and 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 48

36 11.21.41.61.822.22.42.62.850525456586062Temperature (oC)Log[D-value(min)] Figure 1-14. TDT curve from preliminary experiments with Escherichia coli in orange juice cultured at low pH (acid adapted culture). R 2 value of 0.99

PAGE 49

37 Table 1-3. D-values (seconds) for Escherichia coli in orange juice cultured at low pH (acid adapted culture) in preliminary experiments Temperature Replicate 52 o C 55 o C 60 o C 1 424.25 112.21 16.05 2 342.2 100.4 16.46 3 136.9 Average 383.2 116.4 16.3 Std Dev 58.01 18.79 0.29

PAGE 50

38 33.23.43.63.844.205101520ReplicatespH Orange Juice Broth Figure 1-15. pH of broth vs. pH of orange juice product for Saccharomyces cerevisiae preliminary experiments

PAGE 51

39 012345678910050010001500200025003000Time (sec)Log[survivors (cfu/ml)] Rep 1 Rep 2 Rep 3 Rep 4 Figure 1-16. Survivor curves from thermal inactivation experiments at 52 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 52

40 01234567890100200300400500600700Time (sec)Log[survivors (cfu/ml)] Rep 1 Rep 2 Rep 3 Rep 4 Rep 5 Figure 1-17. Survivor curves from thermal inactivation experiments at 55 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 53

41 0123456789050100150200Time (sec)Log[survivors (cfu/ml)] Rep 1 Rep 2 Rep 3 Figure 1-18. Survivor curves from thermal experiments at 58 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 54

42 01234567802040608010Time (sec)Log[survivors (cfu/ml)] 0 Rep 1 Rep 2 Rep 3 Rep 4 Figure 1-19. Survivor curves from thermal inactivation experiments at 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 55

43 02468101205001000150020002500Time (sec)Log [survivors (cfu/ml)] 52 C 55 C 60 C 58 C Figure 1-20. Family of survivor curves at 52 o C, 55 o C, 58 o C and 60 o C with Escherichia coli in orange juice cultured at low pH (acid adapted culture)

PAGE 56

44 Table 1-4. D-values (seconds) from thermal inactivation experiments for Escherichia coli cultured at low pH D-values at various temperatures (seconds) Replicate 52 o C 55 o C 58 o C 60 o C 1 398 151 36 16 2 370 148 32 20 3 308 146 34 18 4 336 147 19 Average 353 148 34 18 Std Deviation 39.08 2.18 2.27 1.52 z-value = 6.0 o C

PAGE 57

45 11.21.41.61.822.22.42.650525456586062Temperature (oC)Log [D-value(sec)] Figure 1-21. TDT curve from the thermal inactivation experiment with Escherichia coli in orange juice cultured at low ph (acid adapted culture). R 2 value of 0.98

PAGE 58

46 Table 1-5. Comparison of TDT kinetic parameters with published data from Mazzotta (2001) and Splittstoesser et. al. (1996) using acid adapted and non-acid adapted Escherichia coli in orange juice D 58 (sec) Acid Adapted This study 34 Mazzotta 300 Non-acid Adapted This study 10 Mazzotta 198 Splittstoesser et al. 60

PAGE 59

CHAPTER 2 ESTIMATING KINETIC PARAMETERS FOR THERMAL INACTIVATION OF Escherichia coli IN ORANGE JUICE USING THE PAIRED EQUIVALENT ISOTHERMAL EXPOSURES (PEIE) METHOD WITH A CONTINUOUS HIGH TEMPERATURE SHORT TIME (HTST) PROCESS TREATMENT Introduction Achieving the best balance between quality retention and safety in heat sensitive products that must be pasteurized is important in the fruit juice processing industry. Recent outbreaks of Escherichia coli O157:H7 and Salmonella in products such as orange juice, apple juice, and apple cider have emphasized the ability of these microorganisms to survive and grow in low-pH environments. Processing these products to sufficiently reduce the probability of microbial survival for food safety and spoilage is an essential design objective for food engineers. However, the popularity of unpasteurized fruit juice is growing because of better flavor and texture retention over heat pasteurized products. Understanding the thermal inactivation behavior of the target microorganisms in the product is one key requirement to achieve a good balance between food safety and quality retention. This behavior can be quantified by estimation of the Dand z-values of the target microorganisms if first-order inactivation kinetics are observed. The temperature-time combination for specific process goals can be determined using these estimated thermal inactivation parameters. The greater the accuracy of these estimations the more precise the temperature-time process condition can be determined 47

PAGE 60

48 for the product. There are several methods used to estimate the thermal inactivation parameters of microbial and chemical constituents. These methods include the following: Isothermal bath immersion with vials Isothermal three-neck flask Isothermal hold tube with sampling ports Paired equivalent isothermal exposures (PEIE) from non-isothermal data Because estimated thermal inactivation parameters can have a significant impact on the design of thermal treatment processes, it is essential to know which method provides the best estimation of the parameters. The purpose of this study was to compare the thermal inactivation kinetic parameters estimated by the traditional method of isothermal bath analysis with those estimated by the PEIE method. It is possible to determine which method provides the best estimation of the kinetic parameters by comparing the number of survivors from a dynamic thermal process predicted mathematically using parameters from each method with actual experimental survivor data. Literature Review Accurate estimation of kinetic parameters describing thermal inactivation of microbial populations is of crucial importance in designing thermal treatments for sterilization or pasteurization of liquid food products. Difficulty in achieving accurate parameter estimation often leads to over processing in order to minimize risk to public health. For products that are sensitive to heat this over processing comes at the expense of flavor and nutrient degradation. In a study performed by Parish (1998) to compare orange juice quality after treatment by thermal and isostatic high pressure pasteurization, the orange juice processed at 75 o C for 10 seconds had a closer sensory score to fresh extracted, frozen orange juice than that processed at 98 o C for 10 seconds. The study also indicated that the flavor degradation after 16 weeks of storage at 4 o C and 8 o C was worse

PAGE 61

49 for the product processed at the higher temperature. The results of this study showed the importance of minimizing the thermal exposure to heat-sensitive products. Greater accuracy in estimating kinetic parameters of thermal inactivation will allow food processors to achieve maximum product quality without compromising food safety. The logarithmic order of bacterial death is commonly described by a straight line on a semilog plot of concentration of viable microorganisms vs. time of exposure to a constant lethal temperature called a survivor curve. Survivor curves and their temperature dependency are used as a mathematical model to determine the temperature-time requirements for a pasteurization process. Commercial pasteurization processes rely on such modeling of microbial population dynamics to design and operate thermal processes for proper application of heat necessary to assure stability and safety of food products, while reducing unnecessary overexposure of the products to heat, which can severely degrade the quality of the products. Consumer demand for high quality processed foods often drives the need for designing processes that are less detrimental to product quality such as flavor and texture, while still reducing the microbial population to levels that ensure safety from food borne illness. First-order kinetics The classical model of a first order reaction has been used for decades to predict the processing temperature-time relationship of microbial thermal inactivation. Food scientists and engineers have used survival curves, obtained from isothermal bath experiments at different temperatures, as a means to estimate the kinetic parameters describing these first-order reactions. These experiments are conducted by inoculating a sample with a specific population of viable microorganisms and submerging vials containing the sample into a constant temperature bath. These vials are removed at

PAGE 62

50 different time intervals to obtain different extents of reaction that can be represented by points on a survivor curve. The significant problems with isothermal bath experiments are: Limited temperature range from which to calculate parameters for a wide range of temperatures and to select parameters with good statistical confidence (Welt et al. 1997). Time lag of heat transfer encountered when the samples are heated from ambient to reaction temperature and when cooled down from reaction temperature. Tedious preparation of small samples required to reduce thermal lags. Need for using buffer solutions rather than actual food product in many cases. Significant difference between experimental and actual processing conditions. Difficulty in obtaining statistically valid data at high temperatures when very rapid reaction rates require short exposure times that cannot be accurately controlled. Because of these problems, the use of kinetic parameters estimated by analysis of data from isothermal batch experiments performed using vials submerged in a constant temperature bath has often lead to inaccurate results when characterizing a continuous ultra high temperature (UHT) or high temperature-short time (HTST) process such as those used in commercial pasteurizations. An alternative technique for conducting isothermal experiments involves using a three-neck flask instead of submerged vials into an isothermal bath. This technique dramatically minimizes the thermal lags experienced by the sample but it still has many of the problems associated with isothermal bath experiments. Wescott et al. (1995) proposed using a continuous thermal process to gather isothermal data for the construction of survivor curves used to determine the thermal inactivation kinetics of microorganisms. Measuring the number of survivors at various locations in the hold tube and assuming a constant temperature (isothermal) over the

PAGE 63

51 entire length of the hold tube, survivor curves can be constructed for various temperatures. The Dor k-values can be determined for each temperature and used to construct a TDT curve or an Arrhenius plot from which a z-value or activation energy value can be determined. This type of analysis of a continuous dynamic process from obtaining isothermal data has been termed traditional analysis of a continuous dynamic thermal process. One problem with this method is the assumption of isothermal conditions along the hold tube. Because UHT/HTST systems operate at very high temperatures the rate of inactivation is very rapid and a small change in the temperature will yield different thermal inactivation parameters from those based upon nominal operating temperature. This method also can be highly dependent upon the experimental technique of the researcher and how well the UHT/HTST system maintains constant temperature along its hold tube. The PEIE Method To facilitate the design of a UHT/HTST process, Swartzel (1984) developed the Equivalent Point Method (EPM) as a practical non-isothermal method for kinetic parameter estimation. The EPM operates on the premise that any number of equivalent isothermal processes may be obtained for a given dynamic thermal process so long as a temperature-time profile is known as well as the extent of reaction. These equivalent temperature-time combinations will fall on a straight line when plotted on a semilog temperature-time graph for any assumed value of activation energy. He further postulated that other straight lines constructed for different values of activation energy would have different slopes and all intersect at a common universal equivalent point from which the true value for the rate constant and activation energy could be determined

PAGE 64

52 when two such equivalent points are found from two different dynamic processes with different extents of reaction. Welt et al. (1997 a, b) discovered that although Swartzels universal equivalent point did not exist, the concept of substituting dynamic processes with equivalent isothermal processes could still be used to obtain kinetic parameters by employing an iterative technique called the Paired Equivalent Isothermal Exposures method (PEIE). They demonstrated the use of this method to estimate values for kinetic parameters that were close to published values determined from traditional analysis of isothermal bath data for Bacillus stearothermophilus spores in pea puree. Vieira et al. (2001) used this method to estimate kinetic parameters for ascorbic acid degradation and later for thermal inactivation parameters of Alicyclobacillus acidoterrestris spores in fruit nectar (Vieira et al. 2002). With the PEIE method, it is no longer necessary to perform isothermal bath experiments in order to estimate the reaction kinetics of reactants, whether they are microorganisms, vitamins, or flavor components. Using a UHT/HTST process that more accurately simulates the conditions the product will experience, the kinetic parameters can be more accurately estimated and the temperature used is only dependent upon the design parameters of the equipment and/or process. The PEIE method is a potential tool for obtaining the kinetic parameters of a first order reaction more accurately than from isothermal bath experiments. Objectives The purpose of this study was to apply the PEIE method to estimate thermal inactivation kinetic parameters of Escherichia coli in orange juice and compare those with parameters estimated using traditional analysis of survivor curves from isothermal experiments. To achieve these goals the objectives of this project were the following:

PAGE 65

53 Estimate the kinetic parameters for thermal inactivation of Escherichia coli in orange juice using the PEIE method with a continuous HTST process treatment. Compare the kinetic parameters estimated from the PEIE method with those estimated from a traditional isothermal bath method. Validate the results by subjecting samples of inoculated product to random dynamic temperature exposures beyond the range of temperatures used for parameter estimation and then comparing the final population of surviving microorganisms predicted from both sets of model parameters with the actual population of microorganisms enumerated in the laboratory. Methods and Materials Preparation of Cultures The strain of Escherichia coli used in these experiments was a preceptol culture ATCC #9637. The microorganism was prepared in the same manner as detailed in Chapter 1. These cells were also subjected to an acid adaptation procedure prior to thermal inactivation as detailed in Chapter 1. Experimental Apparatus The experimental apparatus used in these experiments was the Microthermics UHT/HTST Lab-25 lab-scale pasteurizer unit (Figure 2-1). All the heat exchangers of the apparatus were shell and tube. The unit had two product inlets leading to the product pump (Figure 2-2). Each inlet was equipped with a plug valve to control product flow. The system was started by connecting a product reservoir to one inlet and a water reservoir to the other inlet. The valve to the water reservoir was opened to provide water to the system while operating conditions were being established and stabilized. Once the system had reached steady state (stabilized at the desired operating conditions), the valve to the product reservoir was opened to introduce product as the valve to the water reservoir was closed. The main body of the pasteurizer was divided into three sections consisting of the heater, hold tube, and chiller sections. Both the heater and chiller

PAGE 66

54 sections were shell and tube heat exchangers made of stainless steel tubing with an outer diameter of 0.375 in (0.9525 cm), wall thickness of 0.049 in (0.1244 cm) and length of 228 in (579.12 cm). The hold tubes also had an outer diameter of 0.375 in (0.975 cm) but wall thickness of 0.035 in (0.0889 cm) and length of 200 in (508 cm) for each section of the hold tube for a total length of 400 in (1016 cm). Using hot water as the heating medium, the temperature of the product exiting the heater was controlled by adjusting the steam pressure used to generate the hot water by a manual pressure flow control valve. The hold tube section consisted of a series of tubes whose length could be adjusted by adding extension tubes at the hold tube jumper panel. Hold times varied according to the flow rate of the product and extension tubes used to extend the length of the hold tube section for the appropriate residence time. Adjusting the speed of the product pump controlled the flow rate of the system, which was measured by collecting a volume of product exiting the system in a known period of time. The chiller section used a 50/50 mixture of water and propylene glycol as the cooling medium. To maintain pressure when the product temperatures approached their boiling point in the system, an adjustable back-pressure valve was located after the chiller section prior to the product exiting the system. To monitor the temperature of the product and heating medium, thermocouple probes were located at various points within the flow stream of the product and heating medium. Calibration of Thermocouples Thermocouples were calibrated by comparing the temperature reading from each thermocouple with the temperature reading from a standardized mercury-in-glass tube thermometer (Arthur H. Thomas Company, National Bureau of Standards, Bureau file 117084) in a constant water bath. Correction factors for each thermocouple are shown in

PAGE 67

55 Table 2-1. The average offset for each thermocouple was programmed into the datalogger to eliminate the temperature reading as a significant source of experimental error. Continuous Dynamic Thermal Treatments A commercial brand orange juice concentrate at 44 o Brix was reconstituted using sterile filtered deionized water. The reconstitution was performed following the recipe indicated on the label (1 part concentrate to 3 parts water). Although the orange juice was not reconstituted under aseptic conditions, the resident population of Escherichia coli in the product was negligible when compared with the number of cells in the inoculum, and the product was subjected to a thermal treatment within 30 minutes of reconstitution. The product was inoculated with an acid-adapted Escherichia coli cell suspension prior to thermal exposure to achieve a minimum initial concentration of 1 x 10 8 cfu/mL. Five liters of orange juice were prepared along with 800ml of cell suspensions. The pasteurizer was sanitized by circulating hot water at 83 o C through the heater, hold tube, chiller sections and accessory tubes for a minimum of 30 minutes. Once the sanitation cycle was completed the temperature of the pasteurizer was adjusted to the desired experimental temperature and allowed to reach steady-state conditions, upon which the product flow control valve was opened to allow the inoculated product to flow through the unit. Temperatures at various locations throughout the system were recorded using a datalogger attached to a notebook computer. The pasteurizer unit had three thermocouples installed in the product flow stream and one in the heating medium flow stream of the unit. The thermocouples were located after the heating section, after the hold tube section, after the chiller section, and the flow tube of the heating medium. All thermocouples were copper-constantan type T. The thermal

PAGE 68

56 profile (temperature vs. time) of each experimental run was captured from each port and saved as a text file that was used in the PEIE method. To produce replicate data for each temperature, samples were collected in triplicate for each experimental run, and a minimum of two experimental runs were conducted for each temperature-time combination. An experiment involved a product cycle whereby a batch of product was pumped through the system after using water to achieve a stable steady state condition. Then the product and water reservoir valves were switched to allow water to run through the system at the same conditions while another sample of product was being prepared. Then the valves were switched and the product was pumped through the system and samples were taken once again. Reynolds numbers for each experimental run indicated transitional flow (Table 2-2). Although the PEIE method is not dependent upon the flow behavior of the fluid in the pasteurizer unit, the flow behavior characteristics will influence the designed residence times. Temperature Profiles The temperature was measured at the inlet of the product (initial product temperature), after the heating section (at the entrance to the hold tube), after the hold tube, and after the chiller section. Using the recorded temperature at each point, the heater and chiller portions of the profile were constructed from heat transfer equations, while the hold tube portions were constructed based upon measured data. The standard profile for a shell and tube heat exchanger follows an exponential increase that can be described by Equation 2-1, )-1(-hteBAT (2-1)

PAGE 69

57 where T is the temperature at any point within the heat exchanger at a specific time t, A is the initial temperature of the product, B is the temperature of the product upon exit from the heat exchanger, and h is the rate constant for the temperature change through the heat exchanger. This equation yielded the calculated temperatures along the heater section of the pasteurizer. The hold tube inlet and outlet temperatures were measured directly by thermocouples. The temperature along the chiller section of the pasteurizer was calculated using Equation 2-2. )(-cteBT (2-2) where B is the temperature of the product upon entrance into the chiller section, T is the temperature at any point within the chiller section at a specific time t and c is the rate constant for the temperature change through the chiller section. Knowing the residence time of the product within the heater and chiller section of the pastuerizer, the temperature profile was constructed by determining the parameters of Equations 2-1 and 2-2 using the boundary conditions of each section. The residence times for each section were determined based upon the flow rate of the product and the diameters and lengths of the tubes in all sections of the pasteurizer with the assumption of plug flow for simplicity. The flow rates of the product were determined by measuring the amount collected in a graduated cylinder over a specific time period. Estimating Dand z-Values with the PEIE Method The PEIE method uses the knowledge that for a given dynamic thermal exposure, there exist any number of equivalent isothermal exposures (EIEs) that would yield the same reduction in concentration of reactant. From two different dynamic thermal exposures for a given reactant, the kinetic parameters for thermal inactivation of that

PAGE 70

58 reactant can be estimated. The PEIE method as detailed by Welt et al. (1997a,b) is carried out in Arrhenius kinetics to estimate first order rate constants (k) and activation energy (E a .) These parameters were converted into Dand z-values at the end of the process. The following steps were taken from Welt et al. (1997a, b) and outline the PEIE method used in this work: 1. The temperature histories along with the initial and final concentration of the reactants from at least two distinct dynamic thermal processes were recorded. Distinct means that each process produced a different extent of reaction. 2. One E a value (E a 1) was arbitrarily selected and the other E a value (E a 2) was arbitrarily chosen at 1.5 times E a 1. 3. Using the recorded temperature-time data and the selected E a 1 and E a 2 values, the respective EIEs (equivalent time (t e ) and temperature (T e )) for the pair of dynamic thermal experiments were determined by equation 2-3, where G is the product constituent reduction relationship factor, R is the universal gas constant (J/mole-K), T(t) is temperature-time data, t e is the equivalent time, and T e is the equivalent temperature. eaetaTREtdttTREGexp)(exp0 (2-3) Equation 2-3 was applied twice for each data set using E a 1 first, then E a 2. This application yielded two lines, each of which represented an infinite set of temperature time combinations that were equivalent isothermal exposures for respective Ea-values. The intersection of these two lines gave the equivalent time and temperature for an isothermal process that would yield the same extents of reaction as the dynamic thermal exposure for the reactants characterized by the E a 1 and E a 2. This point is an Equivalent Isothermal Exposure. 4. The isothermal rate constants, k, for each process pair were calculated using the EIE specification (t e and T e ) from step 3, the extent-of-reaction data from step 1 and Equation 2-4, where C o is the initial concentration at time zero and C is the concentration of survivors remaining at the end of the process time. eotCCkln (2-4) Equation 2-5 determined the D-value.

PAGE 71

59 k valueD303.2 (2-5) 5. Each pair of k values calculated from step 4 along with the equivalent temperature from step 3 was used in Equation 2-6 to estimate an E a value. )(ln212121eeeeaTTTTkkRE (2-6) 6. The newly estimated E a value was used as the initial guess (Step 2) for the next iteration. The process was repeated until the estimated E a value from step 5 stopped changing. A TDT curve of D-value versus temperature was plotted to estimate the z-value. An algorithm using a commercial software package (Mathcad for Windows Version 8.0) was used to execute the PEIE steps using the recorded thermal history, and population survivor data (extent of reaction). It is important to note that the PEIE method only works with constituents that follow a first order reaction. Validation Experiments The validation aspect of this study involved comparing the predicted number of survivors for a particular process using the kinetic parameters estimated by the PEIE method and those estimated by the 3-neck flask isothermal method (see Chapter 1) with the actual number of survivors obtained from plate count enumeration of inoculated orange juice. The validation experiments were performed with the same strain of Escherichia coli and the same lab-scale pasteurizer unit. The inoculated orange juice product was subjected to a dynamic process whereby the temperature of the heating medium was varied to give a changing hold tube temperature. Samples of the product were collected at a predetermined interval and serial dilutions were prepared, plated out on nutrient broth and incubated for 48 hours. To observe if any injured cells were able

PAGE 72

60 to recover the plates were incubated for an additional 24 hours and the number of survivors was compared with those from the first 48 hours for any significant differences. There were not significant differences between the two plant counts. The predicted number of survivors was calculated by using numerical integration over the temperature-time profile of each validation process, as follows. Inactivation of vegetative cells at a constant lethal temperature follows a first-order reaction process that is described by Equation 2-7 when C represents the concentration of surviving viable cells, D, decimal reduction time, is the time interval required to reduce the population of viable cells one log cycle (90%) of its former value (D=ln(10)/k), t is the exposure time, and C o is the initial number of viable cells. DtoCC10 (2-7) Since the rate of population reduction is dependent on temperature, Equation 2-8 was used to describe the variation of D with temperature T, zTTooDTD10)( (2-8) where D o is reference D-value at reference temperature T o and z is the temperature interval required to change the value of D by one log cycle. For a non-isothermal process where T varies with time, the lethal effect of the changing temperature on the population can be determined by dividing the temperature history into small time intervals (t) of constant temperature, use Equation 2-8 to compute the D-value for each interval, and estimate the reduction in the population from its former value using Equation 2-7 for each time interval. This process yields Equation

PAGE 73

61 2-9, which can be used to find the change in the initial concentration of survivors over the time interval t for a given temperature history, T (t). tDtttzoTtToCC)(10110 (2-9) For a time increment of differential magnitude, the total lethal effect over the total process time is found by adding the contribution of all the time intervals to yield Equation 2-10. ttdtDoozoTtToCC)(10110 (2-10) Using the D o value obtained from the TDT curve from both methods, Equation 2-10 was solved by numerical integration to estimate the number of surviving viable cells (C) for each validation process. Equation 2-10 is the mathematical model used to predict the relationship between survival response and the temperature history for a given set of kinetic parameters. The predicted number of survivors was compared directly with the number of survivors enumerated from plate count techniques. Results and Discussion Continuous Dynamic Thermal Experiments Parameter Estimation Figures 2-3 through 2-5 show the temperature histories for the continuous dynamic thermal experiments with hold tube temperatures at 58 o C, 60 o C, and 62 o C, respectively. Each temperature included two different residence times in order to get two processes with different extents of reaction. Temperature rate constants determined for the heater and chiller sections were used to construct temperature profiles (Table 2-2 and

PAGE 74

62 2-3). They were used with the measured hold tube temperatures to create a complete profile for use in the PEIE method. These profiles along with the survivor data were used to estimate the thermal inactivation parameters for Escherichia coli in single strength orange juice. Table 2-5 shows the population survivor data for all the continuous thermal treatment experimental runs. The initial population was enumerated by plating out a sample of the untreated inoculated orange juice before and after each experimental run. Since the experimental runs were completed within 30 minutes after the orange juice was inoculated, the inactivation of the cells due to low-pH environment in the orange juice was not a significant source of error. The final values of the activation energy from each set of related kinetic parameters were determined after three iterations of the PEIE method (Table 2-6). The TDT curve for Escherichia coli in orange juice yielded a z-value of 6.16 Celsius degrees with an R 2 value from regression analysis of 0.99 (Figure 2-6). Three experiments were conducted yielding six experimental pairs, and 13 sets of parameters. These parameters were compared with the parameters estimated from the isothermal method with a 3-neck flask described in Chapter 1. Comparing PEIE and 3-Neck Flask Isothermal Methods The isothermal bath temperatures ranged from 52 o C to 60 o C while the continuous dynamic HTST hold tube temperatures ranged from 58 o C to 62 o C. The ranges overlapped between the two processes at 58 o C and 60 o C. The Dand z-values obtained from the traditional method using isothermal bath data and the PEIE method using continuous dynamic data was compared (Table 2-7). At the two overlapping temperatures there was a 16% difference at 58 o C and a 36% difference at 60 o C between

PAGE 75

63 the D-values estimated by the two methods. The PEIE method yielded essentially the same z-value as the 3-neck flask method. There is a slight difference in the slopes between the TDT curves but the most noticeable difference is the shift of each curve (Figure 2-7). This shift reflects the difference in reference D-values and will have an impact on the predicted number of survivors when used in the mathematical model (Equation 2-10). Vieira et al. (2002) observed this phenomena when comparing the kinetic parameters estimated by the PEIE method from continuous dynamic experiments with those estimated by a traditional method using vials submerged in an isothermal bath (Table 2-8 and Figure 2-9). The purpose of that study was to estimate the kinetic parameters for Alicyclobacillus acidoterrestris spores in Cupuacu nector. It is important to note that Vieira et al. estimated a reference D-value that was lower than with the PEIE method, while this study estimated a value that was higher. This difference in the comparisons between method between these studies can be explained by the methodology used to generate the isothermal bath data (submerge vials in a water bath vs. 3-neck flask). One of the significant problems with using vials submerged in an isothermal bath is the thermal lag experienced by the cell suspension. This lag is significant particularly when cooling the vials, where the temperature remains in the lethal range well after the sample has been extracted from the bath leading to additional inactivation beyond the measured time interval, thus over estimating the killing effect within that time interval. For the 3-neck flask method the primary source of error is the inability to extract the sample and quench the thermal inactivation process at the precise time interval planned and recorded.

PAGE 76

64 In this study the samples were extracted from the flask a few seconds prior to the prescribed time interval in order to account for anticipated transit time for injection into the chilled peptone water. Although this technique eliminated the possibility of any thermal inactivation occurring after the planned time interval, the anticipated transit time from the flask to the chilled water may lead to premature withdrawal reducing the lethal effect experienced by the cell suspension by 2 to 3 seconds shorter than the prescribed time interval, thus under-estimating the killing effect within that time interval. These sources of error for both isothermal bath methods can have a significant impact on the accuracy of estimating the thermal inactivation kinetic parameters when operating in the UHT/HTST temperature ranges where the D-values range from a few seconds to less than a second. Validation Experiments To verify which method yielded more accurate results, a series of validation experiments was performed whereby continuous pasteurizations were carried out using the lab scale pasteurizer with single strength orange juice. The purpose of these experiments was to compare the number of survivors predicted mathematically using the kinetic parameters from both the traditional and PEIE methods with the actual number of survivors obtained by plate count enumeration. The temperature histories along with the predicted and measured survivor responses from both sets of experiments are shown in Figures 2-10 and 2-11. The hold tube temperatures were chosen to be above the range in which the parameters were estimated to challenge the robustness of the model. The hold tube residence time was set at 10 seconds for experiment one and 15 seconds for two. Temperatures were recorded throughout the experimental run using a datalogger. The

PAGE 77

65 results show that the model predictions with PEIE parameters were closer to the actual number of survivors than those predicted with the 3-neck flask parameters (Table 2-9). These results were not surprising because of the shift in the TDT curve between the two methods and the implications of this shift as discussed previously. The significance of this finding is that using the 3-neck flask to generate isothermal bath data over-estimates the thermal inactivation rate constants, while using vials submerged in a constant temperature water bath to generate isothermal bath data underestimates the kinetic parameters. For processing thermally sensitive products, this difference can have a significant impact on the quality components of the product, such as flavor and vitamin retention. The microbiological characterization of systems and processes is important to validate lethality. Because of the short times for such high temperatures, using the traditional method with isothermal bath experiments often leads to imprecise kinetic parameter estimation. Cautious extrapolation is needed to relate parameters estimated under laboratory conditions to UHT/HTST process conditions in the manufacturing facility. This extrapolation may lead to further uncertainty. This extrapolation along with the tedious nature of isothermal bath experiments, have made characterizing continuous high temperature processes difficult. The PEIE method offers a valid alternative to isothermal bath experiments for estimating thermal inactivation kinetics of microbiological populations for the characterization of UHT/HTST systems with reasonable degree of confidence. For processing thermally sensitive products this difference can have a significant impact on the quality components of the product, such as flavor and vitamin retention.

PAGE 78

66 For example, if designing a process that will reduce the population of Escherichia coli in orange juice by 6 log cycles at a temperature of 66 o C, the required hold tube residence time would be 11 seconds based upon parameters from the traditional method and 8 seconds based upon those from the PEIE method. The advantage of the PEIE method would be a 10% retention of components such as vitamin C. Using Veira et al. (2002) data for Alicyclobacillus acidoterrestris, a six log cycle reduction in the population at 95 o C would result in a 10.08 minutes difference between the PEIE method and the isothermal method, a significant impact on shelf-life of the product. The PEIE method was developed to overcome some of the problems associated with isothermal bath experiments. The method is easier and faster for estimating kinetic parameters by saving laboratory time and equipment, and the kinetic parameters estimated using this method would provide better results than those from isothermal bath experiments. These parameters can be used in optimization techniques to determine the best balance in thermal processes between food safety and quality. The PEIE method can be applied to estimate kinetic parameters describing thermal inactivation of microorganisms or thermal degradation of quality factors in a more realistic way using real time processing equipment and conditions.

PAGE 79

67 Figure 2-1. Photo of the Microthermics HTST Lab 25 Labscale Pasteurizer

PAGE 80

68 Figure 2-2. Schematic Diagram of the flow of the Microthermics pasteurizer

PAGE 81

69 Table 2-1. Calibration of thermocouples Replicate Thermocouples # Mercury in glass reading ( o C) Thermocouple reading ( o C) Correction Factor ( o C) 1 1 71.80.02 69.51.0 2.31.0 2 72.00.02 70.21.0 1.81.0 3 72.50.02 71.01.0 1.51.0 2 1 750.02 73.41.1 1.61.1 2 750.02 72.71.1 2.31.1 3 750.02 74.11.1 1.51.1 3 1 75.60.02 741.1 1.61.1 2 75.60.02 74.51.1 1.11.1 3 75.60.02 74.11.1 1.51.1 4 1 75.70.02 74.21.1 1.51.1 2 75.60.02 74.21.1 1.41.1 3 75.60.02 74.31.1 1.51.1

PAGE 82

70 Table 2-2. Reynolds numbers for each flow rate for the continuous system Temperature ( o C) Residence Time (sec) Flow Rate (ml/min) Reynolds Number 58 60 480 619 90 320 413 60 30 960 1239 60 480 619 62 15 640 2491 30 960 1032

PAGE 83

71 010203040506070050100150Time(sec)Temperature (oC) 010203040506070050100150200Time (sec)Temperature (oC) Figure 2-3. Thermal profile of product at a hold tube nominal temperature of 58 o C and residence times of 60 and 90 seconds

PAGE 84

72 010203040506070020406080Time(sec)Temperature (oC) 010203040506070050100150Time(sec)Temperature (oC) Figure 2-4. Thermal profile of product at a hold tube nominal temperature of 60 o C and residence times of 30 and 60 seconds

PAGE 85

73 01020304050607001020304Time (sec)Temperature (oC) 0 01020304050607002040608Time (sec)Temperature (oC) 0 Figure 2-5. Thermal profile of product at a hold tube nominal temperature of 62 o C and residence times of 15 and 30 seconds

PAGE 86

74 Table 2-3 Rate constants used in Equation 2-1 for the heater section temperature profile. Temperature ( o C) Residence Time (sec) B h 58 60 58.6 -0.03273 90 58.6 -0.02101 60 30 59.91 -0.06545 60 59.91 -0.03152 62 15 61.34 -0.13558 30 61.34 -0.06779 Table 2-4. Rate constants used in Equation 2-2 for the chiller section temperature profile. Temperature ( o C) Residence Time (sec) B c 58 60 58.27 -0.09105 90 57.92 -0.64 60 30 60.28 -0.185 60 60.62 -0.093 62 15 61.23 -0.3742 30 61.94 -0.18285

PAGE 87

75 Table 2-5. Population survivor data for continuous experiments Hold tube Temperature ( o C) Replication Residence Time (sec) Initial Population (cfu) Number of Survivors (cfu) C/C o 58 1 60 5.6x10 8 5.2x10 6 9.3x10 -3 2 60 4.3x10 8 7.8x10 6 1.8x10 -2 1 90 5.6x10 8 3.0x10 4 5.4x10 -5 2 90 3.9x10 8 2.1x10 4 5.4x10 -5 60 1 30 6.1x10 8 3.1x10 6 5.1x10 -3 2 30 7.3x10 8 4.2x10 6 5.8x10 -3 3 30 3.9x10 8 7.0x10 5 1.8x10 -3 1 60 5.6x10 8 2.5x10 4 4.5x10 -5 2 60 7.0x10 8 7.9x10 4 1.1x10 -4 3 60 5.2x10 8 2.6x10 5 5.0x10 -4 4 60 5.1x10 8 1.4x10 5 2.8x10 -4 62 1 15 7.0x10 8 4.9x10 6 7.0x10 -3 2 15 7.0x10 8 8.0x10 6 1.1x10 -2 1 30 6.9x10 8 3.2x10 3 5.0x10 -6 2 30 1.7x10 8 3.3x10 3 1.9x10 -5

PAGE 88

76 Table 2-6. Estimation of Dand z-values from each iteration of the PEIE method Iteration 1 Iteration 2 Iteration 3 Iteration 4 Initial Ea Guess 20,000 J 62,089 J 267,398 J 342,711 J o C Residence Time (sec) D(sec) k(sec -1 ) D(sec) k(sec -1 ) D(sec) k(sec -1 ) D(sec) k(sec -1 ) 58 60 27.3 0.084 29.53 0.078 29.53 0.078 29.53 0.078 32.0 0.072 34.56 0.067 34.56 0.067 34.56 0.067 90 27.5 0.084 25.5 0.09 27.51 0.084 27.51 0.084 25.6 0.09 27.58 0.083 27.59 0.083 27.59 0.083 60 30 12.0 0.191 13.02 0.177 13.03 0.177 13.03 0.177 12.42 0.185 13.41 0.172 13.41 0.172 13.41 0.172 10.14 0.227 10.94 0.21 10.95 0.21 10.95 0.21 60 12.76 0.18 13.77 0.167 13.77 0.167 13.77 0.167 14.08 0.163 15.20 0.151 15.20 0.151 15.20 0.151 62 15 2.69 0.854 6.95 0.331 6.95 0.331 6.95 0.331 2.99 0.769 7.73 0.298 7.73 0.298 7.73 0.298 30 5.20 0.442 5.63 0.409 5.63 0.409 5.63 0.409 5.89 0.391 6.37 0.362 6.4 0.362 6.4 0.362 Estimated Ea 62,089 J 267,398 J 342,711 J 342,711 J

PAGE 89

77 00.20.40.60.811.21.41.657585960616263Temperature (oC)Log D-value(sec) Iteration 1 Iteration 2 Iteration 3 Figure 2-6. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic parameters from the PEIE method

PAGE 90

78 Table 2-7. Comparison of Dand z-values estimated by traditional method using isothermal treatments and PEIE method using continuous dynamic treatments Isothermal (3-neck flask) Dynamic (PEIE) Temperature ( o C) Average D-value (sec) Standard Deviation D-value (sec) Standard Deviation 52 353 39.08 55 148 2.18 58 34.7 2.27 29.8 3.3 60 18 1.52 13.27 1.54 62 6.93 0.47 z-value ( o C) 5.99 6.16 Table 2-8. Kinetic parameters of thermal inactivation of Alicyclobacillus acidoterrestris spores in Cupuacu nectar using the PEIE method and Isothermal method PEIE Method Isothermal (submerged vials) D 95 o C (min) 5.5 1.2 3.82 0.48 z ( o C) 31 29 10 R 2 0.87 0.98 No. of observations 22 26 *Source: Veira et al. (2002)

PAGE 91

79 00.20.40.60.811.21.41.657585960616263Temperature (oC)Log D-value(sec) Iteration 1 Iteration 2 Iteration 3 Figure 2-7. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic parameters from the PEIE method

PAGE 92

80 00.511.522.535052545658606264Temperature (oC)Log [D-value(sec)] Traditional PEIE Figure 2-8. Comparison of TDT curves based upon data from the traditional and PEIE methods

PAGE 93

81 00.20.40.60.811.21.480859095100105110115Temperature (oC)Log [D-value(min)] PEIE Isothermal Figure 2-9. Comparison of TDT curves based upon data from the traditional and PEIE methods for Alicyclobacillus acidoterrestris spores in Cupuacu nectar (Vieira et. al. 2002) (Estimated curve based upon reference D-value and z-value)

PAGE 94

82 0123456789100102030405060Time (seconds)Log [survivors(cfu/ml)]010203040506070 Isothermal PEIE Experimental Temperature Figure 2-10. Temperature history and measured and predicted survivor responses for validation experiment I (10 second hold tube)

PAGE 95

83 0123456789100510152025303540Time (seconds)Log [survivors (cfu/ml)]010203040506070 Isothermal PEIE Experimental Temperature Figure 2-11. Temperature history and measured and predicted survivor responses for validation experiment II (15 second hold tube)

PAGE 96

84 Table 2-9. Results of validation experiments, comparison of predicted number of survivors for PEIE analysis and Traditional isothermal batch analysis with experimental number of survivors Hold Tube Survivors (cfu) Experiment Time (sec) Temp ( o C) Initial (cfu) PEIE Predicted Isothermal Predicted Experimental I 15 65 5.4x10 8 3.51x10 3 4.95x10 4 5.15x10 3 8.0x10 3 II 10 65 5.4x10 8 1.98x10 3 8.59x10 4 1.0x10 3 1.25x10 3

PAGE 97

CHAPTER 3 ESTIMATION OF KINETIC PARAMETERS FOR THERMAL INACTIVATION OF Alicyclobacillus acidoterrestris IN ORANGE JUICE Introduction The recent discovery of Alicyclobacillus acidoterrestris in high-acid pasteurized fruit juices and its ability to cause spoilage in these products have become concerns for processors in the design of thermal pasteurization processes. Alicyclobacillus acidoterrestris are sporeforming thermophilic bacteria that grow well in low pH environments. These characteristics of the bacteria can be problematic since all shelf stable and refrigerated fruit juices are pasteurized at temperatures below the lethal range of Alicyclobacillus. This inadequate processing can lead to premature spoilage of the product with risk of recall from the marketplace. Accurate estimation of the thermal inactivation kinetic parameters that are used in a model to predict the number of survivors is essential to establish optimum process conditions to assure a low probability of spoilage of the product without over processing the product, which leads to degradation of juice quality important to consumers. It has been shown in chapters 1 and 2 that the Paired Equivalent Isothermal Exposures (PEIE) method is valid and accurate for the estimation of kinetic parameters. The PEIE method is not just limited to using Arrhenius kinetics for estimation of thermal inactivation kinetic parameters (k and Ea). It may also be possible to use the PEIE method with Thermal Death Time (TDT) kinetics (Dand z-value) which are more commonly used by food scientists. Orange juice was the product 85

PAGE 98

86 chosen for this study because of the spoilage problems that have been documented involving Alicyclobacillus acidoterrestris in orange juice. Using the PEIE method to analyze continuous dynamic thermal treatment data, thermal inactivation kinetic parameters were estimated for Alicyclobacillus acidoterrestris in orange juice. Literature Review Occurrences of Alicyclobacillus acidoterrestris in Juice Products Fruit juices with a pH below 4.0 have been considered susceptible to spoilage only by microorganisms of low heat resistance such as molds and acid-tolerant non-sporeforming bacteria (Eiroa et al. 1999). Because of the low resistance to heat by these microorganisms, pasteurization processes designed with temperatures ranging from 85 o C to 95 o C were thought to be sufficient to inactivate these spoilage-causing microorganisms (Eiroa et. al. 1999). The first reported incidence of a food product being spoiled by acidophilic sporeformers was in Germany with apple juice (Walls and Chuyate, 1998). It was determined that this spoilage microorganism was Bacillus acidoterrestris, which was later named Alicyclobacillus acidoterrestris. Spoilage by this microorganism leads to off flavors in the products similar to the taste of phenolic substances, odors of a disinfectant and pronounced cloudiness (Eiroa et. al. 1999; Walls and Chuyate 1998). Fortunately, Alicyclobacillus acidoterrestris does not appear to be pathogenic according to Walls and Chuyate (2000), who conducted pathogenicity studies with the bacteria in mice and guinea pigs. Although Alicyclobacillus acidoterristris is not a safety concern for industry, it is a serous economic issue. During the 1990s Alicyclobacillus acidoterrestris was presenting itself as a spoilage problem in shelf stable juice products (Walls and Chuyate 1998). In 1994 there was a report of off odors in apple juice caused by gram positive rods isolated from the juice and showing characteristics similar to

PAGE 99

87 Alicyclobacillus acidoterrestris (Eiroa et. al. 1999). Spoilage of juices by these bacteria is not readily detected since there is often little sedimentation and no gas produced that would distort the product package. In juice inoculation studies, it was discovered that Alicyclobacillus acidoterrestris grew well in orange, apple, tomato, and grape juice in which the pH of the juices ranged from 3.47 to 4.27 (Walls and Chuyate 2000). Alicyclobacillus acidoterrestris is a new spoilage microorganism that must addressed by the juice industry and other processors of low pH food products. Current temperatures used for pasteurization are insufficient to inactivate spores of these bacteria in fruit juices, yet thermally overprocessing the product can lead to unacceptable quality degradation of the product. Because sporeforming bacteria of importance in foods are rarely as acid tolerant as Alicyclobacillus acidoterrestris, it is important to characterize the thermal inactivation behavior in populations of this microorganism in order to design processes that will reduce the probability of spoilage for shelf stable products while maintaining quality factors acceptable to the consumer. The PEIE method has been reported in recent literature to be useful in obtaining greater accuracy in parameter estimation (Welt et al. 1997a, b). Using the PEIE method to obtain thermal inactivation kinetic parameters for Alicyclobacillus acidoterrestris will insure a more accurate estimation of those parameters. In this work the PEIE method will be carried out in both Arrhenius kinetics and Thermal Death Time (TDT) kinetics using process lethality (F-value), which is more familiar to food scientists. The PEIE Method in Arrhenius Kinetics Recall that a first-order rate process is described by kCdtdC (3-1)

PAGE 100

88 where C is the concentration of a reactant at a time, t, and k is the rate constant of the reaction. Solving Equation 3-1 by integration yields Equation 3-2. )(lnoottkCC (32) The temperature dependency of the rate constant, k, is described by Equation 3-3, the Arrhenius equation, RaRTTREkk11exp (33) where k R is the rate constant at reference temperature T R E a is the activation energy, T is the desired operating temperature, and R is the ideal gas law constant. Under isothermal conditions, the Arrhenius equation that describes the behavior of a reactant that follows a first-order reaction process is shown in Equation 3-4. )(explnoRottTREkCC (34) This equation can be used to determine the extent of reaction for a given constituent at a constant temperature. Under non-isothermal conditions, Equation 3-4 is integrated as shown in Equation 3-5 using the temperature history, T(t), to give the extent of reaction. Combining Equations 3-4 and 3-5 to equate a dynamic process to an dttTREkCCttRoo)(expln (35) isothermal and normalizing to eliminate k R introduces a new factor called G, the decimal reduction factor, and the equation becomes

PAGE 101

89 tteaeaRooTREtdttTREGkCCexp)(expln (36) where t e is the equivalent time needed to obtain the same reduction in the constituent at a constant temperature, T e, for a given dynamic process. The PEIE method uses Equation 3-6 to represent an isothermal process as eaeTREtGexp (37) and taking two sets of E a and G values yields two equations that can be used to solve for t e and T e (using the elimination method) as follows: eaeeaeTREtGTREtG21)ln()ln()ln()ln(21 (38) where 21ln)(12GGREETaae (39) and eaeTREGtexp (310) These isothermal combinations, when plotted on a semilog plot of time versus temperature would appear as a straight line, and at any point on this line the combinations of T e and t e are equivalent isothermal exposures (EIE) yielding the same extent of reaction as the given non isothermal process. As shown in Chapter 2, using these EIEs along with the measured extents of reactions allows the estimation of the kinetic

PAGE 102

90 parameters k and E a These parameters can then be used in the model (Equation 2.10) to predict the number of survivors from a process at a given time and temperature history. The PEIE Method and TDT Kinetics The conceptual approach to the PEIE method is not new to the food industry. A similar means to equate a dynamic process to an isothermal process for microorganisms that has been used in the food industry for many years is the Process Lethality Value (F). Recall, Equations 2-7 and 2-8 from Chapter 2 describe first order thermal inactivation kinetics using thermal death-time (TDT) parameters Dand z-value. Equations 3-11 and 3-12 are restatements of these equations where D is the decimal reduction time (the time needed for a one-log cycle reduction in the population at a given constant temperature. Thus, the DtoCC10 (311) D-value is related to the first order rate constant (k), since both show the relationship between time and population reduction at a given temperature. The relationship between D-value and temperature (the D-values temperature dependency) is assumed to be log-linear and given by Equation 3-12 where D o is reference D-value at reference temperature T o and z is the temperature interval required to change the value of D by one log cycle. zTTooDTD10)( (312) Equation 3-13 determines the F value for a process under isothermal conditions. tFZRTT10 (313) When the product temperature varies the F value delivered by the process must be integrated mathematically as shown in Equation 3-14. This equation determines the F

PAGE 103

91 value delivered by a process that has experienced a dynamic temperature history, in which temperature varies as a function of time, T(t). In TDT kinetics F value is equivalent to the G value in Arrhenius kinetics. dtFttZTtToR)(10 (314) Like the PEIE method, isothermal combinations of temperature and time that would achieve the same F value would appear as a straight line on a semilog plot of time versus temperature known as the thermal death time (TDT) curve. Any point on this line of temperature-time is an EIE. Combining Equations 3-13 and 3-14 to equate a dynamic process to an isothermal the equation becomes eZTTttZTtTtdtFReoR1010)( (315) For a dynamic process where the temperature history has been recorded and a z-value chosen, an accumulated F value can be determined using Equation 3-14. An isothermal process derived from two z-values may be interpreted as an equivalent exposure that would result in the same lethalities for reactant characterized by the respective z-values. These EIEs can be used to solve for t e and T e (using substitution) as follows: 2211)log()log()log()log(ZTTtFZTTtFReeRee (316) where

PAGE 104

92 ReTZZFFT212111log (317) and ZTTeReFt10 (318) The extent of reaction data along with the equivalent times can be used to estimate the D-values at the respective equivalent temperature. Objectives The purpose of this study was to apply the PEIE method to estimate thermal inactivation kinetic parameters of Alicyclobacillus acidoterrestris in orange juice. To achieve this goal the objectives of this project were the following: Estimate the kinetic parameters for thermal inactivation of Alicyclobacillus acidoterrestris in orange juice by the PEIE method in Arrhenius kinetics from data generated by continuous dynamic thermal treatments Estimate the kinetic parameters for thermal inactivation of Alicyclobacillus acidoterrestris in orange juice by the PEIE method in TDT kinetics from the same data. Compare results from both methods. Methods and Materials Preparation of Cultures The strain of Alicyclobacillus acidoterrestris used in this study was obtained from the American Type Culture Collection (ATCC #49025). Working and stock cultures of this strain were made from frozen stock cultures of the original freeze-dried culture obtained from ATCC. The cultures were thawed and streaked onto acidified K agar and

PAGE 105

93 incubated at 45 o C for 3 to 5 days, depending on the date of each experimental run. The K agar was acidified using 2.5 mg of malic acid per 700 ml of media to give a final pH of 3.7 0.1. The cultures, containing vegetative cells and spores were harvested from the agar plates using a sterile pipette to submerge the culture in sterile buffer solution. A sterile spreader was used to gently separate the cells and spores from the agar and suspend them in buffer solution. The solution containing the cell/spore suspension was aseptically poured off the agar plate and into a sterile container of buffer solution. This procedure was repeated with twelve plates and the resulting concentration of cell/spore suspension was used as the inoculum for the orange juice. The inoculum was heat shocked at 75 o C for 10 minutes prior to inoculation into the orange juice to inactivate vegetative cells in order to have a more uniform population of heat resistant spores. Experimental Apparatus The experimental apparatus used in these experiments was the Microthermics UHT/HTST Lab-25 lab-scale pasteurizer unit. A photo and schematic diagram of the unit are shown in Figures 2-1 and 2-2. Refer to Chapter 2 for a detailed description of the pasteurizer unit. Continuous Dynamic Thermal Treatments The orange juice was reconstituted from concentrate using sterile filtered deionized water. The orange juice concentrate was a commercial brand orange juice concentrate at 44 o Brix. The reconstitution was performed following the manufacturers directions on the label (1 part concentrate to 3 parts water). Although the orange juice was not reconstituted under aseptic conditions, the resident population of Alicyclobacillus acidoterrestis in the product was negligible when compared with the number of spores/cells in the inoculum. The product was inoculated with an Alicyclobacillus

PAGE 106

94 acidoterrestis cell suspension prior to thermal exposure to achieve a minimum initial concentration of 1 x 10 6 cfu/mL. Two liters of orange juice were prepared along with 200ml of spore/cell suspensions. The pasteurizer was sanitized by circulating hot water at 83 o C through the heater, hold tube, chiller sections and accessory tubes for a minimum of 30 minutes. After being sanitized, the temperature of the pasteurizer was adjusted to the desired experimental temperature and allowed to reach steady-state conditions. Then the product flow control valve was opened to allow the inoculated product to flow through the unit. Temperatures at various locations throughout the system were recorded using a datalogger attached to a notebook computer. The thermal profile (temperature vs. time) of each experimental run was captured from each port and saved as a text file that was used in the PEIE method. To produce replicate data for each temperature, 2 samples were collected for each experimental run and a minimum of two experimental runs were conducted for each residence temperature-time combination. An experiment involved a product cycle whereby a batch of product was pumped through the system after using water to achieve a stable steady state condition, then the product and water reservoir valves were switched to allow water to run through the system at the same conditions while another sample of product was being prepared. Then the valves were switched and the product pumped through the system and samples taken again. The residence time for each run was 90 seconds, with the exception at 104 o C for Day 2, which was set for residence time of 60 seconds. Temperature Profiles The temperature was measured at the inlet of the heating section (initial product temperature), after the heating section (at the entrance to the hold tube), after the hold

PAGE 107

95 tube (at the entrance to the chiller section), and after the chiller section. Using the recorded temperature at each point, the heater and chiller portions of the profile were constructed from heat transfer equations, while the hold tube portions were constructed based upon measured data. Recall from Chapter 2 that the standard profile for a shell and tube heat exchanger follows an exponential increase that can be described by Equation 3-1, )-1(-hteBAT (3-1) where T is the temperature at any point within the heat exchanger at a specific time t, A is the initial temperature of the product, B is the temperature of the product upon exit from the heat exchanger, and h is the rate constant for the temperature change through the heat exchanger. This equation yielded the calculated temperatures along the heater section of the pasteurizer. The hold tube inlet and outlet temperatures were measured directly by thermocouples. The temperature along the chiller section of the pasteurizer was calculated using Equation 3-2. )(-cteBT (3-2) where B is the temperature of the product upon entrance into the chiller section, T is the temperature at any point within the chiller section at a specific time t and c is the rate constant for the temperature change through the chiller section. Knowing the residence time of the product within the heater and chiller section of the pasteurizer, the temperature profile was constructed by determining the parameters of Equations 3-1 and 3-2 using the boundary conditions of each section. The residence times for each section were determined based upon the flow rate of the product and the diameters and lengths of the tubes in all sections of the pasteurizer with the assumption of plug flow for simplicity.

PAGE 108

96 The flow rates of the product were determined by measuring the amount collected in a graduated cylinder over a specific time period. Results and Discussion Parameter Estimation by PEIE Figures 3-1 and 3-2 show the temperature histories for the continuous dynamic thermal experiments with hold tube temperatures at 95 o C, 100 o C and 104 o C. Tables 3-1 and 3-2 show the temperature profile rate constants determined for the heater and chiller sections, respectively. Once the temperature profiles were constructed, they were used with the measured hold tube temperatures to create a complete profile for use in the PEIE method. These profiles along with the survivor data were used to estimate the thermal inactivation parameters for Alicyclobacillus acidoterretris in single strength orange juice. Table 3-3 shows the population survivor data for all the continuous thermal treatment experimental runs. The initial population was verified by plating out (in triplicate) a sample of the untreated inoculated orange juice before and after each experimental run. Table 3-4 shows the four iterations of the PEIE method needed to converge on a final solution for the EIE's using Arrhenius kinetics. From Table 3-4, k-values of 0.021 seconds -1 0.070 seconds -1 and 0.119 seconds -1 were estimated at equivalent temperatures of 94.4 o C, 100.7 o C, and 104.2 o C respectively. Figure 3-3 shows the Arrhenius curve for Alicyclobacillus acidoterrestris, with an Ea-value of 204 kJ/mol. Parameter Estimation using F value and TDT kinetics Table 3-5 shows the four iterations of the PEIE method needed to converge on a final solution for EIEs using TDT kinetics. From Table 3-5, D-values of 107.4 seconds (1.79 minutes), 34.22 seconds (0.57 minutes), and 17.65 seconds (0.30 minutes) were estimated at equivalent temperatures of 95 o C, 100 o C, and 104 o C respectively. The z

PAGE 109

97 value for Alicyclobacillus acidoterrestis in orange juice from this study was determined to be 13.1 o C, as shown in Figure 3-4. When compared to work performed by other researchers using Alicyclobacillus in orange juice, these values fall within published ranges as shown in Table 3-6. Eiroa et. al. (1999) reported D 95 -values for four strains of Alicyclobacillus acidoterrestris spores in orange juice that ranged from 2.5 min to 8.7 min and z-values that ranged from 7.2 o C to 11.3 o C. Splittstoesser et. al. (1998) reported a D 95 -value for Alicyclobacillus acidoterrestris in concord grape juice at 16 o Brix and a pH of 3.5 of 1.9 minutes and a z-value of 6.9 o C. Komitopoulou et. al. (1999) reported a D 95 -value of 3.9 minutes in single strength orange juice with a pH of 3.9 and a z-value of 12.9 o C and McIntyre et. al. (1995) reported a D 95 -value of 1.0 minute. Komitopoulou et al. (1995) used a wide-necked flask apparatus (equivalent to a three-necked flask apparatus) to generate heat inactivation data whereas Spittstoesser et. al. (1998) and Eiroa et. al. (1999) used glass vials submersed in an isothermal water bath.. Figure 3-4 shows a comparison between the TDT curves generated by the kinetic parameters from Arrhenius kinetics versus those generated by TDT kinetic. There is a slight difference in the offset of the curves and in the slopes of the curves. The effect of this difference would be amplified as temperatures move further away from the range used in this experiment. Jones (1968) and Jonsson et al. (1977) showed that sterilization times based on parameters obtained by both methods could show important discrepancies at higher temperatures. Using the TDT and Arrhenius curves the D-value at both extremes of the temperature range was estimated. From the TDT curve, D-values of 136 seconds and 20 seconds for temperatures of 93 o C and 104 o C respectively were estimated. From the Arrhenius, curve D-values of 156 seconds and 18 seconds at the same

PAGE 110

98 temperatures were estimated. There was a difference of 12.8% in D-values at the lower temperature and 10% at the higher temperature. Using each curve to estimate the F value for a 6 log cycle reduction in the population of Alicyclobacillus in orange juice at a temperature of 90 o C, the TDT method would yield a value of 23 minutes whereas the Arrhenius method would yield a value of 29 minutes. Ocie et al. (1994) conducted a study to compare the TDT and Arrhenius methods for rate constant predictions of Bacillus stearothermophilus and concluded that using the TDT method to generate kinetic parameters could introduce unsafe process times, whereas Jones concluded that the appropriate method to generate parameters depends on which gives the most conservative values for the kinetic parameters. In this study both converged on the same time-temperature combinations in the given temperature range (95 o C 104 o C) that would give the same estimated values for the kinetic parameters. The significance of this finding is that the PEIE method can be successfully executed in both Arrhenius and TDT kinetics. Since TDT kinetics have been used extensively in the food industry for establishing thermal processes and determining the lethality of a process, the reasonable agreement between the kinetic parameters estimated by the PEIE method in this study and those estimated by traditional methods from published work shows that the method is a valid and trustworthy method for estimating thermal inactivation kinetic parameters using data generated from a continuous dynamic thermal treatment process. However, the PEIE method as presented here can only be used to estimate the parameters of first order kinetics with a uniform population of organisms.

PAGE 111

99 0.0020.0040.0060.0080.00100.00120.00050100150200250Time (sec)Temperature (oC) 95 C 100 C 104 C Figure 3-1. Thermal profile of product at a hold tube nominal temperature of 95 o C, 100 o C, and 104 o C for experimental set one

PAGE 112

100 0.0020.0040.0060.0080.00100.00120.00050100150200250Time (sec)Temperature (oC) 95 C 100 C 104 C Figure 3-2. Thermal profile of product at a hold tube nominal temperature of 95 o C, 100 o C, and 104 o C for experimental set two.

PAGE 113

101 Table 3-1. Rate constants used in Equation 2-1 and 2-2 for the heater and chiller sections temperature profile for experimental set 1 Heater profile parameters Chiller profile parameters Temperature B h B c 90 88.9 -0.028 87.8 -0.056 95 94.3 -0.033 94.2 -0.063 100 101.1 -0.046 100.9 -0.069 104 104.2 -0.051 104.1 -0.069 Table 3-2. Rate constants used in Equation 2-1 and 2-2 for the heater chiller sections temperature profile for experimental set 2 Heater profile parameters Chiller profile parameters Temperature B h B c 95 94.9 -0.0325 95.06 -0.065 100 100.4 -0.039 100.83 -0.065 104 104.2 -0.079 104.3 -0.12

PAGE 114

102 Table 3-3. Population survivor data from Ultra High Temperature (UHT) heat treatments with Alicyclobacillus acidoterrestris in orange juice Hold tube Temperature ( o C) Replication Residence Time (sec) Initial Population (cfu) Number of Survivors (cfu) C/C o 95 1 90 3.46x10 6 2.42x10 5 6.9x10 -2 2 90 1.08x10 6 1.06x10 5 9.8x10 -2 3 86 1.98x10 6 8.18x10 4 7.1x10 -2 4 86 7.37x10 5 1.4x10 5 1.9x10 -1 100 1 90 3.46x10 6 1.35x10 3 3.0x10 -42 90 1.08x10 6 5.38x10 2 5.0x10 -4 3 86 1.98x10 6 1.65x10 3 8.0x10 -4 4 86 7.37x10 5 8.4x10 2 1.1x10 -3 104 1 90 3.46x10 6 3.0x10 1 8.7x10 -6 2 90 1.08x10 6 2.7x10 1 2.5x10 -5 3 62 1.98x10 6 4.70x10 2 2.3x10 -4 4 60 7.37x10 5 3.35x10 2 4.5x10 -4

PAGE 115

103 Table 3-4. Estimation of k and Ea values from each iteration of the PEIE method using Arrhenius kinetics Iteration 1 Iteration 2 Iteration 3 Iteration 4 Initial Ea Guess 20,000 J 214,000 J 188,000 J 203,000 J o C Residence Time (sec) k(sec -1 ) k(sec -1 ) k(sec -1 ) k(sec -1 ) 95 102 0.02077 0.02191 0.0223 0.02232 0.01847 0.01948 0.01983 0.01985 99 0.0124 0.01912 0.01886 0.01901 0.01401 0.0216 0.02131 0.02149 100 102 0.04742 0.07426 0.07356 0.07399 0.04373 0.0685 0.06784 0.06824 99 0.04526 0.07131 0.0705 0.07099 0.04298 0.06772 0.06695 0.06742 104 102 0.07044 0.1097 0.1089 0.1094 0.06809 0.106 0.1053 0.1058 57 0.08645 0.1327 0.1324 0.1326 0.08578 0.1317 0.1314 0.1316 Estimated Ea 214,000 J 188,000 J 203,000 J 204,000 J

PAGE 116

104 -4.5-4-3.5-3-2.5-2-1.50.00260.002650.00270.00275Inverse Absolute Temperature (K-1)ln[k(sec-1)] Iteration 1 Iteration 2 Iteration 3 Iteration 4 Figure 3-3. Arrhenius curve for Alicyclobacillus acidoterrestris in orange juice using kinetic parameters from the PEIE method

PAGE 117

105 Table 3-5. Estimation of Dand z-values from each iteration of the PEIE method using TDT kinetics Iteration 1 Iteration 2 Iteration 3 Iteration 4 Initial z-value Guess ( o C) 20 14.79 13.18 13.11 o C Residence Time (sec) D(sec) D(sec) D(sec) D(sec) 95 102 110.8 105.1 103.2 103.2 124.7 118.2 116.1 116 99 185.7 120.4 122.1 121.1 164.4 106.6 108 107.2 100 102 48.56 31.01 31.3 31.12 52.65 33.62 33.94 33.74 99 50.87 32.29 32.66 32.43 53.57 34 34.39 34.15 104 102 32.69 20.99 21.14 21.04 33.82 21.71 21.86 21.77 57 26.63 17.35 17.39 17.36 26.84 17.48 17.52 17.5 Estimated z-value ( o C) 14.79 13.18 13.11 13.11

PAGE 118

106 11.21.41.61.822.22.4949698100102104106Temperature (oC)Log[Dvalue(sec)] Iteration 1 Iteration 2 Iteration 3 Iteration 4 Figure 3-4. TDT curve for Alicyclobacillus acidoterrestris in orange juice using kinetic parameters from the PEIE method

PAGE 119

107 Table 3-6. Comparison of TDT kinetic parameters with published data from various sources using Alicyclobacillus acidoterrestris Study (strain) D 95 (minutes) z-value ( o C) Eiroa et. al. (46) 2.5 7.2 Eiroa et. al (70) 8.7 11.3 Eiroa et. al. (145) 3.8 7.2 Eiroa et. al. (DSM2498) 2.7 7.9 Splittstoesser et. al. 1.9 6.9 Komitopoulou et. al. 3.9 12.9 McIntyre et. al. 1.0 This study 1.8 13.1

PAGE 120

108 11.21.41.61.822.292949698100102104106Temperature (oC)Log[Dvalue(sec)] Arrhenius Kinetics TDT Kinetics Figure 3-5. Comparison of TDT curves based upon data from the PEIE method using TDT kinetics and Arrhenius kinetics

PAGE 121

APPENDIX A MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli USING ARRHENIUS KINETICS Step 1: Retrieal of the dynamic temperature history and survival dataConstants: Universal Gas constant:Rval8.314 J/mol K Reading in temperature and concentration files:Conc58READPRN"c:\phd\58Ca.txt"() Conc58606090905.61084.251085.61083.851085.21067.810631052.1105 Conc60READPRN"c:\phd\60Ca.txt"() Conc60303030606060606.131087.251083.851085.610871085.21085.11083.051064.210671052.461047.91042.641051.39105 Conc62READPRN"c:\phd\62Ca.txt"() Conc62151530307.0510871086.91081.71084.910681063.21033.3103 109

PAGE 122

110 Conc62READPRN"c:\phd\62Ca.txt"() Conc62151530307.0510871086.91081.71084.910681063.21033.3103 Step 2: Selection of two arbitrary Ea values and determination of the corresponding G valuesEa1200000 Ea21.5Ea1 Ea2300000Temperature = 58CTemp58READPRN"c:\PhD\58C Profile.txt"() T5860v()Temp58v1273 T5890v()Temp58v3273 G58_6010116vexpEa1 RvalT5860v()() = G58_6011.836014080274911030 G58_6020116vexpEa2 RvalT5860v()() = G58_6022.959260986642731046 G58_9010174vexpEa1 RvalT5890v()() = G58_9012.809547169462081030 G58_9020174vexpEa2 RvalT5890v()() = G58_9024.55341046 Temperature = 60CTemp60READPRN"c:\PhD\60C Profile.txt"()

PAGE 123

111 T6030v()Temp60v1273 T6060v()Temp60v3273 G60_301060vexpEa1 RvalT6030v()() = G60_3011.507401951543201030 G60_302060vexpEa2 RvalT6030v()() = G60_3023.05261046 G60_6010116vexpEa1 RvalT6060v()() = G60_6013.059830397015291030 G60_6020116vexpEa2 RvalT6060v()() = G60_6026.42321046 Temperature = 62CTemp62READPRN"c:\PhD\62C Profile.txt"() T6215v()Temp62v1273 T6230v()Temp62v3273 G62_151029vexpEa1 RvalT6215v()() = G62_1511.012993149334321030

PAGE 124

112 G62_152029vexpEa2 RvalT6215v()() = G62_1522.43531046 G62_301060vexpEa1 RvalT6230v()() = G62_3012.050786750955811030 G62_302060vexpEa2 RvalT6230v()() = G62_3024.91381046 Step 3: Equate G values to the isothermal form to determine the respective equivalent isothermal exposure for each pair of dynamic thermal experiments.58C; 60 and 90 second residence time TeEa2Ea1 ()RvallnG58_601G58_602 te58_60G58_601expEa1 RvalTe Temp58_60Te Temp58_60330.764te58_6070.674TeEa2Ea1 ()RvallnG58_901G58_902

PAGE 125

113 te58_90G58_901expEa1 RvalTe Temp58_90Te Temp58_90330.814te58_90106.96660C; 30 and 60 second residence time TeEa2Ea1 ()RvallnG60_301G60_302 te60_30G60_301expEa1 RvalTe Temp60_30Te Temp60_30332.853te60_3036.757TeEa2Ea1 ()RvallnG60_601G60_602 te60_60G60_601expEa1 RvalTe Temp60_60Te Temp60_60333.185te60_6069.43662C; 15 and 30 second residence time TeEa2Ea1 ()RvallnG62_151G62_152

PAGE 126

114 te62_15G62_151expEa1 RvalTe Temp62_15Te Temp62_15334.441te62_1517.527TeEa2Ea1 ()RvallnG62_301G62_302 te62_30G62_301expEa1 RvalTe Temp62_30Te Temp60_30332.853te62_3035.721Temp58aTemp58_60Temp58_90 Temp60aTemp60_30Temp60_60 Temp62aTemp62_15Temp62_30 Temp58meanTemp58a() Temp60meanTemp60a() Temp62meanTemp62a() Temp58330.789Temp60333.019Temp62334.426Temp58stdstdevTemp58a() Temp60stdstdevTemp60a() Temp62stdstdevTemp62a() Temp58std0.025Temp60std0.166Temp62std0.016Step 4: Calculate the k values at equivalent temperatures from the equivalent time determined in step 3

PAGE 127

115 Conc58606090905.61084.251085.61083.851085.21067.810631052.1105 Conc60303030606060606.131087.251083.851085.610871085.21085.11083.051064.210671052.461047.91042.641051.39105 Conc62151530307.0510871086.91081.71084.910681063.21033.3103 k58601lnConc5801Conc5802 te58_60 k60301lnConc6001Conc6002 te60_30 k62151lnConc6201Conc6202 te62_15 k62152lnConc6211Conc6212 te62_15 k58602lnConc5811Conc5812 te58_60 k60302lnConc6011Conc6012 te60_30 k586010.066k621510.284k586020.057k621520.255k60303lnConc6021Conc6022 te60_30 k603010.144k603020.14k603030.172k58901lnConc5821Conc5822 te58_90 k60601lnConc6031Conc6032 te60_60 k62301lnConc6221Conc6222 te62_30

PAGE 128

116 k62302lnConc6231Conc6232 te62_30 k58902lnConc5831Conc5832 te58_90 k60602lnConc6041Conc6042 te60_60 k589010.07k623010.344k589020.07k623020.304k60603lnConc6051Conc6052 te60_60 k606010.144k606020.131k606030.109k58ak58601k58602k58901k58902 k60ak60301k60302k60303k60601k60602k60603 k62ak62151k62152k62301k62302 k58meank58a() k60meank60a() k62meank62a() k58stdstdevk58a() k60stdstdevk60a() k62stdstdevk62a() k580.066k600.14k620.297k58std0.006k60std0.019k62std0.032D582.303k58 D602.303k60 D622.303k62 D5834.968D6016.436D627.766Step 5: Estimation of the activation energyEa00Rvallnk58k60 Temp58Temp60 Temp58Temp60 Ea00310049.417

PAGE 129

117 Ea01Rvallnk58k62 Temp58Temp62 Temp58Temp62 Ea01380539.569Ea02Rvallnk60k62 Temp60Temp62 Temp60Temp62 Ea02493535.73Step 6: Used the newly estimated activation energy value as the initial guess i step 2stdeviationAstdevEa() meannAmeanEa() stdeviationA75574.996meannA394708.238

PAGE 130

APPENDIX B MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli USING THERMAL DEATH TIME (TDT) KINETICS Step 1: Retrieal of the dynamic temperature history and survival dataConstants: Universal Gas constant:Rval8.314 J/mol K Reading in temperature and concentration files:Conc58READPRN"c:\phd\58Ca.txt"() Conc58606090905.61084.251085.61083.851085.21067.810631052.1105 Conc60READPRN"c:\phd\60Ca.txt"() Conc60303030606060606.131087.251083.851085.610871085.21085.11083.051064.210671052.461047.91042.641051.39105 Conc62READPRN"c:\phd\62Ca.txt"() Conc62151530307.0510871086.91081.71084.910681063.21033.3103 118

PAGE 131

119 Step 2: Selection of two arbitrary z-values and determination of the correspondi n F valueszvalue15.468 zvalue2zvalue11.5 zvalue28.202Temperature = 58CTemp58READPRN"c:\PhD\58C Profile.txt"() T5890v()Temp58v3273 T5860v()Temp58v1273 F58_6010116t10T5860t()273 ()121 zvalue1 = F58_6011.946521010 F58_6020116t10T5860t()273 ()121 zvalue2 = F58_6021.37706106 F58_9010174t10T5890t()273 ()121 zvalue1 = F58_9013.008281010 F58_9020174t10T5890t()273 ()121 zvalue2 = F58_9022.11369106 Temperature = 60CTemp60READPRN"c:\PhD\60C Profile.txt"()

PAGE 132

120 T6060v()Temp60v3273 T6030v()Temp60v1273 F60_301060t10T6030t()273 ()121 zvalue1 = F60_3012.439021010 F60_302060t10T6030t()273 ()121 zvalue2 = F60_3021.28777106 F60_6010116t10T6060t()273 ()121 zvalue1 = F60_6015.303491010 F60_6020116t10T6060t()273 ()121 zvalue2 = F60_6022.66973106 Temperature = 62CTemp62READPRN"c:\PhD\62C Profile.txt"() T6230v()Temp62v3273 T6215v()Temp62v1273 F62_151029t10T6215t()273 ()121 zvalue1 = F62_1512.269491010 F62_152029t10T6215t()273 ()121 zvalue2 =

PAGE 133

121 F62_1529.59042107 F62_301060t10T6230t()273 ()121 zvalue1 = F62_3014.56951010 F62_302060t10T6230t()273 ()121 zvalue2 = F62_3021.93744106 Step 3: Equate G values to the isothermal form to determine the respective equivalent isothermal exposure for each pair of dynamic thermal experiment s .58C; 60 and 90 second residence time Te58_60logF58_601F58_602 1zvalue1 1zvalue2 121 te58_60F58_60210Te58_60121 zvalue2 Te58_6057.85te58_6068.919Te58_90logF58_901F58_902 1zvalue1 1zvalue2 121 te58_90F58_90210Te58_90121 zvalue2

PAGE 134

122 Te58_9057.898te58_90104.3560C; 30 and 60 second residence time Te60_30logF60_301F60_302 1zvalue1 1zvalue2 121 te60_30F60_30210Te60_30121 zvalue2 Te60_3059.934te60_3035.899Te60_60logF60_601F60_602 1zvalue1 1zvalue2 121 te60_60F60_60210Te60_60121 zvalue2 Te60_6060.274te60_6067.65162C; 15 and 30 second residence time Te62_15logF62_151F62_152 1zvalue1 1zvalue2 121 te62_15F62_15210Te62_15121 zvalue2

PAGE 135

123 Te62_1561.521te62_1517.126Te62_30logF62_301F62_302 1zvalue1 1zvalue2 121 te62_30F62_30210Te62_30121 zvalue2 Te62_3061.497te62_3034.829Te58aTe58_60Te58_90 Te60aTe60_30Te60_60 Te62aTe62_15Te62_30 Te58meanTe58a() Te60meanTe60a() Te62meanTe62a() Te5857.874Te6060.104Te6261.509Te58stdstdevTe58a() Te60stdstdevTe60a() Te62stdstdevTe62a() Te58std0.024Te60std0.17Te62std0.012Step 4: Calculate the k values at equivalent temperatures from the equivalent time determined in step 3Conc58606090905.61084.251085.61083.851085.21067.810631052.1105 Conc62151530307.0510871086.91081.71084.910681063.21033.3103

PAGE 136

124 Conc60303030606060606.131087.251083.851085.610871085.21085.11083.051064.210671052.461047.91042.641051.39105 D58601te58_60logConc5801Conc5802 D60301te60_30logConc6001Conc6002 D62151te62_15logConc6201Conc6202 D58602te58_60logConc5811Conc5812 D62152te62_15logConc6211Conc6212 D60302te60_30logConc6011Conc6012 D5860133.914D621517.936D60303te60_30logConc6021Conc6022 D5860239.693D621528.819D6030115.587D6030216.047D6030313.1D58901te58_90logConc5821Conc5822 D60601te60_60logConc6031Conc6032 D62301te62_30logConc6221Conc6222 D58902te58_90logConc5831Conc5832 D62302te62_30logConc6231Conc6232 D60602te60_60logConc6041Conc6042

PAGE 137

125 D5890131.901D623016.53D60603te60_60logConc6051Conc6052 D5890231.977D623027.392D6060115.526D6060217.138D6060320.535D58aD58601D58602D58901D58902 D60aD60301D60302D60601D60602 D62aD62151D62152D62301D62302 D58meanD58a() D60meanD60a() D62meanD62a() D58stdstdevD58a() D60stdstdevD60a() D62stdstdevD62a() D627.669D5834.371D6016.075D62std0.832D58std3.177D58std3.177k622.303D62 k582.303D58 k602.303D60 k620.3k580.067k600.143Step 5: Estimation of the activation energyz00Te60Te58 logD58()logD60() z006.757z01Te62Te58 logD58()logD62() z015.579z02Te62Te60 logD60()logD62() z024.371

PAGE 138

APPENDIX C MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus acidoterrestris USING ARRHENIUS KINETICS Step 1: Retrieal of the dynamic temperature history and survival dataConstants: Universal Gas constant:Rval8.314 J/mol K Reading in temperature and concentration files:Conc11READPRN"c:\phd\Abacillus Batch 1 Day 1.txt"() Conc11090951003.4671062.4231057.451041.35103 Conc12READPRN"c:\phd\abacillus Batch 2 Day 1.txt"() Conc120951001041.9881068.1831041.67510327 Conc21READPRN"c:\phd\abacillus batch 1 day 2.txt"() Conc210951001041.081061.063105583.3395 Conc22READPRN"c:\phd\abacillus batch 2 day 2.txt"() Conc220951001047.3751051.407105840485 126

PAGE 139

127 Step 2: Selection of two arbitrary Ea values and determination of the corresponding G valuesEa1300000 Ea21.5Ea1 Ea2450000Day 1Temp95READPRN"c:\PhD\95C Profile 1.txt"() T95v()Temp95v1273 G95_10198vexpEa1 RvalT95v() = G95_12.264207947271421041 G95_20198vexpEa2 RvalT95v() = G95_20.000000000000000Temp100READPRN"c:\PhD\100C Profile 1.txt"() T100v()Temp100v1273 G100_10198vexpEa1 RvalT100v() = G100_11.367082529256211040 G100_20198vexpEa2 RvalT100v() = G100_20

PAGE 140

128 Temp104READPRN"c:\PhD\104C Profile 1.txt"() T104v()Temp104v1273 G104_10198vexpEa1 RvalT104v() = G104_12.965607210405721040 G104_20198vexpEa2 RvalT104v() = G104_20.000000000000000Day 2Temp902READPRN"c:\PhD\90C Profile 2.txt"() T902v()Temp902v1273 G902_10194vexpEa1 RvalT902v() = G902_11.276448602079581041 G902_20194vexpEa2 RvalT902v() = G902_20.000000000000000Temp952READPRN"c:\PhD\95C Profile 2.txt"() T952v()Temp952v1273 G952_10194vexpEa1 RvalT952v() =

PAGE 141

129 G952_12.614501353702911041 G952_20194vexpEa2 RvalT952v() = G952_20.000000000000000Temp1002READPRN"c:\PhD\100C Profile 2.txt"() T1002v()Temp1002v1273 G1002_10194vexpEa1 RvalT1002v() = G1002_11.088856151392371040 G1002_20194vexpEa2 RvalT1002v() = G1002_20.000000000000000Temp1042READPRN"c:\PhD\104C Profile 2.txt"() T1042v()Temp1042v1273 G1042_10112vexpEa1 RvalT1042v() = G1042_11.669688311183351040 G1042_20112vexpEa2 RvalT1042v() =

PAGE 142

130 G1002_20.000000000000000Step 3: Equate G values to the isothermal form to determine the respective equivalent isothermal exposure for each pair of dynamic thermal experiment s .95C and 90 second residence time 100C and 90 second residence time Te1Ea2Ea1 ()RvallnG95_1G95_2 TeEa2Ea1 ()RvallnG100_1G100_2 te95_1G95_1expEa1 RvalTe te100_1G100_1expEa1 RvalTe Temp95_1Te1 Temp100_1Te Temp95_1367.252Temp100_1374.206te95_117.094te100_1103.212Te2Ea2Ea1 ()RvallnG952_1G952_2 TeEa2Ea1 ()RvallnG1002_1G1002_2 te95_2G952_1expEa1 RvalTe te100_2G1002_1expEa1 RvalTe Temp95_2Te2 Temp100_2Te Temp95_2367.887Temp100_2373.42te95_224.176te100_2100.684104C and 90 second residence time TeEa2Ea1 ()RvallnG104_1G104_2 te104_1G104_1expEa1 RvalTe

PAGE 143

131 Temp104_1Te Temp104_1377.218te104_1103.652TeEa2Ea1 ()RvallnG1042_1G1042_2 te104_2G1042_1expEa1 RvalTe Temp104_2Te Temp104_2377.235te104_258.108Temp95aTemp95_1Temp95_2 Temp100aTemp100_1Temp100_2 Temp104aTemp104_1Temp104_2 Temp95meanTemp95a() Temp100meanTemp100a() Temp104meanTemp104a() Temp95367.569Temp100373.813Temp104377.227Temp95stdstdevTemp95a() Temp100stdstdevTemp100a() Temp104stdstdevTemp104a() Temp95std0.317Temp100std0.393Temp104std8.469103 Step 4: Calculate the k values at equivalent temperatures from the equivalent time determined in step 3Conc11090951003.4671062.4231057.451041.35103 Conc120951001041.9881068.1831041.67510327

PAGE 144

132 k9511lnConc1101Conc1121 te95_1 k10011lnConc1101Conc1131 te100_1 k10411lnConc1201Conc1231 te104_1 k10412lnConc1201Conc1231 te104_1 k9512lnConc1201Conc1211 te95_1 k10012lnConc1201Conc1221 te100_1 k95110.225k100110.076k104110.108k95120.187k100120.069k104120.108Conc210951001041.081061.063105583.3395 Conc220951001047.3751051.407105840485 k9521lnConc2101Conc2111 te95_2 k10021lnConc2101Conc2121 te100_2 k10421lnConc2101Conc2131 te104_2 k10422lnConc2201Conc2231 te104_2 k9522lnConc2201Conc2211 te95_2 k10022lnConc2201Conc2221 te100_2 k95210.096k100210.075k104210.136k95220.069k100220.067k104220.12609k95ak9511k9512k9521k9522 k100ak10011k10012k10021k10022 k104ak10411k10412k10421k10422 k95meank95a() k100meank100a() k104meank104a() k95stdstdevk95a() k100stdstdevk100a() k104stdstdevk104a() k950.144k1000.072k1040.12k95std0.064k100std0.004k104std0.012

PAGE 145

133 D952.303k95 D1002.303k100 D1042.303k104 Step 5: Estimation of the activation energyEa00Rvallnk95k100 Temp95Temp100 Temp95Temp100 Ea00127557.147 Ea01Rvallnk95k104 Temp95Temp104 Temp95Temp104 Ea0122070.511 Ea02Rvallnk100k104 Temp100Temp104 Temp100Temp104 Ea02175926.015Step 6: Used the newly estimated activation energy value as the initial gues s step 2stdeviationstdevEa() meannmeanEa() stdeviation125800.585meann8766.119

PAGE 146

APPENDIX D MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus acidoterrestris USING THERMAL DEATH TIME (TDT) KINETICS Step 1: Retrieal of the dynamic temperature history and survival dataConstants: Universal Gas constant:Rval8.314 J/mol K Reading in temperature and concentration files:Conc11READPRN"c:\phd\Abacillus Batch 1 Day 1.txt"() Conc11090951003.4671062.4231057.451041.35103 Conc12READPRN"c:\phd\abacillus Batch 2 Day 1.txt"() Conc120951001041.9881068.1831041.67510327 Conc21READPRN"c:\phd\abacillus batch 1 day 2.txt"() Conc210951001041.081061.063105583.3395 Conc22READPRN"c:\phd\abacillus batch 2 day 2.txt"() Conc220951001047.3751051.407105840485 134

PAGE 147

135 Step 2: Selection of two arbitrary Ea values and determination of thcorresponding G valueszvalue113.11 zvalue2zvalue11.5 zvalue219.665Day 1Temp95READPRN"c:\PhD\95C Profile 1.txt"() T95v()Temp95v1273 F95_10198t10T95t()273 ()121 zvalue1 = F95_10.98078F95_20198t10T95t()273 ()121 zvalue2 = F95_24.81836Temp100READPRN"c:\PhD\100C Profile 1.txt"() T100v()Temp100v1273 F100_10198t10T100t()273 ()121 zvalue1 = F100_13.21794F100_20198t10T100t()273 ()121 zvalue2 = F100_210.45777Temp104READPRN"c:\PhD\104C Profile 1.txt"()

PAGE 148

136 T104v()Temp104v1273 F104_10198t10T104t()273 ()121 zvalue1 = F104_15.4672F104_20198t10T104t()273 ()121 zvalue2 = F104_214.82268Day 2Temp952READPRN"c:\PhD\95C Profile 2.txt"() T952v()Temp952v1273 F952_10194t10T952t()273 ()121 zvalue1 = F952_11.06957F952_20194t10T952t()273 ()121 zvalue2 = F952_25.06584Temp1002READPRN"c:\PhD\100C Profile 2.txt"() T1002v()Temp1002v1273 F1002_10194t10T100t()273 ()121 zvalue1 = F100_13.21794

PAGE 149

137 F1002_20194t10T100t()273 ()121 zvalue2 = F100_210.45777Temp1042READPRN"c:\PhD\104C Profile 2.txt"() T1042v()Temp1042v1273 F1042_10112t10T1042t()273 ()121 zvalue1 = F1042_13.06341F1042_20112t10T1042t()273 ()121 zvalue2 = F1042_28.22173Step 3: Equate F values to the isothermal form to determine the respective equivalent isothermal exposure for each pair of dynamic thermal experiments.95C and 90 second residence time 100C and 90 second residence time Te95_1logF95_1F95_2 1zvalue1 1zvalue2 121 Te100_1logF100_1F100_2 1zvalue1 1zvalue2 121 te95_1F95_210Te95_1121 zvalue2 te100_1F100_210Te100_1121 zvalue2 Te95_193.81Te100_1100.869te95_1116.294te100_1110.449Te100_2logF1002_1F1002_2 1zvalue1 1zvalue2 121 Te95_2logF952_1F952_2 1zvalue1 1zvalue2 121

PAGE 150

138 te100_2F1002_210Te100_2121 zvalue2 te95_2F952_210Te95_2121 zvalue2 Te95_294.435Te100_2100.869te95_2113.641te100_2110.449104C and 90 second residence time Te104_1logF104_1F104_2 1zvalue1 1zvalue2 121 te104_1F104_210Te104_1121 zvalue2 Te104_1103.964te104_1108.956Te104_2logF1042_1F1042_2 1zvalue1 1zvalue2 121 te104_2F1042_210Te104_2121 zvalue2 Te104_2104.137te104_259.221Te95aTe95_1Te95_2 Te100aTe100_1Te100_2 Te104aTe104_1Te104_2 Te95meanTe95a() Te100meanTe100a() Te104meanTe104a()

PAGE 151

139 Te9594.122Te100100.869Te104104.05Te95stdstdevTe95a() Te100stdstdevTe100a() Te104stdstdevTe104a() Te95std0.312Te100std3.826106 Te104std0.087Step 4: Calculate the D-values at equivalent temperatures from the equivalent time determined in step 3Conc11090951003.4671062.4231057.451041.35103 Conc120951001041.9881068.1831041.67510327 D9511te95_1logConc1101Conc1121 D10011te100_1logConc1101Conc1131 D10411te104_1logConc1201Conc1231 D9512te95_1logConc1201Conc1211 D10412te104_1logConc1201Conc1231 D10012te100_1logConc1201Conc1221 D951169.731D1001132.394D1041122.386D951283.933D1001235.924D1041222.386Conc210951001041.081061.063105583.3395 Conc220951001047.3751051.407105840485 D9521te95_2logConc2101Conc2111 D10021te100_2logConc2101Conc2121 D10421te104_2logConc2101Conc2131

PAGE 152

140 D9522te95_2logConc2201Conc2211 D10022te100_2logConc2201Conc2221 D10422te104_2logConc2201Conc2231 D9521112.841D1002133.802D1042117.231D9522157.986D1002237.523D1042218.611D95aD9511D9512D9521D9522 D100aD10011D10012D10021D10022 D104aD10411D10412D10421D10422 D95meanD95a() D100meanD100a() D104meanD104a() D95stdstdevD95a() D100stdstdevD100a() D104stdstdevD104a() D95106.122D10034.911D10420.154D95std33.733D100std1.963D104std2.285k952.303D95 k1002.303D100 k1042.303D104 Step 5: Estimation of the activation energyz00Te100Te95 logD95()logD100() z0013.971z01Te104Te95 logD95()logD104() z0113.761z02Te104Te100 logD100()logD104() z0213.335Step 6: Used the newly estimated activation energy value as the initial guess in step 2stdeviationstdevz() meannmeanz() stdeviation0.265meann13.689

PAGE 153

LIST OF REFERENCES Benjamin, M. M. and Datta, A. R. 1995. Acid tolerance of enterohemorrhagic Escherichia coli. Applied and Environmental Microbiology. 61(4): 1669-1672. Besser, R. E., Lett, S. M., Weber, J. T., Doyle, M. P., Barrett, T. J., Wells, J. G., Griffin, P. M. 1993. An outbreak of diarrhea and hemolytic uremic syndrome from Escherichia coli O157:H7 in fresh-pressed apple cider. Journal of the American Medical Association. 269(17): 2217-2220. Blackburn, C. W., Curtis, L. M., Humpheson, L., Billon, C., and McClur, P. J. 1997. Development of thermal inactivation models for Salmonella enteritidis and Escherichia coli O157:H7 with temperature, pH and NaCl as controlling factors. International Journal of Food Microbiology. 38: 31-44. Buchanan, R. L. and Edelson, S. G. 1996. Culturing enterohemorrhagic Escherichia coli in the presence and absence of glucose as a simple means of evaluating the acid tolerance of stationary-phase cells. Applied and Environmental Microbiology. 62(11): 4009-4013. Center for Disease Control (CDC). 1996. Outbreak of Escherichia coli O157:H7 infections associated with drinking unpasteurized commercial apple juice British Columbia, California, Colorado, and Washington, October 1996. Morbidity and Mortality Weekly Review. 45(44): 974-975 Center for Disease Control (CDC). 1999. Outbreak of Salmonella serotype Muenchen infections associated with unpasteurized orange juice United States and Canada, June 1999. Morbidity and Mortality Weekly Review. 48(27):582-585 Deak, T. and Beuchat, L.R. 1993 Yeasts associated with fruit juice concentrates. Journal of Food Protection. 56 (9): 777-782 Diez-Gonzalez, F. and Russell, J. B. 1997. The ability of Escherichia coli O157:H7 to decrease its intracellular pH and resist the toxicity of acetic acid. Microbiology. 143: 1175-1180. Eiroa, M. N. U., Junqueira, Valeria C. A., and Schmidt, F. L. 1999. Alicyclobacillus in orange juice: occurrence and heat resistance of spores. Journal of Food Protection. 62 (8): 883-886. 141

PAGE 154

142 Ingham, S. C. and Uljas, H. E. 1998. Prior storage conditions influence the destruction of Escherichia coli O157:H7 during heating of apple cider and juice. Journal of Food Protection. 61(4): 390-394. Jones, M. C. 1968. The temperature dependence of the lethal rate in sterilization calculations. Journal of Food Technology. 3: 31-38. Komitopoulou, E., Boziaris, I. S., Davies, E. A., Broughton-Delves, J., Adams, M. R. 1999. Alicyclobacillus acidoterrestris in fruit juices and its control by nisin. International Journal of Food Science and Technology. 34: 81-85. Leyer, G. J., Wang, L., and Johnson, E. A. 1995. Acid adaptation of Escherichia coli O157:H7 increases survival in acidic foods. Applied and Environmental Microbiology. 61(10): 3752-3755. Mazzotta, A. S. 2001. Thermal inactivation of stationary-phase and acid-adapted Escherichia coli 0157:H7, Salmonella, and Listeria monocytogenes in fruit juices. Journal of Food Protection. 64(3): 315-320. McIntyre, S. J., Ikawa, N. P., Haglund, J., and Lee, J. 1995. Charateristics of an acidophilic Bacillus strain isolated from shelf stable juices. Journal of Food Protection. 58: 319-321. Miller, L. G. and Kasper, C. W. 1994. Escherichia coli 0157:H7 acid tolerance and survival in apple cider. Journal of Food Protection. 57(6): 460-464. Ocio, M. J., Fernandez, P. S., Alvarruiz, A., Martinez, A. 1994. Comparison of TDT and Arrhenius models for rate constant inactivation predictions of Bacillus sterothermophilus heated in mushroom-alginate substrate. Letters in Applied Microbiology. 19: 114-117. O Hara, G. W. and Glenn, A. R. 1994. The adaptive acid tolerance response in root nodule bacteria and Escherichia coli. Archives of Microbiology. 161: 286-292. Parish, M. E. 1997. Public health and nonpasteurized fruit juices. Critical Reviews in Microbiology. 23(2): 109-119. Parish, M. E. 1998. Orange juice quality after treatment by thermal pasteurization or isostatic high pressure. Lebensmittel-Wissenschaft and Technologie (Food science and technology). 31(5): 439-442. Semanchek, J. J. and Golden, D. A. 1996. Survival of Escherichia coli O157:H7 during fermentation of apple cider. Journal of Food Protection. 59(12): 1256-1259. Splittstoesser, D. F., Lee, C. Y. and Churey, J. J. 1998. Control of Alicyclobacillus in the juice industry. Dairy, Food and environmental Sanitation. 18(9): 585-587

PAGE 155

143 Splittstoesser, D. F., McLellan, M. R. and Churey, J. J. 1996. Heat resistance of Escherichia coli 0157:H7 in apple juice. Journal of Food Protection. 59(3): 226-229. Swartzel, K. R. 1984. A continous flow procedure for reaction kinetic data generation. Journal of Food Science. 49: 803-806. Vieira, M. C., Teixeira, A. A., Silva, C. L. M. 2001. Kinetic parameters estimation for ascorbic acid degradation in fruit nectar using the partial equivalent isothermal exposures (PEIE) method under non-isothermal continuous heating conditions. Biotechnology Progress. 17: 175-181. Vieira, M. C., Teixeira, A. A., Silva, F. M., Gaspar, N. and Silva, C. L. M. 2002. Alicyclobacillus acidoterrestris spores as a target for Cupuacu nectar thermal processing: kinetic parameters and experimental methods. International Journal of Food Microbiology. 77: 71-81. Walls, I. and Chuyate, R. 1998. Alicyclobacillus Historical perspective and preliminary characterization study. Dairy, Food and Environmental Sanitation 18(8): 499-503. Walls, I. and Chuyate, R. 2000. Spoilage of fruit juices by Alicyclobacillus acideterrestris. Food Australia. 52(7): 286-288. Welt, B. A., Teixeira, A. A., Balaban, M. O., Smerage, G. H., Hintinlang, D. E., and Smittle, B. J. 1997a. Kinetic parameter estimation in conduction heating foods subjected to dynamic thermal treatments. Journal of Food Science. 62(3): 529-534 & 538. Welt, B. A., Teixeira, A. A., Balaban, M. O., Smerage, G. H., and Sage, D. S. 1997b. Iterative method for kinetic parameter estimation from dynamic thermal treatments. Journal of Food Science. 61(1): 8-14. Wescott, G. G., Fairchild, T. M., and Foegeding, P. M. 1995. Bacillus cereus and Bacillus stearothermophilus spore inactivation in batch and continuous flow systems. Journal of Food Science. 60(3): 446-450. Winniczuk, P. P. and Parish, M. E. Minimum inhibitory concentrations of antimicrobials against micro-organisms related to citrus juice. 1997. Food Microbiology. 12: 373-381. Zhao, T., Doyle, M. P., Besser, R. E. 1993. Fate of enterohemorrhagic Escherichia coli O157:H7 in apple cider with and without preservatives. Applied and Environmental Microbiology. 59(8): 2526-2530. Zook, C. D. 1997. Isostatic high pressure inactivation kinetics of Saccharomyces cerevisiae ascospores and storage study of pressurized Valencia orange juice. M. S. Thesis, University of Florida.

PAGE 156

BIOGRAPHICAL SKETCH Vertigo Moody was born March 30, 1971 in Ft. Lauderdale, Florida. He began his academic life in 1989 after graduating from Ft. Lauderdale High School. He received a Bachelor of Science degree and a Master of Engineering degree in Agricultural and Biological Engineering at the University of Florida. He was awarded a McKnight Graduate fellowship and a General Electric graduate fellowship to complete his Master of Engineering degree and Doctor of Philosophy degree. 144


Permanent Link: http://ufdc.ufl.edu/UFE0002222/00001

Material Information

Title: Thermal Inactivation Kinetics of Escherichia coli and Alicyclobacillus acidoterrestris in Orange Juice
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: UFE0002222:00001

Permanent Link: http://ufdc.ufl.edu/UFE0002222/00001

Material Information

Title: Thermal Inactivation Kinetics of Escherichia coli and Alicyclobacillus acidoterrestris in Orange Juice
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: UFE0002222:00001


This item has the following downloads:


Full Text












THERMAL INACTIVATION KINETICS OF Escherichia coli AND Alicyclobacillus
acidoterrestris IN ORANGE JUICE















By

VERTIGO MOODY


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


2003















ACKNOWLEDGMENTS

The author would like to express sincere gratitude to his major advisor, Dr. Arthur

A. Teixeira for his confidence and enthusiasm throughout this research project. His

guidance and support were essential for successful completion of this body of work. The

author would also like to express gratitude and appreciation to his supervisory committee

(Dr. Glen H. Smerage, Dr. Mickey Parish, Dr. Robert Braddock, and Dr. Spiros

Svorounous) for their guidance and suggestions related to the research and the

completion of this publication.

Special thanks go to the faculty and staff of the Agricultural and Biological

Engineering Department, especially Dr. David Chynoweth and Dr. Roger Nordstedt for

the use of their lab space and equipment as well as Ms. Veronica Campbell for her

guidance and technical skills in assisting with the laboratory aspect of this research

project. Special thanks go to Dr. Braddock, Rockey Bryan and the staff at the Citrus

Research and Education Center for assisting the author in coordinating visits to the center

to conduct research and for troubleshooting problems with equipment. The author wishes

to thank Dr. Parish and Lorrie Friedrich for their assistance with the microbiological

aspect of this research project. Their help facilitated the completion of this project and

enhanced the skills of the author for handling microorganisms in a laboratory setting.

Finally, the author would like to thank his family and friends for their continued support

and patience throughout this milestone in life.











TABLE OF CONTENTS

Page


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

T A B L E O F C O N T E N T S ................................................ ............................................. iii

LIST OF TABLES ....................................................... ............ .. ............ vi

LIST OF FIGURES .................................................... .......... ................ viii

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

CHAPTER

1 ESTIMATING THERMAL KINETIC PARAMETERS FOR Escherichia coli IN
SINGLE-STRENGTH ORANGE JUICE USING TRADITIONAL ANALYSIS OF
ISOTHERMAL BATH EXPERIMENTAL DATA ..................................................1

Introduction ............... .. ...... ... .. ......... ................. ............... 1
L literature R eview ................. ....... .......................... ...... ........ .......... ...... .
M icrobiology of Fruit Juices ...................................................... .... ........... 2
M echanism of A cid Tolerance ........................................ ......................... 5
Spoilage ............................................................ ............. ........... 6
O bjectiv e s ........................................................... .. ..................................... . 7
M methods and M materials ................... .... ........ ..................... ... ...... .......... .... ....
Scope of W ork ................................................................ ......... ......... .........8........ 8
Preliminary Experiments ...... .... ....... ................... ........................... 9
Preparation of Cultures .. .... .. ..... ..... .. ...... .... ........... ... ............... .. .9
Source of strains ........ ....... ... ................. ...............
Acid adaptation preparation ............... ....... ...................... ............... 11
Experimental Apparatus .................. ...................................... 12
Isothermal Inactivation Experiments.................. ............................. 12
E stim ating D and z-values ........................................ ........................... 13
R results and D discussion ......... .................................................................. 14
Preliminary Experiments ................. .................................. 14
Saccharomyces cerevisiae .......... ............ ....... .... ............... 14
Escherichia coli cultured at neutral pH................. .............................14
Acid-tolerant Escherichia coli cultures................................... ............... 16
Therm al Inactivation of Escherichia coli....................................... ................. 17









2 ESTIMATING KINETIC PARAMETERS FOR THERMAL INACTIVATION OF
Escherichia coli IN ORANGE JUICE USING THE PAIRED EQUIVALENT
ISOTHERMAL EXPOSURES (PEIE) METHOD WITH A CONTINUOUS HIGH
TEMPERATURE SHORT TIME (HTST) PROCESS TREATMENT .....................47

Introdu action ...................................... ................................................. 4 7
Literature Review .................................. .. .. .. ...... .. ............48
F irst-order kinetics........... ........................................................... .... .... ... ... 49
The PEIE M ethod ........... .... ....... ......... ........ ... .... .. .......... .... 51
O b je ctiv e s ............. ...... ........... ................. .................................................5 2
M methods and M materials ............................. .................................. .... ...... ...... 53
Preparation of C ultures......... ................. .................................. ............... 53
Experim ental A pparatus ......................................................... ................ ..... 53
Calibration of Therm ocouples..................................... .......................... ....... 54
Continuous Dynamic Thermal Treatments ............................... ............... .55
T em perature Profi les ........................................ .. ............. ............... .56
Estimating D- and z-Values with the PEIE Method......................................57
V alidation E xperim ents ............................................................ .....................59
R results and D discussion ................... .... ........ .............. ......... .......................61
Continuous Dynamic Thermal Experiments Parameter Estimation.................61
Comparing PEIE and 3-Neck Flask Isothermal Methods .................................62
Validation Experiments ..................................................... 64

3 ESTIMATION OF KINETIC PARAMETERS FOR THERMAL INACTIVATION
OF Alicyclobacillus acidoterrestris IN ORANGE JUICE .........................................85

Intro du action ...................................... ................................................ 8 5
L literature R review ....................... ......... ...... ... .................... ............ .............86
Occurrences of Alicyclobacillus acidoterrestris in Juice Products .....................86
The PEIE M ethod in Arrhenius Kinetics............................................... 87
The PEIE M ethod and TDT Kinetics ...................................... ............... 90
O b j e c tiv e s ........................................................................................................9 2
M methods and M materials ....................................................................... ..................92
Preparation of C ultures...................................... ............ ............................ 92
Experim ental A pparatus ......... .......................................... ..... ...............93
Continuous Dynamic Thermal Treatments ............................... ............... .93
T em p eratu re P rofiles ........................................ ............................................94
R results and D discussion ................................. .............. ................ ......... 96
Param eter Estim ation by PEIE .............................................. ............... .... 96
Parameter Estimation using F value and TDT kinetics................. .......... 96









APPENDIX


A MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli
U SIN G ARRHEN IU S KINETICS ............................................... .....................109

B MATHCAD PROGRAM FOR THE PEIE METHOD WITH Escherichia coli
USING THERMAL DEATH TIME (TDT) KINETICS ............... ..................... 118

C MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus
acidoterrestris USING ARRHENIUS KINETICS ...............................................126

D MATHCAD PROGRAM FOR THE PEIE METHOD WITH Alicyclobacillus
acidoterrestris USING THERMAL DEATH TIME (TDT) KINETICS ...............134

L IST O F R E FE R E N C E S ...................................... ................................. ..................... ..... 14 1

BIOGRAPHICAL SKETCH ............................................................. ............... 144
















LIST OF TABLES


Table p

1-1. Plate counts of survivors grown in standard nutrient broth and pH-modified nutrient
broth for inducing acid tolerance ........................................ ......................... 23

1-2. D-values (seconds) for Escherichia coli in orange juice cultured at neutral pH
(standard culture) in preliminary experim ents ............................... ............... .31

1-3. D-values (seconds) for Escherichia coli in orange juice cultured at low pH (acid
adapted culture) in preliminary experiments.......... .............. .........................37

1-4. D-values (seconds) from thermal inactivation experiments for Escherichia coli
cultured at low pH ........... ....... ....... ..... ......... ...... ............ 44

1-5. Comparison of TDT kinetic parameters with published data from Mazzotta (2001)
and Splittstoesser et. al. (1996) using acid adapted and non-acid adapted
E scherichia coli in orange juice ................................................................... ......46

2-1. Calibration of therm ocouples ............................................................................. 69

2-2. Reynolds numbers for each flow rate for the continuous system.............................70

2-3 Rate constants used in Equation 2-1 for the heater section temperature profile.......74

2-4. Rate constants used in Equation 2-2 for the chiller section temperature profile.......74

2-5. Population survivor data for continuous experiments .................... ................75

2-6. Estimation of D- and z-values from each iteration of the PEIE method.................76

2-7. Comparison of D- and z-values estimated by traditional method using isothermal
treatments and PEIE method using continuous dynamic treatments .....................78

2-8. Kinetic parameters of thermal inactivation ofAlicyclobacillus acidoterrestris spores
in Cupuacu nector using the PEIE method and Isothermal method *......................78

2-9. Results of validation experiments, comparison of predicted number of survivors for
PEIE analysis and Traditional isothermal batch analysis with experimental number
of survive ors ......................................................................... ....................... 84









3-1. Rate constants used in Equation 2-1 and 2-2 for the heater and chiller sections
temperature profile for experimental set 1 .................................. ............... 101

3-2. Rate constants used in Equation 2-1 and 2-2 for the heater chiller sections
temperature profile for experimental set 2 ......................... .. ...................101

3-3. Population survivor data from Ultra High Temperature (UHT) heat treatments with
Alicyclobacillus acidoterrestris in orange juice...............................................102

3-4. Estimation of k and Ea values from each iteration of the PEIE method using
A rrhenius kinetics .................................... ... ... ... .. .. ...............103

3-5. Estimation of D- and z-values from each iteration of the PEIE method using TDT
k in etic s ............................................................................ 10 5

3-6. Comparison of TDT kinetic parameters with published data from various sources
using Alicyclobacillus acidoterrestris.................... .... .... ................... 107















LIST OF FIGURES


Figure Page

1-1. Growth curve showing light absorbance at a wavelength of 600 nanometer vs time
for Saccharomyces cerevisiae in yeast extract peptone dextrose (YEPD) broth.
Sets are runs conducted on separate days.................................... ............... 21

1-2. Growth curve showing absorbance of light at wavelength of 600 nanometer vs time
for Escherichia coli ATCC #9637 in nutrient broth. Sets are experiments
conducted on separate days .................................... .... .................................... 22

1-3. Experimental apparatus (photograph) ........................................... ............... 24

1-4. Experimental apparatus (diagram)................................................. ............... 25

1-5. Survivor curves from preliminary experiments at 500C, 54C and 56C for
Saccharomyces cerevisiae in orange juice cultured at neutral Ph (standard culture)26

1-6. Preliminary experiments survivor curve at 590C for Escherichia coli in orange juice
cultured at neutral pH (standard culture)............... ......... ...... ............... 27

1-7. Preliminary experiments survivor curves at 620C for Escherichia coli in orange
juice cultured at neutral pH (standard culture).......................... .... ............... 28

1-8. Preliminary experiments survivor curves at 640C for Escherichia coli in orange
juice cultured at neutral pH (standard culture)....... .. ........................................ 29

1-9. TDT curve from preliminary experiments with Escherichia coli in orange juice
cultured at neutral pH (standard culture). R2 value of 0.90.............................. 30

1-10. Survivor curves from preliminary experiments at 520C with Escherichia coli in
orange juice cultured at low pH (acid adapted culture). ...................................32

1-11. Survivor curves from preliminary experiments at 55C with Escherichia coli in
orange juice cultured at low pH (acid adapted culture) ....................................33

1-12. Survivor curves from preliminary experiments at 600C with Escherichia coli in
orange juice cultured at low pH (acid adapted culture) ....................................34

1-13. Family of survivor curves from preliminary experiments at 520C, 55C, and 60C
with Escherichia coli in orange juice cultured at low pH (acid adapted culture).....35









1-14. TDT curve from preliminary experiments with Escherichia coli in orange juice
cultured at low pH (acid adapted culture). R2 value of 0.99 .................................36

1-15. pH of broth vs. pH of orange juice product for Saccharomyces cerevisiae
preliminary experiments.............................. ............................... 38

1-16. Survivor curve from thermal inactivation experiments at 520C with Escherichia coli
in orange juice cultured at low pH (acid adapted culture) ....................................39

1-17. Survivor curve from thermal inactivation experiments at 550C with Escherichia coli
in orange juice cultured at low pH (acid adapted culture) ....................................40

1-18. Survivor curve from thermal experiments at 58C with Escherichia coli in orange
juice cultured at low pH (acid adapted culture) .............. ....................... .......... 41

1-19. Survivor curve from thermal inactivation experiments at 600C with Escherichia coli
in orange juice cultured at low pH (acid adapted culture) ....................................42

1-20. Family of survivor curves at 520C, 55C, 58C and 60C with Escherichia coli in
orange juice cultured at low pH (acid adapted culture) ....................................43

1-21. TDT curve from the thermal inactivation experiment with Escherichia coli in
orange juice cultured at low ph (acid adapted culture). R2 value of 0.98 ...............45

2-1. Photo of the Microthermics HTST Lab 25 Labscale Pasteurizer..........................67

2-2. Schematic Diagram of the flow of the Microthermics pasteurizer...........................68

2-3. Thermal profile of product at a hold tube nominal temperature of 58C and
residence tim es of 60 and 90 seconds ........................................... ............... 71

2-4. Thermal profile of product at a hold tube nominal temperature of 60C and
residence tim es of 30 and 60 seconds ........................................... ............... 72

2-5. Thermal profile of product at a hold tube nominal temperature of 62C and
residence times of 15 and 30 seconds ....................................... ................73

2-6. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic
param eters from the PEIE m ethod ................................................ ............... 77

2-7. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic
param eters from the PEIE m ethod ................................................ ............... 79

2-8. Comparison of TDT curves based upon data from the traditional and PEIE methods80

2-9. Comparison of TDT curves based upon data from the traditional and PEIE methods
for Alicyclobacillus acidoterrestris spores in Cupuacu nectar (Vieira et. al. 2002)
(Estimated curve based upon reference D-value and z-value)..............................81









2-10. Temperature history and measured and predicted survivor responses for validation
experiment I (10 second hold tube)....... .......................... ..... ............... 82

2-11. Temperature history and measured and predicted survivor responses for validation
experim ent II (15 second hold tube) ............................................. ............... 83

3-1. Thermal profile of product at a hold tube nominal temperature of 95C, 100C, and
104C for experim ental set one ........................................... ......................... 99

3-2. Thermal profile of product at a hold tube nominal temperature of 95C, 100C, and
104C for experim ental set tw o .............. ..................................... ................100

3-3. Arrhenius curve for Alicyclobacillus acidoterrestris in orange juice using kinetic
param eters from the PEIE m ethod .............................................. ............... 104

3-4. TDT curve for Alicyclobacillus acidoterrestris in orange juice using kinetic
param eters from the PEIE m ethod .............................................. ............... 106

3-5. Comparison of TDT curves based upon data from the PEIE method using TDT
kinetics and A rrhenius kinetics ................................................... .....................108















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

THERMAL INACTIVATION KINETICS OF Escherichia coli AND Alicyclobacillus
acidoterrestris IN ORANGE JUICE

By

Vertigo Moody

December 2003

Chair: Arthur A. Teixeira
Co-chair: Glen H. Smerage
Major Department: Agricultural and Biological Engineering

Growing concern about the safety of unpasteurized low-pH foods has changed the

view of the microbial loads supported by these products. Recent outbreaks of Salmonella

in single-strength unpasteurized orange juice and Escherichia coli 0157:H7in apple juice

have prompted food processors to seek ways of ensuring the safety of their products

without compromising consumer acceptance. Spoilage is also a concern as it relates to

the shelf life of fruit juice products. In order to achieve an optimum balance between

safety, shelf life, and quality, good estimation of thermal inactivation parameters is

essential for designing pasteurization processes that achieve all three goals.

The purpose of this study was to validate a method for estimating thermal

inactivation kinetic parameters of specific microorganisms. The method, called the

Paired Equivalent Isothermal Exposures (PEIE) method, may be applied to products that

are heated under non-isothermal conditions. This method simplifies the estimation of









parameters by eliminating the need to perform tedious isothermal bath experiments, while

still obtaining accurate estimations. The study was performed in three phases:

1) Estimating thermal kinetic parameters for Escherichia coli in single strength orange

juice using traditional analysis of isothermal bath experimental data; 2) Estimating

kinetic parameters for thermal inactivation of Escherichia coli in orange juice using the

PEIE method with end-point data from continuous high-temperature short-time (HTST)

process treatments and validation for each set of kinetic parameters, and 3) Estimating

kinetic parameters for thermal inactivation ofAlicyclobacillus acidoterrestris using the

PEIE method.

Estimating kinetic parameters from isothermal bath and continuous dynamic

thermal treatment data gave parameters that were different. To confirm which

parameters were more accurate, validation experiments were conducted at higher

temperatures. Using the parameters from both methods the number of survivors from

each experiment were compared with those predicted by each set of kinetics parameters.

Results from validation experiments with Escherichia coli showed that model predictions

agreed more closely with experimental data when kinetic parameters used were estimated

by the PEIE method rather than the traditional isothermal bath method. The process

conditions determined from the kinetic parameters estimated by the PEIE method yielded

a 39.7% shorter time than that determined by the isothermal bath method. The PEIE

method was used as the preferred method for estimating the kinetic parameters for

Alicyclobacillus acidoterrestris in single-strength orange juice.
















CHAPTER 1
ESTIMATING THERMAL KINETIC PARAMETERS FOR Escherichia coli IN
SINGLE-STRENGTH ORANGE JUICE USING TRADITIONAL ANALYSIS OF
ISOTHERMAL BATH EXPERIMENTAL DATA

Introduction

Recent outbreaks of Escherichia coli and Samonella in low-pH fruit juices

(including apple and orange) have prompted reevaluation of the ability of pathogenic

microorganisms to survive in these high-acid food products. Unpasteurized fruit juices

have become popular consumer products because flavor and texture quality are better

than in pasteurized juices. Escherichia coli 0157:H7 and Salmonella contaminated

orange and apple juice and apple cider have raised the attention of the Food and Drug

Administration, which previously considered high-acid foods with pH below 4.6 not to be

potentially hazardous to consumers. These outbreaks provide a compelling reason to

study these organisms' tolerance to low pH and to study their effect on the safety and

shelf life of these products. The design of pasteurization processes depends on estimating

the thermal inactivation kinetic parameters. Performing thermal inactivation experiments

on the acid-tolerant bacteria allows engineers to design thermal processes that more

completely reduce the number of pathogenic microorganisms in the product to more safe

levels. Accurate estimation of kinetic parameters is essential to food engineers. The

purpose of this study is to characterize the thermal inactivation behavior of potentially

pathogenic bacteria in orange juice.









Literature Review

Microbiology of Fruit Juices

Up to the latter part of the 20th century it was widely assumed that pathogenic

microorganisms could not survive in low-pH, high-acid foods because of the belief that

organic acids had an inhibitory and sometimes microbicidal effect (Parish 1997). The

Food and Drug Administration generally considers foods with a pH greater than 4.6 to be

potentially hazardous to consumers. Unpasteurized fruit juices have become a popular

consumer food product because their flavor retention is better than that of pasteurized

fruit juices. However, recent outbreaks of foodborne illness stemming from

unpasteurized fruit juices have brought to the forefront the need for pasteurization of all

processed fruit juices. Outbreaks involving Escherichia coli 0157:H7 and Salmonella

enterica in orange and in apple juices and apple cider have changed long held views on

the safety of fruit juices and other low-pH products.

Escherichia coli 0157:H7 was first confirmed as a health concern in juices after an

apple cider related outbreak in 1991 (Besser et. al. 1993). An outbreak of diarrhea and

Hemolytic Uremic Syndrome (HUS) in southern Massachusetts was traced back to

contamination of fresh-pressed apple cider (Besser et. al. 1993). Twenty-three persons

were identified with Escherichia coli 0157:H7 infections between October 23 and

November 24 of 1991. An epidemiological study based on this case showed that when

apple cider, with a pH ranging between 3.7 and 3.9, was inoculated with Escherichia coli

0157:H7, bacteria survived for 20 days at refrigerated conditions (80C) (Besser et. al.

1993). Another outbreak of Hemolytic Uremic Syndrome (HUS) caused by the

consumption of unpasteurized apple juice that was contaminated with Escherichia coli

0157:H7 was documented in 1996 (Parish 1997). In this outbreak a large producer of









fresh unpasteurized fruit products was implicated in the distribution of contaminated

product.

Salmonella has been isolated from apple cider samples (pH from 3.7 to 4.0)

associated with an outbreak of gastroenteritis (Besser et. al. 1993). In 1989 an incident of

typhoid fever caused by consuming orange juice contaminated with Salmonella typhi was

documented in a New York hotel restaurant in which there were 45 confirmed and 24

probable cases of typhoid fever with 21 hospitalizations (Parish 1997). In 1996 (on June

19, in the state of Washington and on June 23, in the state of Oregon) health officials

investigated clusters of outbreaks of diarrhea attributed to Salmonella and associated with

a commercially distributed unpasteurized orange juice (CDC 1999). Samples of the

unpasteurized orange juice yielded cultures of Salmonella when analyzed by the Food

and Drug Administration (FDA). There were approximately 300 confirmed cases

associated with this outbreak (CDC 1999).

These recent outbreaks of food poisoning from Salmonella and Escherichia coli

0157:H7 have called into question the safety of unpasteurized fruit juices and other low-

pH, high-acid food products. Pasteurization is the traditional method of inactivating

pathogenic and some spoilage-causing microorganisms in citrus products. The

inhibitory effect of acid concentration and low pH toward the growth of most pathogenic

bacteria alone does not ensure product safety (Parish 1997). Pathogens such as

Salmonella, Escherichia coli 0157:H7, .\/ngel//, Vibrio, and Staphyloccocus have been

shown to survive from hours to days and even weeks in various fruit juice products

(Parish 1997).









Miller and Kaspar (1994) showed the acid tolerance and survival of Escherichia

coli 0157:H7 in apple cider by testing two different strains. In their study they

inoculated Trypticase soy broth (TSB) adjusted to various pHs, and commercial apple

cider with those strains and observed the survival at each pH. Viable cells of Escherichia

coli 0157:H7 were still detectable in TSB at pH 2 after 24 hours of storage at refrigerated

conditions. In apple cider cells were still detectable after 14 days of storage at 40C.

Leyer et al. (1995) showed that acid-adapted Escherichia coli 0157:H7 survived for 81

hours in apple cider with a pH of 3.42 stored at 6C, whereas the non adapted cells

survived for only 28 hours.

Semanchek and Golden (1996) showed that pathogenic Escherichia coli 0157:H7

is capable of survival in apple cider for at least 10 days at a storage temperature of 20C

with a minimal decrease in population of viable cells. In a study by Zhao et al. (1993)

Salmonella survived in apple juice stored at 40C for more than 30 days at pH 3.6. These

studies revealed that storage conditions affect the resistance to acid of these pathogens.

Storage at refrigerated temperatures increases the time at which cells remain viable in the

product. Zhao et al. (1993) showed that Escherichia coli 0157:H7 was more rapidly

inactivated in apple cider stored at 25C than at 40C. Ingham and Uljas (1998) reported

that 84% to 91% of their inoculum of Escherichia coli cells was still viable in apple

cider, without preservatives, after 21 days when stored at 40C. Similar studies conducted

in different low-pH products showed an increase in the thermotolerance of Escherichia

coli 0157:H7. Leyer et al. (1995) reported that acid-adapted Escherichia coli 0157:H7

in fermented meats showed a higher thermotolerance.









Mechanism of Acid Tolerance

The mechanism of acid tolerance of bacteria is not completely understood. Several

theories have been proposed in an attempt to explain how bacteria are able to adjust and

maintain their internal pH within homeostatic limits. These theories include the buffering

capacity of cytoplasm, the low proton permeability of cells, and the extrusion of protons

from the cytoplasm by a membrane-bound proton pump (Benjamin and Datta 1995).

The antimicrobial effect of acids has been explained by the ability of

undisassociated molecules to enter the cell membrane and release protons. This release

of protons disrupts the electron transport system of the bacterial cell draining cellular

energy resources (Diez-Gonzalex and Russell 1997). The electron transport system is

highly dependent on the maintenance of a constant chemiosmotic potential across the

inner mitochondrial membrane to ensure steady production of adenosine triphosphate

(ATP) in the cellular environment. Bacteria capable of surviving in low pH (such as

lactic acid bacteria) are able to decrease intracellular pH when extracellular pH decreases

to maintain a low transmembrane pH gradient (Diez-Gonzalez and Russell 1997), thus

decreasing the dissipation of the proton-motive force.

Diez-Gonzalez and Russell (1997) studied the ability of Escherichia coli 0157:H7

to change its intracellular pH in response to a change in the extracellular pH as a

mechanism of acid tolerance and the ability to survive in low-pH products. They showed

that Escherichia coli 0157:H7 had a greater ability to control the level of acetate

concentration within its internal environment than a non-pathogenic Escherichia coli

strain. The 0157:H7 strain maintained a maximum internal concentration of acetate less

than 300 mM while the non-pathogenic strain accumulated as much as 500 mM of acetate

internally when the external pH dropped to 5.9. The significance of the concentration of









acetate in the cytoplasm gives insight into the ability of the bacteria to regulate the ions

and thus reduce the impact of dramatic changes in external pH. Protein synthesis appears

to be an essential aspect of the acid tolerance response of cells. O'Hara and Glenn (1994)

showed inhibiting protein synthesis with compounds such as chloramphenicol prevented

the development of acid tolerance in the cells. The nature of these proteins and their role

in the acid tolerance response are not known. They also reported that the capacity to

maintain alkaline intracellular pH is essential for the survival of root nodule bacteria in

acidic environments.

Spoilage

In addition to product safety, the population size of viable microorganisms that

remain in the product also affects the shelf life of the product with significant economic

implications. Pasteurized single-strength juices and frozen juice concentrates are the

predominant types of processed fruit juices commercially available. Yeasts, molds, and

lactic acid bacteria have been implicated in the spoilage of fruit juices (Deak et al. 1993).

Yeasts are the most problematic because of their ability to tolerate low-pH environment.

In particular, Saccharomyces cerevisiae is the most commonly isolated species of yeasts

from fruit juices that is responsible for spoilage. Twenty-five percent of yeast isolates

from frozen concentrate were identified as Saccharomyces cerevisiae in a survey

conducted in 1993 (Deak and Beuchat 1993). Yeasts lead to formation of films,

alteration of color, and change in viscosity. The fermentation caused by yeasts produce

products such as ethanol, carbon dioxide, and ethyl acetate, which alter the flavor of the

products. The production of gases may also compromise the integrity of product

packaging. The aim of pasteurization has been to eliminate the pathogenic









microorganisms, reduce the population of spoilage-causing microorganisms and to

inactivate enzymes for product safety and extended shelf life.

Recent outbreaks of Salmonella and Escherichia coli 0157:H7 in orange and apple

juice and in apple cider provide a compelling reason to understand these microorganisms'

tolerance to low pH in relation to their ability to cause disease and how that tolerance

affects thermal inactivation characteristics in those products for the purpose of food

safety. Estimating the thermal inactivation characteristics of these pathogenic organisms

in low-pH environments has both a food safety and economic impact on the design and

processing of fruit juice products. Because the assumption (that inactivation caused by

acid is sufficient) may no longer be valid, performing isothermal inactivation experiments

on the acid tolerant strains of pathogenic microorganisms such as Escherichia coli allows

engineers to design thermal processes that more completely reduce the number of viable

microorganisms to levels that ensure the safety of the product. Economic impacts of

microorganisms are also important in the food industry from a safety viewpoint and also

from a shelf-life viewpoint. Yeasts such as Saccharomyces cerevisiae are implicated as

the primary microorganisms responsible for spoilage of fruit juices and their limited shelf

life at refrigerated conditions.

Objectives

Because of these impacts on the fruit juice processing industry, the objectives of

this study were the following:

* To characterize the thermal inactivation kinetics of Saccharomyces cerevisiae and
Escherichia coli in orange juice

* To estimate thermal-death-time parameters (D- and z-value) for Escherichia coli
subjected to an acid adaptation procedure vs. standard cultures in orange juice









To compare the estimated parameters for Escherichia coli and Saccharomyces
cerevisiae with published data.

Methods and Materials

Scope of Work

The scope of work undertaken in this study has been divided into two parts to

determine the thermal inactivation kinetics of Escherichia coli and Saccharomyces

cerevisiae in single-strength orange juice. The Saccharomyces cerevisiae strain was a

wild type isolated from orange juice, and the Escherichia coli strain was obtained from

the American Type Culture Collection. Growth curves were created for each

microorganism to determine logarithmic and stationary phases of growth. Preliminary

experiments were used to help determine the temperature range in which thermal

inactivation of both microorganisms would yield measurable numbers of survivors in

order to plot survivor curves.

After the appropriate temperatures were selected, microorganisms were subjected

to different time-temperature combinations in order to estimate the thermal-death-time

(TDT) kinetic parameters. These kinetic parameters were estimated by traditional

methods of analyzing the survivor curves at each constant temperature. This method

entailed estimating the decimal reduction times (D-values) using linear regression to

construct the straight line of best fit on a semilog plot of survivors vs time (survivor

curve). The D-value is the reciprocal slope of this curve expressed as time required for

the curve to cross one log cycle, or time for one log cycle reduction of the population.

A semilog plot of D-values vs. temperatures allows estimation of the z-value, by taking

the reciprocal of the slope of the curve. The z-value is expressed as the number of

degrees of temperature change required for one log cycle change in D-value.









Preliminary Experiments

Analysis of the survivor curves generated from the preliminary experiments helped

determine at which temperatures to conduct the thermal inactivation experiments. For

the Escherichia coli, a procedure was developed and implemented to adapt the cells to

survival in a low-pH medium similar to the pH of single-strength orange juice. This

procedure more closely modeled the conditions experienced by Escherichia coli that

survive in contaminated orange juice.

Two sets of preliminary experiments were conducted. The first set involved the

thermal inactivation of both microorganisms grown in neutral-pH broth. The second set

involved acid-adapted Escherichia coli grown in low-pH broth.

Preparation of Cultures

Source of strains

The strain of Saccharomyces cerevisiae chosen for this study was obtained from

the yeast culture collection maintained in the microbiology laboratory at the University of

Florida's Citrus Research and Education Center, Lake Alfred, FL (Zook 1997). Stock

cultures were streaked onto potato dextrose agar (PDA) and incubated at 300C for 72

hours. A loop full of cells was aseptically transferred to 200 mL screw-cap flasks of yeast

extract peptone dextrose (YEPD) broth and incubated for 48 hours at 300C while

continuously shaken at 120 rpm on a junior orbit table shaker. Small aliquots of this

broth were then put into 1 mL vials placed into a -40C freezer and maintained as a stock

culture. A small loop full of broth was streaked onto slants of PDA refrigerated at 10C

and used as a working culture for a period of 3 weeks. After 3 weeks a new working

culture was created from the stock culture using the above procedure.









Growth curves for this particular strain of Saccharomyces cerevisiae were

documented by Zook (1997). A new set of growth curves was created to verify those

results. A small aliquot of working culture was inoculated into a flask of 200 mL of

YEPD broth and incubated at 300C. One-millimeter samples were withdrawn at

predetermined timed intervals for 30 hours. Turbidity of the samples was measured

optically using a Spectronic 40 spectrophotometer (Figure 1-1). As documented by Zook

(1997), the yeast completed their logarithmic phase after approximately 17 hours of

incubation.

The strain of Escherichia coli (preceptol culture ATCC #9637) used in this study

was obtained from the American Type Culture Collection (ATCC). Working and stock

cultures of this strain were made from the original freeze-dried culture obtained from

ATCC. The reconstituted cultures were inoculated into 200 mL of nutrient broth and

incubated at 370C while shaken at 120 rpm for 48 hours. Small aliquots of broth were

placed in 1 mL vials placed in a -4C freezer and maintained as a stock culture. A small

loop full of broth was streaked onto slants of nutrient agar, incubated for growth and

refrigerated at 100C. These slants were used as the working culture and maintained for a

period of 3 weeks. Thereafter new slants were prepared from stock cultures.

Growth curves for Escherichia coli were created in the same manner as those for

the Saccharomyces cerevisiae. In addition to measuring turbidity, the culture was plated

out after reaching logarithmic phase to estimate the concentration of cells. The average

concentration was 7.6 x 107 colony forming units (cfu)/mL after 25 hours and 13 x 108

cfu/mL after 36 hours. These numbers were used to estimate the proportion of inoculum

to medium in order to maintain a high initial concentration during the thermal









inactivation experiments (Figure 1-2). For the Saccharomyces cerevisiae it was

desirable to use the cells while in the logarightmic phase (Zook 1997); whereas, for the

Escherichia coli cells in the stationary phase were used (Buchanan and Edelson 1996,

O'Hara and Glenn 1994, Parish 1999).

Acid adaptation preparation

During the first set of preliminary experiments with Escherichia coli, thermal

inactivation was conducted by inoculating the medium with standard cultures (strains

grown at approximately neutral-pH conditions). Results showed that these cultures had

no resistance at all to the low-pH conditions of the orange juice at any lethal temperature.

It was reasoned that the cells should be subjected to an acid adaptation procedure in order

to increase their thermal resistance at low pH. This procedure would provide a closer

approximation of the growth environment the microorganisms would experience if

growing in contaminated orange juice.

For the second set of preliminary experiments, the Escherichia coli cells were

subjected to an acid adaptation procedure before thermal inactivation. In this procedure,

200mL of nutrient broth was inoculated with 1 mL of stock culture and incubated at 370C

for 24 hours. After 24 hours 6 mL of sterile 5% citric acid solution was injected into the

broth to lower the pH to approximately 4.5. The broth was then incubated for an

additional 24 hours. Then another 6mL of sterile 5% citric acid was injected into the

broth to lower the pH to approximately 3.5. The broth was then incubated for an

additional 48 hours. After 48 hours of incubation, the cells were ready to be used in the

thermal inactivation experiments. The final pH of the broth was at approximately 3.4.

A sample of broth was extracted and plated out for enumeration and to measure

final pH at each incubation interval. Below pH 3.7 there was a one or two log cycle









reduction in viable cells between the standard culture grown in neutral-pH broth and

those grown in low-pH broth (Table 1-1).

Experimental Apparatus

Heating at constant temperature was accomplished by using a three-neck flask

apparatus to reduce the thermal lags associated with glass or stainless steel tubes

submersed in a constant temperature bath. The flask was equipped with a mercury-in-

glass thermometer, rubber stoppers, a reflux condenser, a set of 9 needles, a 10 mL

syringe, eight 3 mL syringes, and a heating plate (Figure 1-3 and 1-4). The inoculated

orange juice was continuously mixed with a magnetic stirrer. A condenser placed in the

middle neck of the flask recovered evaporated water vapor from the orange juice to

assure a constant volume of inoculum.

Isothermal Inactivation Experiments

The flask, magnet, needles, rubber stoppers, condenser, and syringes were sterilized

before each experimental run. The thermometer was submerged in 10% ethanol alcohol

for 30 min to sanitize. The orange juice was reconstituted using sterile filtered deionized

water. The orange juice concentrate was a commercial brand at 440 Brix. Reconstitution

was performed under aseptic conditions using the recipe shown on the label (1 part

concentrate to 3 parts water). A 100mL sample of reconstituted orange juice was

aseptically poured into the flask. The flask was resealed using the rubber stopper, placed

on a heating plate, and allowed to reach equilibrium at the desired treatment temperature.

Then 7 mLs of inoculum was suctioned into one 10 mL sterile syringe (under aseptic

conditions) and injected into the flask. The effect of injecting the inoculum, which was at

incubation temperature, on the equilibrium temperature of the flask was determined by

allowing a 100mL sample of orange juice to equilibrate at each experimental









temperature. A thermocouple probe was used to measure the temperature drop of the

heated sample as the inoculum was injected into the three-neck flask apparatus. While

maintaining equilibrium conditions the temperature was observed over a period of 30

minutes for any significant change. The results indicated that for each 7 mL of inoculum

injected into the flask the temperature of the orange juice was lowered by precisely 1IC.

This lowered temperature was held constant throughout the experiment, and recorded as

the lethal temperature of exposure for the survivor curve resulting from that experiment.

A sample of inoculum was plated out before thermal inactivation to determine the

dilution of cells to be injected into the 100 mL of orange juice in the flask. After

injecting the inoculum into the flask, the timer was started, and eight successive 1 mL

samples were taken from each run at predetermined time intervals. The extracted 1 mL

samples were quickly transferred by injection into 9 mL of sterile peptone water

maintained in an ice water bath to immediately quench further thermal inactivation.

After the last sample was taken, three dilutions at each time interval were prepared and

plated in duplicate. Isothermal experiments were performed at 520C, 55C, 58C, and

620C.

Estimating D- and z-values

Four replicate experiments were conducted at each temperature. The D-values

obtained from each replicate at the same temperature were averaged for a single

representative D-value at each temperature. Statistical analysis was performed on these

values to determine the standard deviation. The z-value (C) was estimated from the

negative inverse slope of the linear regression line of the log D-value vs temperature.

Statistical analysis was performed using Microsoft Excel spreadsheet program using the 4

replicates at each testing temperature.









Results and Discussion

Preliminary Experiments

Saccharomyces cerevisiae

Survivor curves for preliminary experiments conducted at 500C, 54C, and 56C for

Saccharomyces cerevisiae are shown in Figure 1-5. Note that tailing was observed in all

of the survivor curves. This tailing phenomenon can probably be attributed to the

presence of two variant populations in the inoculum. For Saccharomyces cerevisiae the

two populations consist of spores and vegetative cells. Saccharomyces cerevisiae is

known to produce spores under normal growth patterns (Zook 1997). At the relatively

low temperature used in the preliminary experiments the more heat-resistant spores

remained viable to germinate in the media on enumeration of the survivors while the heat

quickly inactivated the vegetative population of cells.

To assure a more uniform population of yeasts, it would be necessary to separate

the spores from the vegetative cells. This separation requires growing the yeast on media

that encourages sporulation, separating the spores by centrifugation, and verifying

uniformity of population by microscopy. Our laboratory was not equipped for this

purpose, so further work on Saccharomyces cerevisiae was set aside for future study.

Escherichia coli cultured at neutral pH

Temperatures chosen for the preliminary experiments were based upon work by

Line et al. (1991) and Blackburn et al. (1997). Line et al. (1991) estimated the D- and

z-values of Escherichia coli 0157:H7 in ground beef subjected to various temperatures.

Although the heating characteristics for ground beef are different than those of orange

juice, it was useful to know the expected D- and z-values for nonpathogenic Escherichia

coli. Line et al. (1997) estimated D-values of 78.2 min at 51.60C, 4.1 min at 570C, and 18









sec at 62.7C in fatty ground beef. Blackburn et al. (1997) performed experiments with E

coli 0157:H7 in solutions that varied with pH and NaCl concentration. At 0.5% w/w

concentration of NaCl and pH of 4.3 (closest to pH of the orange juice at 3.8) the D-

values at 62.50C were 19 seconds, 34 seconds, 15 seconds, and 33 seconds for each

specific strain of 0157:H7. Using the results from both of these studies, the temperatures

chosen for the preliminary experiments were 590C, 62C, and 64C in an attempt to show

a significant difference between the D-values at each respective temperature.

Survivor curves obtained from preliminary experiments conducted at 590C, 62C,

and 64C with Escherichia coli cultured at neutral pH are shown in Figures 1-6 to 1-8.

The TDT curve resulting from these experiments is shown in Figure 1-9, with a z-value

of 6.4C. As shown in Figures 1-6 and 1-7 nearly all survivor curves showed tails at

590C and 62C. Therefore, D-values were obtained from the initial linear portion of the

curves. Results of these replicates at each temperature are shown in Table 1-2. It should

be noted that at the highest temperature (64C) the effective D-value was 1.2 seconds.

With such a rapid decrease in the population of survivors over a 10 second interval, a

sample extraction interval time of less than 5 seconds was needed to get countable plates

which yielded at least 4 data points for each survivor curve. With the current technique

for conducting isothermal bath experiments, this sample extraction interval was too short

for one individual to perform accurately.

The tailing phenomenon was observed only at the lower temperatures of 590C and

62C. The presence of tails suggested that a small fraction of the population was more

tolerant of these conditions. It was postulated that the two populations likely differed in

their tolerance to the acidic conditions of the orange juice. During this first set of









preliminary experiments the Escherichia coli cells were cultured in neutral-pH broth and

inactivated in low-pH orange juice. Existence of an acid-tolerant culture within the

inoculum was suspected to account for the appearance of tailing. Since acid will

inactivate vegetative cells the combination of it and the heat quickly kills the population

that is relatively susceptible to acid, whereas the more resistant population persists. The

lower temperatures used during the preliminary experiments were not high enough to

inactivate the remaining resistant population of Escherichia coli, yet this was the

population of greatest concern. Therefore, it became necessary to achieve a more heat-

resistant acid-tolerant population.

Acid-tolerant Escherichia coli cultures

To test this hypothesis of the existence of acid tolerant subpopulations in the

inoculum, a second set of preliminary experiments for the Escherichia coli was

conducted using acid-tolerant cultures. Figures 1-10 to 1-12 show survivor curves

obtained from these preliminary experiments for the acid-tolerant cultures at 520C, 55C,

and 60C (Figure 1-13 shows the family of curves). Figure 1-14 shows the TDT curve

resulting from these experiments at low pH. Table 1-3 lists the D-values obtained from

analysis of the survivor curves at each temperature. The acid-adapted cultures displayed

more resistance to heat than the non-acid-adapted Escherichia coli cultures. A

comparison of the Escherichia coli grown in nutrient broth where the pH had not been

adjusted vs adjusted pH nutrient broth showed a clear distinction between the thermal

resistances of the cultures. The tailing observed in the survivor curves of the Escherichia

coli grown in neutral broth did not show up in the survivor curves of the Escherichia coli

grown in low-pH broth. At each replicate a sample was taken at a sufficiently long

interval and plated out. The plates showed no growth at any of the temperatures for the









isothermal experiments conducted with the acid adapted cultures. At 52C, 55C, 58C,

and 60C the extended interval where no growth appeared on the plates was 56 min,15

min, 3 min, and 1.5 min, respectively. These results show that a more uniform

population existed among the cells of the acid-adapted Escherichia coli. The acid

adaptation procedure was successful in achieving its goals (elimination of the tailing

phenomenon and higher thermal resistance). The difference in the thermal resistance

between the two cultures along with the elimination of the tailing phenomenon

demonstrated the importance of acid adaptation of the inoculum when working with low-

pH fruit juices such as orange juice.

Thermal Inactivation of Escherichia coli

Based on results from the acid-tolerant preliminary experiments the best

temperatures selected to give a significant difference between D-values were 52C, 55C,

58C, and 60oC. At these temperatures the extraction intervals ranged from 7 minutes to

10 seconds. These times were appropriate to allow a sample to be taken at precise time

intervals.

Since pH was a major factor contributing to thermal inactivation of Escherichia

coli, it was important to measure the pH for consistency during each experimental run.

The pH of the orange juice used in the isothermal inactivation experiments vs the pH of

the growth broth before inoculation of the Escherichia coli into the orange juice is shown

in Figure 1-15. The pH of the orange juice ranged from 3.74 to 4.11 (a difference of

0.36) whereas the pH of the broth ranged from 3.29 to 4.09 (a difference of 0.8). For

the orange juice the difference between the minimum and the maximum pH yielded no

change in the number of survivors. To account for the difference in pH ranges, dilutions









were plated out at one above and one below the target dilution. This method would also

account for any variation in the initial concentration of cells.

The isothermal survivor curves for Escherichia coli at 520C, 55C, 58C, and 60C

are shown in Figures 1-16 through 1-19, respectively (Figure 1-20 shows the family of

curves). Table 1-4 shows the results of the thermal inactivation experiments for

Escherichia coli. The D-values were determined by taking an average of all the D-values

for all the replications at each temperature. The standard deviation for D-values at each

temperature was within 10% of the average value, thus the variation in the D-values

among replications was not a significant source of experimental error. The TDT curve

for the z-value ofEscherichia coli in orange juice is shown in Figure 1-21. The z-value

for this microorganism in orange juice was found to be 6.0C. This value agrees closely

with the z-value from the preliminary experiments with the acid tolerant cultures. The R2

-value from regression analysis was 0.98.

These results were compared with those reported in the literature for the thermal

inactivation of Escherichia coli in orange juice (Table 1-5). The cultures in this study

were subjected to an acid-adaptation laboratory procedure before inoculation using a non-

acid-resistant, low-heat-resistant strain of generic Escherichia coli, whereas Mazotta

(2001) and Splittstoesser et al. (1996) used a naturally-occuring, acid-tolerant, pathogenic

strain isolated from patients who had consumed contaminated product and showed

clinical symptoms of Eshcherichia coli infection. Because of the natural genetic

differences between generic and pathogenic strains of Escherichia coli, difference in heat

resistance results among the three studies were expected. More importantly the Mazotta

(2001) and Splittstoesser et al. (1996) study was expected to produce TDT kinetics









different than those estimated in this study. Mazotta used single-strength orange juice

adjusted to a pH of 3.9 with 1 N NaOH while Splittstoesser and colleagues used freshly

prepared apple cider and commercial brand apple juice concentrates. Similar to this

study, Mazotta conducted two sets of experiments using acid adapted and non-acid

adapted cultures. Both this study and Mazotta's showed a significant difference in the

heat resistance between acid adapted and non-acid adapted cultures. This difference has

a significant impact on the kinetic parameters estimated by thermal inactivation

experiments with orange juice. Table 1-5 shows the D-values for Escherichia coli from

all three studies. For both our study and Mazotta's study, thermal inactivation kinetic

parameters differ significantly between cultures grown in standard broth and those grown

in pH-adjusted broth. In both studies acid-adapted cultures were at least twice as

resistant as the non-acid-adapted cultures to thermal inactivation.

The acid tolerance of Escherichia coli is important to their survival in low-pH

products and may prove to be an important component of virulence for this species of

bacteria (as it is able to survive the acidic conditions of the stomach, which relates to the

infective dose). The acid tolerance of Escherichia coli significantly affects its thermal

inactivation characteristics. Our study shows the value of acid adaptation before

performing thermal inactivation experiments in low-pH products. The traditional

recommended pasteurization treatment for orange juice (98C for 10 seconds)

significantly affects the flavor of orange juice when compared with fresh untreated

orange juice (Parish 1998). Parish (1998) showed that a 23 degree decrease in the

temperature with the same treatment time had an impact on the sensory characteristics of

orange juice.









Most consumers prefer unpasteurized orange juice products to pasteurized

products. However the recent outbreaks of disease associated with unpasteurized fruit

juices has magnified the risk to consumer of these products. Data in this study suggest

that a minimal treatment process can achieve the necessary reduction in population of

pathogenic Escherichia coli in orange juice to a level that is safe for the consumer. With

parameters estimated in this study the calculated thermal process time that will reduce the

population of the acid-adapted Escherichia coli by 6 log cycles at a hold tube temperature

of 67C is 11 seconds; whereas for the non-acid-adapted culture it would be 3.2 seconds,

and could result in an unsafe product. The same difference in process time between acid-

adapted and non-acid-adapted cultures was shown for the strain used in Mazotta's study.

The thermal process time for a 6.0 log cycle reduction of the acid-adapted culture at 67C

is 22.81 seconds; whereas for the non-acid-adapted culture the thermal process time at the

same hold tube temperature is 13.74 seconds. These process times differ by 39.7%.

Results of both studies emphasize the importance of conducting experiments with

cultures that are similar to those found in the product. Using the thermal inactivation

kinetics from the non acid-adapted cultures from both studies leads to a significant

difference in the final population of microorganisms present in the product.














































5 10 15 20


Time (Hrs)


-$-Set One Rep 1
--Set One Rep 3
--iSet Two Rep 2


--Set One Rep 2
Set Two Rep 1


Figure 1-1. Growth curves showing light absorbance at a wavelength of 600 nanometer
vs time for Saccharomyces cerevisiae in yeast extract peptone dextrose
(YEPD) broth. Sets are runs conducted on separate days


0.5






0
0















































0 10 20 30 40
Time (Hrs)


i-4Set two Rep 1
Set two Rep 2


50 60 70 80


'-Set two Rep 1 Set three
--Set two Rep 1


Figure 1-2. Growth curves showing absorbance of light at wavelength of 600 nanometer
vs time for Escherichia coli ATCC #9637 in nutrient broth. Sets are
experiments conducted on separate days


0.6


0.5





0.4


E
0
o
S0.3
C,



0.2





0.1





0









Table 1-1. Plate counts of survivors grown in standard nutrient broth and pH-modified
nutrient broth for inducing acid tolerance


Acid-adapted Culture


Non-acid-adapted
Culture


Incubation Total Amount of pH of broth Plate Count pH of broth Plate
Hours Acid added (mL) (cfu) Count
(cfu)
48 3 6.729 2.8 x 109 8.1 4.3 x 109
2.5 x 109 3.2 x 109
72 6 4.760 2.2 x 109 8.2 3.6 x 109
2.4 x 109 1.7 x 109
96 10 3.694 1.6 x108 8.4 3.1 x109
1.4 x 108 1.2 x 109
120 12 3.360 1.2 x 107 8.4 1.4 x 109
1.5 x 107 7.6 x 108






































































. I I


Figure 1-3. Experimental apparatus (photograph)








Reflux Condenser


Thermometer






Inoculation
Needle















Heating Plate
Figure 1-4. Experimental apparatus (diagram)


Extraction
Needles















































200


300


400


500


600


700


Time (sec)


--50 C --54 C -A-56 C


Survivor curves from preliminary experiments at 500C, 54C and 56C for
Saccharomyces cerevisiae in orange juice cultured at neutral Ph (standard
culture)


8



7



6



05

o

4

0
-J


100


Figure 1-5.













































150
Time (sec)


-+-Run 1
'-Run 3


-,- Run 2


Figure 1-6. Survivor curves from preliminary experiments at 590C for Escherichia coli in
orange juice cultured at neutral pH (standard culture)


100


200


250


300











10






8










0
E6
o



S4
o
-J





2






0


150
Time (sec)


-Run 1


-* Run 2


-Run 3


Figure 1-7. Survivor curves from preliminary experiments at 620C for Escherichia coli in
orange juice cultured at neutral pH (standard culture)


200


250


300
















































0 5 10 15


Time (sec)


-mRun 1


-Run 2


Figure 1-8. Survivor curves from preliminary experiments 64C for Escherichia coli in
orange juice cultured at neutral pH (standard culture)










1


0.9


0.8 -


0.7


0.6


- 0.5

0

o 0.4


0.3


0.2


0.1


0 -
58 59 60


61 62 63 64 65


Temperature (oC)



Figure 1-9. TDT curve from preliminary experiments with Escherichia coli in orange
juice cultured at neutral pH (standard culture). R2 value of 0.90









Table 1-2. D-values (seconds) for Escherichia coli in orange juice cultured at neutral pH
(standard culture) in preliminary experiments


Temperature


Replicate 590C 620C 64C

1 6.25 4.81 1.1

2 7.14 3.57 1.3

3 6.55 2.95 NA

Average 6.64 3.77 1.2

Std Deviation 0.45 0.94 0.14

z-value = 7.0C
















































2000
Time (sec)


--Run 1


--Run 2


Figure 1-10. Survivor curves from preliminary experiments at 520C with Escherichia
coli in orange juice cultured at low pH (acid adapted culture).


1000


3000


4000














































400
Time (sec)


- Rep 1


- Rep 2


--Rep 3


Figure 1-11. Survivor curves from preliminary experiments at 550C with Escherichia
coli in orange juice cultured at low pH (acid adapted culture)


200


600


800
















































Time (sec)


-$-Rep 1


- Rep 2


Figure 1-12. Survivor curves from preliminary experiments at 600C with Escherichia
coli in orange juice cultured at low pH (acid adapted culture)


100


150















10



9



8



7



6



4-



o 4
-J



3



2



1



0
0 500 1000 1500 2000 2500 3000

Time (sec)

-*-52 C --55 C 60 C



Figure 1-13. Family of survivor curves from preliminary experiments at 520C, 550C, and
60C with Escherichia coli in orange juice cultured at low pH (acid adapted
culture)











2.8



2.6



2.4



2.2




E 2




0
-1.8 \
o


1.6



1.4



1.2



1
1 --I I I I 1 1
50 52 54 56 58 60 62

Temperature (oC)


Figure 1-14. TDT curve from preliminary experiments with Escherichia coli in orange
juice cultured at low pH (acid adapted culture). R2 value of 0.99









Table 1-3. D-values (seconds) for Escherichia coli in orange juice cultured at low pH
(acid adapted culture) in preliminary experiments

Temperature

Replicate 520C 55C 60C

1 424.25 112.21 16.05

2 342.2 100.4 16.46

3 136.9

Average 383.2 116.4 16.3


Std Dev


58.01


18.79


0.29










4.2





4





3.8





C3.6





3.4





3.2





3


10
Replicates


-- Orange Juice


-U- Broth


Figure 1-15. pH of broth vs. pH of orange juice product for Saccharomyces cerevisiae
preliminary experiments















































1500
Time (sec)


-'Rep 1


*Rep 2


Figure 1-16. Survivor curves from thermal inactivation experiments at 520C with
Escherichia coli in orange juice cultured at low pH (acid adapted culture)


500


1000


2000


2500


3000


Rep 3


Rep 4







40



9



8



7



6





0

4


-J
3



2



1



0
0 100 200 300 400 500 600 700

Time (sec)

--eRep 1 --Rep 2 -- Rep 3 --Rep 4 RK-ep 5


Figure 1-17. Survivor curves from thermal inactivation experiments at 550C with
Escherichia coli in orange juice cultured at low pH (acid adapted culture)







41



9



8



7



6











3-




2



1
0
4-

-J
3-












0-
0 50 100 150 200

Time (sec)

-- Rep 1 -- Rep 2 Rep 3



Figure 1-18. Survivor curves from thermal experiments at 58C with Escherichia coli in
orange juice cultured at low pH (acid adapted culture)











8




7




6




5
E
I3


04



0)
- 3


100


Time (sec)

-*-Rep -i- Rep 2 -- Rep 3 -- Rep 4




Figure 1-19. Survivor curves from thermal inactivation experiments at 600C with
Escherichia coli in orange juice cultured at low pH (acid adapted culture)



















10






8


E



o 6



0
O
-1

4






2


1000


1500


2000


2500


Time (sec)

--52C 55 C --60C 58C




Figure 1-20. Family of survivor curves at 520C, 550C, 58C and 600C with Escherichia
coli in orange juice cultured at low pH (acid adapted culture)









Table 1-4. D-values (seconds) from thermal inactivation experiments for Escherichia
coli cultured at low pH

D-values at various temperatures (seconds)

Replicate 520C 55C 580C 60C

1 398 151 36 16

2 370 148 32 20

3 308 146 34 18

4 336 147 19

Average 353 148 34 18

Std Deviation 39.08 2.18 2.27 1.52

z-value = 6.0 C











2.6




2.4




2.2




2

C.,
u,

| 1.8
0



1.6




1.4




1.2




1
1 --I I I I 1 1
50 52 54 56 58 60 62

Temperature (oC)


Figure 1-21. TDT curve from the thermal inactivation experiment with Escherichia coli
in orange juice cultured at low ph (acid adapted culture). R2 value of 0.98






46


Table 1-5. Comparison of TDT kinetic parameters with published data from Mazzotta
(2001) and Splittstoesser et. al. (1996) using acid adapted and non-acid
adapted Escherichia coli in orange juice

D58 (sec)
Acid Adapted This study 34
Mazzotta 300


Non-acid Adapted This study 10
Mazzotta 198
Splittstoesser et al. 60
















CHAPTER 2
ESTIMATING KINETIC PARAMETERS FOR THERMAL INACTIVATION OF
Escherichia coli IN ORANGE JUICE USING THE PAIRED EQUIVALENT
ISOTHERMAL EXPOSURES (PEIE) METHOD WITH A CONTINUOUS HIGH
TEMPERATURE SHORT TIME (HTST) PROCESS TREATMENT

Introduction

Achieving the best balance between quality retention and safety in heat sensitive

products that must be pasteurized is important in the fruit juice processing industry.

Recent outbreaks of Escherichia coli 0157:H7 and Salmonella in products such as

orange juice, apple juice, and apple cider have emphasized the ability of these

microorganisms to survive and grow in low-pH environments. Processing these products

to sufficiently reduce the probability of microbial survival for food safety and spoilage is

an essential design objective for food engineers. However, the popularity of

unpasteurized fruit juice is growing because of better flavor and texture retention over

heat pasteurized products. Understanding the thermal inactivation behavior of the target

microorganisms in the product is one key requirement to achieve a good balance between

food safety and quality retention. This behavior can be quantified by estimation of the D-

and z-values of the target microorganisms if first-order inactivation kinetics are observed.

The temperature-time combination for specific process goals can be determined using

these estimated thermal inactivation parameters. The greater the accuracy of these

estimations the more precise the temperature-time process condition can be determined









for the product. There are several methods used to estimate the thermal inactivation

parameters of microbial and chemical constituents. These methods include the following:

* Isothermal bath immersion with vials
* Isothermal three-neck flask
* Isothermal hold tube with sampling ports
* Paired equivalent isothermal exposures (PEIE) from non-isothermal data

Because estimated thermal inactivation parameters can have a significant impact on

the design of thermal treatment processes, it is essential to know which method provides

the best estimation of the parameters. The purpose of this study was to compare the

thermal inactivation kinetic parameters estimated by the traditional method of isothermal

bath analysis with those estimated by the PEIE method. It is possible to determine which

method provides the best estimation of the kinetic parameters by comparing the number

of survivors from a dynamic thermal process predicted mathematically using parameters

from each method with actual experimental survivor data.

Literature Review

Accurate estimation of kinetic parameters describing thermal inactivation of

microbial populations is of crucial importance in designing thermal treatments for

sterilization or pasteurization of liquid food products. Difficulty in achieving accurate

parameter estimation often leads to over processing in order to minimize risk to public

health. For products that are sensitive to heat this over processing comes at the expense

of flavor and nutrient degradation. In a study performed by Parish (1998) to compare

orange juice quality after treatment by thermal and isostatic high pressure pasteurization,

the orange juice processed at 75C for 10 seconds had a closer sensory score to fresh

extracted, frozen orange juice than that processed at 980C for 10 seconds. The study also

indicated that the flavor degradation after 16 weeks of storage at 40C and 8C was worse









for the product processed at the higher temperature. The results of this study showed the

importance of minimizing the thermal exposure to heat-sensitive products. Greater

accuracy in estimating kinetic parameters of thermal inactivation will allow food

processors to achieve maximum product quality without compromising food safety.

The logarithmic order of bacterial death is commonly described by a straight line

on a semilog plot of concentration of viable microorganisms vs. time of exposure to a

constant lethal temperature called a survivor curve. Survivor curves and their

temperature dependency are used as a mathematical model to determine the temperature-

time requirements for a pasteurization process. Commercial pasteurization processes rely

on such modeling of microbial population dynamics to design and operate thermal

processes for proper application of heat necessary to assure stability and safety of food

products, while reducing unnecessary overexposure of the products to heat, which can

severely degrade the quality of the products. Consumer demand for high quality

processed foods often drives the need for designing processes that are less detrimental to

product quality such as flavor and texture, while still reducing the microbial population to

levels that ensure safety from food borne illness.

First-order kinetics

The classical model of a first order reaction has been used for decades to predict the

processing temperature-time relationship of microbial thermal inactivation. Food

scientists and engineers have used survival curves, obtained from isothermal bath

experiments at different temperatures, as a means to estimate the kinetic parameters

describing these first-order reactions. These experiments are conducted by inoculating a

sample with a specific population of viable microorganisms and submerging vials

containing the sample into a constant temperature bath. These vials are removed at









different time intervals to obtain different extents of reaction that can be represented by

points on a survivor curve. The significant problems with isothermal bath experiments

are:

* Limited temperature range from which to calculate parameters for a wide range of
temperatures and to select parameters with good statistical confidence (Welt et al.
1997).

* Time lag of heat transfer encountered when the samples are heated from ambient to
reaction temperature and when cooled down from reaction temperature.

* Tedious preparation of small samples required to reduce thermal lags.

* Need for using buffer solutions rather than actual food product in many cases.

* Significant difference between experimental and actual processing conditions.

* Difficulty in obtaining statistically valid data at high temperatures when very rapid
reaction rates require short exposure times that cannot be accurately controlled.

Because of these problems, the use of kinetic parameters estimated by analysis of

data from isothermal batch experiments performed using vials submerged in a constant

temperature bath has often lead to inaccurate results when characterizing a continuous

ultra high temperature (UHT) or high temperature-short time (HTST) process such as

those used in commercial pasteurizations. An alternative technique for conducting

isothermal experiments involves using a three-neck flask instead of submerged vials into

an isothermal bath. This technique dramatically minimizes the thermal lags experienced

by the sample but it still has many of the problems associated with isothermal bath

experiments.

Wescott et al. (1995) proposed using a continuous thermal process to gather

isothermal data for the construction of survivor curves used to determine the thermal

inactivation kinetics of microorganisms. Measuring the number of survivors at various

locations in the hold tube and assuming a constant temperature isothermall) over the









entire length of the hold tube, survivor curves can be constructed for various

temperatures. The D- or k-values can be determined for each temperature and used to

construct a TDT curve or an Arrhenius plot from which a z-value or activation energy

value can be determined. This type of analysis of a continuous dynamic process from

obtaining isothermal data has been termed traditional analysis of a continuous dynamic

thermal process.

One problem with this method is the assumption of isothermal conditions along

the hold tube. Because UHT/HTST systems operate at very high temperatures the rate of

inactivation is very rapid and a small change in the temperature will yield different

thermal inactivation parameters from those based upon nominal operating temperature.

This method also can be highly dependent upon the experimental technique of the

researcher and how well the UHT/HTST system maintains constant temperature along its

hold tube.

The PEIE Method

To facilitate the design of a UHT/HTST process, Swartzel (1984) developed the

Equivalent Point Method (EPM) as a practical non-isothermal method for kinetic

parameter estimation. The EPM operates on the premise that any number of equivalent

isothermal processes may be obtained for a given dynamic thermal process so long as a

temperature-time profile is known as well as the extent of reaction. These equivalent

temperature-time combinations will fall on a straight line when plotted on a semilog

temperature-time graph for any assumed value of activation energy. He further

postulated that other straight lines constructed for different values of activation energy

would have different slopes and all intersect at a common universal "equivalent point"

from which the true value for the rate constant and activation energy could be determined









when two such "equivalent points" are found from two different dynamic processes with

different extents of reaction.

Welt et al. (1997 a, b) discovered that although Swartzel's universal "equivalent

point" did not exist, the concept of substituting dynamic processes with equivalent

isothermal processes could still be used to obtain kinetic parameters by employing an

iterative technique called the Paired Equivalent Isothermal Exposures method (PEIE).

They demonstrated the use of this method to estimate values for kinetic parameters that

were close to published values determined from traditional analysis of isothermal bath

data for Bacillus \ieail theii mephihi spores in pea puree. Vieira et al. (2001) used this

method to estimate kinetic parameters for ascorbic acid degradation and later for thermal

inactivation parameters ofAlicyclobacillus acidoterrestris spores in fruit nectar (Vieira et

al. 2002). With the PEIE method, it is no longer necessary to perform isothermal bath

experiments in order to estimate the reaction kinetics of reactants, whether they are

microorganisms, vitamins, or flavor components. Using a UHT/HTST process that more

accurately simulates the conditions the product will experience, the kinetic parameters

can be more accurately estimated and the temperature used is only dependent upon the

design parameters of the equipment and/or process. The PEIE method is a potential tool

for obtaining the kinetic parameters of a first order reaction more accurately than from

isothermal bath experiments.

Objectives

The purpose of this study was to apply the PEIE method to estimate thermal

inactivation kinetic parameters of Escherichia coli in orange juice and compare those

with parameters estimated using traditional analysis of survivor curves from isothermal

experiments. To achieve these goals the objectives of this project were the following:









* Estimate the kinetic parameters for thermal inactivation of Escherichia coli in
orange juice using the PEIE method with a continuous HTST process treatment.

* Compare the kinetic parameters estimated from the PEIE method with those
estimated from a traditional isothermal bath method.

* Validate the results by subjecting samples of inoculated product to random
dynamic temperature exposures beyond the range of temperatures used for
parameter estimation and then comparing the final population of surviving
microorganisms predicted from both sets of model parameters with the actual
population of microorganisms enumerated in the laboratory.

Methods and Materials

Preparation of Cultures

The strain of Escherichia coli used in these experiments was a preceptol culture

ATCC #9637. The microorganism was prepared in the same manner as detailed in

Chapter 1. These cells were also subjected to an acid adaptation procedure prior to

thermal inactivation as detailed in Chapter 1.

Experimental Apparatus

The experimental apparatus used in these experiments was the Microthermics

UHT/HTST Lab-25 lab-scale pasteurizer unit (Figure 2-1). All the heat exchangers of

the apparatus were shell and tube. The unit had two product inlets leading to the product

pump (Figure 2-2). Each inlet was equipped with a plug valve to control product flow.

The system was started by connecting a product reservoir to one inlet and a water

reservoir to the other inlet. The valve to the water reservoir was opened to provide water

to the system while operating conditions were being established and stabilized. Once the

system had reached steady state (stabilized at the desired operating conditions), the valve

to the product reservoir was opened to introduce product as the valve to the water

reservoir was closed. The main body of the pasteurizer was divided into three sections

consisting of the heater, hold tube, and chiller sections. Both the heater and chiller









sections were shell and tube heat exchangers made of stainless steel tubing with an outer

diameter of 0.375 in (0.9525 cm), wall thickness of 0.049 in (0.1244 cm) and length of

228 in (579.12 cm). The hold tubes also had an outer diameter of 0.375 in (0.975 cm) but

wall thickness of 0.035 in (0.0889 cm) and length of 200 in (508 cm) for each section of

the hold tube for a total length of 400 in (1016 cm). Using hot water as the heating

medium, the temperature of the product exiting the heater was controlled by adjusting the

steam pressure used to generate the hot water by a manual pressure flow control valve.

The hold tube section consisted of a series of tubes whose length could be adjusted by

adding extension tubes at the hold tube jumper panel. Hold times varied according to the

flow rate of the product and extension tubes used to extend the length of the hold tube

section for the appropriate residence time. Adjusting the speed of the product pump

controlled the flow rate of the system, which was measured by collecting a volume of

product exiting the system in a known period of time. The chiller section used a 50/50

mixture of water and propylene glycol as the cooling medium. To maintain pressure

when the product temperatures approached their boiling point in the system, an adjustable

back-pressure valve was located after the chiller section prior to the product exiting the

system. To monitor the temperature of the product and heating medium, thermocouple

probes were located at various points within the flow stream of the product and heating

medium.

Calibration of Thermocouples

Thermocouples were calibrated by comparing the temperature reading from each

thermocouple with the temperature reading from a standardized mercury-in-glass tube

thermometer (Arthur H. Thomas Company, National Bureau of Standards, Bureau file

117084) in a constant water bath. Correction factors for each thermocouple are shown in









Table 2-1. The average offset for each thermocouple was programmed into the

datalogger to eliminate the temperature reading as a significant source of experimental

error.

Continuous Dynamic Thermal Treatments

A commercial brand orange juice concentrate at 440 Brix was reconstituted using

sterile filtered deionized water. The reconstitution was performed following the recipe

indicated on the label (1 part concentrate to 3 parts water). Although the orange juice

was not reconstituted under aseptic conditions, the resident population of Escherichia coli

in the product was negligible when compared with the number of cells in the inoculum,

and the product was subjected to a thermal treatment within 30 minutes of reconstitution.

The product was inoculated with an acid-adapted Escherichia coli cell suspension prior to

thermal exposure to achieve a minimum initial concentration of 1 x 108 cfu/mL. Five

liters of orange juice were prepared along with 800ml of cell suspensions. The

pasteurizer was sanitized by circulating hot water at 83C through the heater, hold tube,

chiller sections and accessory tubes for a minimum of 30 minutes.

Once the sanitation cycle was completed the temperature of the pasteurizer was

adjusted to the desired experimental temperature and allowed to reach steady-state

conditions, upon which the product flow control valve was opened to allow the

inoculated product to flow through the unit. Temperatures at various locations

throughout the system were recorded using a datalogger attached to a notebook computer.

The pasteurizer unit had three thermocouples installed in the product flow stream and one

in the heating medium flow stream of the unit. The thermocouples were located after the

heating section, after the hold tube section, after the chiller section, and the flow tube of

the heating medium. All thermocouples were copper-constantan type T. The thermal









profile (temperature vs. time) of each experimental run was captured from each port and

saved as a text file that was used in the PEIE method. To produce replicate data for each

temperature, samples were collected in triplicate for each experimental run, and a

minimum of two experimental runs were conducted for each temperature-time

combination. An experiment involved a product cycle whereby a batch of product was

pumped through the system after using water to achieve a stable steady state condition.

Then the product and water reservoir valves were switched to allow water to run through

the system at the same conditions while another sample of product was being prepared.

Then the valves were switched and the product was pumped through the system and

samples were taken once again. Reynolds numbers for each experimental run indicated

transitional flow (Table 2-2). Although the PEIE method is not dependent upon the flow

behavior of the fluid in the pasteurizer unit, the flow behavior characteristics will

influence the designed residence times.

Temperature Profiles

The temperature was measured at the inlet of the product (initial product

temperature), after the heating section (at the entrance to the hold tube), after the hold

tube, and after the chiller section. Using the recorded temperature at each point, the

heater and chiller portions of the profile were constructed from heat transfer equations,

while the hold tube portions were constructed based upon measured data. The standard

profile for a shell and tube heat exchanger follows an exponential increase that can be

described by Equation 2-1,

T= A+ B(1-e-ht) (2-1)









where T is the temperature at any point within the heat exchanger at a specific time t, A is

the initial temperature of the product, B is the temperature of the product upon exit from

the heat exchanger, and h is the rate constant for the temperature change through the heat

exchanger. This equation yielded the calculated temperatures along the heater section of

the pasteurizer. The hold tube inlet and outlet temperatures were measured directly by

thermocouples. The temperature along the chiller section of the pasteurizer was

calculated using Equation 2-2.

T = B(e-ct) (2-2)

where B is the temperature of the product upon entrance into the chiller section, T is the

temperature at any point within the chiller section at a specific time t and c is the rate

constant for the temperature change through the chiller section. Knowing the residence

time of the product within the heater and chiller section of the pastuerizer, the

temperature profile was constructed by determining the parameters of Equations 2-1 and

2-2 using the boundary conditions of each section. The residence times for each section

were determined based upon the flow rate of the product and the diameters and lengths of

the tubes in all sections of the pasteurizer with the assumption of plug flow for simplicity.

The flow rates of the product were determined by measuring the amount collected in a

graduated cylinder over a specific time period.

Estimating D- and z-Values with the PEIE Method

The PEIE method uses the knowledge that for a given dynamic thermal exposure,

there exist any number of equivalent isothermal exposures (EIEs) that would yield the

same reduction in concentration of reactant. From two different dynamic thermal

exposures for a given reactant, the kinetic parameters for thermal inactivation of that









reactant can be estimated. The PEIE method as detailed by Welt et al. (1997a,b) is

carried out in Arrhenius kinetics to estimate first order rate constants (k) and activation

energy (Ea.) These parameters were converted into D- and z-values at the end of the

process. The following steps were taken from Welt et al. (1997a, b) and outline the PEIE

method used in this work:

1. The temperature histories along with the initial and final concentration of the
reactants from at least two distinct dynamic thermal processes were recorded.
Distinct means that each process produced a different extent of reaction.

2. One Ea value (Eal) was arbitrarily selected and the other Ea value (Ea2) was
arbitrarily chosen at 1.5 times Eal.

3. Using the recorded temperature-time data and the selected Eal and Ea2 values, the
respective EIEs (equivalent time (te) and temperature (Te)) for the pair of dynamic
thermal experiments were determined by equation 2-3, where G is the product
constituent reduction relationship factor, R is the universal gas constant (J/mole-K),
T(t) is temperature-time data, te is the equivalent time, and Te is the equivalent
temperature.


G = exp dt = t, exp a (2-3)
f R R*T(t) RR-T,

Equation 2-3 was applied twice for each data set using Eal first, then Ea2. This
application yielded two lines, each of which represented an infinite set of
temperature time combinations that were equivalent isothermal exposures for
respective Ea-values. The intersection of these two lines gave the equivalent time
and temperature for an isothermal process that would yield the same extents of
reaction as the dynamic thermal exposure for the reactants characterized by the Eal
and Ea2. This point is an Equivalent Isothermal Exposure.

4. The isothermal rate constants, k, for each process pair were calculated using the
EIE specification (te and Te) from step 3, the extent-of-reaction data from step 1
and Equation 2-4, where Co is the initial concentration at time zero and C is the
concentration of survivors remaining at the end of the process time.


In
k = (2-4)
te


Equation 2-5 determined the D-value.









2.303
D value = 03 (2-5)
k

5. Each pair of k values calculated from step 4 along with the equivalent temperature
from step 3 was used in Equation 2-6 to estimate an Ea value.


RTln
R. In k

Ea 2 (2-6)
(el Te2)



6. The newly estimated Ea value was used as the initial guess (Step 2) for the next
iteration. The process was repeated until the estimated Ea value from step 5
stopped changing. A TDT curve of D-value versus temperature was plotted to
estimate the z-value.

An algorithm using a commercial software package (Mathcad for Windows

Version 8.0) was used to execute the PEIE steps using the recorded thermal history, and

population survivor data (extent of reaction). It is important to note that the PEIE method

only works with constituents that follow a first order reaction.

Validation Experiments

The validation aspect of this study involved comparing the predicted number of

survivors for a particular process using the kinetic parameters estimated by the PEIE

method and those estimated by the 3-neck flask isothermal method (see Chapter 1) with

the actual number of survivors obtained from plate count enumeration of inoculated

orange juice. The validation experiments were performed with the same strain of

Escherichia coli and the same lab-scale pasteurizer unit. The inoculated orange juice

product was subjected to a dynamic process whereby the temperature of the heating

medium was varied to give a changing hold tube temperature. Samples of the product

were collected at a predetermined interval and serial dilutions were prepared, plated out

on nutrient broth and incubated for 48 hours. To observe if any injured cells were able









to recover the plates were incubated for an additional 24 hours and the number of

survivors was compared with those from the first 48 hours for any significant differences.

There were not significant differences between the two plant counts. The predicted

number of survivors was calculated by using numerical integration over the temperature-

time profile of each validation process, as follows.

Inactivation of vegetative cells at a constant lethal temperature follows a first-order

reaction process that is described by Equation 2-7 when C represents the concentration of

surviving viable cells, D, decimal reduction time, is the time interval required to reduce

the population of viable cells one log cycle (90%) of its former value (D=ln(10)/k), t is

the exposure time, and Co is the initial number of viable cells.


C = C10 (2-7)

Since the rate of population reduction is dependent on temperature, Equation 2-8 was

used to describe the variation of D with temperature T,


D(T) = Dol0 z (2-8)

where Do is reference D-value at reference temperature To and z is the temperature

interval required to change the value of D by one log cycle.

For a non-isothermal process where T varies with time, the lethal effect of the

changing temperature on the population can be determined by dividing the temperature

history into small time intervals (At) of constant temperature, use Equation 2-8 to

compute the D-value for each interval, and estimate the reduction in the population from

its former value using Equation 2-7 for each time interval. This process yields Equation









2-9, which can be used to find the change in the initial concentration of survivors over the

time interval At for a given temperature history, T (t).


At

Ct+At = C0 LD10 j (2-9)

For a time increment of differential magnitude, the total lethal effect over the total

process time is found by adding the contribution of all the time intervals to yield

Equation 2-10.


dt
t T(I)-T O
-C = C 10 (2-10)
to

Using the Do value obtained from the TDT curve from both methods, Equation 2-

10 was solved by numerical integration to estimate the number of surviving viable cells

(C) for each validation process. Equation 2-10 is the mathematical model used to predict

the relationship between survival response and the temperature history for a given set of

kinetic parameters. The predicted number of survivors was compared directly with the

number of survivors enumerated from plate count techniques.

Results and Discussion

Continuous Dynamic Thermal Experiments Parameter Estimation

Figures 2-3 through 2-5 show the temperature histories for the continuous

dynamic thermal experiments with hold tube temperatures at 58C, 60C, and 62C,

respectively. Each temperature included two different residence times in order to get two

processes with different extents of reaction. Temperature rate constants determined for

the heater and chiller sections were used to construct temperature profiles (Table 2-2 and









2-3). They were used with the measured hold tube temperatures to create a complete

profile for use in the PEIE method. These profiles along with the survivor data were used

to estimate the thermal inactivation parameters for Escherichia coli in single strength

orange juice.

Table 2-5 shows the population survivor data for all the continuous thermal

treatment experimental runs. The initial population was enumerated by plating out a

sample of the untreated inoculated orange juice before and after each experimental run.

Since the experimental runs were completed within 30 minutes after the orange juice was

inoculated, the inactivation of the cells due to low-pH environment in the orange juice

was not a significant source of error.

The final values of the activation energy from each set of related kinetic parameters

were determined after three iterations of the PEIE method (Table 2-6). The TDT curve

for Escherichia coli in orange juice yielded a z-value of 6.16 Celsius degrees with an R2

value from regression analysis of 0.99 (Figure 2-6). Three experiments were conducted

yielding six experimental pairs, and 13 sets of parameters. These parameters were

compared with the parameters estimated from the isothermal method with a 3-neck flask

described in Chapter 1.

Comparing PEIE and 3-Neck Flask Isothermal Methods

The isothermal bath temperatures ranged from 52C to 600C while the continuous

dynamic HTST hold tube temperatures ranged from 58C to 620C. The ranges

overlapped between the two processes at 580C and 60C. The D- and z-values obtained

from the traditional method using isothermal bath data and the PEIE method using

continuous dynamic data was compared (Table 2-7). At the two overlapping

temperatures there was a 16% difference at 580C and a 36% difference at 600C between









the D-values estimated by the two methods. The PEIE method yielded essentially the

same z-value as the 3-neck flask method. There is a slight difference in the slopes

between the TDT curves but the most noticeable difference is the shift of each curve

(Figure 2-7). This shift reflects the difference in reference D-values and will have an

impact on the predicted number of survivors when used in the mathematical model

(Equation 2-10).

Vieira et al. (2002) observed this phenomena when comparing the kinetic

parameters estimated by the PEIE method from continuous dynamic experiments with

those estimated by a traditional method using vials submerged in an isothermal bath

(Table 2-8 and Figure 2-9). The purpose of that study was to estimate the kinetic

parameters for Alicyclobacillus acidoterrestris spores in Cupuacu nector. It is important

to note that Vieira et al. estimated a reference D-value that was lower than with the PEIE

method, while this study estimated a value that was higher. This difference in the

comparisons between method between these studies can be explained by the

methodology used to generate the isothermal bath data (submerge vials in a water bath

vs. 3-neck flask). One of the significant problems with using vials submerged in an

isothermal bath is the thermal lag experienced by the cell suspension. This lag is

significant particularly when cooling the vials, where the temperature remains in the

lethal range well after the sample has been extracted from the bath leading to additional

inactivation beyond the measured time interval, thus over estimating the killing effect

within that time interval. For the 3-neck flask method the primary source of error is the

inability to extract the sample and quench the thermal inactivation process at the precise

time interval planned and recorded.









In this study the samples were extracted from the flask a few seconds prior to the

prescribed time interval in order to account for anticipated transit time for injection into

the chilled peptone water. Although this technique eliminated the possibility of any

thermal inactivation occurring after the planned time interval, the anticipated transit time

from the flask to the chilled water may lead to premature withdrawal reducing the lethal

effect experienced by the cell suspension by 2 to 3 seconds shorter than the prescribed

time interval, thus under-estimating the killing effect within that time interval. These

sources of error for both isothermal bath methods can have a significant impact on the

accuracy of estimating the thermal inactivation kinetic parameters when operating in the

UHT/HTST temperature ranges where the D-values range from a few seconds to less

than a second.

Validation Experiments

To verify which method yielded more accurate results, a series of validation

experiments was performed whereby continuous pasteurizations were carried out using

the lab scale pasteurizer with single strength orange juice. The purpose of these

experiments was to compare the number of survivors predicted mathematically using the

kinetic parameters from both the traditional and PEIE methods with the actual number of

survivors obtained by plate count enumeration. The temperature histories along with the

predicted and measured survivor responses from both sets of experiments are shown in

Figures 2-10 and 2-11. The hold tube temperatures were chosen to be above the range in

which the parameters were estimated to challenge the robustness of the model. The hold

tube residence time was set at 10 seconds for experiment one and 15 seconds for two.

Temperatures were recorded throughout the experimental run using a datalogger. The









results show that the model predictions with PEIE parameters were closer to the actual

number of survivors than those predicted with the 3-neck flask parameters (Table 2-9).

These results were not surprising because of the shift in the TDT curve between the

two methods and the implications of this shift as discussed previously. The significance

of this finding is that using the 3-neck flask to generate isothermal bath data over-

estimates the thermal inactivation rate constants, while using vials submerged in a

constant temperature water bath to generate isothermal bath data underestimates the

kinetic parameters.

For processing thermally sensitive products, this difference can have a significant

impact on the quality components of the product, such as flavor and vitamin retention.

The microbiological characterization of systems and processes is important to validate

lethality. Because of the short times for such high temperatures, using the traditional

method with isothermal bath experiments often leads to imprecise kinetic parameter

estimation. Cautious extrapolation is needed to relate parameters estimated under

laboratory conditions to UHT/HTST process conditions in the manufacturing facility.

This extrapolation may lead to further uncertainty. This extrapolation along with the

tedious nature of isothermal bath experiments, have made characterizing continuous high

temperature processes difficult. The PEIE method offers a valid alternative to isothermal

bath experiments for estimating thermal inactivation kinetics of microbiological

populations for the characterization of UHT/HTST systems with reasonable degree of

confidence.

For processing thermally sensitive products this difference can have a significant

impact on the quality components of the product, such as flavor and vitamin retention.









For example, if designing a process that will reduce the population of Escherichia coli in

orange juice by 6 log cycles at a temperature of 66C, the required hold tube residence

time would be 11 seconds based upon parameters from the traditional method and 8

seconds based upon those from the PEIE method. The advantage of the PEIE method

would be a 10% retention of components such as vitamin C. Using Veira et al. (2002)

data for Alicyclobacillus acidoterrestris, a six log cycle reduction in the population at

95C would result in a 10.08 minutes difference between the PEIE method and the

isothermal method, a significant impact on shelf-life of the product.

The PEIE method was developed to overcome some of the problems associated

with isothermal bath experiments. The method is easier and faster for estimating kinetic

parameters by saving laboratory time and equipment, and the kinetic parameters

estimated using this method would provide better results than those from isothermal bath

experiments. These parameters can be used in optimization techniques to determine the

best balance in thermal processes between food safety and quality. The PEIE method can

be applied to estimate kinetic parameters describing thermal inactivation of

microorganisms or thermal degradation of quality factors in a more realistic way using

real time processing equipment and conditions.













.1 z /'T,


*


Figure 2-1. Photo of the Microthermics HTST Lab 25 Labscale Pasteurizer


?le













r----------------------------..-




5 -
2e
Pue 5 I
il fI Ws


Figure 2-2. Schematic Diagram of the flow of the Microthermics pasteurizer









Table 2-1. Calibration of thermocouples


Replicate Thermocouples Mercury in glass Thermocouple Correction Factor
# reading (C) reading (C) (C)


71.8+0.02
72.00.02
72.50.02

750.02
750.02
750.02

75.60.02
75.60.02
75.60.02

75.70.02
75.60.02
75.60.02


69.5+1.0
70.21.0
71.01.0

73.41.1
72.7+1.1
74.1+1.1

741.1
74.5+1.1
74.11.1

74.21.1
74.21.1
74.3+1.1


2.31.0
1.8+1.0
1.5+1.0

1.61.1
2.3+1.1
1.51.1

1.61.1
1.1+1.1
1.51.1

1.51.1
1.41.1
1.51.1






70


Table 2-2. Reynolds numbers for each flow rate for the continuous system

Temperature (C) Residence Time (sec) Flow Rate (ml/min) Reynolds Number
58 60 480 619
90 320 413

60 30 960 1239
60 480 619

62 15 640 2491
30 960 1032













60

S50
0

30

2 30
E
' 20


100


Time(sec)


70

60

50
0
- 40

C 30
2
T 20


0 50 100


150 200


Time (sec)


Figure 2-3. Thermal profile of product at a hold tube nominal temperature of 58C and
residence times of 60 and 90 seconds


150












60

( 50
o
,- 40

S30
20
E 20
I-


0 20 40
Time(sec)


70

60

06 50
50


40
G 30
2
T 20


60 80


50 100
Time(sec)


150


Figure 2-4. Thermal profile of product at a hold tube nominal temperature of 60C and
residence times of 30 and 60 seconds






























10 20 30 40


Time (sec)


40
Time (sec)


Figure 2-5. Thermal profile of product at a hold tube nominal temperature of 62C and
residence times of 15 and 30 seconds


0
S40


I 30

20


10









Table 2-3 Rate constants used in Equation 2-1 for the heater section temperature profile.

Temperature Residence Time (sec) B h
(C)
58 60 58.6 -0.03273
90 58.6 -0.02101

60 30 59.91 -0.06545
60 59.91 -0.03152

62 15 61.34 -0.13558
30 61.34 -0.06779


Table 2-4. Rate constants used in Equation 2-2 for the chiller section temperature profile.

Temperature Residence Time (sec) B c
(C)
58 60 58.27 -0.09105
90 57.92 -0.64

60 30 60.28 -0.185
60 60.62 -0.093

62 15 61.23 -0.3742
30 61.94 -0.18285








Table 2-5. Population survivor data for continuous experiments


Hold tube Replication Residence Initial Number of C/Co
Temperature (C) Time (sec) Population Survivors (cfu)
(cfu)


5.6x108
4.3x108
5.6x108
3.9x108

6.1x108
7.3x108
3.9x108
5.6x108
7.0x108
5.2x108
5.1x108

7.0x108
7.0x108
6.9x108
1.7x108


5.2x106
7.8x106
3.0x104
2.1x104

3.1x106
4.2x106
7.0x105
2.5x104
7.9x104
2.6x105
1.4x105

4.9x106
8.0x106
3.2x103
3.3x103


9.3x10-3
1.8x10-2
5.4x10-5
5.4x10-5

5.1x10-3
5.8x10-3
1.8x10-3
4.5x10-5
1.1x10-4
5.0x10-4
2.8x10-4

7.0x10-3
1.1x10-2
5.0x10-6
1.9x105









Table 2-6. Estimation of D- and z-values from each iteration of the PEIE method


Iteration 1


Iteration 2


Iteration 3


Iteration 4


Initial Ea 20,000 J 62,089 J 267,398 J 342,711 J
Guess
C Residence D(sec) k(sec-1) D(sec) k(sec-1) D(sec) k(sec-1) D(sec) k(sec-1)
Time (sec)
58 60 27.3 0.084 29.53 0.078 29.53 0.078 29.53 0.078
32.0 0.072 34.56 0.067 34.56 0.067 34.56 0.067

90 27.5 0.084 25.5 0.09 27.51 0.084 27.51 0.084
25.6 0.09 27.58 0.083 27.59 0.083 27.59 0.083

60 30 12.0 0.191 13.02 0.177 13.03 0.177 13.03 0.177
12.42 0.185 13.41 0.172 13.41 0.172 13.41 0.172
10.14 0.227 10.94 0.21 10.95 0.21 10.95 0.21

60 12.76 0.18 13.77 0.167 13.77 0.167 13.77 0.167
14.08 0.163 15.20 0.151 15.20 0.151 15.20 0.151

62 15 2.69 0.854 6.95 0.331 6.95 0.331 6.95 0.331
2.99 0.769 7.73 0.298 7.73 0.298 7.73 0.298

30 5.20 0.442 5.63 0.409 5.63 0.409 5.63 0.409
5.89 0.391 6.37 0.362 6.4 0.362 6.4 0.362

Estimated Ea 62,089 J 267,398 J 342,711 J 342,711 J











1.6




1.4




1.2




1-




S0.8

0
-J
0.6




0.4




0.2




0
57 58 59 60 61 62 63

Temperature (oC)

-Iteration 1 -m- Iteration 2 -A- Iteration 3

Figure 2-6. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic
parameters from the PEIE method









Table 2-7. Comparison of D- and z-values estimated by traditional method using
isothermal treatments and PEIE method using continuous dynamic treatments


Isothermal
(3-neck flask)


Dynamic
(PEIE)


Temperature Average D- Standard D-value Standard
(C) value (sec) Deviation (sec) Deviation
52 353 39.08
55 148 2.18
58 34.7 2.27 29.8 3.3
60 18 1.52 13.27 1.54
62 6.93 0.47
z-value (C) 5.99 6.16


Table 2-8. Kinetic parameters of thermal inactivation ofAlicyclobacillus acidoterrestris
spores in Cupuacu nectar using the PEIE method and Isothermal method *


D950C (min) 5.5 + 1.2 3.82 + 0.48
z(oC) 31 + 29+10
R2 0.87 0.98
No. of observations 22 26
*Source: Veira et al. (2002)


PEIE Method


Isothermal (submerged vials)










1.6




1.4




1.2




1-




E 0.8

0

0.6




0.4




0.2




0
57 58 59 60 61 62 63

Temperature (oC)

--Iteration 1 --- Iteration 2 -A- Iteration 3


Figure 2-7. TDT curve for acid tolerant Escherichia coli in orange juice using kinetic
parameters from the PEIE method

















2.5





2





S1.5


01


1


0.5





0
50 52 54 56


58 60 62


Temperature (oC)


* Traditional


m PEIE


Figure 2-8. Comparison of TDT curves based upon data from the traditional and PEIE
methods

















1.2 t


S0.8



0
o 0.6
o
,-J



0.4




0.2


90 95 100 105 110 115

Temperature (oC)


- 4- PEIE


-- Isothermal


Comparison of TDT curves based upon data from the traditional and PEIE
methods for Alicyclobacillus acidoterrestris spores in Cupuacu nectar
(Vieira et. al. 2002) (Estimated curve based upon reference D-value and z-
value)


Figure 2-9.

































- Isothermal
- -PEIE
Experimental
-Temperature


Time (seconds)


Figure 2-10. Temperature history and measured and predicted survivor responses for
validation experiment I (10 second hold tube)































M M M M M


- Isothermal
- PEIE
Experimental
-Temperature


0 5


10 15 20 25 30 35 40


Time (seconds)


Figure 2-11. Temperature history and measured and predicted survivor responses for
validation experiment II (15 second hold tube)







84


Table 2-9. Results of validation experiments, comparison of predicted number of
survivors for PEIE analysis and Traditional isothermal batch analysis with
experimental number of survivors


Experiment Time Temp (C) Initial PEIE Isothermal Experimental
(sec) (cfu) Predicted Predicted
I 15 65 5.4x108 3.51x103 4.95x104 5.15x103
8.0xl03

II 10 65 5.4x108 1.98x103 8.59x104 1.0x103
1.25x103


Hold Tribe


~llrvivc~r8 (cfill
















CHAPTER 3
ESTIMATION OF KINETIC PARAMETERS FOR THERMAL INACTIVATION OF
Alicyclobacillus acidoterrestris IN ORANGE JUICE

Introduction

The recent discovery of Alicyclobacillus acidoterrestris in high-acid pasteurized

fruit juices and its ability to cause spoilage in these products have become concerns for

processors in the design of thermal pasteurization processes. Alicyclobacillus

acidoterrestris are sporeforming thermophilic bacteria that grow well in low pH

environments. These characteristics of the bacteria can be problematic since all shelf

stable and refrigerated fruit juices are pasteurized at temperatures below the lethal range

of Alicyclobacillus. This inadequate processing can lead to premature spoilage of the

product with risk of recall from the marketplace. Accurate estimation of the thermal

inactivation kinetic parameters that are used in a model to predict the number of survivors

is essential to establish optimum process conditions to assure a low probability of

spoilage of the product without over processing the product, which leads to degradation

of juice quality important to consumers. It has been shown in chapters 1 and 2 that the

Paired Equivalent Isothermal Exposures (PEIE) method is valid and accurate for the

estimation of kinetic parameters. The PEIE method is not just limited to using Arrhenius

kinetics for estimation of thermal inactivation kinetic parameters (k and Ea). It may also

be possible to use the PEIE method with Thermal Death Time (TDT) kinetics (D- and z-

value) which are more commonly used by food scientists. Orange juice was the product









chosen for this study because of the spoilage problems that have been documented

involving Alicyclobacillus acidoterrestris in orange juice. Using the PEIE method to

analyze continuous dynamic thermal treatment data, thermal inactivation kinetic

parameters were estimated for Alicyclobacillus acidoterrestris in orange juice.

Literature Review

Occurrences of Alicyclobacillus acidoterrestris in Juice Products

Fruit juices with a pH below 4.0 have been considered susceptible to spoilage only

by microorganisms of low heat resistance such as molds and acid-tolerant non-

sporeforming bacteria (Eiroa et al. 1999). Because of the low resistance to heat by these

microorganisms, pasteurization processes designed with temperatures ranging from 85C

to 95C were thought to be sufficient to inactivate these spoilage-causing microorganisms

(Eiroa et. al. 1999). The first reported incidence of a food product being spoiled by

acidophilic sporeformers was in Germany with apple juice (Walls and Chuyate, 1998). It

was determined that this spoilage microorganism was Bacillus acidoterrestris, which was

later named Alicyclobacillus acidoterrestris. Spoilage by this microorganism leads to off

flavors in the products similar to the taste of phenolic substances, odors of a disinfectant

and pronounced cloudiness (Eiroa et. al. 1999; Walls and Chuyate 1998).

Fortunately, Alicyclobacillus acidoterrestris does not appear to be pathogenic

according to Walls and Chuyate (2000), who conducted pathogenicity studies with the

bacteria in mice and guinea pigs. Although Alicyclobacillus acidoterristris is not a safety

concern for industry, it is a serious economic issue. During the 1990's Alicyclobacillus

acidoterrestris was presenting itself as a spoilage problem in shelf stable juice products

(Walls and Chuyate 1998). In 1994 there was a report of off odors in apple juice caused

by gram positive rods isolated from the juice and showing characteristics similar to









Alicyclobacillus acidoterrestris (Eiroa et. al. 1999). Spoilage of juices by these bacteria

is not readily detected since there is often little sedimentation and no gas produced that

would distort the product package. In juice inoculation studies, it was discovered that

Alicyclobacillus acidoterrestris grew well in orange, apple, tomato, and grape juice in

which the pH of the juices ranged from 3.47 to 4.27 (Walls and Chuyate 2000).

Alicyclobacillus acidoterrestris is a new spoilage microorganism that must addressed by

the juice industry and other processors of low pH food products.

Current temperatures used for pasteurization are insufficient to inactivate spores of

these bacteria in fruit juices, yet thermally overprocessing the product can lead to

unacceptable quality degradation of the product. Because sporeforming bacteria of

importance in foods are rarely as acid tolerant as Alicyclobacillus acidoterrestris, it is

important to characterize the thermal inactivation behavior in populations of this

microorganism in order to design processes that will reduce the probability of spoilage

for shelf stable products while maintaining quality factors acceptable to the consumer.

The PEIE method has been reported in recent literature to be useful in obtaining greater

accuracy in parameter estimation (Welt et al. 1997a, b). Using the PEIE method to obtain

thermal inactivation kinetic parameters for Alicyclobacillus acidoterrestris will insure a

more accurate estimation of those parameters. In this work the PEIE method will be

carried out in both Arrhenius kinetics and Thermal Death Time (TDT) kinetics using

process lethality (F-value), which is more familiar to food scientists.

The PEIE Method in Arrhenius Kinetics

Recall that a first-order rate process is described by

dC
d= kC (3-1)
dt








where C is the concentration of a reactant at a time, t, and k is the rate constant of the

reaction. Solving Equation 3-1 by integration yields Equation 3-2.


ln C2 -k (t -to (3-2)
The temperature dependency of the rate constant, k, is described by Equation 3-3, the

Arrhenius equation,

<-E 1 1; (3-3)
k = kR .exp JE (3-3)

where kR is the rate constant at reference temperature TR, Ea is the activation energy, T is

the desired operating temperature, and R is the ideal gas law constant. Under isothermal

conditions, the Arrhenius equation that describes the behavior of a reactant that follows a

first-order reaction process is shown in Equation 3-4.


In \ C =-kR exP{ (t -t) (3-4)
Co R -T
This equation can be used to determine the extent of reaction for a given constituent at a

constant temperature. Under non-isothermal conditions, Equation 3-4 is integrated as

shown in Equation 3-5 using the temperature history, T(t), to give the extent of reaction.

Combining Equations 3-4 and 3-5 to equate a dynamic process to an


ln = -R.exp -E /t (3-5)
Co t R-T(t)
isothermal and normalizing to eliminate kR introduces a new factor called G, the decimal

reduction factor, and the equation becomes