The microbial ecology of a rum production process

Victoria Green

A thesis submitted as a fulfilment for the degree of

Doctor of Philosophy

University of New South Wales

School of Chemical Sciences and Engineering
Sydney, Australia

March 2015
This thesis is, lovingly, dedicated to

Graham Harold Fleet

“Fluffy”

Supervisor

Mentor

Friend
PLEASE TYPE
THE UNIVERSITY OF NEW SOUTH WALES
Thesis/Dissertation Sheet

Surname or Family name: Green

First name: Victoria Other name/s: Irene

Abbreviation for degree as given in the University calendar: PhD

School: Chemical Engineering Faculty: Engineering

Title: The microbial ecology of a rum production process

Abstract 350 words maximum: (PLEASE TYPE)

Rum is an alcoholic beverage made from the distillate of a microbial fermentation of sugar cane molasses. Although
commercial rum production started in the 16th century, the microbial ecology of the process has remained relatively
unexplored. This thesis reports an investigation of the microorganisms associated with rum production at a distillery in
Queensland, Australia. Samples of raw materials (molasses, dunder, water, additives), starter cultures and fermenting
molasses were systematically examined for the populations and species of yeasts and bacteria.
2
Molasses contained low populations(< 10 CFU/mL) of yeasts, Bacillus species and lactic acid bacteria. Saccharomyces
cerevisiae, used as a starter culture, was the main yeast of molasses fermentation, growing to populations of about 107
CFU/mL. Lactic acid bacteria were consistently isolated from the molasses fermentation and reached populations of 107
CFU/mL. The main species isolated were Lactobacillus fermentum, Lactobacillus plantarum, Lactobacillus brevis and an
unidentifiable Lactobacillus spp. These species were indigenous contaminants within the processing environment,
colon ising sites that escaped effective cleaning and sanitation operations. Species of Clostridium, Zymomonas and
Propionibacterium were not detected in the production system. Dunder, which originated from the distillation operation,
was considered to be sterile, but developed a population of lactic acid bacteria (the unidentifiable Lactobacillus spp.) on
storage. Dunder had significant concentrations of organic acids and amino acids. At concentrations of 10% and above, it
significantly inhibited the growth of S. cerevisiae in molasses medium and to a lesser extent lactic acid bacteria.
Laboratory scale molasses fermentations and distillations were performed to investigate the effect of lactic acid bacteria
on the growth of S. cerevisiae, process efficiency and production of flavour volatiles. Both single and mixed cultures
using S. cerevisiae, L. fermentum, L. plantarum and Lactobacillus spp. were undertaken. The bacteria did not restrict the
growth of S. cerevisiae but enhanced utilization of molasses sugars and ethanol production. The work presented in this
thesis is the first comprehensive and systematic study, of its type, into the microbial ecology of a rum distillery.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in
part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all
property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral
theses only).

Signature Witness

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for
restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional
circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

ORIGINALITY STATEMENT

'I hereby declare that this submission is my own work and to the best of my
knowledge it contains no materials previously published or written by another
person, or substantial proportions of material which have been accepted for the
award of any other degree or diploma at UNSW or any other educational
institution, except where due acknowledgement is made in the thesis. Any
contribution made to the research by others, with whom I have worked at
UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that
the intellectual content of this thesis is the product of my own work, except to
the extent that assistance from others in the project's design and conception or
in style, presentation and linguistic expression is acknowledged.'

Signed

Date
 ii

ACKNOWLEDGEMENTS

First, and foremost, I would like to extend my gratitude and sincere appreciation to
my supervisor, Professor Graham Fleet. He has generously given his guidance,
expertise, time and mentorship consistently over the last 9 years. I would also like to
acknowledge my co-supervisor Dr Graeme Bushell for his advice and technical
expertise.

I am very grateful to the Australian Research Council (ARC) and The Bundaberg
Distilling Company for funding this research.

The technical staff of the Department of Food Sciences and Technology, specifically
Camillo Taraborrelli, for his assistance, support and friendship. My thanks to Paul
and Phil in the workshop for assisting me with my construction requirements. The
administration staff, particularly Ling, Ann, Sandra, John and Rahul, of the School
Chemical Engineering for their help and assistance over the years.

Undertaking a PhD, both performing the laboratory work, and writing up, can be an
isolating, exhausting and mind-numbing experience. I am extraordinarily fortunate to
have had many people to lean on during the last 9 years.

My fellow PhD students, both from UNSW and other universities, gave me the
knowledge that anything and everything is an acceptable feeling during the course of
a PhD. From feelings of excitement and motivation to those of fatigue, hopelessness
and near surrender, having those people, who have experienced it, by your side is
priceless. To my PhD friends, Hugh, Jeremy, James, Matthew, Arjuna, Ai Lin and
Alex, thank you for keeping me sane, giving me your advice and encouragement.

To my surrogate Sydney family, the UNSW Hockey Club, thank you for being there. I
left my home, knowing no-one in Sydney. I left Sydney feeling as though I had lived
there for most of my life. Whether it was training, playing, coaching or socialising,
there is definitely something special about the Black and Gold.

My thanks to everyone I’ve worked with over the years while juggling part time/full
time work and study. Thank you for your patience and understanding, your support
and your laughter. I’ve had the pleasure of working with some amazing people at 3M
and ACTGAL.
 iii

Heartfelt thanks to my family, extended family and friends. I am grateful for the
encouragement, friendship, care and love you have all shown me over the years. I
treasure our times together and look forward to many more now that I have, those
things called, weekends back!

My dearest friend Elmo, no matter where you have been on this Earth physically,
you’ve always been next to me in spirit. Thank you for having faith.

To my husband, Matt; your support and love through all that has happened these
last 7 years knows no limits. Words cannot express my thanks.

Finally, Mum and Dad, thank you for your continued love. Mum, your letters were a
great comfort when I moved to Sydney. Dad, who would have thought reading the
newspaper would lead to this?

Victoria Green (2015)

 iv

LIST OF TABLES AND FIGURES

Title Page
Number
Table 2.1 Rum production statistics for 2011 by leading brand names 14
Figure 2.1 Outline of the process of rum production from molasses 15
Table 2.2 Chemical composition of blackstrap molasses 18
Table 2.3 Dunder definitions used for rum production 23
Table 2.4 Chemical properties of dunder (distillery wastewater, 25
molasses fermentations)
Figure 2.2 Pot stills used in the production of distilled spirits 32
Figure 2.3 Distillation at the Bundaberg Distilling Company 34
Figure 2.4 Interactions occurring during maturation of raw rum spirits in 36
an oak barrel
Table 2.5 HACCP principles 40
Table 2.6 Important flavour volatiles in rum 45
Table 2.7 Microorganisms found in raw materials, such as molasses 48
and sugar cane juice, associated with alcoholic fermentations
for rum production
Table 2.8 Microbial species associated with the fermentation of 52
molasses or sugar cane juice for rum or rum agricole
production
Table 2.9 Properties required by rum yeasts to be commercially 59
successful as a starter culture
Table 3.1 Sampling sites for microbiological study of a rum distillery 66
Figure 3.1 Flow diagram of rum production process showing sites of 72
sampling for microbiological analysis as given in Table 3.1
Figure 3.2 Process steps at the Bundaberg Distilling Company 73
Figure 3.3 Raw materials used in the production of rum 75
Figure 3.4 Yeast propagation steps undertake at the Bundaberg distillery 76
Table 3.2 Morphology and sequence identification of yeasts associated 81
with rum production
Figure 3.5 Scanning electron microscope image of Saccharomyces 82
cerevisiae isolated consistently from the rum distillery
Table 3.3 Occurance of yeast species in samples of molasses taken at 83
the Bundaberg distillery
Table 3.4 Populations of yeast species throughout the rum production 84
process; conducted in 2008
Figure 3.6 Yeast populations during production of rum as determined by 86
culture on MEA & WLA
Table 3.5 Supporting yeast population data from mini-ecological survey 87
performed in 2010
Table 3.6 Identification of bacterial isolates from Bundaberg distillery 88
Figure 3.7 Lactobacillus plantarum isolated from in-process samples of 89
rum distillery fermentations
Figure 3.8 Lactobacillus fermentum isolated from rum fermentations 90
Table 3.7 Occurrence of bacterial species in samples of molasses 92
taken at the Bundaberg distillery
Table 3.8 Populations of bacterial species throughout the rum 94
production process; conducted in 2008
Figure 3.9 Changes in the populations of individual bacterial species 96
during fermentation of molasses
Table 3.9 Supporting bacterial population data from mini ecological 97
survey (2010)
 v

Title Page
Number
Figure 3.10 Flow diagram of production process showing sites where 110
lactic acid bacteria were isolated
Table 4.1 Populations of bacteria (CFU/mL) in rum fermentation buffer 118
tanks as determined by culture on WLS, Raka Ray and MRS
agar plates
Table 4.2 Prevalence of species of lactic acid bacteria in rum 119
fermentations
Table 4.3 Frequency of isolation of different species of lactic acid 120
bacteria from rum fermentations on three different media
Figure 4.1 Populations of Lactobacillus fermentum, Lactobacillus 121
plantarum, Lactobacillus brevis and Lactobacillus spp in
samples of fermented molasses
Table 5.1 Yeast and bacterial counts of dunder sourced directly from 139
the rum distillery and analysed by direct plating and
enrichment
Figure 5.1 Micrographs of dunder examined under phase contrast and 140
eipifluorescence microscopy
Table 5.2 Yeast and bacterial counts of dunder aseptically sampled 141
from the rum distillery and examined after transport to UNSW
– approximately 4-7 days
Table 5.3 Morphology, rDNA sequencying and API identification of 142
bacteria isolated from dunder
Figure 5.2 Bacterial isolate from dunder sample 143
Table 5.4 Carbohydrate fermentation reactions of dunder isolates as 144
determined by API 50CHL analyses
Figure 5.3 Effect of ethanol concentration on the growth of a 146
Lactobacillus spp. isolated from heated dunder
Figure 5.4 Effect of temperature on the growth of Lactobacillus spp. 147
isolated from dunder
Figure 5.5 Growth of Lactobacillus spp. in MRS broth 148
Figure 5.6 Populations of Saccharomyces cerevisiae, Lactobacillus 150
plantarum, Lactobacillus fermentum and Lactobacillus spp. in
molasses medium containing dunder at different
concentrations
Table 5.5 Effect of dunder concentration in molasses on the growth 151
properties of Saccharomyces cerevisiae and several species
of Lactobacillus
Table 5.6 Chemical properties of dunder samples taken at the 153
Bundaberg distillery
Table 5.7 Comparison of free amino acid concentrations in fresh and 155
stored dunder
Table 5.8 Volatile compounds identified from stored dunder 156
Table 6.1 Laboratory fermentations of molasses with inoculated yeast 165
and lactic acid bacteria
Figure 6.1 Components of the batch distillation column 167
Figure 6.2 Secondary distillation apparatus used for distillation of 169
primary distillates
Figure 6.3 Growth profiles and changes in concentrations of fructose, 174
glucose, sucrose and ethanol during fermentation of
molasses medium
Figure 6.4 Growth profiles of mixed culture fermentations of molasses 177
medium
Figure 6.5 Fermentation of molasses medium by individual microbial 178
 vi

Title Page
Number
species
Figure 6.6 Fermentation of molasses medium by mixed microbial culture 179
Table 6.2 Some volatiles produced during molasses medium 181
fermentation by single and mixed cultures of S. cerevisiae
and lactic acid bacteria
Figure 6.7 Key alcohols produced during molasses medium 182
fermentations by single and mixed cultures
Table 6.3 Higher alcohols detected in the distillates of laboratory 183
fermentations of molasses medium with combinations of
Saccharomyces cerevisiae and species of lactic acid bacteria
Figure 6.8 Selected volatile compounds detected in laboratory scale 184
distillations of molasses medium fermented using differing
starter culture combinations
Table 6.4 Esters detected in the distillates of laboratory fermentations of 186
molasses medium with combinations of Saccharomyces
cerevisiae and species of lactic acid bacteria
Table 6.5 Other acetates detected in the distillates of laboratory 187
fermentations of molasses medium with combinations of
Saccharomyces cerevisiae and species of lactic acid bacteria
 v

TABLE OF CONTENTS
DECLARATION i
ACKNOWLEDGEMENTS ii
LIST OF FIGURES AND TABLES iv
TABLE OF CONTENTS v
ABSTRACT 1

CHAPTER 1 INTRODUCTION 2

CHAPTER 2 LITERATURE REVIEW 6

2.1 Background 6
2.1.1 Origins of the Rum Industry 8
[Link] The Rum Styles 10
[Link] History in Australia 11
[Link] Rum Production Statistics 13
2.2 The Process of Rum Production 14
2.2.1 Production 16
[Link] Physical and Chemical Properties of
17
Molasses
[Link] Molasses Properties and Microbial Growth 19
[Link] Preparation of Molasses for Use in
21
Fermentation
[Link] Dilution Waters 22
[Link] Dunder 23
2.2.2 Preparation of Fermentation Media 26
[Link] Yeast Inoculum for Fermentation 26
[Link] Fermentation 27
2.2.3 Distillation of Ferment 29
[Link] Batch distillation 31
[Link] Continuous distillation 33
2.2.4 Ageing and Maturation 35
2.2.5 Cachaça 38
2.2.6 Quality Assurance and Control in Rum Production 39
2.3 Chemistry of Rum Flavour and Quality 41
2.4 The Role of Microorganisms in Rum Fermentations 46
2.4.1 Microbial Ecology of Molasses 47
2.4.2 Microbial Ecology of Molasses Fermentation for Rum
51
Production
[Link] Yeasts 52
[Link] Bacteria 54
2.4.3 Microbiology of Dunder 56
2.4.4 Use of Starter Cultures in Molasses Fermentation for
56
Rum Production
 vi

2.4.5 Microbiology of Other Distilled Beverages; Cachaça and

59
Whisky
[Link] Microbiology of Cachaça Fermentation 59
[Link] Microbiology of Whisky Fermentation 61
2.5 Summary 63

CHAPTER 3 THE MICROBIAL ECOLOGY OF A PROCESS FOR

RUM 64
PRODUCTION
3.1 Introduction 64
3.2 Materials and Methods 65
3.2.1 Process Outline and Sample Selection 65
3.2.2 Isolation, Enumeration and Identification of
66
Microorganisms
[Link] Yeasts 66
[Link] Bacteria 68
3.2.3 Electron Microscopic Examination 71
3.3 Results 72
3.3.1 Outline of Rum Production Process 72
[Link] Molasses and Molasses Preparation 74
[Link] Preparation of Yeast 75
[Link] Fermentation 77
[Link] Buffer Tanks 77
[Link] Distillation, Maturation and Packaging 78
[Link] Cleaning and Sanitation 79
3.3.2 Ecological Investigation of Yeast and Bacteria
80
throughout the Rum Production Process
[Link] Identification of Yeasts 81
[Link] Identification of Bacteria 87
3.4 Discussion 98

CHAPTER 4 OCCURRENCE OF BACTERIA IN THE

FERMENTATION 113
OF MOLASSES FOR RUM PRODUCTION

4.1 Introduction 113

4.2 Materials and Methods 114
4.2.1 Samples from Buffer Tanks 114
4.2.2 Analysis of Bacteria 115
[Link] Identification of Bacterial Isolates 116
4.3 Results 116
4.3.1 Populations of Bacteria in Rum Fermentations 116
4.3.2 Bacterial Species associated with Rum Fermentations 118
4.4 Discussion 122
 vii

CHAPTER 5 THE MICROBIAL ECOLOGY AND COMPOSITION OF

127
DUNDER USED IN RUM PRODUCTION

5.1 Introduction 127

5.2 Materials and Methods 128
5.2.1 Dunder Samples 128
5.2.2 Microbiological Analysis of Dunder 129
5.2.3 Identification of Bacterial Isolates 130
[Link] Phenotypic Tests 130
[Link] Extraction and Sequencing of ribosomal
131
DNA
5.2.4 Microscopic Examination 131
[Link] Dunder Samples 131
[Link] Bacterial Isolates 132
5.2.5 Isolation of Bacteria from Stored and Heated Dunder 132
[Link] Characterisation of Bacteria Isolates from
132
Heated Dunder
[Link] Temperature Tolerance 133
[Link] Ethanol Tolerance 133
5.2.6 Effect of Dunder on the Growth of Yeasts and Bacteria
134
in Molasses
5.2.7 Chemical Composition of Dunder 134
[Link] Sample Preparation 135
[Link] pH 135
[Link] Organic acids 135
[Link] Sugars 136
[Link] Amino acids 136
[Link] Ethanol 137
[Link] Volatile Compounds 137
5.3 Results 139
5.3.1 Yeast and Bacterial Populations in Dunder Samples 139
5.3.2 Identification of Bacteria Isolated from Dunder 141
5.3.3 Isolation and Identification of Bacterial Species from
143
Heated Dunder
5.3.4 Growth of the Lactobacillus spp. Isolate from Heated
Dunder in the Presence of Ethanol and at Different 145
Temperatures
5.3.5 Effect of dunder on the growth of Saccharomyces
cerevisiae 149
and Lactobacillus species in molasses medium
5.3.6 Chemical Composition of Dunder 152
[Link] Organic Acids 153
[Link] Amino Acids 154
[Link] Volatile Compounds 155
5.4 Discussion 156
 viii

CHAPTER 6 GROWTH OF YEASTS AND BACTERIA

DURING FERMENTATION OF MOLASSES FOR RUM 163
PRODUCTION

6.1 Introduction 163

6.2 Materials and Methods 164
6.2.1 Species Selection 164
6.2.2 Molasses Medium and Fermentation 164
6.2.3 Distillation of Fermented Molasses 165
[Link] Primary Distillation 166
[Link] Secondary Distillation 169
6.2.4 Microbiological Analyses 170
6.2.5 Chemical and Physical Analyses 170
[Link] Sample Preparation 170
[Link] pH and Brix 170
[Link] Sugars, organic acids and ethanol 171
[Link] Volatile Compounds 172
6.3 Results 173
6.3.1 Growth of Individual Cultures of S. cerevisiae and
173
species of Lactic Acid Bacteria
6.3.2 Growth of Mixed Cultures of S. cerevisiae and species
of 176
Lactic Acid Bacteria
[Link] Organic acids 177
[Link] Volatile compounds 180
6.4 Discussion 188

CHAPTER 7 CONCLUSIONS 194

CHAPTER 8 BIBLIOGRAPHY 198

CHAPTER 9 APPENDICES 211

 1

ABSTRACT

Rum is an alcoholic beverage made from the distillate of a microbial fermentation of

sugar cane molasses. Although commercial rum production started in the 16 th
century, the microbial ecology of the process has remained relatively unexplored.
This thesis reports an investigation of the microorganisms associated with rum
production at a distillery in Queensland, Australia. Samples of raw materials
(molasses, dunder, water, additives), starter cultures and fermenting molasses were
systematically examined for the populations and species of yeasts and bacteria.

Molasses contained low populations (< 102 CFU/mL) of yeasts, Bacillus species and
lactic acid bacteria. Saccharomyces cerevisiae, used as a starter culture, was the
main yeast of molasses fermentation, growing to populations of about 107 CFU/mL.
Lactic acid bacteria were consistently isolated from the molasses fermentation and
reached populations of 107 CFU/mL. The main species isolated were Lactobacillus
fermentum, Lactobacillus plantarum, Lactobacillus brevis and an unidentifiable
Lactobacillus spp. These species were indigenous contaminants within the
processing environment, colonising sites that escaped effective cleaning and
sanitation operations. Species of Clostridium, Zymomonas and Propionibacterium
were not detected in the production system. Dunder, which originated from the
distillation operation, was considered to be sterile, but developed a population of
lactic acid bacteria (the unidentifiable Lactobacillus spp.) on storage. Dunder had
significant concentrations of organic acids and amino acids. At concentrations of
10% and above, it significantly inhibited the growth of S. cerevisiae in molasses
medium and to a lesser extent lactic acid bacteria. Laboratory scale molasses
fermentations and distillations were performed to investigate the effect of lactic acid
bacteria on the growth of S. cerevisiae, process efficiency and production of flavour
volatiles. Both single and mixed cultures using S. cerevisiae, L. fermentum, L.
plantarum and Lactobacillus spp. were undertaken. The bacteria did not restrict the
growth of S. cerevisiae but enhanced utilization of molasses sugars and ethanol
production. The work presented in this thesis is the first comprehensive and
systematic study, of its type, into the microbial ecology of a rum distillery.
 2

CHAPTER 1 - INTRODUCTION

Rum is a distilled alcoholic beverage made from fermented sugar cane juice, sugar
cane syrup or molasses (Nicol, 2003; Piggott, 2009). Most rums are produced from
sugar cane molasses, although an allied product, cachaça (Faria, et al 2003; Rosa,
et al, 2009), is produced in Brazil from sugar cane juice. This thesis is focused on
rum produced from molasses.

Rum production evolved from the cane sugar industry, and has been produced in
countries of the Caribbean and West Indies region since the 16th century. It is now
produced in many other countries, including Australia. It is a significant commodity in
international trade and contributes substantially to the export economies of
numerous developed and developing countries. Rum represented about 8% of the
global distilled beverage industry in 2011, with production of about 1.2 billion litres
annually valued at about $US2.1 billion (Datamonitor, 2011; Collicutt, 2009a).

The basic process for rum production consists of the following operations:
preparation of the raw material (molasses, sugar syrup or sugar cane juice);
fermentation of this material; distillation of the fermented product; collection of the
distillate; maturation of the distillate in wooden barrels; and packaging of the final
product. Detailed descriptions of the process can be found in Lehtonen and
Suomalainen (1977) and Nicol (2003), and more general overviews are given in
I’Anson (1971) and Kampen (1975).

The scientific basis of rum production has been investigated and described in the
literature since the early 1900s (Greig, 1885; Pairault, 1903; Allan, 1906; Ashby,
1909; Arroyo, 1945a). There has been little advance in understanding the biology of
the process since that time, with particular regard to the microbiology and
biotechnology involved. This contrasts with other alcoholic beverages such as beer
 3

(Campbell, 2003), wine (Parish & Fleet, 2013) and whisky (Walker, 2012) where
major advances have been made in understanding the basic biology and, now,
advanced molecular biology of these processes. Current knowledge and
understanding about the microbiology and biotechnology of rum production from
molasses has been reviewed by Lehtonen and Suomalainen (1977), Fahrasmane
and Ganou-Parfait (1998) and by Fleet and Green (2010) as part of this thesis.

Microbial fermentation is a key biotechnological process in the production of rum.

The microorganisms that grow throughout this process have a major influence on the
flavour and quality of the final product, and the efficiency of the overall process
(Lehtonen & Suomalainen, 1977; Ganou-Parfait et al., 1989; Nicol 2003). Until
relatively recently, the fermentation of molasses for rum production was conducted
as a traditional, spontaneous process through the growth of indigenous microflora
(Fahrasmane & Ganou-Parfait, 1998). Yeasts were predominant in the process as
agents of alcoholic fermentation with Saccharomyces cerevisiae being the main
species isolated from these fermentations, but Schizosaccharomyces pombe was
also found in some cases. Eventually, strains of S. cerevisiae were developed for
use as starter cultures to conduct these rum fermentations (Arroyo, 1945a;
Fahrasmane & Ganou-Parfait, 1998). Despite their prominent role, very little
research has been done to understand the kinetics of growth of yeasts during
molasses based rum fermentations and to understand how they impact on rum
flavour.

Early literature (Allan, 1906; Ashby, 1907; Hall et al., 1935) as well as more recent
literature (Ganou-Parfait, Fahrasmane & Parfait, 1987; Fahrasmane & Ganou-
Parfait, 1998) also reported the association of bacteria with rum fermentations and
their contribution of distinctive flavour characteristics to some products. In this
context, species of Clostridium, Zymomonas, Bacillus, propionic acid bacteria and
lactic acid bacteria have been mentioned. However, there appears to be no detailed
studies of their growth during fermentation, how they interact with the growth of
yeasts and how they impact on rum quality.

Throughout the literature, dunder is mentioned as a unique raw material used in rum
production (Wustenfeld & Haeseler, 1953; I’Anson, 1971; Kampen, 1975; Wilkie et
 4

al, 2000). Dunder is the liquid residue, depleted of volatile compounds, remaining in
the bottom of the still at the end of distillation of the fermented molasses. It is added
in various proportions back to fresh molasses medium for subsequent rum
fermentation. Frequently, the product is added after it has been stored and
developed an indigenous microbial flora. In these circumstances, it is thought to
contribute flavour enhancing bacteria and yeasts to the fermentation process
(I’Anson, 1971; Murtagh, 1999; Broom, 2003; Nicol, 2003). Despite the widespread
use of dunder in rum production, very little research has been conducted to
understand its microbiology or chemical composition.

The research project described in this thesis originated from an enquiry by The
Bundaberg Distilling Company located in Bundaberg, Queensland, Australia.
Production of rum first started at The Bundaberg Distilling Company in 1888 and it
has continued to the present date. The company produces approximately 5 million
litres of rum annually, valued at about $200 million dollars and accounts for 95% of
the dark rum market share in Australia. In 2000, it was purchased by the
multinational, alcoholic beverage company, Diageo plc.

The process for rum production at this distillery has evolved and developed
somewhat empirically over the years, and the final rum product has a unique or
distinctive flavour character in relation to rums, globally (Broom 2003) .The company
is aware of this quality trait, but is uncertain as to what factors might lead to this
property. Future development and expansion of rum production at this distillery
requires a more thorough understanding of the science and technology of the overall
operation, so that systems can be improved or developed to better manage
production efficiency and product quality. As a basis for further development of its
business, the Bundaberg Distilling Company has sought a more detailed
understanding of the microbiology and biotechnology of its process. It is with this
background that the microbiological investigations reported in this thesis were
undertaken.

The overall objective of this thesis is to systematically investigate the microbial

ecology of the rum distillery located at Bundaberg, Queensland, Australia - from raw
materials (molasses, dunder, yeast and water) through yeast propagation and
 5

fermentation. This basic information about the microbiology of the process should
provide platform knowledge which the company can use to optimize its process and
product quality.

Given the limited information that exists about the microbiology and biotechnology of
molasses based rum fermentation, in general, it is expected the results of the thesis
will have broader scientific and technological interest.

Chapter 2 of this thesis gives some background literature and information regarding
the history and production of rum, the microbiology of rum production and the
chemical composition of rum.
Chapter 3 reports a systematic investigation of the microbial ecology of the rum
production process at the Bundaberg distillery
Chapter 4 presents a more detailed investigation of the association of lactic acid
bacteria with the rum production process
Chapter 5 reports the microbiology and chemistry of dunder and examines its effect
on yeast and bacterial growth during molasses fermentation.
Chapter 6 combines the significant microbiological findings of Chapters 3, 4 and 5 to
conduct laboratory scale fermentations and distillations to determine the effects of
various controlled microbial combinations on molasses rum fermentation.
Chapter 7 summarises important conclusions of the project with recommendations
for further research and development.
 6

CHAPTER 2 LITERATURE REVIEW

2.1 BACKGROUND

Rum is a distilled alcoholic beverage made from sugar cane juice, sugar cane
syrup or molasses (Nicol, 2003; Piggott et al, 1995). It is often matured in barrels or
vats. Many countries have defined rum within their food and beverage legislations
and examples for Australia, the USA and the European Union are given in below.

In Australia, rum is defined as;

‘a potable alcoholic distillate…which, unless otherwise required by this

Standard, contains at least 37% alcohol by volume, produced by
distillation of fermented liquor derived from food sources, so as to
have the taste, aroma and other characteristics generally attributable
to [rum]’

(FSANZ, 2000)

In the USA, rum is;

‘an alcoholic distillate from the fermented juice of sugar cane, sugar cane
syrup, sugar cane molasses, or other sugar cane by-products, produced at
less than 190° proof in such manner that the distillate possesses the taste,
aroma and characteristics generally attributed to rum, and bottled at not less
than 80° proof; and also includes mixtures solely of such distillates.’

(USA Government, 2010)

European (EU) definition;

‘Rum
(1) A spirit drink produced exclusively by alcoholic fermentation and
distillation, either from molasses or syrup produced in the manufacture of
cane sugar or from sugar cane juice itself, and distilled at less than 96%
vol., so that the distillate has the discernible specific organoleptic
 7

characteristics of rum
(2) The spirit produced exclusively by alcoholic fermentation and distillation
of sugar cane juice, which has the aromatic characteristics specific to
rum, and a content of volatile substances equal to or exceeding 225 g hl-
1
of alcohol of 100% vol. (2250ppm). This spirit may be marketed with
the word ‘agricultural’ qualifying the designation ‘rum’ accompanied by
any of the geographical designation of the French Overseas
Departments as listed in Annex II
(3) Bottled at a minimum alcoholic strength of 37.5% v/v’

(The Council of European Communities (EEC), 1989)

The origin of the word “rum” is uncertain and has been attributed to a number of
sources (Clutton, 1974; Lehtonen & Suomalainen, 1977; Broom, 2003; Coulombe,
2004). There have been suggestions that the word is derived from the last part of
the word “Saccharum” which is the taxonomic name for a sugar cane variety.
Another derivation is from rumbullion, a British slang word meaning “great tumult or
uproar”. Possibilities also include; the Spanish word for ‘ron’, the French word for
aroma, ‘arôme’, and a Dutch word for drinking glasses, “roemer”, that sailors
colloquially called “rummers”. Despite the uncertain origin of the word, it was
documented in 1654 within a law that was passed in Connecticut confiscating
“Barbados liquors, commonly called Rum, Killdevil, or the like.” As this law showed,
not only was the origin of the word ambiguous but there were multiple names by
which the rum beverage was known, including kill-devil, brebaje, ron, Nelson’s
Blood, grog, rumbullion, taffia, guildhive, Demon Water, Pirate’s Drink, Barbados
water and Navy Neaters. The word was also used in an order penned by the
Governor-in-Council of Jamaica regarding rum and its consumption on 3rd of July
1661 (Thomson, 1885).

Rum is the product of a microbiological fermentation and this thesis is focused on

the microbiology of that process. The literature review presented in this section
provides some background information on the history of the rum industry and the
process of rum production. This is followed by a compilation and critical evaluation
of the research conducted on the microbiology of the process and how this
microbiology impacts on rum quality and production efficiency. The gaps in
 8

knowledge and needs of further research are identified and discussed. In contrast
to many other alcoholic beverages, such as beer, wine, and whisky, there are only
a few general discussions or reviews on the science and technology of rum
production and these include Arroyo (1945a), Fahrasmane & Ganou-Parfait (1998),
Lehtonen & Suomalainen (1977), Nicol (2003) and Fleet & Green (2010).

2.1.1 Origins of the Rum Industry

Discussions on the history of rum production have been given by Clutton (1974),
Lehtonen & Suomalainen (1977), Nicol (2003), and Broom (2003). The origin of
rum can be traced back to around 2000BC from obscure references in the sacred
Hindu texts known as “The Vedas”, which refer to two beverages made from sugar
cane by-products, one from molasses (gandi) and the other from cane juice
(sidhu). From that time, over 3500 years ago, until the present day, rum production
has been intertwined with the development and expansion of the sugar industry
(Nicol, 2003; Coulombe, 2004).

Originally grown in Asia, specifically the East Indies (Indonesia and Papua New
Guinea), sugarcane was taken west by trading between the Chinese and Arabs.
Grown in Northern Africa by Arabs, the “sweet spice” was introduced to the
western European world during the Crusades of the 11th century. Sugar was
reportedly first available in England in the late 11th century. Imported to Europe
from the East at great cost, there was an increasing need for the European
countries to instigate production of their own supplies. By the 15th century, the
Portuguese had colonised Madeira where sugarcane was successfully grown. In
1492, Columbus sailed to the Caribbean region with sugarcane cuttings, hopeful of
starting a sugar industry in the Americas (Nicol, 2003). The cultivation of the
cuttings was extremely successful and, on his second voyage, Columbus, bound
for Cuba and Hispaniola, took more cuttings. Sugar cane cultivation quickly spread
throughout the Caribbean region, including locations such as Jamaica, Martinique,
Puerto Rico and Cuba, Barbados, Trinidad, Haiti, Guadeloupe, the Virgin Islands,
 9

the Dominican Republic, Guyana and Brazil. Slaves from Africa were introduced to
work on the plantations and harvest the sugarcane (Broom, 2003; Nicol, 2003;
Coulombe, 2004).

Once harvested, the sugarcane was crushed and the resulting juice collected and
boiled. After boiling, the liquid was allowed to cool, and crystalline sugar formed in
lumps which were extracted, leaving behind a thick, dark syrup-like juice termed
‘melazas’ or molasses, in English. This waste product was also sold as a
sweetener because the demand for sugar outweighed its production in the 16th
century. Any molasses that was unsold was usually set aside, frequently with the
addition of water, reboiled and left, whereupon it underwent a natural fermentation.
An alcoholic spirit was derived from the ferment by distillation and generally, was of
dark colouration and relatively unpleasant in flavour. Nevertheless, it appealed to
those such as the African slaves, who could not afford finer liquors and its sale
returned a healthy profit to plantation owners. By the end of the 19th century, most
plantations that could afford distillation equipment were producing liquor, and
approximately 200 000 gallons (757 082 litres) per year was being produced on the
island of Barbados alone. The distilled spirit was then, either consumed locally or,
purchased by naval ships that frequented the seas nearby. The presence of these
naval ships discouraged intervention by pirates and lead to a long association
between the rum producers and naval companies, with the eventual export of rum
to other countries, including Europe, by the end of the 17th century (Bluhm, 1983;
Broom, 2003).

By the end of the 17th century the Caribbean region had become the world centre
for rum production, and a thriving industry had been established. Rum had now
become a global commodity. Rum’s popularity drew from its stark contrast to the
more traditional European style spirits such as gin and Scotch whisky. Throughout
the 18th century, Colonial America produced more rum than the entire Caribbean
region and the popularity of the spirit in that country as an alcoholic beverage
continued until the Great Depression of the 1930’s. The Depression caused sugar
 10

and rum sales to dramatically decrease, forcing distilleries to close. The

introduction of American Prohibition, prohibited the manufacture, sale and transport
of alcohol in the United States from January 1920 to December 1933, also saw a
decrease in overall rum production around that time. These circumstances caused
distillers to form alliances to better regulate the sale and price of rum in America
and the Caribbean, and set the foundations for the modern commercialisation of
rum production. This cooperation of groups of distillers was the forefront of the “big
brand” era that has characterized the rum industry for the last one hundred years
(Broom, 2003).

World War II lead to a resurgence in Caribbean rum popularity, mainly due to

American and British distilleries ceasing production of drinking spirits such as gin
and whisky as they switched their activities to production of industrial alcohol for
“the war effort”. Such transitions enhanced the importation of rum from the
Caribbean region. Unfortunately, this popularity did not last long after the war when
consumer tastes shifted from the heavy, sweet drinks towards lighter Scotch
whisky varieties and the newer import of vodka, a subtly flavoured colourless spirit.
This shift in public perception and taste allowed a lighter-bodied, white rum
(Bacardi) to establish itself at the forefront of the rum industry. Bacardi, although
established in 1862 in Cuba, became a household name in the mid 20th century
worldwide. It remains the leading rum brand worldwide (see Table 2.1 in section
[Link]).

[Link] The Rum Styles

From its humble beginnings as plantation waste given to slaves as a form of

payment, rum has matured into a popular choice of distilled beverage worldwide. It
was the first distilled beverage consumed in America and while its popularity
waned for a large part of the 20th century, it has had a revival worldwide, mainly
due to the production of white rum and its use in cocktails. Rum linkage to the
expansion and industrialisation of the sugar industry and the former colonisation of
 11

the Caribbean basin by different European countries has meant that rum
production is a process impacted by history and international influences, leading to
many regional variations. The Spanish colonies Cuba, Panama, the Dominican
Republic and Puerto Rico manufactured subtle, light rums. The islands of
Barbados, Bermuda, Jamaica and other British colonies islands produced a fuller
tasting rum with a darker colour. French colonial islands such as Guadeloupe,
Haiti, and Martinique, produced a very different type of rum (Rhum Agricole)
(Kampen, 1975; Broom, 2003). This rum is generally produced from sugar cane
juice rather than molasses and, consequently, exhibits a flavour closer in
characteristics to the sugar cane and is usually more expensive. Cachaca is a spirit
produced in Brazil from sugar cane juice, similar to rhum agricole, but it is often
triple distilled for a smoother, lighter flavour (I’Anson, 1971; Clutton, 1974).
Consequently, rum is a term that represents various styles of product.

[Link] History in Australia

Sugar cane was first introduced to Australia and Norfolk Island in 1788 but this
tropical grass did not fare well in the southern parts of these new British colonies.
The first successful crop was harvested more than 50 years later, in 1862, by
Captain Louis Hope at a plantation east of Brisbane. With the introduction of
regulations encouraging the production of sugar, it did not take long before sugar
was a major crop (Kerr, 1983). By 1890, sugar cane was grown as far north as
Mackay along the east coast of Queensland, Australia. North Queensland has
since developed into a major region for Australian sugar production, producing a
total of 4.75 million tonnes of sugar annually, valued at $AUD 1.75 billion. Of this,
85% is exported and sold internationally. Australia is the second biggest raw sugar
producer worldwide after Brazil. Other raw sugar producing countries include India,
Thailand, Guatemala, Cuba and South Africa. (Queensland Sugar Limited, 2010).

The rum industry in Australia has an interesting association with the development
of the country since its establishment as a British colony in 1788. In the very early
 12

stages, rum was imported from other colonies of the British Empire. Rum can be
linked to several historical events throughout the colony. One of the most
interesting events occurred during Captain Bligh’s office as governor of England’s
southern penal colony. He was deposed by the New South Wales Corps (the “Rum
Corps”) because of his confrontational approach to several influential colonists who
defied government regulations and engaged in private trading enterprises for
personal profit. The Rum Rebellion occurred on 26 January 1808, with the NSW
Corps marching on Government House and arresting Bligh for being “unfit to
govern”. This was the only time in the history of Australia that any government was
changed by force (Fitzgerald & Hearn 1988; Bligh & Currey, 2003).

Major General Lachlan Macquarie was appointed Governor in 1810. On his arrival,
Macquarie found a hospital consisting of tents and other temporary structures at
The Rocks at Sydney Cove. Macquarie set about earmarking property for a new,
permanent, hospital but the British Government refused to allocate funds for the
building. Macquarie approached a consortium of businessmen for this funding and
in exchange for building the hospital, the businessmen received the exclusive
rights on rum imports (60,000 gallons/ year) to sell to colonists and convict labour.
Macquarie’s hospital was known as “The Rum Hospital”. It still exists today as the
Sydney Hospital and is the oldest hospital in Australia (Lewis, 1992).

In 1869, a man named James Stewart, bought a small boat, converted it to steam
power, and installed a pot still on the deck. The Walrus, as it was called, also had a
small sugar mill for crushing up to two tons of cane a day. Stewart obtained a
distillery licence and, on 14 April 1869, the Pioneer Floating Sugar Company
started moving along rivers to small sugar plantations crushing cane and producing
rum from the molasses. Due to the difficulty of keeping adequate account of the
Pioneer Floating Sugar Company, the licence was not renewed in 1872 by the
authorities. Stewart, however, continued to operate the distillery unofficially until
1883 (Lewis, 1992).
 13

In 1884, the price of sugar fell, however, distilleries were encouraged to continue
operation by the government (mainly due to the considerable tax revenue
applicable on the product). Rum was easier to transport to other parts of Australia,
mainly the east coast, due to the longer shelf life compared to beer and wine. In
1888, production of rum started at a distillery in Bundaberg, Queensland. The
Bundaberg Distilling Company was originally run by the operators of local sugar
mills in the region, concerned with how to deal with waste molasses from sugar
production (Kerr, 1983). The distillery has been closed on two separate occasions
due to fires; however, it continues to run today (Kerr, 1983)

Other distilleries that have produced rum in Australia, either previously or at time of
print are; Beenleigh Distillery (Beenleigh, Queensland), The Lark Distillery (Hobart,
Tasmania) ([Link]), The Hoochery (Ord River, Western
Australia)([Link]), The Kimberley Rum Company (Swan River,
Western Australia)([Link]) and small boutique distilleries nationwide.

[Link] Rum Production Statistics

Alcoholic beverage production worldwide has been estimated to exceed 120 billion
litres in 2008. The total market revenue for 2011 was estimated at US $1009.7
billion (Marketline, 2012). Any estimates are deemed to be on the low side due to
sale statistics being closely guarded within the industry or difficult to calculate for
some countries (Piggott, 2003).

Rum represented 7.7% of the global distilled beverage industry in 2011

(Datamonitor 2011). In 2007, the rum market was valued at $US 2.1 billion in the
USA, accounting for 18% of spirit sales (Drinks International, 2011). North America
accounts for the highest share of global rum sales (55%). Countries in the
Caribbean region remain the major producers of rum, but it is produced in other
countries where sugar cane is grown, such as the Philippines, India, Brazil, Fiji and
Australia. Table 2.1 shows some statistics on rum production from 2011, giving the
 14

brand and company name along with location of production.

Table 2.1 Rum production statistics for 2011 by leading brand names (Euromonitor, 2011)
Brand Country Company Sales
(million
9L cases)a
Bacardi Cuba (headquarters in Bermuda) Bacardi 19.6
Tanduay Philippines Tanduay Distillers 18.7
McDowell’s No.1 Celebration India United Spirits 15.6
Captain Morgan Puerto Rico (American) Diageo PLC 9.2
Havana Club Cuba Pernod Ricard 3.8
Contessa India Radico Khaitan 2.4
Old Cask India United Spirits 2.2
Old Port Rum India Armut Distilleries 2.0
Montilla Brazil Pernod Ricard 1.6
Cacique Venezuela Diageo PLC 1.7
a
9L cases are the common way of measuring volume of sales in the alcohol industry. 9L case
refers to 12 bottles of 750mL.

Australian rum production is dominated by two companies. Of these, the

Bundaberg Distilling Company is the market leader with estimates of $200m net
sales in 2008 (Main, 2010) and a 95% share of the dark rum market (McNicoll,
2006).
Bundaberg Distilling Company’s nearest competitor is Inner Circle (Lion Nathan).
First produced by the Colonial Sugar Refinery (CSR) Company in 1901, Inner
Circle has a worldwide reputation as a high class, quality rum. Production of this
rum previously occurred off shore in Fiji; however, the recent acquisition of the old
Beenleigh distillery, by Lion Nathan Ltd, has seen production return to Australia
([Link]).

2.2 THE PROCESS OF RUM PRODUCTION

The basic process of rum production consists of: preparation of the raw materials
for fermentation; alcoholic fermentation; distillation of the ferment; collection of the
distillate; maturation of the distillate; blending; packaging; and sale (Figure 2.1).
The main raw material can be either molasses or sugar cane juice. Most rum is
produced from molasses and this process will be the focus of this review. Rum
production from sugar cane juice will be briefly described in Section 2.2.1. General
descriptions of the process have been given by Arroyo (1942 & 1945a), Clutton
(1974), Kampen (1975), Lehtonen & Suomalainen (1977), Bluhm (1983) and Nicol
 15

(2003) and these articles can be consulted for greater detail.

Figure 2.1 Outline of the process of rum production from molasses

 16

2.2.1 Production

Molasses is the major by-product of the extraction of sugar from sugar producing
crops. There are two main sources of molasses; cane molasses or sugar beet
molasses. However, by strict definition, only sugar cane molasses can be utilised
to produce rum (Kampen, 1975; Nicol, 2003).

Sugar cane (Saccharum officinarum) is a tall, tropical grass which has stalks that
grow to heights of 2-6 metres The stems contain approximately 15% sugar
(sucrose) solution but, depending on sugar cane cultivar, farming methods and
factors such as climate, up to 50-60% sugar content can be present. Details on the
botany of the sugar cane plant, its cultivation and its processing into sugar are
given in Deerr (1921), Birch et al (1979) and Galloway (1989).

Sugar cane crops, today, are generally harvested by machines, whereas some
years ago this was done manually by workers with machetes. It was often the
practice to burn the sugar cane field just prior to harvesting to remove some of the
fibrous plant material and to sanitize the soil However, this practice is now less
common, and the cane is directly harvested without burning. The sugar cane juice
is extracted from the harvested crops by crushing the stems and pressing into a
dry mass (Nicol, 2003). The juice is boiled to reduce the water content to a
desirable consistency for sugar extraction, leaving ‘first molasses’ after the initial
boiling. ‘Second molasses’ is produced from boiling of the first molasses, where
raw sugar begins to crystallize and separate. Blackstrap molasses is the final
molasses produced from further rounds of boiling and crystal extraction of the
second molasses. Detailed descriptions of sugar cane processing and molasses
production are given in Lehtonen & Suomalainen (1977), Nicol (2003) and James
(2008).

Due to its availability and comparatively low cost, blackstrap is the most commonly
employed molasses in rum production (Lehtonen & Suomalainen 1977). Because
of the additional steps involved in its production, blackstrap molasses has a more
 17

pronounced, slightly bitter, flavour compared to first and second molasses. Some
rum producers may also mix in a small portion of refinery molasses to their
blackstrap stocks. Refinery molasses is produced during the refining of raw sugar
to white sugar, which liberates extra molasses originally bound to the raw sugar
crystals.

Rum distilleries are usually built in close proximity to a sugar mill, either directly
adjacent or only a short distance away. Consequently, molasses is delivered to the
rum distillery storage tanks either by pipe, still hot from the sugar mill, or by road
tankers. Molasses is stored in either large concrete lined wells, stainless steel
tanks or a combination of both. Due to its nature (low water content, high sugar
content, low oxygen, low pH), molasses can be stored for long periods of time (up
to a couple of years) with minimal processing required both prior to storage and
also prior to fermentation. Thus, a supply is available throughout the entire year,
regardless of when the sugar cane crushing season took place. Rum producers will
usually possess multiple storage facilities, used on a rotational schedule, to avoid
the use of ‘fresh’ molasses and also avoid the influences of seasonal variation and
periodic supply (Fahrasmane & Ganou-Parfait, 1998).

Since the early studies of Arroyo (1947) on molasses and rum production, it has
been realized that molasses quality will impact on rum quality and efficiency of the
production process. Quality assurance testing of molasses for its chemical,
physical and microbiological properties is now a routine part of rum production. The
following section describes the key physical and chemical properties of molasses.
The microbiology of molasses will be considered in a later section (Section
[Link]).

[Link] Physical and Chemical Properties of Molasses

Molasses is a black, viscous liquid containing about 55% w/v total fermentable
sugars, of which about 35% is sucrose and 20% is a mixture of glucose and
fructose. In addition to water and these sugars, it contains small amounts of many
 18

other compounds, such as; nitrogenous substances, phosphorus, metal ions,

vitamins, gums and colloidal constituents. Table 2.2 summarizes the composition
of molasses as reported by Bluhm (1983), Curtin (1973), Lehtonen and
Suomalainen (1977), Nicol (2003) and others.

a
Table 2.2 Chemical composition of blackstrap molasses
Properties Amount Properties Amount
Water (%) 10-20 Sulfur (%) 0.5
Brix density 79.5-89 Sodium (%) 0.06-0.2
pH 5.3 Formic acid (%) 0.1
Total Solids (%) 75.0 Acetic acid (%) 0.2
Sucrose (%) 33.8-36.4 Aconitic acid (%) 0.8
Reducing sugar (%) 19.6-65 Lactic acid (%) 0.05
Total Sugar (%) 46.0-60.9 Malic acid (%) detected
Total nitrogen (%) nd-1.5 Citric acid (%) detected
Total Fat (%) 0.0 Biotin (µg/g) 0.36-3.2
Boron, ppm 410 Chlorine (µg/g) 745.0
Iron, ppm 158-249 Pantothenic Acid (µg/g) 20-120
Manganese, ppm 35-57 Pyridoxine (µg/g) 6.5
Copper, ppm 28-36 Riboflavin (µg/g) 1.8-2.5
Zinc, ppm 10-20 Thiamine (µg/g) 0.9-8.3
Nickel, ppm 1 Inositol(µg/g) 6000
Lead, ppm 0.78 Nicotinamide (µg/g) 20-25
Cobalt, ppm 0.54 Folic Acid (µg/g) 0-0.04
Potassium (%) 2.6-5.0 Crude Protein (%) 3.0
Magnesium (%) 0.3-1.0 Gums (%) nd-6.3
Calcium (%) 0.2-2.0 Ash (%) 8.1
Phosphorus (%) 0.08-0.2 Sulfated ash (%) 7-11.57
Chlorine (%) 1.4 Apparent purity 28.67
a
Data obtained from Nelson & Greenleaf (1927), Arroyo (1947), Burrows (1970), Wythes et al
(1978), Bluhm (1983), Curtin (1973), Murtagh (1995 a, b), Nicol (2003), Bortolussi & O’Neil
(2006), Amorim et al. (2009)
nd –not detected
detected – component detected not quantified

As mentioned previously, molasses is rich in fermentable sugars and these are the
main chemical components of this raw material (Table 2.2). Traditionally, distillers
have used °Brix as a measurement of sugar content in molasses and molasses
quality. For production of good quality rum with desirable flavours and ethanol
yield, molasses with a °Brix of 87.6 has been recommended. Lesser quality rums
are obtained from molasses with °Brix less than 85.4 and °Brix greater than 88.2
(Lea & Piggott, 2003). However, the °Brix value does not give an accurate
correlation with total sugar content (Baker 1979; Nicol, 2003) because it is a
 19

measure of soluble solids and molasses contains many soluble solids that are not
sugars. Total fermentable sugar is easily quantified by high performance liquid
chromatography (HPLC) and such measurements would be the basis for better
quality grading of molasses than °Brix (Baker 1979; Nicol, 2003).

Molasses selection is critical in order to obtain efficient fermentations with high

levels of ethanol, and the production of rum with a desirable flavour profile. The
fermentation medium of molasses needs to contain appropriate levels of nitrogen,
phosphorus, vitamins and minerals in addition to fermentable sugars (Walker,
2004).

Molasses has a pH of 5.0 to 5.5 due to the presence of numerous organic acids,
the most prevalent of which are acetic, malic, lactic and citric acids. Total nitrogen
represents no more than 1.5% of molasses and consists of free amino nitrogen
(ammonia and amino acids) and crude protein (about 3%). The gums of molasses
can constitute up to 6% and are represented by hemicelluloses, pectins and
dextrins which are found in sugar cane, and levans that may be produced by
bacteria during the sugar cane milling process (Schooness & Pillay, 2004). Several
vitamins have been found in molasses (Table 2.2), with inositol and pantothenic
acid being the most prevalent.

Many factors affect the composition and quality of molasses, and these include the
soil type, ambient temperature, moisture, season of production, cultivar and
cultivation of sugar cane, the sugar refining process and conditions of molasses
storage (Murtagh, 1995 a, b; Bortolussi and O’Neil, 2006). Thus, variation may be
found in nutrient content, flavour, colour, viscosity and total sugar content.

[Link] Molasses Properties and Microbial Growth

Molasses contains sufficient carbon and other micronutrients such as metal ions
and vitamins to allow microbial growth. However, it has a very high concentration
 20

of soluble sugars (50-60%), low water activity (Aw 0.76), low pH, and is a poor
source of nitrogen and phosphorus. Consequently, it is an extremely stressful
environment for microbial growth and survival (Lehtonen & Suomalainen, 1977;
Kampen, 1975) and only a few, well adapted species will tolerate these conditions
(see 2.4.1).

Upon dilution with water, the stresses of high sugar content and low water activity
are relieved and it becomes more favourable for microbial growth. For rum
production and fermentation by the distiller’s strains of the yeast Saccharomyces
cerevisiae, it is usually diluted to about 15-40 °Brix or 15-30% fermentable sugars
(Lehtonen & Suomalainen, 1977). In some cases, additional supplies of nitrogen,
vitamins and other micronutrients may be added to encourage growth of the
distiller’s yeast (Lehtonen & Suomalainen 1977; Nicol, 2003) .Fermentations
lacking these nutrients tend to be slow, often unable to progress, giving a condition
that is known as becoming “stuck”.

Stress on the yeast cells can also cause them to produce undesirable flavours.
Therefore, it is of vital importance to rum quality and production efficiency to
ensure that the molasses medium contains sufficient levels of nitrogen and
phosphorus (Cacho & Murphy 1988). In order to rectify these deficiencies, rum
producers will often supplement the medium with ammonium phosphate or
ammonium sulphate at levels between 0.03 -0.06% (w/v) (Bluhm, 1983).
Commercial yeast nutrients such as Fermaid A® may also be used
([Link]). It is important to carefully control the addition of these
supplements since excess nitrogen in the medium may inhibit yeast growth and
affect production of flavour metabolites (Walker, 2004).

Molasses contains several substances that could be inhibitory to yeast growth if

they are present at high concentrations. Hydroxymethyl furfural (HMF) is a product
of the overheating of cane molasses during milling at a low pH. HMF levels of 0.4%
can be tolerated by yeasts but higher amounts should be avoided. Several low
 21

molecular weight organic acids can be present in molasses. Acetic acid may be
found at levels of up to about 1%; however, levels of 0.25% or more are known to
be inhibitory to yeasts. Butyric acid is, similarly, inhibitory to yeast at 0.1%-0.5%
and may be found at levels up to 1.0%. Similarly valeric acid is inhibitory to yeast at
0.1%-0.5% and has been found at levels of 0.1% (Lehtonen & Suomalainen,
1977). Consequently, it is necessary to prepare and store molasses under
conditions that prevent the development of these inhibitors.

[Link] Preparation of Molasses for Use in Fermentation

As mentioned previously, rum distillers generally store molasses in bulk quantities

in wells or tanks, and draw from these supplies to prepare it for fermentation. Such
preparation includes clarification, adjustment of pH, heating to inactivate
microorganisms, dilution with water, addition of nutrients for yeast growth and
addition of dunder (Clutton, 1974; Nicol, 2003; Lehtonen & Suomalainen, 1977;
Kampen, 1975). The following sections describe the operations for preparing the
molasses for fermentation.

Immediately prior to use in fermentation, molasses is clarified by a combination of

chemical or physical processes to partially remove suspended solids. This is
usually done by addition of flocculating agents and allowing the solids to sediment.
Centrifugation may also be used to clarify the molasses. At this stage, the pH is
adjusted to values around 5.0-5.5 by addition of sulphuric acid, and the mixture is
given a mild heat pasteurization treatment (80°C). Duration of pasteurisation is
often unrecorded and as such will vary among distilleries. It is important to remove
colloidal material from the molasses, otherwise it will cause severe fouling of the
distillation columns, thereby leading to inefficiency of the stills and production
downtime due to increased frequency of cleaning (Murtagh 1995a).

After settling or centrifugation of the solid materials, the clarified liquid is pumped
off, and then diluted with potable water to give a final concentration of 100-150
 22

g/litre for fermentable sugars (15 to 20 °Brix). Yeast nutrients such as ammonium
sulphate and vitamin mixtures may be added at this time to ensure complete
fermentation (Paturau, 1989; Nicol, 2003). Also, at this stage, dunder is added to
the mix. The addition of dunder is a traditional, unique part of the rum making
process and, in this context, is described in the literature from studies published
since the early 1900s (I’Anson, 1971; Broom, 2003; Nicol, 2003). The role of
dunder and its use in rum production is discussed in more detail in a later section
(Section [Link]). The diluted molasses containing added nutrients and dunder is
then cooled to about 30°C by heat exchange and pumped to large vessels for
fermentation.

[Link] Dilution Waters

Water is used to dilute the molasses for preparation of the fermentation medium
(Arroyo, 1942 & 1945a). Waters used for dilution purposes may originate from
various locations including town water, treated distillery waste waters, rivers and
creeks and rainwater storage tanks. As such, the quality may vary greatly and
microbial testing and chemical analysis should be undertaken. Chemical testing
should determine ion content, heavy metal presence and hardness. Microbiological
testing for total viable count, coliforms and E. coli and Clostridium perfringens
should be done as a minimum (Fahrasmane & Ganou-Parfait, 1998). Detailed
testing procedures and results are difficult to find in the rum literature as water
testing has been deemed more important for dilution of the finished matured spirit
rather than that used for dilution of fermentation medium. Arroyo (1945a)
recognised the importance of using good quality water, during dilution of molasses,
to limit the potential production of off-odours by contaminating microflora during
subsequent fermentations. His research also discussed the potential of increasing
the mineral content of molasses because of its effect on yeast growth and activity.
 23

[Link] Dunder

Dunder is the liquid residue, depleted of volatile compounds, obtained from the
stills after distillation of the fermented molasses (Kampen, 1975). As a result,
dunder is rich in suspended solids, non volatile compounds, particularly non
volatile acids, and non viable microbial biomass. Dunder is either added to the
fermentation medium fresh from the primary distillation columns or after prolonged
storage in dunder pits. Although the literature reports widespread use of dunder in
rum production, there is a diversity of descriptions as to what it actually represents.
Table 2.3 gives a summary of some of the descriptions used in the literature in
reference to dunder.

Table 2.3 Definitions of dunder used for rum production

Author Year Definition
Wustenfeld 1953 Spontaneous, bacterial soured cane molasses slops from
& Haeseler, fermentation plant. It is important for producing characteristic rum
esters
I’Anson 1971 Added to fermented molasses medium. It is the lees of previous
distillations which has been allowed to age and undergo bacterial
fermentation. It is a strong smelling liquid consisting of high acid
and ester content and causes the distillate to have a rich and fruity
aroma. This aroma is characteristic of Jamaican rums.
Kampen 1975 Bottom product of a distillation column; rich in both yeast nutrients
and acids
Murtagh 1995b Old stillage that has been stored in open tanks to allow
development of strong bacterial flora
Fahrasmane 1997 Denoted as “stillage” being the residual liquor from distillation also
& Ganou- known as slops or “spent wash”
Parfait
Wilkie et al 2000 Also termed distillery wastewater, distillery pot ale, distillery slops,
distillery spent wash, dunder. Mosto, vinasse and thin stillage
The aqueous by-product from the distillation of ethanol following
fermentation of carbohydrates.
Broom 2003 The acidic spent lees (the non-alcoholic residue left in the still after
distillation) which has been aged outside in dunder pits to
concentrate the acetic/butyric acids and the ester content
Nicol 2003 Residue of wash distillations and is allowed to ferment naturally in
a “dunder pit”. It is a naturally developed inoculum for
fermentations containing wild yeasts and anaerobic bacteria.
Australian 2006 Sugar cane lees used to promote fermentation of rum. Residue
Government left in the still after distillation and is generally used in the process
of slow fermentation
Melamane, 2007 Also known as distillery wastewater, stillage, distillery pot ale,
Strong & distillery slops, distillery spent wash, mosto, vinasse, thin stillage
Burgess Aqueous by-product of the distillation of ethanol, wine and some
waste biological material
 24

There appears to be no standard procedure for the use of dunder in rum

production. It is mostly added to molasses fermentations at the time molasses is
diluted with water. If dunder is added directly from the stills, it is considered to be
sterile because of its prior heat treatment in the stills. However, it is the practice
with some rum producers to store the dunder until needed in covered wells
(Olbrich, 1963; I’Anson, 1971; Murtagh, 1995b; Broom, 2003). During storage, a
complex microflora of yeast and bacteria are likely to grow within the dunder and,
therefore, will impact on the microbial ecology of the molasses fermentation and,
consequently, rum flavour.

The amount of dunder added to the molasses varies with the rum producer and
can range from 0- 50% (I’Anson, 1971; Fahrasmane & Ganou-Parfait, 1998) but
mostly it is added at about 0-10%. The reasons for its use are also varied and
include; lowering molasses pH, providing an added source of nutrients for
microorganisms that conduct molasses fermentation, recycling part of the process
water to minimise water usage and disposal costs, and providing a source of wild
yeasts and bacteria for molasses fermentation, if stored dunder is used (Olbrich,
1963; I’Anson, 1971).

Lowering the pH is thought to encourage the growth of Schizosaccharomyces

yeasts in comparison to Saccharomyces yeasts during fermentation, and additional
nutrients from the dunder may encourage the growth of bacterial species that could
favourably impact on rum flavours (Kampen, 1975; Faharasmane & Ganou-Parfait,
1998).

Despite the use of dunder in rum production and its potential impact on product
quality and process efficiency, there is little published information on its
microbiological and chemical properties. Essentially, dunder is a microbial (mostly
yeasts) culture that has been heated to 80-110°C for an extended period. Volatile
metabolites would have been driven off, leaving a concentration of non-volatile
metabolites (e.g. non-volatile acids). When examined under the microscope, it
shows masses of yeast and bacterial cells, along with other particulate debris
 25

(Kampen, 1975). Theoretically, the microbial cells should be dead due to the high
temperature of distillation. The heating process would have extracted and partially
degraded the proteins, lipids, polysaccharides and nucleic acids of these
microorganisms. However, if the dunder is subsequently stored before use in rum
fermentations, it is likely to become contaminated and support the growth of a
specific microbiological flora, as already mentioned. It is broadly mentioned
throughout the literature that dunder is acidic (with high levels of butyric and acetic
acid) and is enriched in nutrients such as amino acids, vitamins and peptides
(Nicol, 2003).

Table 2.4 is an accumulation of data from articles describing the properties and
composition of dunder (or distillery wastewater). As limited information is available,
there are large variations in ranges for some of the chemical properties listed.

Table 2.4 Chemical properties of dunder (distillery wastewater molasses fermentations)

Properties Value Properties Value
a,c d*
pH 3.0-5.0 Protein (w/w) 21.0
d*
Electrical conductivity 346 Methionine (w/w) 4
d*
Volatile fatty acids (g/L) 1.6 Tryptophan (w/w) 2.4
a,c d*
BOD (g/L) 8.7- 48 Lysine (w/w) 10
a,b,c b,d*
Total chemical oxygen demand(g/L) 51.2-120 Calcium (w/w) 1.46-3
d*
Volatile solids (g/L) 50 Iron (w/w) 0.35
b d*
Volatile suspended solids (g/L) 1.8-2.8 Phosphorus (w/w) 0.23
a d*
Total solids (g/L) 28.1-111 Phenol (w/w) 0.23
a,c d*
Total nitrogen (g/L) 0.6-2.2 Crude Fibre (w/w) 5.5
a d*
Suspended solids (g/L) 1.0-7.0 Ash content (w/w) 56
d*
Nitrates (g/L) 1.5- 4.9 Glucan (w/w) 3.6
a d*
Calcium (g/L) 0.6-2.6 Mannan (w/w) 2.8
a,b d*
Magnesium (MgO) (g/L) 0.15-1.1 Glycogen (w/w) 1.2
a,b d*
Sulphate (g/L) 1.3 – 5.2 Thiamine (w/w) 0.06
a d*
Reducing sugar (g/L) 1.0-15.0 Ascorbic Acid (w/w) 3
a b
Ammonia (mg/L) 40-200 Acetic Acid (g/L) 1.34-1.6
a,b b
Potassium (K2O) (g/L) nd-20.0 Propionic Acid (g/L) 0.09-0.12
a b
Iron (mg/L) 120 Butyric Acid (g/L) 0.29-0.66
a
Copper (mg/L) 3.8
a b c d
Basu (1975); Bories et al (1988); Wilke et al (2000); Rameshwari & Karthikeyan (2005)
*
( composition given as percentage dried yeast sludge)
 26

2.2.2 Preparation of Fermentation Media

[Link] Yeast Inoculum for Fermentation

The microbiology of rum fermentation is discussed later in Section 2.4. Modern rum
distilleries conduct the fermentation by inoculation with starter cultures of selected
strains of the yeast, Saccharomyces cerevisiae. These strains may be maintained
as pure cultures “in house” and propagated to inoculum volumes as needed on
site. Some distillers may purchase their yeast as active dry cultures from
specialized companies and rehydrate them for direct inoculation into the
fermenters according to the manufacturer’s instructions. Companies that supply
distiller’s yeasts for rum production are Lallemand ([Link]) and AB
mauri ([Link]). The process for yeast propagation on site is briefly
outlined here and described in Murtagh (1995 a) and Nicol (2003).

Yeast propagation is initiated by preparing an inoculum from a stock culture of the

selected yeast strain. This stock culture will be maintained on site and securely
stored in established culture collections for retrieval as needed. Purity of the culture
is verified by agar plating, from which a small volume (100 – 500 mL) of liquid
culture is prepared under strict aseptic conditions. This culture is used to inoculate
about fifty litres of sterile medium (autoclaved) and incubated with aeration to
increase the numbers of growing cells prior to aseptic transfer of this culture to a
larger volume (500 litres). Culture transfer to progressively larger volumes is
conducted to provide a yeast inoculum that gives 10-30% of the final molasses
fermentation and a starting yeast population of approximately 106 cells/mL
(Murtagh, 1995a).

The propagation medium is usually similar to the molasses fermentation medium

(with yeast nutrient addition) to ensure that the yeast is well adapted to that
condition. While medium used in the early stages of propagation can be sterilized
by autoclaving, the final stages of propagation usually require large volumes of
molasses medium that is not autoclaved and has been processed in a manner
 27

similar to the final molasses fermentation medium to contain about 10-15%

fermentable sugars (Murtagh 1995b, Nicol 2003).

Glucoamylase and possibly yeast foods are also added to the propagation
medium; however, few guidelines are available. The temperature at which
propagation is conducted is normally monitored and controlled. There are
variations in temperature from plant to plant (depending on yeast strain). Generally,
propagation is conducted at a temperature at least 2-5°C below that of normal
fermentation temperature. Once the yeasts have entered their active stage of
growth (log phase), they are transferred from the propagator into the batch
fermenter. There are typically 4 different propagation systems that can be used in
industry; continuous, semicontinuous, multiple batch and single batch (Nicol,
2003).

Throughout propagation, quality control measures should ensure cell viability and
culture purity. Yeast cell viability can be quickly determined using methylene blue
(or other cellular stain) to differentiate between viable and non-viable cells. Counts
performed in combination with a haemocytometer can give relatively quick
approximations compared to cultural plating methods. Cultural purity can also be
quickly determined using microscopy; however, cultural plating or real time PCR
can also be used (Simpson, 1973; Nicol, 2003)

[Link] Fermentation

Fermentation by microorganisms, principally yeasts, is the key operation in rum

production. The profile of flavour volatiles that distinguish rum from other distilled
alcoholic beverages is produced during the fermentation of molasses by the
microorganisms that grow. Without this fermentation, there would be no rum.
Yeasts conduct an alcoholic fermentation of the sugars in molasses, metabolizing
them into mainly ethanol and carbon dioxide, and a vast array of small amounts of
secondary end products. In some cases, bacteria may be associated with the
 28

fermentation (Lehtonen & Suomalainen, 1977; Ganou-Parfait et al 1991.;

Fahrasmane & Ganou-Parfait, 1998). The microbiology and biochemistry of the
fermentation will be discussed in more detail in Section 2.4.

In earlier times, fermentation was conducted in large, open concrete tanks or large
wooden vessels. Today, most rum fermentations are conducted in large (up to
100,000L) stainless steel closed “cyclindro-conical” vessels. These vessels are
equipped with stirring and sparging devices, temperature control, and cleaning in
place (CIP) facilities (I’Anson, 1971; Nicol, 2003; Broom, 2003; Piggott, 2009).
Although the fermentations may be gently stirred to keep the yeast cells in
suspension, they are not aerated.

Rum fermentations can vary in length from 24 hours up to 10 days. Most distilleries
run fermentations to a standardised time for each specific rum. Longer
fermentations are used to produce the heavier flavoured rums, while shorter
fermentations (24 -30 h) are used to produce lighter style rums. Fermentations are
generally conducted at 28-35°C to maximize their rate of completion, and are
considered complete when the desired alcohol levels (%) have been reached
(approximately 5-7%). Some distilleries use change in final gravity or ° Brix, from
set up to determine the completion of fermentation; however, it should be noted
that all three units are related (Destruhaut et al, 1985; Fahrasmane & Ganou-
Parfait, 1998).

Since the fermentation process generates heat, it is necessary to cool the

fermenters so the temperature does not exceed 37°C. At temperatures exceeding
this value, the yeast becomes sensitive to the increasing levels of ethanol and may
be inactivated. If this occurs, the fermentation will stop and remain incomplete or
“stuck” (Arroyo, 1945a; I’Anson, 1971). Such occurrences lead to major
inefficiencies. Temperature control is needed to ensure that the temperature limits
are not violated and that yeasts are not killed in the process. Cooling can be
provided to help control temperature fluctuations. This can be in the form of;
 29

internal cooling coils or panels, double jacketed walls with cooling in the outer
walls, recirculating spirals, plate and frame or shell in tube heat exchangers
(I’Anson, 1971; Broom, 2003)

2.2.3 Distillation of Ferment

Distillation is a critical process in rum production that separates, concentrates and

selects the volatile components of the fermented molasses (Bluhm, 1983). The
volatile fraction of the fermented molasses consists predominately of ethanol, and
lesser amounts of higher alcohols, organic acids, esters, phenols and some
carbonyl and nitrogenous compounds. The use of the collective term “congeners”
has been applied to describe all volatile components of rum other than ethanol
(Nykanen & Suomalainen, 1983; Murtagh, 1995 a, b).

Distillation is used to capture most of the ethanol and refine flavour by selecting for
the types and concentrations of other, desirable volatile compounds; however, it
does not create these base components. Creation of the desirable flavour volatiles
occurs mainly during the fermentation of molasses, but some may occur in the
molasses before fermentation (Yokota & Fagerson, 1971). Distillation can produce
new compounds via esterification, dehydration etc from the base components
produced in fermentation. The use of pot distillation (outlined in Section [Link]) is
known to increase furfural concentration (Madrera et al, 2003).

The technology for the commercial distillation of alcoholic beverages, including

rums, varies with the manufacturer but the general principles of the process are
common and are described in Piggott (2009a, b). Some other reviews and
discussions of the distillation process as it relates to rum production are given by
Murtagh (1995a), Reche et al (2007) and Sampaio et al (2008).

Distillation is based on the principle that different components within a liquid

mixture, such as the fermented molasses, have different temperatures at which
 30

they boil and transform to a vapour or gaseous phase. Heat is applied to the liquid
mixture. Simply, the smaller, more volatile components are vaporised and boil off
first and are progressively followed by the less volatile components. Non-volatile
substances are left in the liquid mixture. As the temperature of the vaporised
fraction is decreased, the individual components revert to their liquid phase and
can then be collected in this state. This process of transformation back to the liquid
phase is called condensation (Wankat, 2007).

The basic apparatus for distillation consists of a vessel (still) in which the liquid
mixture is heated. The base of the vessel is attached or connected to columns, into
which the volatiles vaporize and eventually condense back to a liquid, and capture
vessels for collecting the condensed liquid (distillate). The columns may be
differentially cooled to encourage condensation (Kampen, 1975; Nicol, 2003).

Distillation efficiency can be increased by the addition of a reflux step. This is a

method of returning a proportion of the condensed distillate back into the distillation
column. The down flowing reflux liquid enters the column and cools. It condenses
the rising vapours and works to increase the separation efficiency of the distillation
column. Increasing the amount of reflux for a column will improve the separation of
lower boiling components from higher boiling compounds, resulting in a distillate
with a higher composition of a desired product (Wankat, 2007).

Separation of the volatile compounds is based on volatility differences and occurs

through heating (and cooling). The degree of separation of the desired component
may be affected by various operating conditions such as pressure, temperature,
the initial feed composition and liquid phase conditions. Controlling the distillation
process is crucial for the production of product with consistent and desirable quality
(Nicol, 2003; Wankat, 2007).

Two types of distillation processes are used in the production of rum: batch
distillation and continuous distillation. Batch distillation, in pot stills, is used to
 31

manufacture rums with stronger, heavier flavours such as those of Barbados,

Bermuda, Jamaica and other English speaking regions in the Caribbean.
Continuous distillation, using column stills, is used for the production of lighter style
rums of the former Spanish colonies (such as Cuba, Panama). Some distilleries
use a combination of both techniques (I’Anson, 1971; Clutton, 1974; Lehtonen &
Suomalainen, 1977; Bamforth, 2007; Sampaio et al 2008).

[Link] Batch Distillation

Pot stills are the earliest known distillation apparatus and, until the mid 1800s, all
rums were produced by this process. Pot stills consist of three parts; the kettle
(boiler), condenser and gooseneck, similar to those used in whisky distillation (see
Figure 2.2 and Figure 2.3(a)). The kettle is the base of the vessel into which
fermented molasses is transferred and heated by steam injection, either directly or
indirectly through heating coils. Low boiling components, including ethanol, will
begin to vaporise and pass through the gooseneck into the condenser (or retort
depending on still). Limited reference is made to specific temperatures throughout
the literature. This may have been influenced by two factors. Traditionally pot stills
are heated by steam or fire, making temperature difficult to control. Distillers
wishing to keep production practices secret from competitors may also have some
bearing on the lack of records (Piggott, 2009 a, b).

The liquid distillate obtained from this type of process is also known as “single
distillate” since it is processed through the still only once, giving a product of about
40-60% alcohol by volume (Nicol, 2003). This process gives a heavy pot still rum.
Typically, however, this liquid is processed a second time, thus producing a double
distillate which is cleaner and stronger than the single distillate. This re-distillation
enables further separation of the desirable volatile compounds. This occurs due to
the increased ethanol concentration of the primary distillate compared to the
molasses fermentation. This increase in ethanol concentration decreases the
boiling temperatures, thus ensuring greater variation as to when different volatile
 32

compounds are liberated into the vapour phase. The process can be repeated
several times, thereby obtaining a cleaner, stronger more rectified spirit each time.
Distillation is performed batch by batch and is very labour intensive. Pot distillation
is usually performed in conjunction with the addition of dunder in the fermentation,
thus producing a heavy, high ester rum (Nicol, 2003; Piggott, 2009a, b).

The first fractions (first 5 minutes of process) to be collected as distillate contain

about 88% alcohol by volume (ABV), but also contain some pungent less desirable
flavour volatiles. This fraction, often referred to as the first cut or low wines, may be
discarded. The final fractions of batch distillation will contain much less ethanol
(less than 40-45% ABV) as it has already been distilled out, along with other
volatiles with less desirable flavour attributes. Such fractions are often referred to
as late cuts, talls or feints, and may also be discarded. The “centre” fraction (also
called “hearts” or “spirit” or “middles”) usually contains the most desirable flavour
volatiles and is the cut that is collected (85% ABV at the beginning). As distillation
proceeds, the concentration of ethanol in the distillate decreases. Generally,
collection of the “hearts” is stopped when the alcohol content of this fraction is
about 40-43% ABV (Nicol, 2003; Piggott, 2009a)

Figure 2.2 Pot stills used in the production of distilled spirit. (a) crude rum still (Artisan
Distiller, 2011) (b) Scotch whisky production at Roseisle distillery, Scotland. (Sword 2010)
 33

[Link] Continuous Distillation

Continuous distillation gives a distillate with more consistent composition than

batch distillation. Detailed history and theory of continuous distillation can be found
in I’Anson (1971), Nicol (1989, 1993), Murtagh (1995a), and Wankat (2007 –
Chapter 3). The distillation column or tower consists of two sections (see Figure
2.3(b)). The portion of the tower above the molasses feed entry point is defined as
the ‘rectifying section’ of the tower. The part of the tower below the feed entry point
is referred to as the ‘stripping section’ of the tower. Throughout the column there
are a number of horizontal trays placed at different levels. Pre-heated fermented
liquid is usually introduced at the top (or at least half way up the column).

As the liquid makes its way down the column, it is heated by rising vapour. This
liquid-vapour contact occurs on the horizontal trays which commonly have holes
punched through the metal, or specialised “bubble caps” which also allow for liquid-
vapour contact. As the liquid flows on to the tray, the rising vapour is forced to
come in contact with it. During this contact, heat is exchanged and the more
volatile components tend to concentrate into the vapour. After repetitive liquid-
vapour contacts over the height of the column, the most volatile compounds rise to
the top of the column. This partial separation allows column distillation units to be
more efficient at separating fermentation components than pot stills. Once the
fermented medium reaches the bottom of the still, it contains no alcohol and is
removed, as dunder, through a release valve. Careful control of the heating rate of
the column, and the degree of reflux in the rectifying section, allows column
operators to dictate the ethanol concentration of the primary distillate taken from
the column (I’Anson, 1971; Murtagh, 1995b; Wankat, 2007). The highest
temperature (usually in excess of 90°C - 95°C) in the tower will occur at the base,
and the temperature in the tower will regularly and progressively decrease from the
bottom to the top of the tower. To produce the temperature variations, reboilers
(heat exchangers) are often used to heat and partially vaporise the liquid streams
in the lower sections of the column (I’Anson 1971; Wankat, 2007).
 34

The rectifying sections of both pot stills and continuous columns are often under
reflux, which is where condensed liquid collecting near the base is continually
pumped and fed back into the rectifying section near the top. It then simply
combines with the liquid phase flowing down through the column. This enhances
the interaction of the vapour and liquid phases and achieves greater separation of
the volatile components. Distillation systems for rum production usually operate as
two separate processes: primary distillation and secondary distillation. (Wankat,
2007).

A B

Figure 2.3 Distillation at The Bundaberg Distilling Company, Bundaberg, Australia.

(a) Pot still, (b) column still. (Photo D. Flett)

Primary distillation operates similarly to pot distillation, generally referred to as “low

wines” with a concentration of approximately 50% ABV. Primary distillation allows
subsequent distillations (secondary distillation) to achieve greater separation of the
volatile compounds. Unlike primary distillation, the distillate obtained from the
secondary distillation is collected in fractions. These fractions are either based on
the collection time or the ethanol concentration of the distillate collected from the
rectifying section. If sufficient separation of the volatile components is achieved,
each fraction should contain significantly different flavour profiles. The first fraction,
rich in highly volatile components, is called the heads or ‘high feints’ and is usually
discarded because it contains high concentrations of aromatic esters and acids.
The second fraction is the final rum distillate or ‘raw rum’ or ‘hearts’ which is only
 35

collected up to a certain point, once again based on either collection time or

percentage of ethanol, depending on distiller. Collection of distillate beyond this
point is undesirable as fusel oils constitute a large proportion of the less volatile
components (I’Anson 1971; Nicol 2003; Wankat, 2007).

2.2.4 Ageing and Maturation

Freshly produced rum distillate has some strong, raw flavours that are not
appreciated by all consumers. Consequently, the fresh distillate can be subjected
to a process of maturation where it is stored in large wooden barrels, generally
made of oak, that hold approximately 120 – 150 L (Quesada Granados et al,
2002). It is important to note that some provincially produced rums are not subject
to ageing and are available for consumption immediately after distillation (Broom,
2003). During maturation, a range of physical and chemical interactions takes
place between the barrel wood, the surrounding atmosphere and the maturing
spirit. These interactions transform both the flavour and composition of the
alcoholic beverage. The effects and time required for maturation are variable and
are influenced by a wide range of factors, particularly the type of barrel used
(Kampen, 1975; Nicol, 2003).

Maturation is not the same as ageing. Maturation is the end stage which is reached
after ageing (the means to obtain the maturation). A mature rum is not defined as
one that has spent a fixed period of time in a barrel. Rather, maturity is measured
by the rum possessing distinctive characteristics acquired during ageing. Such
characteristics are body, colour, aroma and taste. Distillate straight from a still is
clear but maturation in wooden barrels gives the finished rum its yellow/golden
colour, depending on how long it is aged. Ageing is simply the time the rum is
stored in the wooden barrel (Nicol, 2003).

While there have been extensive studies done on the ageing of distilled alcoholic
beverages such as whisky and brandy (Mosedale, 1995; Singleton, 1995;
 36

Mosedale & Puech, 1998), there are few reports on rum maturation (de Torres et
al, 1987; Thompson, 2009). Similar to whisky and brandy, the constituents of rum
continually undergo changes during storage in wood.

Figure 2.4 shows the diversity of chemical reactions that may occur during
maturation. These include: (i) direct extraction of chemical constituents from the
wood, (ii) decomposition of oak on a molecular level and interaction of resulting
compounds with the distillate, (iii) reactions between the constituents extracted
from the wood and those in the rum distillate, (iv) reaction between wood
compounds within the raw rum, (v) reactions between raw rum compounds, (vi)
evaporation of volatile compounds through the cask, and (vii) interaction between
the raw spirit and air present in the cask/vat. (Mosedale & Puech, 1998)

Figure 2.4 Interactions occurring during maturation of raw rum spirit in an oak barrel

The length of time that raw rum is aged depends on the type of rum being
produced and also, to a certain extent, the market in which the rum will be sold.
“White” rums, those that are clear and used as the basis for cocktails, are generally
not aged for extended periods, except for where the law requires a minimum
ageing period. These rums are usually aged in old, well used barrels and are
submitted to charcoal filtering to remove any colour prior to bottling. Amber and
dark rums are typically aged for a period of time between 12 months and 25+
years. Typically the longer the rum is in oak, the darker the rum (Broom, 2003;
 37

Piggott, 2009b).

The rum industry uses two main types of barrels: those which have already been
used in curing whisky and new oak barrels. The main reasons for using pre-used
whisky barrels is that they are cheaper than new barrels and they have previously
been ‘cured’ or charred.

Specific types of wood are used in barrel construction but American oak (Quercus
alba) is common. It is known to produce specific flavours including vanilla, coconut
and spice (Mosedale & Puech, 1998; Quesada Granados et al, 2002). Cask/barrel
manufacture is not standardised, with large discrepancies between American and
European cooperages. Barrels are very expensive and in high demand,
consequently, rum distilleries may buy old barrels that have been used for the
maturation of other alcoholic beverages; such as whisky, wine and brandy. Barrels
are heated or charred (burning of the inside surface of the barrel) prior to raw spirit
being stored. This charring changes the physical and chemical composition of the
wood by caramelising sugars, increasing vanillins and helping to extract tannins
(Mosedale, 1995; Mosedale & Puech, 1998; Broom, 2003).

The quality of ageing barrels is a very important parameter in the production of

rum. Distillers, however, have little control over some of the factors affecting this
quality such as: character of the soil and climate in which the trees were grown,
age of the trees when cut, manner in which they were cut, part of tree from which
the staves were derived, variety and amount of resins present, and period of
ageing given to the boards prior to barrel construction (Mosedale, 1995).

Many countries have legislation governing the minimum ageing time which must
occur prior to a product being sold as “rum”. For example, to be sold in Australia,
the Dominican Republic and Panama, rum must be aged for a minimum of 2 years.
Mexico legislation requires a minimum of 8 months. However, some countries are
less stringent and allow rum to be sold without any ageing such as rhum agricole
 38

and Brazilian cachaça (Broom 2003).

2.2.5 Cachaça

By definition, cachaça is a type of rum, as it is produced from sugar cane juice

(Marini et al, 2009). Cachaça is made from fermented sugar cane juice. It is the
national drink of Brazil with approximately 1.5 billion litres consumed annually
(2007) with approximately 1% exported. Economically it is of great importance in
Brazil, generating approximately 400,000 jobs and second only to beer in
consumption (Souza, 2010). Consumption is concentrated mainly within the low-
income population.

As with other rums, cachaça production consists of fermentation, distilling and

ageing. While production has traditionally been artisanal, there has been a move
towards more industrial production in recent years. National surveys have shown
that a greater export market may be achievable if quality of the finished product
were to be improved (Souza, 2010).

Few other spirits have the post distillation sensory quality that cachaça possesses:
consequently, cachaça is rarely aged. This is mainly due to the target market
discussed previously, and need for a cheaper product. Distinction between unaged
(white) and aged (gold) cachaça is similar to other distilled beverages, with
improvements with flavour mellowing and typical subtle woody and vanillin notes
developing along with gold-yellow colouration during ageing (Broom, 2003; Marini
et al, 2009)

White cachaça is usually bottled immediately after distillation (sometimes aged for
up to 12 months). Gold cachaça is aged in wooden barrels and should be
consumed straight or ‘neat’. Ageing is usually undertaken until 3 years; however,
there are some premium cachaça that are aged for 15 years (Souza, 2010).
 39

Changes in distillation techniques such as moving from single distillations to double

distillations, as used in molasses rum production, are currently being undertaken in
Brazil (Souza, 2010). Closer examination of the microflora involved with the
fermentation and their effects on cachaça quality have also been conducted by
some groups (Schwan et al, 2001; Dato et al, 2005; Gomes et al, 2010) and are
discussed in more detail in Section [Link].

2.2.6 Quality Assurance and Control in Rum Production

As a finished product, rum needs to meet the acceptance criteria of consumers.

Such criteria will be determined by appearance, flavour and aroma. The rum
should be free of technical faults and meet any technical and legal specifications.
In addition to these criteria, there needs to be consistency in quality, which would
require little variation from batch to batch production.

The basic principles of good manufacturing practice and quality assurance and
control as applied to food and beverages in general, would also apply to the rum
industry. General descriptions of these principles, including the Hazard Analysis
and Critical Control Point (HACCP) concept are given in Jouve (2000), Bernard &
Scott (2007) and Jukes (2009). HACCP is now a key component of good
manufacturing practice. Although HACCP has been more broadly applied to the
management of safety issues in food and beverage production, it can also be
adapted to quality management.

The seven principles of HACCP are listed in Table 2.5 and these principles should
be applied to the entire rum production process. There is very little published
literature on the systematic application of quality management systems to rum
production.
 40

Table 2.5 HACCP principles (Bernard & Scott; 2007)

Seven Principles of HACCP
(i) Conduct a hazard analysis
(ii) Determine the Critical Control Points (CCPs)
(iii) Establish critical limits for each control point
(iv) Establish a method of monitoring CCPs
(v) Establish corrective actions
(vi) Establish procedures to verify effectiveness of HACCP system
(vii) Establish documentation and recording procedures

Connolly (1997) has mentioned the general importance of quality management in

the production of distilled alcoholic beverages. With respect to rum, Nicol (2003)
has listed specific operations in the process where some control is needed and
these are:
- quality of molasses as obtained from the supplier,
- molasses storage,
- preparation of molasses,
- preparation of inoculum cultures for fermentation,
- control of fermentation process,
- control of the distillation process,
- management of ageing/maturation,
- packaging,
- application of effective cleaning and sanitation throughout production
process.

For each of these steps in the rum production chain, it is necessary to identify and
develop specific management criteria. This would require, for example, the
development of specifications for all raw materials, identification of any hazards
along the production chain that may impact on quality, the determination of critical
control points and control limits, description of cleaning and sanitation procedures,
and systematic documentation of the management plan. A detailed discussion of
quality management for a rum production process will be given in Chapter 3.
 41

2.3 Chemistry of Rum Flavour and Quality

The distinctive feature of rum and its main quality criterion is attributed to its
flavour. As mentioned previously, this flavour is determined by the diversity and
concentration of volatile products collected as a consequence of distillation of the
fermented molasses. The main component of rum is ethanol and its concentration
during fermentation reaches a maximum of 6-8% v/v. After distillation and bottling,
the final concentration can vary between 37 and 80 mg/ ml (i.e. 37.0- 80.0 %),
depending on brand (Nicol, 2003).The remaining components represent the
congeners of which about 100 have been identified. These products have their
origin as components of the molasses, components produced by metabolism of
molasses by microorganisms and the changes which occur to them during
maturation.

The chemistry of rum flavour has been studied since the early 1900s. Lehtonen
and Suomalainen (1977) have given the most comprehensive discussion of the
chemical constituents of rum flavour (aroma) and the factors that affect their
production. More general discussions are provided in Clutton (1974) and Nicol
(2003). Rums are categorized as heavy, light or medium style products according
to their flavour or aroma profile, this being determined by the microbiology of the
fermentation process and the distillation process (Lehtonen and Suomalainen,
1977). A comprehensive summary of chemical compounds found in rums is
included as Appendix A.

Early studies on the chemistry of rum flavour using “classical” analytical methods,
such as functional group analysis, established the presence of various volatile
acids, esters, aldehydes, higher alcohols and furfurals (Simmonds, 1919; Arroyo
1942; Arroyo, 1945a). The most significant advances in understanding the
chemistry of rum flavour came with the development and application of gas
chromatography and later gas chromatography linked with mass spectrometry
(GC-MS) or high performance liquid chromatography (HPLC) (Allan, 1972;
 42

Cardoso et al, 2003).

The first study on a heavy-body rum by gas chromatography combined with

infrared and mass spectrometry was done by Maarse & ten Noever de Brauw
(1966) with approximately 75 flavour compounds being identified. The main
components were esters, aldehydes, acetals and alcohols. The authors noted
surprise at the large proportion of esters and acetals. However, there was concern
that the acetals may not have been present in the rum itself, but developed as a
result of sample preparation. Propanol, isobutanol and isoamylalcohol were the
main alcohols present, aside from ethanol. Furans, heterocyclic organic
compounds consisting of a five membered aromatic ring, were expected to be
detected; however, there was interest in one specific furan, 2-methyl-
tetrahydrofuran-3-one, which at the time, was also discovered as an aromatic
compound present in coffee.

Nykanen (1986) studied the fatty acids of whisky, cognac, brandy and rum. Rum
was found to contain the largest amount of volatile acids, being in the order of
600mg/L. Between 75% and 90% of this amount was due to acetic acid. Rum was
also found to contain more butyric and propionic acid than whisky or cognac, with
propionic acid predominating. The principal higher fatty acids present were
myristic, palmitic and palmitoleic.

Liebich et al. (1970) identified over 200 flavour compounds in a Jamaican rum
using GC-MS techniques. The compounds identified were characterised as various
esters, acids, alcohols, phenols, lactones, carbonyl compounds, acetals, pyrazine
derivatives and hydrocarbons. As a part of this research, the development of an
imitation rum was completed, with the best imitation using as many of the 200
volatile components available. It was concluded that all 200 compounds were
needed for a good, full rum flavour.

Pino et al (2002, 2007, 2012) have performed research on aroma compounds in

 43

rums since 2002. Initial studies (Pino et al 2002) determined fatty acid ethyl esters
in Cuban rums. Using headspace solid phase microextraction (HS-SPME), this
preliminary work was developed further with the recovery of 184 volatile
compounds from six different samples of four different brands of Cuban rum (Pino
et al 2007). The volatile compounds included 64 esters, 47 benzoic compounds,
16 terpenoids, 14 alcohols, 10 acetals, 9 aldehydes, 6 phenols, 6 furans, 3 acids
and 3 benzopyrans. More recently, Pino et al (2012) investigated an aged,15 year
rum from Cuba using gas chromatography – flame ionisation detection (GC-FID),
GC-MS and GC-O (Gas chromatography – olfactometry). GC-O enables the
panellist to sniff the sample after injection and evaluate the odour. The retention
time of the odour is recorded and compared to previous runs on GC-MS and GC-
FID. This study identified and quantified a total of 116 volatiles. Alcohols,
unsurprisingly, made up the greatest proportion of volatile compounds (82.4%),
with 3-methylbutan-1-ol and 2-methylpropan-1-ol being the major alcohols present.
The next prevalent compounds were esters with ethyl acetate, ethyl lactate, ethyl
decanoate and diethyl succinate present in the greatest concentrations. It was
considered that these alcohols and esters were mainly produced during microbial
fermentation of the molasses (Pino et al; 2012). Pino et al (2012) also isolated
lactones (3) and phenols (13) which were thought to predominately come from the
oak barrels during maturation. While 12 acids were detected, they were considered
to have little impact on rum flavour due to their low concentrations and low flavour
thresholds. Of these acids, the main ones found were acetic acid, octanoic acid,
decanoic acid and hexanoic acid.

Sampiao et al (2008) compared the key flavour volatiles (61 analytes) in 44

samples of Cuban and non-Cuban rums. Of these, 8 compounds (isoamyl alcohol,
n-propyl alcohol, copper, iron, furfuraldehyde, benzaldehyde, epicatechin and
vanillin) were able to comprehensively discriminate between the Cuban and non-
Cuban rums. Using statistical methods, the results were clustered into two distinct
groups. Further comparison was performed between Cuban rums and those from
around the world. Benzaldehyde, isoamyl alcohol, copper and furfuraldehyde had
 44

higher statistical loading values characteristic of Cuban rums when compared to

compounds whose statistical loading values were lower and thus coincided with
rums from other countries. The analysis also highlighted differences between
distillations performed either by pot stills or column stills.

The only previous research conducted on the flavour volatiles of Bundaberg Rum,
was performed by Allan (1972) who identified 77 compounds consisting of
alcohols, esters, carbonyl compounds, acetals, sulphur compounds and
hydrocarbons. This research identified a new compound (1-octen-3-ol) which had
not been identified in a rum previously. The importance of methyl-salicylate to rum
flavour was also discussed.

Each study mentioned previously identified key flavour compounds, indicative of

rum flavour. Appendix A tabulates all compounds found in these studies and
includes, where available, the concentrations quantified. A summary table (Table
2.6) highlights the most important sensory components, their detected
concentrations and source. A comparison to well known sensory ‘notes’ is included
to identify unique flavours or aromas.
 45

Table 2.6 Important flavour volatiles in rum

Compound ppm Isolated from Sensory notes
Esters
c,f
Ethyl acetate 73-200 Sugar cane molasses, Jamaican rum Sweet – pear drops
c,f
Ethyl formate 10 Sugar cane molasses, Jamaican rum Characteristic smell of rum
c,f
Isoamyl acetate 5-10 Sugar cane molasses, Jamaican rum Banana/pear
c
Ethyl propionate 50 Jamaican rum Fruity
c
Ethyl n-butyrate 220 Jamaican rum Fruity - pineapple
c
Ethyl n-valerate 40 Jamaican rum Pleasant fruity - apple
c,h,i
Ethyl n-hexanoate 1-40 Jamaican rum, Cuban Rum Powerful fruity - pineapple
c,h,i,k
Ethyl n-octanoate 15-50 Jamaican rum, Cuban Rum Floral fruity - wine
c,h,i,k
Ethyl n-decanoate 25-130 Jamaican rum, Cuban Rum Sweet, oily, nut-like, yeasty
c,h,i
Ethyl n-dodecanoate 12-15 Jamaican rum, Cuban Rum Floral fruity - waxy
c
Ethyl n-hexadecanoate 5-50 Jamaican rum Mild, waxy, sweet
c,f
ethyl linoleate 50 Jamaican rum Faintly fruity
c
Ethyl lactate 10 Jamaican rum Tart butterscotch, pineapple
c,e
Methyl salicylate 0.5-25 Jamaican rum, Australian rum Winter green
Acids
c,f
Acetic acid 10.3-35 Sugar cane molasses Sour vinegar
c,f
n- hexanoic acid 0.3-15 Jamaican rum Fruity cheese
b,c,f
n-butyric acid 0.3-7.5 Sugar cane molasses, high Creamy sour cheese, fruity
concentrations in Jamaican rum
c,f
n-valeric acid 0.04-7.5 Sugar cane molasses Earthy cheese
b,c,f
n-propionic acid 0.2-1.5 Sugar cane molasses Dairy, fruity
c
n-octanoic acid 4.3-7.5 Jamaican rum Rancid vegetable
Alcohols
c,f,j
1- propanol 7.5-420 Sugar cane molasses, fermented Alcoholic ripe fruit
molasses
c,f,j
1-butanol 10 Fermented molasses Medicinal
c
Isobutanol 100 Jamaican rum Fruity wine-like
c,f
2-methyl-1-butanol 200-210 Sugar cane molasses, fermented Roasted wine, whisky-like
molasses
c,f
3-methyl-1-butanol 860-1000 Sugar cane molasses, fermented Fusel, whisky-like
molasses
Phenolic compounds
c,f
4-methylguaiacol 0.05 Decarboxylation by yeast and Sweet, coffee, cocoa
bacteria of vanillin
a,c,d,f,g,k
Eugenol nd-1.36 Extracted from oak during maturation Spicy, woody clove
c,f,j,k
Vanillin 0.25 Cane molasses Vanilla, sweet
Extracted from oak during maturation
Cuban Rum
a b c d
Maarse and ten Noever de Brauw (1966), Nykanen (1968), Liebich et al. (1970), Timmer et al
e f g h
(1971), Allan (1972), Lehtonen and Suomalainen (1977), Lehtonen (1983), Pino et al. (2002),
i j k
Pino et al (2007), Sampaio et al (2008), Pino et al (2012).
 46

2.4 The Role of Microorganisms in Rum Production

The earliest known scientific study into the microbiology of rum was undertaken by
Bryan Higgins, an Irish naturalist and philosopher in chemistry (Keattch, 1991). He
resided in Jamaica from 1797-1799. It took a further 100 years for Greig (1895) to
publish an article entitled “The Jamaica Yeasts”, a paper outlining preliminary
studies of a yeast found in association with Jamaican molasses used at a
Jamaican rum distillery. Studies on the microbiology of rum production date back to
the 1890s with some of the earliest works being reported by Greig (1895), Pairault
(1903) Allan (1906), and Ashby (1909). Despite more than 100 years of research
since that time, microbiological understanding of the process remains very limited,
and is significantly lagging compared with knowledge of the microbiology of other
alcoholic beverages such as beer, wine or whisky.

There are two main points in the production chain where microorganisms impact
on the process and determine rum quality and process efficiency. These points are
storage and preparation of the raw materials (principally molasses, sugar cane
syrup or sugar cane juice), and the process of fermentation. Since this thesis
concerns molasses style rums, the following sections will focus on the microbiology
of molasses and the microbiology of molasses fermentation. Because dunder plays
a unique role in the process, a section will also consider its microbiology. A final
section will give a brief overview of the microbiology of cachaça style rums.

In developing these sections, particular consideration will be given to:

 Which microbial species occur throughout the production chain
 The survival and growth behaviour of these microbial species throughout
production and the various factors that affect this behaviour
 How the biochemistry of this microbial growth changes the chemical
composition of molasses and impacts on the flavour and sensory quality of
the final rum product.
 47

2.4.1 Microbial Ecology of Molasses

The chemical composition and properties of molasses as they relate to rum

production have been described in Section [Link]. The high concentrations of
sugars, low water activity and relatively low pH make these raw materials
unfavourable environments for growth and survival of microorganisms.
Nevertheless, the literature contains sporadic reports on the isolation of yeasts and
bacteria from these materials. The microbiological content of molasses, initially,
consists only of endospore forming bacterial species as a result of the high
temperatures involved in sugar extraction. An indigenous microflora will establish
through contaminations occurring during transportation, storage, contact with
processing equipment and general exposure to the elements (Moroz, 1963;
Watson, 1993). Molasses is stored in various kinds of containers, depending on
manufacturing conditions, from open air wells to secure stainless steel tanks.
During this time, microbial contaminants have the potential to grow and produce
metabolic end products that may impact on rum quality. This indigenous microflora,
itself, is diverse and may impact on the ecology of the fermentation process.

While there are isolated reports on the recovery of microorganisms from molasses,
there have not been any systematic investigations of the microbiology of molasses
or syrups during storage for rum production. Table 2.7 summarises literature
detailing microorganisms present in molasses (and other related products such as
sugar cane and cane juice). Sugar cane and sugar cane juice have been included
as there are limited previous studies detailing the microorganisms naturally present
in molasses.
 48

Table 2.7 Microorganisms found in raw materials, such as molasses and sugar cane juice,
associated with alcoholic fermentations for rum production.
Reference Raw material Species isolated
Zygosaccharomyces Penicillium,
Hall et al (1935) Molasses nussbaumeri, Mucor,
Zygosaccharomyces major, Clostridium saccharolyticum
Zygosaccharomyces
Zygosaccharomyces globiformis
Owen (1949) Molasses nussbaumeri
Schizosaccharomyces pombe
Zygosaccharomyces major
Candida guilliermondii Saccharomyces cerevisiae
Candida intermedia var. Saccharomyces carlsbergensis var.
ethanophila alcoholophila
Candida mycoderma Saccharomyces marxianus
El-Tabey Sugar cane Candida tropicalis Saccharomyces microellipsodes
Shehata (1960) juice Endomyces magnusii Saccharomyces rosei
Kloeckera apiculata Saccharomycodes ludwigii
Pichia fermentans Torulopsis glabrata
Pichia membranaefaciens Torulopsis stellata
Saccharomyces acidifaciens Torulopsis stellata var. cambresieri
Kampen (1975) Molasses Saccharomyces cerevisiae
Candida krusei Saccharomyces cerevisiae
Candida pseudotropicalis Saccharomyces chevalieri
Candida tropicalis Saccharomyces rouxii
Parfait and Molasses/ cane
Hansenula anomala Torulopsis candida
Sabin (1975) juice
Hansenula minuta Torulopsis glabrata
Saccharomyces aceti Torulopsis globosa
Saccharomyces acidifaciens Torulopsis stellata
Molasses/ raw
Tilbury (1980) Saccharomyces heterogenicus
cane sugar
Ganou-Parfait Propionibacterium jensenii
Molasses Leuconostoc paramesenteroides
et al (1989) Lactobacillus fructivorans
Bonilla-Salinas Schizosaccharomyces pombe Torulaspora delbrueckii
Molasses
et al. (1995) Saccharomyces cerevisiae Cryptococcus albidus var. albidus
Todorov &
Molasses Lactobacillus plantarum
Dicks (2005)
Note: Bacterial species given in the table are shown in bold font

There is a noticeable difference between fresh “green” molasses and “aged”

molasses. It has been debated that the molasses itself undergoes spontaneous
chemical alteration which leads to this “ageing” (Owen, 1911). Because of the heat
processes involved, freshly produced molasses, or syrup, contain few
microorganisms (Owen, 1911; Browne, 1929). Browne (1929) analysed molasses
samples over a 14 year period and failed to detect the presence of any yeasts,
moulds or bacteria.
 49

More recently, several research groups have revealed that molasses harbours
complex microflora, including yeast and bacteria (Hall et al., 1935; Owen, 1949;
Bonilla-Salinas et al., 1995; Fahrasmane & Ganou-Parfait, 1998)

Hall et al. (1935) examined the microflora of Barbados molasses. Using

microscopy, they observed various yeast and bacterial cells but were unsuccessful
in obtaining culturable colonies of yeast when directly plated on to malt extract agar
or cane sirup [sic] agar (an agar made by 1:2 of cane sirup [sic] to nutrient agar pH
to 6.8-7.0). Hall et al. (1935) improved their isolation technique by the addition of
an enrichment step prior to plating. This extra step allowed for 11 cultures of two
yeasts (Zygosaccharomyces nussbaumeri and Zygosaccharomyces major) to be
obtained from two samples of molasses (1932 and 1933). A study was
concurrently run to isolate the bacteria present in Barbados molasses. Clostridium
saccharolyticum was consistently isolated and it was concluded that it was an
important organism in the development of flavour volatiles in rum fermentations
(Hall et al, 1935). Mould was frequently isolated after incubation at a variety of
temperatures and on a variety of media. The most predominant moulds were
considered to be in the genera Penicillium and Mucor. Species of these moulds
were not determined.

Owen (1949) discussed four origins of microorganisms in relation to their presence

in molasses. These included the “epiphytic” flora of the sugar cane plant itself,
microorganisms endemic to the soil that the plants grow in, microflora introduced to
the mills and refineries via contamination through air particulates, and
microorganisms introduced through a secondary raw material used to extract sugar
from molasses and sugar cane juice (for example, starch used in production of
confectionary sugars). While Owen did not perform any experiments at this time,
his theories were informative and provided a “check list” of potential sources of
microflora present in distilleries. To date, no one has done a complete survey,
including all of these origin points, of a distillery.
 50

El-Tabey Shehata (1960) undertook research into the natural yeast microflora of
sugarcane and the associated juice (both fresh and fermented) at sugar factories in
Brazil. Forty three yeast cultures were recovered from the 14 samples taken from
five different factories. Isolated from sugar cane plants and fresh juice, species of
Saccharomyces, Candida, Pichia and Torulopsis were the most predominant.
Yeasts from only three genera were isolated from fermenting juice
(Saccharomyces, Candida and Schizosaccharomyces), while the most frequently
isolated yeasts from fresh juice were S. cerevisiae and Candida krusei.

From the 1970’s to the mid 1990’s several research groups investigated both yeast
and bacterial species in molasses, sugar cane and sugar cane juice. Parfait and
Sabin (1975) examined the prevalence of yeast species in both molasses and cane
juice. Fourteen species were isolated and identified (Table 2.7). As part of a larger
study (discussed further in Section 2.4.2), Ganou-Parfait et al (1989) isolated three
bacteria (Propionibacterium jensenii, Lactobacillus fructivorans and Leuconostoc
paramesenteroides) from molasses.

Bonilla-Salinas et al (1995) isolated and identified 13 yeast strains from sugar cane
molasses produced in Mexico. The most frequently isolated yeast was
Schizosaccharomyces pombe (isolated 7 times). This was the first study to identify
killer strains in molasses, including Schizosaccharomyces pombe which had not
previously been known to have killer strains. This research may help further work
into starter cultures for the rum industry as discussed later in Section 2.4.4.

Previous research has concentrated on isolation and identification of both yeast

and bacterial species. There has been limited research in quantifying population
levels of microorganisms in molasses when used for rum production. Bacteria are
generally thought to exist in populations between 102-103 bacteria/g (Fahrasmane
and Ganou-Parfait, 1998) while, to date, there is no published population data for
yeast species isolated from molasses.
 51

2.4.2 Microbial Ecology of Molasses Fermentation for Rum

Production

Molasses has a high sugar content and a relatively low pH which naturally selects
for the growth of yeasts. Consequently, yeasts are the most prevalent
microorganisms of molasses based rum fermentations, with strains of
Saccharomyces cerevisiae being the most frequently isolated (Lehtonen &
Suomalainen, 1977; Nicol, 2003). Nevertheless, various bacterial species have
been associated with these fermentations and need to be considered as part of the
microbial ecology. Table 2.8 lists various studies that have reported the isolation of
yeasts and bacteria from molasses based rum fermentations and sugar cane juice
rhum agricole fermentations.
 52

Table 2.8 Microbial species associated with the fermentation of molasses or sugar cane juice
for rum or rum agricole production
Fermentation
Reference Species isolated
medium
Greig (1895) Molasses Schizozaccharomyces mellacei
Kampen
Molasses Lactobacillus ssp
(1975)
Parfait & Schizosaccharomyces pombe
Molasses Saccharomyces cerevisiae
Sabin (1975) Schizosaccharomyces ssp
Ganou- Bacillus ateriimus
Bacillus mesentericus
Parfait et al. Molasses Bacillus cereus
Bacillus subtilis
(1987) Bacillus megatherium
Schizosaccharomyes pombe Schizosaccharomyces
Fahrasmane Schizosaccharomyces malidevorans
Molasses
et al (1988) japonicus Saccharomyces cerevisiae
Clostridium spp. Bacillus spp.
Micrococcus luteus Lactobacillus fermentum
Micrococcus varians Lactobacillus fructivorans
Bacillus cereus Lactobacillus hilgardii
Bacillus subtilis Lactobacillus viridescens
Bacillus megaterium Leuconostoc mesenteroides
Bacillus sphaericus Leuconostoc
Brevibacterium incertae sedis paramesenteroides
Corynebacterium incertae Clostridium butyricum
Ganou-
Molasses/ sugar sedis Clostridium beijerinckii
Parfait et al.
cane/Soils/Waters Erysipelothrix Clostridium acetobutylicum
(1989)
Kurthia zopfli Clostridium felsineum
Listeria Clostridium puniceum
Microbacterium lacticum Clostridium
Propionibacterium thermosulfurigenes
acidipropionici Clostridium
Propionibacterium jensenii thermohydrosulfuricum
Propionibacterium Clostridium sporogenes
freeudenreichii Clostridium bifermentans
Saccharomyces cerevisiae
Saccharomyces carlsbergensis
Fahrasmane Saccharomyces chevalieri
Schizosaccharomyces pombe
& Ganou- Molasses and Saccharomyces rouxii
Hansenula anomala
Parfait sugar cane juice Saccharomyces aceti
Torulopsis glabrata
(1998) Saccharomyces microellipsodes
Torulopsis stellata
Saccharomyces delbrueckii
Note: Bacterial species given in the table are shown in bold font

[Link] Yeasts

Yeasts are essential to fermentation of molasses for rum production. Greig’s

(1895) article “The Jamaica Yeast” described a yeast found in Jamaican molasses
 53

used in rum production. This yeast was later described by Jorgensen and Holm as
Schizosaccharomyces mellacei (Jorgensen, 1909). Early studies noted two broad
groups of yeasts associated with rum fermentations, notably, the more prevalent
Saccharomyces yeasts where cell division occurred by budding, and the
Schizosaccharomyces yeasts where cell division occurred by fission. The two
types could be readily distinguished by microscopic examination (Kayser, 1913).

Arroyo (1945) discussed at length the impact of fission yeasts versus budding
yeasts on rum production, but precise details about the species involved and their
growth kinetics were rarely mentioned. Heavy style rums, namely, those more
traditional in production and flavour would require fermentation with a
predominance of Schizosaccharomyces or fission yeast and minor contributions
from budding yeasts. Such fermentations were generally slower, and produced
rums with higher contents of esters, higher alcohols, organic acids and aldehydes
which combine to contribute stronger organoleptic tastes and aromas. For
production of lighter rums, Arroyo (1945a) postulated a faster fermentation using,
budding Saccharomyces yeasts. This type of yeast would produce fewer
congeners such as organic acids, aldehydes, esters etc., thus giving a rum with a
“cleaner” taste. These conclusions were supported by later studies of Parfait and
Sabin (1975) and Fahrasmane et al. (1985).

Parfait and Sabin (1975), while investigating fermentation in the French West
Indies, reported that Schizosaccharomyces spp. were predominant in fermentation
media used for heavy flavour rum production, while Saccharomyces spp. were
predominant in wild fermentation and seeded media. The two different species of
yeast exerted different effects on flavour development during fermentation. Fission
type yeasts such as Schizosaccharomyces pombe produced heavy flavoured
rums, while faster fermenting budding type yeasts such as Saccharomyces
cerevisiae produced light flavoured rums.

Fahrasmane et al. (1985) found that Saccharomyces cerevisiae produced greater

 54

concentrations of higher alcohols and short chain fatty acids than

Schizosaccharomyces strains. S. cerevisiae also fermented the molasses medium
more rapidly than Schizosaccharomyces strains.

[Link] Bacteria

Bacteria can impact favourably and detrimentally to the production of rum. Greig
(1893) was the first to document the presence of bacteria in rum fermentations.
While he did not conduct specific studies on these bacteria, he suggested that their
growth be suppressed by the inoculation of yeast cultures. The possibility of a
positive role of bacteria in rum fermentations was first suggested by Allan (1906).
While investigating Jamaican rum fermentations, Allan (1906) highlighted the
presence of two Bacillus species. He suggested that such bacteria might utilise
dead yeast cells as nutrients at the end of yeast fermentation, and that organic acid
and higher alcohol production by these bacteria would contribute the characteristic
rum flavour.

Several authors (Kampen, 1974; Ganou-Parfait et al, 1987 & 1989; Fahrasmane,
1988) have reported the qualitative presence of bacterial species in molasses rum
fermentations but more detailed information about populations and frequency of
occurrence was scant and inconsistent. Bacterial populations, when reported, were
generally low (102-103 CFU/mL) and reflected a diversity of species within the
genera Bacillus, Brevibacterium, Clostridium, Corynebacterium, Lactobacillus and
Propionibacterium (Ganou-Parfait et al, 1987 & 1989; Fahrasmane et al, 1988).

Clostridium saccharobutyrium (isolated from sugar cane bagasse) has been shown
to have beneficial effects by increasing the rate of formation of alcohol during yeast
fermentation of molasses (Fahrasmane et al, 1988). Also, these workers proposed
that Clostridium species found in rum distilleries could contribute to the production
of volatile acids such as acetic, butyric, caproic, heptanic and propionic acids.
These acids are important precursors for ester formation that is essential for rum
 55

aroma and flavour.

Ganou-Parfait et al (1989) investigated the bacterial species associated with rum

production. This research comes the closest to a complete survey based on the
four potential origins of microflora postulated by Owen (1949). The study examined
the source of the bacteria (cane, molasses, water, and soils), their optimum growth
temperatures, optimum pH, optimal growth period during fermentation, metabolic
products and effect (either positive or negative) on the yield of rum and its
organoleptic qualities. It was suggested that the development of heavy rum
aromas relies significantly on the presence and activity of bacterial species. The
presence of propionic acid bacteria, such as Propionibacterium jensenii, is
associated with the production of propionic acid. High levels of this acid
differentiate rum from other distilled alcoholic beverages (Ganou-Parfait et al,
1989). High populations of propionic acid bacteria can, however, be detrimental to
flavour due to the production of highly acidified rum (Fahrasmane & Ganou-Parfait,
1998). Lactic acid bacteria, such as Lactobacillus species, produce metabolites
which can acidify the medium and develop more complex organoleptic properties
such as aldehydes, esters and diacetyl (Jay, 1982). Some lactic acid bacteria can,
however, be detrimental to finished rum by producing dextrins and other
polysaccharides during sugar processing and molasses storage. These bacteria
utilise sugars (e.g. metabolise sucrose to polysaccharides) that otherwise would
have been fermented to produce ethanol and other flavour volatiles.

From the limited research conducted to date, it is evident that bacteria are part of
the microbial ecology of rum fermentation. Their potential impacts can be
summarized as: positive and unique contributions to rum flavour; detrimental
effects on rum flavour; and reduction in process efficiency by modulating yeast
growth and their production of ethanol. The extent to which they will contribute
positively or negatively to the process will be determined by the species that are
present and their ability to compete with the growth of yeasts. More research is
needed to better define and understand the role of bacteria in rum fermentations
 56

and will be a focus of this thesis.

2.4.3 Microbiology of Dunder

The use of dunder in rum production has been described in Section [Link]. This
section considers its microbiology.

Most definitions accept dunder as the material remaining in the still after distillation
of the fermented molasses (Section [Link]). Yeast and bacterial cells associated
with the fermentation will be dead as they would have been inactivated by the heat
(>80°C) during the distillation process. Consequently, dunder originating directly
from the still should be sterile (Kampen, 1975). If it is subsequently stored, it is
likely to become contaminated and support microbial growth, the extent of which
will depend on the conditions of storage such as time, temperature and hygiene of
the environment. There are several reports (Wustenfeld & Haeseler, 1953; I’Anson,
1971; Murtagh, 1995b) that refer to the contamination and growth of
microorganisms in stored dunder, but no scientific evidence has been presented to
support these claims or to provide data about the species and populations of
microorganisms that might be present. In some cases, such growth in dunder is
encouraged to serve as a source of unique microflora for inoculation into the
molasses fermentation (Wustenfeld & Haeseler, 1953; I’Anson, 1971; Murtagh,
199b, Nicol, 2003)

Given the wide spread use and perceived importance of dunder in rum production,
more research is needed about its microbiological status and will be examined in
Chapter 5 of this thesis.

2.4.4 Use of Starter Cultures in Molasses Fermentation for

Rum Production

As mentioned in Section 2.1.1, rum fermentation is a spontaneous process

 57

conducted by indigenous yeasts and bacteria originating from uncontrolled “wild”

fermentations. The concept of using starter cultures for rum fermentations has
been considered since the late 1890s. Arroyo (1945a) gives an historical account
of the early studies concerning the use of starter cultures in rum production.

The notion of using starter cultures to better control alcoholic fermentations is well
established in the brewing and wine industries (Fleet, 1998; Mateo et al, 2001).
The ability to control such fermentations meant that greater batch to batch
similarity was achieved, giving products which remained consistent in flavour and
aroma. This led to greater consumer loyalty to particular brands and, consequently,
increased profits. The rum industry, with the exception of small artisanal distilleries,
has followed other fermented beverages towards more commercially driven
production capabilities. The production of clearer, lighter rums came as a direct
result of the starter culture era. Nevertheless, the philosophy to rum production
became divided over the use of yeast starter cultures.

Pairault (1903) was the first to suggest that starter culture yeasts for rum
production should be selected on the basis that they are well adapted to growth
and fermentation in molasses. This concept was further supported by Kayser
(1913), who was behind the push for pure culture fermentations containing only the
selected yeasts. Both Pairault (1903) and Kayser (1913) recognized that bacteria
were also endemic to many rum processes but were of the view that they impacted
negatively on production efficiency and quality. Consequently, use of yeast starter
cultures would overcome these issues. Prompted by these researchers, various
distilleries isolated and identified the main yeast strains responsible for their
fermentations, and these included strains of [Link] and Schiz. pombe. The
prevalence of either (or both) of these species of yeast in fermentations saw many
distilleries adopt them as starter cultures (Arroyo, 1945a).

During the 1970s, rum distilleries commenced using dried baker’s yeast as a
cheap, readily available and easily stored alternative to laboratory maintained
starter cultures. Fermentation efficiency was boosted and a more consistent
 58

product was able to be produced. The use of dunder was no longer required for
these types of fermentations due to baker’s yeasts sensitivity to the high acidity
provided by dunder. The advent of lighter, clear rums was, thereby, initiated
(Fahrasmane & Ganou-Parfait, 1998).

While many considered yeasts to be the main agent of rum fermentation and the
main target for use as starter cultures (Arroyo, 1945a), there were other
researchers who considered that bacteria were also important to the fermentation.
Allan (1906) and Ashby (1909) studied Jamaican rums and advocated the
importance of bacteria in wild type fermentations for the development of the
complex rum aroma found in heavy bodied rums of the region.

Rocques (1927) showed that, while the alcohol yield and efficiency of production
increased with the use of yeast starter cultures, the finished rum was low in acid
and ester concentrations, high in higher alcohols, and consequently lacked the
characteristic flavour and aroma attributed to rum. Such issues caused many
distilleries to revert to older practices of wild, uncontrolled fermentations.

To accommodate the concern about bacterial contributions to rum fermentation,

Arroyo (1942, 1945a, b) suggested three possible approaches to the use of yeast
starter cultures. These were; use with a sterilised molasses medium in aseptic
fermenters (a completely controlled environment); use with a partially sterilised
fermentation medium; or use with a non-sterilised fermentation medium. The
aseptic fermenter would give results such as those obtained by Rocques (1927),
containing ethanol but fewer distinguishing “rum-like” aromas. Partial sterilisation of
fermentation media may lead to the growth of any bacterial species present,
depending on the length of fermentation. The final scenario, the non-sterilised
fermentation media, would rely on the selected starter culture out-competing the
natural microflora already present and should maximize the growth of the bacterial
species found in rum fermentations and their contribution to more complex
characteristics of rum flavour (Arroyo, 1945c).
 59

Table 2.9 outlines desirable traits that rum distillery yeasts should possess for
maximum efficiency during fermentation, and these have been used to develop
specific starter cultures for rum production. Several companies sell distillers yeasts
for rum production and these are mentioned in Section [Link].

Table 2.9 Properties required by rum yeasts to be commercially

successful as a starter culture
Properties
 Conduct strong rapid alcoholic fermentation of molasses ( grow at
appropriate pH , Aw , temperature, give high ethanol production)
 Initiate rapid growth with minimal lag phase so as to outcompete
indigenous yeasts and bacteria
 Ferment molasses to give desirable sensory metabolites consistent
with good rum characteristics
 Able to grow under large scale propagation protocols
 Be tolerable to preservation by freezing and lyophilisation
 Must not be pathogenic
 Must not form toxins or antibiotics

2.4.5 Microbiology of Other Distilled Beverages; Cachaça and

Whisky

As mentioned in the previous sections, progress in understanding the microbiology

and biochemistry of rum fermentation has been very limited despite some 100
years of research. In contrast, there has been significant recent research into the
microbiology of cachaça production and whisky production, the developments of
which may guide future research on rum microbiology. These developments are
summarized in this section.

[Link] Microbiology of Cachaça Fermentation

As mentioned previously (Section 2.2.5), cachaça is a rum-like, distilled alcoholic

beverage produced from fermented sugar cane juice and is produced mainly in
Brazil. Sugar cane is crushed to extract the juice which is then immediately
fermented. Fermentation may occur as a spontaneous process from the growth of
 60

indigenous yeasts that occur as natural contaminants, or it may be conducted by

inoculation with a traditional starter preparation, or selected strains of yeasts,
usually, S. cerevisiae. Fermentation usually takes about 24-48 h after which the
fermented material is distilled to give the cachaça product. Details of the process
are discussed by Faria et al (2003).

Research into the microbiology of cachaça fermentations has been steadily

growing in depth and popularity since the 1990s, especially among Brazilian
researchers. Whether conducted by traditional or inoculated processes, the
fermentation is dominated by the growth of S. cerevisiae and this species is
considered to be the key driver of the fermentation. Usually, this species grows to
maximum populations of 108-109 CFU/ml within 36-48 h, after which the
fermentation is terminated. Some key studies on the yeast ecology of the
fermentation are those of El-Tabey Shehata (1959), Morais et al. (1997), Schwan
et al (2001), Pataro et al (2000), de Araujo Vicente et al. (2006) and Duarte et al
(2013). These studies also report the contribution of other yeasts to the
fermentation. Such species are, usually, from the genera Candida, Debaryomyces,
Schizosaccharomyces, Kluyveromyces, Pichia and Hanseniaspora but were
isolated less frequently than S. cerevisiae. They were mostly found during the early
stages of cachaça fermentations after which they died off, giving way to the
dominance of S. cerevisiae

Phenotypic and molecular analyses of isolates of S. cerevisiae from cachaça

fermentations show substantial diversity within the strains examined, suggesting
adaptation to particular process conditions and geographical locations (Badotti et al
2010, 2013). This variation can account for the different flavour profiles of
cachacas obtained from the different distilleries (Badotti et al 2010) and has led to
the selection of particular strains for development as unique starter cultures in
recent years (Oliveira et al, 2004; Gomes et al, 2007; Marini et al, 2009, Campos et
al, 2010).

Although yeasts are primarily responsible for cachaça fermentations, Schwan et al

 61

(2001) also noted the presence of bacteria, especially lactic acid bacteria,
throughout these fermentations. Subsequent studies by Duarte et al (2011)
investigated cachaça fermentations inoculated with a mixture of S cerevisiae and
Lactobacillus fermentum and demonstrated how the flavour profile of cachaça
could be modulated by the growth of bacteria. Consequently, further research is
needed to better understand the potential role of bacteria in cachaça
fermentations.

[Link] Microbiology of Whisky Fermentation

Whisky is produced by fermenting the liquid extract obtained by mashing of malted

barley. This liquid contains the sugars and other nutrients that are metabolised by
microorganisms to produce ethanol and other aromatic compounds. The fermented
extract is distilled, matured and blended to give the final product. Details of the
process have been reviewed by Piggott and Conner (2003) and Walker (2012).
Microbial fermentation of the wort or malt extract is a key operation in the
production chain, and the microbiology of this step has been discussed by
Suomalainen (1971) Priest and Pleasants (1988), Barbour and Priest (1988) Fleet
(1998), van Beek and Priest (2002) and Collicutt (2009 a, b).

Early studies revealed the predominance and importance of yeasts in the

fermentation and key species identified were isolates of Saccharomyces cerevisiae
(Sharp & Watson, 1979). Subsequently, isolates of this yeast were selected,
purified and developed as strains of distiller’s yeasts that became commercially
available for use as starter cultures to guide and better control the fermentation
(Jones, 1998). The predominant strains have been commercially marketed by
Kerry Ingredients and Flavours as “M” and “MX” strains, by Mauri Products as
“Pinnacle” yeast, and by Lallemand Ethanol Technology as DistillaMax©. The
cultures of Kerry Ingredients and Flavours and Mauri Products are used in more
traditional processes to give a fuller bodied product, while Lallemand’s,
DistillaMax© is marketed for the grain whisky or neutral spirit fermentations
 62

(Walker, 2012; Russel & Stewart, 2014).

In recent years, lactic acid bacteria have been found to have a significant role in
whisky fermentation and, depending on how the process is managed, they can
grow to populations of 106-108 CFU/ml after 48-72 h of fermentation. The main
species isolated from such fermentations are Lactobacillus fermentum,
Lactobacillus plantarum, Lactobacillus brevis, Weissella confuse and Lactobacillus
paracasei. The potential impacts of these bacteria on the fermentation have been
considered as:(i) a decrease in ethanol production by the yeasts due to sugar
consumption by the bacteria;( ii) an enhancement of whisky flavours due to
production of higher alcohols and other congeners; and (iii) a detrimental effect if
their growth is unregulated, leading to stuck alcoholic fermentation and the
production of “off” flavours (Simpson et al, 2001; van Beek & Priest, 2003; Cachat
& Priest, 2005; van Beek & Priest, 2000, 2001 & 2003).

Although cachaça and whisky fermentations were once thought to be

predominantly processes attributed to yeasts, principally strains of S. cerevisiae,
there is increasing evidence that bacteria, especially lactic acid bacteria, may
positively contribute to these fermentations. Although cachaça, whisky and rum
production are based on different raw materials, the microbial ecology of the
processes has many similarities. On this basis, it is likely that bacteria, especially
lactic acid bacteria, may play a more significant role in rum production than
previously thought.
 63

2.5 SUMMARY
A systematic investigation into the microbial species which impact on the
processes involved in the manufacture of rum will be undertaken. Rum is produced
by microbial fermentation of molasses, a waste product generated by the cane
sugar industry. The rate and extent of this fermentation determines process
efficiency. Several yeast and bacterial species contribute to rum fermentation but
the related ecology is poorly defined and understood. Through a combination of
ecological studies, controlled fermentations and distillations, and chemical
analysis, the impact of particular microbial species on rum flavour will be
determined.
 64

CHAPTER 3.
THE MICROBIAL ECOLOGY OF A PROCESS FOR
RUM PRODUCTION

3.1 INTRODUCTION

Rum is a distilled alcoholic beverage obtained from the fermentation of sugar cane
molasses or sugar cane juice (Arroyo, 1945a; Clutton, 1974; Kampen, 1975;
Lehtonen & Suomalainen, 1977; Bluhm, 1983 and Nicol, 2003). In contrast to many
other alcoholic beverages such as beer (Iserentant, 1995), wine (Fleet, 1998;
Mateo et al, 2001) and distilled products such as whisky (Suomalainen, 1971;
Priest & Pleasants, 1988; Barbour & Priest, 1988; Fleet, 1998; van Beek & Priest,
2002 and Collicutt, 2009b), the microbial ecology of the process has not been
extensively studied. The little information known about the microbiology of rum
fermentation has been summarised in reviews by Lehtonen & Suomalainen (1977)
and Nicol (2003). It is concluded that the yeast, S. cerevisiae, is the key organism
that conducts the alcoholic fermentation of molasses and it has been developed as
a starter culture for this purpose but, in some processes, indigenous strains of
Schizosaccharomyces pombe may also make a contribution (Hall, 1935; Parfait &
Sabin, 1975; Ganou-Parfait et al, 1989; Bonilla-Salinas et al, 1995; Todorov &
Dicks, 2004). A few studies have reported the presence and significance of
bacteria during rum fermentations but details about their frequency of occurrence
and growth during such fermentations have not been investigated (Kampen, 1974;
Ganou-Parfait et al, 1987 & 1989; Fahrasmane et al, 1988). Such species include
Clostridium saccharobutyricum (Fahrasmane et al, 1988), Propionibacterium
jensenii (Ganou-Parfait et al, 1989) and lactic acid bacteria (Kampen, 1975;
 65

Ganou-Parfait et al, 1989). It has been suggested that these bacteria may
contribute positively or negatively to the quality of the final product, but definitive
conclusions about their contributions require more detailed investigation. Bacterial
populations, when reported, were generally low (102 – 103 CFU/mL).

To provide a greater knowledge about the microbiology of the rum fermentation

process, this Chapter aims to systematically study the microbial ecology of a rum
distillery located at Bundaberg, Queensland, Australia; from raw materials
(molasses, dunder, yeast and water) through propagation and fermentation.

3.2 MATERIALS AND METHODS

3.2.1 Process Outline and Sample Selection

The process of rum production at the Bundaberg distillery, Queensland, was

investigated. This process is outlined in Figure 3.1 and is described in detail in
Section 3.3.1 of the Results section.

Samples for microbiological examination were taken at the sites listed in Table 3.1
and shown in Figure 3.1. They were taken during commercial operation of the
facility over a four year period (March 2006 to April 2010). Samples (50 mL) were
taken in duplicate under aseptic conditions and stored at 4°C until microbiological
analysis, within 24 h. Some sites had fixed sampling ports that were sterilized by
flushing with 70% ethanol, after which sample was then flushed through the port in
order to obtain a representative fraction. Some samples were collected using 10
mL syringe - vaccutainers (Becton Dickinson). In these cases, samples from three
vaccutainers were transferred into sterile plastic containers (Sarstedt) and mixed to
produce a representative 30 mL sample.
 66

Table 3.1 Sampling sites for microbiological study of a rum distillery

Stage/Sampling point
Molasses preparation
Molasses supply
storagetank
tank
Dunder
Dunder and mud line
Dilution tank
Clarifier
Surge tank (50° Brix)
Surge tank (30° Brix)
Floc addition
Town water
Yeast propagation in molasses medium
Yeast
m0llaspropagation Vessel A
molasses medium (early)
vessel 1
Yeast propagation Vessel A (late)
Yeast propagation Vessel B before transfer to Yeast vessel C
Yeast propagation Vessel C before transfer to Yeast vessel D
Yeast vessel D prior to transfer to Fermenter
Molasses fermentation
Fermenter sample post transfer of yeast (0 h)
Fermenter sample (6 h)
Fermenter sample (12 h)
Fermenter sample (18 h)
Fermenter sample (24 h)
Fermenter sample (36 h)
Fermented molasses in buffer tank A
Fermented molasses in buffer tank B

3.2.2 Isolation, Enumeration and Identification of

Microorganisms
The microbial flora of samples was examined by (1) culture plating on agar media;
(2) enrichment culture followed by plating on agar media

[Link] Yeasts
The samples were serially diluted in 0.1 % Bacteriological Peptone Water, and 0.1
mL spread inoculated, in duplicate, onto plates of Malt Extract Agar (MEA) (Oxoid)
and Wallerstein Differential Nutrient Agar (WL) (Oxoid) supplemented to contain
100 µg/mL of oxytetracycline (Sigma) to restrict bacterial growth. Plates were
incubated at 25°C for 48 h, after which time colonies were counted. Predominant
colony morphologies were noted and approximately five representative isolates of
each type were purified by streaking onto plates of MEA. Stock cultures were
stored at -80°C under 30 % glycerol until used for identification.
 67

To determine the presence of low populations of yeasts in some samples,

subsamples (1 mL) were subject to enrichment culture in 50 mL of Malt Extract
(ME) broth (Oxoid) for 24-48 h at 25°C. The cultures were then tested for the
presence of viable yeasts by streak plating samples onto MEA.

Extraction and sequencing of ribosomal DNA

DNA was extracted from disrupted cells by a modification of the method described
by Kowalchuk et al (1997). Yeast isolates were grown overnight in ME broth.
Samples (1-2 mL) of this culture were centrifuged (2 min at 10,000 g) at room
temperature in a Beckman Microfuge 18 Centrifuge (Beckman Coulter Inc.,
Fullerton, CA, USA) to obtain a cell pellet. The cell pellet, contained within a 2 mL
screw capped microtube, was resuspended in 0.2 mL of extraction buffer (100 mM
Tris-HCl [pH 8], 50 mM EDTA [pH 8], 100 mM NaCl, 1% sodium dodecyl sulphate
[SDS]), to which was added 0.3 g zirconia/silica beads (diameter 0.5 mm; Daintree
Scientific, Tasmania, Australia), and 0.2 mL of phenol:chloroform:isoamyl alcohol
(24:24:1 (v:v:v) Sigma-Aldrich, Australia). The samples were shaken at 500 rpm for
1 min in a mini-bead beater (Biospec Products, OK, USA). After centrifugation for
10 min (15,000 g at 4°C), 0.4 mL of the upper layer was removed. DNA was
precipitated by addition of 0.4 mL of isopropanol and the mix allowed to stand for
24 h at -20°C. After centrifugation for 10 min (15,000 g), the pellet was washed
once with 70% ethanol and allowed to air dry. The dried pellet was dissolved in 50
µL of TE buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8]) and stored at -20°C until use
in PCR.

Extracted DNA was used as template DNA in the polymerase chain reaction (PCR)
to amplify sections of the 26S ribosomal RNA region. The extracted DNA was
amplified by PCR using universal primers NL1 (5’-
GCATATCAATAAGCGGAGGAAAAG-3’) and NL4 (5’-
GGTCCGTGTTTCAAGACGG-3’) for identification by partial 26S rDNA sequence
analysis (Kurtzman and Robnett, 1998). These primers were obtained from Sigma
Genosys, NSW, Australia.
 68

The PCR reaction mixtures contained 10 mmol l-1 Tris HCl (pH 8.3), 50 mmol l-1
KCl, 0.2 µmol l-1 of each primer, 200 µmol l-1 of each dNTP (Roche Diagnostics,
Indianapolis, IN, USA), 1.0 mmol l-1 MgCl2, 1.25 U of Gold Taq DNA Polymerase
(AmpliTaq™, Roche Molecular Systems, Branchburg, NJ, USA) and 10 ng of
purified template DNA in 50 µl final volume. Amplification was performed under the
following programme: initial denaturation at 95°C for 7 min, 36 cycles at 95°C for 1
min (denaturation), 52°C for 2 min (annealing) and 72°C for 2 min (extension), with
the final extension conducted for 10 min at 72°C. Confirmation of amplicons was
done by agarose gel electrophoresis after which they were used for sequencing
with the ABI PRISM® BigDye™ Terminators v3.1 Cycle Sequencing kit (Applied
Biosystems, Foster City, CA, USA). The products were sequenced at the
Ramaciotti Centre for Gene Function Analysis, UNSW, Australia. The resulting
sequences underwent DNA similarity searches with the NCBI Blast program using
sequences retrieved from the Genebank Database (Karlin & Altschul, 1990).

[Link] Bacteria
Total bacteria and lactic acid bacteria
The samples were serially diluted in 0.1 % Bacteriological Peptone Water, and 0.1
mL spread inoculated, in duplicate, onto plates of de Man, Rogosa and Sharpe
agar (MRS) (Oxoid), Wallerstein Differential Nutrient Agar with Supplement (WLS)
(Oxoid), Plate Count Agar (PCA) (Oxoid) and Raka Ray (Oxoid) agar, each
supplemented to contain 10 µg/mL of cycloheximide (Sigma) to restrict yeast
growth. Plates were incubated either microaerophillically and anaerobically at 30°C
for 48 h, after which time colonies were counted. Predominant colony
morphologies were noted and five representative isolates of each type were
purified by streaking onto plates of MRS or WLS agar. Stock cultures were stored
at -80°C under 30% glycerol until used for identification.

To determine the presence of low populations of bacteria in some samples,

subsamples (1 mL) were subject to enrichment culture in 50 mL of MRS broth
 69

(Oxoid) for 24-48 h at 30°C. The cultures were then tested for the presence of
viable bacteria by streak plating samples onto MRS agar.

Zymomonas and Propionibacterium

The samples were serially diluted in 0.1 % Bacteriological Peptone Water, and 0.1
mL spread inoculated, in duplicate, onto plates of Sodium Lactate Agar (SLA) and
Universal Beer agar (modified for rum fermentations), each supplemented to
contain 10 µg/mL of cycloheximide (Sigma) to restrict yeast growth. SLA was made
according to Atlas (2006) with the composition per litre consisting of: agar 15.0 g,
Pancreatic digest of casein 10.0 g, sodium lactate 10.0 g, yeast extract 10.0 g and
K2HPO4 0.25 g. Plates were incubated at 30°C under anaerobic conditions for 5-7
days. Universal Beer agar (Oxoid) was prepared according to directions provided
by Oxoid with the addition of 250mL/L of either dunder or fermentation medium to
batches. Plates were incubated anaerobically for 3-7 days. Anaerobic conditions
were obtained using AnaeroGen© (Oxoid) in appropriate sealed containers.

Clostridium
The samples were serially diluted in 0.1% Bacteriological Peptone Water, and 1
mL was used to prepare duplicate pour plates of Differential Reinforced Clostridial
Agar (RCA) (Difco). The plates were incubated anaerobically at 30°C for 2-10 days
(plates were checked for growth every 48 h). Anaerobic conditions were obtained
using AnaeroGen© (Oxoid) in appropriate sealed containers.

Identification of bacterial isolates.

Bacteria were identified by a combination of phenotypic methods and sequencing
of the 16 S ribosomal DNA.

Phenotypic tests
Phenotypic characterization included microscopic examination for cell morphology,
Gram staining, and tests for oxidase. Isolates were then selected for identification
using API CHL50 test strips (Biomerieux, Durham NC). Cultures were grown at
30°C for 24 h on MRS agar prior to collection and suspension of cell biomass in
 70

sterilised distilled water according to kit instructions. API kits were inoculated with
the biomass and incubated at 30°C for 24 h and observed for reactions. Kits were
incubated for a further 24 h, reactions recorded and data processed using
APIweb™ ([Link] to give genus and species identification.

Extraction and sequencing of ribosomal DNA

DNA was extracted from disrupted cells by a modification of the method described
by Kowalchuk et al (1997). Bacterial isolates were grown overnight in MRS broth.
Samples (1-2 mL) of this culture were centrifuged (10-15 min at 15,000g) in a
Beckman Microfuge 18 Centrifuge (Beckman Coulter Inc., Fullerton, CA, USA) to
obtain a cell pellet. To lyse the cell pellet, 0.3 g zirconia/silica beads (diameter 0.1
mm; Daintree Scientific, Tasmania, Australia), 0.5 mL of extraction buffer (100 mM
Tris-HCl [pH 8], 50 mM EDTA [pH 8], 100 mM NaCl, 1% sodium dodecyl sulphate
[SDS]) and 0.5 mL phenol:chloroform:isoamyl alcohol (24:24:1 (v:v:v) Sigma-
Aldrich, Australia) were mixed in a 2 mL screw capped microtube. The samples
were shaken at 5000 rpm for 30 s in a mini-bead beater (Biospec Products, OK,
USA). After centrifugation for 10 min (15,000 g at 4°C), 0.4 mL of the upper layer
was removed. DNA was precipitated by addition of 0.6 mL of isopropanol and the
mix allowed to stand for 24 h at -20°C. After centrifugation for 10 min (15,000 g),
the pellet was washed once with 70% ethanol and allowed to air dry. The dried
pellet was dissolved in 50 µL of TE buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8])
and stored at -20°C until use in PCR.

Extracted DNA was used as template DNA in the polymerase chain reaction (PCR)
to amplify sections of the 16S ribosomal RNA region. The extracted DNA was
amplified by PCR using universal primers for the 968 (968f 5’-
AACGCGAAGAACCTTAC) and 1401 (1401r 5’-CGGTGTGTACAAGACCC)
region, Escherichia coli numbering (Bae et al 2004). These primers were obtained
from Sigma Genosys, NSW, Australia. The PCR reaction mixtures contained 10
mmol l-1 Tris HCl (pH 8.3), 50 mmol l-1 KCl, 0.2 µmol l-1 of each primer, 200 µmol l-1
of each dNTP (Roche Diagnostics, Indianapolis, IN, USA), 1.5 mmol l-1 MgCl2, 1.25
 71

U of Gold Taq DNA Polymerase (AmpliTaq™, Roche Molecular Systems,

Branchburg, NJ, USA) and 10 ng of purified template DNA in 50 µl final volume.
Amplification was performed under the following programme: initial denaturation at
95°C for 7 min, 10 cycles at 94°C for 30 s (denaturation), 50°C for 30 s (annealing)
and 72°C for 45 s (extension), followed by 20 cycles of 94°C for 30 s
(denaturation), 50°C for 30 s (annealing) and 72°C for 30 s (extension), with the
final extension conducted for 10 min at 72°C. Confirmation of amplicons was done
by agarose gel electrophoresis after which they were used for sequencing with the
ABI PRISM® BigDye™ Terminators v3.1 Cycle Sequencing kit (Applied
Biosystems, Foster City, CA, USA). The products were sequenced at the
Ramaciotti Centre for Gene Function Analysis, UNSW, Australia. The resulting
sequences underwent DNA similarity searches on NCBI Blast program using
sequences retrieved from the Genebank Database (Karlin & Altschul, 1990).

3.2.3 Electron Microscopic Examination

Pure cell cultures, grown in MRS broth (bacteria) or ME broth (yeast) were
sedimented by centrifugation (5 minutes at 1 000g). The supernatant was removed
and cell pellets were fixed in 5% (v/v) glutaraldehyde in PBS (0.1M, pH 7.2)
overnight at room temperature. Cell pellets were then washed three times for 10
min each with PBS (0.1M, pH 7.2). Coverslips (12 x 12 mm) were coated with 0.1%
ethylene imine polymer solution (Fluka, Switzerland). Cell suspensions (20 µL)
were added to the coverslips and cells were left to adhere to coverslips for 15 min
at room temperature. The coverslips were then washed in buffer three times for 10
min each (0.1M phosphate buffer, pH 7.2) and post fixed in 1% osmium tetroxide
(OsO4) in 0.1M PBS (pH 7.2) for 1 h. Following post-fixation, the samples were
washed in the same buffer in triplicate for 5 min each and then dehydrated through
a graded series of ethanol of 30%, 50%, 70%, 90% and 100% for 10 min each.
The samples were then transferred to a critical point dryer (Emitech K850
(ProSciTech, Australia)). The coverslips were then mounted on aluminium stubs
and sputter coated with gold (Emitech K550). Slides were observed using a JEOL
JSM 7100 Field Emission Scanning Electron Microscope. This work was done in
 72

conjunction with the Microscopy Unit, Department of Biological Sciences,

Macquarie University, NSW, Australia.

3.3 RESULTS

3.3.1 Outline of rum production process

Production at the Bundaberg distillery, generally runs 24 hours a day, 6 days a

week. It involves sequential operations that are outlined in Figure 3.1. Details of
these operations are given in the following sections. Figure 3.2 gives a pictorial
presentation of some facilities in the Bundaberg Distillery. Figure 3.3 shows
photographs of molasses and dunder.

Figure 3.1 Flow diagram of rum production process showing sites of sampling for
microbiological analysis as given in Table 3.1
 73

Figure 3.2 Process steps at the Bundaberg Distilling Company. (a) Cane harvest, (b)
sugar mill, (c) molasses, (d) molasses storage tank, (e) yeast vessel A, (f) yeast vessels, (g)
small fermenters, (h) large stainless steel fermenters, (i) buffer tank, (j) pot still, (k)
continuous still, (l) maturation vat, (m) bottling.
 74

[Link] Molasses and Molasses Preparation

Molasses is the main raw material of the rum production process. It is purchased
from an adjacent sugar mill, delivered hot (> 60°C) to the distillery via pipelines and
stored in large open tanks (approximately 25m x 15m x 10m) within covered
warehouses (Figure 3.2 c, d). The tanks are constructed of concrete. The
molasses may remain in these tanks for up to 10 months or longer to provide a
constant supply for rum production. The molasses is stored, without temperature
control, at ambient temperature that ranges from 20-40°C according to season.
Laboratory staff analyse the molasses for fermentable sugars (sucrose, fructose
and glucose), total sugars, pH and °Brix on a daily basis. Specific gravity is
measured twice weekly. Further chemical analysis is performed sporadically by an
external laboratory.

As needed, molasses is pumped from the storage tanks to a clarification vessel

where it is diluted with water and dunder to give a final °Brix value of 45-50°.
Dunder is added to reach a final concentration of 7.5% of the fermentation medium
volume. Flocculant (HISET P730SP, Dai-ichi Kogyo Seiyaku) is added at a ratio of
500 mL per 300 L through a venturi mixer to facilitate the sedimentation of mud
and unwanted particulate matter. The mixed materials are heated to 70-80°C and
allowed to “stand” for approximately 1h during which the solids sediment. The
clarified molasses mixture is then pumped through a heat exchanger to cool it to
35°C. The Brix is further reduced from 50° to 30° by the addition of potable water
and the mix is transferred into a surge tank. From here, the molasses, water and
dunder mixture (now called fermentation medium) is used to supply the yeast
propagation vessels and the main rum fermentations vessels.

Usually, the dunder is added directly from the stills at 100°C. Routine
microbiological testing by distillery laboratory staff has shown it to be sterile - viable
organisms not detected in 1.0 mL of sample.
 75

Ammonium sulphate, magnesium sulphate and ammonium dihydrogen phosphate

are added in early propagation steps to provide additionally nutrients for yeast
growth.

Figure 3.3 Raw materials used in the production of rum; molasses (left), dunder (right).
Sourced from Bundaberg Distilling Company, Bundaberg Australia

[Link] Preparation of Yeast

A selected strain of the yeast Saccharomyces cerevisiae is used to conduct the

fermentation. The strain is stored in several commercial culture collections
worldwide. Every 6-12 months, a fresh culture is acquired from these collections
from which a pure culture is prepared on MEA agar. It is maintained on slopes of
this medium for the next 3 months as the operational culture to propagate yeast
inocula for the main rum fermentations. Propagation is a sequential process that
aims to adapt the yeast to the rum fermentation medium and to progressively
amplify the biomass for large scale fermentation. This process is outlined in Figure
3.1. The early stages of the process (up to and including vessel B) use molasses
medium that has been sterilized by heating at 121°C for 20 min. However, medium
 76

sterilization was not possible in the later, larger scale stages. In these cases, the
molasses medium, as prepared for the rum fermentation stage, was used.
Fermentation medium in the early stages contained added ammonium salts to
boost assimilable nitrogen content. The commercial product Fermaid (Lallemand
Inc., Canada) was added to the medium in the later stages as an additional source
of assimilable nitrogen and vitamins. At each stage of the process, yeast cell
counts need to reach 105-106 cells/mL before transfer to the next stage and these
are monitored by microscopic analyses using a haemcytometer. Such analyses
also gave an indication of culture purity.

Yeast biomass from the slope is used to inoculate 250 mL of molasses

fermentation medium (15°Brix). This culture is incubated for 24 h at 30-32°C after
which it is transferred to 2 L of fermentation medium (24°Brix) and incubated at 30-
32°C for a further 24 h. This culture is then used in a four stage propagation
process whereby it is sequentially inoculated into vessels labelled A to D (Figure
3.1). This propagation sequence is summarised in Figure 3.4.

Figure 3.4 Yeast propagation steps undertaken at the Bundaberg distillery

 77

Yeast vessel A consists of 200 L of sterilised fermentation medium (121°C for 20

min at 150 KPa) and is incubated for 12-14 h. The entire volume of this vessel is
then transferred, into 400L of sterilised fermentation medium (23°Brix, 32°C) in
vessel B. This culture is incubated for 2-3 h, after which it is transferred to vessel
C. Incubation in vessel C (3400 L at 15°Brix, unsterilised fermentation medium) is
conducted for 1 hour before half of the volume (1700 L) is transferred into vessel D
(total volume 3400 L). Conditions in yeast vessel D are similar to those in yeast
vessel C. After 1 hour in yeast vessel D, the entire volume is transferred into a
fermenter from which samples are used to inoculate the main fermentation tanks.
Propagation only lasts for up to 1 hour before being transferred into a fermentation
tank.

[Link] Fermentation

Fermentation medium at 28-32 °Brix is pumped into the fermentation tanks which
range in size from small (41000 L) to large (79000 L) tanks (Figure 3.1 and 3.2
g&h). The fermentation medium is mildly agitated during the filling stage (lasts up
to 6 h) as the fermenters are filled from the top by pipes that are set against the
walls of the tanks and distribute the medium in such a way that it flows in a vortex
like manner. This filling operation provides sufficient aeration of the medium to
allow a rapid start of the fermentation. Yeast from propagation vessel D is
inoculated into the fermentation tanks to give an initial population of approximately
105-106 cells/mL and an initial °Brix of 28–32. Fermentation is conducted at 30-
35°C for 36 h without additional aeration.

[Link] Buffer Tanks

When fermentation is complete, fermented medium (now called “wash”) from the
various fermentation vessels is pumped to one of two 90 000L tanks called buffer
tanks (A or B) (Figures 3.1 and 3.2i). These tanks serve to mix the wash from the
different fermenters and to provide a continuous supply of fermented wash to the
still house. In practice, wash is continuously removed from the buffer tanks but this
is not necessarily synchronised with inflow of new wash from any particular
 78

fermenter. Removal of wash from the buffer tanks to the distillation units can stop
and start depending on the distillation needs. As a consequence, there is a mixing
of new wash with wash already in the buffer tank and, therefore, it is very difficult to
know the proportion of newly fermented wash to any fermented wash that was
already present in the tank for some period of time. Adding to this imprecision was
the irregularity of the cleaning and sanitation of the buffer tanks which was
sporadic and, generally, was only done when distillery technicians gauged it
necessary due to changes in distillation efficiency. A cessation of distilling
operations could cause fermented wash to remain in the buffer tanks for prolonged,
indefinite, periods of time.

[Link] Distillation, Maturation and Packaging

Distillation separates, concentrates and selects for desired volatile compounds

produced during fermentation. A two stage distillation process was used at the
Bundaberg Distillery. This consists of an initial continuous distillation process,
followed by a pot distillation process. Fermented wash from the buffer tanks is
pumped to the top of the column or “continuous” still (Figure 3.3k) and flows down
through several levels of horizontal trays. Steam is used to heat the liquid in the
bottom of the column. Alcohol vapour emerges from the top of the column in a
steady stream. The vapour is condensed and the distillate is stored in a “receiver”
until there is sufficient volume to be transferred to a pot still (Figure 3.2j). The
collected liquid is referred to as “low wines”. Once the pot still is filled, steam is
used to heat the liquid in the bottom of the still. The pot still operates similarly to a
kettle, with vapour leaving the pot still and condensing to yield a distillate that is
transferred to a vessel called the raw rum receiver. The raw rum distillate is then
stored in oak vats (Figure 3.2l) for a period of at least 2 years for maturation. After
maturation, the rum is packaged for sale (Figure 3.2m).
 79

[Link] Cleaning and Sanitation

Due to the potentially hazardous environment of volatile vapours, physical

cleaning, such as scrubbing and other abrasively dependent cleaning methods
were only used in some locations and usually only once the plant had been shut
down. These shut downs happened sporadically, usually when efficiency had
decreased significantly, and could last anywhere between a couple of hours and
days. As a result, most cleaning and sanitation on site was clean-in-place (CIP).
CIP cleans consist of computer controlled, highly pressurised nitric acid, caustic,
sanitiser (when required) and water rinses. The acid and caustic cleans both occur,
at approximately 74 - 77°C, with rinses in between, and the total time for the CIP
clean is approximately 2h 45min.

There are cleaning schedules for the Bundaberg distillery. Depending on which
piece of equipment or area is needed to be cleaned there are daily, weekly,
monthly or ‘as required’ cleans. For example, pipe work relating to the emptying of
yeast vessel D to fermenters may be cleaned frequently, (daily) depending on
frequency of filling. Yeast vessels are cleaned at the completion of each
propagation step, in preparation for the next set of propagation steps, usually daily.
Each yeast vessel has an isolated CIP circuit so they can be cleaned individually of
each other. Cleaning includes a combination of acid, caustic and sanitiser.
Fermenters and associated pipes are cleaned between fermentations,
approximately every 36-40 hours. Buffer tanks are rinsed with water weekly
however are only thoroughly cleaned sporadically, when a build-up of fermentation
sediment occurs and the buffer tank capacity is reduced. The heat exchanger used
to heat the fermentation medium post clarification was cleaned only 2-3 times a
year and usually only when efficiency is severely reduced (as determined by the
production engineers, on a case by case basis, established on expected
production efficiency levels).
 80

3.3.2 Ecological Investigation of Yeasts and Bacteria throughout

the Rum Production Process.

Samples of materials were systematically collected throughout the process chain

and examined for their populations of yeasts and bacteria. Predominant yeasts and
bacteria were isolated and identified to genus and species. The first part of this
section describes the occurrence and growth of yeasts throughout the process,
and the second part shows the occurrence and growth of bacteria along the
process chain. Results presented in these sections are the means of data obtained
from the analyses of duplicate biological samples that were each analysed for
microbial populations by plating in duplicate on the appropriate culture medium.

[Link] Identification of yeasts

Throughout the course of this investigation, five yeast species were isolated on
either the MEA or WL agar as evident from their different colony morphologies.
These species were Saccharomyces cerevisiae, Zygosaccharomyces rouxii,
Schizosaccharomyces pombe, Zygosaccharomyces bailii and Ogataea
thermomethanolica. Their colony and cellular morphologies as well as identification
by sequencing of the rDNA are given in Table 3.2.
 81

Table 3.2. Morphology and sequence identification of yeasts associated with rum production

Colony Light Species

Accession Confirmed
morphology on microscopy identification by a
number identity
MEA images rDNA sequencing

Saccharomyces
cerevisiae DQ466538.1 100

Schizosaccharomyces
pombe DQ466539.1 99

Ogataea
thermomethanolica
AB200287 89

Zygosaccharomyces
rouxii AJ783431.1 100

Zygosaccharomyces
bailii
GI4038807 100

a
confirmed identity by sequencing of the 26S rDNA (% match according to the BLAST database)

Overall, S. cerevisiae was the main yeast isolated and corresponded to the starter
culture used by the company to conduct the rum fermentations. Figure 3.5 shows a
scanning electron micrograph of this yeast.
 82

Figure 3.5 Scanning electron microscope image of Saccharomyces cerevisiae isolated

consistently from the rum distillery (Electron Microscope images obtained at Macquarie
University Microscopy Unit, NSW, Australia).

Yeasts in molasses.

Over a four year period, random samples of molasses were taken and examined
for the presence of yeasts by culture on plates of MEA and WL agar (Table 3.3).
These samples were taken directly from the storage wells or from holding tanks
just prior to clarification. Yeast counts were very low at less than 1000 CFU/mL
and, on some occasions, no viable yeasts could be detected (< 5 CFU/mL).

Saccharomyces cerevisiae was the most frequently isolated species, being found
in 11 out of the 16 samples followed by Schizosaccharomyces pombe in six out of
16 samples. Zygosaccharomyces rouxii was found in four of the samples while
Zygosaccharomyces balii (sample 9 at 3.1 x 101 CFU/mL) and Ogataea
thermomethanolica (sample 11 at 1.4 x 101 CFU/mL) were found only once each
(data not shown in table).
 83

Table 3.3 Occurrence of yeast species in samples of molasses taken at the Bundaberg
distillery

Sample Sample Population (CFU/mL)

number date S. cerevisiae Schiz. pombe Zygo. rouxii
1 10/08/06 nd nd nd
2 10/08/06 nd nd nd
2 1 1
3 22/01/07 1.1 x 10 0.1 x 10 1.0 x 10
1 1
4 09/03/07 0.8 x 10 nd 2.2 x 10
5 31/01/08 nd nd nd
1
6 04/02/08 0.7 x 10 nd nd
2 1
7 10/03/08 1.1 x 10 3.8 x 10 nd
2
8 10/03/08 2.4 x 10 nd nd
1 1 1
9 12/03/08 3.8 x 10 2.5 x 10 3.1 x 10
2
10 12/03/08 2.0 x 10 nd nd
2 1 1
11 12/03/08 1.4 x 10 3.5 x 10 1.4 x 10
2 1
12 12/03/08 2.0 x 10 1.0 x 10 nd
1
13 09/11/08 0.9 x 10 nd nd
1 1
14 09/11/08 1.3 x 10 0.7 x 10 nd
15 27/04/10 nd nd nd
16 27/04/10 nd nd nd
nd – not detected, less than 5 CFU/mL

Occurrence and growth of yeasts throughout the rum production process

Preliminary studies in 2006 revealed no detectable yeasts (< 5 CFU/mL) in

samples of the raw materials (molasses, town water, dunder) and consequently
they were not detected in freshly prepared fermentation medium as taken from the
surge tanks just prior to transfer to the fermenters. High populations (107-108
CFU/mL) were found in the yeast propagation vessels and these consisted of only
S. cerevisiae, the starter culture. These same high populations of S. cerevisiae
were found in samples taken from two different fermentation tanks and samples of
buffer tanks A and B. No yeasts other than S. cerevisiae were detected in these
locations. It was concluded from these preliminary trials that rum fermentations at
the distillery were conducted by the selected starter culture of S. cerevisiae that
was routinely propagated at the distillery and inoculated into the fermentation
medium.
 84

A more detailed investigation of yeasts throughout the production process was

conducted in 2008 and the results are presented in Table 3.4.

Table 3.4 Populations of yeast species throughout the rum production process;
conducted in 2008.
Stage/sampling point Total population Number of different
species
3
Molasses storage tank 2.0 x 10 2
2
Molasses supply tank 5.5 x 10 2
Dunder nd (<10) -
Dunder and mud line nd (<10) -
Dilution tank nd (<10) -
Clarifier nd (<10) -
Surge tank (50° Brix) nd (<10) -
2
Surge tank (30° Brix) 4.5 x 10 2
Floc nd (<10) -
Acid water nd (<10) -
8 a
Yeast; laboratory propagation 1.6 x 10 1
5 a
Yeast vessel A (early 1000h) 5.9 x 10 1
8 a
Yeast vessel A (2300h) 2.4 x 10 1
8 a
Yeast vessel B before transfer to C 5.1 x 10 1
7 a
Yeast vessel C before transfer to D 9.1 x 10 1
* *
F5 F9
7 7 a
D prior to transfer to fermenter 4.6 x 10 7.1 x 10 1
7 7 a
Fermenter sample after transfer of yeast 1.9 x 10 1.7 x 10 1
7 7 a
6h 6.3 x 10 2.5 x 10 1
7 7 a
12h 7.1 x 10 7.5 x 10 1
7 7 a
18h 6.8 x 10 7.1 x 10 1
7 7 a
24h 4.0 x 10 5.6 x 10 1
7 7 a
36h 6.1 x 10 7.8 x 10 1
7 a
Buffer tank A 6.2 x 10 1
7 a
Buffer tank B 5.1 x 10 1
a
1 - Yeast species identified as [Link] introduced as starter culture.
*
F5 & F9 refer to Fermenter 5 and Fermenter 9. These fermenters were the two observed for this study. F5 is
an example of an old fermenter (smaller in volume 41000L) and F9 is a newer fermenter (larger in volume
79000L).

Two species of yeast were consistently isolated at low populations (approximately

103 CFU/mL) from the molasses storage tanks. These were identified as
Saccharomyces cerevisiae and Schizosaccharomyces pombe. These species
carried over into the molasses supply tank, but were not isolated from the
molasses preparation after clarification and transfer to the first surge tank.
 85

The culture of yeast propagated in the laboratory gave a yeast population

exceeding 108 CFU/mL, from which only Saccharomyces cerevisiae was isolated.
This population was diluted to about 105-106 CFU/mL on addition to yeast
propagation vessel A, but multiplied to populations greater than 108 CFU/mL during
subsequent incubation in this vessel (Table 3.4). Such populations and species
homogeneity were maintained on subsequent transfer to propagation vessels B, C
and D. This population was diluted to about 107 CFU/mL on inoculation into the
fermenters. Several fermentation tanks are usually inoculated at this stage, and
two of these, tank F5 and tank F9 were subsequently monitored for yeast growth
(Table 3.4). These tanks represented an older and newer style of fermentation
tank. The yeast quickly grew to about 7.0 x107 CFU/mL, and remained at such
populations until the end of fermentation at 36 h. Similar population data were
obtained for both fermentation tanks and for analysis on either MEA or WL agar
(Figure 3.6).

Saccharomyces cerevisiae was the only species isolated throughout fermentation.

On completion of fermentation, the fermented wash from the different fermenters
was pumped into either of two buffer (or mixing) tanks where the yeast population
remained at about 5.0 x107 CFU/mL and consisted of Saccharomyces cerevisiae
(Table 3.4).
 86

Figure 3.6 Yeast populations during production of rum as determined by culture on, MEA (a),
WLA (b). (I) Yeast propagation vessel A early, (II) yeast propagation vessel A before transfer to vessel
B, (III) yeast propagation vessel B, (IV) yeast propagation vessel C, (V) yeast propagation vessel D, (VI)
0 h (post transfer from yeast vessel into fermenter), (VII) 6 h, (VIII) 12 h, (IX) 18 h, (X) 24 h, (XI) 36 h, (XII)
buffer tank A, (XIII) buffer tank B

After the in depth survey performed in 2008, there was a requirement for a small,
targeted ecological survey to understand some critical points from the 2008 survey.
The data for yeast analyses are shown in Table 3.5 and confirm the single
presence of S. cerevisiae in the yeast propagation vessels and buffer tanks at
levels of 108-109 CFU/mL.
 87

Table 3.5 Supporting yeast population data from mini-ecological survey performed in 2010.
Stage/sampling Point Total Number of
population different
(CFU/mL) species
Molasses (2 tanks) <5 N/A
Dunder <5 N/A
Surge tank – in <5 N/A
Surge tank – out <5 N/A
Wort cooler <5 N/A
6
Yeast vessel A (early) 2.0 x 10 1
8
Yeast vessel A (late) 3.5 x 10 1
6
Fermenter 5 (12 h) 6.3 x 10 1
6
Fermenter 9 (12 h) 2.5 x 10 1
8
Buffer tank A 7.7 x 10 1
8
Buffer tank B 7.5 x 10 1

[Link] Identification of Bacteria

Bacteria were routinely isolated by plating samples onto MRS, Raka Ray and WLS
agar plates from which isolates were obtained and identified. To determine the
morphological characteristics of each species, isolates were streaked onto both
MRS and WLS agar. This helped to determine how the differing media impacted on
the appearance of the colonies.

Morphological and physiological properties and rDNA sequencing were used to

characterise and identify the bacterial species. Biochemical testing using API
CHL50 test strips (Biomerieux, Durham NC) was conducted to further support
identifications. Throughout the course of the ecological surveys, eight different
species of bacteria were isolated from various sampling points during the
production process. Table 3.6 gives a list of these bacteria as identified with API
strips and rDNA sequencing.
 88

Table 3.6 Identification of bacterial isolates from Bundaberg distillery

Identification API CH50 Identity
Accession
Species by rDNA
a number
sequencing
Lactobacillus plantarum 98.0 GU064445.1 L. plantarum (good - 97.7%)
Lactobacillus
99.0 HM004216.1 L. fermentum (good – 96.9%)
fermentum
L. paracasei spp paracasei
Lactobacillus paracasei 99.0 AB330933.1
(good - 95.0%)
Uncultured
98.0 GQ082129.1 L. acidophilus (good - 92.7%)
Lactobacillus
Lactobacillus brevis 98.0 EF076751.1 L. brevis (good - 97.5%)
Lactobacillus Acceptable ID to genus
98.0 AY733085.1
oligofermentans Lactobacillus
b
Bacillus cereus 99.0 DQ 884352.1 Not performed
b
Bacillus subtilis 97.0 EU016526.1 Not performed
Lactobacillus farraginis 94.0 AB262733.1 L. buchneri (very good - 99.9%)
a
confirmed identity by sequencing of the 16S rDNA (% match according to the BLAST database)
b
API CH50 specific for lactic acid bacteria

Two bacterial species were repeatedly isolated over the course of the investigation.
These were Lactobacillus fermentum and Lactobacillus plantarum. Figures 3.7 and
3.8 show the colony morphology of these isolates on MRS agar, and cellular
morphology as determined by examination with phase contrast microscopy and
scanning electron microscopy.
 89

Fig 3.7 Lactobacillus plantarum isolated from in-process samples of rum distillery fermentations. (A)
appearance on MRS agar, (B) examined under light microscopy (x1000, Leica), (C) and (D) scanning
electron microscope. (Electron Microscope images obtained at Macquarie University Microscopy Unit,
NSW, Australia.)
 90

Figure 3.8 Lactobacillus fermentum isolated from a rum fermentations. (A) appearance on MRS agar,
(B) examined under light microscopy (x1000, Leica), (C) and (D) scanning electron microscope.
(Electron Microscope images obtained at Macquarie University Microscopy Unit, NSW, Australia.)

Zymomonas and Propionibacterium

Analyses for Zymomonas and Propionibacterium species were conducted by
culturing samples, taken throughout the process during the 2008 and 2010 surveys
(Table 3.1), on Universal Beer agar and Sodium Lactate agar, respectively. No
isolates of these bacteria were detected (< 50 CFU/mL) in any of the samples.
Reference cultures of these species (Zymomonas mobilis, ATCC 10988 and
Propionibacterium freundenreichii, UNSW 035900, respectively) grew on these
media as controls to demonstrate that the conditions of cultivation were effective.
 91

Clostridium

Analyses for Clostridium species, were conducted for the 2008 and 2010 surveys,
(Table 3.1). Samples were tested using Differential Reinforced Clostridial Agar
(RCA) (Difco). No isolates of this species were detected (<1 CFU/mL) in any of the
samples. Reference cultures of Clostridium beijerinckii (UNSW 060600) grew on
RCA as controls.

Bacteria associated with molasses

During 2006-2010, random samples of molasses were taken and examined for the
presence of bacteria by culture on plates of MRS, Raka Ray and WLS agar (Table
3.7). These samples were taken directly from the storage wells or from holding
tanks just prior to clarification. Bacterial counts were very low, at less than 1000
CFU/mL and, on some occasions, no viable bacteria could be detected (< 5
CFU/mL).

Bacillus subtilis was the most frequently isolated species, being found in 11 out of
the 16 samples, followed by Lactobacillus spp. in six out of 16 samples. Bacillus
cereus was found in one of the samples (1.0 x 101CFU/mL). The Lactobacillus spp.
were not identifiable by API testing and were given as equivalent to an unculturable
Lactobacillus spp. by rDNA sequencing.
 92

Table 3.7 Occurrence of bacterial species in samples of molasses taken at the Bundaberg
distillery
Population (CFU/mL)
Sample Sample
number date Lactobacillus
B. cereus B. subtilis
spp.
1 10/08/06 nd <5 nd <5 nd <5
2 10/08/06 nd <5 nd <5 nd <5
2
3 22/01/07 nd <5 3.5 x 10 nd <5
2 2
4 09/03/07 nd <5 2.5 x 10 1.4 x 10
2 2
5 31/01/08 nd <5 2.5 x 10 1.1 x 10
2
6 04/02/08 nd <5 2.5 x 10 nd <5
7 10/03/08 nd <5 nd <5 nd <5
3 1
8 10/03/08 nd <5 1.5 x 10 1.0 x 10
2 2
9 12/03/08 nd <5 2.5 x 10 2.2 x 10
1 1
10 12/03/08 nd <5 8.0 x 10 4.0 x 10
1
11 12/03/08 nd <5 9.0 x 10 nd <5
12 12/03/08 nd <5 nd <5 nd <5
1
13 09/11/08 nd <5 nd <5 3.0 x 10
1 1
14 09/11/08 1.0 x 10 2.0 x 10 nd <5
2
15 27/04/10 nd <5 4.5 x 10 nd <5
2
16 27/04/10 nd <5 3.0 x 10 nd <5
nd – not detected, less than 5 CFU/mL

Occurrence and growth of bacteria throughout the rum production process

Preliminary studies in 2006 revealed a complex, and largely unknown, population

of bacteria present in the fermentations at the Bundaberg distillery. These studies
included examination of samples by phase contrast microscopy, and exploratory
plating onto culture media routinely used by the quality assurance laboratory at the
Bundaberg distillery.

When examined by phase contrast microscopy, samples of fermenting molasses

medium from the fermenters and buffer tanks consistently showed the presence of
bacterial cells in addition to yeast cells. Typically, the bacterial populations were
approximately 5-10 times less than the yeast population and consisted mainly of
long, thin rod-shaped cells. Samples for culture were taken at the following sites;
town water supply, dunder, molasses, molasses medium in surge tank, yeast
propagation vessel C, fermenter 3, fermenter 8 and buffer tanks A and B. These
 93

samples were examined by spread plate culture on each of the following media -
PCA, Raka Ray, MRS and WLS. Using these media, no bacteria (CFU/mL <100)
were detected in samples of town water, dunder, molasses, and molasses medium
in the surge tank. However, bacteria were detected in samples taken from the
yeast propagation vessel C, at populations of 104-105 CFU/mL and in samples
taken from the two fermentation vessels and the two buffer tanks at populations of
105-106 CFU/mL. It was concluded, from these preliminary studies, that rum
fermentations conducted at the distillery consistently exhibited the presence and
growth of bacteria in addition to the yeast starter culture (S. cerevisiae).

A more detailed investigation of bacteria throughout the production process was

conducted in 2008 and the results are presented in Table 3.8. The predominant
bacterial species changed as the sample points progressed through the production
process. Low populations of Bacillus species and the unidentified Lactobacillus
spp. were found in molasses samples and these carried through into the molasses
medium present in the surge tanks. No bacteria were isolated from samples of
town water, flocculating agent or dunder used in preparation of the molasses
medium. Also, no bacteria were detected in the steps used to prepare the yeast
starter culture at the laboratory stage. However, lactic acid bacteria were
consistently isolated from the larger yeast propagation vessels in the factory.
These included four species, L. plantarum, L. fermentum, L. brevis and the
unidentified Lactobacillus. Bacteria were not detected in yeast propagation vessel
A, initially, but after 12-14 h bacterial populations reached 1.5 x 101 CFU/mL.
These populations increased in yeast propagation vessel B to about 10 6 CFU/mL
and continued at this level in yeast vessel C. A decline in these populations
occurred in conjunction with dilution into yeast vessel D and subsequent transfer to
a larger vessel for fermentation. Four species were consistently isolated from these
yeast propagation vessels. These were predominantly L. plantarum and L.
fermentum, the species identified as an unculturable Lactobacillus, and to a lesser
extent, L brevis.
 94

Table 3.8 Populations of bacterial species throughout the rum production process;
conducted in 2008

Sample point Species isolated Average population CFU/mL

3
Bacillus subtilis 1.5 x 10
2
Molasses Lactobacillus spp. 9.0 x 10
1
Bacillus cereus 1.0 x 10
Dunder none nd (<10)
Dunder and mud line none nd (<10)
Dilution tank none nd (<10)
1
Bacillus subtilis 1.6 x 10
Clarifier Lactobacillus spp. nd (<10)
1
Bacillus cereus 2.2 x 10
1
Bacillus subtilis 3.0 x 10
Surge tank (50° Brix)
Lactobacillus spp. nd (<10)
1
Bacillus subtilis 4.5 x 10
Surge tank (30° Brix)
Lactobacillus spp. nd (<10)
Floc none nd (<10)
Acid water none nd (<10)
Yeast starter culture none nd (<10)
Yeast
(slant) propagation
slant) (slant) none nd (<10)
(laboratory) A B
1 4
Lactobacillus plantarum 1.3 x 10 8.1 x 10
4
Yeast propagation Lactobacillus fermentum nd (<10) 1.1 x 10
4
vessel A/B Lactobacillus brevis nd (<10) 1.3 x 10
3
Lactobacillus spp. nd (<10) 1.3 x 10
C D
6 5
Lactobacillus plantarum 1.9 x 10 1.0 x 10
5 4
Yeast propagation Lactobacillus fermentum 5.1 x 10 3.2 x 10
vessel C/D Lactobacillus brevis nd (<10) nd (<10)
5 3
Lactobacillus spp. 1.5 x 10 3.0 x 10
F5 F9
6 7
Lactobacillus plantarum 2.2 x 10 1.1 x 10
Fermentation vessels – Lactobacillus fermentum 1.4 x 10
6
2.6 x 10
6
4 4
Early Lactobacillus brevis 1.6 x 10 5.0 x 10
4 5
(0-12 h) Lactobacillus spp. 3.0 x 10 1.2 x 10
6 7
Lactobacillus plantarum 3.7 x 10 3.6 x 10
Fermentation vessels – Lactobacillus fermentum 2.5 x 10
5
4.0 x 10
6
5
Late Lactobacillus brevis nd (<10) 3.5 x 10
4 5
(18 – 36 h) Lactobacillus spp. 5.0 x 10 5.0 x 10

A B
7 6
Lactobacillus plantarum 1.6 x 10 2.5 x 10
Buffer tanks Lactobacillus fermentum 5.0 x 10
6
4.5 x 10
6

Lactobacillus plantarum and L. fermentum were consistently isolated from the

molasses medium throughout the fermentation process, reaching populations of
106-107 CFU/mL or slightly higher in the case of L. plantarum. Such populations
carried over into the buffer tanks (Table 3.8). Their contributions are shown in more
 95

detail for two different fermentation vessels in Figure 3.9 (a, b, c, d). Similar
population data trends were obtained for culture on either WLS or MRS agar. The
unidentifiable Lactobacillus spp. and L. brevis were also prevalent throughout the
fermentation process, but their populations were generally lower (10 4 - 106
CFU/mL) and they were not as consistently isolated, as evident in Figure 3.9 (e, f,
g, h). Also, the two culture media, MRS and WLS agar, did not always give similar
population data for these two species. Their less frequent presence and lesser
populations reflected in their inconsistent detection in samples taken from the
buffer tanks (Fig 3.9).

Lactobacillus parcasei, Lactobacillus oligofermentans, Lactobacillus farraginis and

Lactobacillus buchneri were randomly isolated from the fermenting molasses
medium but at low populations (< 4.1 x 104 CFU/mL).
 96

Figure 3.9 Changes in the populations of individual bacterial species (a,b L. fermentum, c,d L.
plantarum, e,f Lactobacillus spp. g,h L. brevis) during fermentation of molasses, on two different media
(■) MRS (▲) WLS from a small fermenter (a,c,e,g) and large fermenter (b, d, f, h). (I) yeast propagation
vessel A early, (II) yeast propagation vessel A before transfer to vessel B, (III) yeast propagation vessel
B, (IV) yeast propagation vessel C, (V) yeast propagation vessel D, (VI) 0 h (post transfer from yeast
vessel into fermenter), fermenter at (VII) 6 h, (VIII) 12 h, (IX) 18 h, (X) 24 h, (XI) 36 h, (XII) buffer tank A,
(XIII) buffer tank B
 97

After the in depth survey performed in 2008, there was a need for a targeted
survey of a few critical points in the production process. The data for bacterial
analyses are shown in Table 3.9. These findings confirmed those obtained in the
earlier investigations of 2008 and 2006. These were: absence of bacteria in dunder
samples; presence of Bacillus and Lactobacillus in samples of molasses medium
from the surge tank; presence of L. plantarum and L. fermentum in the yeast
propagation system; and prevalence of L. fermentum, L. plantarum and to a lesser
extent the unidentifiable Lactobacillus spp. in the fermentation process as evident
from data obtained from buffer tanks

Table 3.9 Supporting bacterial population data from mini ecological survey (2010)

Stage/sampling Point Species isolated Total population

(CFU/mL)
Dunder (after still) N/A nil
Dunder (after clarifier) N/A nil
1
Wort cooler Bacillus subtilis 5.0 x 10
Lactobacillus spp.
2
Surge tank (in) Bacillus subtilis 2.4 x 10
1
Lactobacillus spp. 6.0 x 10
2
Surge tank (out) Bacillus subtilis 4.2 x 10
1
Lactobacillus spp. 3.0 x 10
Yeast propagation Vessel A N/A nil
(early)
4
Yeast propagation Vessel A L. fermentum 1.3 x 10
(late) L. plantarum 4.1 x 10
4

6
Buffer tank A L. fermentum 7.3 x 10
6
L. plantarum 3.5 x 10
5
Lactobacillus spp. 2.0 x 10
7
Buffer tank B L. fermentum 1.0 x 10
6
L. plantarum 2.2 x 10
 98

3.4 DISCUSSION
This chapter reports the first systematic investigation of the microbiology of rum
production in an Australian rum distillery. Its main aim was to understand which
microbial species are predominant during the molasses fermentation, how these
species grow and develop throughout the process and the likely sources of these
microorganisms. With this basic ecological information, quality assurance
procedures could then be refined and developed to better manage the process.

The Bundaberg Distilling Company

Production of rum first started at The Bundaberg Distilling Company, Bundaberg,

Queensland in 1888. It was set up and run by the operators of local sugar cane
processing mills as a means to manage waste molasses from sugar production
(Kerr, 1983). As rum production was not a main business focus of the sugar cane
farmers and millers, the distilling operation evolved and developed somewhat
empirically over the years to become the largest rum producing operation in
Australia. In 2000, it was purchased by the British based multinational, alcoholic
beverage company, Diageo plc. Presently, the Bundaberg Distilling Company
produces heavy body, dark rums that are highly appreciated by local consumers
for their unique flavours, but are less valued in an international context (Broom,
2003). Approximately half a million 9L cases valued at $200 million are produced
annually (Main, 2010). Bundaberg Rum is an iconic brand name to Australian
consumers, recently celebrating 125 years of production (Kerr, 1983).

Future development and expansion of the distillery requires a more thorough

understanding of the science and technology of the overall operation so systems
can be improved or developed to better manage production efficiency and product
quality. The company is aware that its product has unique character in relation to
rums, globally, but is uncertain as to what factors lead to this property. It is with this
background that the microbiological investigations reported in this chapter were
undertaken.
 99

Production process
The basic process followed those described in the literature for rums produced
using sugar cane molasses as the main raw material (Anson, 1971; Kampen,
1975; Lehtonen & Suomalainen, 1977; Nicol, 2003; James, 2008). Dunder was
added to the molasses fermentation medium and the company used a fermentation
process that relied on inoculation with a specific strain of Saccharomyces
cerevisiae. The overall operation, plant equipment and quality assurance practices
had evolved over many years. Consequently, there was a mixture of old and new
equipment, (such as the fermentation vessels), myriads of pipelines to
accommodate new additions as the production expanded, staff who had been
associated with the company for many years, and entrenched quality assurance
procedures. Although basic microbiological testing was routinely done, the staff
had no formal microbiological training or qualifications. There was an assumption
that the fermentation process was due to the inoculated yeast because a starter
culture of this yeast was used in conjunction with heat treatment (80°C) of the
molasses medium for fermentation and the regular use of CIP systems.

Molasses

The stored molasses at the Bundaberg distillery had an initial °Brix of

approximately 84°Brix, water activity (Aw) of 0.7 - 0.8, fermentable sugar content of
60-80% and a pH 5.1 – 5.7. This is consistent with literature values (Lehtonen &
Suomalainen, 1977; Kampen, 1975, Nicol, 2003). The molasses was stored in
large in-ground pools/tanks with periodic topping up from external suppliers as
deemed necessary, due to production requirements. These pools were open to the
atmosphere from where external contamination could occur. The tanks were rarely
emptied and cleaned, neither were the associated pipes and pumps. Culturable
microbial populations were very low (<103 CFU/mL), with many molasses samples
giving no detectable microorganisms on various occasions. Possibly, those
samples were obtained immediately after a topping up operation when the
temperature of the incoming molasses would be high (>60°C) as a consequence of
the sugar processing operation.
 100

Studies performed by Browne (1929) over a 14 year period, failed to isolate any
yeast, moulds or bacteria from molasses samples. Other researchers found low
populations of bacteria (Fahrasmane & Ganou-Parfait, 1998) similar to those
obtained in this study. In attempts to improve the recovery of culturable organisms
from the molasses samples (compared with results obtained on PCA, MRS, Raka
Ray, WL and WLS), some samples were simultaneously cultured on PCA, into
which either 5% w/v sucrose, 5% w/v molasses or 5% v/v dunder had been
incorporated. However, trials with these media did not improve the growth of the
bacteria on the plates or lead to the isolation of additional species.

The molasses tanks, with storage over many years without cleaning, represent a
unique microbial ecosystem, presenting a specific environment of very high sugar
concentration, low Aw, very little available nitrogen sources and the presence of
microbial inhibitors (Bluhm, 1983; Curtin, 1975; Lehtonen & Suomalainen, 1977;
Nicol, 2003). Such conditions are generally hostile to the survival and growth of
microorganisms and explain the very low populations observed in this study and by
others (Fahrasmane & Ganou-Parfait, 1998).

Yeasts have been frequently isolated from stored molasses, although at low levels
< 103 CFU/g. As might be expected, the most frequent isolates have been osmo-
tolerant species of Zygosaccharomyces and Schizosaccharomyces (Hall et al,
1935; Parfait & Sabin, 1975; Bonilla-Salinas, 1995) along with S. cerevisiae which
is also known to tolerate high concentrations of sugars (El-Tabey Shehata et al.,
1959; Bonilla-Salinas, 1995). Similar data were found during this study (Table 3.3).

The bacteria of molasses used for rum production have not been systematically
studied. The few studies available report very low populations (102 - 103 CFU/ mL)
and random isolations of Clostridium saccharolyticum, Propionibacterium jensenii,
Lactobacillus fructivorans, Leuconostoc paramesenteroides and Lactobacillus
plantarum (Hall et al, 1935; Ganou-Parfait et al, 1989; Fahrasmane & Ganou-
Parfait, 1998; Todorov & Dicks, 2004). The most frequently isolated bacteria from
molasses in the present study were species of Bacillus - especially B. subtilis. This
is not unexpected as it is a spore forming organism that could withstand the heat
 101

processing of the sugar cane, and the spores are likely to survive in stored
molasses (Owen, 1911). Previous research into molasses based rum
fermentations did not find Bacillus species in the molasses raw material. However,
Bacillus species have been isolated from molasses fermentation media used in the
production of rums (Ganou-Parfait et al, 1987; Fahrasmane et al, 1988; Ganou-
Parafit et al, 1989).

Although not consistently found, an unidentifiable species of Lactobacillus was

often isolated from molasses samples (Table 3.7). It was the only species of lactic
acid bacteria found in such samples and could be part of an indigenous flora that
has developed over time in the molasses storage tanks of the Bundaberg distillery.

Molasses is a complex and unique environment, consequently, there may be

species present that cannot be isolated by standard cultural procedures. Such
species might be detected by using non-culture based, molecular technologies
such as PCR-DGGE or the more recently developed high throughput sequencing
technologies that are now being applied to study the microbiology of foods and
beverages (Amann et al, 1995; Muyzer & Smalla, 1998; Torsvik et al, 1998; Rappe
& Giovannoni, 2003; Ercolini, 2004; Tyson & Banfield, 2005; Bokulich & Mills,
2012). Facilities for the use of these technologies were not available at the time the
research for this project was conducted and further research applying such
methods will be needed to better understand the microbial ecology of molasses.

Indigenous microflora present in the molasses could impact on rum quality and
flavour, especially if they grow during molasses storage. As mentioned previously,
molasses is a hostile environment for microorganisms and naturally restricts
microbial growth. The low populations of microorganisms found in molasses in this
study are unlikely to have a significant metabolic impact on molasses composition
and quality. The molasses storage wells/pits at the Bundaberg distillery were
housed within large buildings (sheds) to protect them from the external
environment. Nevertheless, the surface of the pits remained exposed to the
atmosphere within these housings. Within the company records, there were reports
where water had leaked onto the surface of the molasses pits during extreme
 102

weather conditions such as strong storms and tropical rainfall. These

circumstances are likely to create micro-environments such as diluted molasses
“hot spots” where the Aw would be decreased and the resultant growth of
microorganisms could be significant. In such cases, the microbial growth could
produce metabolites with the potential to impact on rum flavour and quality.
Consequently, monitoring of molasses storage conditions would be an important
critical point to be incorporated into the company quality assurance and HACCP
management systems.

Preparation of Molasses Fermentation Medium

The aim of molasses preparation was to produce a fermentation medium that had
the appropriate concentrations of fermentable sugars and to inactivate
contaminating microorganisms present in the molasses. Molasses was diluted with
potable water and dunder to a final value of 30°Brix for the fermentation mash. This
would consist of approximately 15-20% fermentable sugars that would give a fully
fermented product of approximately 7-10% ethanol (Bundaberg in-process
documentation, 2009). For the Bundaberg distillery, the diluted molasses
incorporated dunder at a final concentration of 7.5%, and the mix was clarified by
flocculation and then heated at 70-80°C and left to “stand” for approximately 1 h.

Culturable yeasts and bacteria were not detected in samples of the dunder,
sourced directly from the distillation process, and they were not detected in the
flocculating agent added to the mixture. After clarification and heating, culturable
microorganisms were detected at low populations in the molasses fermentation
medium. Only Bacillus species at 101-102 CFU/mL were detected at this stage and
this is consistent with their presence in molasses and heat resistance due to the
production of endospores. After this stage, the molasses medium was passed
through a heat exchanger to cool it down to 50-60°C and then further diluted with
potable, UV-sterilized water to give the final Brix value of 30°Brix in the surge tank.
Dilution at this stage facilitated the earlier clarification operation, as a more
concentrated mixture was used at that point and this avoided the processing and
heating of much larger volumes in that operation. However, samples taken from
 103

this surge tank gave low populations of yeast (102-103 CFU/mL), Bacillus species
(101-102 CFU/mL) and the Lactobacillus spp. (2010 data) suggesting contamination
from unclarified molasses. After the surge tank, the molasses medium was used in
the fermentation tanks and in the yeast propagation stages.

Yeast and propagation process

The Bundaberg distillery used a specific strain of S. cerevisiae which was stored in
several commercial culture collections located around the world. These collections
routinely monitored and checked species and strain purity and homogeneity. This
ensured the security and continued availability of this critical microorganism. As
such, the Bundaberg distillery could be confident that every starter culture provided
by the external culture collection was the same strain of S. cerevisiae with its
specific and desired characteristics – thereby, maintaining consistency in product
character and flavour, and minimising the risks of spoilage and failed fermentations
(Verstrepen et al 2003). The frequency (6-12 months) of renewal of the starter
culture from the culture collection, means that the iterations of propagation will not
be too far removed from the stock starter culture. This minimizes chances of
contamination and evolutionary change. In controlled fermentations, the
introduction of a starter culture is undertaken to provide a yeast strain which will
impart desirable characteristics to the ferment (Fleet, 2008). This starter culture is
added at populations high enough to out-compete any wild yeast present in the raw
materials e.g. molasses. This was seen in the Bundaberg distillery where
populations of wild yeast were not detected during the propagation and
fermentation stages.

The propagation process was a series of stepwise procedures aimed to nurture the
starter culture from the storage slope to growth in a chemically and nutritionally
complex fermentation medium to high populations of 10 6-108 CFU/mL. It involved a
laboratory phase where the yeast was cultured in media that had been sterilized by
autoclaving (Figure 3.4). For scale up to larger volumes in vessels A-D, molasses
medium from the surge tank was used. Fermaid, ammonium sulphate, magnesium
sulphate and ammonium dihydrogen phosphate were added, usually during early
 104

propagation steps, to provide extra nitrogen sources and key nutrients and
vitamins for yeast growth and survival. Analyses of samples from the various
stages of propagation during the 2006, 2008 and 2010 investigations consistently
showed the presence of only S. cerevisiae at populations of 106-107 CFU/mL or
sometimes higher. Therefore, it may be concluded that, with respect to the yeast,
the propagation system used at the distillery was performing correctly. It is
assumed that the isolates of yeast obtained throughout the process and identified
as S. cerevisiae by rDNA sequencing were the same strain. Further research using
molecular methods to determine strain homogeneity, such as examining restriction
patterns, chromosomal profiles or mtDNA restriction analysis (Fernandez-Espinar
et al, 2001) would be needed to confirm this assumption.

Unexpectedly, bacteria were found in samples taken throughout the yeast

propagation process. They were not detected in samples taken from the
propagation stages prepared in the laboratory where autoclaved media were used.
However, they first became evident in yeast propagation vessels A and B that
contained heat treated molasses fermentation medium, as prepared for the
fermentation process. It was generally assumed by distillery staff that this medium
would be sterile. However, this study showed this not to be the case, as bacteria
were present in samples taken at this stage for both the 2008 and 2010
investigations. These bacteria were not Bacillus species as might be expected.
Rather, they were a mixture of Lactobacillus species, but predominantly L.
plantarum and L. fermentum, which increased in populations to about 105-106 by
the end of the propagation stages. Consequently, the starter culture used to
inoculate the fermentation vessels was not a pure culture of the selected strain of
S. cerevisiae and contained significant populations of up to four species of
Lactobacillus. Knowing and understanding the source of this bacterial
contamination is important and will be discussed in a later section.

Fermentation Process

Before inoculation, the molasses fermentation medium showed non-detectable or

low populations (approx. 102 CFU/mL) of yeasts. After inoculation from propagation
 105

vessel D, the initial yeast population was approximately 10 7 CFU/mL and only S.
cerevisiae was found. This species grew during fermentation and was the only
species isolated throughout the fermentation and at the end of fermentation.
Maximum populations of 107-108 CFU/mL were achieved during fermentation.
These population data are consistent with what has been reported previously for
starter culture inoculated, molasses rum fermentations (Arroyo, 1945a, b). Any
contaminating yeast species (e.g. Schizosaccharomyces pombe or
Zygosaccharomyces species) that might have originated from the molasses were
overwhelmed by the inoculation process and did not establish themselves during
fermentation. Also, both Schizosaccharomyces pombe and the
Zygosaccharomyces species have much slower growth rates than S. cerevisiae
(Lehtonen & Suomalainen, 1977, Fahrasmane et al, 1988) and this would
contribute to their inability to compete with S. cerevisiae during these rum
fermentations.

It may be concluded from the yeast analyses, that the inoculated strain of S.
cerevisiae is the principal yeast responsible for the rum fermentations conducted at
the Bundaberg distillery. Its biochemical activities and particular profile of
secondary metabolites will have a strong impact on the final quality and flavour of
the rum. Although the rum from this distillery may be considered as a heavy style
product, there was no quantifiable contribution from Schizosaccharomyces pombe,
which is frequently associated with the fermentation and production of these
particular rum styles (Fahrasmane & Ganou-Parfait, 1998; Lehtonen &
Suomalainen, 1977).

Lactic acid bacteria were consistently isolated from the molasses rum
fermentations conducted at the Bundaberg distillery. They were found at initial
populations of 104 – 106 CFU/mL at the commencement of fermentation and this is
consistent with their presence in the yeast starter culture after propagation as
discussed previously. Thereafter, they grew throughout fermentation, reaching
maximum populations of 107 – 108 CFU/mL. The most prevalent species
associated with the fermentation were Lactobacillus plantarum and Lactobacillus
 106

fermentum, but two other species, an unidentifiable Lactobacillus species and

Lactobacillus brevis also contributed to the process, but less consistently and at
lower overall populations.

It is evident from the onsite, 2006, 2008 and 2010 ecological surveys of the rum
fermentation process that lactic acid bacteria are an integral, but uncontrolled part
of the overall operation at the Bundaberg distillery. The extent to which they grow
during fermentation (populations of 107 – 108 CFU/mL) is quantitatively significant
and their biochemical activities are most likely to impact on process efficiency and
the flavour and quality of the rum product. Because of this importance, further
studies on the association of these bacteria with the process will be presented and
discussed in Chapters 4 and 6.

As discussed in Chapter 2, sporadic studies have reported the association of

bacteria with molasses based rum fermentations (Chapter 2, Table 2.8).
Organisms of particular concern in these early studies were species of Clostridium
and Propionibacterium, with only a brief mention of lactic acid bacteria (Hall, et al
1935; Fahrasmane et al. 1988; Ganou-Parfait et al., 1989). Such bacteria were
considered to have the potential to positively or negatively impact on product
flavour and quality (Allan, 1906; Fahrasmane et al. 1988). Despite this significance,
systematic studies of the contribution of bacteria to rum fermentations have not
been reported. The data presented in this Chapter represent the first detailed
investigation of the contribution of bacteria to molasses based rum fermentations
and show a definitive presence of lactic acid bacteria. In this context, the findings
are consistent with those for cachaça fermentations (Schwan et al., 2001; Duarte
et al.; 2011) and whisky fermentations (Simpson et al., 2001; van Beek & Priest,
2000, 2002 & 2003; Cachat & Priest, 2005) which report similar associations of
lactic acid bacteria and conclude the potential of these species to impact on
process efficiency and product quality.

Apart from Bacillus species, no bacteria other than lactic acid bacteria were found
in the Bundaberg distillery molasses or fermentations despite using specific media
and culture conditions to monitor the possible presence of Clostridium,
 107

Propionibacterium and Zymomonas species, as mentioned in some early studies

(Fahrasmane et al. 1988; Ganou-Parfait et al., 1989). In particular, Clostridium
saccharolyticum has previously been isolated from rum fermentations and has,
along with Bacillus species (B. butyricus and B. amylobacter), been linked to the
development of important flavour volatiles (Allan, 1906; Hall et al. 1935, Ganou-
Parfait et al, 1989). The contribution of Clostridium species to the production of
volatile acids (acetic, butyric, caproic, heptanoic and propionic acids), which are
precursors of some esters considered essential for rum aroma and flavour, has
previously been postulated (Hall et al., 1935). These esters are of particular
importance in the dark, heavy style rums. The absence of these bacterial species
from the fermentations at the Bundaberg distillery, which produces such rums, is of
note.

Cleaning and Sanitation

Although the Bundaberg distillery implemented a well-established, CIP program at

the completion of each fermentation cycle, there were some locations and sections
of pipe work where cleaning and sanitation were not systematically applied. As
mentioned previously, the molasses storage pits and associated pumps and pipe
work were rarely cleaned. Provided there was no “infection” within these pits, this
would not present a hazard since the molasses is heat treated at the dilution and
preparation stage. Ideally, all locations after this stage should be subject to regular
and effective cleaning and sanitation. However, this was found not to be the case.
As the operation expanded and developed over the years, it was inevitable, that
added lines, and pipework, created areas that escaped the CIP system or reduced
the CIP effectiveness. After onsite inspections, several locations were identified
where the CIP system was not active or regularly applied. These included a
secondary dunder storage vessel and associated pipework located near the
molasses clarifier tank, various heat exchangers associated with molasses medium
preparation, and pipework associated with the yeast propagation vessels and
feeding lines to inoculation ports of the fermentation tanks, junctions or “elbows”
within the pipe work, provided “dead spaces” with reduced fluid movement. With
 108

the exception of the secondary dunder storage tank, these critical zones were
located after the molasses had been diluted and heated, but prior to the
fermentation tanks (Figure 3.1). The location of these sites would favour the
continued growth of adequate populations of indigenous bacteria that repeatedly
“inoculate” the prepared fermentation medium.

There were some locations that were only cleaned “as required”, based on the
judgements and expertise of production staff. These decisions were linked to
production efficiency and the need to minimize any downtime due to cleaning.
Such locations included buffer tanks, clarifier and heat exchangers. For example,
the buffer tanks were not cleaned on a daily basis, because they fed the distillation
units on a continuous, 24 hour basis. Heat exchangers were often only cleaned at
the time they became ineffective due to blockages caused by sediment of mud and
scale. These blockages also caused leakages and cross contamination (visible
when cooling water on one side of the exchanger became discoloured from hot
molasses or dunder on the other side). As a consequence, heat exchangers were
monitored by recording pH and colouration of cooling water. Cooling water was
continuously circulated via heat exchangers through-out the production facility, as
a means of conserving energy and minimising production costs. Thus, heat
recovered from the distillation units and dunder storage vessels was used to heat
the molasses for preparation of the molasses fermentation medium and then the
water was chilled to cool the molasses fermentation medium to the temperature
required for fermentation. The same water was continuously recycled throughout
the entire production system using a network of heat exchangers.

Quality Assurance Program

The Bundaberg distillery had an onsite quality assurance laboratory. This

laboratory was responsible for yeast propagation, raw material testing, checking
production efficiency and sensory testing. The staff had no formal qualifications in
microbiology and their skills were mostly based in analytical chemistry and
company experience passed down over the years. Process efficiency was mostly
monitored by regular chemical testing such as the concentration of sugars,
 109

(sucrose, fructose and glucose) in the molasses and sugar and ethanol content in
the fermented molasses and dunder. Sensory testing of the distillates was done on
a daily basis.

Good aseptic technique was followed in preparation of the yeast starter culture,
and yeast growth and purity in the starter preparations and fermentations were
monitored by microscopic observation and counting with a haemocytometer slide.
High populations of large cells of the yeast (S. cerevisiae) were dominant in
samples examined under the microscope but the staff were not trained to see the
smaller sized, bacterial cells which were always present in such slide preparations.
Consequently, bacterial populations within the production process at the distillery
were not studied in any detail prior to this research project. Some culture plating for
yeasts and bacteria were conducted but the rationale for this testing was not well
understood. Samples of fermented molasses were diluted 10 3-104 fold before
plating, consequently, species present at low populations would not be detected.
There was a general assumption that because a pure yeast starter was used and
that the molasses fermentation medium had been heated to 70-80°C for 40-60 min,
the inoculated S. cerevisiae would be the only species present during fermentation.

Possible Sources of Lactic Acid Bacteria in the Rum Production Process

This was the first systematic ecological survey of an Australian rum distillery. Raw
materials, in-process and fermentation samples were examined. The starter
culture, Saccharomyces cerevisiae, was the only yeast found during fermentation
stages. Investigations showed that the predominant bacteria were lactic acid
bacteria. Four species of lactic acid bacteria were repetitively found in the
fermentation samples with final populations reaching approximately 10 7 CFU/mL.
The source of where these bacteria enter the production process is critical in order
to obtain a thorough understanding of the microbial ecology of the distillery.

Samples of various raw materials and intermediate production components were

taken throughout the study, with sampling sites becoming more specific and
specialised as potential bacterial contamination sites were eliminated. This
 110

systematic sampling often coincided with new information obtained from production
staff regarding the complex array of pipework found in the production areas of the
distillery.

Figure 3.10 is a reproduction of the process outline given in Figure 3.1, showing
sites where lactic acid bacteria were isolated and key heat exchange units.

Figure 3.10 Flow diagram of production process showing sites where lactic acid bacteria
were isolated and key heat exchangers,(*) Lactobacillus spp., ($) L. plantarum, (#) L.
fermentum, (@) L. brevis, (^) all Bacillus species.

The unidentified Lactobacillus spp. was randomly found in samples of molasses,

but at low populations and it carries through into the molasses fermentation
medium and process. However, it is not the dominant species within the process.
Aspects of its heat resistance will be discussed in Chapter 5. The main entry point
 111

for L. fermentum, and L. plantarum, the two dominant bacteria of the process, (as
well as L. brevis and the Lactobacillus spp.) is the scaled up yeast propagation
stages where molasses fermentation medium was used. Sections of pipelines used
to supply molasses fermentation medium to the propagation vessels were
identified as those that escape effective CIP. It was not possible to dismantle and
microbiologically examine these sections of pipeline during the course of this
project. It is likely that biomasses of lactic acid bacteria have accumulated in
sections of these lines and continually contaminate the molasses fermentation
medium feeding into the yeast propagation vessels. Further research is needed to
investigate this possibility.

Another likely source of contamination in the process arises from the dunder
storage vessel, located between the feeding pipes from the distillation unit and the
clarifier. Dunder is held in this tank on a transient basis as it is directly fed into the
clarifier as needed to prepare the molasses fermentation medium. The temperature
of dunder in this tank is in excess of 60°C. Samples of dunder taken from this tank
throughout the course of the project were always at this temperature and, as
expected, gave no culturable microorganisms. Dunder from this source is also
used to give up its heat via heat exchangers to “cooling” waters that are
continuously circulated throughout the production lines. It was revealed in the late
stages of the project that these heat exchangers were not included in the CIP
system and were only cleaned when they were seen to malfunction through
accumulated materials and leakage into the cooling waters, as evidenced by a
brown colouration of these waters. It was also revealed that the dunder storage
tank is not within the CIP system since it is continuously hot, but there are times
when production stops for one, two or more days when it cools down and develops
off odours. When production restarts, this material is then fed into the system along
with the microbial contaminants that have grown in the cooled dunder. In Chapter
5, it will be demonstrated that species of lactic acid bacteria develop in this dunder
on cooling and storage. These contaminants will then be distributed throughout the
production system through the heat exchangers that are not systematically
maintained and cleaned. In this way, it may be concluded that lactic acid bacteria
 112

become indigenous residential microflora of the production process, having the

physiological ability to survive and grow in the environment that molasses
presents. Such colonization is likely to occur in those sections of the process that
escape effective CIP. Because of production pressures and timing constraints, it
was not possible to undertake specific microbiological analyses of the dunder
storage tank and associated heat exchangers to substantiate the above conclusion
and this remains a direction for further research.

Conclusion

Microbiological surveys of the Bundaberg distillery over the period 2006-2010

showed that the inoculated yeast S. cerevisiae was responsible for the alcoholic
fermentation of molasses for the production of rum.

Lactic acid bacteria were consistently found as major contributors to the

fermentation process and originated as indigenous contaminants within the
operation. Most likely sources were cross contamination from fermented molasses
medium to raw materials that were being prepared for new fermentations. This in
all likelihood arose due to a combination of factors such as under maintained
equipment cleans and sporadic plant shutdowns causing temperature variations
and unintentional lag times critical for bacterial growth.

Because these bacteria were present at quantitatively significant populations and

likely to impact on product quality and process efficiency further research is
needed to better understand their contribution to the process. Chapter 4 of this
thesis will examine the bacterial ecology at the end of fermentation in more detail
by studying the ecology of the buffer tanks. Chapter 5 will investigate the
microbiological and chemical characteristics of the raw material, dunder and
Chapter 6 will explore the impact of key bacterial species on the efficiency of yeast
fermentation in a controlled environment and how this may affect production
efficiency and flavour development.
 113

CHAPTER 4

OCCURRENCE OF BACTERIA IN THE

FERMENTATION OF MOLASSES FOR RUM
PRODUCTION

4.1 INTRODUCTION
Alcoholic fermentation of molasses is a main operation in the production of rum.
Yeasts are the primary microorganisms of this fermentation. There are numerous
reports in the literature that bacteria may contribute to this fermentation (Allan,
1906; Hall et al, 1935; Arroyo, 1945a; Ganou-Parfait et al 1987 &1989), but a
systematic, ecological study of the contribution of these bacteria is lacking. In
Chapter 3, it was shown that bacteria, principally lactic acid bacteria, were part of
the ecology of molasses fermentation for rum production at the Bundaberg
Distilling Company. During fermentation, their populations increased to maximum
levels of approximately 106-108 CFU/mL. The ecological investigations reported in
Chapter 3 focussed on the entire production chain and, consequently, were limited
to a few sampling times. Additional studies are required to confirm that lactic acid
bacteria and, possibly, other species of bacteria are a consistent part of the
microbial ecology of molasses fermentation.

In rum distilleries, it is not uncommon to use several tanks for the fermentation of
molasses. At the completion of fermentation, the contents of these tanks are mixed
into one tank called a buffer tank, from which the fermented molasses is fed into
the distillation process. Such tanks are a quality control mechanism to help ensure
a more uniform composition of fermented molasses for distillation. Due to mixing of
 114

the contents of several fermenters into the one vessel, these buffer tanks provide a
good sampling window to determine the bacterial populations and species in the
production system.

To study these bacteria in more detail and to gain a greater knowledge of their
prevalence in rum production over an extended period of time, the bacterial
populations and species in buffer tanks at the Bundaberg distillery were
determined over a period of 18 months.

4.2 MATERIALS AND METHODS

As part of its quality assurance program, the Bundaberg distillery routinely monitors
the buffer tanks, for their populations of yeasts and bacteria, for each batch of rum
production. These analyses are conducted by trained staff in a dedicated
microbiology laboratory. Bacterial populations are monitored by the pour plate
method using two culture media (i) Wallerstein Differential Nutrient Agar with
Supplement (WLS) (Oxoid) and (ii) Raka Ray agar (Oxoid), each supplemented to
contain 10 µg/mL of cycloheximide (Sigma) to restrict yeast growth. After on-site
inspection, review of these procedures and training of staff, it was concluded that
the plated cultures obtained by this routine quality assurance monitoring would be
suitable for more detailed enumeration, isolation and identification of the specific
bacterial populations associated with the rum process. For some fermentations, the
laboratory staff were requested to include additional analyses by culturing samples
from buffer tanks on de Man, Rogosa and Sharpe agar (MRS) agar also
supplemented to contain 10 µg/mL of cycloheximide (Sigma). Details of the
procedures are given in the following sections.

4.2.1 Samples from Buffer Tanks

Samples (30 mL) of fermented molasses were aseptically collected from the
sampling ports of the buffer tanks, taken to the quality assurance laboratory on site
 115

at the Bundaberg distillery, stored at 5°C, and analysed for the presence of
bacteria within 1 h of collection. Samples for this study were collected over a period
of approximately 18 months during August 2006-March 2008.

4.2.2 Analysis of Bacteria

Samples were serially diluted in 9 mL of Ringers solution and 1 mL sub-samples

examined in duplicate for the populations of viable bacteria by the pour plate
method using WLS agar (Oxoid) and Raka Ray agar (Oxoid). As per standard
operating procedures used at the Bundaberg distillery, only the 105 dilution was
used to generate the recorded population. Some samples were also examined, in
duplicate, using MRS agar (Oxoid). Plates were incubated aerobically at 35°C.
After incubation, total colony counts were calculated and the plates were
aseptically packed in a cool box with ice bricks and shipped by courier to the
University of New South Wales, Sydney. They were received within 2 days and
then stored at 5°C until further examination, generally, within 2 days. Colonies on
the plates were examined for their morphology and cellular morphology, and
counts made for each specific colony type.

Representative isolates (5 of each colony type - if possible - each shipment) were

isolated and purified by streaking onto plates of MRS and WLS agar. Plates were
incubated microaerophilically using candle jars at 30°C for 48 hours. Stock cultures
on slopes were prepared from these plates and stored under glycerol at -20°C and
-80°C.

On several occasions samples from buffer tanks were analysed for the presence of
Propionibacterium and Zymomonas species by culture on Sodium Lactate Agar
and Universal Beer Agar respectively, as described in Section [Link] Chapter 3.
Some samples were also analysed for the presence of Clostridium species using
duplicate pour plates of Differential Reinforced Clostridial Agar (RCA) (Difco). All
three of these media were incubated anaerobically at 30°C for 3-7 days
 116

(Propionibacterium and Zymomonas species) or 2-10 days (Clostridium species)

and plates were checked for growth every 48 h. Anaerobic conditions were
obtained using AnaeroGen© (Oxoid) in appropriate sealed containers.

[Link] Identification of Bacterial Isolates

Bacteria were identified by a combination of phenotypic methods and sequencing

of the 16 S ribosomal DNA according to the methods described in Section 3.2.2 of
Chapter 3. A minimum of 10 isolates of each species were isolated from plates of
each medium type (WLS, Raka Ray and MRS) at different stages during the
sampling period and their identities confirmed by rDNA sequencing.

4.3 RESULTS
Selection of isolation media

As mentioned already, the Quality Assurance Laboratory at the Bundaberg

Distilling Company use WLS and Raka Ray media for the routine monitoring of
bacteria in fermentation samples taken from buffer tanks. As part of this project,
MRS agar was used as an additional medium for this purpose. Colony
development on these media was slow and small for some species (e.g. the
Lactobacillus spp.). In an attempt to improve the growth of these isolates, their
growth on other media was examined at the outset of the project. These media
included Plate Count Agar (PCA) (Oxoid) into which either 5% sucrose, 5%
molasses or 5% dunder had been incorporated. However, trials with these media
did not improve the growth of the bacteria on the plates or lead to the isolation of
additional species.

4.3.1 Populations of Bacteria in Rum Fermentations

Throughout the study, a total of 66 sampling dates were examined. Table 4.1
shows the total populations as enumerated on either WLS, Raka Ray or MRS
media. The first 44 of these samples were cultured on both WLS and Raka Ray
 117

media. A further 22 samples were simultaneously plated on MRS agar as an

additional check on media performance.

All of the sampled dates had populations of bacteria exceeding 106 CFU/ mL (on at
least one medium), with many samples (59%) exceeding 107 CFU/mL and several
(14%) exceeding 108 CFU/mL. These high populations were obtained on all three
media, with no one medium giving consistently higher or lower populations for the
same sample. The screening method, undertaken by the Bundaberg laboratory
staff, had a limit of detection of <5 x 104 CFU/mL. No consistent trends in
population variations were observed in the samples taken over the 18 month time
frame.
 118

Table 4.1 Populations of bacteria (CFU/mL) in rum fermentation buffer tanks as determined
by culture on WLS, Raka Ray and MRS agar plates
Sample WLS Raka Ray MRS Sample WLS Raka Ray MRS
1 3.0 x 106 2.3 x 106 -a 34 1.4 x 108 2.6 x 107 -a
6 6 a 8 7
2 5.1 x 10 3.8 x 10 - 35 4.6 x 10 1.0 x 10 -a
4 6 a 6 6
3 < 5.0 x 10 4.5 x 10 - 36 8.0 x 10 9.0 x 10 -a
6 6 a 7 6
4 7.3 x 10 6.5 x 10 - 37 2.2 x 10 7.0 x 10 -a
8 6 a 6 6
5 1.8 x 10 2.5 x 10 - 38 9.0 x 10 2.5 x 10 -a
4 6 a 7 7
6 < 5.0 x 10 2.9 x 10 - 39 4.2 x 10 4.9 x 10 -a
5 7 a 4 7
7 5.0 x 10 1.1 x 10 - 40 < 5.0 x 10 1.9 x 10 -a
6 4 a 4 6
8 1.5 x 10 < 5.0 x 10 - 41 < 5.0 x 10 5.5 x 10 -a
4 6 a 4 6
9 < 5.0 x 10 1.0 x 10 - 42 < 5.0 x 10 4.0 x 10 -a
4 6 a 7 4
10 < 5.0 x 10 1.5 x 10 - 43 2.4 x 10 < 5.0 x 10 -a
7 6 a 6 4
11 1.3 x 10 1.5 x 10 - 44 4.0 x 10 < 5.0 x 10 -a
6 6 a 6 7
12 3.0 x 10 7.5 x 10 - 45 4.5 x 10 2.1 x 10 4.5 x 106
6 6 a 4 7
13 4.0 x 10 3.5 x 10 - 46 < 5.0 x 10 1.8 x 10 1.5 x 106
4 6 a 7 6
14 < 5.0 x 10 5.0 x 10 - 47 2.7 x 10 3.0 x 10 1.4 x 107
4 6 a 8 6
15 < 5.0 x 10 3.0 x 10 - 48 1.0 x 10 1.0 x 10 3.0 x 106
4 7 a 7 7
16 < 5.0 x 10 6.1 x 10 - 49 5.6 x 10 1.9 x 10 3.2 x 107
6 4 a 6 7
17 4.5 x 10 < 5.0 x 10 - 50 1.0 x 10 1.2 x 10 1.0 x 106
7 4 a 6 7
18 1.5 x 10 < 5.0 x 10 - 51 3.0 x 10 1.8 x 10 1.0 x 105
6 4 a 6 7
19 3.5 x 10 < 5.0 x 10 - 52 3.5 x 10 2.9 x 10 1.0 x 106
7 4 a 6 7
20 2.2 x 10 < 5.0 x 10 - 53 7.5 x 10 4.7 x 10 3.5 x 106
5 4 a 7 7
21 5.0 x 10 < 5.0 x 10 - 54 3.6 x 10 1.1 x 10 3.0 x 106
7 4 a 7 6
22 3.0 x 10 < 5.0 x 10 - 55 3.2 x 10 4.0 x 10 3.0 x 106
6 4 a 6 4
23 3.5 x 10 < 5.0 x 10 - 56 4.0 x 10 < 5.0 x 10 1.0 x 106
7 7 a 7 7
24 8.1 x 10 3.4 x 10 - 57 8.4 x 10 3.3 x 10 8.0 x 106
8 6 a 8 7
25 1.1 x 10 3.0 x 10 - 58 1.1 x 10 1.9 x 10 6.0 x 106
4 7 a 6 7
26 < 5.0 x 10 3.2 x 10 - 59 1.5 x 10 1.7 x 10 1.0 x 106
6 4 a 6 6
27 1.0 x 10 < 5.0 x 10 - 60 3.0 x 10 4.5 x 10 1.5 x 106
4 7 a 4 4
28 < 5.0 x 10 2.5 x 10 - 61 < 5.0 x 10 < 5.0 x 10 3.8 x 108
4 7 a 4 4
29 < 5.0 x 10 1.4 x 10 - 62 < 5.0 x 10 < 5.0 x 10 >3.0 x 108
6 6 a 4 4
30 4.0 x 10 4.5 x 10 - 63 < 5.0 x 10 < 5.0 x 10 1.0 x 106
7 4 a 6 7
31 1.5 x 10 < 5.0 x 10 - 64 2.0 x 10 2.1 x 10 5.5 x 106
4 6 a 7 7
32 < 5.0 x 10 6.5 x 10 - 65 1.0 x 10 1.2 x 10 1.5 x 106
4 7 a 8 7
33 < 5.0 x 10 1.9 x 10 - 66 1.3 x 10 2.2 x 10 9.5 x 106
a
These samples were not tested on MRS agar
The results given are the mean counts of duplicate analyses.

4.3.2 Bacterial Species associated with Rum Fermentations

The colonies that developed on the WLS, Raka Ray and MRS isolation plates were
consistently Gram positive, catalase negative rods and were considered as
presumptive lactic acid bacteria. The colonial and cellular morphologies of the
isolates, on these media, were similar to those described in Chapter 3. Four
different morphological types were consistently present and were identified as L.
plantarum, L. fermentum, L brevis and an isolate identified as an “unculturable
 119

Lactobacillus” also described in Chapter 3. The “unculturable Lactobacillus” was a

rod type isolate that was not precisely identified by sequence data but is
presumptively described here as Lactobacillus spp. On some sporadic occasions
(about twice over the 18 month survey period), colonies that did not match any of
these four species were also noted on the isolation plates. The populations of
these colonies (<105 CFU/mL) were less than those of the four main species found
and, consequently, were not studied any further.

Lactobacillus fermentum was the most frequently occurring species and was
isolated from 54 of the 66 samples (Table 4.2). Lactobacillus plantarum was the
next most frequently isolated species, being present in 42 samples, followed by
Lactobacillus brevis (27 samples) and Lactobacillus spp. (27 samples). Not all
species occurred in the one sample at the same time but two to three species were
generally present. There were only 13 occasions during the sampling period where
only one species was present. Table 4.2 shows the population range for the
different species. While most samples gave species in the population range 105-
106 CFU/mL, some had L. fermentum and L. plantarum at population levels of 107-
108 CFU/mL.

Table 4.2 Prevalence of species of lactic acid bacteria in rum fermentations

Species Population range (CFU/mL) Total
5 6 6 7 7 8
10 -10 10 -10 10 -10 occurrence
L. fermentum 33 18 3 54 (0.82)
L. plantarum 30 9 3 42 (0.64)
L. brevis 25 2 0 27 (0.41)
Lactobacillus spp. 22 5 0 27 (0.41)

From observations of the colony types recovered on each of the three isolation
media, some differences were noted in the frequency of occurrence for the four
species. Table 4.3 shows the number of times each species was isolated on each
 120

medium. Lactobacillus plantarum, L. fermentum and L. brevis and the Lactobacillus

spp. always grew on at least two of the three media used. However, Raka Ray
agar had a higher frequency of recovery of L. plantarum, L. brevis and the
Lactobacillus spp. Lactobacillus fermentum was more frequently recovered on
WLS medium. Lactobacillus fermentum and L. brevis were not found in any
samples plated onto MRS agar, although they could be isolated from these same
samples when plated onto either Raka Ray or WLS media. Lactobacillus spp. and
L. plantarum were the only two species isolated on all three media.

Table 4.3. Frequency of isolation of different species of lactic acid bacteria from rum
fermentations on three different media
Medium Number of L. plantarum L. fermentum L. brevis Lactobacillus
times growth spp.
occurred on
medium
Raka Ray 50 33/50 (0.66) 24/50 (0.48) 30/50 (0.60) 18/50 (0.36)
WLS 47 15/47 (0.32) 45/47 (0.96) 2/47 (0.04) 3/47 (0.06)
MRS 22 12/22 (0.55) nd nd 10/22 (0.45)
(nd) not detected (species were not detected in populations above the limit of detection)
Note: Frequencies were calculated by dividing the number of times a species was isolated by the
total number of samples with growth above the limit of detection for the test method.

Although analyses of the data obtained over the 2006-2008 sampling period
enabled some broad conclusions to be made as shown in Tables 4.1 and 4.2, it
was not possible to draw consistent conclusions about the populations for
individual species. These populations fluctuated randomly from one sample to the
next. This behaviour is illustrated in Figure 4.1 which shows the populations for
individual species in samples collected over four different months in [Link]
data also confirm the greater prevalence and higher populations of L fermentum
and L. plantarum that were observed over the entire sampling period.
 121

Figure 4.1 Populations of Lactobacillus fermentum, Lactobacillus plantarum,

Lactobacillus brevis and Lactobacillus sp in samples of fermented molasses taken from
buffer tanks over a four month period- February, March, June, August during 2007. Data
represent the highest population observed on either of the three media used to analyse the
sample.

Propionibacterium and Zymomonas

Buffer tank samples are not routinely tested for the specific presence of
Propionibacterium and Zymomonas species. However, on five random occasions
throughout the sampling period, samples of fermented molasses from the buffer
tanks were analysed for the presence of these bacteria by culture on Sodium
Lactate Agar and Universal Beer Agar, respectively as described in Chapter 3
(Section [Link]). None of the samples gave detection (< 50 CFU/mL) of these
bacterial species.

Clostridium
Buffer tank samples are not routinely tested for the specific presence of
Clostridium species. However, similarly to Propionibaciterium and Zymomonas
species, five random occasions throughout the sampling period, samples of
 122

fermented molasses from the buffer tanks were analysed for the presence of these
bacteria by culture on Differential Reinforced Clostridial Agar (RCA) as described
in Chapter 3 (Section [Link]). None of the samples gave detection (< 1 CFU/mL)
of these bacterial species.

4.4 DISCUSSION
As described in Chapter 3, rum fermentations at the Bundaberg distillery are
conducted in a battery of different fermenters, the use of which varies according to
the volume of product required by marketing forces and other operational factors.
Fermentation in each vessel usually lasts for about 24-36 h, after which the
contents are discharged into one of two holding tanks, called buffer tanks. The fully
fermented molasses in the buffer tanks is then continuously fed into the distillation
units. As mentioned previously, this preliminary mixing or blending process also
brings some uniformity or consistency into the product going to distillation.
As a consequence, the buffer tanks should provide a good overall picture of the
microbial ecology of the fermentation process. Theoretically, microorganisms
present in the buffer tanks should be the same as those present towards the end of
fermentation in the individual fermentation vessels. Therefore, long term analyses
of the microorganisms in the buffer tanks should provide a good, representative
picture of the microbial ecology of the total operation, and was the focus of the
investigations reported in this Chapter. In previous analyses, no variation was
observed in the yeasts associated with the rum fermentations, this being the S.
cerevisiae starter culture that was consistently used by the Bundaberg distillery.
Therefore, analyses were confined to examination of the bacteria.

In designing the experimental protocol for this longitudinal study, several technical
limitations, which could impact on the reliability of the data obtained, were
recognized at the outset. These were:
 that the sampling, preparation of microbiological media, and cultural
analyses would be conducted by technical staff at the Bundaberg distillery
as part of their day to day work tasks, and that the reliability of their
 123

contribution might be affected by unforeseen variables (e.g. pressures of

day to day operation, staff absenteeism, changes within staff roles etc.)
 transportation of the cultured organisms on plates of isolation media from
the Bundaberg distillery to the laboratories at the University of New South
Wales (time delay and conditions of transport).

Although these variables could not be fully controlled, their impacts were minimized
by visits to the Bundaberg distillery to train and ensure that staff were competent in
the aseptic collection of samples, taking representative samples, preparation of
microbiological media, plating and culturing of samples on the appropriate media,
and packaging of cultured samples for transport. There were several samples in
Table 4.1 where no bacteria were detected on one plating medium (< 50x104
CFU/mL) but the same sample gave greater than 106 CFU/mL when plated on
another medium. Such gross discrepancies are most likely due to analytical error
within the laboratory, rather than an effect of plating medium composition.

On receipt at the University of New South Wales, samples were carefully checked
for condition and any possible contamination, and were used within 24 hr. Samples
that did not arrive within 2 days or were visibly damaged or contaminated were
discarded. Moreover, to minimize any such influences, a large number of samples
was taken and examined, this being some 66 samples that gave approximately
320 primary isolation plates from which cultures were isolated and examined.

Four species of bacteria were consistently isolated over an 18 month period. These
were Lactobacillus fermentum, Lactobacillus plantarum, Lactobacillus brevis and
Lactobacillus spp. These were the same species that were found in the survey
studies reported in Chapter 3, but within a very short time frame. The populations
found for these bacteria ranged between < 5.0 x 104 CFU/mL and 3.8 x 108
CFU/mL and these data are also similar to those reported in Chapter 3. It is
relevant to note that the culture technique used in this Chapter was based on pour
plating while that reported in Chapter 3 used spread plating. Both approaches gave
 124

similar ecological data with respect to the species and populations obtained,
thereby reinforcing the reliability of the findings.

Reliability of the ecological data obtained was strengthened by the use of three
media to isolate and culture the lactic acid bacteria. Although Raka Ray medium
had the greatest number of isolations of lactic acid bacteria from the samples, and
all four species could be isolated from this medium, it was less efficient at isolating
L. fermentum than WLS medium. However, WLS was not particularly efficient at
isolating L. brevis or the Lactobacillus spp., which was best isolated on MRS agar
(Table 4.3). Conversely, MRS agar did not give reliable isolation of L. fermentum or
L. brevis. There are numerous microbiological media available for the isolation and
enumeration of lactic acid bacteria. This is due to the diversity of physiological
properties exhibited by this group and, consequently, their presence in a large
variety of matrices (Stamer, 1979; Carr et al, 2002), MRS agar is widely used for
the isolation of lactic acid bacteria from many environmental/food samples
(Stamer, 1979; Carr et al, 2002; Taskila et al, 2010), but, based on the data
obtained in this study, it would not be appropriate as a sole medium for the
analyses of molasses rum fermentations, and would lead to an underestimation of
the ecology. The data of this Chapter clearly demonstrate the importance of using
at least two, if not three, different isolation media in order to obtain reliable
ecological data about the contribution of lactic acid bacteria to molasses rum
fermentations. Media used for the isolation of LAB from rum fermentations have
not been found in previous literature. . However, WL and Raka Ray media have
been used for the isolation of these bacteria from cachaça fermentations ((Schwan,
2001; Badotti, 2010) and MRS based media have been used for their isolation from
whisky fermentations (van Beek, 2002).
Similar to the data of Chapter 3, Zymomonas, Propionibacterium and Clostridium
species were not detected in samples from the buffer tanks.

Although lactic acid bacteria were consistently associated with the rum
fermentations at significant populations of 105 - 106 CFU/mL or more, there were
 125

notable fluctuations in the presence and populations of individual species (Fig 4.1).
The unidentifiable Lactobacillus spp. was not isolated from any sample during
March 2007. However, it might have been present at lower populations, namely,
less than the detection limit (5.0x104 CFU/mL) of the plating method used by the
laboratory at the Bundaberg distillery. Such variations with the populations of the
lactic acid bacteria are likely to affect the consistency of product quality and will
need to be managed in quality assurance programs. As mentioned in Chapter 3,
the different species of lactic acid bacteria have probably established themselves
as part of an indigenous, residential microflora within the network of production
facilities and pipelines. The extent and consistency of this colonisation will be
determined by the frequency and effectiveness of cleaning and sanitation
procedures and interruptions to the process that are likely to affect the growth and
survival of these bacteria in the dunder holding tanks. Further research is
recommended to confirm these possibilities.

As discussed in Chapter 3, the ecological data obtained are based on standard

cultural techniques and further research using molecular based culture
independent methods is recommended.

Conclusion

It can be concluded from the data of this Chapter, that lactic acid bacteria are a
consistent component of the microbial ecology of molasses fermentation at the
Bundaberg distillery. As discussed in Chapter 3, they originate as indigenous
contaminants, having established themselves at locations in the process pipelines
over many years. It is highly suspected that irregularities in the dunder
management procedures gives rise to their presence after which they are
inadvertently distributed to other locations within the process chain where
residential colonisation has occurred. More detailed investigation of the
microbiology of dunder will be presented in Chapter 5. The populations of lactic
acid bacteria at the completion of fermentation are consistently high (approximately
106-107 CFU/mL) and, therefore, quantitatively significant in terms of their
 126

metabolic contributions. Consequently, it is likely that they will affect the efficiency
of the fermentation process and rum quality. Aspects of these influences will be
investigated in Chapter 6.
 127

CHAPTER 5
THE MICROBIAL ECOLOGY AND COMPOSITION
OF DUNDER USED IN RUM PRODUCTION

5.1 INTRODUCTION
Rum is a distilled alcoholic beverage obtained from the fermentation of sugar cane
molasses or sugar cane juice. Generally, production consists of preparation of the
molasses or sugar cane juice, microbiological fermentation, distilling the ferment,
maturation of the distillate, packaging of the product, and sale (Kampen, 1975;
Lehtonen & Suomalainen, 1977; Bluhm, 1983; Fahrasmane & Ganou-Parfait,
1998; Nicol, 2003). Historically, dunder has been an ingredient of the process, and
continues to be used in modern rum production (I’Anson, 1971; Kampen, 1975;
Bluhm, 1983). Dunder is most often described as the residue in the still after
distillation of the fermented molasses (Kampen, 1975). Dunder is taken from the
Spanish word “redundar”, meaning overflow (Martini, 2009). It is usually considered
as a waste material and discarded into external pits or tanks where it is stored for
collection as a feed for animals or as a fertilizer (Nicol, 2003). There seems to be
no standard procedure for its use in rum production. It is mostly added to molasses
fermentations at the time molasses is diluted with water. The amount added varies
and can range from 0% to 50% but, mostly, it is added at about 0-10% (Olbrich,
1963; I’Anson, 1971; Fahrasmane & Ganou-Parfait, 1998). The reasons for its use
are also varied and are described as: to lower molasses pH; provide an added
source of nutrients for microorganisms for molasses fermentation; recycle part of
the process water; and provide a source of wild yeasts and bacteria for molasses
fermentation. Lowering the pH is thought to encourage the growth of
Schizosaccharomyces yeasts in comparison to Saccharomyces yeasts during
fermentation, and additional nutrients may encourage the growth of bacterial
species that could favourably impact on rum flavours (Kampen, 1975;
 128

Faharasmane & Ganou-Parfait, 1998). Frequently, dunder is stored before use and
quickly undergoes a natural fermentation by yeasts and bacteria, the ecology of
which is not well defined but, nevertheless, may impact on molasses fermentation
and rum quality. Sometimes, this dunder has been used as a key source of
microbial inocula for rum fermentations (Lehtonen & Suomalainen, 1977).

Despite the use of dunder in rum production and its potential impact on product
quality and process efficiency, there is little published information on its
microbiological and chemical properties. When examined under the microscope, it
shows masses of yeast and bacterial cells, along with other particulate debris
(Kampen, 1975). Theoretically, the microbial cells should be dead due to the high
temperature of distillation. However, if the dunder is stored before use in rum
fermentations, it is likely to become contaminated and support the growth of a
specific microbiological flora. It is broadly mentioned throughout the literature that
dunder is acidic (with high levels of butyric and acetic acid) and is enriched in
nutrients such as amino acids, vitamins and peptides (Nicol, 2003). Dunder that is
immediately recycled back into fermentations without being stored is generally
considered to be sterile and is used in this way to reduce water consumption within
the distillery and save on such costs.

The aim of this Chapter is to determine the microbial ecology and chemical
composition of dunder and to investigate its effect on the growth of microorganisms
during molasses fermentation.

5.2 MATERIALS AND METHODS

5.2.1 Dunder Samples

Samples (approximately 1000 mL) of dunder were collected on site from the
Bundaberg Distilling Company over the period (March 2006 to September 2009).
They were aseptically taken during commercial operation of the facility from
sampling ports already built into the production lines. Dunder was allowed to flush
through the sample ports prior to collection of the material for analysis, to ensure
 129

representative samples were obtained. The dunder was not stored and came from
a pipeline that was directly connected to the base of the distillation column. At the
site of collection, its temperature was approximately 60°C. For chemical analysis,
samples were frozen and stored at -20°C and thawed before examination. For
microbiological analysis, they were stored at 5°C and examined within 24 hr. This
dunder, collected and analysed on site is referred to as “fresh” dunder in the
following sections.

After aseptic collection, some dunder samples were sent by courier to UNSW,
Sydney. The time between collection of the samples and their receipt at UNSW
was generally less than 7 days, and they were not refrigerated during this time,
although they were packed with ice bricks to maintain as cool an environment as
possible without freezing. On receipt, it was noted that the samples were
undergoing microbial fermentation. These samples were stored at 4°C and then
subjected to microbiological examination. This dunder, collected and transported to
UNSW for further analysis is referred to as “stored” dunder in the following
sections.

5.2.2 Microbiological Analysis of Dunder

Microbial flora of dunder was determined by (i) culture plating on agar media, (ii)
enrichment culture followed by plating on agar media.

Samples (1.0 mL) of dunder were serially diluted in 0.1% Bacteriological Peptone
Water, and 0.1 mL spread inoculated, in duplicate, onto plates of various agar
media. Yeasts were analysed by plating onto Malt Extract Agar (MEA) and WL
Nutrient agar (WL) supplemented to contain 100 µg/mL of oxytetracycline (Sigma)
to restrict bacterial growth. The plates were incubated at 25°C for 48-72 h and
checked for yeast colonies. Colonies with different morphologies were noted and
separately counted to estimate their population. Representatives of the different
colony types were isolated and purified by streak culture onto plates of MEA.
 130

Bacteria were analysed by inoculation onto plates of Plate Count Agar (PCA), de
Man, Rogosa, Sharpe Agar (MRS agar), Raka Ray (RR) and WL Nutrient Agar
(WLS) each supplemented to contain 10 µg/mL of cycloheximide (Sigma) to restrict
yeast growth. Plates were incubated at 30°C for 48 h, after which time colonies
were counted. Predominant colony morphologies were noted, counted, and
representative isolates were purified by streaking onto plates of MRS Agar (without
cycloheximide).

To determine the presence of low populations of yeasts and bacteria in dunder,

subsamples (1 mL) were subject to enrichment culture in test tubes of MRS and
Malt Extract (ME) broth (50 mL) for 48 h at 30°C. The cultures were then tested for
the presence of viable bacteria or yeasts by streak plating samples onto MRS agar
for bacteria and MEA for yeasts.

All culture media were obtained from Oxoid (Basingstoke, UK) and prepared
according to the manufacturer’s instructions. Bacterial and yeast isolates from the
above samples were stored at -80°C under 30% glycerol until used for
identification.

5.2.3 Identification of Bacterial Isolates

Bacteria were identified by a combination of phenotypic methods and sequencing

of the 16 S ribosomal DNA.

[Link] Phenotypic Tests

Phenotypic characterization included microscopic examination for cell morphology,

Gram staining, and tests for oxidase and catalase reactions that were done
according to standard procedures described in Smitbert and Krieg (1994). Isolates
were then selected for identification using API CHL50 test strips (Biomerieux,
Durham NC). Cultures were grown at 30°C for 24 hours on MRS Agar prior to
collection and suspension of cell biomass in sterilised distilled water according to
kit instructions. API kits were inoculated with the biomass and incubated at 30°C
for 24 hours and observed for reactions. Kits were incubated for a further 24 hours,
 131

reactions recorded and data processed using APIweb™

([Link] to give bacterial genus and species identification.

[Link] Extraction and Sequencing of ribosomal DNA

Bacterial DNA was extracted, amplified and sequenced according to the methods
found in Chapter 3, Section [Link].

The products were sequenced at the Ramaciotti Centre for Gene Function
Analysis, UNSW, Australia. The resulting sequences underwent DNA similarity
searches on NCBI Blast program using sequences retrieved from the Genebank
Database (Karlin & Altschul, 1990).

5.2.4 Microscopic Examination

[Link] Dunder Samples

Samples of dunder were examined under normal light/ fluorescence microscopy to

determine their general microbial load and composition. These samples were (1)
dunder samples taken directly from the distillery, not stored and not undergoing
any fermentation and (2) dunder samples that had been couriered to UNSW and
were undergoing fermentation. Microbial cells in dunder were sedimented by
centrifugation (2 minutes at 10 000g) washed twice by aspiration with 1 x PBS
(phosphate buffered saline) and stained for 10 min with SYBR Green I Nucleic acid
stain (Invitrogen) (1:1000 in PBS). Cells were centrifuged (2 minutes at 10 000g)
and resuspended in 1 mL 1x PBS. Samples of the cell suspension were transferred
to a slide and analysed using a BH 2 epi-fluorescence microscope 100X oil,
objective (N.A 1.4) (Olympus, Japan), Excitation 488nm – Emission (Jin, X et al.,
1994). This work was done in conjunction with the Microscopy Unit, Department of
Biological Sciences, Macquarie University, NSW, Australia.
 132

[Link] Bacterial Isolates

Bacterial isolates were examined by microscopy according to the methods

described in Chapter 3, Section 3.2.3.

5.2.5 Isolation of Bacteria from Stored and Heated Dunder

In principle, dunder samples taken after distillation should be sterile because of the
time and temperature of the distillation process. This was a continuous feed
process at 5 000 mL/ minute at 103°C. Dunder sampled and examined
immediately after distillation showed no viable microorganisms (see Results).
However, dunder samples aseptically collected and shipped to UNSW were
consistently fermenting on arrival, suggesting the presence of some heat resistant
microorganism(s) that recovered from heating stress and subsequently grew.
Consequently, experiments were conducted to test for the presence of heat
resistant species in these stored and fermenting samples. Samples (500 mL) of
these fermenting dunders were transferred to a pre-sterilised glass bottle with
loosened cap and heated on a hot plate with magnetic stirring at 100°C for 15 min.
After heating, the bottle of dunder, covered with cap to prevent
recontamination),was incubated at 25°C for 4 days Samples (10 mL) were taken
immediately after heating and thereafter at 2 and 4 days during incubation and
spread plated (0.1 mL) onto MRS agar and MEA to check for the presence of
viable bacteria and yeasts. Plates were incubated at 30°C for 48 hours, after which
time any colonies present were counted and isolated onto MRS agar. Isolates were
purified by restreaking and stored as described previously, and identified to genus
and species by phenotypic and nucleic acid sequencing as described in Section
[Link].

[Link] Characterisation of Bacterial Isolates from Heated Dunder

Isolates from Section 5.2.2 were subjected to further characterization because they
were not conclusively identified by the phenotypic and DNA sequencing
 133

determination described in section [Link]. Further characterization included

growth at different temperatures and pH and ethanol tolerance.

Stock cultures were grown in MRS broth for 24-48 hours prior to inoculation into
MRS broth prepared according to the relevant test. Initial population was
approximately 107 CFU/mL.

[Link] Temperature Tolerance

Growth at different temperatures was conducted in MRS broth (100 mL) in 120 mL
bottles. The medium was inoculated with 100 µL of stock culture and incubated at
either 20°C, 25°C, 30°C, 40°C, 50°C or 60°C in triplicate. Growth was monitored by
measurement of increase in optical density (OD) (600 nm) (UV1201, Shimadzu)
due to biomass formation. Samples of culture were taken every 8 h for
measurement. One bottle of medium (blank) was also incubated at each
temperature to ensure any discolouration of the medium due to temperature would
be taken into account. Experiments were done in triplicate on two separate
occasions and average values are reported.

[Link] Ethanol Tolerance

Ethanol tolerance was determined by measuring growth in MRS broth (final volume
100mL) to which ethanol had been added after autoclaving, to give final
concentrations of 0%, 1%, 2.5%, 5%, 7.5% and 10% v/v of ethanol. Media were
inoculated with 100 µL of stock culture and incubated at 30°C. Growth was
monitored by OD measurement as described previously, and experiments done in
triplicate on two separate occasions and average values are reported.
 134

5.2.6 Effect of Dunder on the Growth of Yeasts and Bacteria in

Molasses

Molasses and dunder were obtained from the Bundaberg Distilling Company. The
molasses was diluted with water to give the concentration normally used in rum
fermentation, namely, 30°Brix or 15% fermentable sugars. Part of the water was
replaced with dunder to give diluted molasses containing 10%, 25% and 50% v/v of
dunder, as well as the control, 0% dunder, where no dunder was added. The
molasses medium was adjusted to pH 5.5 by addition of 5M HCl. The medium was
dispensed as 1000 mL volumes in 1L Schott bottles and sterilized by autoclaving.

Fermentations of the molasses media were conducted with Saccharomyces

cerevisiae, Lactobacillus plantarum, Lactobacillus fermentum and Lactobacillus
spp. (Chapter 3 &4). Inoculum cultures of S. cerevisiae were prepared from fresh
slants of MEA and cultured in MEB at 30°C for 24-48 h. This culture (5 mL) was
inoculated into 100 mL of molasses medium and incubated at 30°C for 24-36 h.
The resulting culture (2 mL) was used to inoculate molasses medium containing
different concentrations of dunder. Inoculum cultures of each of the Lactobacillus
species were similarly prepared except that they were pre-cultured in MRS broth.

The molasses cultures were incubated at 30°C for 48 h without shaking and
samples (10 mL) were taken aseptically every 6 h until 48 h for measurements of
yeast or bacterial growth by plate culture. Samples were serially diluted in 0.1%
Bacteriological Peptone Water, and 0.1 mL spread inoculated, in duplicate, onto
plates of MEA for yeasts and MRS agar for the Lactobacillus species. All plates
were incubated at 30°C for 48 h and colonies were counted.

5.2.7 Chemical Composition of Dunder

Samples of dunder were analysed for pH, sugars, organic acids and free amino
acids.
 135

[Link] Sample preparation

Samples to be analysed for sugars and organic acids were prepared according to a
modified version of Ardhana and Fleet (2003). Twenty grams of sample (dunder)
were diluted with 100 mL of distilled water. The homogenate was centrifuged
(Beckman-Coulter, Fullerton, CA) at 4°C 20,000 x g for 20 min and the supernatant
collected. The supernatant was filtered under vacuum through a 0.45µm non-
sterilised filter membrane (Millipore).

[Link] pH

These analyses were done using an Activon© optical pH probe for pH on

uncentrifuged samples, taken directly from the either the fresh or stored dunder
sample. Analyses were done in duplicate and average values are reported.

[Link] Organic acids

The concentration of individual organic acids was determined by HPLC analysis.

Chromatography instrumentation consisted of a pump (Waters 600), autoinjector
(Waters 717, Autosampler PLUS), a PDA-UV detector (Waters 996 PDA)
programmed to extract at a wavelength of 210 nm, a column heater and controller
(Waters Temperature Control Module) and a 300 x 7.8 mm Aminex Ion Exclusion
HPX-87H stainless steel column (BioRad, Richmond CA). This system was
controlled by a computer running Waters Millenium ™ software and data recorded
simultaneously. The column was eluted at 65°C using 0.08% H3PO4 (Ajax
Chemicals, Sydney, Australia) at a flow rate of 0.5 mL/min. Stock solutions of
acetic acid (UNIVAR), oxalic acid 99% (BDH), tartaric acid 99% (Sigma Aldrich),
malic acid 99% (Sigma Aldrich), lactic acid 99% (BDH), propionic acid 99% (Sigma
Aldrich), butyric acid (as 4-phenyl butyric acid, 99%) (Sigma Aldrich), succinic acid
99% (Cole Palmer) and citric acid 99% (Sigma Aldrich) were used to produce a
mixed standard calibration curve to identify and quantify the concentration of each
acid. The calibration curve consisted of three points at 1.5%, 1.0% and 0.5%.
Individual vials of each acid were analysed to determine exact retention time
 136

(location) in the chromatogram produced by the HPLC. Analyses were done in

duplicate and average values are reported.

[Link] Sugars

The concentration of individual sugars was determined by HPLC analysis.

Chromatography instrumentation consisted of a pump (Waters 600), autoinjector
(Waters 717, Autosampler PLUS), a RI detector (Waters RI) coupled with a PDA-
UV detector (Waters 996 PDA) programmed to extract at 210 nm, a column heater
and controller (Waters Temperature Control Module) and a 300 x 7.8 mm Aminex
Ion Exclusion HPX-87H stainless steel column (BioRad, Richmond CA). This
system was controlled by a computer running Waters Millenium ™ software and
data recorded simultaneously. The column was eluted at 65°C using 0.005 M
H2SO4 (Ajax Chemicals, Sydney, Australia) at a flow rate of 0.6 mL/min. Stock
solutions of sucrose (Sigma Aldrich), d-fructose (Sigma Aldrich) and glucose
(sigma Aldrich) were used to produce a mixed calibration curve (to identify and
quantify the concentration of each sugar. The calibration curve consisted of three
points at 60 mg/mL, 50mg/mL and 40mg/mL. Individual vials of each sugar were
analysed to determine exact retention time (location) in the chromatogram
produced by the HPLC. Analyses were done in duplicate and average values are
reported.

Lactose was used as an internal standard to determine any loss through the
sample preparation method. Five mL of the stock solution (32g/L) was added prior
to sample preparation.

[Link] Amino acids

Dunder samples (fresh and stored) were sent to the Australian Proteome Analysis
Facility (Macquarie University, North Ryde, NSW) for analysis of free amino acids.
Samples (4 µL) were analysed, in duplicate, using the Waters AccQ-Tag Ultra
chemistry with quantification by Waters ACQUITY UPLC. The analysis was
facilitated using infrastructure provided by the Australian Government through the
 137

National Collaborative Research Infrastructure Strategy (NCRIS). Analyses were

done in duplicate and average values are reported.

[Link] Ethanol

Ethanol concentrations were determined from samples prepared as described in

Section [Link]; however, vacuum filtration was not used. Aliquots (2 mL) of
supernatant after initial centrifugation were subjected to a secondary centrifugation
step (4°C, 15,000 x g for 10 min). Ethanol concentrations were determined in this
supernatant using an enzymatic assay kit (r-biopharm AG, Germany). Analyses
were done in duplicate and average values are reported.

[Link] Volatile Compounds

Volatile compounds were determined from samples prepared as described in

Section [Link], but vacuum filtration was not used. Aliquots of supernatant after
initial centrifugation were subjected to a secondary centrifugation step (4°C, 15,000
x g for 10 min). Samples were then dispatched to The Australian Wine Research
Institute under refrigerated temperature conditions. There they underwent in-house
sample preparation as described below.

Samples were prepared in 2 dilutions 1/20 and 3/10 with buffer (10% potassium
hydrogen tartrate, pH adjusted with tartaric acid to 4.5). Samples were prepared
and analysed in a randomised order with a blank run every 10 samples.

The analysis was performed on an Agilent 7890 gas chromatograph equipped with
Gerstel MPS1 multi-purpose sampler and coupled to an Agilent 5975C VL mass
selective detector. Instrument control and data analysis was performed with Agilent
G1701A Revision E.02.00 ChemStation software. The gas chromatograph was
fitted with a 30 m x 0.18 mm Resteck Stabilwax – DA column (crossbond carbowax
polyethylene glycol) of 0.18 µm film thickness that had a 5m x 0.18mm retention
gap. Helium (Ultra High Purity) was used as the carrier gas with linear velocity
24.6cm/s, flow rate 0.78mL/min in constant flow mode. The oven temperature was
 138

started at 33°C, held at this temperature for 4 min then increased to 60°C at
4°C/min, then heated at 8°C/min to 230°C and held at this temperature for 5 min.

The conditions of large volume headspace sampling used were as follows: the vial
and its contents were heated to 40°C for 10 minutes with agitation (speed 750 rpm,
on time 80 s, off time 1 s). A heated (55°C) 2.5 mL syringe penetrated the septum
(27.0 mm) and removed 2.5 mL of headspace (fill speed 200 µL/s). The contents of
the syringe were then injected into a Gerstel PVT (CIS 4) inlet fitted with a Tenax
TA inlet liner (0.75 mm I.D., pre-conditioned in the GC inlet at 200°C for 1 hour and
then ramped to 350°C to remove all contaminates before first injection).

The inlet conditions used were as follows: Prior to injection the inlet was cooled to
0°C with liquid nitrogen. While maintaining 0°C, the sample was introduced to the
inlet at 25.0µL/s (penetration 22.0 mm) using split mode (split ratio 33:1, split flow
25.78 mL/min). Following capture of analytes on the Tenax liner the injector was
heated to 330°C at 12°C/min (pressure 24.6).

The mass spectrometer conditions used were as follows: The mass spectrometer
quadrupole temperature was set at 150°C, the source was set at 250°C and the
transfer line was held at 280°C. Positive ion electron impact spectra at 70eV were
recorded in selective ion monitoring (SIM) mode and Scan mode simultaneously
(relative EM volts) with a solvent delay of 4.0 min.

All quantitative data processing was performed with Agilent G1701A Revision
E.02.00 ChemStation software. Analyses were done in duplicate and average
values are reported.
 139

5.3 RESULTS

5.3.1 Yeast and Bacterial Populations in Dunder Samples

Five samples of dunder collected and analysed on site at the rum distillery on
separate occasions gave no viable yeasts (< 5 CFU/mL) when plated onto MEA
and WL agar and no viable bacteria (< 5 CFU/mL) when plated onto MRS, WLS,
PCA and RR agars (Table 5.1). The same samples were also analysed by
enrichment culture in ME broth and MRS broth and gave no viable yeasts or
bacteria after enrichment (Table 5.1).

Table 5.1. Yeast and bacterial counts of dunder sourced directly from the rum distillery and
analysed by direct plating and enrichment.
Dilutions Population
Sample Media used
Performed (CFU/mL)
0 -1 -2
MEA, WL, 10 , 10 , 10 <5
1 0 -1 -2
MRS, PCA 10 , 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
2 0 -1
RR, MRS 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
3 0 -1
RR, MRS, WLS 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
4 0 -1
MRS, WLS, PCA 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
5 0 -1
MRS, WLS, PCA 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
4* 0 -1
MRS, WLS, PCA 10 , 10 <5
0 -1
MEA, WL 10 , 10 <5
5* 0 -1
MRS, WLS, PCA 10 , 10 <5
Note: Samples 4* and 5* were subjected to enrichment in MRS or ME broth for 24 hours prior to
plating onto agar.

When examined under the light microscope, these samples of dunder showed high
concentrations of yeast cells, and lesser populations of small to long rod like
bacterial cells (Figure 5.1). The staining with SYBR ® Green 1 dye confirmed the
phase contrast results, with large amounts of yeast cells visible and smaller
quantities of bacteria also present (Figure 5.1 c & e).
 140

Figure 5.1 Micrographs of dunder examined under phase contrast and epifluorescence
microscopy (magnification x 1000) (a) fresh dunder; (b) stored dunder; (c) stored dunder
stained with SYBR green I dye and exposed to epifluorescence; (d) stored dunder without
epifluorescence excitation; (e) fresh dunder stained with SYBR green I dye and exposed to
epifluorescence; (f) fresh dunder without epifluorescence excitation

Dunder samples received after courier transport to UNSW were fermenting on

arrival. The plastic bottles containing the dunder samples were visibly swollen and
the dunder was gassy. Table 5.2 shows the results of microbiological analysis of
 141

these samples. No yeasts (< 5 CFU/mL) were detected after culture on MEA and
WL agar. However, all samples showed the presence of bacteria when plated onto
MRS agar. Total populations ranged from 1.8 x 106 CFU/mL to 1.9 x 108 CFU/mL.
Based on colony morphologies, at least two different species were present in these
samples.

Table 5.2. Yeast and bacterial counts of dunder aseptically sampled from the rum distillery
and examined after transport to UNSW-approximately 4-7 days.
Bacteria (CFU/mL) Yeast
Sample Date
A B (CFU/mL)
8 2
1 04/12/06 3.4 x 10 <5.0 x 10 <5
7 2
2 29/01/07 8.9 x 10 <5.0 x 10 <5
6 5
3 28/03/07 1.8 x 10 1.2 x 10 <5
8 6
4 30/03/07 1.9 x 10 4.5 x 10 <5
7 6
5 16/05/07 5.2 x 10 1.5 x 10 <5
7 2
6 16/07/07 5.6 x 10 <5.0 x 10 <5
7 2
7 20/09/07 2.7 x 10 <5.0 x 10 <5
a 6 3
8 15/07/09 2.0 x 10 2.3 x 10 <5
b 6 3
9 15/07/09 2.7 x 10 2.8 x 10 <5
6 2
10 12/01/10 4.9 x 10 <5.0 x 10 <5
A- Cream colony 1-3 mm diameter, defined edge, circular
B- Pale brown irregular edge 1-2 mm diameter
a b
sample taken early (8:30am), sample taken late (2pm)

5.3.2 Identification of Bacteria Isolated from Dunder

Dunder samples contained two main types of colonies of bacteria (Table 5.2).
These were isolated and characterised. Table 5.3 shows the colonial and cellular
morphologies of the two isolates. Both isolates were distinctive as long, thin rod-
shaped cells. Both isolates tested positive for the Gram reaction and negative for
the catalase and oxidase reactions. Based on these criteria, the isolates were
tentatively classified within the genus of Lactobacillus for further testing.
Sequencing of the rDNA of both of the isolates gave identical non-specific results
with closest matches (~97-99%) to an “unculturable Lactobacillus species”.
Phenotypic identification using API CHL50 test strips gave both isolates as
Lactobacillus acidophilus.
 142

Table 5.3. Morphology, rDNA sequencing and API identification of bacteria isolated from dunder
Morphological API
Plate culture Light microscopy Sequencing
description identification

Unculturable
Lactobacillus
Pale brown circular
Lactobacillus
acidophilus
colony, 2-3 mm in
(GQ082129.1) 92.7%
diameter
99.0% (Good ID)

Unculturable
Lactobacillus
Cream, circular
Lactobacillus
acidophilus
colony, 2-4 mm in
(GQ082129.1) 92.7%
diameter
98.0% (Good ID)
 143

5.3.3 Isolation and Identification of Bacterial Species from Heated

Dunder
It was noted that, after heating, dunder samples sometimes re-fermented,
suggesting the presence of heat resistant organisms. To investigate this
observation more thoroughly, dunder samples were heated to boiling as described
in Section [Link], allowed to cool and then sampled for microbiological analysis
during subsequent storage at 25°C. Out of five different dunder samples treated in
this way, two re-fermented, from which one bacterial species was isolated as
shown in Figure 5.2.

Figure 5.2. Bacterial isolate from dunder samples which had re-fermented after boiling. (A)
Colony on MRS agar (B) Cells examined under light microscopy (x 1000, Leica) (C) Cells
examined with Scanning Electron Microscopy and Transmission Electron Microscopy (D)
(Electron Microscope images were taken at Macquarie University Microscopy Unit, NSW,
Australia)
 144

This organism had a pin-head sized, cream colony with dark centres on MRS agar
and characteristic long thin rods under microscopic examination .The isolate was
Gram positive, catalase negative and oxidase negative. It was tentatively assigned
to the genus of Lactobacillus for further testing. rDNA sequencing of the isolates
gave non-specific results with closest matches (~97-99%) to an “uncultureable”
Lactobacillus species. Phenotypic identification API CHL50 test strips gave a
closest match 92.7% to Lactobacillus acidophilus. Table 5.4 shows the API
carbohydrate fermentation profile tested on four different isolates of the same
colonial appearance in comparison to data obtained from the literature for L.
acidophilus

Table 5.4 Carbohydrate fermentation reactions of dunder isolates as

determined by API 50CHL analyses
Fermentation of:: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
Galactose + + + + + + + + (+) + v
Glucose + + + + + + + + + + +
Fructose + + + + + + + + + + +
Mannose + + + + + + + + + + v
N-acetyl glucosamine + + + + + + + + (+) + (+)
Amygdalin + + + + + + + + v (-) -
Arbutin + + + + + + + + v - -
Aesculin + + + + + + + + + + v
Salicin + + + + + + + + + v v
Cellobiose + + + + + + + + + (+) v
Maltose - - - - - - - - (+) + (+)
Lactose + + + + + + + + v + v
Starch + + + + + + + + + + +
Trehalose + + + + + + + + (+) - v
Raffinose - - - - - - - - (-) (+) (-)
Amygdalin - - - - - - - - (-) (+) -
Gentiobiose + + + + + + + + + + -
D-tagatose + + + + + + + + v (-) -
Note: Data in columns (1), (2), (3), (4) were results determined in this study (Chapter
5). Column (5) & (6) results obtained from isolates matching the same rDNA results
during ecological survey performed in Chapter 3, (7) & (8) results obtained from
isolates matching the same rDNA results during buffer tank survey performed in
Chapter 4, (9), (10) and (11) were sourced from API 50 CHL literature regarding L.
acidophilus (as per API CHL information brochure, Biomerieux, France). Characters
are scored as + (85-100%), (+) 75-84% positive, v variable (26-74% positive), (-) 16-
25% positive and – 0-15% positive.

Note: (1)-(8) results according to API CHL 50 (Biomerieux) were L. acidophilus

92.7% Good ID,
 145

The morphology of the isolate as a rod shaped bacterium often in pairs or short
chains was confirmed by scanning and transmission electron microscopy (Figure
5.2). This corresponds to the identification and classification of the “unculturable”
Lactobacillus species isolated during the ecological survey performed at the
Bundaberg distillery as described in Chapter 3 and one of the bacterial species
isolated from the buffer tanks in Chapter 4.

5.3.4 Growth of the Lactobacillus spp. Isolate from Heated Dunder

in the Presence of Ethanol and at Different Temperatures
Some growth experiments were undertaken to determine the tolerance of the
isolate to ethanol and temperature conditions that occur in the distillery
environment.

Figure 5.3 shows growth of the isolate in MRS broth adjusted to contain ethanol
concentrations up to 10%. Growth was completely inhibited in the presence of
7.5% and 10% ethanol concentrations. Growth rate and growth yield were
substantially decreased in the presence of 2.5% and 5% ethanol. In the case of 5%
ethanol, the growth rate was approximately 50% that of the control. The addition of
1% ethanol, impacted slightly on growth of the bacterium, with final cell density
being slightly less than that obtained in the absence of ethanol.
 146

Figure 5.3 Effect of ethanol concentration on the growth of a Lactobacillus spp. isolated
from heated dunder.

The effect of varying temperatures on the growth of Lactobacillus spp. is shown in

Figure 5.4. Fastest growth occurred at 30°C followed by 40°C. Significant lag
phases were obtained for growth at 20°C and 25°C, although by 24 h, final cell
densities were similar to those found for growth at 30°C and 40°C. No growth was
observed at 50°C or 60°C.
 147

Figure 5.4 Effect of temperature on the growth of Lactobacillus spp. isolated from dunder.

The appearance of the culture medium during growth at 40°C was clearly different
to that for growth at lower temperatures (Figure 5.5). At 40°C, the cell biomass
tended to come out of a turbid suspension and clump at the bottom of the vessel
(Figure 5.5 b), whereas the biomass was more uniformly suspended and turbid for
growth at 30°C (Figure 5.5 a). Samples from these cultures also revealed different
appearances when examined by light and electron microscopy. The bacterial cells
at 40°C were more clumped and aggregated (Figure 5.5 d) than those grown at
30°C (Figure 5.5 c). This change in cell behaviour was also observed during
electron microscopy imaging of the 30°C (Figure 5.5 e) and 40°C (Figure 5.5 f)
samples.
 148

Figure 5.5 Growth of Lactobacillus spp. in MRS broth at 30°C and 40°C. Culture medium at
30°C (A) and 40°C (B). Cell morphology (light microscopy x 1000, Leica), of culture at 30°C
(C) and 40°C (D) and when examined under the Scanning Electron Microscope 30°C (E) and
40°C (F) (Electron Microscope images were taken at Macquarie University Microscopy Unit,
NSW, Australia).
 149

5.3.5 Effect of Dunder on the Growth of Saccharomyces

cerevisiae and Lactobacillus species in Molasses Medium

As mentioned previously, dunder is added to molasses for rum fermentations.

Such addition could affect the growth of yeasts and bacteria associated with the
fermentation. To examine this possibility, controlled fermentation experiments were
conducted with molasses to which different concentrations of dunder were added.
These fermentations were conducted as pure cultures with the distillers yeast,
Saccharomyces cerevisiae and several species of Lactobacillus that were
previously found to be associated with rum fermentations (Chapter 3). Experiments
were performed in duplicate and results expressed as averages ± 6%.

The growth kinetics of the distillery yeast, S. cerevisiae, under varying

concentrations of dunder in a molasses medium similar to that used in rum
production are shown in Figure 5.6 a. In the absence of dunder, the population of
S. cerevisiae increased from the initial inoculum level of 102 CFU/mL to a
maximum of 1.5 x 107 CFU/mL after 40h. With the addition of 10% dunder, the
growth rate of S. cerevisiae was decreased and the maximum population after 48 h
was decreased to 4.0 x 106 CFU/mL. Yeast growth was very restricted in the
presence of 25% dunder, and reached only a maximum population of 5.0 x 10 3
CFU/mL after 48h. In the presence of 50% dunder, no yeast growth occurred and
the initial population declined to about 102 CFU/mL . However, the yeast cells
started to recover and grow after 30 h. Doubling time and lag time, as shown in
Table 5.5, increased as dunder concentration increased. The addition of 25%
dunder gives the lowest specific growth rate (0.09 h -1) while 10% dunder
concentration gave the highest specific growth rate (during a secondary
exponential phase) of 0.25 h-1.
 150

Figure 5.6 Populations of (a) Saccharomyces cerevisiae, (b) Lactobacillus plantarum,

(c) Lactobacillus fermentum and (d) Lactobacillus spp. in molasses medium
containing dunder at different concentrations

The impact of varying concentrations of dunder on the growth of L. plantarum is

shown in Figure 5.6b and Table 5.5 . In the absence of dunder, L. plantarum
populations increased exponentially to 1.7 x 109 CFU/mL in 48 h. The addition of
dunder at levels of 10% decreased growth slightly, producing a short lag phase,
and giving a final population of 6.5 x 108 CFU/mL at 48 h. Cultures exposed to 25%
and 50% dunder experienced a significant lag phase, lasting between 16 and 24
h, respectively. After the extended lag phase, growth in the presence of 25% or
50% dunder peaked at 2.9 x 107 CFU/mL and 2.5 x 107 CFU/mL, respectively at 48
h.
 151

Table 5.5. Effect of dunder concentration in molasses on the growth properties of

Saccharomyces cerevisiae and several species of Lactobacillus .
Specific Maximum
Dunder Doubling Lag Time
Species Growth Rate Population
Concentration -1 Time (h) (h)
(h ) (CFU/mL)
0%
a
0.15 , 0.21
b
2.4 <8 1.15 x 10
7

10%
a
0.13 , 0.25
b
2.5 <8 4.0 x 10
6

S. cerevisiae
25% 0.09 7 40 5.0 x 10
3

50% 0.13 n/a 48 6.8 x 10

2

0% 0.22 1.5 <8 1.7 x 10

9

10% 0.15 1.5 <8 6.5 x 10

8

L. plantarum
25% 0.26 2 16 2.9 x 10
7

50% 0.32 1.5 24 2.5 x 10

7

0% 0.13 6.5 <16 4.3 x 10

8

10% 0.09 4 <16 3.8 x 10

8

L. fermentum
25%
a
0.08 , 0.13
b
2 <24 2.6 x 10
8

50% 0.38 1 32 3.3 x 10

7

0%
a
0.25 , 0.13
b
2 <8 1.8 x 10
8

10% 0.20 2.5 <8 2.2 x 10

8
Lactobacillus
spp. 25%
a
0.18 , 0.13
b
3.1 <8 3.8 x 10
7

50%
a
0.09 , 0.25
b
2 <8 2.7 x 10
7

a
denotes first exponential phase
b
denotes secondary exponential phase
Specific growth rates were calculated by linear regression of natural logarithm plot slope of cell
population versus time (h) (Stanbury & Whitaker, 1984; Zwietering et al., 1990).

Figure 5.6c shows the impact of varying concentrations of dunder on the growth of
L. fermentum. In the absence of dunder, L. fermentum populations increased
exponentially to 4.3 x 108 CFU/mL by 40 h. The addition of dunder at levels of 10%
affected growth marginally and final populations reached 3.8 x 10 8 CFU/m within
40 h. Cultures exposed to 25% and 50% dunder experienced a significant lag
phase, lasting between 24 and 36 h, respectively (Table 5.5). After the extended
lag phase, growth in 25% dunder gave a maximum population of 2.6 x 108
CFU/mL at 48 h. Populations of L. fermentum grown in 50% dunder peaked at 40 h
(3.3 x 107 CFU/mL).
 152

Figure 5.6d shows survival and growth of Lactobacillus spp. in varying

concentrations of dunder. The growth of a Lactobacillus spp. in molasses medium
was similar in cultures containing no dunder and 10% dunder and gave maximum
populations of approximately 2.0 x 108 CFU/mL at 48 h. Higher concentrations of
dunder did not give noticable lag phases but growth rate of the organism was
decreased, more so for the 50% dunder, and final maximum populations were
decreased to 2.7 x 107 CFU/mL.

Table 5.5 summarizes the effect of dunder on the growth rates, doubling time, lag
time and maximum cell populations of S. cerevisiae and the different species of
Lactobacillus.

5.3.6 Chemical Composition of Dunder

Samples of both “fresh” and “stored” dunder (see Section 5.2.1 for descriptions)
were examined for pH, ethanol (%), sugars, organic acids and free amino acids
(Table 5.6). Volatile compound analysis was performed on stored dunder only.
Eight samples of fresh dunder and 12 samples of stored dunder collected on
different occasions between 2006 and 2010 were analysed. Data presented in
Table 5.6 show the ranges for these samples and averages plus the range for
organic acids.
 153

Table 5.6 Chemical properties of dunder samples taken at the Bundaberg distillery

Fresh Stored Fresh dunder Stored dunder

Properties Properties
dunder dunder mean (range) mean (range)
pH 4.6 - 4.9 4.3 - 4.8 Organic
acids
(mg/g)
0.002 0.002
Ethanol (%) nd - 0.2 nd - 0.2 Oxalic
(0.001 – 0.002) (0.001 – 0.002)
0.08 1.16
Citric
(0.54 – 1.49) (0.55 - 1.86)
Sugars 0.33 0.42
Tartaric
(mg/g) (0.09 – 0.46) (nd – 0.73)
0.94 6.93
Sucrose nd nd Malic
(0.66 – 1.29) (1.3 – 10.3)
1.29 2.63
Glucose nd nd Lactic
(0.42 – 2.01) (1.98 – 4.42)
4.18 4.91
Fructose nd nd Acetic
(0.82 – 6.25) (2.32 – 9.85)
5.47 5.23
Propionic
(5.36 – 5.71) (5.09 – 5.39)
9.95 11.26
Butyric
(9.78 – 10.11) (11.02 – 11.88)
16.80 16.03
Succinic
(16.57 – 18.69) (15.89 – 16.63)
Note: nd – sugars analysis had a limit of detection of <0.1 mg/g, organic acid analysis (tartaric
acid) limit of detection < 0.001 mg/g

Fresh dunder was acidic and had a pH of between 4.6 and 4.9 while stored dunder
was slightly more acidic (4.3 - 4.8). There were no detectable levels of sugars
(glucose, fructose or sucrose) in either stored or “fresh” dunder. The ethanol
content of dunder was very low being with a maximum level of 0.02%, this being
similar for both fresh and stored dunder.

[Link] Organic Acids

Nine organic acids were detected in both fresh and stored dunders. For all of the
acids, there was considerable variation in the concentration of individual acid
found, depending on sample. The main acids in fresh dunder were succinic
followed by butyric and propionic. These were also the main acids in stored dunder
 154

and they occurred at similar levels as found in fresh dunder. However, the stored
dunder gave higher concentrations of lactic and acetic acids compared with the
fresh dunder.

[Link] Amino Acids

Two samples each of fresh and stored dunder were analysed for amino acids.
Results are shown in Table 5.7, with means calculated from duplicate analyses
performed by the Australian Proteome Analysis Facility Ltd, Macquarie University,
NSW, Australia.

Twenty one amino acids were detected in dunder, with aspartic, asparagine,
alanine and glutamic having the highest concentrations (Table 5.7). For many of
the acids, there were notable decreases in concentration during storage and this is
reflected in a decrease in the overall total concentration of amino acids during
storage. However increases were evident for alanine, methionine and ornithine.
 155

Table 5.7 Comparison of free amino acid concentrations in fresh and stored dunder

Amino Acid Fresh Stored Difference

a
(µg/mL)* 1 2 Average 1 2 Average (%)

Histidine 14.2 16.2 15.2 11.8 11.8 12.4 -18.4

Asparagine 93.9 652.4 374.15 97.7 97.7 143.75 -61.6
Serine 25.6 29.8 27.7 11.9 11.9 17.65 -36.3
Glutamine 9.4 4.5 6.95 3.4 3.4 6.2 -10.8
Arginine 12.0 10.5 11.25 5.7 5.7 7.15 -36.4
Glycine 408.8 21.8 16.9 4.4 25.7 15.05 -10.9
Aspartic acid 33.9 520.5 464.65 265.5 425.9 345.7 -25.6
Glutamic Acid 14.6 160.0 96.95 55.6 42.4 48.9 -49.6
Threonine 62.0 18.7 16.65 5.0 6.2 5.6 -67.0
Alanine 62.0 158.0 110.0 167.5 447.1 307.3 179.4
Proline 20.3 45.9 33.1 22.2 31.9 27.05 -18.3
Ornithine 7.3 9.7 8.5 3.1 23.1 13.1 54.1
Cysteine 1.5 1.1 1.3 1.0 1.2 1.1 -15.4
Lysine 9.1 9.6 9.35 4.4 4.5 4.45 -52.4
Tyrosine 8.6 10.7 9.65 8.9 nd 4.45 -53.9
Methionine 1.5 1.2 1.35 2.9 1.9 2.4 77.8
Valine 29.3 29.4 29.35 1.9 3.9 2.9 -90.1
Isoleucine 18.2 18.1 18.15 2.9 1.3 2.1 -88.4
Leucine 11.8 9.1 10.45 nd 3.0 1.5 -85.6
Phenylalanine 12.1 11.0 11.55 6.1 4.1 5.1 -55.8
Tryptophan 2.9 3.5 3.2 3.6 3.5 3.55 10.9
Total 809.1 1741.6 1275.35 685.4 1269.2 977.3 n/a
*Calculation based on free amino acid molecular weight.
a
Difference % compared to fresh dunder.
Data are the averages of duplicate analyses

[Link] Volatile Compounds

As dunder is expected to contain no volatiles, due to it being the remaining liquid

after boiling in the bottom of the distillation column, volatiles were only analysed in
samples of stored dunder (Table 5.9). Few volatiles were found, the most
significant of which were 3-methylbutanol, 1-butanol and the esters, ethyl
butanoate and ethyl hexanoate.
 156

Table 5.8 Volatile compounds identified from stored dunder

Ethyl Esters (µg/L)
ethyl acetate -
ethyl propanoate -
ethyl 2-methylpropanoate 7
ethyl butanoate 137
ethyl 2-methylbutanoate -
ethyl 3-methylbutanoate -
ethyl hexanoate 99
Acetates
2-methylpropyl acetate -
2-methylbutyl acetate 37
3-methylbutyl acetate -
2-phenylethyl acetate 15
hexyl acetate -
Alcohols
2-methylpropanol -
1-butanol 2108
2-methylbutanol -
3-methylbutanol 5857
hexanol 155

5.4 DISCUSSION
As discussed in Chapter 2, and as mentioned in several review papers (I’Anson,
1971; Fahrasmane & Ganou-Parfait, 1998; Wilkie et al., 2000), dunder is a raw
material that is uniquely used in rum production. The reasons for its use have
probably evolved empirically over the centuries but these are briefly summarized
as: recycling available fresh water, reduce distillery waste, reduce disposal costs,
provide acidity required to lower molasses pH, provide a source of wild yeasts and
bacteria for the next molasses fermentation. It is a by-product of each individual
distillery and, consequently, its chemical and biological properties are likely to vary
from one distillery to the next, depending on how it is processed and stored. There
appears to be no standard procedure for its production and use from distillery to
distillery. The Bundaberg distillery uses it at 7.5% volume in the fermentation
medium. The reasons for its use were not clearly stated, but most likely to
decrease water usage and decrease waste output, while also helping to acidify the
next batch of molasses medium. The process at this distillery was managed as a
pure culture yeast (S. cerevisiae) fermentation, and it was generally accepted that
 157

the process proceeded to this expectation. The dunder used in the process was
taken directly from the still, was assumed to be sterile and, consequently, would
not give additional microbial diversity to the fermentation process. A similarly
obtained by-product, “backset” stillage, has been used in whisky production
(Kelsall & Piggot, 2009).

Microbiology of the dunder

Dunder samples taken directly after the still for use in preparation of the molasses
fermentation medium was sterile, as expected, from the heating process. No
yeasts, aerobic bacteria or anaerobic bacteria could be detected in such samples
by direct plate culture (< 5 CFU/mL) or plate culture after enrichment. However
under microscopic examination they showed dense quantities of yeasts cells and,
to a lesser extent, bacterial cells. This is consistent with observations in previous
literature (Kampen, 1975) but it is always assumed to contain dead cells.

Nevertheless, similar samples, aseptically taken and transported to UNSW were

consistently fermenting on arrival (after 1-2 days by courier). Microbiological
analysis of these samples, taken over an 18 month period gave isolation of the
same bacterial species, identified as an “unculturable” Lactobacillus
(GQ082129.1). While there were two different colony types, both were identified as
the same species, and most likely represented different strains. This bacterial
species returned the same sequencing and BLAST identification data, as well as
API CHL50 profiles, that were also obtained for isolates of lactic acid bacteria
described in Chapter 3 (molasses, clarifier, surge tank, yeast propagation,
fermenters and buffer tanks) and Chapter 4 (buffer tanks). The isolation of this
species of Lactobacillus from the entire production facility shows an endemic
contamination and may mean that it is an important species for production
efficiency or flavour development. Further research regarding the impact of this
species on the fermentation of rum at the Bundaberg distillery will be described in
Chapter 6.
 158

An explanation of the development of this species in samples of dunder is puzzling.

First, samples were collected aseptically which means that this species survived
heating of the fermented molasses during distillation. The molasses medium is
heated to 103°C at the base of the still, however this is a continuous distillation
process with no specific time spent at these high temperatures. Typically, for
bacterial cells to be killed, high temperatures must be maintained for particular
durations, such as 15-20 min. This may not be the case during the distillation step
at the Bundaberg distillery. Similarly, the high biomass density of yeast (especially)
and bacterial cells at the end of fermentation, may also influence the ability of cells
to survive, by conferring protection. The increased density of cells may inhibit the
heat exchangers from having prolonged contact to all cells present in the
fermented medium prior to distillation and may mean that some cells are left
unaffected by the heating process. There is evidence that this species produces
some extracellular material (Figure 5.5) that may afford some protection during the
heating steps. The production of extracellular material, such as
exopolysaccharides, known to occur in some lactic acid bacteria may protect the
microorganism from inactivation by external factors such as attack by
bacteriophages, heat or desiccation (Sutherland, 1998). The synthesis of
exopolysaccharides is an important element in the production of biofilms, as the
exopolysaccharides are the substances that are used to adhere the cell mass to
surfaces such as plastics, soils and, in the case of the Bundaberg distillery, metal
surfaces such as pipes and plates of heat exchangers (Trachoo, 2003). The ability
of such microorganisms to produce biofilms would increase their ability to
withstand the stresses associated with processing environments, such as heating
seen at the Bundaberg distillery. Costerton et al. (1995), showed that biofilm
forming bacteria are significantly more resistant (up to 500 times) to antimicrobial
agents than non-biofilm producing cultures. Biofilms also protect microbial growth
as the exopolysaccharides can form a physical barrier reducing the effectivity of
biocides, such as chlorine, making CIP cleans less effective (Trachoo, 2003).

It was unusual that enrichment cultures of dunder samples taken directly after
distillation did not show the presence of the Lactobacillus spp. although it
 159

developed in dunder samples during transport to UNSW. Possibly, the dunder

itself—rich in amino acids and vitamins, after the heating and destruction of yeasts,
provides the survival and growth factors for this species. Confirmation that a heat
resistant species occurs in the dunder was obtained by heating dunder and finding
subsequent re-growth, although this was not consistently found. This may have
been due to various factors, such as variable heat resistance of the cell or
production of exopolysaccharides (as discussed previously). Further research is
needed for more complete taxonomic identification and physiological
characterisation of this organism but this was outside the scope and time frame of
this project. The isolation of a new or novel species from this distillery environment
is a most likely possibility but would not necessarily be a unique observation.
Cachat & Priest (2005) isolated a novel species, Lactobacillus suntoryeus sp., from
a malt whisky distillery and similar strains had also been isolated, in the same
study, from a Japanese malt whisky fermentation.

Chemical composition of dunder

The chemical composition of dunder was described in Chapter 2, Table 2.4 where
only general proximate compositions have been reported, mainly from the point of
view of the chemical oxidation and biological oxidation demand properties related
to waste disposal. There is little doubt that dunder is an acidic product, rich in
proteins, amino acids and vitamins that would arise from heating of the high
amount of yeast biomass in the fermented molasses (Basu, 1975; Bories et al.,
1988; Wilke, 2000). The acidity of the product was confirmed in this study where
the pH was between 4.3 and 4.9. This is consistent with the predominance of
organic acids (Table 5.6) and is also consistent with previous reports, describing
the main acids found in dunder as acetic, propionic and butyric acids (Bories et al.,
1988). These acids would arise from the sugar processing operations that lead to
the molasses, and acids produced by microbial metabolism during molasses
fermentation. As shown in Table 5.7 and mentioned in earlier reports (Nicol,
2003), dunder contained significant levels of free amino acids which would serve
as a source of nutrients - especially for lactic acid bacteria which are nutritionally
 160

fastidious (Carr et al., 2002). Such levels decreased during dunder storage (Table
5.7) which is consistent with the presence and growth of lactic acid bacteria in
stored dunder (Table 5.2). The dunder samples examined in this study did not
show the presence of any fermentable sugars, most likely as any fermentable
sugars remaining after the molasses fermentation stage would probably be utilized
during extended storage of the fermented product in the buffer tanks.

As dunder is expected to contain no volatiles, due to it being the remaining liquid

after boiling in the bottom of the distillation column, volatiles were only analysed in
samples of stored dunder (Table 5.9). Given there are hundreds of volatile
compounds that have been previously characterised from rum (Appendix A), only a
few key compounds were studied. However it should be noted that the presence of
volatile compounds would not be limited to the few compounds highlighted in Table
5.9 and, analytically, the method chosen could have identified numerous other
compounds if deemed necessary. Of those volatiles examined, few were found, the
most significant of which were 3-methyl-1-butanol, 1-butanol and the esters, ethyl
butanoate and ethyl hexanoate. All four of these have been previously associated
with molasses based rum fermentations (Chapter 2, Table 2.6). 3-methyl-1-butanol
has been found in sugar cane molasses and fermented molasses and is known to
have whisky-like sensory characteristics (Lehtonen, 1983 a, b; Pino et al 2002).
The compound, 1-butanol, previously found by Lehtonen, (1983); Pino et al (2002);
Pino et al, (2012) exhibits medicinal sensory notes. Ethyl hexanoate, previously
found in both Jamaican and Cuban rums, ethyl butanoate has also been found
previously in Jamaican rum, and both are known to demonstrate fruity notes similar
to those of pineapple (Pino et al., 2002; Allan, 1972; Pino et al., 2012). Work
reported by Fahrasmane & Ganou-Parfait (1998) outlined the importance of
Clostridium species in the production of higher alcohols, formic, propionic and
butyric acid at the end of fermentation of molasses medium. The combination of
ethanol and butyric acid (also known as butanoic acid) may lead to the production
of ethyl butanoate. No Clostridium species were found in the ecological surveys
undertaken in Chapters 3 and 4. Further research is need to determine if the
 161

Lactobacillus spp. that develops in stored dunder is responsible for the production
of these volatiles.

Effect of dunder on growth of yeasts and bacteria during molasses

fermentation

As mentioned previously, one of the reasons for adding dunder to molasses rum
fermentations is to serve as a potential source of nutrients for yeast growth since
molasses may be deficient in assimilable nitrogen (Kampen, 1975; Fahrasmane &
Ganou-Parfait, 1998). The data of Table 5.7 confirm that it is a good source of
amino acids. However, the growth experiments of Figure 5.6 suggest that dunder
was not necessary to promote the growth of the culture yeast S. cerevisiae. In
contrast, dunder decreased the growth of the yeast especially at concentrations
above 10%. Consequently, its use in this context could be counterproductive, and
decrease fermentation efficiency - especially if other sources of nutrients, such as
Fermaid, are added to the molasses medium, as is the case at the Bundaberg
distillery. The use of dunder at 7.5 % at this distillery is at the borderline of having a
negative impact and, therefore, its use may need re-consideration. At levels above
10%, it clearly has a negative impact on yeast growth. This negative influence
could be related to the presence of relatively high levels of acetic and butyric acids
which have been reported to decrease/ inhibit yeast growth (Lehtonen &
Suomalainen, 1977).

The three lactic acid bacteria, isolated from the Bundaberg distillery (L. fermentum,
L. plantarum and Lactobacillus spp.), also grew best when no dunder was added to
the fermentation medium. Interestingly, but not surprisingly, Lactobacillus spp.,
which was isolated from dunder, grew well at all four concentrations, including 50%
dunder. This is not unexpected as it can remain viable and grow in dunder samples
during storage. Both L. fermentum and L. plantarum also had the potential to grow
in the presence of high concentrations (25%, 50%) of dunder, but relatively long
lag phases were evident compared with the Lactobacillus spp. (Table 5.5). This
 162

suggests the presence of some substances in dunder that are inhibitory to their
growth. Further research is necessary to determine the nature of the substances in
dunder that are inhibitory to yeast and bacterial growth.

The microbiological and chemical studies of dunder have shown that it is a highly
variable, acidic and complex raw material. The presumption by many researchers
and distillery staff that it is sterile if collected close to the still has been put into
question by the microbiological work performed in this Chapter. Its necessity as an
ingredient for the supposed benefit of nutrients and growth factors is also
debatable as simple growth experiments showed no added benefit for its inclusion,
with potential detrimental effects to yeast growth and fermentation efficiency.
 163

CHAPTER 6 GROWTH OF YEASTS AND BACTERIA

DURING FERMENTATION OF MOLASSES FOR
RUM PRODUCTION

6.1 INTRODUCTION
Although rum fermentation is performed by yeasts, primarily strains of
Saccharomyces cerevisiae, bacteria are also present. These bacteria contribute as
natural contaminants, in an uncontrolled, and largely unknown, way. The data of
Chapters 3 and 4 demonstrated a consistent presence of lactic acid bacteria
throughout fermentations conducted in the Bundaberg Distilling Company. Three
species, Lactobacillus fermentum, Lactobacillus plantarum, and an unidentified
Lactobacillus spp., were predominant in these fermentations, frequently reaching
populations as high as 106-107 CFU/mL. These populations are ecologically
significant and are likely to impact on the growth and metabolic activities of S.
cerevisiae during the fermentation and, also, have their own metabolic imprint on
the production of flavour volatiles that could uniquely contribute to rum flavour and
sensory character. Although such influences of lactic acid bacteria have been
described for some other alcoholic fermentations such as wine (Swiegers et al
2005), whisky (Simpson et al, 2001; van Beek & Priest, 2003; Cachet & Priest,
2005; van Beek & Priest, 2000, 2001 & 2003), and cachaça (Schwan et al, 2001;
Duarte et al, 2011) they have not been investigated with respect to rum production.

The objectives of this chapter are to determine the effect of these lactic acid
bacteria on the growth of yeasts, process efficiency and production of flavour
volatiles during rum fermentation. It will examine molasses fermentations under
controlled conditions with single and mixed populations of the yeast, S. cerevisiae,
and the lactic acid bacteria, L. fermentum, L. plantarum and the unidentified
Lactobacillus spp.
 164

6.2 MATERIALS AND METHODS

6.2.1 Species Selection
Species chosen for this investigation were isolated and identified from rum
fermentations conducted at the Bundaberg Distilling Company as described in
Chapter 3. These were the distillery strain of Saccharomyces cerevisiae, and the
bacterial species Lactobacillus plantarum, Lactobacillus fermentum and the
Lactobacillus spp. These lactic acid bacteria were also isolated and identified in
Chapter 4 (from buffer tanks at the end of the fermentation process). The
Lactobacillus spp. was also isolated and examined in detail in Chapter 5.

6.2.2 Molasses Medium and Fermentation

Molasses and dunder were obtained from the Bundaberg Distilling Company. The
molasses was diluted with water to give the concentration normally used in rum
fermentation, namely, 30° Brix or 15 % fermentable sugars. Part of the water was
replaced with dunder to give 7.5% dunder in the final molasses medium, as used at
the Bundaberg Distilling Company for rum production. The molasses medium was
adjusted to pH 5.5 by addition of 5M HCl. The medium was dispensed as 8L
volumes in 10L sealed stainless steel buckets with lids and sterilised by
autoclaving at 121°C for 15 min.

Fermentations of the molasses medium were conducted with Saccharomyces

cerevisiae, Lactobacillus plantarum, Lactobacillus fermentum or the Lactobacillus
spp. as single cultures and in combinations as shown in Table 6.1.
 165

Table 6.1. Laboratory fermentations of molasses with inoculated yeast and lactic
acid bacteria
Individual fermentations
Saccharomyces cerevisiae
Lactobacillus plantarum
Lactobacillus fermentum
Lactobacillus spp.
Mixed fermentations
Saccharomyces cerevisiae and L. fermentum
Saccharomyces cerevisiae and Lactobacillus spp.
Saccharomyces cerevisiae, L. fermentum and Lactobacillus spp.

Inoculum cultures of S. cerevisiae for the fermentations were prepared from fresh
slants of MEA and cultured in MEB at 30°C for 24-48 h, while inoculum cultures of
individual bacteria were prepared from slants of MRS agar and cultured into MRS
broth at 30°C for 24 h. Each of these cultures (5 mL) were inoculated into separate
100 mL of molasses medium and incubated at 30°C for 24-36 h. This culture was
used to inoculate individual 500 mL of molasses medium and incubated for a
further 24 h at 30°C. The 500 mL volume of culture was then used to inoculate the
8 L of molasses medium for the fermentation trials.

The fermentations were incubated at 30°C for 48 h without shaking and samples
(100 mL) were aseptically taken every 8 h until 48 h. These samples were used
immediately for microbiological analyses as described in Section 6.2.4 and the
remaining part of the sample was stored at -20°C until chemical analysis. Each of
the fermentation trials listed in Table 6.1 was conducted in duplicate.

6.2.3 Distillation of Fermented Molasses

Distillation at the Bundaberg distillery, was described in Chapter 3, Section [Link].

To attempt to replicate the two stage process undertaken, laboratory distillation
was first done in a custom made distillation column. This was known as the
“primary” distillation. The distillate collected from this primary distillation was then
subjected to further distillation using laboratory distillation glassware to replicate
the “secondary” pot still distillation undertaken at the Bundaberg distillery.
 166

[Link] Primary Distillation

Distillation was undertaken, using a custom made distillation column, designed to

simulate critical properties of the primary distillation process at the Bundaberg
Distilling Company. This distillation column (Figure 6.1) was constructed by the
Workshop, School of Chemical Engineering, UNSW.

The distillation unit consisted of a stainless steel vessel (Fig 6.1 a) that had an
electric heating element (Fig 6.1 b) inserted directly through the side of the vessel.
The height at which this heating element was located, restricted the vessel to a
minimum operational volume of 6L. The distillation column (Fig 6.1 d) was mounted
on top of this vessel. Components of the distillation unit were joined using stainless
steel connection flanges (Fig 6.1 i) with a heat stable rubber seal used between
each component to ensure the column was airtight.
 167

Figure 6.1 Components of the batch distillation column used for large scale distillation of
fermented molasses. (a) stainless steel vessel, (b) electric heating element unit, (c) thermocouple
readout display, (d) stainless steel column with fibre glass insulation, (e) external copper condenser
(water cooled), (f) stainless steel reflux section with internal copper coiled condenser (water
cooled), (g) thermocouple port, (h) inlet and outlet of cooling water supply, (i) stainless steel
connection flanges

The stainless steel column had a 1/8” Swagelok fitting halfway along its length,
which acted as a thermocouple port allowing the temperature of the column to be
monitored (Fig 6.1 c, g). The column had 31 mm thick fibreglass insulation around
its circumference (Fig 6.1 d), which prevented the rising vapour from condensing
on the inside walls and allowed the column to reach thermal equilibration
efficiently.
 168

An extension of the main section of the column was mounted on the top with an
internal copper coiled condenser (Fig 6.1 f). The presence of the internal
condenser meant that the extended section of the column was under reflux.

The outlet at the top of the reflux section was attached to a second external copper
condenser (Fig 6.1 e) which acted as an added quality and safety precaution,
preventing the loss of any vapour from the top of the column. Both condensers
were water cooled, which was carried under pressure from the laboratory tap using
10 mm PVC piping (Fig 6.1 h). Since the flow rate through the internal condenser
needed to be controlled, the in-flow of water was fed through a rotameter (not
shown), which allowed the rate to be easily manipulated and held constant.

The distillate released from the external condenser was collected in Schott bottles
which were sitting in ice (Fig 6.1 j). Distillates were collected and stored in air tight
containers prior to undergoing secondary distillation (Section [Link]) and then
analysis of volatile compounds.

Fermented molasses medium (48 hr sample, 8 litres) was transferred to the base
vessel. Fermentation solids, including spent yeasts, were not separated and were
transferred into the vessel. Distillation was commenced by turning on the heating
element, where the temperature increased to approximately 100°C The ethanol
concentration of the distillate could be estimated from the thermocouple reading at
the top of the internal condenser (Fig 6.1 g) using standard chemical data from the
equilibration between the liquid and vapour phases of an ethanol solution at
varying temperatures (CRC, 1977). In these calculations, the temperature of
condensation was simply taken as the thermocouple reading at the top of the
internal condenser. The final temperature required at the top of the internal
condenser to give an ethanol concentration of about 50% (v/v) was 93.5- 94°C. As
such, distillation was considered to be complete once the temperature reached
93.5°C. Time was not used as a method of measuring completion as time would
vary depending on the composition of the ferment. The volume of distillate varied
depending on the microbiological content of the ferment, S. cerevisiae containing
ferments returned volumes of between one and two litres while bacteriological
 169

based ferments only produced 200-500mL of distillate. The distillates were stored
in Schott bottles, wrapped with parafilm to ensure an airtight seal.

[Link] Secondary Distillation

Distillates obtained from the primary distillation ([Link]) were subject to a

secondary distillation step to refine the distillate. This done was to simulate the two
step process of distillation at the Bundaberg distillery. Distillation was performed
using scientific glassware for laboratory distillations, representing a pot distillation.
A 1 L round bottom reaction flask (Fig 6.2 a) was used as the pot, heated with a
heating mantle (Fig 6.2 b). A thermometer was attached to monitor the temperature
of evaporated primary distillate (Fig 6.2 c). The condenser (Fig 6.2 d), with
continuous flowing cold water (Fig 6.2 e), was attached and condensate was
collected at specific cuts, based on temperatures, in a receiver flask sitting in ice
(Fig 6.2 f).

Figure 6.2 Secondary distillation apparatus used for distillation of primary distillate. (a)
reaction flask, (b) heating mantle, (c) thermometer, (d) condenser, (e) cooling water, (f) receiver
flask
 170

Samples obtained from the primary distillation stage underwent further separation.
Separation was, as previously mentioned, monitored via temperature. Distillates
collected prior to 78°C were collected as heads, those collected between 78°C and
89°C were the “hearts” and above 89°C represented the “tails”. Distillates were
stored in glass vessels. The hearts were analysed for volatile compounds (method
described in further detail in Section [Link]).

6.2.4 Microbiological Analyses

Yeast growth was monitored by spread inoculating 0.1 mL of serially diluted

samples, in duplicate, onto plates of MEA (Oxoid) (supplemented with 100 µg/mL
oxytetracycline for mixed fermentations). Bacterial growth was monitored by
spread plating 0.1 mL of serially diluted samples, in duplicate, onto plates of MRS
agar (Oxoid) (supplemented with 10 µg/mL cycloheximide for mixed
fermentations). Inoculated plates were incubated for 48 hours at 30°C after which
colonies were counted. In mixed culture fermentations, the colonies of the different
bacterial species were readily distinguishable by their morphology on MRS agar as
reported in Chapter 3.

6.2.5 Chemical and Physical Analyses

[Link] Sample Preparation

Samples of fermented molasses medium for chemical analyses were clarified by

centrifugation at 4°C 20,000g for 20 min (Beckman-Coulter, Fullerton, CA) and the
supernatant collected. The supernatant was then filtered under vacuum through a
0.45 µm filter membrane (Millipore). Samples were stored, at -20°C until analysed.
The filtration step was not done for samples used for the analysis of ethanol and
volatile compounds.

[Link] pH and Brix

These analyses were done on unclarified samples taken directly from the
fermentations using an Activon© optical pH probe for pH and a PAL-1 (ATAGO,
 171

Japan) refractometer for Brix. Analyses were done in duplicate and average values
are reported.

[Link] Sugars, organic acids and ethanol

The concentrations of sugars and organic acids were determined by HPLC as

described in Chapter 5, Section 5.2.6. A coefficient of variation for both the sugars
and organic acid methods was a maximum of + 25% for each analysis.

Ethanol concentrations were determined using an enzymatic UV assay kit with

appropriate standards and control as described by the manufacturer (Cat no.
10176290035; r-biopharm, Roche, Germany). Analyses were done in duplicate and
average values are reported.

[Link] Volatile Compounds

Two methods of analysis for volatile compounds were performed. The first method
was a detailed method, able to analyse both fermented molasses samples and
distillates. The second method analysed only distillates and was performed on
equipment used by laboratory staff at the Bundaberg distillery.

Method 1
Volatile compounds were measured in samples of fermented molasses taken
directly from fermentation vessels and distillates of fermented samples.

Volatiles in samples were measured by head space analysis at the commercial

facility of the Metabolomics Unit of the Australian Wine Research Institute (AWRI),
South Australia. Samples were optimized for analysis by dilution in 10% potassium
hydrogen tartrate buffer, pH 4.5.

The analysis was performed on an Agilent 7890 gas chromatograph equipped with
Gerstel MPS2 multi-purpose sampler and coupled to an Agilent 5975C VL mass
selective detector. Instrument control and data analysis were performed with
Agilent G1701A Revision E.02.00 ChemStation software. The gas chromatograph
was fitted with a 30 m x 0.18 mm Resteck Stabilwax – DA (crossbond carbowax
 172

polyethylene glycol) column with a 5m x 0.18 mm retention gap. Helium (Ultra High
Purity) was used as the carrier gas with linear velocity 24.6 cm/s, flow-rate 0.78
mL/min in constant flow mode. The oven temperature started at 33°C, was held at
this temperature for 4 min, then increased to 60°C at 4°C/min, followed by heating
to give increases of 8°C/min to 230°C when it was held for 5 min.

The conditions of large volume headspace sampling were as follows: The vial and
its contents were heated to 40°C for 10 minutes with agitation (speed 750 rpm, on
time 80 s, off time 1 s). A heated (55°C) 2.5 mL syringe penetrated the septum
(27.0 mm) and removed 2.5 mL of headspace (fill speed 200 μL/s). The contents of
the syringe were then injected into a Gerstel PVT (CIS 4) inlet fitted with a Tenax
TA inlet liner (0.75 mm I.D., pre-conditioned in the GC inlet at 200°C for 1 hour and
then ramped to 350°C to remove contaminates prior to first injection). Prior to
injection, the inlet was cooled to 0°C with liquid nitrogen. The sample was
introduced at 25.0 μL/s (penetration 22.0 mm) using split mode (split ratio 33:1,
split flow 25.78 mL/min). Following capture of analytes on the Tenax liner, the
injector was heated to 330°C at 12°C/min (pressure 24.6 kPa).

The mass spectrometer conditions were as follows: the quadrupole temperature

was set at 150°C, the source was set at 250°C and the transfer line was held at
280°C. Positive ion electron impact spectra at 70eV were recorded in selective ion
monitoring (SIM) mode and Scan mode simultaneously (relative EM volts) with a
solvent delay of 4.0 min. Quantitative data processing was performed with Agilent
G1701A Revision E.02.00 ChemStation software. Samples were prepared and
analysed in a randomized order with a blank run every 10 samples. Analyses were
done in duplicate and average values are reported.

Method 2
Methyl salicylate concentrations were measured in samples of distillates obtained
during distillation in the laboratory, with a sample of commercially available
matured Bundaberg Rum run as a comparative sample.
 173

Samples were analysed using a Varian ® CP-3800 Gas Chromatograph, coupled

with a Flame Ionisation Detection system (GC-FID). A Varian ® fused silica, CP-
WAX 52CB column (30 m x 0.53 mm) was used with nitrogen as the carrier gas.
For each analysis, 1.0µL of sample was injected into the column at 250°C (split
ratio 30). The temperature program was as follows; initial 30°C for 9 min, ramp
20°C/min to 70°C and hold for 3 min, ramp 25°C/min to 190°C and hold for 5 min.
A final ram of 20°C/min to 220 and a final hold for 2.5 min. Simultaneously,
pressure was also adjusted automatically using the following program; initial
pressure 3psi for 9 min, increase 2 psi/min to 5 psi, hold for 14 min and a final
increase by 2 psi/min to 9 psi and hold for 1.8 min. The FID operated at 300°C.

6.3 RESULTS
6.3.1 Growth of Individual Cultures of S. cerevisiae and Species of
Lactic Acid Bacteria

Saccharomyces cerevisiae and the species of lactic acid bacteria were inoculated
into the molasses medium at initial levels of approximately 106 CFU/mL to simulate
what is likely to occur in the commercial operations at the Bundaberg distillery, and
their growth profiles followed during 48 h of fermentation.

Figure 6.3 shows the growth of single cultures of the different species in molasses
medium. Data presented are the averages of analyses from the duplicate
fermentations. The raw data for these fermentations are given in Appendix B.
Saccharomyces cerevisiae grew from initial populations of about 6 x 105 CFU/mL
to maximum populations of approximately 107 CFU/mL at the end of fermentation.
The main fermentable sugar of the medium was sucrose at initial concentrations of
127 mg/g. It was progressively utilized during fermentation, in parallel with the
growth of the yeast and production of ethanol. It was not completely utilized and
residual levels of about 54 mg/ml were present at the end of fermentation (Fig 6.3
a). Smaller amounts of glucose (62 mg/g) and fructose (54 mg/g) were present in
the medium and were partially utilized throughout fermentation. As sugars were
 174

utilised, the ethanol content increased from about 0.0% to 5.1% w/v. The initial
amount of ethanol in the medium came from the starter inoculum. The fermentation
medium became marginally more acidic, with pH changing from 5.3 to a final pH of
5.2.

Figure 6.3 Growth profiles and changes in concentrations of fructose, glucose, sucrose and
ethanol. during fermentation of molasses medium with (a) S. cerevisiae, (b) L. fermentum (c) L.
plantarum (d) Lactobacillus spp. Data are the mean values of analyses from duplicate
fermentations. Standard errors were less than ±6% (population log CFU/mL), ±15% (ethanol) and
±25% (sugars) of the mean values.

Lactobacillus fermentum (Figure 6.3 b), L. plantarum (Figure 6.3 c) and the
Lactobacillus spp. (Figure 6.3 d) grew from initial populations of about 106-107
CFU/mL to maximum levels of approximately 5 x 108 CFU/mL. Despite this growth,
the fermentable sugars, sucrose, glucose and fructose were only slightly utilised
during fermentation (initial total sugar concentrations of about 260 mg/g to a final
total sugar concentration of about 220 mg/g). In addition, the fermentations for L.
 175

plantarum produced some unusual data such as a small increase in fructose

levels after 8 hr of fermentation, followed by its utilisation, and an increase in
sucrose concentration towards the end of fermentation. These same trends were
found in each of the duplicate fermentations (Appendix B). The concentration of
ethanol increased from an initial value of approximately 0.0% to only 0.3% by 48 h.
During these bacterial fermentations, the pH became more acidic, decreasing from
about pH 5.3 to 4.2 by 48 h for all three species.

6.3.2 Growth of Mixed Cultures of S. cerevisiae and Species of

Lactic Acid Bacteria

Molasses medium was fermented with mixed cultures of S. cerevisiae and different
species of lactic acid bacteria as given in Table 6.1. Figure 6.4 shows the growth of
these mixed cultures of the different species in molasses medium. Data presented
are the averages of analyses from the duplicate fermentations. The raw data for
these fermentations are given in Appendix B.

Figure 6.4 (a) shows the mixed culture of S. cerevisiae and L. fermentum. The
populations of S. cerevisiae increased from initial values of about 106 CFU/mL at
the beginning of fermentation to about 107 CFU/mL at the end of fermentation and
those of L. fermentum increased from approximately 107 CFU/mL to greater than
108 CFU/mL. This growth was accompanied by an almost complete utilization of
both fructose and glucose and partial utilization of sucrose. As sugars were
utilised, the ethanol content increased from 0.1 % to 7.9 %. The fermentation
medium became slightly more acidic, with pH changing from 5.2 to 4.7.

A mixed culture of S. cerevisiae and the Lactobacillus spp. is shown in Figure

6.4(b). Populations of the yeast increased from about 106 CFU/mL to about 107
CFU/mL and those the Lactobacillus spp. increased from approximately 107
CFU/mL to a final population of 5.0 x 108 CFU/mL. Glucose and fructose were only
partially utilized (fructose 51 mg/g – 26 mg/g; glucose 60 mg/g – 31 mg/g) during
fermentation. Sucrose levels decreased progressively from initial values of
approximately 140 mg/ml to 80 mg/ml. Ethanol content rose, in conjunction with the
 176

sugar utilisation, from 0.1% to final content of 6.0%. The fermentation medium
became slightly more acidic, with pH changing from 5.2 to 4.5.

Figure 6.4 Growth profiles of mixed culture fermentations of molasses medium (a) S. cerevisiae
and L. fermentum, (b) S. cerevisiae and Lactobacillus spp. (c) S. cerevisiae, L. fermentum and
Lactobacillus spp. showing changes in concentrations of fructose, glucose, sucrose and ethanol.
 177

The mixed culture, consisting of S. cerevisiae and both lactic acid bacteria (L.
fermentum and Lactobacillus spp.) is shown in Figure 6.4 (c). Populations of S.
cerevisiae increased from about 106 CFU/mL to 6.3 x 107 CFU/mL after 48 h. Total
bacterial populations were 1.6 x 107 CFU/mL initially and increased to above 5 x
108 CFU/mL at the completion of fermentation. Individual bacterial populations
were not determined. Sucrose was progressively utilised throughout the
fermentation from initial concentrations of 148 mg/g to final concentrations of 64
mg/g. Fructose was partially utilised (initial concentration 43 mg/g) with a final
concentration of 27 mg/g. Glucose concentrations also decreased, in conjunction
with the other sugars, from 73 mg/g to a final concentration of 38 mg/g. Ethanol
content increased from 0.1% to a final level of 7.7% after 48 h. As with all previous
fermentations, the fermentation medium became slightly acidic, with pH changing
from 5.2 to 4.2.

[Link] Organic acids

Samples taken during the laboratory scale fermentations performed above (Section
6.3.1 & 6.3.2) were analysed for organic acids. At the beginning of fermentation,
the fermentation medium contained oxalic acid at 0.28 ± 0.08 mg/mL; citric acid at
1.60 ± 0.8 mg/mL; tartaric acid at 3.33 ± 1.2 mg/mL; malic acid at 4.28 ± 3.8
mg/mL; lactic acid 3.27 ± 0.7 mg/mL; acetic acid 2.01 ± 0.36 mg/mL; propionic acid
at 0.23 ± 0.18; butyric acid 9.46 ± 2.5 mg/mL; and succinic acid 20.81 ± 2.0
mg/mL. There were considerable variations in determining the concentrations of
these organic acids despite repeated assays. Possibly, some constituents in the
molasses were causing interference with the HPLC method used. One-way single
factor analysis of variance and t-test were used to determine significant differences
between means using Microsoft Excel. Significant differences in the concentrations
of organic acids were considered when p<0.05. Appendix C shows the raw data
with notations on all results that had significant changes (p<0.05)

Figure 6.5 shows changes in the concentrations of organic acids during

fermentations with individual cultures of S. cerevisiae and the species of lactic acid
 178

bacteria. The concentrations of oxalic, citric, tartaric and propionic acids did not
show major changes during fermentation and are not shown in the figure.

Fermentations undertaken by, S. cerevisiae, showed a decrease in malic acid at 48

hours while lactic, acetic, butyric and succinic acids all showed small but fluctuating
increases between 24-48 h. The L. fermentum fermentation gave decreases in
malic and succinic acids but only the 24 h data were statistically significant. There
were increases in concentrations of lactic and butyric acids at late stage
fermentation (>24 h). Both L. plantarum and Lactobacillus spp. fermentations gave
decreases in succinic acid concentrations and increases in the concentrations of
lactic and acetic acids. Lactobacillus spp. fermentation, unlike the other three
fermentations, had a significant decrease of butyric acid (48 h).

Figure 6.5 Fermentation of molasses medium by individual microbial species (a) S. cerevisiae, (b)
L. fermentum (c) L. plantarum (d) Lactobacillus spp. showing changes in concentrations of five
organic acids (malic, lactic, acetic, butyric and succinic) at 4 time stations (0h, 16h, 24h, 48h). One-
way single factor analysis of variance and t-test were conducted and (*) indicates significant
differences between T = 0h and subsequent time points, at a 95% confidence level.
 179

Changes in the concentrations of organic acids during mixed culture fermentations

are shown in Figure 6.6. There were clear increases of lactic, butyric and succinic
acids in all three fermentations, although not always statistically significant.

Figure 6.6 Fermentation of molasses medium by mixed microbial cultures (a) S. cerevisiae & L.
fermentum, (b) S. cerevisiae & Lactobacillus spp. (c) S. cerevisiae, L. fermentum & Lactobacillus
spp. showing changes in concentrations of five organic acids (malic, lactic, acetic, butyric and
succinic) at 4 time stations (0h, 16h, 24h, 48h). One-way single factor analysis of variance and t-test
were conducted and (*) indicates significant differences between T = 0h and subsequent time
points, at a 95% confidence level.
 180

[Link] Volatile compounds

The concentrations of volatile compounds produced during the laboratory

fermentations were determined in (a) samples of the fermentation culture taken at
the mid-stage (16h/24h) and end (48h) of fermentation and (b) in distillates of the
fermented medium at 48 h.

Table 6.2 shows some main volatiles produced during molasses medium
fermentation by single and mixed cultures S. cerevisiae and lactic acid bacteria.

The main volatiles (apart from ethanol, discussed previously) produced during
fermentation were the higher alcohols; 3-methylbutanol, 2-methylpropanol and 2-
methybutanol. These were principally produced by S. cerevisiae (Figure 6.7) with
very little produced by either of the three species of lactic acid bacteria, especially
when the basal levels of these alcohols in the fermentation medium (0 h data) are
taken into consideration. These three higher alcohols were also predominant in the
mixed yeast- lactic acid bacteria fermentations, although the mixed fermentations
with Lactobacillus spp. and both L fermentum and Lactobacillus spp. decreased
the levels of 2-methylbutanol and 3-methylbutanol by about 50%.

Ethyl acetate and ethyl propanoate were the main esters found in these ferments,
being produced almost exclusively by S. cerevisiae (Table 6.2). Single cultures of
the lactic acid bacteria did not produce these acetates or other esters. 3-methyl
butyl acetate and phenyl ethyl acetate were also measured but were not detected
in any of the fermentations except trace amounts for the late stage of the S.
cerevisiae fermentation or if combined with the bacteria in fermentation. Production
of ethyl acetate and ethyl propanoate also occurred in the mixed culture
fermentations with some apparent increases for ethyl acetate in these cases.
Insufficient samples were analysed to make firm conclusions about these
increases.
 181

Table 6.2 Some volatiles produced during molasses medium fermentation by single and mixed cultures of S. cerevisiae and lactic acid
bacteria
Sample Alcohols (µg/L) Ethyl Esters (µg/L)
Time
2-methyl 1- 2-methyl 3-methyl Ethyl Ethyl Ethyl Ethyl
Starter culture station hexanol
propanol butanol butanol butanol acetate propanoate butanoate hexanoate
(h)
Fermentation medium,
648 - - 5824 49 - -
no culture
16 2001 - 797 7491 49 - 5 - -
S. cerevisiae
48 11996 815 9809 29462 88 479 69 8 -

16 975 30 44 5981 102 - - - -

L. fermentum
48 1746 526 256 6299 123 - - - -

16 782 - - 5764 51 - - - -
L. plantarum
48 1069 - 148 6271 77 7 - - -

Lactobacillus 16 1093 78 213 6140 131 - - - -

spp.
48 2003 607 368 6490 130 - - - -

S. cerevisiae & 24 3766 38 2374 11850 48 60 - - -

L. fermentum
48 14162 790 10151 33972 88 767 11 11 -

S. cerevisiae & 24 1651 - 647 7121 54 - - - -

Lactobacillus
spp 48 7692 790 5897 21618 142 578 - 9 35
S. cerevisiae, 24 3290 68 1637 9144 101 - - - -
L. fermentum &
Lactobacillus
48 7219 814 4768 19335 169 640 27 7 36
spp
-below limit of quantification
 182

Figure 6.7 Key alcohols (A) 2-methylpropanol, (B) 2-methylbutanol and (C) 3-methylbutanol,
produced during molasses medium fermentation by single and mixed cultures. (i) S. cerevisiae (16
h), (ii) S. cerevisiae (48 h), (iii) L. plantarum (16 h), (iv) L. plantarum (48 h), (v) L. fermentum (16 h),
(vi) L. fermentum (48 h), (vii) Lactobacillus spp. (16 h), (viii) Lactobacillus spp. (48 h), (ix) S.
cerevisiae & L. fermentum (24 h), (x) S. cerevisiae & L. fermentum (48 h), (xi) S. cerevisiae &
Lactobacillus spp. (24 h), (xii) S. cerevisiae & Lactobacillus spp. (48 h), (xiii) S. cerevisiae, L.
fermentum & Lactobacillus spp. (24 h), (xiv) S. cerevisiae, L. fermentum & Lactobacillus spp. (48 h).

As mentioned in Section [Link], distillates of the laboratory fermentations were

analysed for volatiles by two methods, one as used at the AWRI (to determine
levels of specific alcohols, ethyl esters and acetates, known to be important in rum
 183

flavour and composition). and the other as routinely used by the quality assurance
laboratory at the Bundaberg distillery to detect, among other volatile compounds,
the specific presence of methyl salicylate, a methyl ester that the Bundaberg
distillery deemed to be of importance in the flavour of Bundaberg Rum.

Table 6.3 Higher alcohols detected in the distillates of laboratory fermentations of molasses
medium with combinations of Saccharomyces cerevisiae and species of lactic acid bacteria
Sample Alcohols (µg/L)
2-methyl 2-methyl 3-methyl
Starter culture 1-butanol hexanol
propanol butanol butanol

S. cerevisiae 113024 7105 90196 220906 422

L. fermentum 31599 15372 14548 25769 2270

L. plantarum 300291 6754 209481 566322 7008

Lactobacillus
28162 12706 12432 22453 1747
spp.
S. cerevisiae & L. $
135117 7718 102468 280548 692
fermentum
S. cerevisiae &
56371 6106 46204 124944 783
Lactobacillus spp
S. cerevisiae, L.
fermentum & 42058 4378 26621 79127 507
Lactobacillus spp
$
Bundaberg Rum 91552 9382 93009 364888 900

-below limit of quantification

$ above calibration limit

The main higher alcohols found in the distillates were 2-methylpropanol, 2-

methylbutanol and 3-methylbutanol, with lesser amounts of butanol and hexanol
(Table 6.3). With the exception of butanol, the highest amounts of these alcohols
were found in distillates of the fermentation with L. plantarum, followed by S.
cerevisiae as illustrated in Figure 6.8. The lowest amounts were present in
distillates of fermentations with L. fermentum and Lactobacillus spp. Distillates from
the combined fermentations of S. cerevisiae with Lactobacillus spp. or S.
cerevisiae, L. fermentum and Lactobacillus spp. gave lesser amounts of these
alcohols than that from a single S. cerevisiae fermentation, but higher amounts
 184

were found in distillates of from the S. cerevisiae and L. fermentum fermentation.

Parallel analyses of samples of a finished product of Bundaberg rum confirmed the
predominance of 3-methyl butanol, 2-methyl butanol and 2-methyl propanol in this
product along with lesser amounts of 1-butanol and hexanol.

Figure 6.8 Selected volatile compounds (A) 2-methylpropanol, (B) 1-butanol, (C) 2-methyl
butanol, (D) 3-methyl butanol, (E) ethyl propanoate, (F) methyl salicylate, detected in laboratory
scale distillations of molasses medium fermented using differing starter culture combinations (i)
S. cerevisiae, (ii) L. fermentum, (iii) L. plantarum, (iv) Lactobacillus spp. (v) S. cerevisiae & L.
fermentum, (vi) S. cerevisiae & Lactobacillus spp. (vii) S. cerevisiae, L. fermentum &
Lactobacillus spp. (viii) Bundaberg Rum
 185

Tables 6.4 and 6.5 show the concentrations of esters and other acetates in the
distillates. The main esters found in the distillates were ethyl butanoate, ethyl
acetate, ethyl propanoate, ethyl 2-methylpropanoate, ethyl butanoate, ethyl 2-
methylbutanoate, ethyl 2-methylbutanoate, ethyl 3-methylbutanoate and ethyl
propanoate as also found for the sample of Bundaberg rum. Highest
concentrations were produced individually by S. cerevisiae and L. plantarum with
much lesser amounts being produced by either L. fermentum or Lactobacillus spp.
Notably, higher concentrations of ethyl 2-methylpropanoate, ethyl 2-
methylbutanoate and methyl salicylate were produced by the lactic acid bacteria
than S. cerevisiae. The Lactobacillus spp. gave three time more methyl salicylate
than either L. fermentum or L. plantarum, (Figure 6.8/Table 6.4). None of the lactic
acid bacteria individual fermentations produced ethyl-3-methylbutanoate. The
combination of S. cerevisiae and L. fermentum gave the highest concentrations of
most volatiles studied with the exceptions being ethyl-3-methylbutanoate and 2-
methyl-propylacetate.
 186

Table 6.4 Esters detected in the distillates of laboratory fermentations of molasses medium with combinations of Saccharomyces
cerevisiae and species of lactic acid bacteria
Methyl Ester
Sample Ethyl Esters (µg/L)
(µg/L)
Ethyl 2- Ethyl 3-
Ethyl Ethyl Ethyl 2-methyl Ethyl Ethyl Methyl
methyl methyl #
Starter culture acetate propanoate propanoate butanoate hexanoate salicylate
butanoate butanoate

S. cerevisiae 113024 7105 90196 220906 422 28 148 4.8

L. fermentum 31599 15372 14548 25769 2270 - - 21.6

L. plantarum 300291 6754 209481 566322 7008 - 1039 26.0

Lactobacillus
28162 12706 12432 22453 1747 - - 61.8
spp.
S. cerevisiae & $
135117 7718 102468 280548 692 15 262 5.7
L. fermentum
S. cerevisiae &
Lactobacillus 56371 6106 46204 124944 783 20 115 4.4
spp
S. cerevisiae, L.
fermentum &
42058 4378 26621 79127 507 21 88 14.9
Lactobacillus
spp
$
Bundaberg Rum 91552 9382 93009 364888 900 32 123 21.9

-below limit of quantification

$ above calibration limit
# methyl salicylate results obtained via method 2 (Chapter 6, Section [Link])
 187

Table 6. 5 Other acetates detected in the distillates of laboratory fermentations of molasses

medium with combinations of Saccharomyces cerevisiae and species of lactic acid bacteria
Sample Acetates (µg/L)
2-methyl 2-methyl 3-methyl 2-phenyl
Hexyl
Starter culture propyl butyl butyl ethyl
acetate
acetate acetate acetate acetate

S. cerevisiae 59 105 545 48 38

L. fermentum - 10 - 35 -

L. plantarum 91 239 1794 - 19

Lactobacillus
8 12 - 28 -
spp.
S. cerevisiae & L.
49 105 932 - -
fermentum
S. cerevisiae &
25 43 171 - 31
Lactobacillus spp
S. cerevisiae, L.
fermentum & 23 31 138 25 28
Lactobacillus spp

Bundaberg Rum 5 32 258 - 5

-below limit of quantification

 188

6.4 DISCUSSION

The investigations reported in this Chapter were designed to determine if the

species of lactic acid bacteria described in Chapters 3, 4 and 5 impacted on the
growth of the starter yeast, S. cerevisiae, during fermentation of molasses medium
for rum production. As one practical measure of such impact, an aim was to
examine key volatiles in distillates of the fermentations. The production of such
distillates in sufficient volume necessitated relatively large laboratory scale
fermentations (approximately 8 L volumes for each fermentation) and the
development of a distillation column and process with sufficient capacity to manage
such volumes. These scale up operations brought unforeseen challenges in the
preparation of media and starter cultures for the fermentations, management of the
fermentations and, finally, the distillation process. In addition, the molasses
medium (molasses plus dunder) was a complex matrix that contained unknown
factors which, on some occasions, caused unexpected interference with the
methods of chemical analyses. Although individual fermentations were conducted
in duplicate and various microbiological and chemical analyses were conducted in
at least duplicates some variable data were obtained. It is with these limitations
that conclusions of the investigations are discussed.

As an individual culture, S. cerevisiae fermented the molasses medium with an

expected profile - growth to populations of about 107 CFU/mL, utilization of sugars
(sucrose, glucose and fructose) and production of ethanol. By the end of
fermentation, approximately 140 mg/g of fermentable sugars had been utilised
giving 5-6% ethanol. However, only about 50 % of the available sugars were
utilized, raising the question as to what might be limiting the fermentation and
production efficiency. The molasses medium had a high initial content of organic
acids (approximates; oxalic acid at 0.3mg/mL; citric acid at 2.0mg/mL; tartaric acid
at 3.0 mg/mL; malic acid at 4.0 mg/mL; lactic acid 3.0 mg/mL; acetic acid 2.0
mg/mL; propionic acid at 0.2; butyric acid 9.0 mg/mL; and succinic acid 20.0
mg/mL). This is consistent with previously reported values for molasses (Chapter
2, Table 2.2; Bruijn & Vanis, 1972; Pislor et al., 2009) and dunder, (Chapter 2,
 189

Table 2.4; Chapter 5, Table 5.6; Bories et al.,1988) which represented 7.5% of the
molasses medium composition. These background levels made it difficult to assess
the small changes to their concentrations that might have occurred due to yeast
metabolism (Figure 6.5 & Figure 6.6), but the apparent increases in succinic and
acetic acids and decrease in malic acid are consistent with S. cerevisiae
metabolism (Swiegers et al., 2005). The main higher alcohols produced by S.
cerevisiae from sugar metabolism are 1-propanol, 2-butanol, 1-butanol, 2-methyl
butanol, 3-methylbutanol, 2-methylpropanol, amyl alcohol (Timmer, et al., 1971;
Suomalainen & Lehtonen, 1979; Swiegers et al., 2005, Abbas, 2006). Previous
research, summarised in Chapter 2, Table 2.6 lists the important higher alcohols in
rum as 1- propanol, 1-butanol, 2-methyl propanol, 2-methylbutanol, 3-
methylbutanol (Liebich et al.,1970; Lehtonen & Suomalainen, 1977; Sampaio et al,
2008). Esters of significance in rum flavour are summarised in Chapter 2, Table
2.6. Of particular importance to the Bundaberg distillery is methyl salicylate which
has previously been detected in Jamaican rum (Liebich et al, 1970) and a previous
study of Bundaberg Rum (Allan, 1972). It is known to give “winter green” sensory
notes.

All three species of lactic acid bacteria grew well in the molasses medium,
reaching populations over 108 CFU/mL which is about 10 fold higher than the
maximum population for S. cerevisiae (Figure 6.3). Despite such strong growth,
these bacteria used little of the fermentable sugars, raising the question as to what
they utilize as major growth substrates. As discussed in Chapter 5, these are likely
to be the free amino acids that arise from the added dunder. Nevertheless some
small amounts of fermentable carbohydrates were utilized and probably accounted
for clear increases in the concentration of lactic acid and to a lesser extent acetic
acid that was observed in ferments with these bacteria. Such acid production
probably accounted for the decrease in pH of the fermentation to about 4.2 during
these fermentations. Depending on species and strain, some lactic acid bacteria
also utilize organic acids such as malic and citric (Bartowsky & Henschke, 1995;
Fleet, 2003) and there was evidence for malic acid utilization by L. fermentum
possibly by malolactic fermentation. As expected, none of the lactic acid bacteria
 190

produced any significant amounts of ethanol. Analysis of ferments showed,

generally lower concentrations of higher alcohols when compared to those
produced by S. cerevisiae. Hexanol was the exception to this trend, with both L.
fermentum and Lactobacillus spp. producing greater concentrations than the yeast
starter culture. Of the higher alcohols analysed in this study, the most prevalent
were 2-methylpropanol, 2-methylbutanol and 3-methylbutanol, confirming their
presence in rum fermentations. Similar proportions to those from pure S. cerevisiae
fermentation distillates were seen in the sample of finished Bundaberg Rum
product. The concentrations of four key higher alcohols present in Bundaberg rum,
when compared to previous research summarised in Chapter 2, Table 2.6, are
average to low compared to other rums. 2-methyl propanol and 1- butanol are both
within previously reported ranges of rums or fermented molasses. 2-methyl butanol
and 3-metyl butanol, however, are both below the previously reported values. The
exception is the fermentation of molasses medium by L. plantarum, where 2-methyl
butanol is within the narrow range reported by Liebich et al (1970) and Lehtonen
and Suomalainen (1977). 2-methylbutanol sensory characteristics are described as
whisky-like or roasted wine.

A small about of the ethyl ester, ethyl acetate, was detected at 48h in the L.
plantarum fermentation. There is an unusual variation between the higher alcohols
detected in the ferment of L. plantarum when compared to those detected in the
associated distillate. An analytical error during preparation of L. plantarum distillate
may have been at fault for the higher results reported for the when distillate
compared with those obtained from the ferment.

Conclusions about the effect of the lactic acid bacteria on the growth of S.
cerevisiae may be compromised to some extent by the relatively high initial
inoculum of the bacteria relative to S. cerevisiae - about 10 fold higher (Figure 6.4).
Nevertheless, in all cases the yeast reached maximum populations of about 107
CFU/mL by 48 h which is similar to the maximum level when cultured singly. In
comparison to the observations during the ecological studies undertaken at the
Bundaberg distillery, initial populations of yeast were about 107 CFU/mL while
 191

bacterial populations were variable, being between 101 – 107 CFU/mL. The growth
of the lactic acid bacteria was not compromised by the presence of the yeast or the
concentrations of ethanol that developed during these laboratory fermentations
(Figure 6.4). There was evidence for all three combined fermentations that
presence of the bacteria increased sugar fermentation and ethanol production by
S. cerevisiae. This was more apparent in the S. cerevisiae & Lactobacillus spp.
and the S. cerevisiae, L. fermentum and Lactobacillus spp. combinations, where
final ethanol concentrations were 6.0% and 7.7 % respectively. The reasons for
this are not clear but could be due to the lower pH of these fermentation resulting
from their production of lactic and acetic acids. As mentioned previously, the
molasses medium is a complex matrix that could have inhibitors of yeast
metabolism (Kampen, 1975; Lehtonen & Suomalainen, 1977) and it is possible that
the bacterial biomass may adsorb or remove such inhibitors. It was mentioned in
Chapter 5 that dunder had a stronger inhibitory effect on yeasts than the lactic acid
bacteria experienced. Further research is needed to investigate this phenomenon
as there is a clear benefit to fermentation efficiency and ethanol production by
these mixed fermentations

When compared to the distillate produced from the single culture S. cerevisiae
ferment, both mixed ferments consisting of S. cerevisiae & Lactobacillus spp. and
S. cerevisiae, L. fermentum and Lactobacillus spp. produced lower levels of most
volatile compounds analysed (Figure 6.3, 6.4, 6.5).The exceptions were the
production of hexanol and ethyl 2-methylbutanoate. The fermentation consisting of
S. cerevisiae, L. fermentum and Lactobacillus spp. produced more methyl
salicylate when compared to the S. cerevisiae only distillates. The esters of ethyl
acetate and ethyl butanoate (also known as ethyl n-butyrate) have similar
concentrations for some of the laboratory distillates to those seen in previous
research (Chapter 2, Table 2.6). The S. cerevisiae and S. cerevisiae & L.
fermentum fermentations, along with the sample of Bundaberg rum were within (or
close to) the previously reported ranges. The variation when compared with
previously reported results is most likely due to most research on flavour volatiles
in rum concentrating on Jamaican based distilleries. Methyl salicylate, specifically
 192

analysed due to its importance to the Bundaberg distillery and possible contribution
to the distinctive character of its rum, has an odour threshold of 0.04µg/L (Amerine,
et al. 1965). The distillates of the S. cerevisiae & L. fermentum ferments went
against this trend displayed by the other mixed ferments and generally produced
higher levels of volatiles than the yeast distillates, with a few exceptions, 2-methyl
propylacetate, 2-phenyl ethylacetate, hexyl acetate and ethyl 3-methylbutanoate.

Conclusion points

The use of laboratory scale fermentation and distillation apparatus has been used
to determine, on a smaller scale, the key elements of microbiological and chemical
characteristics of fermentation (Makanjuoula et al., (1992) – whisky). However,
there have only been a couple of research groups that have previously investigated
rum specific experiments using cane molasses (Arroyo, 1945; and a series of
papers by researchers at a “Rum Pilot Plant” at the University of Puerto Rico –
Cacho et al., 1986; Cacho & Murphy, 1988).

The type of yeast used for fermentation has a significant effect on the formation of
alcohols. Schizosaccharomyces pombe, for example, produces relatively little of
the higher alcohols (Parfait and Jouret, 1975). The laboratory scale fermentations
undertaken during this research did not use a variety of yeasts, only the starter
culture obtained from the Bundaberg distillery. The occurrence of co-flocculation,
such as those seen in conjunction with brewery yeasts (Zarattini, et al, 1993;
Domingues et al, 2000) may also be occurring in at the distillery.

Project limitations surround distillation which would impact on the results. There is
considerable technicality in the art of distillation which was limited in this study due
to the size of fermentations used and the physical limitation of the laboratory scale
distillation unit. Limitations also exist as there was no study of the impact of
ageing/maturation in this research study. There is considerable research into the
effects of aging in oak of raw distillates which is known to mellow and also develop
some of those flavours (de Torres et al., 1987; Mosedale, 1995; Mosedale &
Puech, 1998; Singleton, 1995). Other factors such as blending, charcoal filtering
 193

and dilution were also outside of the limitations of this study. The distillates
produced in this study, can, however be used as a guide to understand what
attributes the finish product can potentially exhibit.

The impact of L. plantarum would have been interesting to observe, either

individually, in conjunction with S. cerevisiae and/or as a more complex mixture
with the two other lactic acid bacteria. Lactobacillus plantarum was assessed, and
excluded, prior to chemical analysis based purely on population levels and
frequency of incidence in the ecological surveys performed in Chapters 3 and 4.
Perhaps those lower populations were still important for flavour development.
 194

CHAPTER 7 CONCLUSIONS

This thesis describes a systematic investigation of the microbial ecology of molasses

fermentation associated with the production of rum at the Bundaberg Distilling
Company, Bundaberg, Australia. A combination of cultural and molecular methods
was used to isolate, identify and enumerate the microbial species associated with
the rum production process. Raw materials and in-process samples from the
different stages of the process were examined in surveys conducted during 2006 to
2010. These analyses were followed up with laboratory investigations to gain a
deeper understanding of the microbiology of the process.

Chapter 2 examined the background literature on the microbiology and

biotechnology of rum production, in general. Although significant advances had been
made in describing the chemistry of rum flavour and quality, detailed studies
describing the occurrence and growth of microorganisms during molasses based
rum fermentations were lacking. While the yeast, S. cerevisiae (and to a lesser
extent Schizosaccharomyces pombe) was mainly responsible for the fermentation
various qualitative observations suggested a possible role of bacteria in the process.

Chapter 3 presented an overview of the chain of operations in the production

process, along with data obtained from on-site surveys of the microbial ecology of
the operation, conducted in 2006, 2008 and 2010. It was concluded from these
studies that the yeast S. cerevisiae was the predominant microorganism in the
fermentations. This was consistent with its preparation and inoculation as a starter
culture. In addition to this yeast, lactic acid bacteria were consistently isolated from
the fermenting molasses medium and grew to populations of 107- 108 CFU/mL. The
main species found were Lactobacillus fermentum, Lactobacillus plantarum,
Lactobacillus brevis and a Lactobacillus spp. These bacteria entered at the stage of
propagation. An overview of the cleaning and sanitation operations was undertaken
and revealed, while generally efficient, there were some irregularities and oversights.
It was concluded that these bacteria probably originated, at some time, as
contaminants from the molasses or improperly stored dunder and established
residence as endemic, indigenous flora in the myriad of pipework and other areas,
 195

such as heat exchangers, which were difficult to thoroughly clean using CIP
programs. Heat exchangers which circulate cooling and heating waters throughout
the production chain, appear to be of importance in cross contamination of the
endemic microflora from molasses, dunder and contaminated parts of the system to
freshly prepared fermentation medium. Further targeted investigation is
recommended to determine the precise locations of these contamination points in
the production chain. With such information, revised cleaning and sanitation
programs and revised quality assurance and control programs could be developed to
better manage process microbiology.

Chapter 4 confirmed the endemic association of lactic acid bacteria with the
molasses fermentation process by examining the microbial flora in buffer tanks, at
the completion of fermentation over an extended period of time (8 months). It
confirmed the main species present as Lactobacillus fermentum, Lactobacillus
plantarum, Lactobacillus brevis and a Lactobacillus spp. and that these bacteria
occurred at populations of 105 – 108 CFU/mL, which are quantitatively significant
and, therefore, likely to impact on process efficiency and rum product flavour and
quality. Fluctuations in the presence and populations of the individual species were
observed and would lead to variations in product quality. Such fluctuations confirmed
the randomness of the indigenous flora contamination and the need for research to
better understand the circumstance that lead to their residence within the distillery.
Analyses conducted in this chapter also confirmed the absence of other bacteria
(e.g. Zymomonas, Propionibacterium and Clostridium species) from the
fermentations. However, the microbiological analyses conducted in this Chapter as
well as those presented in Chapter 3 were based on standard procedures of cultural
isolation and may fail to detect the presence of some species. It is recommended
that further research and examination of buffer tank samples be conducted with
culture independent molecular methods to eliminate the possibility that other
bacterial species may also contribute to the molasses fermentations.

Chapter 5 established a greater understanding of dunder, both microbiologically and

chemically. Although routine analysis of dunder taken immediately after the still
suggested that it was sterile and free of viable microorganisms, it inevitably
developed microbial growth on storage. A strain of unidentifiable Lactobacillus spp.,
similar to that isolated from the molasses fermentations reported in Chapters 3 and
 196

4, was consistently isolated from stored dunder. This strain produced extracellular
material on heat stress that may confer an advantage of heat tolerance. This may be
a novel species and further research is recommended to determine the taxonomic
status of this isolate and its physiological characteristics. Chemical analyses of the
dunder confirmed its acidic nature and that it was rich in its composition of organic
acids and free amino acids. These components, especially the amino acids, may
form the substrates for the growth of lactic acid bacteria. However, addition of
dunder to the molasses medium did not enhance the growth of the rum fermentation
yeast, S .cerevisiae, or strains of lactic acid bacteria (L. fermentum, L. plantarum
Lactobacillus spp.). In contrast, at concentrations of 10% and above, it decreased
the growth of these organisms, and would detract from the efficiency of molasses
fermentation. The reasons for the use of dunder in molasses based rum
fermentations needs to be critically questioned. While the data of this Chapter have
advanced knowledge about dunder, it is largely at an exploratory level, and further
more detailed research is needed to better understand its microbiological and
chemical properties.

Chapter 6 combined the learned knowledge of the microflora of the Bundaberg

distillery with small scale laboratory based fermentations and distillations to
determine any impact of bacterial contamination on yeast based molasses
fermentations. Using the bacterial species isolated in Chapters 3, 4 and 5, this
impact was measured by comparing fermentations of bacteria, or bacteria and yeast
combinations, with a baseline yeast only fermentation. Cultural and chemical
analyses showed that generally lactic acid bacteria had no effect on growth of yeasts
and led to more sugar being used by yeast, higher ethanol concentrations and the
production of important organic acids, key higher alcohols, esters and acetates.
Especially when those higher alcohols, esters and acetates are integral to the
flavours of dark rums, such as Bundaberg Rum. The unidentifiable Lactobacillus spp.
produced the highest concentrations of the desired methyl ester, methyl salicylate, a
volatile compound of industrial significance to the Bundaberg distillery.

Finally, the fermentations occurring at the Bundaberg distillery are not pure yeast
culture fermentations as the company expected. It is a mixed fermentation of the
starter culture yeast and uncontrolled contributions from an indigenous flora of lactic
acid bacteria. The lactic acid bacteria probably contribute to the unique and
 197

distinctive flavour of Bundaberg rum, which the Bundaberg Distilling Company is

adamant about preserving. It is recommended that the Bundaberg Distilling
Company conduct further research to better understand the origin and presence of
these lactic acid bacteria and to more precisely determine how the individual species
might affect process efficiency and product flavour. Quality assurance and control
programs could then be developed to manage the positive or negative impacts of
these species. This could also lead to selection of strains of these bacteria for more
detailed characterisation and, the possibility of their development as specific starter
cultures for use along with S. cerevisiae to have greater control over process
efficiency and product quality and consistency.
 198

CHAPTER 8 REFERENCES

Websites:
AB Mauri ([Link])
Inner Circle ([Link])
Lallemand ([Link])
Lark Distillery ([Link])
The Hoochery ([Link])
The Kimberley Rum Company ([Link])

Journals/Books:
Abbas, C., (2006). Production of Antioxidants, Aromas, Colours, Flavours and
Vitamins by Yeasts. In: The Yeast Handbook Editors Querol, A & Fleet, G.H
Springer, Verlag Berlin, Heidelberg
Albet, J., Pislor, E., & Pontalier, PY., (2009). Process for recovering carboxylic acids
from sugar cane industry by-products. In: 2nd International Congress on Green
Process Engineering, 2nd European Process Intensification Conference, 14-17 June
2009. Venice, Italy.
Al-Kandari, D. & Jukes, D. J. (2009). A situation analysis of the food control systems
in Arab Gulf Cooperation Council (GCC) countries. Food Control, 20 (12). p 1112-
1118.
Allan, C.H., (1906). Report on manufacture of Jamaica. West Indian Bulletin 7 p 141-
142
Allan, D.A., (1972). The composition of an Australian Rum (Thesis) University of
Queensland.
Allan, D.A., (1973). A New Approach to the Identification of Flavour Compounds in
Rum. Australian Wine, Brewing and Spirit Review January 25 p 42-44.
Amann, R.I., Ludwig, W., & Schleifer, K-H., (1995). Phylogenetic Identification and In
Situ Detection of Individual Microbial Cells without Cultivation Microbiological reviews
59 (1) p 143-169
Amerine, M.A., Pangborn, R.M., & Roessler, E.B (1965) Principles of sensory
evaluation of food. Academic Press, New York, NY.
Amorim, H.V., Basso, L.C., & Lopes, M.L., (2009). Sugar cane juice and molasses,
beet molasses and sweet sorghum: Composition and usage. Chapter 5 In: In: The
Alcohol Textbook (Ed. Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C.) 5th
Edition. Nottingham University Press, Lallemand Ethanol Technology, Nolfolk. UK.
 199

Ardhana, M.M., & Fleet, G.H., (2003). The microbial ecology of cocoa bean
fermentations in Indonesia. International journal of food microbiology 86 (1) p 87-99
Arroyo, R., (1942). The manufacture of rum. Sugar June p 29-31
Arroyo, R., (1943). Ethanol Fermentation of Molasses Sugar 38 p 18-19
Arroyo, R., (1945a). Studies on Rum Bulletin University of Puerto Rico Agricultural
Experiment Station, Rio Piedras
Arroyo, R., (1945b). The production of Heavy Bodied Rum Sugar 40 p 34-39
Arroyo, R., (1945c). Production of Heavy Rums. United States Patent Office Serial
No. 2386924 issued October 16, 1945
Arroyo, R., (1947). The economics of rum production. The International Sugar
Journal p 292-294
Arroyo, R., (1949). A new rum distillation process Sugar 44 (7) p 34-36
Artisan Distiller, (2011). [Link]
[Link]/phpBB3/[Link]?f=1&t=3814 [29 August 2011]
Ashby, S.F., (1907). The study of fermentations in the manufacture of Jamaica rum.
International Sugar Journal 11 p 243-251
Australian Government, (2006). The alcohol industry – excise technical guidelines.
Badotti, F., Belloch, C., Rosa, C.A., Barrio, E., & Querol, A., (2010). Physiological
and molecular characterisation of Saccharomyces cerevisiae cachaça strains
isolated from different geographic regions in Brazil World Journal of Microbiology
and Biotechnology 26 p 579-587
Badotti, F., Vilaca, S.T., Arias, A., Rosa, C.A., & Barrio, E., (2013). Two
interbreeding populations of Saccharomyces cerevisiae strains coexist in cachaça
fermentations from Brazil. Federation of European Microbiological Societies 14 p
289-301
Baker, P., (1979). Proceedings American Feed Manufacturers Association Ninth
Annual Liquid Feed Symposium. American Feed Manufacturers Association
Arlington VA
Bamforth, C.W., (2007). Flavoured Spirits in Food, Fermentation and Micro-
organisms Wiley-Blackwell, Chichester
Banat, I.M., Nigam, P., Singh, D., Marchant, R.,& McHale, A.P., (1998). Review:
Ethanol production at elevated temperatures and alcohol concentrations: Part 1 –
Yeast in general World Journal of Microbiology and Biotechnology 14 p 809-821
Barbour, E.A., & Priest, F.G., (1988). Some effects of Lactobacillus contamination in
scotch whisky fermentations Journal of the Institute of Brewing 94 (2) p 89-92
Bartowsky, E.J. & Henschke, P.A., (1995). Malolactic fermentation and wine flavour
The Australian Grapegrower and Winemaker Annual Technical Issue p 83-94
 200

Basu, A.K., (1975). Characteristics of distillery wastewater Journal WPCF 48 (8) p

2184-2190
Bernard, D & Scott, V.N., (2007). Hazard analysis and critical control point system:
use in controlling microbiological hazards, In: Food Microbiology: Fundamentals and
Frontiers. (Doyle, M.P., Beuchat, L.R eds) 3rd Edition. Washington DC: ASM Press.
Berry, D.R., (1979). The physiology of whisky fermentation Brewers’ Guardian June
p 25-32
Birch, G.G. & Parker, K.J. & National College of Food Technology (Great Britain).,
(1979). Sugar: science and technology edited by G.G. Birch and K.J. Parker Applied
Science Publishers, London
Bligh, W., & Currey, J., (2003). Account of the rebellion of the New South Wales
Corps: communicated to the Rt Hon. Lord Castlereagh and Sir Joseph Banks,
Bart/William Blight; edited, with an introduction by John Currey Banks Society
Publications Malvern, Victoria.
Bluhm, L., (1983). Distilled beverages, Biotechnology (Rehm, J.J., & Reed, G, Eds)
Vol.5. Vele Chimie Weinheim, p 447-475
Bokulich, N.A., & Mills, D.A., (2012). Next-generation approaches to the microbial
ecology of food fermentations Biochemistry and Molecular Biology reports 45 (7) p
377-389
Bonilla-Salinas, M., Lappe, P., Ulloa, M., Garcia-Garibay, M., & Gomez-Ruiz, L.,
(1995). Isolation and identification of killer yeasts from sugar cane molasses Letters
in Applied Microbiology 21 p 115-116
Bories, A. & Raynal, J., (1988). Anaerobic Digestion of High-Strength Distillery
Wastewater (Cane Molasses Stillage) in a Fixed-Film Reactor, Biological Wastes 23
p 251-267
Bortolussi & O’Neil (2006). Variation in molasses composition from eastern
Australian sugar mills. Australian Journal of Experimental Agriculture 46 p 1455-
1463
Broom, D., (2003). Rum 1st Edition Abbeville Press Inc. U.S.
Browne, C.A., (1929). The spontaneous decomposition of molasses Industrial
Engineering and Chemistry 21 p 600-606
Bruijn, J., & Vanis, M., (1972). The determination of nitrogen free organic acids in
natal cane molasses. Proceedings of the The South African Sugar Technologists’
Association June p 64
Burrows, S. (1970). Baker’s yeast. In: The Yeasts (ed. A.H. Rose and J.S. Harrison)
Volumen 3, Yeast technology, p 349 – 420. Academic Press, London.
Bundaberg in process documentation (2009) (In house: unpublished)
 201

Cachat, E., & Priest, F.G., (2005). Lactobacillus suntoryeus sp. nov., isolated from
malt whisky distilleries International Journal of Systemic and Evolutionary
Microbiology 55 p 31-34
Cacho, E., & Murphy, N.F., (1988). The use of distillers’ dry yeast as inoculum for
rum fermentation on pilot plant scale. Journal of Agriculture of the University of
Puerto Rico. 72 (3) p 13
Cacho, E., Murphy, N.F., & Fontane, E., (1986). Pilot plant fermentation of high test
molasses compared with blackstrap molasses using factorial design. Journal of
Agriculture of the University of Puerto Rico. 72 (1) p 9-17
Campbell, I., (2003). Wild yeasts in brewing and distilling. In Brewing Microbiology p
247-266. Springer US.
Campos, C.R., Silva., C.F., Dias, D.R., Basso, L.C., Amorim, H.V., & Schwan, R.F.,
(2010). Features of Saccharomyces cerevisiae as a culture starter for the production
of the distilled sugar cane beverage, cachaça in Brazil Journal of Applied
Microbiology 108 p 1871-1879
Cardoso , D.R., Andrade – Sobrinho, L. G., Leite-Neto, A.F., Reche, R. V., Isique,
W.D., Ferreira, M.M., Lima-Neto, B.S. & Franco, D.W., (2003). Comparison between
Cachaca and Rum Using Pattern Recognition Methods Journal of Agriculture and
Food Chemistry 52 p 3429-3433
Carr, F.J., Chill, D., and Maida, N., (2002). The Lactic Acid Bacteria: A Literature
Survey Critical Reviews in Microbiology 28 (4) p 281 -370
Clutton, D.W., (1974). Rum The Flavour Industry November/December p 286-288
Cocolin, L., Bisson, L.F., Mills, D.A., (2000). Direct profiling of the yeast dynamics in
wine fermentations FEMS Microbiology Letters 189 p 81-87
Cocolin, L., Pepe, V., Comitini, F., Comi, G., & Ciani, M., (2004). Enological and
genetic traits of Saccharomyces cerevisiae isolated from former and modern
wineries FEMS Yeast Research 5 p237-245
Collicutt, H., (2009a). The worldwide beverage distilling industry, Chapter 3 In: The
Alcohol Textbook (Ed. Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C.) 5th
Edition. Nottingham University Press, Lallemand Ethanol Technology, Nolfolk. UK.
Collicutt, H., (2009b). Whisky: Grain mashing and fermentation. Chapter 29 In: The
Alcohol Textbook (Ed. Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C.) 5th
Edition. Nottingham University Press, Lallemand Ethanol Technology, Nolfolk. UK.
Connolly, C., (1997). Chapter 21 – Bacterial contaminants and their effects on
alcohol production The Alcohol Textbook 3rd Edition Nottingham University Press,
Lallemand Ethanol Technology, Norfolk. UK.
Costerton, J.W., Lewandowski, Z., Caldwell, D.E., Korber, D.R. & Lappin-Scott,
H.M., (1995). Microbial Biofilms Annual Review of Microbiology 49 p 711-745
 202

Couallier, E.M., Payot, T., Bertin, A.B., & Lameloise, M.L., (2006). Recycling of
Distillery Effluents in Alcoholic Fermentation – role in Inhibition of 10 Organic
Molecules Applied Biochemistry and Biotechnology 133 p 217-236
Coulombe, C.A., (2004). Rum: The Epic Story of the Drink that Conquered the World
Citadel Press Books, New York, NY.
Curtin, L.V., (1973). Effect of processing on the nutritional value of molasses In:
Effect of Processing on the Nutritional Value of Feeds. National Academy of
Sciences, Washington DC
Datamonitor (2011) [Link] [accessed 04/11/11]
Dato, M.C.F., Junior, J.M.P., & Mutton, M.J.R., (2005). Analysis of the secondary
compounds produced by Saccharomyces cerevisiae and wild yeast strains during
the production of cachaça. Brazilian Journal of Microbiology 36 p 70-74
de Araújo Vicente, M., Fietto, L. G., de Miranda Castro, I., Dos Santos, A. N. G.,
Coutrim, M. X., & Brandao, R. L., (2006). Isolation of Saccharomyces cerevisiae
strains producing higher levels of flavoring compounds for production of “cachaça”
the Brazilian sugarcane spirit. International journal of food microbiology, 108 (1), p
51-59.
de Souza Liberal, A. T., Basílio, A. C. M., do Monte Resende, A., Brasileiro, B. T. V.,
Silva‐Filho, D., De Morais, J. O. F., & De Morais, M. A. (2007). Identification of
Dekkera bruxellensis as a major contaminant yeast in continuous fuel ethanol
fermentation. Journal of applied microbiology, 102 (2), p 538-547.
de Torres, C.S., de Ayala, Z., & Rivera, M.I., (1987). Aging of high-test molasses
rums. The Journal of Agriculture of the University of Puerto Rico 71 (2) p 137-158
Deerr, N., (1921). Cane sugar London, Norman Rodger
Deibner, L., (1957). Factors regulating the maturation and aging of alcoholic
beverages. American Journal of Enology and Viticulture 8 (3) p 94-104
Destruhaut, C., Fahrasmane, & L., Parfait, A., (1986). Technologie rhumiere. Etude
des rhums de sirop. Rapport de convention I.N.R.A.
Dolan, T.C.S., (1976). Some Aspects of the Impact of Brewing Science on Scotch
Malt Whisky Production Journal of Institute of Brewing 82 p 177-181
Domingues, L., Vincente, A.A., Lima, N., & Teixeira, J.A. (2000). Applications of
yeast flocculation in biotechnological processes. Biotechnology and Bioprocess
Engineering 5 (4) p 288 - 305
Drinks International (2011). Published in Smith, H. The Millionaires Club, ed. Smith,
H., June 2012. West Sussex, UK.
du Plessis, H.W., Dicks, L.M.T., Pretorius, I.S., Lambrechts, M.G., & du Toit, M.,
(2004) Identification of lactic acid bacteria isolated from South African brandy base
wines International Journal of Food Microbiology 91 p 19-29
 203

Duarte, W.F., Amorim, J.C., & Schwan, R.F., (2013). The effects of co-culturing non-
Saccharomyces yeasts with S. cerevisiae on the sugar cane spirit (cachaça)
fermentation process. Antonie van Leeuwenhoek 103 p 175-194
Duarte, W.F., de Sousa, M.V.F., Dias, D.,R., 7 Schwan, R.F. (2011). Effect of Co-
Inoculation of Saccharomyces cerevisiae and Lactobacillus fermentum on the
Quality of the Distilled Sugar Cane Beverage Cachaça. Journal of food science 76
(9) p 1307-1318
El-Tabey Shehata, A.M., (1960). Yeasts Isolated from Sugar Cane and Its Juice
During the Production of Aguardente de Cana Applied Microbiology 8 (2) p 73-75
Ercolini, D., (2004). PCR-DGGE fingerprinting: novel strategies for detection of
microbes in food Journal of Microbiology Methods 56 297 -314
Euromonitor (2011). [Link]/alocoholic-drinks [accessed 04/11/11]
Fahrasmane, L & Ganou-Parfait, B., (1998). Microbial flora of rum fermentation
media Journal of Applied Microbiology 84 p 921-928
Fahrasmane, L. & Ganou-Parfait, B., (1997). De la canne au rhum. Editions Quae
Fahrasmane, L., Ganou-Parfait, B., & Parfait, A., (1988). Research Note: Yeast flora
of Haitian rum distilleries MIRCEN Journal 4 p 239-241
Fahrasmane, L., Parfait, A., Jouret, C., & Galzy, P., (1985). Production of Higher
Alcohols and Short Chain Fatty Acids by Different Yeasts Used in Rum
Fermentations Journal of Food Science 50 p 1427-1430
Faria, J.B., Loyola, E., Lopez, M.G. & Dufour, J.P., (2003). Cachaca, piso and
tequila. In Fermented Beverage Production, second edition, Edited by Lea, A.G.H
and Piggot, J.R, Kluwer Academic New York p 355-363
Fernandez-Espinar, M.T., Lopez, V., Ramon, D., Bartra, E., & Querol, A., (2001).
Study of the authenticity of commercial wine yeast strains by molecular techniques
International Journal of Food Microbiology 70 p 1-10
Fitzgerald, R., & Hearn, M., (1988). Bligh, Macarthur and the Rum Rebellion.
Kangaroo Press
Fleet, G.H., & Green, V., (2010). The microbiology and biotechnology of rum
production Distilled Spirits, Volume 3: New Horizons: Energy, Environment and
Enlightenment 3 p 203
Fleet, G.H., & Heard, G.M., (1993). Yeasts: growth during fermentation Wine
microbiology and biotechnology p 27-54
Fleet, G.H., (1998). Microbiology of alcoholic beverages, in: Wood, B.J. (Ed.) 2nd ed,
Microbiology of Fermented Foods Vol.1 Blackie Academic and Professional, London
p 217-262
Fleet, G.H., (2008). Wine yeasts for the future FEMS Yeast Research 8 (7) p 979-
995
 204

FSANZ (2000). Standard 2.7.5 – Spirits. Date standards made: 20/12/00, Date
current version prepared 11/07/11
Galloway, J.H., (1989). The sugar cane industry. A historical geography from its
origins to 1914. Cambridge University Press, UK.
Ganou-Parfait, B., Fahrasmane, L. & Parfait, A., (1989). Les bacteries aerobies des
milieux fermentaires a base de jus de canne a sucre. Industries Alimentaires et
Agricoles 106 p 579-585
Ganou-Parfait, B., Fahrasmane, L., Parfait, A., (1987). Bacillus spp in sugar cane
fermentation media Belgian Journal of Food Chemistry and Biotechnology 42 (6) p
192 - 194
Ganou-Parfait, B., Valladon, M. & Parfait, A., (1991). Contribution a la bacteriologie
des eaux de fabrication de distilleries en Guadeloupe. In Actes de la premiere
rencontre (p 296 -302). Presented at Rencontres Internationales en langue
Francaise su la Canne a Sucre, Monpellier, FRA
Gomes F.C.O., Silva, C.L.C., Marini, M.M., Oliveira, E.S., Rosa, C.A., (2007). Use of
selected indigenous Saccharomyces cerevisiae strains for the production of the
traditional cachaça in Brazil Journal of Applied Microbiology 103 p 2438-2447
Gomes, F. C., Silva, C. L., Vianna, C. R., Lacerda, I. C., Borelli, B. M., Nunes, Á. C.,
Franco,G>R., Maurao, M.M. & Rosa, C. A. (2010). Identification of lactic acid
bacteria associated with traditional cachaça fermentations. Brazilian Journal of
Microbiology, 41(2), p 486-492.
Gomes, F. C.O., Pataro, C., Guerra, J.B, Neves, M. J., Correa, S.R., Moreira, E.S.A
& Rosa, C.A (2002). Physiological diversity and trehalose accumulation in
Schhizosaccharomces pombe strains isolated from spontaneous fermentations
during the production of the artisanal Brazilian cachaça. Canadian Journal of
Microbiology 48 p 399-406
Gomez, S.M., (1998). Rum Aroma Descriptive Analysis Thesis. Louisiana State
University and Agricultural and Mechanical College
Greig, P., (1895). The Jamaica yeast. Bulletin of Botanical Department Jamaica 2
p157-160
Hall, H.H., James L.H., & Nelson, E.K., (1935). Microorganisms causing fermentation
flavors in cane sirups, especially Barbados “molasses” Journal of Bacteriology 33 (6)
p 577-585
I’Anson, J.A.P., (1971.) Rum Manufacture Process Biochemistry July, p 35-39
Iserentant, D., (1995). Beers: recent technological innovations in brewing. In
Fermented Beverage Production p 45-65. Springer US.
James, G., (2008). Sugar cane- World Agriculture Series 2nd Edition John Wiley &
Sons, NY
 205

James, T.C., Campbell, S., Donnelly, D., & Bond, U., (2003). Transcription profile of
brewery yeast under fermentation conditions Journal of Applied Microbiology 94 p
432-448
Jay, J.M., (1982). Antimicrobial properties of diacetyl. Applied and Environmental
Microbiology 44 p 525 - 532
Jin, X., Yue, S., Wells, K.S., & Singer, V.L., (1994). SYBR Green I: A New
Fluorescent Dye Optimized for Detection of Picogram Amounts of DNA in Gels.
Biophysical Journal 66 A159
Jones, R.P. (1989). Biological principles of the effects of ethanol. Enzymatic
Microbiology Technology 11 p130-153
Jorgensen, A., (1909). Die Mikroorganismen der Garungsindustrie, 5th Edition, P.
Parcy, Berlin.
Jouve, J. L. (2000). Good manufacturing practice, HACCP, and quality systems. The
microbiological safety and quality of food. Gaithersburg (MD): Aspen Publishers,
1627-1655.
Jukes, D (2014). Public Health Measures: Fundamentals of Food Legislation,
Encyclopaedia of Food Safety p 33-38, Elsevier Inc
Kampen, W.H. (1975). Technology of the Rum Industry Sugar y Azucar July p 36-43
Karlin, S., & Altschul, F., (1990). Methods for assessing the statistical significance of
molecular sequencing features by using general scoring schemes. Proceedings of
the National Academy of Science USW 87 p 2264-2268
Kayser, E., (1913). Contribution a l’etude des ferments de la fermentation des
rhums. Annales des Sciences Agricoles p 297 - 322
Keattch, C. & Dollimore, D., (1991) Studies in the history and development of
thermogravimetry. Journal of thermal analysis 37 (9) p 2089-2102
Kelsall, D.R., & Piggot, R., (2009) Grain milling and cooking for alcohol production:
Designing for the options in dry milling Chapter 13 In: The Alcohol Textbook (Ed.
Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C.) 5th Edition. Nottingham
University Press, Lallemand Ethanol Technology, Nolfolk. UK.
Kerr, J., (1983). Southern Sugar Saga, Bundaberg Sugar Company Limited,
Bundaberg, Australia
Kowalchuk, G.A., Stephen, J.R., De Boer, W., Prosser, J.I., Embley, T.M., &
Woldendorp, J.W., (1997). Analysis of ammonia-oxidising bacteria of the beta-
subdivision of the class Proteobacteria in coastal sand dunes by denaturing gradient
gel electrophoresis and sequencing of PCR-amplified 16S ribosomal DNA
fragments. Applied Environmental Microbiology 63 p 1489-1497
Lachance, M., (1995). Yeast communities in a natural tequila fermentation Antonie
van Leeuwenhoek 68 p 151-160
 206

Lea, A.G.H., & Piggott, J.R., (2003). Fermented Beverage Production, Second
edition. Kluwer Academic, New York
Lehtonen, M. & Suomalainen, H., (1977). Rum. In: Economic Microbiology, Edited by
Rose, A.H., Academic Press, London, p 595-633
Lehtonen, M. (1983b). High Performance Liquid Chromatographic Determination of
Nonvolatile Phenolic Compounds in Matured Distilled Alcoholic Beverages Journal of
Assocation of Official Analytical Chemists 66 (1) 71-78
Lehtonen, M., (1983a). Gas-Liquid Chromatographic Determination of Volatile
Phenols in Matured Distilled Alcoholic Beverages Journal of Association of Official
Analytical Chemists 66 (1) p 62-70
Lehtonen, M., Gref, B.K., Puputti, E.V. & Suomalainen, H., (1977) 3-Ethyl-3-
methylbutyric Acid, a new Volatile Fatty Acid Found in Rum Journal of Agriculture
and Food Chemistry 25 (4) 953-955
Lewis, M.J. (1992) A rum state: alcohol and state policy in Australia, 1788-1988.
Canberra, AGPS Press.
Liebich, H.M., Koenig, W.A., & Bayer, E., (1970). Analysis of the Flavor of Rum by
Gas-Liquid Chromatography and Mass Spectrometry Journal of Chromatographic
Science 8 p 527-533
Luxner, L., (2005). Havana Club 44th among world’s 100 top spirits brands. Cuba
News, April. [Link]
Maarse, H. & Noever de Brauw, M.C., (1966). The Analysis of Volatile Components
of Jamaica Rum Journal of Food Science 31 p 951-955
Madrera, R. R., Gomis, D.B., Alonso, J.J.M., (2003) Influence of Distillation System,
Oak Wood Type and Aging Time on Volatile Compounds of Cider Brandy Journal of
Agricultural and Food Chemistry 51 p 5709-5714
Main, A., (2010) Diageo’s still in good spirits despite taxes The Australian July 19,
2010
Makanjuoula, D.B., Tymon, A & Springham, D.G. (1992) Some effects of lactic acid
bacteria on laboratory-scale yeast fermentations. Enzyme and microbial technology.
14 (5) p 350-357
Marini, M.M., Gomes, F.C.O., Silva, C.L.C., Cadete, R.M., Badotti, F., Oliveira, E.S.,
Cardoso, C.R., & Rosa, C.A., (2009). The use of selected starter Saccharomyces
cerevisiae strains to produce traditional and industrial cachaça: a comparative study
World Journal of Microbiology and Biotechnology 25 p 235-242
Mateo, J.J., Jimenez, m.m Pastor, A., & Huerta, T. (2001) Yeast starter cultures
affecting wine fermentation and volatiles. Food Research International, 34 (4). p 307-
314
McNicoll, D.D. (2006) A rum do. Accessed on 20th April 2007 from
[Link]
 207

Melamane, X.L., Strong, P.J., Burgess, J.E., (2007) Treatment of Wine Distillery
Wastewater: A Review with Emphasis on Anaerobic Membrane Reactors., South
African Journal of Enology and Viticulture 28 (1) p 25-36
Morais, P.B., Rosa, C.A., Linardi, V.R., Pataro, C., & Maia, A.B.R.A., (1997). Short
Communication: Characterization and succession of yeast populations associated
with spontaneous fermentations during the production of Brazilian sugar-cane
aguardente. World Journal of Microbiology & Biotechnology 13 p 241-243
Moroz, R. (1963). Methods and procedures for the analysis of microorganisms in
sugar. Principles of sugar technology, ed. P. Honig, p 373-449.
Mosedale, J.R & Puech, J.L., (1998). Wood maturation of distilled beverages, Trends
in Food Science & technology 9 (3) p 95-101.
Mosedale, J.R. (1995). Effects of oak wood on the maturation of alcoholic beverages
with particular reference to whisky Forestry 68 (3) p 203-230
Murtagh, J., (1995a). Molasses as a Feedstock for alcohol production. The Alcohol
Textbook: A reference for the Beverage, Fuel and industrial Alcohol Industries,
Nottingham University Press. Nottingham, RU. p 89-96
Murtagh, J., (1995b). Feedstocks, fermentation and distillation for production of
heavy and light rums. The Alcohol Textbook: A reference for the Beverage, Fuel and
industrial Alcohol Industries, Nottingham University Press. Nottingham, RU. P 243-
255
Muyzer, G., & Smalla, K., (1998). Application of denaturing gradient gel
electrophoresis (DGGE) and temperature gradient gel electrophoresis (TGGE) in
microbial ecology Antonie van Leeuwenhoek 73 p 127-141
Nelson, E.K., & Greenleaf, C.A., (1927). An Investigation of Cane-Molasses Distillery
Slop – with special reference to certain organic acids industrial and Engineering
Chemistry 21 (9) p857-859
Nicol D.A (2003). Chapter 12: Rum, Fermented Beverage Production Edited by
A.G.H Lea and J.R. Piggott 1995
Nicol, D. (1989). Batch distillation in The Science and Technology of Whiskies, (Ed.
J.R. Piggott, R. Sharp and R.E.B. Duncan). Longman, Harlow p 235-263
Nicol, D. (1993). Distillation, brewing Science and Technology, Series ii, Vol.4,
Engineering (inc distillation), institute of Brewing, London, p 303-342
Nykanen, L., & Suomalainen, H., (1983) Aroma of Beer, Wine and Distilled Alcoholic
Beverages D. Reidel Publishing Company, Boston
Nykanen, L., (1986). Formation and Occurrence of Flavor Compounds in Wine and
Distilled Alcoholic Beverages American Journal of Enology and Viticulture 37 (1) p
84-96
Olbrich, H., (1963). The Molasses Book. Fermentation Technologist, Institute fur
Zuckerindustrie, Berlin (Germany).
 208

Oliveira, E.S., Rosa., C.A., Morgano, M.A., & Serra, G.E., (2004). Fermentation
characteristics as criteria for selection of cachaça yeasts World Journal of
Microbiology and Biotechnology 20 p 19-24
Owen, W.L. (1911). A recently Discovered Bacteria Decomposition of Sucrose.
Industrial & Engineering Chemistry, 3 (7) p 481-486
Owen, W.L., (1939). Process of Fermenting Molasses and Like Mashes United
States Patent Office Patent number 2164255 Patented June 27, 1939
Owen, W.L., (1949). The Microbiology of Sugars, Syrups and molasses. Barr-Owen
research Enterprises, Baton Rouge.
Pairault, M.E.A. (1903). Le rhum et sa fabrication. Collection des grandes cultures
colonials, Gauthier-Villars.
Parfait, A., & Sabin, G. (1975). Les fermentations traditionnelles des melasses et des
jus de canne aux Antilles francaises. Industries Alimentaires et Agricoles 92 p 27-34
Pataro, C., Guerra, J.B., Petrillo-Peixoto, M.L., Mendonca-Hagler, L.C., Linardi, V.R.
& Rosa, C.A. (2000). Yeast communities and genetic polymorphism of
Saccharomyces cerevisiae strains associated with artisanal fermentation in Brazil
Journal of Applied Microbiology 89 p 24-31
Pataro, C., Santos, A., Correa, S., Morais, P.B., Linardi, V.R., & Rosa, C.A. (1998).
Physiological characterisation of yeasts isolated from artisanal fermentation in an
Aguardente distillery Revista de Microbiologica 29 p 104-108
Paturau, J. M. (1989).. By-products of the cane sugar industry. An introduction to
their industrial utilization. Elsevier Science Publishers [Link], E. A. (1903). Le
rhum et sa fabrication. Gauthier-Villars.
Pesticide Research Institute for the USDA National Organic Program (2013).
Technical Evaluation Report, Vinasse, Crops August 16, 2013
Piggot, J.R., Paterson, A., Fleming, A.M., & Sheen, M.R., (1991/2). Consumer
perceptions of Dark Rum explored by Free Choice Profiling Food Quality and
Preference 3 135-140
Piggott, J. R., & Conner, J. M. (2003). Whiskies. In Fermented beverage production
(pp. 239-262). Springer US.
Piggott, J. R., Conner, J. M., & Paterson, A. (1995). Flavour development in whisky
maturation. Developments in Food Science, 37, 1731-1751.
Piggott., R., (2009a) Beverage alcohol distillation. In The Alcohol Textbook 5th
Edition (p 431 – 444). Nottingham University Press, UK.
Piggott., R., (2009a) Rum: Fermentation and Distillation. In The Alcohol Textbook 5th
Edition (p 473 - 480). Nottingham University Press, UK.
Pino, J. (2007) Characterisation of rum using solid-phase microextraction with gas
chromatography-mass spectrometry. Food Chemistry 104 p 421 - 428
 209

Pino, J., Martin, M.P., Mestres, M., Perez, J., Busto, O., & Guasch, J., (2002)
Headspace solid-phase microextraction of higher fatty acid ethyl esters in white rum
aroma Journal of Chromatography A 954 p 51-57
Pino, J.A., Tolle, S., Gok, G.R., & Winterhalter, P., (2012) Characterisation of odour-
active compounds in aged rum Food Chemistry 132 p 1436-1441
Pisolor, E., Pontalier, P.Y., & Albet, J. (2009) Process for Recovering Carboxylic
Acids from Sugar Cane Industry By-Products. Available at: [Link]
[Link]/2928/1/Albet_2928.pdf
Pretorius, I.S., du Toit, M., & van Rensburg, P., (2003). Designer Yeasts for the
Fermentation Industry of the 21st Century. Food Technology and Biotechnology 41
(1) p 3-10
Priest, F.G (2004). Lactic acid bacteria – the uninvited but generally welcome
participants in malt whisky fermentation Microbiology Today 31 p 16-18
Priest, F.G. & Pleasants, J.G. (1988). Numerical taxonomy of some leuconostocs
and related bacteria isolated from Scotch whisky distilleries. Journal of Applied
Bacteriology 64 p 379 - 387
Queensland Sugar Limited (2010) Annual report, 2009/2010 ([Link]/news-
media/publications)
Quesada Granados, J., Merelow Guervos, J.J., Oliveras Lopez, M., J., Gonzalez
Penalver, J., Olalla Herrera, M., Blanca Herrera, R., & Lopez Martinez, M.C. (2002).
Application of artificial aging techniques to samples of rum and comparison with
traditionally aged rums by analysis with artificial neural nets. Journal of agricultural
and food chemistry, 50 (6) p 1470 - 1477
Rameshwari, K.S. & Karthikeyan, S., (2005). Distillery Yeast Sludge (DYS) as an
Alternative Feed Resource in Poultry International Journal of Poultry Science 4 (10)
p 787-789
Rappe, M.S. & Giovannoni, S.J., (2003). The Uncultured Microbial majority Annual
review of Microbiology 57 p 369=394
Reche, R. V., Leite Neto, A. F., Da Silva, A. A., Galinaro, C. A., De Osti, R. Z., &
Franco, D. W. (2007). Influence of type of distillation apparatus on chemical profiles
of Brazilian cachaças. Journal of agricultural and food chemistry, 55(16), p 6603-
6608.
Rocques, X., (1927). Enquete sur les caracteres specifiques des rhums des colonies
francaises. Annales des Falsifications 15, p71-79, 148-155, 224-232, p 399-403.
Rosa, C. A., Soares, A. M., & Faria, J. B. (2009). Cachaça production. The alcohol
textbook. Nottingham: Nottingham University, p 484-497.
Rosa, C.A., Soares, A.M., & Faria, J.B., (2009). Cachaça production In:The alcohol
textbook. (Ed. Ingledew, W.M., Kelsall, D.R., Austin, G.D., Kluhspies, C.) 5th Edition.
Nottingham University Press, Lallemand Ethanol Technology, Nolfolk. UK.
 210

Russell, C., Bhandari, R.R., Walker, T.K. (1954). Vitamin requirements of thirty-four
Lactic Acid Bacteria associated with Brewery Products Journal of general
Microbiology 10 p 371-376
Russell, I., & Stewart, G. (Eds.). (2014). Whisky: technology, production and
marketing. Elsevier
Sampaio, O.M., Reche, R.V., Franco, D.W. (2008). Chemical Profile of Rums as a
Function of their Origin. The Use of Chemometric Techniques for their Identification
Journal of Agricultural and Food Chemistry 56 p 1661-1668
Schoonees, B. M., & Pillay, V. (2004). A New Filtration System for Gum Analysis in
Raw Sugar and Molasses. In Proc S Afr Sug Technol Ass p 78
Schwan, R.F., Mendonca, A.T., da Silva Jr, J.J., Rodrigues, V., & Wheals, A.E.,
(2001). Microbiology and physiology of Cachaça (Aguardente) fermentations Antonie
van Leewenhoek 79 p 89-96
Sharp, R., & Watson, D., (1979). Improving yeast strains for whisky production.
Brew. Dist. Int, 9(6), p 50-53.
Simmonds, C. (1919). Alcohol, Its Production, Properties, Chemistry, and Industrial
Applications: With Chapters on Methyl Alcohol, Fusel Oil, and Spirituous Beverages.
Macmillan.
Simpson, A.C., (1973). Quality Control in the Distilling Industry Process Biochemistry
February p 18-19
Simpson, K.L., Pettersson, B., & Priest, F.G., (2001). Characterization of lactobacilli
from Scotch malt whisky distilleries and description of Lactobacillus ferintoshensis
sp. nov., a new species isolated from malt whisky fermentations. Microbiology 147 p
1007-1016
Singleton, V.L., (1995). Maturation of Wines and Spirits: Comparisons, Facts and
Hypotheses American Journal of Enology and Viticulture 46 (1) p 98-115
Smibert, R. M., & N. R. Krieg. (1994). Phenotypic characterization, p. 615. In P.
Gerhardt, R. Murray, W. Wood, and N. Krieg (ed.), Methods for general and
molecular bacteriology. ASM Press, Washington, DC.
Souza, P.A., Scanavini, H.F.A., Meirelles, A.J.A, & Alcarde, A.R., (2010). Cachaca
production through double distillation technique. Worldwide Distilled Spirits
Conference, New Horizons: Energy, Environment and Enlightenment. Ed. G.M.
Walker and P.S. Hughes.
Stamer, J.R (1979). The Lactic Acid Bacteria: Microbes of Diversity Food
Technology January 60-65
Stanbury, P.F. & Whitaker, A (1984). Principles of fermentation technology .
Pergamon Press, Oxford
Stewart, G. G., Hill, A. E., & Russel, I (2013). 125th Anniversary Review:
Developments in brewing and distilling yeast strains Journal of Institute of Brewing
119 p 202-220
 211

Suomalainen, H., (1971) Yeast and its Effect on the Flavour of Alcoholic Beverages.
Journal of the Institute of Brewing 77 (2) p 164-177
Suomalainen, H., Nykanen, L., & Eriksson, K., (1974) Composition and Consumption
of Alcoholic Beverages – A Review American Journal of Enology and Viticulture 25 p
179-187
Sutherland, I.W., (1998). Novel and established applications of microbial
polysaccharides. Tibtech. 16 p41-46
Swiegers, J.H., Bartowsky, E.J., Henschke, P.A., Pretorius, I.S. (2005). Yeast and
bacteria modulation of wine aroma and flavour. Australian Journal of grape and wine
research 11 (2) p139-173
Sword, K., (2010). Available from: <[Link]
.asp?article_id=rosei> [11 May 2010]
Taskila, S., Tuomola, M., Kronlof, J., & Neubauer, P., (2010). Comparison of
Enrichment Media for Routine Detection of Beer Spoiling Lactic Acid Bacteria and
Development of Trouble-shooting Medium for Lactobacillus backi. Journal of the
Institute of Brewing 116 (2) p 151-156
The Council of European Communities (EEC), (1989). Appellations of Origin (Spirit
Drinks) Council Regulations, 29/05/1989, No 1576/89
Thompson, I.L., (2009). Studies on the Maturation of Jamaican Rum. Thesis. The
University of the West Indies, St. Augustine, Trinidad and Tobago.
Tilbury, R.H., (1980). Xerotolerant (osmophilic) yeast. Biology and actifices of yeasts
p 153-179 Academic Press, London
Timmer, R., Heide, R.T., Wobben, H.J., & Valois, P.D., (1971) Phenolic compounds
in rum Journal of Food Science 36 (3) p462-463
Todorov, S.D., & Dicks, L.M.T., (2005). Lactobacillus plantarum isolated from
molasses produces bacteriocins active against Gram-negative bacteria Enzyme and
Microbial Technology 36 p 318-326
Tokuoka, K., (1993). A Review: Sugar- and salt-tolerant yeasts Journal of Applied
Bacteriology 74 p 101-110
Torsvik, V., Daae, F.L., Sandaa, R., & Øvreas, L., (1998). Review Article: Novel
techniques for analysing microbial diversity in natural and perturbed environments.
Journal of Biotechnology 64 p53-62
Trachoo, N., (2003). Biofilms and the food industry. Biofilms 25 (6) p 808
Tyson, G.W., & Banfield, J.F., (2005). Cultivating the uncultivated: a community
genomics perspective. TRENDS in Microbiology 13 (9) p411-415
USA Government (2010). 27 CFR 5.22
 212

van Beek, S & Priest, F.G (2003). Bacterial Diversity in Scotch Whisky
Fermentations as Revealed by Denaturing Gradient Gel Electrophoresis American
Society of Brewing Chemists 61 (1) 10-14
van Beek, S. & Priest, F.G (2002).. Evolution of the Lactic Acid Bacterial Community
during Malt Whisky Fermentation: a Polyphasic Study Applied and Environmental
Microbiology 68 (1) 297-305
van Beek, S., & Priest, F.G., (2000). Decarboxylation of Substituted Cinnamic Acids
by Lactic Acid Bacteira Isolated during Malt Whisky Fermentation Applied and
Environmental Microbiology 66 (12) p 5233-5328
Verstrepen, K.J., Derdelinckz, G., Verachtert, H., & Delvaux, F.R. (2003). Yeast
flocculation: what brewers should know Applied Microbiology and Biotechnology 61
p 197-205
Walford, S.N. (1996). Composition of Cane Juice, Proceedings of the South African
Sugar Technologists’ Association 70 p 265
Walker, G., Brosnan, J., Bringhurst, T., & Jack, F. (2012). Chapter 16 – Selecting
new distilling yeasts for improved fermentation and for sustainability Distilled Spirits:
Science and Sustainability p 1-11
Walker, G.M. (2004). Metals in Yeast Fermentation Processes. In. Laskin, A.L.,
Bennett, J.W. & Gadd, G.M. (eds) ‘Advances in Applied Microbiology’ Volum 54.
Elsevier Academic Press, London p 197-227
Wankat, P (2007). Separation Process Engineering: Including Mass transfer analysis
(3rd Edition) Prentice Hall,
Watson, D.C., (1993). Spirits Ullmann’s Encyclopedia of Industrial Chemistry
Wilkie, A. C., Riedesel, K.J., & Owens, J.M. (2000). Stillage characterisation and
anaerobic treatment of ethanol stillage from conventional and cellulosic feedstocks
Biomass and Bioenergy 19 p 63-102
Wobben, H.J., Timmer, R., ter Heide, R. & de Valois, P.J. (1971). Nitrogen
Compounds in Rum and Whiskey Journal of Food Science 36 464-465
Wustenfeld, H. & Haeseler, G., (1953). Trinkbranntweine und Likore: ihre
Herstellung, Untersuchung und Beschaffenheit Edited by Haeseler, G., Paul Parey
Wythes, J.R., Wainwright, D.H., & Blight, G.W. (1978). Nutrient composition of
Queensland molasses. Australian Journal of Experimental Agriculture and Animal
Husbandry 18 p 629-634
Yokota, M., & Fagerson, I.S. (1971). The Major Volatile Components of Cane
Molasses Journal of Food Science 36 p 1091-1093
Zarattini, R. de A., Williams, J.W., Ernandes, J.R., & Stewart, G. G., (1993).
Bacterial-induced flocculation in selected brewing strains of Saccharomyces. Cereal
Biotechnology 18 p 65-70
 213

Zwietering, M.H., Jongenburger, I., Rombouts, F.M., & Riet, K.V. (1990). Modeling of
the Bacterial Growth Curve, Applied and Environmental Microbiology 56 (6) p 1875
 211

CHAPTER 9 APPENDICIES
 212

APPENDIX A
Chemical components found in rum

The following table was collated from all available references discussing chemical
composition of rum and shows the complexity of rum composition. Concentrations
have been given where possible, while some components were only detected
qualitatively and not quantitatively. Components with a concentration range were
generally detected in more than one study, however results varied; possibly due to
different styles of rums tested.

The table has been organised in to chemical groups/families and ordered within each
group greatest concentration to least. It is important to note, however, that some
components have a greater sensory impact than others regardless of concentration.

Component Concentration Component Concentration

ppm (mg/l) ppm (mg/l)
Esters
Ethyl n-butyrate a,b,j,k 220 Isobutyl acetate a,j,k 1.5
Ethyl acetate a,b,j,k 73-200 Ethyl palmitoleate a 1.5
a,i,j,k a
Ethyl n-decanoate 25-130(3.79-103.3) Ethyl heptenoate 1.2
a,b,j a,b,c,i,j,k
Ethyl propionate 50 Ethyl n-hexanoate 1-40(nd–0.15)
Ethyl linolate a 50 Isoamyl n-hexadecanoate a 1
a,b,j,k a,k
Ethyl n-valerate 40 β-phenylethyl acetate 1
Ethyl n-octanoate a,b,i,j,k 15-50(0.65-14.14) n-hexyl acetate a,b 1
a a
Isoamyl n-ocatnoate 15 Ethyl n-pentadecanoate 0.8
Ethyl n-dodecanoate a,i,j 12-15(0.03-1.59) Diethyl succinate a,b,k 0.8
a,b,k a,b,k
Ethyl lactate 10 Ethyl benzoate 0.5
Ethyl formate a,b,j, k 10 n-propyl acetate a,b,j 0.5
a,b,k a
Ethyl isobutyrate 8 Ethyl sterate 0.5
a a
Ethyl oleate 8 Isoamyl n-valerate 0.5
Ethyl n-hexadecanoate a,j 5-50 Isobutyl propionate a,j 0.5
a,j a
Isoamyl acetate 5-10 Ethyl (β-phenylpropionate) 0.5
Ethyl n-tetradecanoate a,j 5-6 Methyl n-hexandecanoate a 0.3
a,b a,b
Ethyl n-heptanoate 2.5 Ethyl n-nonanoate 0.3
Ethyl isovalerate a,k 2.5 Ethyl (3-methoxy-4-hydroxybenzoate)a,k 0.3
a a,d
Ethyl (2-methylbutyrate) 2.5 Methyl n-decanoate 0.2
Monoethyl succinate a 2 n-propyl propionate a 0.2
a a
Isoamyl formate 2 Isoamyl n-tetradecanoate 0.2
Isoamyl n-butyrate a,j 1.5 Isoamyl n-decanoate a,j 0.2
 213

Component Concentration Component Concentration

ppm (mg/l) ppm (mg/l)
Esters (cont.) Alcohols (cont.)
β-phenylethyl n-decanoate a 0.2 2-hexanol a 0.02
a,j,k d
Methyl salicylate 0.15-25 Methyl butenol -
a d,j
Isobutyl n-decanoate 0.15 2-methyl-1-propanol -
a k
An ethyl dodecanoate 0.15 4-methylpentan-1-ol -
a k
An ethyl dodecanoate 0.13 3-methyl-2-buten-1-ol -
Isoamyl n-dodecanoate a,j 0.13 2-heptanol k -
a k
Ethyl n-undecanoate 0.13 (Z)-4-hepten-2-ol -
a k
Isoamyl propionate 0.1 (z)-3-hexen-1-ol -
a k
Methyl n-octanoate 0.1 2-octanol -
a k
β-phenylethyl n-octanoate 0.1 1-heptanol
Ethyl (2-furancarboxylate) a 0.1 1-octen-3-ol k
Methyl n-dodecanoate a 0.1 1-octanol k
n-propyl n-hexadecanoate a 0.05 2-nonanol k
Ethyl n-heptadecanoate a 0.05 (z)-2-octen-1-ol k
n-propyl n-butyrate a,j 0.03 1-decanol k
a,b,j
n-butyl acetate 0.01 Phenols
k
ethyl 2-ethoxyacetate - Isoeugenol a 0.08
b a,k
Ethyl (3-methylbutyrate) - 4-methyl-2-methoxyphenol 0.05
b,k a,k
3-methylbutyl acetate - 4-ethyl-2-methoxyphenol 0.02
Ethyl-3-hydroxybutanoate b,k - 4-ethylphenol a,e,f,k nd-1.96
b e,f
Isobutyl valerate - 2 ethyl-2-methoxy-phenol nd-1.83
b a,b,e,f,j,k
3-methylbutyl butanoate - Eugenol nd-1.36
d e,k
Ethyl nonadecanoate - 2-methoxy-4-propylphenol nd-0.84
k a,e,f,k
Diethyl carbonate - 2-methoxyphenol nd-0.83
Ethyl 3-ethoxypropanoate k - Phenol a,e,f,k tr-0.26
e,f
Alcohols 4 methylphenol nd-0.2
a,b,k e,f
Ethanol (see below) 2 methylphenol nd-0.1
a,b,d,j,k e
3-methyl-1-butanol 860-1000 3 methylphenol nd-0.05
a,b,d,j b f
2-methyl-1-butanol 200-210 2-methoxy-4-methylphenol , -
Isobutanol a,b,k 100 3-methyl-3-buten-1-olk -
a,b,j,k
n-butanol 10 Lactones
a,b,d,j,k
n-propanol 7.5-420 δ-octalactone a 0.1
a,b,k a,k
2-butanol 6 ɣ-nonalactone 0.025
1-hexanol a,j,k
1.3-2.5 ɣ-decalactone a 0.025
1-pentanol a,b,j 1-2.5 δ-decalactone a 0.025
a,j,k a
β-phenylethanol 0.5-<1.0 ɣ-dodecalactone 0.025
a a
Menthol 0.5 δ-dodecalactone 0.025
a,k
2-pentanol 0.5 Carbonyl compounds
a
2-methyl-2-butanol 0.2 Furfural a,b,k 25
a,k
3-pentanol 0.13 4-ethoxy-2-pentanone a,k 7.5
j,k a,k
Phenylmethanol <0.1 3-penten-2-one 7
a a
Methanol 0.08 4-ethoxy-2-butanone 5
 214

Component Concentration Component Concentration

ppm (mg/l) ppm (mg/l)
Carbonyl compounds (cont.) Acetals (cont.)
2-methyl-3-tetrahydrofuranone a,k 5 1-ethoxy-1-(2-methylpropoxy)-ethane a 1.5
a,b,k a
Benzaldehyde 2 1,1-di-(2-methylpropoxy)-propane 1.3
a,b,k a
Isovaleraldehyde 1.8 1,1-diethoxypropane 1.2
2-methylbutyraldehyde a 1.5 1-ethoxy-1-propoxyethane a 1.2
a,b,k a
5-methylfurfural 1.2 1-ethoxy-1-propoxy-2-methylpropane 1
2-pentanone a,k 1.2 1-(2-methylpropoxy)-1-(3- 0.8
a,b a
Acetylfuran 1.2 methylbutoxy)-2-methylpropane
a,b,j
Acetaldehyde 0.5-150 1,1-di-(2-methylpropoxy)-3- 0.8
Isobutyraldehyde a
0.25 methylbutanea
Vanillin a,k 0.25 Diethoxymethane a 0.8
a,k
Acetone 0.25 1,1-di-(2-methylpropoxy)-2- 0.5
a,k a
2-butanone 0.03 methylpropane
a,k
Diacetyl 0.03 1-ethoxy-1-(3-methylbutoxy)-propane a 0.5
a
o-hydroxyacetonphenone 0.01 1-ethoxy-1-butoxyethane a 0.5
a,b a
Propionaldehyde 0.01 1-ethoxy-1-pentoxyethane 0.5
but-2-enal b - 1-butoxy-1-(3-methylbutoxy)-ethane a 0.5
b a
Hexanal - 1-methoxy-1-ethoxyethane 0.3
1,1,3 triethoxypropane k - 1-ethoxy-1-(2-methylpropoxy)-propane a 0.25
b a
Prop-2-enal - 1,1-dipropoxyethane 0.25
b a
2-ethoxypropanal - 1-ethoxy-1-(2-methylpropoxy)-pentane 0.25
Acetals 1-propoxy-1-(3-methylpropoxy)-3- 0.25
a,b,j,k a
1,1-diethoxyethane 50-320 methylbutane
a, k
1,1-diethoxy-3-methylbutane 13 1-(2-methylpropxy)-1-(3-methylbutoxy)- 0.25
1-ethoxy-1-(3-methylbutoxy)-ethane a 10 3-methylbutanse a
1,1-di-(3-methylbutoxy)-ethane a 7.5 1,1-di-(3-methylbutoxy)-2- 0.12
a a
1,1-diethyoxy-2-methylpropane 6 methylpropane
1-(2-methylpropoxy)-1-(3-methylbutoxy)- 5 1-ethoxy-1-(2-methylpropoxy)-2- 0.12
a a
ethane methylpropane
1-ethoxy-1-(3-methylbutoxy)-2- 5 1-propoxy-1-(3-methylbutoxy)2- 0.12
a a
methylpropane methylpropane
1-ethoxy-1-(2-methylpropoxy)-3- 5 1-propoxy-1-(2-methylpropoxy)-2- 0.12
methylbutane a methylpropane a
1-ethoxy-1-(3-methylbutoxy)-3- 4 1-propoxy-1-(3-methylbutoxy)-propane a 0.12
a a
methylbutane 1,1-dipropoxy-3-methylbutane 0.12
a a
1,1-diethoxy-2-methylbutane 2.5 1,1-di-(3-methylbutoxy)-3-methylbutane 0.12
a
1-ethoxy-1-(2-methylbutoxy)-ethane 2.5 1-propoxy-1-(2-methylpropoxy)-3- 0.08
a methylbutane
1-propoxy-1-(3-methylbutoxy)-ethane 2.5
1-(3-methylbutoxy)-1-(2-methylbutoxy)- 2.3 1,1-di-(2-methylpropoxy)-pentane a 0.05
a
ethane 1-ethoxypropane k
-
a b
1,1-diethoxy-2-propanone 1.5 1,1-ethoxy-hexoxyethane -
a k
1,1-di-(2-methylpropoxy)-ethane 1.5 1-1 diethoxymethane -
 215

Component Concentration Component Concentration

ppm (mg/l) ppm (mg/l)
Acids Hydrocarbons
a,b,d,j,k
Acetic acid 10.3-35 3,8,8-trimethyl-tetrahydronaphthalene a 5
a,b,c,j,k
n-octanoic acid 4.3-7.5 (8.9-24.1) 3,8,8-trimethyl-tetrahydronaphthalene a 1.2
c,j,k a
Dodecanoic 4.0(6.5-12.0) 3,8,8-trimethyl-dihyrdronaphthalene 0.25
hexadecanoic acid c,j 1.4(0.5-4.0) n-pentane a 0.05
c,j a
Tetradecanoic acid 0.7(0.4-1.7) 2-methylpentane 0.01
Isovaleric acid a,j,k 0.7 (3.0-6.5) n-hexane a 0.05
a,c,j,k a
n-decanoic acid 0.6-13.4 n-heptane 0.05
2-methylbutyric acid a,k 0.6 2-methylhexane a 0.025
a, a
Heptenoic acid 0.5 2,3-dimethylpentane 0.025
a a
2-furancarboxylic acid 0.4 Methylcyclohexane 0.003
Benzoic acid a 0.4 Nitrogen compounds
n-hexanoic acid a,b,c,j,k 0.3-15 (4.5-6.6) Pyridine g -
a,b,j,k
n-butyric acid 0.3-7.5 (8.0-15.3) 2-methylpyridine g -
a g
3-furancarboxylic acid 0.3 1,3-thiazol -
4-methyl pentanoic j 0.3 Methylpyrazine g -
a,c,j g
n-heptanoic acid 0.28-2.5 (tr-2.4) 3-methylpyridine -
Propionic acid a,b,c,k 0.2-1.5(7.4-30.2) 2,5-dimethylpyrazine g -
a,b,j,k g
Isobutyric acid 0.2-0.9 (2.9-4.3) 2,6-dimethylpyrazine -
a g
β-ethoxypropionic acid 0.15 2,3-dimethylpyrazine -
Isohexanoic acid a 0.09 2-methyl-6-ethylpyrazine g -
c,j g
Octadecanoic acid 0.08(0.2-0.5) 2-methyl-5-ethylpyrazine -
n-nonanoic acid a,c 0.06 (tr-0.5) Trimethylpyrazine g -
a,b,j g
n-valeric acid 0.04-7.5 (1.5-6.5) 2,5-dimethyl-3-ethylpyrazine -
Isooctanoic acid a 0.03 Miscellaneous
Oleic acid c (0.2-1.2) α-ionone a* 0.3
c
Hexadec-9-enoic acid (0.2-1.0) Unsaturated decalactone a* 0.2
c a,j
Linoleic acid (nd-2.2) Dimethyl sulphide 0.12
Undecanoic acid c (nd-0.7) Methylethyl sulphide a 0.12
c a*
Pentadecanoic acid (nd-0.1) sec. butyl acetate 0.1
c a*
Tridecanoic acid (nd-0.1) isobutyl n-butyrate 0.1
2-ethyl-3-methylbutyric acid h - methylbutenol a* 0.1
k a
Butanoic acid - Diethyl ether 0.05
Alkylpyrazines 2-methylfuran a 0.05
a a*
2,6-dimethylpyrazine 1.5 Methyl acetate 0.025
2,5-dimethylpyrazine a 0.8 4-methyl-3-pentene-2-one a* 0.025
a a*
2-methylpyrazine 0.5 2-hexanone 0.025
2-methyl-3-ethylpyrazine a 0.25 A hydroxymethylbenzaldehyde a* 0.01
a a*
2,5-dimethyl-3-ethylpyrazine 0.2 3-hexanone 0.005
a a*
3,5-dimethyl-2-ethylpyrazine 0.1 Ethyl dodecadienoate 0.03
2-methyl-6-ethylpyrazine a 0.1 Isobutyl formate a* 0.025
Note: Ethanol concentration depended on sampling point etc. Pino data included samples containing 80% ethanol
v/v.
a * b c
Liebich et al (1970) tentatively identified in this work, Maarse and ten Noever de Brauw (1966), Nykanen et al
d e f g h
(1968), Gracia et al (2007), Lehtonen (1983), Timmer et al (1971), Wobben et al (1971), Lehtonen et al
i j k
(1977), Pino et al (2002), Allan (1972), Pino et al (2012) (-) denotes identification made however no
quantification performed.
 216

APPENDIX B
Individual culture fermentations
S. cerevisiae
Time Log Ethanol Sugars (mg/g)
pH °Brix
Station CFU/mL (%) Fructose Glucose Sucrose
0 5.5 5.3 31.3 0.0 54.0 63.5 123.5
8 5.4* 5.3 31.4 0.3 51.8 48.4 135.0
16 6.0 5.3 31.4 0.4 46.7 47.1 136.7
24 6.4 5.3 30.0 2.3 42.3 66.9* 108.4
1

32 6.6 5.3 28.6 3.6 37.7 43.9 74.3

40 6.9 5.2 28.8 3.6 37.2 39.7 61.0
48 6.8 5.2 27.2 5.5 28.1 34.9 55.9
0 6.1 5.3 31.2 0.0 53.8 61.1 125.9
8 5.2* 5.3 31.1 0.3 37.0 66.1 121.2
16 6.0 5.2 31.0 0.7 28.6 55.7 158.3*
24 6.3 5.2 29.3 2.8 30.8 55.3 115.0
2

32 6.5 5.2 27.9 4.1 26.1 48.1 58.2

40 7.1 5.2 27.3 4.4 25.6 44.7 52.9
48 6.8 5.2 26.3 4.6 25.5 41.5 51.5
0 5.8 5.3 31.3 0.0 53.9 62.3 126.6
8 5.3* 5.3 31.3 0.3 44.4 57.3 128.1
Average

16 6.0 5.3 31.2 0.6 37.6 51.4 136.7

24 6.4 5.3 29.7 2.6 36.6 55.3 111.7
32 6.6 5.3 28.3 3.9 31.9 46.0 66.3
40 7.0 5.2 28.1 4.0 31.4 42.2 57.0
48 6.8 5.2 26.8 5.1 26.8 38.2 53.7
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.

L. fermentum
Time Log Ethanol Sugars (mg/g)
pH °Brix
Station CFU/mL (%) Fructose Glucose Sucrose
0 6.9 5.3 30.4 0.0 47.5 74.2 141.6
8 6.9 5.2 30.2 0.2 47.9 73.9 137.2
16 7.6 5.2 29.7 0.3 45.7 73.2 138.6
24 8.1 4.9 29.5 0.3 42.0 73.6 145.1*
1

32 8.4 4.5 29.8 0.3 48.8 72.6 131.5

40 8.3 4.3 29.5 0.3 41.5 86.8* 122.5
48 8.2 4.2 29.4 0.3 42.3 77.7 122.5
0 7.0 5.3 30.3 0.0 56.1 79.2 115.6
8 7.2 5.3 30.1 0.2 59.1 74.8 109.2
16 7.4 5.1 29.6 0.2 56.5 73.8 103.7
24 8.4 4.9 29.3 0.3 50.1 72.4 136.5
2

32 8.2 4.6 29.6 0.3 45.4 83.8* 126.4

40 8.4 4.3 30.0 0.3 44.3 87.0* 115.8
48 8.3 4.1 28.9 0.3 43.4 77.9 104.4
0 7.0 5.3 30.4 0.0 51.8 76.7 128.6
8 7.1 5.3 30.2 0.2 53.5* 74.4 123.2
Average

16 7.5 5.2 29.7 0.3 51.1 73.5 121.2

24 8.3 4.9 29.4 0.3 46.1 73.0 136.5*
32 8.3 4.6 29.7* 0.3 47.1 72.6 129.0
40 8.4 4.3 29.8* 0.3 42.9 86.9* 119.2
48 8.3 4.2 29.2 0.3 42.9 77.8 113.5
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.
 217

L. plantarum
Time Log Ethanol Sugars (mg/g)
pH °Brix
Station CFU/mL (%) Fructose Glucose Sucrose
0 6.0 5.1 30.2 0.0 52.8 76.6 122.6
8 6.0 5.1 30.1 0.2 72.0* 45.5 116.5
16 6.9 5.0 30.0 0.1 47.5 47.6 119.1
24 7.7 4.8 30.2 0.2 40.1 52.1 123.4
1

32 8.4 4.6 30.1 0.2 40.1 48.2 110.9

40 8.2 4.4 29.9 0.2 37.1 48.3 131.0*
48 8.6 4.4 29.8 0.2 36.6 45.3 140.8*
0 6.6 5.1 31.2 0.0 58.8 82.8 132.8
8 6.5 5.1 31.1 0.1 94.6* 49.1 129.7
16 8.1 5.0 31.0 0.2 60.7 47.3 101.9*
24 8.0 4.8 30.6 0.1 43.3 47.0 114.0
2

32 8.7 4.6 30.7 0.2 41.4 48.0 103.1

40 8.7 4.4 30.3 0.2 41.1 43.9 111.5*
48 8.6 4.4 30.0 0.2 42.2 40.9 128.4*
0 6.3 5.1 30.7 0.0 55.8 79.7 127.7
8 6.2 5.1 30.6 0.2 83.3* 47.3 123.1
Average

16 7.5 5.0 30.5 0.1 54.1 47.5 119.1

24 7.8 4.8 30.4 0.2 41.7 49.6 118.7
32 8.5 4.6 30.4 0.1 40.8 48.1 107.0
40 8.4 4.4 30.1 0.2 39.1 46.1 121.3*
48 8.6 4.4 29.9 0.2 39.4 43.1 134.6*
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.

Lactobacillus spp.
Time Log pH °Brix Ethanol Sugars (mg/g)
Station CFU/mL (%) Fructose Glucose Sucrose
0 6.5 5.3 31.1 0.0 53.7 82.1 124.5
8 7.2 5.3 30.9 0.2 65.1* 85.4 110.8
16 7.5 5.1 31.4 0.2 56.6 85.5 120.5
24 7.6 4.7 30.7 0.3 34.1* 91.4* 123.6
1

32 8.3 4.4 30.2 0.3 51.0 88.0 118.3

40 8.2 4.3 30.6 0.3 45.2 83.1 118.6
48 7.9 4.2 30.2 0.2 42.6* 88.6* 127.9
0 7.3 5.3 31.6 0.1 55.5 81.4 130.3
8 7.2 5.3 31.1 0.2 55.8 84.7 109.3
16 7.8 5.2 31.0 0.2 55.5 94.1* 136.6
24 7.9 4.6 29.3 0.2 49.1 90.7* 131.5
2

32 8.1 4.4 27.9 0.2 50.6 83.1 132.9

40 8.2 4.3 27.3 0.2 41.8 83.0 125.2
48 8.9 4.1 26.3 0.2 36.5 78.2 135.5*
0 6.9 5.3 31.3 0.1 54.6 81.8 127.4
8 7.2 5.3 31.0 0.2 55.8 85.1 110.1*
Average

16 7.7 5.2 31.2 0.2 56.1 85.5 128.6

24 7.7 4.7 30.0 0.3 49.1 91.1* 127.6
32 8.2 4.4 29.1 0.3 50.8 85.6 125.6
40 8.2 4.3 29.0 0.3 43.5 83.1 121.9
48 8.4 4.2 28.3 0.2 39.6 78.2 127.9
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.
 218

Mixed culture fermentations

S. cerevisiae & L. fermentum
Time Bacteria Yeast pH °Brix Ethanol Sugars (mg/g)
Station Log Log (%)
Fructose Glucose Sucrose
CFU/mL CFU/mL
0 7.3 5.0 5.2 30.4 0.1 54.4 62.9 137.4
8 7.4 5.3 5.1 29.1 0.7 26.3 59.3 102.3*
16 7.6 6.0 5.0 28.5 0.8 20.4 57.3 170.6*
1

24 7.7 6.8 4.9 26.7 2.0* 22.8* 48.5 166.7*

36 8.0 7.1 4.8 24.6 5.6 15.9 21.5 113.7
48 8.0 7.2 4.6 23.2 7.4 5.1 18.8 110.7
0 7.4 7.4* 5.2 30.6 0.1 51.6 79.4 134.7
8 7.1 5.8 5.1 30.1 0.6 41.9 45.0 139.5
16 7.5 6.0 5.0 29.6 0.8 62.3* 79.7* 140.3
2

24 7.4 6.6 4.9 27.8 0.3* 21.9 54.3 129.9

36 7.7 7.1 4.8 25.4 5.3 11.6 35.4 139.7*
48 7.9 7.2 4.7 23.6 8.4 7.9 27.0 108.1
0 7.3 5.0 5.2 30.5 0.1 53.0 71.1 136.1
8 7.3 5.6 5.1 29.6 0.7 34.2 52.2* 139.5
Averages

16 7.6 6.0 5.0 29.1 0.8 20.4 57.3 140.3

24 7.6 6.7 4.9 27.3 2.0 21.9 51.5 129.9
36 7.9 7.1 4.8 25.0 5.5 13.8 28.5 113.7
48 8.0 7.2 4.7 23.4 7.9 6.6 24.5 109.4
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.

S. cerevisiae & Lactobacillus spp.

Bacteria Yeast Sugars (mg/g)
Time Ethanol
Log Log pH °Brix
Station (%) Fructose Glucose Sucrose
CFU/mL CFU/mL
0 7.4 5.9 5.2 31.9 0.1 42.5 66.6 180.8
8 7.3 5.1 5.2 31.4 0.4 35.4 63.2 137.2
16 7.6 5.7 5.2 31.2 0.2 37.5 45.6 110.8
1

24 7.2 5.6 5.0 28.1 0.9 46.0* 36.2 104.3

36 8.7 6.7 4.5 28.2 5.7 30.5 33.7 81.8
48 8.6 7.1 4.3 27.9 6.0 26.2 29.4 81.4
0 7.1 5.8 5.2 31.3 0.1 58.3 52.5 102.3
8 7.5 4.7 5.2 31.1 0.4 46.8 53.9 142.0
16 7.5 5.7 5.2 31.0 0.2 37.8 46.5 122.7
2

24 8.1 6.2 5.0 29.7 0.8 38.1 40.5 86.1

36 8.5 6.9 4.5 27.8 6.7* 45.5* 35.3 89.3
48 8.5 7.1 4.6 27.4 13.2* 35.2* 31.7 127.3*
0 7.3 5.9 5.2 31.6 0.1 50.4 59.6 141.6
8 7.4 4.9* 5.2 31.3 0.4 41.1 58.6 139.6
Averages

16 7.6 5.7 5.1 31.1 0.2* 37.7 46.1 116.8

24 7.8 5.8 5.0 28.9 0.9 38.1 38.4 95.2
36 8.6 6.8 4.5 28.0 6.2 30.5 34.5 85.6
48 8.6 7.1 4.5 27.7 6.0 26.2 30.6 81.4
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.
 219

S. cerevisiae, L. fermentum & Lactobacillus spp.

Bacteria Yeast Sugars (mg/g)
Time Ethanol
Log Log pH °Brix
Station (%) Fructose Glucose Sucrose
CFU/mL CFU/mL
0 7.2 5.9 5.2 27.5 0.1 42.4 73.2 148.3
8 7.1 5.9 5.2 27.6 0.3 38.7 48.4* 72.6*
16 7.9 5.8 5.0 23.9 1.2 36.8 64.4 77.4
1

24 8.4 6.5 4.5 26.1 0.8* 47.1* 24.4 58.9

36 8.3 7.0 4.3 23.1 3.7 37.8* 27.6 65.3
48 8.6 7.7 4.1 22.2 7.7 24.5 24.4 63.8
0 7.2 5.5 5.2 29.5 0.1 37.9* 73.3 17.9*
8 7.4 5.6 5.1 29.8 0.3 34.0 73.2 81.3
16 7.6 6.1 4.9 29.3 0.8 32.7 72.4 80.2
2

24 8.3 6.5 4.5 25.8 0.6* 30.6 62.6 79.1

36 8.6 7.3 4.4 27.4 4.5 30.0 61.4 80.0*
48 8.7 7.8 4.2 26.2 13.7* 28.4 51.6 66.5*
0 7.2 5.7 5.2 29.5 0.1 42.4 73.3 148.3
8 7.3 5.8 5.1 28.7 0.3 36.4 73.2 81.3
Averages

16 7.8 6.0 5.0 26.6 1.0 34.8 68.4 78.8

24 8.4 6.5 4.5 26.0 0.7* 30.6 43.6 69.0
36 8.5 7.2 4.4 25.3 4.1 30.0 44.6 65.3
48 8.7 7.8 4.2 24.2 7.7 26.5 38.0 63.8
*indicates the results identified as outliers and are excluded from graphical representations
discussed in Chapter 6.
APPENDIX C – ORGANIC ACIDS
Raw data with notations (*) on results that had significant changes (p<0.05) after a one way single factor analysis of variance and t-
test were performed to determine significant differences between means.
Individual culture fermentations (results expressed as averages of duplicate fermentations)

S. cerevisiae

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic Total
0 5.32 0.21 0.82 3.44 6.85 2.92 1.80 0.23 9.37 20.03 45.67
8 5.32 0.21 0.60 3.74 7.07* 2.49 1.95 0.20* 9.33 18.26 43.85
16 5.26 0.19* 0.66* 3.64* 6.97 2.56* 1.75* 0.19* 9.03 17.91 42.90
24 5.25 0.22 0.90 4.15 6.85 3.17* 2.23 0.23* 11.86* 25.64* 55.25
32 5.26 0.23* 0.83 3.91* 5.68 3.10* 2.34 0.27 13.12* 24.27* 53.75
40 5.17 0.24 0.82 3.75 5.95 2.94 2.50* 0.25* 10.82 21.88 49.15
48 5.18 0.21 0.76 3.52* 5.54* 2.75 2.20* 0.23* 11.46 21.81 48.48

L. fermentum

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic Total
0 5.25 0.34 0.88 4.16 8.10 2.97 1.95 0.20 10.14 20.81 49.55
8 5.22 0.36 1.18 4.77 8.93* 3.21 1.85 0.56 8.09 18.37 47.32
16 5.17 0.20 1.32 5.06 4.57 3.24 2.40* 0.09* 9.06 17.98 43.92
24 4.89 0.34 1.06 3.89 3.94* 4.58* 1.64 0.19* 10.15* 16.11* 41.90
32 4.54 0.38 1.20 3.91 3.46* 4.49* 1.83 0.85 11.75* 17.39 45.26
40 4.30 0.42* 1.37 4.49 5.93 3.47 1.85 0.60 10.87 18.90 47.90
48 4.15 0.43* 1.34 3.85 5.32 3.90 2.31 0.70 11.43 18.45 47.73

220
Individual culture fermentations (cont.)

L. plantarum

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic Total
0 5.10 0.27 1.07 3.87 4.30 3.00 1.94 0.12 6.97 20.25 41.79
8 5.09 0.27 1.02* 3.82* 4.51 3.44 1.80 0.09 6.03 24.12* 45.10
16 5.01 1.39 1.72 3.18 3.82 5.20* 1.14 0.06 5.87 22.39 44.77
24 4.81 1.31* 1.59 3.10 3.53* 6.81* 1.03* 0.17 2.86 20.68* 41.08
32 4.55 1.33* 1.64 3.00 7.26 6.07* 1.50 0.17 3.51 21.30 45.78
40 4.41 1.28* 1.71 2.50 7.82 5.20 2.89* 0.07 4.86 19.41 45.74
48 4.35 1.19 1.49 2.10 8.67* 7.37 2.99 0.12 6.14 17.65* 47.72

Lactobacillus spp.

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic total
0 5.30 0.30 1.07 4.29 5.00 3.08 2.07 0.12 10.40 22.78 49.11
8 5.27 0.32 0.87* 4.82* 9.35 4.80 3.96 0.26 13.89 22.60 60.87
16 5.16 0.25* 0.84* 4.45 9.38* 4.50 3.70 0.23 13.15* 22.56* 59.06
24 4.65 0.29 1.91 4.49 6.01 4.00 4.13* 0.42 12.55 20.48* 54.28
32 4.41 0.30 0.95 4.50 8.13* 5.80 4.11 0.16* 11.06 18.35 53.36
40 4.26 0.31 1.63 4.12* 4.33 6.10* 5.07* 0.32 5.67* 18.11 45.66
48 4.16 0.33* 1.02 4.12 6.45 6.70* 4.15 0.07* 6.51* 16.39* 45.74

221
Mixed culture fermentations

S. cerevisiae & L. fermentum

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic total
0 5.08 0.30 2.07 2.91 2.54 3.97 2.16 0.30 9.06 21.57 44.88
8 5.06 0.30 2.20 2.98 2.51 5.63 2.25* 0.30 8.51 22.25* 46.93
16 5.01 0.29 2.20 2.99 2.42 5.25* 2.21* 0.44 9.77 24.02 49.59
24 4.92 0.39 1.71* 2.73 1.05* 6.26* 1.40 0.96 9.86 23.12 47.48
36 4.79 0.32 2.30 3.15 1.88 6.41* 2.03 0.61 11.45 29.91* 58.06
48 4.68 0.30 2.35* 3.22 9.47* 6.63* 2.00 0.41 10.07 31.44* 65.89

S. cerevisiae & Lactobacillus spp.

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic total
0 5.18 0.33 2.56 3.37 1.52 3.21 2.37 0.22* 10.47 19.85 43.90
8 5.18 0.33 2.58 3.40 1.34 6.65* 2.25 0.24 10.33* 24.48 51.60
16 5.18 0.32 2.69* 3.54 1.76* 7.32 2.38 0.24 11.23 25.07 54.55
24 4.98 0.33 2.60 3.35 1.40 7.42 2.11* 0.25 10.85 27.08 55.39
36 4.49 0.33 2.55 3.44 1.14* 8.86* 2.16 0.34 12.64 26.10 57.56
48 4.40 0.32 2.69 3.56 1.17 12.62* 2.56* 0.43 15.66* 30.60* 69.61

222
Mixed culture fermentations (cont.)

S. cerevisiae, L. fermentum & Lactobacillus spp.

Time pH Organic Acids

Station Oxalic Citric Tartaric Malic Lactic Acetic Propionic Butyric Succinic total
0 5.21 0.22 1.49 2.02 1.63 3.72* 1.79 0.31 9.80* 20.41 41.39
8 5.15 0.18 1.35 1.87 2.26* 5.07 1.75 0.25 7.61* 20.44* 40.78
16 4.95 0.28 1.74 2.03 1.86 5.62 1.97 0.27 9.17 36.20 59.14
24 4.52 0.24 1.24 1.76 3.27 5.92 1.97 0.32 8.58* 50.14* 73.44
36 4.33 0.12 1.41 1.93 2.66 5.69 1.74 0.43 10.86* 43.11 67.95
48 4.16 0.11 1.41 1.98 2.27 9.18* 1.37* 0.54 13.96* 31.09* 61.91

223