HURDLE TECHNOLOGIES: MICROBIAL INACTIVATION BY PULSED ELECTRIC
FIELDS DURING MILK PROCESSING
A Thesis
Presented to
The Faculty of Graduate Studies
of
The University of Guelph
by
OSCAR RODRIGUEZ GONZALEZ
In partial fulfillment of requirements
for the degree of
Doctor of Philosophy
December, 2010
©Oscar Rodriguez Gonzalez, 2010 ABSTRACT
HURDLE TECHNOLOGIES: MICROBIAL INACTIVATION BY PULSED ELECTRIC
FIELDS DURING MILK PROCESSING
Oscar Rodriguez Gonzalez Advisor:
University of Guelph, 2010 Dr. Mansel W. Griffiths
The application of non-thermal processes pulsed electric fields (PEF) and cross-flow micro-filtration (CFMF) continuous to be studied with the purpose of controlling microorganisms in milk. Trends suggesting increased adoption include the study of
Food Safety Objectives as a safety criterion, the promotion of sustainable processing, and the implementation of hurdle strategies. While the advance of gentle processing is counteracted by the risk of enhanced resistance due to microbial stress response, several techniques can be applied to quantitatively assess its impact. The objective of this project was to evaluate the effectiveness of microbial inactivation by PEF and
CFMF at various steps of milk processing including shelf-life, its comparison with high temperature short time (HTST) pasteurization, and the quantitative assessment of the cross protection resistance to PEF of Escherichia coli O157:H7.
Some differences in mesophilics inactivation were observed in milks (fat contents between 1.1% and 3.1%). Increasing the PEF inlet temperature decreased the treatment time by three or two-fold. The combination of CFMF/PEF yielded similar microbial reductions as CFMF/HTST. Higher inactivation of the coliforms was achieved in homogenized cream (12% fat) compared to non-homogenized. The linear relation between electrical conductivity and nutrient content (fat and solids content) was established. In a parallel study the PEF/CFMF sequence resulted in higher inactivation of mesophilics compared to CFMF/PEF and HTST. The shelf life was acceptable for
CFMF/PEF and HTST after 7 days, while enterics and psychrotrophs grew more after
PEF/CFMF, thermodurics did after HTST.
The growth and stress of Escherichia coli O157:H7 in lactose containing broths was monitored by absorbance and fluorescence expression of stress reporters. Growth was explained using a secondary model, and stress response using mechanistic and probabilistic models. PEF inactivation was evaluated following the Weibull distribution after the cells reached stationary phase or maximum fluorescence expression. Similar resistances were observed within the cells grown in lactose broth at 10, 25 or 40°C, as within stressed cells (starved or cold shocked). Cells grown at 45 °C were more resistant compared to the cells grown in acid, high salt concentration while the ones grown at cold temperatures were the weakest. ACKNOWLEDGMENTS
To all my colleagues, advisors, family, and friends for their support.
To the Dairy Farmers of Ontario and the Natural Sciences and Engineering Research
Council for their financial assistance.
i! ! ABBREVIATIONS
Abbreviation Significance
AA Acetic acid
ABS Absorbance
AIC Akaike information criterion
ANOVA Analysis of variance
ATCC American Type Culture Collection
BIC Bayesian information criterion
CFMF Cross-flow microfiltration
CFU Colony forming units
CIPEC Canadian Industry Program for Energy Conservation
ESL Extended shelf life
FIL-IDF International Dairy Federation
FSO Food safety objective
G/NG Growth/no growth
GFP Green fluorescence protein
GWP Global warming potential
HEST High electric field short time
HHP Hydrostatic high pressure
HPH High pressure homogenization
HPHT High pressure high temperature
ii! ! HSP Heat shock proteins
HTST High temperature short time
LA Lactic acid
LACB Lactose broth
LB Luria Berthani Broth, Miller
MF Microfiltration
MPN Most probable number
MSE Mean square errors
MTS Manothermosonication
NACMCF National Advisory Committee on Microbiological Criteria for Foods
NF Nanofiltration
OD Optical density
PATS Pressure assisted thermal sterilization
PC Performance criterion
PEF Pulsed electric fields
PFN Pulse forming network
PI Process integration
PI Propidium iodide
PL Pulsed light
PO Performance objective
RFEF Radio frequency electric fields
RNA Ribonucleic acid
RO Reverse osmosis
iii! ! SEM Standard error of the mean
TAL Thin agar layer
TAMC Total aerobic mesophilic counts
TEBC Total enteric bacteria counts
TLABC Total lactic acid bacteria counts
TMP Trans-membrane potential
TMYC Total moulds and yeasts counts
TP Traditional pasteurization
TPC Total psychrotroph counts
TSB Tryptic soy broth
TSBL Tryptic soy broth with added lactose (0.5%)
TSC Total staphylococci counts
TTMC Total thermoduric mesophilic counts
TTPC Total thermoduric psychotroph counts
TTTC Total thermoduric thermophilic counts
UF Ultrafiltration
UHT Ultra high temperature
UTMP Uniform trans-membrane pressure
WeLL Weibullian-log logistic model
iv! ! NOMENCLATURE
Symbol Significance units
a decay rate kV/mm
ABS absorbance -
Af Accuracy factor -
- aw water activity
b number of bacteria at the end of time interval log CFU/ml
B number of bacteria at the beginning of time interval log CFU/ml
b inactivation rate !s-1
bE regression coefficient for electric field -
bt regression coefficient for treatment time -
C storage capacity or charging capacitance F
Cp specific heat capacity kJ/kg°C
cr cell radius m
ctrl non-transformed cells -
d electrode gap m
DX°C decimal reduction time (at given temperature X°C) s
E electric field strength kV/m
EC0 critical value of E for 100% survival kV/mm
E critical E value constant C00 kV/mm f repetition rate or frequency pulses per second, Hz f(t) density function -
v! ! Fat fat fraction %
FSO food safety objective CFU/l
G generation time min, h
g rate constant -
gfpUV plasmid containing cells -
H0 the initial level of the hazard CFU/l
I current signals A
i intercept of relation between % fat and electrical %fat Fat conductivity
intercept of relation between % solids and electrical i %sol Sol conductivity
k constant factor -
k (E,Tin) inactivation constant, -
k1,2,3…8 dimensionless empirical parameters -
log (N/N0) reduction in log survivor count -
m shape parameter -
!" mass flow rate kg/s
n number of pulses -
N the number of observations used in the calculation -
Obs observed -
P pressure Pa
p shape parameter -
pH pH h-1
PO performance objective -
vi! ! Pre predicted -
q concentration of limiting substrate -
R response ratio -
r rate of response -
RC PEF chamber resistance !
RFI relative fluorescence intensity -
S surviving fraction -
s slope of relation between % fat and electrical S / m %fat Fat conductivity
SFI specific fluorescence intensity -
Sol solids fraction %
s slope of relation between % solids and electrical S / m %sol Sol conductivity
t time s
T temperature °C
tc the mean of the microbial population resistance s
Tc temperature level of inactivation onset °C
tC0 critical value of tt for 100% survival s
Tin inlet temperature °C
TMP trans-membrane potential V
tt treatment time "s
U charging voltage kV
V Volume l
!" volumetric flow rate m3/h
vii! ! v rate of increase of the limiting substrate -
3 VC volume of PEF treatment chamber m
w total applied energy kJ
wpl specific energy per pulse kJ/pulse
wsp specific energy input kJ/kg, kJ/l
x microbial population log CFU/ml
! scale parameter -
" shape parameter -
! gamma function -
# pH parameter -
$P pressure rise Pa
$T temperature rise °C
pump and motor efficiency, 0.85 for positive pumps, and % - 0.65 for efficient centrifugal pumps
& upper asymptote log CFU/ml
' electrical conductivity of the product S/m
( growth rate h-1
-1 (max maximum growth rate h
) density kg/l
* maximum slope of the inactivation curve s-1
+I CFU/l total (cumulative) increase of the hazard on a log scale the total (cumulative) reduction of the hazard on a log +R CFU/l scale
log time at which the maximum slope is reached " s (position of the curve)
viii! ! !w pulse width s
-1 -1 ! Ratkowsky parameter °C h
angle between the electric field and the point of interest ! - in the membrane (cos ! = 1 at the poles)
" a form factor (3/2) for spherical cells -
# lower asymptote log CFU/ml
" lower asymptote log CFU/ml
$ temperature parameter -
" water activity parameter -
ix! ! !
TABLE OF CONTENTS
ACKNOWLEDGMENTS………………………………………………………………………...i
ABBREVIATIONS…………………………………………………………………………....….ii
NOMENCLATURE…………………………………………………………………………...... v
1. Introduction ...... 1
1.1. Non-thermal dairy processing ...... 1
1.1.1. Advances in critical milk processes ...... 2
1.1.2. Microfiltration ...... 5
1.1.3. Emerging technologies ...... 7
1.1.4. Regulatory status of pasteurization technologies ...... 11
1.1.5. Sustainable dairy processing ...... 14
1.2. Pulsed electric fields ...... 17
1.3. Combined processes ...... 25
1.3.1. Advanced combined processes ...... 25
1.4. Microbial stress ...... 28
1.4.1. Microbial stress responses ...... 29
1.4.2. Cross protection ...... 31
1.4.3. Risks associated to PEF ...... 32
1.5. Microbial physiology and inactivation ...... 32
1.5.1. Mechanism of action ...... 36
1.6. Advances in microbial modeling ...... 37
1.6.1. Modeling growth ...... 40
1.6.2. Modeling inactivation ...... 42
x! ! !
1.7. Research needs ...... 48
2. Factors affecting the inactivation of native bacteria in milk processed by pulsed electric fields and cross-flow microfiltration ...... 49
2.1. Introduction ...... 49
2.2. Materials and Methods ...... 51
2.2.1. Conventional and alternative milk processing ...... 51
2.2.2. PEF treatment ...... 52
2.2.3. Conventional heat treatment of milk ...... 56
2.2.4. Microbiological Analysis ...... 56
2.2.5. Physico-chemical properties ...... 57
2.2.6. Statistical analysis ...... 58
2.3. Results and Discussion ...... 58
2.3.1. Bacteria reduction in PEF-treated milk of different fat contents ...... 58
2.3.2. Effect of the PEF inlet temperature on the reduction of microorganisms ...... 64
2.3.3. Milk processing with a combination of CFMF and PEF ...... 68
2.3.4. Inactivation of native bacteria in cream by PEF ...... 71
2.3.5. Effect of processing on the physicochemical properties of milk ...... 75
2.4. Conclusions ...... 79
3. Microbial inactivation and shelf life comparison of ‘cold’ hurdle processing with pulsed electric fields and cross-flow microfiltration, and conventional thermal pasteurisation in milk ...... 81
3.1. Introduction ...... 81
3.2. Materials and methods ...... 83
3.2.1. Milk preparation and analysis of microbial inactivation ...... 83
3.2.2. Nonthermal processing with PEF ...... 85
3.2.3. Nonthermal processing with CFMF ...... 87
xi ! !
3.2.4. Nonthermal hurdle processing ...... 87
3.2.5. Conventional heat pasteurisation ...... 88
3.2.6. Microbiological shelf life ...... 88
3.2.7. Electrical conductivity and pH ...... 90
3.2.8. Statistical analysis ...... 90
3.3. Results ...... 91
3.3.1. Microbial inactivation with PEF, CFMF, CFMF/PEF, PEF/CFMF, and TP .... 91
3.3.2. Microbial shelf stability of milk processed with PEF/CFMF and TP ...... 92
3.4. Discussion ...... 98
4. Resistance of temperature, pH, osmotic and starvation stress induced Escherichia coli O157:H7 to pulsed electric fields in milk ...... 104
4.1. Introduction ...... 104
4.2. Materials and Methods ...... 106
4.2.1. Microorganism ...... 106
4.2.2. Growth medium preparation ...... 107
4.2.3. Construction of stress reporters ...... 108
4.2.4. Monitoring of the physiological state ...... 109
4.2.5. PEF treatment ...... 110
4.2.6. Microbial counts ...... 111
4.2.7. Mathematical Analysis ...... 111
4.3. Results and discussion ...... 115
4.3.1. Growth environment ...... 115
4.3.2. Stress response ...... 118
4.3.3. Cross-protection to physical hurdles ...... 121
4.4. Conclusions ...... 130
xii ! !
5. General conclusion ...... 131
5.1. Applications and future research ...... 135
6. References ...... 137
7. Annex ...... 176
xiii ! !
LIST OF FIGURES
Figure 1-1. Relative milk components and particles size...... 4
Figure 1-2. Cross-flow micro-filtration with uniform trans-membrane pressure...... 6
Figure 1-3. Block diagrams of pulse power systems and their pulse shapes...... 18
Figure 1-4. Types of treatment chambers...... 22
Figure 1-5. Basic components of a high intensity pulsed electric field treatment system...... 23
Figure 1-6. Common types of microbial inactivation curves...... 44
Figure 2-1. Milk process sequence in conventional and alternative fluid milk processing ...... 53
Figure 2-2. Relation between electrical conductivity and fat or solids content ...... 77
Figure 3-1. Schematic drawing of the pulsed electric fields (PEF) treatment chamber used for skim milk processing and sequence for combined processing of PEF with cross-flow microfiltration (CFMF) ...... 84
Figure 4-1. Growth rates of Escherichia coli O157:H7 in various growth media, predicted by the Cardinal and Ratkowski models...... 117
Figure 4-2. . Response ratios of Escherichia coli O157:H7 fluorescence gene promoters fitted by mathematical models...... 120
Figure 4-3. Inactivation trends of Escherichia coli O157:H7 by pulsed electric fields (PEF) processing after growth at challenging conditions and PEF treatment at 1.8 kV/mm and 20 Hz or following environmental stress and PEF treatment at 2.4 kV/mm and 20 Hz...... 126
xiv ! !
LIST OF TABLES
Table 2-1. Applied intensitites of high intensity pulsed electric fields and resulting temperature profiles ...... 54
Table 2-2. Variation in microbial populations after pulsed electric fields (PEF) and high temperature short time (HTST) processing of fluid milks ...... 60
Table 2-3. Effect of the pulsed electric fields inlet processing temperature on the reduction of different bacteria groups in skim milk ...... 65
Table 2-4. Reduction of microbial populations after processing of skim milk with cross- flow microfiltration (CFMF) and pulsed electric fields (PEF) or high temperature short time (HTST) pasteurization in a combined approach (CFMF/PEF; CFMF/HTST) ...... 69
Table 2-5. Effect of pulsed electric fields processing on different groups of native bacteria in cream ...... 72
Table 2-6. Linear regression coefficients of the relation between fat or solids and conductivity of milk fractions ...... 76
Table 3-1. Comparison of reductions in total mesophilic aerobic counts in skim milk obtained following pulsed electric fields (PEF), cross-flow microfiltration (CFMF), thermal pasteurisation (TP); combined CFMF/PEF, and PEF/CFMF treatment ...... 94
Table 3-2. Microbial counts in skim milk before and after pulsed electric fields (PEF) and cross-flow microfiltration (CFMF) hurdle treatment (PEF/CFMF) or thermal pasteurisation (TP) during storage ...... 96
Table 4-1. Mathematical models used to analyze Escherichia coli O157:H7 during stress induced growth and its resistance to subsequent pulsed electric field processing in milk...... 113
2 Table 4-2. Models goodness-of-fit (AIC, BIC and adjusted R ), and coefficients (Rmax, r, ti, g, m), resulting from analysis of fluorescence response of Escherichia coli O157:H7 stress gene promoters after induced stress (starvation or cold-shock)...... 119
Table 4-3. Growth and stress conditions of Escherichia coli O157:H7 and its relative resistance to pulsed electric fields (PEF)...... 122
Table 4-4. Weibull and Weibullian-log logistic (WeLL) inactivation parameters describing Escherichia coli O157:H7 inactivation in milk by pulsed electric fields (PEF) processing...... 124
xv ! !
!
1. Introduction
1.1. Non-thermal dairy processing
Because it has been generally recognized that heat pasteurization does not necessarily achieve commercial sterility (NACMCF 2006), and that excessive heat treatment can be associated to other health risks resulting from the formation of toxic substances, a group of technologies labeled non-thermal makes continuous advances
(Knorr and others 2007). In that regard, minimal processing of liquid foods creates products with improved sensory and nutritional attributes due to the reduced thermal impact (Walkling-Ribeiro and others 2010a). Parallel to the increase in milk production, new pasteurization processes are being implemented in emerging markets. While the application of ultra-high temperature (UHT) treatment of milk has increased because it benefits the distribution chain due to its shelf stability, other technologies like ultra- filtration (UF) have been introduced to satisfy consumer quality demands (FIL-IDF
2006).
Emerging technologies have the potential to improve the safety of dairy products, and these have been the subject of several books and book chapters (e.g. Pereira and
Vicente 2010, Tewari and Juneja 2007, Sun 2005, Barbosa-Cánovas and others
2005a).
Based on the physics involved in their mode of operation, processes can be grouped as those involving momentum, energy (heat) or mass transfer (Ibarz and
Barbosa-Cánovas 2003), or involving electro-thermal, electro-magnetic, and physical
"! ! ! mechanisms (FDA 2000). As quoted by McBee (1996), understanding the fundamental physics of a process allows a better perspective of the operations, and probably reveals unutilized forms of energy (See Annex 1 for an illustration of the processes related to the electromagnetic spectrum).
1.1.1. Advances in critical milk processes
Milk is generally subjected to cream separation because of the tendency of the fat globules to rise to the top of their container. Based on the principle of fat separation by sedimentation forces, centrifugal separators are continuously being developed to achieve separations 4500 times faster than gravity (Nielsen 2000). This technique has been further developed to remove microorganisms in a process called bactofugation, which consists of a centrifuge bowl with continuous discharge where milk flows through distribution holes into disk stacks to be ejected as a heavy phase (transport liquid) and a bacterially clarified product. This process results in a reduction in spores of up to 99.5% when two phases are used in-line (Van den Berg 2001a).
The high fat fraction of the milk (cream) is usually subjected to a homogenization process, where the fat globule diameter is decreased to delay the formation of a fat layer during storage (Kessler 2002). High-pressure homogenizers (HPH) disrupt cells or particles by passing the medium through an adjustable, restricted orifice discharge valve. For the purpose of fat globule disruption, milk is homogenized in two stages using pressures around 15 MPa followed by a lower pressure at the second stage. For microbial cell disruption, pressures between 55 and 200 MPa are applied (Geciova and others 2002).
#! ! !
A basic HPH unit consists of a positive displacement pump that forces the suspension through the center of a valve seat and across the seat face. As the fluid flows radially across the valve and strikes an impact ring, disruption results from tearing apart the cell wall. Based on a similar principle a microfluidizer causes impaction of two streams of suspension at high velocity against a stationary surface with the energy input dissipated almost instantaneously at the point of impact. The operation is a function of flow rate, disruption unit, and the backpressure. While researchers like Hayes and others (2005) suggest that a significant microbial reduction can be achieved by HPH
(250 MPa and 45°C), other investigators like Smiddy and others (2007) state that HPH does not extend the shelf-life of milk, even when combined with relatively high temperatures (250 MPa at 70°C).
The low fat fraction (skimmed milk) can be further separated by membrane processes, which are classified according to the particle size filtered as: microfiltration
(MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO). While the pore sizes for MF are generally higher than 0.2 !m, and for UF between 3 and 300 nm, the molecular weights of the molecules separated by RO range between 1000 and 5000 kDa and for NF between 100 to 500 kDa as they separate based on the solubility of the components (Marcelo and Rizvi 2009). An illustration of the various particles commonly found in milk is illustrated in Figure 1-1.
$! ! !
M04?-205-/0;< Y231/.;
,-./[email protected]/3.04/01 \;<01 J;231?2-. J-1.;8;231?2-. E322?2-.!]!801.;H-./01?2-/3
:HH.;U08-/3! #$& #$% #$" #$) )!U!#$) V30=@/!6W-2/;<49
E-430?8-
L-<;( J01.;( N1.33<4!F!4/.-0<3.4 J38D.-<3! I02/.-/0;< I02/.-/0;< H.;134434 P3Q3.43! K2/.-(I02/.-/0;< E2;/@!F!G3H/@!I02/3.4 "! R48;404
N'!13224
M0.?434 A3-4/!13224 J01.;;.=-<0484 B-1/3.0- J;?2G4
,-./0123!4053!6789 #$() #$(" #$(% #$(& $'# #'$ #$ #$& #$% #$"
,-./0123!4053! # & % " ) + * Z 6:.84/.;<=49 #'$ #$ #$ #$ #$ #$ #$ #$ #$ #$
Figure 1-1. Relative milk components and particles size.
Adapted from Walstra and others (2006), CDR (2004), Kessler (2002), Nielsen (2000), and Jelen (1992).
! ! !
1.1.2. Microfiltration
As the design of membrane equipment advances, different approaches are being taken to satisfy the processing needs. Membrane systems can be batch or continuous.
Continuous systems can be single-pass, plug-flow or multi-stage with recycling (Nielsen
2000). A further development to achieve higher permeation rates and minimum fouling in a continuous system was the concept of cross flow microfiltration (CFMF). As explained by Kessler (2002), it consists of applying a high tangential velocity in the filtration module (high wall shear stress), a low trans-membrane pressure drop, and the highest appropriate temperature (50-55°C). By allow ing a low and uniform trans- membrane pressure (UTMP) of 30 to 40 kPa, rapid clogging and fouling could be prevented and a 1 to 10 ratio between the retentate and permeate could be obtained
(Pedersen 1992). With this design, a stable flux level of at least 500 L/m2/h (normally
700 L/m2/h) for more than 6 hours can be achieved. An illustration of the CFMF with
UTMP membrane system is shown in Figure 1-2.
CFMF is applied in the dairy industry in three major areas: removal of bacteria, whey defatting, and micellar casein enrichment of milk for cheese making (Marcelo and
Rizvi 2009). For the removal of microorganisms it has been shown that CFMF can remove between 99.1 and 99.9% of the total population, 99.95% of Bacillus spores, and
98.4% of other bacterial spores with minimal changes in protein and dry matter
(Pedersen 1992).
%! ! !
Permeate Retentate
' '
3.6 Bar 4.0 Bar Permeate Retentate circulation circulation pump pump
1.8 Bar 2.2 Bar
' ' 1.4 !m
Feed pump
Skim milk " 50 C
Figure 1-2. Cross-flow micro-filtration with uniform trans-membrane pressure.
Adapted from Pedersen (1992).
&! ! !
Because the economics of UTMP are determined by the design that minimizes the load of the pump on the permeate side, manufacturers like Tetra-Pak (Alfa-Laval) offer modules that are entirely packed with plastic spherical particles of defined parameters on the permeate side, while APV-CREPACO achieve the same objective by adding more tubes wrapping individual membrane elements (Kessler 2002). The energy consumed during a pressure-driven process can be calculated based on the operating pressure even when it is only part of the total energy consumption. In cross-flow based processes an important part of the energy consumed is at the recycling pumps that continuously flush the membrane. This energy can be estimated based on Eq. 1-1
(Nielsen 2000).
!"#$% & '!(#, )&*+( Eq. 1-1!
For comparison purposes, the energy consumption by the main four types of cross flow separation processes has been estimated as follows (Nielsen 2000): 9 to 12 kWh/m3 for micro-filtration (32.4 to 43.2 kJ/L), 2 to 5 kWh/m3 for ultra-filtration (7.2 to
18.0 kJ/L), 4 to 6 kWh/m3 for nano-filtration (14.4 to 21.6 kJ/L), and 5 to 10 kWh/m3 for reverse osmosis (18.0 to 36.0 kJ/L).
1.1.3. Emerging technologies
Pressures higher than HPH (up to 1000 MPa) can be applied in specially designed equipment. This equipment can be categorized in two types: batch (solid foods) and in-line (liquid foods). While in-line systems are more attractive for fluid milk processing, the costs of high pressure processing can be five times higher than
(! ! ! conventional heat treatment, with the initial investment being 75-80% of the cost.
Another problem to solve is the cycle time, some semi-continuous systems save energy and time by coupling vessels using the energy stored in a pressured one to pressurize a second one (Van den Berg and others 2001b). As the research on high pressures, especially the application of hydrostatic high pressure (HHP), advances beyond the antimicrobial effects, attributes such as the preservation of the level of free fatty acids related to decreased rancidity, the fat globule cold agglutination, and differences in rennet coagulation after HHP represent exciting opportunities for the dairy industry
(Norton and Sun 2008). Also, since the structure of milk proteins can be modified by
HHP, research has been focused on the modification of the properties of foams, emulsions and gels and the application in cheese production both prior to cheese making and during ripening (Needs 2001).
Innovative electro-magnetic processes include ionizing radiation (x-rays, gamma rays, and beta rays), thermal radiation (ultraviolet to infrared), magnetic fields, and the application of electricity. Ionizing radiation was traditionally generated from radioactive isotopes using nuclear reactors (Pillai and Braby 2007), but with the development of linear accelerators its application and safety has been improved, and recently is more similar to the electricity-based pasteurization equipment (AEB 2009a, Miller 2005). Its application in the dairy industry includes package sterilization. Pulsed light technology
(PLT) is based on the application of electromagnetic radiation by means of waves at different wavelengths (#), frequencies ($) and energy (E). The types of light include: UV
(# = 180-400 nm, subdivided into UVA = 315-400 nm, UVB = 280-315 nm, UVC = 180-
280 nm), visible light (# = 400-700 nm), and infrared rays (# = 700-1100 nm). Similar to
)! ! ! ionizing radiation a few applications have been developed specifically for dairy products.
UV light is a well-established technique applied for sterilizing packaging films; therefore, its future is promising (Palmieri and Cacae 2005), and there is a project team developing a prototype to assess the commercial viability of non-thermal UV processing
(Barnes 2009).
Radio-frequency is being explored for its beneficial non-thermal effects when applied in the form of radio-frequency electric fields (RFEF) at higher voltages (%5 kV/cm) and lower frequencies (<18 MHz), for that purpose a process applying 3.0 kV/mm between 20 and 100 kHz has been developed to inactivate microorganisms. The setup of RF is similar to pulsed electric fields (PEF) as they consist in moving a fluid through two electrodes charged with an electric field, the difference is that the PEF power supply is in the forms of pulses formed in a pulse-forming network (PFN), while in
RFEF the power supply continuously provides the high voltage making it a simpler method. RFEF is claimed that it will require the same energy as PEF, and the two technologies cost approximately twice that of conventional pasteurization (Geveke
2005). These two technologies (PEF and RFEF) are being studied in parallel by the US
Department of Agriculture, Food Safety and Intervention Technologies Unit (USDA
2010).
An additional non-thermal technology that is advancing relatively slowly involves the use of magnetic fields generated by superconducting coils that produce DC fields, and is defined as a region of space in which a magnetic body is capable of magnetizing the surrounding bodies (Barbosa-Cánovas and others 2005b, FDA 2000).
*! ! !
Electro-thermal processes are continuously being developed to optimize the application of thermal treatments. This group of processes includes: infrared radiation, hertzian waves (microwaves and radiofrequency), and ohmic heating. Infrared radiation consists of transferring thermal energy in the form of electromagnetic waves as short waves (0.7 to 2.0 !m), medium waves (2.0 to 4.0 !m), and long waves (0.004 to 1.0 mm), generated by a hot source like a quartz lamp, quartz tube or metal rod (Vicente and Castro 2007). Hertzian waves include electromagnetic waves in the range between
150 kHz and 30 GHz and include microwaves applied at 915 or 2450 MHz, and RF at
13.56, 27.12, or 40.68 MHz (Tewari 2007, Orsat and Raghavan 2005). Ohmic heating is generated by using an unrestricted frequency and waveform to generate heat due to the electrical resistance of the foods (Vicente and Castro 2007). As these technologies are improved, their advantage compared to traditional heating is that the heat is generated at the molecular level resulting in instantaneous heating. While the processing time can be reduced, thus increasing operational efficiency, the major cost is the initial investment (Tewari 2007, Vicente and Castro 2007, Sumnu and Sahin 2005, Orsat and
Raghavan 2005).
As technologies that pursue commercial sterilization advance, established concepts like UHT continue to be improved based on its two basic principles: 1) for the same bacterial destruction a high-temperature short-time treatment results in less chemical change than a low-temperature long-time treatment, and 2) minimum times and temperatures are dictated by the need to inactivate thermophilic bacterial spores, while the maximum times and temperatures are determined by the need to minimize undesirable chemical reactions (Datta and Deeth 2007).
"+! ! !
1.1.4. Regulatory status of pasteurization technologies
According to the Canada National Dairy Code (CFIS 2005) pasteurization is the process of heating every particle of a dairy product in equipment that is designed and operated to meet or exceed the required time and temperature relationships specified.
The minimum pasteurization requirements for milk based products (<10% fat) are 63 °C x 30 min (batch/vat) or 72 °C x 15 s (HTST), for pr oducts with higher fat content or sugar (>10%), the temperature requirements are higher (66 °C x 30 min or 75 °C x 15 s), as for frozen dairy product mixes time-temperature combinations of 69 °C x 30 min,
83 °C x 15 s or 80 °C x 25 s are required. The Onta rio Milk Act Regulation 761 R.R.O.
1990 (OMAFRA 1990) states that skim milk and whey should be pasteurized at least at
69 °C x 30 min, milk for the purpose of cheese maki ng at 62 °C x 30 min, or 72 °C x 16 s, and ice milk mix or ice cream mix by heating the mix to 69 °C x 30 min or 80 °C x 25 s, and cream to manufacture butter for 77 °C x 10 m in.
Initiatives such as the Guide to Energy Efficiency Opportunities in the Dairy
Processing Industry aim to assist in the identification of energy efficiency improvements in dairy plants and the development and achievement of voluntary sector energy efficiency targets. The Canadian Industry Program for Energy Conservation (CIPEC) promotes new technologies relevant to energy/water efficiency including non-thermal pasteurization methods including MF, HHP, PEF, and UV (Wardrop 1997).
In the United States milk safety is guided by the pasteurized milk ordinance (FDA
2009), and by definition pasteurization is the process of heating every particle of milk or milk product in a properly designed and operated equipment to one of the required temperatures (63 °C x 30 min, 72 °C x 15 s, 89 °C x 1 s, 90 °C x 0.5 s, 94 °C x 0.1 s, 96
""! ! !
°C x 0.05 s and 100 °C x 0.001 s), with a requireme nt to increase the temperature by 3
°C for milks with fat content higher than 10%. The International Dairy Federation (FIL-
IDF) defines pasteurization as a process applied to a product with the object of minimizing possible health hazards arising from pathogenic microorganisms associated with milk. This is generally achieved by applying a heat treatment, which results in minimal chemical, physical and organoleptic changes to the product. Various temperature and time combinations are used by the dairy industry including: thermization, which requires heating at 63 to 65°C for 15 to 20 seconds to maintain milk quality at the plant pending final use; pasteurization to destroy pathogenic microorganisms conducted at 63°C x 30 minutes or 72 °C x 15 s, and commercial sterilization, which is achieved by heating to about 115°C x 10-20 min, or %135°C for a few seconds (Cerf 1986).
Depending on the intensity of heat treatment, pasteurization can be low (i.e. 74
°C x 15 s) and used to primarily kill most microorg anisms and some enzymes without causing other changes, or high (i.e. 90 °C x 15 s) and used to kill all vegetative microorganisms, and inactivate most enzymes, but making some serum proteins insoluble. Commercial sterilization is intended to kill all microorganisms, and all enzymes, but chemical changes including browning reactions, and formation of formic acid may occur. These changes can be minimized by heating at high temperatures for a few seconds (Walstra and others 2006).
The term pasteurization has been revised and broadened by the National
Advisory Committee on Microbiological Criteria for Foods (NACMCF). While NACMCF recognized that pasteurization does not necessarily achieve commercial sterility and
"#! ! ! pasteurized foods must be frozen or refrigerated to preserve product quality, NACMCF agreed to define pasteurization as follows: Any process, treatment, or combination thereof that is applied to food to reduce the most resistant microorganism(s) of public health significance to a level that is not likely to present a public health risk under normal conditions of distribution and storage (NACMCF 2006). NACMCF also recommend that regulatory agencies establish a Food Safety Objective (FSO) and/or a performance standard for food pathogen combinations that can be used as the basis for judging equivalency when a process is evaluated as an alternative to traditional pasteurization. The FSO concept has been previously proposed by the International
Commission on Microbiological Specifications for Foods (ICMSF 2002), and its application is currently being reviewed by a collaborative group facilitated by the
International Life Sciences Institute (Bourdichon and others 2009, Membré and others
2009).
The Performance Criterion (PC) could be defined as the effect on frequency and/or concentration of a hazard in a food that must be achieved by the application of one or more control measures to provide or contribute to a Performance Objective (PO) or a FSO (NACMCF 2006). In milk a FSO can be considered as achieving a low probability that a serving of milk contains a viable vegetative cell of a pathogen (e.g. < 1 cell per 1,000 servings), and a PC as a treatment sufficient to reduce the levels of C. burnettii by greater than 6-log cycles (inactivation of greater than 100,000 guinea pig infectious doses, assuming that each guinea pig infectious dose contains at least one, or, more likely, more than one viable rickettsia). A PC can be defined by Eq. 1-2:
"$! ! !
-. / 0 120 34567 Eq. 1-2
The future for regulatory acceptance of non-thermal technologies is promising.
Two of the leading authorities assessing the risk that novel technologies, such as PEF, may present to public health are the European Food Safety Authority and the U.S. Food and Drug Administration. Therefore, there is hope in the industrial sector that this regulatory scrutiny will accelerate the commercialization of emerging technologies and its products (Smith 2007).
1.1.5. Sustainable dairy processing
Because of the increasing public awareness of the environmental impact caused by human activities, the dairy industry has taken the lead in analyzing the different factors involved in the production of its goods. A recent publication of the International
Dairy Federation (FIL-IDF 2009) shows that a relatively large part of the environmental impact occurs at the dairy farm; accounting for 1.0 of the global warming potential
(GWP) expressed as kg of CO2 per kg of milk compared to 0.2 for all the other manufacturing and distribution stages to the consumer. Among dairy products the highest GWP is attributed to cheese, eight times higher than milk and yoghurt. The non- renewable energy use at the dairy farm can be up to 1.5-times higher compared to all other phases. For dairy products the lowest consumption of non-renewable energy is for milk, followed by yoghurt, cream, cheese, butter, and milk powder.
Sustainable processing is highly related to efficient processing, and, therefore, economics. Early analyses of the distribution in energy consumption in a typical milk processing line were made by Miller (1986). In his review he states that 22.6% of the
",! ! ! electricity and 34.4% of the fuel were consumed at the pasteurization operation in the manufacture of bottled milk. Pasteurization was the second highest operation in electricity consumption after cold storage (40.9%), and also second in fuel consumption after bottle washing (58%). It was estimated that, in total, 88 kJ/kg of fuel and 20 kJ/kg of electrical energy were consumed in the manufacture of pasteurized bottled milk.
A more recent review of the pasteurized milk production process was conducted by Nicol and others (2005), who observed that the energy consumption for pasteurized milk production in a line that uses 13 MPa for homogenization and 90% regeneration at the heat exchanger consumed a total of 133.6 kJ/L of thermal energy and 11.5 kJ/L of electrical energy per hour in a 24,000 L/h processing plant. The energy consumption can be decreased with higher heat exchanger regeneration efficiency or partial homogenization.
As the design of dairy processing equipment advances, the study of heat recovery systems at the dairy plant is being expanded to the transfer of heat between operations by applying a technique called process integration (PI) or pinch analysis, and consists of analyzing energy uses in the whole process rather than a unit operation. The issues to be considered when conducting this analysis are: type of process (batch or continuous), contamination avoidance, fouling, and compatibility with existing equipment
(Hanneman and Robertson 2005).
In a modern milk production line and in advanced heat pasteurization equipment the regeneration section is used to pre-heat the milk for separation and homogenization prior to pasteurization (Muir and Tamime 2001). A thermal pasteurization cycle can be summarized as 5-65-75-15-5 °C, needing a pre-heatin g and post-cooling of 10°C only, a
"%! ! !
PEF treatment can be applied as: 5-35-50-5 °C; cons equently, it needs a pre-heating of
30°C and post-cooling of 45°C making regeneration n ot feasible (Hoogland and de
Haan 2007).
As reviewed by Toepfl and others (2006) there is a potential to apply PEF for environmentally friendly and energy efficient processing, the strategies analyzed can be related to dairy processing in the sense that the pasteurization of liquid food can be improved by achieving a sufficient reduction of microbes with low energy consumption by combining PEF with mild heat, and the waste water treatment can be improved by disintegrating excess sludge and destroying cells to release intracellular material to be utilized to initiate biodegradation and autolysis of cells. Therefore, new technologies have a niche in sustainable food processing, as individual techniques like RFEF and
PEF are compared by their energy usage (Geveke 2005).
Non-thermal processes are finding their niche in sustainable dairy processing with UV technology being studied for its potential to cut demand for steam needed for pasteurization (ICUSD 2009) as part of the U.S. Dairy Sustainability Initiative (ICUSD
2008). Electron beams are being used as an alternative for package sterilization in extended shelf-life (ESL) milk. For example, milk produced by Dairy Stix Ltd (UK) is packed using a Form-Fill-Seal process sterilized with this technology, which is claimed to have a supply chain and economic advantage compared to similar products like the traditional hydrogen peroxide treatment (AEB 2009ab).
"&! ! !
1.2. Pulsed electric fields
Fundamentally pulsed electric field technology consists of delivering an electric field in pulses to a food medium. To generate a high voltage of pulsed electrical field of several kV/cm within food, a large flux of electrical current must flow through a piece of food in a treatment chamber for a very short period of time (Zhang and others 1995).
The high voltage power supply for the system can be either an ordinary source of direct current or a capacitor charging power supply with inputs of high frequency alternating current (AC).
Pulses are formed by the pulse-forming network (PFN), which consists of one or more power supplies with the ability to charge voltages (up to 60 kV), switches (ignitron, thyratron, tetrode, spark gap, semiconductors), capacitors (0.1-10 !F), inductors (30
!H), resistors (2 & – 10M&), and a treatment chamber (Barbosa-Cánovas and
Altunakar 2006). A simple PFN is a resistance-capacitance circuit in which a power supply charges a capacitor that can deliver its stored energy to a resistive load
(treatment chamber) by activation of a switch. The energy storage capacitor (C0) determines the total power of the system, the resistor the limiting charging current, and the discharge switch determines the repetition rate (Figure 1-3). A simple PFN (Figure
1-2A) generates exponential decay pulses, whereas a more complex PFN can deliver square wave pulses, bipolar pulses, or instantaneous reversal pulses. A PFN forming square wave pulse requires matching impedance, which can be difficult to realize in practice. An example with three capacitor-inductor units is illustrated in Figure 1-2B
(Barbosa-Cánovas and Altunakar 2006, Zhang and others 1995).
"(! ! !
Charging Discharge Resistor Switch Exponential (Rs) decay
Treatmen t Power Energy Chamber Supply Storage Bipolar Capacitor (CO) A Voltage(kV)
time (!s)
Inductor Square (L) "#! RO LL wave
Treatmen t Chamber Bipolar CO CO CO
B Voltage(kV)
time (!s)
Figure 1-3. Block diagrams of pulse power systems and their pulse shapes.
Adapted from: DeHaan (2007), Barbosa-Cánovas and Altunakar (2006), Toepfl (2006), Heinz and others (2002), Zhang and others (1995).
! ! !
The most important parameter influencing microbial inactivation is electric field strength (E), as pore formation occurs when the threshold value of the membrane potential is surpassed (Toepfl 2006). E can be calculated based on Eq. 1-3 (Ho and
Mittal 2000). The minimum E is influenced by cell size and orientation and limited by the dielectric strength of the material (Heinz and others 2002, Ho and Mittal 2000).
Treatment intensity is influenced by the number of pulses and treatment time, which can be calculated using Eq. 1-4 as in Ho and Mittal (2000) and Walkling-Ribeiro (2008).
8"9, : Eq. 1-3!
;< "=&>? Eq. 1-4
Of the two types of pulse waveforms (exponential decay and square wave) exponential is the least efficient (Figure 1-3) as the exposure time is determined as the time required for a reduction of the electric field to 37% of its initial peak strength (Zhang and others 1995) with an excess of heat dissipated without electroporation (Walkling-
Ribeiro 2008). Bipolar pulses are more effective than monopolar because repeated polarity causes alteration in the direction of charges inside and outside the cell membrane, thus creating a mechanical oscillation and therefore membrane breakdown
(Walkling-Ribeiro 2008). In batch chambers the pulses per volume are better defined than for continuous flow where the average number is considered, but it is also important to consider the chamber geometry as sections where the electric field differs can be found at the treatment zone (Toepfl 2006).
Specific energy input has been suggested as intensity parameter, and the energy dissipated per pulse can be estimated based on voltage and current signals determined
"*! ! ! close to the electrodes, assuming an adiabatic system for all the waveforms as in Eq. 1-
5 (Jaeger and others 2009, Aronsson and others 2005, Picart and others 2002).
@"9&3&;< Eq. 1-5
For exponential decay pulses, specific energy input can be estimated based on the energy stored at the capacitor (Eq. 1-6), taking into consideration that there are electricity losses in the PFN and cables that need to be taken into account (Toepfl 2006,
Ho and Mittal 2000, Zhang and others 1995).
C @AB " ##9 &D(,E( &F&G% Eq. 1-6
For exponential decay and square wave pulses specific energy (wsp) can be calculated based on measured media conductivity and electric field strength (Eq. 1-7).
K # , ( # # ( # (C( @AB " FG% H. I J &8 ; :; Eq. 1-7
Also, the effective energy per pulse (wpl) can be predicted from the temperature increase assuming adiabatic conditions using Eq. 1-8 (El-Hag and Jayaram 2006, Heinz and others 2002, Ho and Mittal 2000):
'J " L@BM &=NLQ O&P&DBN Eq. 1-8!
Treatment chambers have been fabricated with different geometries and can be categorized as depending on the direction of the electric field and the liquid food flow.
The simplest form is the cross-field design (Figure 1-4) where the electric field moves across the electrode gap (H in Figure 1-4) allowing an extension in volume at the width
#+! ! ! of the chamber (W) and length (L) of the electrodes. This design is applied to electroporate cells in the laboratory in batches. In continuous systems liquid food can be treated as they flow through electrodes in a co-axial or co-field design (Van den Bosch
2007). While cross-field chambers provide the most uniform electric field, treatment is reduced in boundary regions, especially in batches without mixing, and affects the microbial inactivation kinetics. In continuous chambers this non-uniformity can be prevented by adding multiple treatment zones (Toepfl 2006).
Based on the principles reviewed above, a PEF processing system can be built by adding other components including a pump, a control and monitoring system (an oscilloscope), a pre-treatment heating/cooling chamber, a post-treatment cooling chamber and a temperature monitoring system (Barbosa-Cánovas and Altunakar 2006,
Heinz and others 2003) as in Figure 1-5.
Several authors have reviewed the application of PEF to milk processing
(Sobrino-L'pez and Martín-Belloso 2009, Sampedro and others 2005). The milk micronutrients that have been studied for their variation after PEF treatment are vitamin content (B1: thiamine, B2: riboflavin, D: cholecalciferol, and E: tocopherol), ascorbic acid, vitamin A, with approximately 90% of the ascorbic acid present in milk retained after PEF and the rest remained unaltered at electric fields between 1.8 and 2.7 kV/mm
(Deeth and others 2007). It has also been observed that the macronutrients, including fat content and protein integrity, are unaffected at field strengths as high as 80 kV/cm
(Deeth and others 2007). The most common physicochemical properties that remain unchanged are pH, and titratable acidity. Color, moisture and particle size are other quality aspects that remain unaltered.
#"! ! !
Batch Continuous Co-linear Cross-field Co-Axial Parallel Plate (Co-Field)
Anode
H + W R + + R
""! L L L
Cathode
Insulator Food Electrode
Figure 1-4. Types of treatment chambers.
Adapted from Mastwiijk and others (2007), van den Bosh (2007), Barbosa-Canovas and Altunakar (2006), and Zhang and others (1995).
! ! !
Control and Monitoring System
High Voltage Pulse Generating System T F T T T In, 1 In, 2 Max Out
Treatmen t Chamber
"#! Product Product Inlet Outlet Pump
Heating/ Cooling Cooling Chamber Chamber (optional)
Figure 1-5. Basic components of a high intensity pulsed electric field treatment system.
Adapted from Barbosa-Cánovas and Altunakar (2006), and Heinz and others (2003).
! ! !
Enzymes are used to monitor milk pasteurization as their sensitivity to heat varies and can be predicted (Griffiths 1986). The efficacy of HTST pasteurization can be checked by the inactivation of alkaline phosphatase and pasteurization at higher temperatures (> 80 °C) by lactoperoxidase (Walstra and others 2006). Because of this, studies have been conducted on alkaline phosphatase with inactivation up to 59% in raw milk at an electric field of 2.2 kV/mm (Castro and others 2001). Other enzymes, for example, plasmin, peroxidase, microbial proteases and lipases, have also been studied
(Elez-Martínez and others 2007, Sampedro and others 2005).
Studies in products prepared with PEF-treated milk like cheese and yogurt have concluded that they were similar to those that had been heat-treated and had a good acceptability (Sobrino-L'pez and Martín-Belloso 2009, Sampedro and others 2005).
The application of PEF in cheese production has been studied by Sepúlveda-Ahumada and others (2000), who PEF-treated milk to manufacture cheddar cheese. They observed that some flavors of the cheese produced from PEF-treated milk were better than cheese made with pasteurized milk at 63°C for 30 minutes. Also, hardness and springiness increased in the cheese made from PEF-treated milk while the other textural attributes (adhesiveness and cohesiveness) remained the same.
The mechanism of cell inactivation by PEF is reviewed in section 1.6, and the mathematical models applied to predict inactivation are reviewed in section 1.7.
#,! ! !
1.3. Combined processes
As encompassed by the hurdle technology concept, the use of an intelligent mix of hurdles can improve the microbial stability, safety, sensory, nutritional quality and economic aspects of foods (Leistner and Gould 2002). While initially hurdle technology was aimed to improve microbial quality and safety of foods, the next step included sensory and nutritional quality in the optimization of processing. As reviewed by Leistner and Gould (2002), new approaches include: sustainable processing, predictive modeling, barriers to food (i.e. packaging), and medical approaches (human natural barriers, and defense mechanisms) to hurdle technology. In developed countries the objective is to minimally process foods, while in developing ones the aim is to keep products stable and tasty without refrigeration (Leistner 2000).
As microorganisms have been subjected to a combination of hurdles that have a different spectrum of antimicrobial action (multitarget preservation), it has been noticed that some combinations can have a synergistic, additive, or antagonistic effect (Lee
2004, Raso and Barbosa-Cánovas 2003). While more than 77 hurdles have been recognized for foods and can be classified as physical, physicochemical, microbiologically and miscellaneous hurdles (Leistner 2002, Leistner 1995), the use of combinations is limited by the desired attributes of the final product.
1.3.1. Advanced combined processes
Successful combinations of physical hurdles that have been patented are mano- thermo-sonication (MTS), the bactocatch system, and high-pressure thermal sterilization.
#%! ! !
MTS was patented by the University of Zaragoza as a procedure for the destruction of microorganisms and enzymes under controlled conditions of pressure, heat and ultrasound. The process is continuous and the temperature can be increased to nearly 105°C to inactivate the most resistant mi croorganisms (Ordoñez Pereda and others 1994).
The TetraPak patented Bactocatch process, consists of removing the cream from the milk by centrifugation to then be passed through a 1.4 !m pore size membrane where 99.99% bacterial reduction can be achieved at 50°C with UTMP of 50 kPa and a cross flow velocity of 7.2 m/s, while the retentate and cream go to a UHT treatment
(115°C to 130°C for 4 to 6 seconds) and then are co mbined with the nearly bacteria-free skim milk (Marcelo and Rizvi 2009). Also, the CFMF process has advanced in parallel with the progress in membrane science. In addition to cross-flow microfiltration, dynamic filtration has been also proposed to be incorporated in the processing sequence (Holm and others 1989, Degen and others 1995).
A method has been described to sterilize low acid foods by combining ultra-high pressure and pre-heating to raise the temperature sterilization conditions, to take advantage of the adiabatic temperature resulting from hydrostatic pressure. High pressure-high temperature (HPHT) treatments include pressures up to 600 MPa and temperatures between 90 and 130°C, while in pressur e assisted thermal sterilization
(PATS) the pressure can be raised to 1000 MPa (Wilmarante and Farid 2008, Barbosa-
Cánovas and Juliano 2008, Farid 2009, Wilson and Baker 2000). Continuous HHP systems, where the food is moved through an open-end tube while a high pressure
#&! ! ! pump maintains the pressure difference at 100 MPa or more, have also been combined with PEF (Van den Berg and others 2001b).
Similar to HHP several studies have demonstrated that the lethality of PEF can be enhanced with increases in temperature (Jaeger and others 2009, Amiali and others
2007, Alvarez and Heinz 2007, Bazhal and others 2006, Heinz and others 2003, Smith and others 2002, Reina and others 1998, Jayaram and others 1991). Therefore, a treatment called high electric field short time (HEST) has been proposed by Sampedro and others (2007) that incorporates PEF in the lines of HTST treatment. The idea is to reduce the consumption of PEF energy to warm up the product initially. Advanced systems that are based on this approach include that used by Sepúlveda and others
(2005) that recovers the heat produced in the treatment chamber from the outgoing milk stream, and incorporate it into the incoming milk stream by using a tube-in-tube heat exchanger to increase the temperature to around 50 °C (for a maximum temperature of
65 °C for 15 seconds). This factor is also consider ed when analyzing the economic aspects of scale-up (Hoogland and de Haan 2007). The drawback of this proposition has been that the temperature may rise to high levels due to the ohmic heating resulting from the PEF, and it could potentially damage the product (Alvarez and Heinz 2007), therefore the product needs to be cooled quickly and the conditions properly controlled.
From a technical perspective PEF is one of the most flexible technologies to be combined with other hurdles. It has been successfully combined with ultrasound, thermo-sonication and hydrostatic high pressure (Ross and others 2003, Raso and
Barbosa-Cánovas 2003). Its combination with other technologies to enhance processing performance has also been explored, as in electrically enhanced microfiltration, which
#(! ! ! consists of applying an electric field to the membrane with the purpose of preventing fouling. The process has been called cross-flow electro-filtration (Wakeman 1998,
Visvanathan and Ben Aim 1989).
1.4. Microbial stress
As hurdle technology seeks to achieve microbial inactivation based on the concepts of metabolic exhaustion (as microorganisms use their repair mechanisms to overcome their hostile environment they spend energy and die), and multi-target preservation (the perturbation of different targets within the microbial cells in one hit), the risk of increased microbial resistance while applying minimal processing exists
(Leistner 2000).
As bacteria respond to stresses they may become more resistant or even more virulent, thus the need to understand these microbial responses. As the meaning of the word stress varies depending on the context, in the study of biology it is applied to the imposition of detrimental nutritional conditions, toxic chemicals, or physical conditions
(Yousef and Courtney 2003). The common stresses that microorganisms encounter in the food-processing environment include various chemicals (acid and alkaline treatments, ozone, phosphate and others chemicals), sanitizers (chlorine and chlorinated compounds, non-chlorinated compounds), metal ions, antibiotics, and others like starvation and osmotic and oxidative stress (Ravishankar and Juneja 2003). Other hurdles that have been analyzed because of their potential induction of stress
#)! ! ! resistance include freezing, dehydration, HHP, PEF, and irradiation (Juneja and Novak
2003).
1.4.1. Microbial stress responses
Because of the potential risk of unfavorable microbial stress response, studies have focused on how microorganisms sense and respond to mild stresses, especially pathogenic microbes like E. coli and L. monocytogenes, which may respond with increased virulence (Hill and others 2002).
At the molecular level, microorganisms like E. coli have developed transduction systems to sense environmental stresses and to control the expression of genes involved in defense mechanisms. By using these adaptive networks microorganisms modify their environments and/or survive the stress condition (Chung and others 2006).
A common regulatory mechanism involves the modification of sigma (() factors, whose primary role is to bind to core RNA polymerase granting promoter specificity, and they are the most extensively studied because of their various functions. They fall into a number of categories including: the main housekeeping ( factor (responsible for the transcription of the majority of the promoters), and alternate ( factors that have different specificities like the expression of regulons involved in heat-shock response, chemotactic response, sporulation, and general stress response (Abee and Wouters
1999). In E. coli, as with other other enteric bacteria, (S (RpoS) is the master regulator of the general stress response (Chung and others 2006).
Based on the growing knowledge of microbial stress response regulation processes, several approaches have been taken to monitor changes at the molecular
#*! ! ! level including the detection of stress response genes (using molecular probes), mRNA analysis (Northern analysis, microarray-genome-wide expression or reverse transcription polymerase chain reaction), the detection of stress proteins (using gel electrophoresis or antibodies), or biosensors (Yousef and Courtney 2003). Quantitative methods to monitor changes in relative stress resistance will be discussed below.
Heat shock proteins (HSP) can be monitored in real time by constructing biosensors that report the induction of genes related to their synthesis (Zhang and
Griffiths 2003). Microbial biosensors are basically analytical devices that combine a biological sensing element with a transducer (energy converter) to produce a signal proportional to the concentration of the chemical of interest. For that purpose microorganisms have been integrated with a variety of transducers like: amperometric, potentiometric, calorimetric, conductimetric, colorimetric, luminescence and fluorescence (Lei and others 2006). While stress response can be monitored quantitatively by measuring the stress conditions, and variation in inactivation trends based on the population growth, other methods like biosensors can contribute to the development of mechanistic models where the stress factor can be considered
(Gawande and Griffiths 2005ab). In addition to monitoring stress response the application of biosensors vary widely, from understanding mechanism of action to internalization studies (Jablasone and others 2005, Kumar and others 2006).
Quantitatively it has been demonstrated that exposure to sub-lethal levels of a stress causes an increased resistance to levels of the same stress that were lethal under normal conditions. For example, Farber and Brown (1990) heat shocked L. monocytogenes at 40, 44, 48 and 52 °C observing that prior stres s for up to 120 min at
$+! ! !
48 °C was needed to double or triple the D64°C values. Murano and Pierson (1992) heat- shocked E. coli O157:H7 (32 °C to 45 °C for up to 15 minutes) caus ing resistance to a higher temperature treatment (55 °C). Juneja and ot hers (1998) observed a lag period
(increase of 4 to 6 min) in the inactivation of E. coli O157:H7 at 60 °C when it was previously exposed to heat shock (46 °C) in a beef gravy system.
1.4.2. Cross protection
The concept of stress adaptation resulting from increased resistance due to non- related stresses (heterologous) is called cross-protection (Ravishankar and Juneja
2003). The term was initially proposed by Lou and Yousef (1996), who observed an increase in heat resistance (D56°C ) in L. monocytogenes at several growth phases, following exposure to starvation, acid treatment, hydrogen peroxide (H2O2), and ethanol.
Subsequent studies such as the work of Ryu and Beuchat (1998) showed that acid adaptation and shock (pH’s down to 3.8) induced resistance to heat treatments at
52 °C to 54 °C. Zhang and Griffiths (2003) observed that starvation (37 °C to 5 °C) of E. coli O157:H7 induced heat resistance at 52 °C and Gawande and Griffiths (2005b) observed increased cryotolerance (-18 °C) due to st arvation stress. Depending on the treatment conditions and hurdle applied, injury may result instead of protection; for example, García and others (2005) observed an initial resistance due to acid shock (pH
4.0 in citrate buffer) and an increase in inactivation by PEF (after 2 h at pH 4.0) in gram negative microorganisms. For a complete review see Rodríguez-Romo and Yousef
(2005).
$"! ! !
1.4.3. Risks associated to PEF
Little research has analyzed the risks of adaptation responses due to PEF. An unusual risk related to the application of PEF is the possibility of transferring extracellular DNA material that, once incorporated in more resistant microorganisms, may confer pathogenic traits. In an assessment of pulsing non-electrocompetent and electrocompetent Lactobacillus casei in buffer containing the plasmid pIA)8 that carries chloramphenicol resistance, it was found that there was no transfer between cells that were not pre-conditioned for electroporation, therefore the risk of cross-transformation may be low (Rodrigo and others 2007).
1.5. Microbial physiology and inactivation
As mentioned before, the concept of stress is broad, and includes moderate stresses, extreme stresses, and severe stresses that may cause injury or death (Yousef and Courtney 2003). The study of injury leads to a review of the concepts of viability and culturability, as microorganisms can be non-viable (but virulent) for a period of time and then recover.
Traditional methods to count microbes are based on cultivating microbes, which is limited to the ability of microbes to reproduce. Therefore, the terminology surrounding viability varies and has been reviewed by Barer and Hardwood (1999). The application of novel methods to assess physiological states that may not be observed by conventional techniques represents progress in the microbial analysis field.
$#! ! !
Liquid media continue to be developed as part of injury recovery protocols, and highly selective enrichment methods continue to advance. For example, in the case of
L. monocytogenes a broth named Pennsylvania State University broth has been optimized, and applied successfully in the recovery of HHP injured L. monocytogenes
(Bull and others 2005). The most probable number technique (MPN), based on the statistical analysis of the probabilities of microbial growth in liquid media placed in multiple tubes, increases the sensitivity of the estimations to around < 5CFU/g
(Davidson and others 2004), and has been applied to increase the spread plate detection below 25 CFU/ml when enumerating E. coli O157:H7 after pulsing apple juice and apple cider with light (PL) (Sauer and Moraru 2009). MPN has been applied in combination with the pour plate technique by Elwell and Barbano (2006) to analyze the low total counts in CFMF processed milk. One weakness of this method is that it requires large volumes of broth (100 to 500 ml) in order to increase the detection limit.
By using automated methods like the Bioscreen, which measures optical density
(OD) continuously in 200 wells, the recovery of populations in liquid media has also been demonstrated in work such as that of Stephens and others (1997) who generated growth curves of heat injured Salmonella, and applied the MPN technique to estimate their inoculum level. The MPN method was also automated by placing the dilutions in
96-well microtitre plates (200 !L/well). An advanced version of the 3-tube or 5-tube
MPN method is the completely mechanized, automated and hands-off TEMPO from bioMerieux in which the operator only needs to make a 1:10 dilution of the food or water, which, after connection to the 48 chamber card, is incubated overnight and a
$$! ! ! fluorescent signal proportional to the number of cells present is automatically read
(Fung 2009).
Modifications to solid media techniques include the thin agar layer method (TAL) evaluated by Kang and Fung (1999) in L. monocytogenes and by Wu and Fung (2001) in E. coli O157:H7, L. monocytogenes, S. Typhimurium, and S. aureus, to recover heat injured cells, consists of recovering the injured microbial population in non-selective media temporarily and adding selective media to recover the target population.
García and others (2006) monitored injury in E. coli after applying PEF by adding sodium chloride (4%) to Nutrient Agar, or incubating in peptone water containing sodium azide (500 !g/ml) chloramphenicol (100 !g/ml), rifampicin (10 !g/ml), cerulenin (70
!g/ml) or penicillin G (100 !g/ml) for 2 hours prior to plating on Nutrient Agar.
Cytometry is a process for measuring the physical and chemical characteristics of biological cells. Cytometry principles are applied in the dairy industry in the form of direct microscopic count, direct epifluorescent filter technique, and the Bactoscan technique (Suhren and others 1990). The most highly developed technique is flow cytometry, which is composed of a light source, a flow chamber, an optical system, light detectors and a computer system (Ormerod 2008 and Shapiro 2003).
While flow cytometry is a standard method to assess milk quality in developed countries, several studies related to dairy products have been conducted based on its principles ranging from the detection of L. monocytogenes by means of immunofluorescence (Donnelly and others 1986, Donnelly and others 1988), monitoring the injury caused by bacteriocins in L. monocytogenes (Swarts and others 1998), enumeration of total bacteria (Gunasekera and others 2000), gene expression of non-
$,! ! ! culturable E. coli and Pseudomonas putida after pasteurization (Gunasekera and others
2002), the application of fluorescence-in-situ-hybridization to detect E. coli (Gunasekera
2003), detection of respiring E. coli O157:H7 (Yamaguchi and others 2003), gram differentiation of a mixture of microorganisms (Holm and Jespersen 2003), and to characterize the microflora of milk (Holm and others 2004ab).
Milk is difficult to study because of its high content of macromolecules (especially fat and proteins) that make it opaque. For that reason, clarification protocols, including savinase (EC 3.4.21.52), and Triton X-100 to remove the fat, and the enzymes proteinase K (EC 3.4.21.52), and subtilisin (EC 232-752-2) to digest the protein
(Gunasekera 2000, Yamaguchi and others 2003) have been employed.
An advantage of flow cytometry over related methods is the ability to sort target populations for further analysis, the physiological states of these clusters can be characterized in medium as shown by Nebe-von Caron and others (1998). The physiological stages that can be monitored by this technique range from the reproduction of the cells (cell division), metabolic activity (enzyme activity, esterase, dehydrogenase, membrane potential and pump activity), membrane integrity (selective membrane permeability, exclusion membrane), and membrane permeability in dead cells (Nebe-von Caron and others 1998).
The principle of using fluorescent dyes with the ability to permeate membranes has been applied to study the mechanism of action of novel processes such as HHP.
For example, Ulmer and others (2000) studied cell viability and sublethal injury by measuring the uptake of the membrane-damaged permeable dye propidium iodide (PI) using a spectrafluorimetric microtiter plate reader after HHP treatment of Lactobacillus
$%! ! ! plantarum in model beer. In a subsequent study, the permeability of L. plantarum in model beer after PEF was determined (Ulmer and others 2002). This application has been extended to the measurement of membrane damage and viability with the nucleotide-binding dyes PI and SYTO 9 by PEF in Lactobacillus leichmannii, L. monocytogenes, and E. coli O157:H7 using a spectrofluorimeter combined with fluorescence microscopy (Unal and others 2002). Flow cytometry and PI permeability has been applied by Wouters and others (2001b) to study membrane damage of exponential and stationary phase cells of L. plantarum. Recently Aronsson and others
(2005), studied membrane permeability by PI, and cell content leakage (ATP) by means of luminescence. García and others (2007) utilized a spectrofluorophotometer to monitor PI intake by two gram negative (E. coli and Salmonella Senftenberg) and two gram positive (L. plantarum and L. monocytogenes) bacteria and related it to the intensity of PEF treatment. As stated by Saulis and others (2007), advancement in these methodologies is leading to a better understanding of the detailed molecular membrane processes of electroporation occurring in the cells that will result in more robust kinetic models to predict these mechanisms.
1.5.1. Mechanism of action
It is known that electric pulses can produce micropores in cell membranes in an effect called electroporation. At non-lethal intensity levels the phenomenon is called reversible electrical breakdown as cells can reseal their pores. Reversible electrical breakdown is of importance to genetic engineering and cell hybridization (Jayaram
2000).
$&! ! !
There are two schools of thought regarding the electrical membrane breakdown phenomenon. In the classical electromechanical models the membrane is considered as a homogeneous bulk elastic layer, which under an external electrical field is stressed by a compressive force that creates a restoring electric strain. When a critical trans- membrane voltage at which electric compression exceeds the electric strain is reached rupture occurs. These models do not consider the proteins between the lipid bilayer or ionic channels that cross the membrane. From a chemical perspective membrane rupture is due to an imbalance in molecular/ionic concentration resulting from charge transport across the membrane where the membrane rupture occurs as a secondary process. With the increase of trans-membrane voltage, temporary aqueous pathways across the membrane are believed to appear (Jayaram 2000).
Electroporation can be divided into three main stages: 1) initial stage (within ns to ms): with the creation of pores when an electric pulse is applied, 2) evolution of the pore population (ns to ms): change in number of pores and their sizes during an electric treatment, 3) post-treatment stage (ms to h): cell death or return to its initial viable state due to pore resealing. Theoretical models to explain the mechanisms of pore formation have been developed (Saulis and Wouters 2007).
1.6. Advances in microbial modeling
In terms of food safety, engineering integrates microbial modeling with process modeling. While modeling the physics of a process can be relatively simple, as it is governed by physical and chemical laws, modeling a biological system can represent
$(! ! ! special challenges. It is important to emphasize that the ultimate purpose of processing models is to optimize system performance; microbial models are developed to ensure that a microbial pathogen population does not exceed a certain level (Marks 2008).
Microbial physiology, functional genomics, molecular biology and mathematical modeling can also help to identify molecule types of living systems in terms of their individual properties, while the emerging field of systems biology can help us to understand the functional properties of the system as a whole. Although advances in some areas, like comparative genomics, can help to define the genes and molecules that help microorganisms to survive environmental stresses, the progress in single-cell analysis techniques (i.e. flow cytometry) help to correctly analyze genetic and physiological characteristics that can be related to the whole population (Brul and
Westerhoff 2007).
A model is a combination of descriptions, mathematical functions or equations, and specific starting conditions, and can be in either of two classes: descriptive or explanatory. Descriptive models (black box, observational, empirical or inductive) are data-driven, and approaches like polynomial functions, artificial neural nets and principal component analysis are used to classify the data. Explanatory models (white box, mechanistic or deductive) relate the data to fundamental scientific principles, or at least to measurable physiological processes (McKellar and Lu 2004a).
Microbial growth or survival can be modeled using two basic approaches, by considering the effect at the single cell level (stochastic or vitalistic) or at the population level (deterministic or mechanistic). The vitalistic approach is based on the assumption that individuals in a population are not identical, while the central dogma of the
$)! ! ! mechanistic approach is that the destruction “is an orderly time process presenting a close analogy to a chemical reaction” (Mathys 2008, Baranyi and Pin 2004).
Researchers are constantly trying to unify models to better describe microbial events with smooth transitions between growth and inactivation. Therefore advanced models based on population dynamics considering events at the cellular level continuos to be studied and developed (Peleg 2006, Corradini and others 2010).
Traditionally investigators working on mechanistic models rely on the fact that these models are capable of accurately predicting the microbial inactivation response to thermal treatment conditions outside the range in which experimental data is obtained.
Process engineers have designed ultra-high temperature (UHT) or high temperature short-time (HTST) and pasteurization processes that operate at higher temperatures based on this extrapolation capability (Teixeira 2007).
Microbial models can be also classified as: primary models that describe how the number of microorganisms in a population changes with time under specific conditions, secondary models that relate the primary model parameters to environmental or intrinsic variables, or tertiary that combine primary and secondary models with a computer interface providing a complete prediction tool (Marks 2008).
As reviewed by Griffiths (1994), predictive microbiology in the dairy industry has been mainly applied in the estimation of microbial growth during storage (growth in raw and pasteurized milk) and thermal processing (inactivation of microorganisms and enzymes). In milk pasteurization the areas of focus include thermal inactivation models for key enzymes and pathogens, and whole line risk assessment (McKellar 2003).
$*! ! !
1.6.1. Modeling growth
Microorganisms multiply when exposed to favorable environments. Vegetative bacterial cells, yeasts and molds reproduce in a process called transverse binary fission, or simply binary fission. In this process a cell divides asexually into two cells with each cell becoming a replica of the original cell (Ray 2004).
After inoculation in fresh medium a population remains temporarily with no division, but may be growing in volume or mass, synthesizing enzymes, proteins and increasing metabolic activity in what is called the lag phase. In the laboratory a growing bacterial population doubles at regular intervals by geometric progression: 20, 21, 22,
23… 2n, where n = the number of generations in what is called exponential growth. The generation time is the time interval required for the cells to divide, and can be expressed mathematically as in Eq. 1-9 (Todar 2008).
R";, = Eq. 1-9
The number of bacteria at the end of the time interval, and can be predicted using Eq. 1-10.
S"T&EU Eq. 1-10
In the context of hurdle technology growth/no growth (G/NG) or boundary models are useful to determine the intrinsic and extrinsic environmental factors of the food product that determine the growth during distribution, storage or consumer use or abuse
(Leistner 2002). G/NG models continuously advance in parallel to new food preservation techniques and the understanding of physiological processes (McMeekin
,+! ! ! and others 2000) with the goal to determine the borderline conditions that prevent growth of a given microorganism without inactivation intervention. Automated equipment that detects growth allows quantification once a threshold value of the measured response is clearly greater than the noise seen over time at no growth conditions
(Legan and others 2002).
As mentioned above (section 1.5), media for assessing microbial recovery can be both liquid and solid, with solid media preferred to enumerate individual cells based on the formation of colonies (FIL-IDF 1990). Because the construction of models using viable count data is time consuming and expensive, researchers are constantly exploring rapid methods to accumulate sufficient data for modeling. One of the simplest methods to monitor growth is optical density with instruments like the Bioscreen
(McKellar and Knight 2000). Turbidimetric data can also be collected using microplates, or simple spectrophotometry in tubes or cuvettes (Dalgaard and Koutsoumanis 2001).
Thus, there has been a trend to model growth based on turbidimetric measurements
(Dalgaard and others 1994). Several models have been evaluated to determine growth rates and lag times including: the non-transformed four-parameter logistic model, the
Richards model, the modified Gompertz model and the exponential model (Dalgaard and Koutsoumanis 2001).
Because of the availability of computational tools like the Microsoft® Excel add- on DMFit (Institute of Food Research, Norwich, UK), the preferred method to analyze growth curves is the one of Baranyi and others (1993). The model is continuously being developed by including physiological parameters (Baranyi and Pin 1994), and
,"! ! ! comparing with other models (Baranyi and others 1999) based on their bias and accuracy. The model in its simplest form can be expressed by Eq. 1-11.
#XYZ[&<( V#;( "V. &W Eq. 1-11
Other primary models that have been used to describe microbial growth in time include the Logistic, Gompertz, Weibull, Hills and McKellar models (López and others
2004, McKellar and Lu 2004b), and are reviewed in Chapter 4.
Growth curve parameters like lag phase or time to detection (td), and maximum growth rate (!max) considering pH, and water activity (aw), have been determined for E. coli after PEF treatment by Aronsson and others (2004). The secondary models applied to describe the variation of these parameters with respect to environmental conditions
(i.e. temperature, pH, water activity, and other factors like % CO2) include the following types: square-root, gamma, cardinal, Arrhenius type, polynomial and artificial networks
(Ross and Dalgaard 2004). As they are applied to experimental data the models are continuously adapted to better describe microbial growth conditions.
1.6.2. Modeling inactivation
As research on microbial inactivation progresses it is constantly reported that when analyzing microbial populations that are subjected to a lethal treatment for a period of time a variety of patterns result from the relation between the surviving microbial populations and time. These patterns can be described as follows: tailing pattern, increase in population due to spore activation, a shoulder or lag prior to inactivation, or a sigmoidal pattern with both a lag and a tail (Marks 2008). Geeraerd
,#! ! ! and others (2005) categorized the inactivation curves as log-linear, log-linear with shoulder and/or tailing, Weibull type, and biphasic models to present the Microsoft®
Excel add-in GInaFit (Bioprocess Technology and Control, Leuven Katholieke
Universiteit. Leuven, Belgium) to predict inactivation curves. An illustration of the common types is depicted in Fig. 6.
In order to explain the kinetics of microbial inactivation by PEF it is important to consider the factors affecting the mechanism of action. As reviewed before, it is generally known that microbial membranes that electrically insulate an electrical field produce an accumulation of charges with opposite polarity on either side of the structure, and pore formation occurs when a certain threshold value of electric field (E) is exceeded (Heinz and others 2002). A trans-membrane potential (TMP) of at least 1 V
(E = 2.5 kV/mm) is needed to achieve irreversible pore formation (Ho and Mittal 2000), and can be explained by solving Maxwell’s equation in spherical coordinates assuming simplifying restrictions as illustrated in Eq. 1-12 (Heinz and others 2002, Ho and Mittal
2000, Fox 2007).
\]^ " _ & ` & ab &acdef!!! ! ! ! ! ! ! Eq. 1-12 !
From this knowledge it can be deduced that E and treatment time are important factors to consider when describing microbial inactivation kinetics. By applying simple kinetics the Bigelow model (Eq. 1-13) has been used to describe thermal inactivation
(Fernández and others 1999). ghi 6 " / ;, j Eq. 1-13
,$! ! !
# $ (N) 10 (N) 10 Log Log ""!
time time
linear sigmoidal biphasic biphasic with shoulder convex linear with tailing concave linear with shoulder
Figure 1-6. Common types of microbial inactivation curves.
Adapted from Marks (2008), Geeraerd and others (2005).
! ! !
The first attempts to describe the kinetics of inactivation by PEF were made by
H*lsheger and others (1981). They described PEF inactivation separately based on electric field value (Eq. 1-14), and treatment time (Eq. 1-15), and combined both parameters in a simplified model (Eq. 1-16).
kl m " /no &kl#`/`p.( Eq. 1-14
kl m " /nq &kl#rr, p.( Eq. 1-15
t#otouv(,w m"#rsrp.( E q . 1 - 1 6
Peleg (1995) described the inactivation based on Fermi’s equation as described in Eqs. 1-17 and 1-18.
# ( t#wy&z( e`p. l "`p.v &x Eq. 1-17
t#w|&z( {#l( "{. &x Eq. 1-18
The Weibull distribution (Eq. 1-19), which is a two parameter model based in a distribution of probabilities, has been further developed to describe microbial inactivation kinetics (Fernández and others 1999, Peleg and Cole 2000, Couvert and others 2005). kl m#r( "/#r}, (~ Eq. 1-19
With the scale and shape parameters, a density function can be generated using
Eq. 1-20, and a mean time value (tc) can be calculated with Eq. 1-21. The mean time
,%! ! ! value is an indicator of microbial resistance and can be used to compare microorganisms or processing conditions.
Ö Ü F#;( "tÄ &Å&#;ÇtÉ( &ÑWLt#< ( Ná Eq. 1-20
tÉ rà " } &â#ä2ã ( Eq. 1-21
The variations in the scale (b) and shape parameters (n) of the Weibull model have been further developed to include food composition. For example, Rodrigo and others (2003) developed a model to quantify the effect of carrot juice addition to orange juice on the inactivation of E. coli. In a later study by García and others (2005a) the n values were related to E and pH, and the variation was explained by a fitting of the
Gompertz equation.
As the b value represents the time necessary to inactivate 0.434 log10 cycles and the n parameter accounts for concavity of the survival curve, Alvarez and others (2003) suggested that a zPEF value can be calculated from the relation between the b values
(from the primary modeling of inactivation with time) and E. In their study, they found no influence from E and n. The zPEF value has been used to compare Salmonella
Enteritidis (Alvarez and others 2003) and Staphylococcus aureus (Rodríguez-Calleja and others 2006) strains.
Assuming a symmetric curve the log-logistic model (Eq. 1-22) described by Cole and others (1993), has been applied by Raso and others (2000) to predict the inactivation of S. Senftenberg by PEF.
/kcå6"Lç2#é/ç(NLä2WQ #è&ê&#ëtíìî <(#,(ïtñ( N Eq. 1-22
,&! ! !
These models (Bigelow, H*lsheger, Gompertz, Peleg and Weibull), have been compared by various authors (Rodrigo and others 2001, Sampedro and others 2006) with the Weibull having a better accuracy factor in the first study, and the H*lsheger, and Weibull having lower Mean Square Errors (MSE) than the Bigelow in the second one.
Because of the integrative characteristic of the PEF energy parameter, a trend to relate microbial inactivation to the PEF energy applied was initiated by Heinz and others
1999, who related field strength settings and pulsing time based on energy levels and concluded that a minimum of 0.4 kJ/kg per pulse were needed to inactivate Bacillus subtilis. Subsequent studies like that of Alvarez and others (2003) with L. monocytogenes, substitute treatment time by specific energy, which was then included in the Weibull distribution model. In a related study, the energy efficiency of apple juice pasteurization by pulsed electric fields and temperature was estimated by Heinz and others (2003). By studying the relation between microbial inactivation and applied energy to the treatment temperature and further modeling the inactivation rate and temperature, a log-linear function was generated (Eqs. 1-23, and 1-24).
ó#8òJôU( "
| ù | WLöyõ#ö|&ú(õ#öù&íz#ûü†((NLQ Éõ#ö°&ú(õLö¢&ú NõLö£&ú Nõ#ö§&íz#ûü†((õ#ö•&íz#ûü†(( N Eq. 1-23
kcå#¶¶, .( "/ß#`ò \®z( &©™´ Eq. 1-24
When evaluating models, the accuracy of the predictions can be determined by considering the accuracy factor as in Eq. 1-25 (Fernández and others 1999).
,(! ! !
Ø∞±≤≥¥µ,π ∂∑∏ 0Ñ á ¨≠ "äÆ † Eq. 1-25
Secondary models include the surface response design applied by Sobrino-
López and Martín-Belloso (2008), where the influence of nisin dose, lysozyme dose,
PEF treatment time, and pH on the inactivation of Staphylococcus aureus were modeled.
1.7. Research needs
As PEF is a flexible technology that can be combined with other existing processes (like microfiltration), promising combinations can be explored for further development.
As milk processing has adopted various strategies for the control of microorganisms from farm to table, the exploration of the inactivation of various microbial groups could guide dairy processors in the alternative use of the PEF technology.
The quantitative description of changes in microbial populations at the molecular level by applying the principles of spectrophotometry, and fluorescence detection of biosensors or permeable dyes can contribute to the safe design of alternative microbial processes that include the risk of protective effects of the growing in extreme environments on microbial resistance.
,)! ! !
2. Factors affecting the inactivation of native bacteria in milk processed by pulsed electric fields and cross-flow microfiltration
2.1. Introduction
To control of microorganisms naturally present in milk the dairy industry depends on heat treatment and cold preservation. Standards like the Pasteurized Milk Ordinance serve as the reference in which time-temperature combinations for pasteurization of products with different composition is described together with other guidelines (FDA
2009). It is generally recognized that heat pasteurization does not ensure commercial sterility (NACMCF 2006) and an excessive heat treatment can be associated with other health risks (e.g. formation of acrylamide, heterocyclic aromatic amines, or furans) resulting from the formation of toxic substances (DFG 2007). Ross and others (2003) and Raso and Barbosa-Cánovas (2003) reviewed the combination of existing non- thermal technologies to enhance microbial inactivation in food without damaging its quality. Among these technologies, pulsed electric fields (PEF) have been used as an alternative to thermal treatment in order to reduce microbial loads in milk (Michalac and others 2003, Fernández-Molina and others 2005, Barbosa-Cánovas and Bermúdez-
Aguirre 2010). In addition to its ability to inactivate microorganisms, Brans and others
(2004) reported that high intensity PEF treatment has also been applied to prevent migration of charged components to the membrane surface during cross-flow microfiltration (CFMF).
Because PEF is effective when applied with moderate temperatures (Toepfl and others 2006, Walkling-Ribeiro and others 2010), a process called high electric field short time (HEST) treatment was proposed by Sampedro and others (2007), and consists in
,*! ! ! the application of PEF in conjunction with high temperature short time (HTST) treatment as a pre-heating step. Recent studies have explored hurdle processing of milk with PEF and other processing technologies, for example, Walkling-Ribeiro and others (2009) combined PEF with heat as a subsequent process, Noci and others (2009) combined
PEF with thermosonication, and Smith and others (2002) combined PEF with moderate heat and antimicrobial agents, such as nisin and lysozyme.
Due to the variety of processing and product parameters which affect the application of PEF, reviews by Sampedro and others (2005), and Sobrino-López and
Martín-Belloso (2009) suggested that different approaches are necessary to analyze the effect of milk nutrients on the effectiveness of pulsed electric fields for milk processing and preservation. Processing conditions (i.e. electric field strength, pulse number, pulse width, treatment time and processing temperatures), microbial analysis (i.e. microorganisms and methodology), and the complexity of the food system (i.e. characteristics of milk products) were identified as main factors determining the overall efficacy of PEF treatment of milk.
Cross flow of skim milk through a filter with a pore size ranging between 0.8 and
1.4 !m under uniform trans-membrane pressure is the basis for CFMF (Gésan-Guiziou,
2010). Up to 99.7% of the bacteria can be reduced in skim milk with a CFMF system
(Pedersen 1992). Further development led to the Bactocatch system, consisting of a centrifugal unit for cream separation, a microfiltration unit to separate vegetative cells and spores, and an ultra-high-temperature component for the processing of the cream- retentate mixture (Kessler 2002).
%+! ! !
The objective of this study was the evaluation of the impact of a PEF system at different steps of milk processing for the reduction of the microbial load in milk with different fat levels. Moreover, the impact of CFMF applied prior to PEF treatment, and the effects of cream homogenization were investigated as well as the correlation between the fat and solids content and electrical conductivity at the different stages of milk processing.
2.2. Materials and Methods
2.2.1. Conventional and alternative milk processing
The sequences of conventional and alternative processing of milk used are shown in Figure 2-1. Conventional processing of milk was carried out with raw milk obtained from the Elora-Posonby Dairy Research Station, University of Guelph and stored at 4°C. Raw milk was pre-heated in a pilot s cale, dual stage heat exchanger unit to 50°C (UHT/HTSTLab-25EDH, Micro-Thermics, Raleigh , NC, USA) prior to separation
(Westfalia LWA 205, Centrico Inc., San Francisco, CA, USA) of skim milk and cream at flow rates of 180 and 18 L/h, respectively. Subsequently, the milk was standardized by mixing the appropriate volumes of cream and skim milk to three final fat contents (1.1%,
2.0%, 3.1%). Homogenization (NS2006H, GEA NiroSoavi S.p.A, Hudson, WI, USA) was conducted at two pressure stages of 20 MPa (first stage) and 5 MPa (second stage), a processing frequency of 15 Hz and initial flow rates of 120 L/h inlet, and 60 L/h outlet in order to achieve a minimum stuffing pressure of 0.4 MPa. The homogenization process was combined with the heat exchanger, which was also used at this processing stage to
%"! ! ! pre-heat milk, which exited the homogenizer at 12°C , to 50°C. By contrast, in an alternative processing sequence separated skim milk was passed through a CFMF pilot plant system (MFS-1, TetraPak, Aarhus, Denmark) with a membrane pore size of 1.4
!m at temperatures of 35°C with respective inlet and outlet pressures of 90 (at 120 L/h), and 20 kPa (at 12 L/h). The skim milk CFMF retentate was transferred in aseptic containers for PEF processing. In the other stream, standardized cream (12%) was exposed to a pasteurization treatment with PEF after homogenization under the same conditions described above. Both processing approaches were concluded with PEF or
HTST milk pasteurization under the processing conditions described below.
2.2.2. PEF treatment
For non-thermal processing of standardized milk and cream a PEF system was used.
Electric exponential decay pulses with an average pulsed width of 1.5 !s were formed with a generator unit (PPS 30, University of Waterloo, Waterloo, ON, Canada). PEF was applied in a co-axial treatment chamber consisting of one vertically arranged central, rod-shaped stainless steel grounded electrode, which passed into the chamber stand at the bottom, and three ring-shaped stainless steel charging electrodes centered around the grounded electrode and mounted in circular ring-shaped chamber metal segments.
Each of the electrode containing metal segments was, in turn, threaded into surrounding plastic segments on the top and bottom functioning as insulating spacers. The electrode gap between the charged and grounded electrodes of the PEF chamber was 0.25 cm and the total chamber volume and the effective electrode treatment area amounted to
600 ml and 147 cm2, respectively.
%#! ! !
Raw Raw
Skim SEPT Cream Skim SEPT Cream
STND CFMF Skim STND HOMO Skim Cream Upstream
Downstream
"#! PAST
Milk STND Cream HOMO
PAST PAST
Milk Milk ConvetionalProcessing Milk Processing Milk Alternative
Figure 2-1. Milk process sequence in conventional and alternative fluid milk processing
Products (ovals) and processes (rectangles) compared in this study are highlighted in gray. SEPT = Separation, STND = Standardization, HOMO:
Homogenization, PAST = Pasteurization, CFMF = Micro-filtration. Adapted from Pedersen 1992, and Walstra and others 2006.
! ! !
Table 2-1. Applied intensitites of high intensity pulsed electric fields and resulting temperature profiles
!
Flow Fre- Treatment Electric Temperature Treatment Energy Rate quency time Field Inlet Outlet 1.2 L/h 25 Hz 1913 -s 3.2 kV/mm 739 kJ/L 14.3 °C 46.0 °C Milks 2.4 L/h 12 Hz 459 -s 4.8 kV/mm 399 kJ/L 13.3 °C 55.1 °C (1.0 to 3.8%) 1.2 L/h 12 Hz 918 -s 4.8 kV/mm 798 kJ/L 17.0 °C 55.1 °C Skim milk 2.4 L/h 20 Hz 765 -s 4.0 kV/mm 462 kJ/L 5.9 °C 43.0 °C Low 1.2 L/h 20 Hz 1530 -s 4.0 kV/mm 924 kJ/L 5.0 °C 47.0 °C temperature 2.4 L/h 12 Hz 459 -s 5.6 kV/mm 543 kJ/L 6.3 °C 56.1 °C Skim milk 4.8 L/h 20 Hz 383 -s 4.0 kV/mm 231 kJ/L 34.3 °C 50.3 °C Mild 3.6 L/h 20 Hz 510 -s 4.0 kV/mm 308 kJ/L 34.0 °C 53.5 °C temperature 4.8 L/h 12 Hz 230 -s 5.6 kV/mm 272 kJ/L 34.3 °C 56.3 °C 1.2 L/h 25 Hz 1913 -s 3.2 kV/mm 739 kJ/L 19.1 °C 42.3 °C Skim milk 2.4 L/h 12 Hz 459 -s 4.8 kV/mm 399 kJ/L 18.9 °C 56.2 °C CFMF + PEF 1.2 L/h 12 Hz 918 -s 4.8 kV/mm 798 kJ/L 20.1 °C 61.2 °C 2.4 L/h 20 Hz 765 -s 4.0 kV/mm 462 kJ/L 5.3 °C 42.3 °C Cream (12.2%) 1.2 L/h 20 Hz 1530 -s 4.0 kV/mm 924 kJ/L 5.5 °C 42.1 °C Standardized 2.4 L/h 12 Hz 459 -s 5.6 kV/mm 543 kJ/L 5.7 °C 50.4 °C 2.4 L/h 20 Hz 765 -s 4.0 kV/mm 462 kJ/L 4.8 °C 30.2 °C Cream (12.3%) 1.2 L/h 20 Hz 1530 -s 4.0 kV/mm 924 kJ/L 5.0 °C 43.5 °C Homogenized 2.4 L/h 12 Hz 459 -s 5.6 kV/mm 543 kJ/L 5.0 °C 50.5 °C
%,! ! !
Milk was exposed to electric field strengths between 3.2 and 5.6 kV/mm, in a pulse frequency range between 12 and 25 Hz, for total PEF treatment times from 230 to
1913 !s, which corresponded to product flow rates between 1.2 and 4.8 L/h (Table 2-1).
Specific energy density (wsp) ranged from 543 to 1848 kJ/L and was calculated based on capacitance, according to the equation by Zhang and others (1995), which was adapted by Walkling-Ribeiro (2009), and is illustrated in Eq. 2-1:
C C @AB " #9 &;< &D(#, E&O∫( " #9 &;<(#, E&O∫ &1∫( Eq. 2-1!
Skim milk was pumped (Masterflex pump drive 7524-40 and pump head 77201-
60, Cole Parmer Instrument Co., Vernon Hills, IL, USA) through silicone tubing to the
PEF chamber. To evaluate the effect of temperature, the product was either pre-cooled for 83 (at 2.4 L/h) or 110 s (at 1.2 L/h) in a stainless steel coil submerged in ice, entering at 14 °C and exiting at 4 °C, prior to entering the PEF chamber at 5 °C, or pre-heated for 165 (3.6 L/h) or 330 s (4.8 L/h), entering the same coil submerged in a water bath set at 37.5 °C (Isotemp 10L, Fisher Scientific, Ham pton, NH, USA) entering at 14 °C and exiting at 34 °C prior to entering the chamber as described by Sampedro and others (2007), and Heinz and others (2003), and illustrated in Table 2-1. The milk temperature at the chamber outlet following PEF processing did not exceed 61 °C and the product was cooled subsequently to 12 °C in a c oil, which was submerged in a refrigerated water bath set at -12 °C (NESLAB RTE-7 , Thermo Scientific, Newington,
NH, USA). Milk samples were stored in a styrofoam container containing ice prior to microbiological and physico-chemical analyses. For temperature measurement, a wireless temperature data logger (OM-SQ2020-2F8, Omega, Stamford, CT, USA) with
%%! ! ! thermocouples was connected to stainless steel flow-through chambers, which, in turn, were connected to the tubing were used. Pulses were monitored with a two-channel digital storage oscilloscope using a bandwidth of 100 MHz and a sample rate of 1 GS/s
(TDS1012B, Tektronix Inc., Beaverton, OR, USA).
2.2.3. Conventional heat treatment of milk
High-temperature short-time (HTST) pasteurization of milk was conducted in the aforementioned pilot scale dual stage heat exchanger using a pre-heating temperature of 50 °C in the first stage of the unit prior to re aching final temperatures of 72, 85, or 95
°C for holding times of 15, 20 or 15 s respectively .
2.2.4. Microbiological Analysis
After the milk was collected from the farm, the mesophilic counts were found to be below 2,500 CFU/ml, coliform counts averaged about 300 CFU/ml and psychrotrophic counts were below 300 CFU/ml. In order to achieve a high bacterial concentration of native microorganisms in raw milk, a portion of the raw milk (10%) was incubated at 18 °C for 24 h. This portion was used as an inoculum for the remainder of the refrigerated raw milk with final populations of about 7.0 log CFU/ml for total counts
(mesophiles), from which 6.0 log CFU/ml were coliforms, and 6.0 log CFU/ml were psychrotrophs. The upstream processing conditions (separation, standardization and homogenization) decreased the initial microbial population by less than 1.0 log CFU/ml.
Approximately 1.5 ml of milk were collected per sample after each processing stage, and the ice cooled samples were serially diluted in Ringers solution (BR0052
%&! ! !
Oxoid, Thermo Fisher Scientific, Basingstoke, UK) before spread-plating on the relevant media. Mesophilic micororganisms were recovered on plate count agar (247940, BD
Difco, Sparks MD, USA) after incubation for 48 h at 32 °C as recommended by Laird and others (2004), enteric microorganisms were enumerated by plating on MacConkey agar (R453802 Remel, Thermo Fisher Scientific, Lenexa, KS, USA) before incubation for 24 h at 35 °C as described by Henning and other s (2004), and psychrotrophs were also recovered on plate count agar (247940, BD Difco, Sparks, MD, USA) after incubation for 10 d at 7 °C as suggested by Frank a nd Yousef (2004). Recovered microorganisms were counted and log reductions were evaluated subsequently.
2.2.5. Physico-chemical properties
Electrical conductivity and pH were measured after processing at room temperature with a handheld conductivity meter (CON 11, Oakton Instruments, Vernon
Hills, IL, USA), and a pH meter (AB 15 Accumet Basic, Fisher Scientific, Hampton, NH,
USA), respectively. Compositional analysis of milk was conducted at the Laboratory
Services Division, University of Guelph using a MilkoScan (FT 120 type 71200, Foss
Analytics, Copenhagen, Denmark). Fat or solids content were related to electrical conductivity as in Eq 2-2 and 2-3.
I"#ªºΩ< &5æ;( 2øºΩ
I"#ª¿¡M &6hg( 2ø¿¡M Eq. 2-3
%(! ! !
2.2.6. Statistical analysis
Statistical analyses were conducted using R version 2.10.1 (R Foundation for
Statistical Computing, Vienna, Austria) in conjunction with the Agricolae (Mendiburu
2009) and Rcommander (Fox 2009) packages for each set of data with a minimum of two batches (n=2) applied with two repetitions per treatment. The effect of the PEF processing conditions on milk was compared for each of the three microbial populations
(mesophiles, coliforms and psychrotrophs) separately using a multi-way analysis of variance and multiple comparisons of treatments by means of Tukey (HSD) with an alpha value of 0.05. Differences between the upstream and downstream products with regard to the relationship between electrical conductivity and composition (percentage of fat or solids) of milk were evaluated following an analysis of covariance. Means of pH values were compared for each product and processing step. An analysis of covariance
(ANCOVA) was conducted following the procedure described by Faraway (2002) to compare the relation between fat or solids content and conductivity between the different products.
2.3. Results and Discussion
2.3.1. Bacteria reduction in PEF-treated milk of different fat contents
Prior PEF processing, for homogenized milks with fat contents of 1.0, 1.8, and
3.1%, the following microbial counts were obtained: 6.9, 6.2, and 6.1 log CFU/ml for total mesophilic counts; 5.7, 6.2, and 6.1 log CFU/ml coliform counts; and 6.2, 5.9 and
%)! ! !
5.9 log CFU/ml psychrotoph counts for the respective fat levels. Before HTST treatment was conducted the total mesophile counts were 7.8, 6.2 and 7.6 log CFU/ml, total coliforms were 6.0, 6.2 and 5.8 log CFU/ml and the total psychrotophs were 6.2, 5.8, and 6.2 log CFU/ml in milk with fat levels of 1.1, 2.2, and 3.1%, respectively. The reduction in counts of total mesophiles, psychrotrophs and coliforms achieved in milk at the above fat levels following PEF applied at different electric fields and treatment times is shown in Table 2-2. Application of PEF to milk led to a reduction of total mesophiles ranging from 1.8 to 3.8 log CFU/ml (from 399 to 798 kJ/L) at fat levels between 1.1 and
3.1%. A trend towards higher reductions of total mesophiles (P<0.05) was observed at lower fat concentrations when a higher electric field (4.8 kV/mm) and a longer treatment time (918 !s) process was applied.
A high electric field strength (4.8 kV/mm) and low pulse frequency (12 Hz), resulting in a high energy level (798 kJ/L) proved to be more effective for the reduction of mesophiles in milk than processing the latter at a lower electric field strength (3.2 kV/mm) and higher frequency (25 Hz), which resulted in similar energy density (739 kJ/L).
Reduction of total mesophilic bacteria in milk pasteurized with HTST treatment ranged from more than 3.8 to 4.7 log CFU/ml using 72 °C for 15 s, while HTST pasteurization at 85 °C for 20 s led to a maximum r eduction of up to 4.9 log CFU/ml
(Table 2-2).
%*! ! !
Table 2-2. Variation in microbial populations after pulsed electric fields (PEF) and high temperature short time (HTST) processing of fluid milks
PEF Intensity HTST Intensity Initial 3.2 kV/mm 4.8 kV/mm 4.8 kV/mm 72 °C 85 °C Population x 1913 !s x 459 !s x 918 !s x 15 s x 20 s (log CFU/ml) Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM Mean SEM
Reduction in mesophilics (log CFU/ml) %Fat 1.1 7.3 ± 0.47A 2.2 ± 0.08a 2.3 ± 0.06a 3.8 ± 0.09c 4.7 ± 0.09d 4.9 ± 0.07d 2.0 6.4 ± 0.43B 2.0 ± 0.09a 2.3 ± 0.21a 3.1 ± 0.20b > 3.8 ± 0.00 > 3.8 ± 0.00 A a a bc d d "#! 3.1 6.9 ± 0.81 1.8 ± 0.15 2.1 ± 0.25 3.2 ± 0.18 4.7 ± 0.07 4.6 ± 0.03 % Fat Reduction in coliforms (log CFU/ml) B 1.1 5.8 ± 0.24 2.6 ± 0.15abc 3.0 ± 0.11abc > 3.3 ± 0.00 3.5 ± 0.05c 3.5 ± 0.03c B 2.0 6.1 ± 0.23 2.4 ± 0.41ab 2.6 ± 0.09abc 2.3 ± 0.10a > 3.8 ± 0.00 > 3.8 ± 0.00 3.1 6.3 ± 0.71A 3.4 ± 0.09bc 3.4 ± 0.17bc 3.0 ± 0.35abc 3.3 ± 0.00abc 3.4 ± 0.03bc % Fat Reduction in psychrotrophs (log CFU/ml) A 1.1 6.2 ± 0.10 > 3.9 ± 0.00 > 3.9 ± 0.00 > 3.9 ± 0.00 > 3.8 ± 0.00 > 3.8 ± 0.00 C 2.0 5.9 ± 0.16 > 3.5 ± 0.03 > 3.3 ± 0.23 > 3.5 ± 0.00 > 3.4 ± 0.00 > 3.4 ± 0.00 3.1 6.1 ± 0.16B > 3.5 ± 0.00 > 3.1 ± 0.40 > 2.8 ± 0.67 > 3.8 ± 0.00 > 3.8 ± 0.00 A,B,C,a,b,c,d,eSuperscripted letters indicate means within the microbial population that are statistically different, with uppercase indicating the initial population, and lowercase the reduction in the populations compared separately within microbial groups. SEM = Standard error of the mean. > indicates values where the microbial count detection limit was reached.The energy inputs by PEF correspond to 1478 kJ/L (3.2 kV/mm x 1913 !s), 798 kJ/L (4.8 kV/mm x 459 !s), and 1597 kJ/L (4.8 kV/mm x 918 !s). The temperature time combinations correspond to energies of 265 kJ/L (72 °C x 15 s), and 316 kJ/L (85 °C x 2 0 s ) .
! ! !
Similar trends showing higher reductions at lower fat levels in milk treated with
PEF were also observed for psychrotroph counts. The reduction in counts observed in this case ranged from greater than 2.8 log CFU/ml in 3.1% fat milk using 4.8 kV/mm for
918 !s to more than 3.9 log CFU/ml in milk containing 1.1% fat at all electric field and treatment time combinations studied (P<0.05). A similar trend was observed for PEF inactivation of psychrotrophs at the different fat levels (P%0.05). Reductions higher than
3.4 log CFU/ml were achieved by HTST pasteurization for all fat levels (1.1%, 2.0%, and 3.1%) and time-temperature combinations (72 °C for 15 s; 85 °C for 20 s; 95 °C for
15 s).
A statistical difference (P%0.05) was found between the minimum (2.3 log
CFU/ml) and the maximum reductions (3.4 log CFU/ml) for coliform bacteria in 2.0% fat milk when processing with PEF at 4.8 kV/mm for 459 !s, and in milk with 3.1% fat using
3.2 kV/mm for 1913 !s, respectively. The minimum and maximum reductions achieved with HTST milk processing for coliform bacteria were 3.3 (3.1% fat) and more than 3.8
(2.2% fat) log CFU/ml, respectively. Although these differences were statistically significant (P<0.05), there were no indications that higher temperatures or higher fat content could affect thermal inactivation of the coliforms. This may indicate that the changes observed after PEF treatment in the mesophilic population could have been influenced by variations in psychrotroph count or other groups of microorganisms not examined in the present study, however, this trend was not observed for HTST-treated microorganisms.
&"! ! !
The influence of milk fat content on PEF effectiveness for microbial inactivation has been studied by several researchers, who differ in their conclusions (Sobrino-López
& Martín-Belloso 2009). Grahl and Märkl (1996) PEF-treated Escherichia coli in ultra high temperature (UHT) milk containing 1.5 and 3.5% fat content in a batch vessel with
20 pulses of 1.4 or 1.2 kV/mm, and at 2.2 kV/mm for 45.7 !s or 30 !s, respectively, and applied regression analysis between survivors and electrical field (BE) and survivors and treatment time (Bt), as described by H*lsheger (1981) in order to observe the variation in the regression coefficients. The BE varied between -4.8 and -3.1 mm/kV, and the Bt between -6.0 and -4.0 !s-1 for 1.5% and 3.5% fat milk, respectively, which indicated a protective effect of fat against the high voltage pulses. Reina and others (1998) investigated Listeria monocytogenes in milk with fat contents of 0.2%, 2.0%, and 3.5% and PEF-processed in a continuous system (0.42 L/h at a pulse frequency of 1700 Hz) with an electric field of 3.0 kV/mm for 100, 300 and 600 !s at 25 °C. These researchers
(Reina and others 1998) reported no significant differences in inactivation based on the different milk fat levels. In their investigation of PEF processing variables, Sobrino-
López and others (2006) studied the inactivation of Staphylococcus aureus in milk in a continuous PEF system (at a pulse frequency of 100Hz and temperatures kept below 25
°C). The analysis looked into different fat content s (0.0%, or 3.0%), electric field strengths (2.5 and 3.5 kV/mm), pulse numbers (50 and 150), pulse widths (4 and 8 !s), and pulse types (monopolar and bipolar). A response surface quadratic model revealed that fat content did not affect the inactivation of S. aureus, although there were differences in the inactivation rates when the electric field intensity and pulse width varied. The authors (Sobrino-López and others 2006) suggested that the conductivity of
&#! ! ! the samples (0.60 S/m for skim milk and 0.55 S/m for whole milk) did not influence the inactivation trends.
Some studies suggest that the fat content of the milk may alter the effectiveness of microbial reduction by PEF in milk, and that analysis of the processing conditions, as performed by Grahl and Märkl (1996) and Sobrino-L'pez (2006) can help to better assess the relative effect of the variation in fat content. As suggested by Walkling-
Ribeiro and others (2009), the protective effect may be noticed above a certain fat level, which may depend on specific processing conditions. Therefore, the application of models like surface response analysis or tertiary models considering composition will aid in the optimization of PEF milk processing. In relation to the microbial groups, it is important to note that the reduction in microbial populations in raw milk varies depending on the population number and diversity, with resistant populations determining the surviving population. For example, Sepúlveda and others (2009) treated whole milk by PEF using processing parameters of 3.5 kV/mm for 11.5 !s, a flow rate of
72 L/h and a maximum processing temperature of 65 °C (10 s). Under these conditions the researchers (Sepúlveda and others 2009) achieved a reduction of mesophilic, enteric, and psychrotrophic bacteria of 1.5, more than 1.5, and more than 0.4 log
CFU/ml, respectively. In the present study higher microbial reductions may be expected from HTST as temperature increased, but the presence of thermoduric microbial populations even at the highest temperature setting may have caused a high post- pasteurization population confirmed following procedures as in Frank & Yousef (2004) in a confirmatory experiment.
&$! ! !
2.3.2. Effect of the PEF inlet temperature on the reduction of microorganisms
The milk temperature prior to PEF processing as well as the overall exposure time of milk to the pulse treatment affected the survival of mesophiles, coliforms and psychrotrophs (Table 2-3). An increase of the PEF processing inlet temperature (from
5.0-6.3 °C to 34.0-34.3 °C), and doubling the milk flow rate (1.2-2.4 L/h to 3.6-4.8 L/h), while maintaining the maximum outlet temperature at 56°C, caused a 3.3 log CFU/ml reduction of total mesophilic bacteria when an electric field strength of 5.6 kV/mm, for
230 !s was applied. Application of a lower electric field strength of 4.0 kV/mm led to an increase in bacterial inactivation from 1.5 to 2.8 log CFU/ml in half the time (765 !s to
383 !s), suggesting an additive treatment effect. These results confirm that the microbial inactivation by PEF is effective at mild temperatures (between 34.3 and 50.3
°C), and that the PEF treatment of a colder product (5.0 to 5.9 °C) requires additional energy to cause a rise in temperature. The overall effect was to produce a final outlet temperature of 43 °C. For total coliforms, a minimu m reduction of 2.0 log CFU/ml was achieved at the lowest PEF treatment conditions (4.0 kV/mm for 765 !s with 924 kJ/L), and an increase in milk temperature at the PEF chamber inlet from 5.9 to 34.3 °C together with an increase of the product flow from 2.4 L/h (765 !s) to 4.8 L/h (383 !s), resulted in a greater inactivation (P < 0.05) of coliform bacteria of up to 2.7 log CFU/ml.
More severe PEF treatment conditions (543 kJ/L from 5.6 kV/mm for 230 !s, 616 kJ/L from 4.0 kV/mm for 510 !s, 1086 kJ/L from 5.6 kV/mm for 459 !s, or 1848 kJ/L from 4.0 kV/mm for 1530 !s) led to coliform reductions in milk of >3.9 log CFU/ml (Table 2-3).
&,! ! !
Table 2-3. Effect of the pulsed electric fields inlet processing temperature on the reduction of different bacteria groups in skim milk
Mesophilics Coliforms Psychrotrophs Intensity Mean SEM Mean SEM Mean SEM Inlet Microbial population Temperature (Log CFU/ml) (°C) " Initial population 6.2 ± 0.13A 6.3 ± 0.18A 6.2 ± 0.08A Reduction in microbial population
(Log CFU/ml) 4.0 kV/mm x 765 !s 1.5 ± 0.09a 2.0 ± 0.41a 3.1 ± 0.18a Low 4.0 kV/mm x 1530 !s 3.7 ± 0.10c > 3.9 ± 0.00 > 3.8 ± 0.00 (5.0-6.3 °C) 5.6 kV/mm x 459 !s 3.4 ± 0.05bc > 3.9 ± 0.00 > 3.8 ± 0.00
4.0 kV/mm x 383 !s 2.8 ± 0.21b 2.7 ± 0.10a 2.9 ± 0.29a Mild 4.0 kV/mm x 510 !s 3.0 ± 0.23bc > 3.9 ± 0.00 > 3.8 ± 0.00 (34.0-34.3 °C) 5.6 kV/mm x 230 !s 3.3 ± 0.07bc > 3.9 ± 0.00 > 3.8 ± 0.00
A,a,b,c,d,eSuperscripted letters indicate means within the different groups of bacteria that are statistically different with uppercase indicating the initial population, and lowercase the reduction in the populations comparated separately within microbial groups. SEM = Standard error of the mean. .Maximum outlet temperature: 56.3 °C. > indicates va lues where the microbial count detection limit was reached. The corresponding energy inputs were 924 kJ/L (4.0 kV/mm x 765 !s), 1848 kJ/L (4.0 kV/mm x 1530 !s), 1086 kJ/L (5.6 kV/mm x 459 !s), 462 kJ/L (4.0 kV/mm x 383 !s), 616 kJ/L (4.0 kV/mm x 510 !s), and 543 kJ/L (5.6 kV/mm x 230 !s).
&%! ! !
Psychrotrophic bacteria in milk were reduced by the applied PEF conditions in a similar manner as the coliform bacteria but a higher reduction in count of 3.1 log CFU/ml was achieved at milder PEF treatment conditions (using 4.0 kV/mm for 765 !s at kJ/L).
Reina and others (1998) also observed that while applying 30 kV/mm at 1,700
Hz, an increase in temperature to 50°C was needed t o achieve a 3 to 4 log CFU/ml (at
300 and 600 !s respectively) reduction in count compared to less than 2 and 3 log
CFU/ml in a temperature range between 25 and 43 °C for 600 !s at the same treatment times. Recent studies (Sepúlveda and others 2005, Sampedro and others 2007, and
Jaeger and others 2009) have taken a similar approach of decreasing the PEF energy input by increasing the product inlet temperature. Sampedro and others (2007) reduced the energy to inactivate 1.8 log CFU/ml of Lactobacillus plantarum in an orange juice milk-based beverage (50% orange juice, 20% skimmed milk) using between 352 and
890 kJ/L by increasing the product inlet temperature from 35 to 55°C while applying 4.0 kV/mm for up to 130 !s (pulse width of 2.5 !s and frequency between 110 and 356 Hz).
Similarly, Jaeger and others (2009) found that increasing the product inlet temperature to 30°C produced greater inactivation of Lb. rhamnosus (up to 2.0 log CFU/ml reduction) in milk (4.3% fat) when applying 3.0 kV/mm for 25 !s (square wave pulses of
3 !s at a frequency of 56 Hz, flow rate of 5 L/h and energy input of 100 kJ/kg) in a continuous system. Advanced systems for pre-heating milk to be PEF treated include the one used by Sepúlveda and others (2005) to inactivate raw milk microflora. In their case they recovered the heat produced in the treatment chamber from the outgoing milk stream, and incorporated into the incoming milk stream by using a tube-in-tube heat
&&! ! ! exchanger to increase the temperature to around 50 °C (for a maximum temperature of
65 °C for 15 seconds). Hoogland and de Haan (2007) estimated that large-scale PEF treatments could be applied using a heating/cooling cycle of 5-35-50-5 °C with regeneration most likely becoming not economically feasible.
From a microbiological perspective a possible synergistic effect of temperature and PEF treatment could be attributed to an increase in microbial membrane fluidity at higher temperatures caused by phase transition of the phospholipids from gel to liquid- crystalline structure, leading to more pronounced electroporation effects (Heinz and others 2003, and Jaeger and others 2009). In complex food systems like milk the liquefaction of colloidal particles (e.g. fat globules) that affect resistivity could also add to potential treatment synergism. Therefore, a decrease in sample viscosity (and increase in conductivity) is a factor that may increase milk PEF treatment efficiency as it improves the initial separation and homogenization operations (Walstra and others
2006, Kessler 2002). As excessive heat treatment causes chemical and physical changes in milk (Walstra 2006), consequently, PEF treatment duration is also limited to outlet temperatures ideally below the minimum low-temperature long-time pasteurization conditions (63 °C for 30 min) at ver y long treatment times (e.g. 108 !s) or shorter treatment times (e.g. 105 !s), which is well below HTST pasteurization requirements (72 °C for 15 s).
&(! ! !
2.3.3. Milk processing with a combination of CFMF and PEF
In order to enhance the processing efficacy, CFMF and PEF were used in a hurdle approach (CFMF/PEF), resulting in overall higher reductions of the bacterial load in milk (Table 2-4).
Milk processing with CFMF/PEF achieved >4.7, >4.7, and >3.7 log CFU/ml reductions for counts of total mesophilic, coliform and psychrotroph bacteria, respectively, using the most severe PEF conditions (at 4.8 kV/mm for 918 !s with an energy density of 798 kJ/L), which resulted in the highest milk temperature at the PEF chamber outlet (61 °C). A reduction to half of the energy density led to the same results at a lower outlet temperature of the product (56 °C ). Equivalent results were obtained for a combination of CFMF and HTST (CFMF/HTST), with the latter used at 72 °C for 15 s, which reduced total mesophilic, coliform and psychrotroph bacteria by >4.4, >4.3, and
>4.1 log CFU/ml, respectively (P % 0.05).
Early studies on milk microfiltration by Olesen and Jensen (1989) reported a reduction in the total count by 3.9 log CFU/ml, the spore count by 3.5 log CFU/ml and lactate fermenting bacteria by >1.8 log CFU/ml (lower initial count) using CFMF exclusively. A review by Saboya and Maubois (2000) cites that the average microbial reduction in fluid milk by CFMF is above 3.5 log CFU/ml with a variation among various microbial species including spores (>4.5 log CFU/ml), Listeria monocytogenes (3.4 log
CFU/ml), Brucella abortus (4.0 log CFU/ml), Salmonella Typhimurium (3.5 log CFU/ml), and Mycobacterium tuberculosis (3.7 log CFU/ml).
&)! ! !
Table 2-4. Reduction of microbial populations after processing of skim milk with cross-flow microfiltration (CFMF) and pulsed electric fields (PEF) or high temperature short time (HTST) pasteurization in a combined approach (CFMF/PEF; CFMF/HTST)
Mesophilics Coliforms Psychrotrophs
Treatment Intensity Mean SEM Mean SEM Mean SEM Microbial Population (Log CFU/ml) Initial population 6.9 ± 0.47A 7.0 ± 0.36A 6.3 ± 0.39B Reduction in microbial population
(Log CFU/ml) CFMF Standard† 2.1 ± 0.11a 2.9 ± 0.07b 3.2 ± 0.09b CFMF/PEF 3.2 kV/mm x 1913 !s 4.2 ± 0.10b > 4.7 ± 0.00 > 3.7 ± 0.00 CFMF/PEF 4.8 kV/mm x 459 !s > 4.7 ± 0.00 > 4.7 ± 0.00 > 3.7 ± 0.00 CFMF/PEF 4.8 kV/mm x 918 !s > 4.7 ± 0.00 > 4.7 ± 0.00 > 3.7 ± 0.00 CFMF Standard† 2.0 ± 0.18a 2.3 ± 0.08a 2.0 ± 0.03a CFMF/HTST 72 °C x 15 s > 4.4 ± 0.00 > 4.3 ± 0.00 > 4.1 ± 0.00 CFMF/HTST 85 °C x 20 s > 4.4 ± 0.00 > 4.3 ± 0.00 > 4.1 ± 0.00 CFMF/HTST 95 °C x 15 s > 4.4 ± 0.00 > 4.3 ± 0.00 > 4.1 ± 0.00
A,B,a,b,c,d,eSuperscripted letters indicate means within the groups of bacteria that are statistically different with uppercase indicating the initial population, and lowercase the reduction in the populations comparated separately within microbial groups. > indicates values where the microbial count detection limit was reached. †The standard CFMF conditions were inlet and outlet pressures of 90 (at 120 L/h), and 20 kPa (at 12 L/h) at 35 °C with a membra ne pore size of 1.4 !m, with an estimated energy consumption of 37 kJ/L. SEM = Standard error of the mean. The energy inputs by PEF correspond to 1478 kJ/L (3.2 kV/mm x 1913 !s), 798 kJ/L (4.8 kV/mm x 459 !s), and 1597 kJ/L (4.8 kV/mm x 918 !s). The temperature time combinations correspond to energies of 265 kJ/L (72 °C x 15 s), 316 kJ/L (85 °C x 20 s), and 355 kJ/L (95 °C x 15 s).
&*! ! !
Madec and others (1992) observed that the microbial reduction in skim milk can be constant and independent from the level of contamination, for example, the reduction in log CFU/ml obtained for Listeria innocua ranged between 1.8 and 1.9 at contamination levels between 102 and 106 CFU/ml and that of Salmonella abortus ovis between 2.4 and 2.5. In their study raising the CFMF processing temperature from 35 to
50 °C increased the decimal reduction of Salmonella, from 2.5 to 3.3, but not that of
Listeria. Elwell and Barbano (2006) reported a reduction in the total counts of milk of 3.8 log CFU/ml using CFMF, and a higher microbial load reduction of 5.6 log CFU/ml was caused by HTST when applied at 72 °C for 15 s. In t his study, because of the low initial population the quantifiable coliform reduction was %1.2 log CFU/ml with CFMF as well as with added heat pasteurization.
The reductions observed in the mesophilic counts in this study are comparable to those reported by Madec and others (1992) with Listeria innocua, and the reductions achieved for coliforms and psychrotrophs are closer to the average decimal reduction cited by Saboya and Maubois (2000), but considering that the processing temperature was 35 °C in our study, there is further room for a n increase in the treatment efficacy by taking advantage of the combined effect of moderate heating (50 °C) on thermally sensitive microbial groups. An equivalent CFMF unit to that utilized in this study (MFS-1
Tetrapak Processing) was used by Hoffman and others (2006) who observed almost no changes in the composition of the milk by CFMF processing, and the effects of thermal load in the cream (monitored by furosine content), and whey protein fractions
(denaturation of +-lactalbumin, )-lactoglobulin and immunoglobulins) were reduced by
(+! ! ! applying HTST instead of UHT to produce extended shelf life (ESL) milk. Our study is the first considering the combination of PEF and CFMF, and is a complement to the study by Walkling-Ribeiro and others (2010), which considered the recovery of several microbial populations after combining both technologies.
2.3.4. Inactivation of native bacteria in cream by PEF
The effect of PEF on the inactivation of different groups of milk bacteria in standardized and homogenized cream containing 12% fat is presented in Table 2-5.
The results indicate that there are no significant differences (P % 0.05) in the reduction of the total mesophile population (0.7 to 3.4 vs. 0.1 to 2.9 log CFU/ml) or psychrotrophs
(0.5 to >3.8 log CFU/ml vs. 0.2 to >3.1 log CFU/ml) for standardized and homogenized cream at the three electric field strength and treatment time combinations tested (4.0 kV/mm for 765 !s; 4.0 kV/mm for 1530 !s; and 5.6 kV/mm for 459 !s). However, for coliform and psychrotroph bacteria higher inactivation in homogenized cream than for standardized cream (2.8 vs. 1.6 log CFU/ml and 2.0 vs. 0.5 log CFU/ml) were obtained when 4.0 kV/mm were applied for 1530 !s. For standardized and homogenized cream,
PEF processing conditions using 5.6 kV/mm for 459 !s with 543 kJ/L generally proved to be more effective (P<0.05) than PEF treatment with 924 (4.0 kV/mm for 765 !s) or
924 (4.0 kV/mm for 1530 !s) kJ/L. The higher reduction obtained with 543 kJ/L compared with the one achieved at 924 kJ/L can be attributed to the higher electric field
(5.6 kV/mm compared to 4.0 kV/mm), which also caused a higher temperature rise (50-
51 °C vs. 42-44 °C, respectively) in a shorter trea tment time (459 vs. 1530 !s, respectively).
("! ! !
Table 2-5. Effect of pulsed electric fields processing on different groups of native bacteria in cream
Pre- Mesophilics Coliforms Psychrotrophs Intensity treatment Mean SEM Mean SEM Mean SEM Microbial Population
(Log CFU/ml) STND 6.5 ± 0.10A 6.3 ± 0.05A 6.2 ± 0.13A Initial population HOMO 5.6 ± 0.10B 5.5 ± 0.05B 5.5 ± 0.05B Reduction in microbial population
(Log CFU/ml) 4.0 kV/mm x 765 !s 0.7 ± 0.00ab 0.4 ± 0.05a 0.5 ± 0.06a STND 4.0 kV/mm x 1530 !s 1.7 ± 0.05b 1.6 ± 0.07b 0.5 ± 0.00a 5.6 kV/mm x 459 !s 3.4 ± 0.06c 3.6 ± 0.15a > 3.8 ± 0.00 4.0 kV/mm x 765 !s 0.1 ± 0.05a 0.2 ± 0.09a 0.2 ± 0.03a HOMO 4.0 kV/mm x 1530 !s 1.2 ± 0.49b 2.8 ± 0.19c 2.0 ± 0.40b 5.6 kV/mm x 459 !s 2.9 ± 0.15c > 3.1 ± 0.03 > 3.1 ± 0.00
A,B,a,b,c,d,eSuperscripted letters indicate means within the group of bacteria that are statistically different. SEM = Standard error of the mean, STND = standardized, HOMO = homogenized. > indicates values where the microbial count detection limit was reached. Homogenization was conducted at two pressure stages of 20 MPa and 5 MPa respectively. The respective energy inputs were 924 kJ/L (4.0 kV/mm x 765 -s), 1848 kJ/L (4.0 kV/mm x 1530 -s), and 1086 kJ/L (5.6 kV/mm x 459 -s).
(#! ! !
The microbial reductions achieved in this study were higher than those obtained by Mañas and others (2001), and Picart and others (2002) who achieved not more than
2.0 log CFU/ml reduction in E. coli counts by treating cream of higher fat content in batch PEF systems. Mañas and others (2001) conducted their study with full-fat (33%) cream using PEF at 3.34 kV/mm for 28 to 98 !s and suggested that a fat content above whole milk did not enhance the protection of microorganisms against inactivation by
PEF. Similarly, Picart and others (2002) obtained a maximum reduction of 2.0 log
CFU/ml using PEF at 3.7 kV/mm for 250 !s for processing of 20% fat cream inoculated with a cocktail of three Listeria innocua strains of different heat resistances. Toepfl and others (2007) suggested that, theoretically, the lethal effect of PEF might be impaired by agglomeration of microbial cells or between cells and insulating particles present in the food. The latter authors estimated that the variations within these factors may occur in zones where the peak electric field can be reduced by a factor of 0.45, thus, the development of a challenging processing scenario for the worst case is suggested in order to determine the most effective electric field strength. In laboratory observations by Toepfl (2006), an increase in fat content up to 10% by addition of cream to milk did not have an impact on the effectiveness of inactivation of Lb. rhamnosus or
Pseudomonas fluorescens (reduced by >3.0 and 4.0 log CFU/ml, respectively) when applying 2.16 kV/mm (up to 150 kJ/kg). It can be speculated that this effect was due to a substantial amount of inoculated microorganisms remaining in the aqueous phase of milk (Toepfl 2006).
($! ! !
Otunola and others (2008) compared the inactivation of E. coli in the presence of indigenous microorganisms following PEF treatment at 4.0 kV/mm for 120 !s of milk with 2% fat and cream with 18% fat. They observed reductions in the bacterial counts of
1.5 and 1.1 log CFU/ml, respectively, which was not significant. The difference was more noticeable at a shorter treatment time (30 !s) where a 0.9 log CFU/ml reduction in count was achieved in 2% milk compared to 0.2 log CFU/ml in the 18% cream. These results suggest a tendency towards a faster initial decrease in bacterial numbers in the
2% milk with a remaining resistant population in both products that caused a tailing in the inactivation graph at a treatment time of 120 !s (4.0 kV/mm). Therefore, at a specific electric field, high fat dairy products may require longer treatment times for microbial inactivation to occur.
While it is possible to achieve a relatively high microbial reduction in cream at a given electric field, high fat dairy products may require longer treatment times for microbial inactivation to occur. For instance, at the highest PEF setting used in this study (5.6 kV/mm for 459 !s), total counts were reduced by 3.4 log CFU/ml in standardized cream, and 2.9 log CFU/ml in homogenized cream (12% fat), similar
(P%0.05) to the 3.7 log CFU/ml reduction in counts observed in skim milk (Tables 2-3 and 2-5). Also, the difference was reflected in a lower temperature rise resulting from the PEF treatment of these products (viz. 45 °C in standardized and 46 °C in homogenized 12% fat cream vs. 50 °C in 0.2% fat mi lk). Therefore, the protective effect of cream is related to the PEF conditions, and an increase in the intensity of the treatment conditions may be needed for creams, similar to the commercial practice applied for heat pasteurization, where cream needs to be pasteurized or sterilized at 5
(,! ! ! to 7 °C above the temperature levels used for whole milk (Kessler 2002). In the United
States the pasteurized milk ordinance (FDA 2009) requires an increase of 3 °C to the pasteurization temperature for milks with fat content higher than 10%.
2.3.5.Effect of processing on the physicochemical properties of milk
Conductivity is one of the most critical physical properties to examine as it affects the energy input and change in temperature at a given dose. This has led researchers to its measurement in various food products and further relation with temperature
(Amiali and others 2006, Ruhlman and others 2001).
When analyzed separately, the relations between fat content and electrical conductivity of raw and skim milk showed a low correlation (R2 of 0.542 and 0.457 respectively) compared to cream (R2 = 0.824). In the other hand, when the information for all the non-homogenized products were combined a correlation of 0.962 was obtained, while when all the products were included a correlation of 0.824 was obtained
(Table 2-6). This decrease in correlation between fat content and electrical conductivity may have been caused by variations in the physicochemical properties due to homogenization, and heating as explained by the Milkoscan manual (FOSS 1998). The correlation of individual products after homogenization was not analyzed as they resulted in low correlations, which as in the raw and skim can be attributed to a higher variability in conductivity in a small range of fat contents (3.7 % to 4.0 % and 0.05 % to
0.30 % fat respectively).
(%! ! !
Table 2-6. Linear regression coefficients of the relation between fat or solids and conductivity of milk fractions
Fraction Slope Lower Upper Intercept Lower Upper Adjusted RSE Limit Limit Limit Limit R2 Conductivity (S/m) as a function of fat content (%) Raw 0.000 0.000 0.000 0.50 0.50 0.50 0.542 0.000 Skim 0.000 0.000 0.000 0.50 0.50 0.50 0.457 0.000 Cream -0.006 -0.007 -0.005 0.45 0.42 0.50 0.820 0.032 Combined -0.007 -0.008 -0.007 0.51 0.50 0.51 0.962 0.022 Standardized -0.006 -0.012 0.000 0.51 0.49 0.52 0.088 0.017 Homogenized -0.006 -0.021 0.008 0.49 0.45 0.52 -0.005 0.046 Pulsed -0.008 -0.034 0.019 0.45 0.38 0.51 -0.040 0.049 Heated 0.000 0.000 0.000 0.40 0.40 0.40 0.639 0.000 "#! Combined -0.008 -0.015 -0.002 0.48 0.47 0.50 0.050 0.033 Non- -0.007 -0.008 -0.007 0.50 0.50 0.51 0.970 0.018 homogenized All -0.007 -0.008 -0.007 0.48 0.47 0.48 0.824 0.036 combined Conductivity (S/m) as a function of solids content (%) Skim 0.003 -0.020 0.027 0.47 0.25 0.68 -0.021 0.017 Retentate 0.077 0.030 0.125 0.08 -0.54 0.19 0.736 0.017 Permeate Pulsed 0.064 -0.032 0.161 -0.04 -0.50 0.43 0.706 0.005 Heated 0.053 0.037 0.068 0.02 -0.09 0.14 0.966 0.013 Combined 0.053 0.049 0.056 0.02 -0.01 0.06 0.921 0.019
RSE = Residual Standard Error.
! ! !
0.7 0.7 y = -0.007x + 0.50 A y = 0.05x + 0.02 0.6 0.6 B Adjusted R² = 0.970 Adjusted R² = 0.921
0.5 0.5
0.4 0.4
0.3
0.3 (S/m) Conductivity Conductivity (S/m) Conductivity ""! 0.2 0.2
0.1 0.1 0.0 10.0 20.0 30.0 40.0 50.0 60.0 2.0 4.0 6.0 8.0 10.0 12.0 % Fat % Solids !
Figure 2-2. Relation between electrical conductivity and fat or solids content (A = non-homogenized products; B = non-fat milk products)
! Lines represent the range where a linear regression was established. Homogenization was conducted at two pressure stages of 20 MPa and 5 MPa respectively.
! ! !
A similar trend was observed for the solids content, which resulted in R2 ranging from 0.706 to 0.966 in the CFMF products, and combined, gave a correlation of 0.921, the range of solids analyzed was 4.1 % to 9.0 %. As mentioned above, the analysis of the relation between conductivity and solids content did not result in a correlation either.
From the slopes it can be estimated that for each 10% increase in the fat level the conductivity of the non-homogenized products decreased by 0.07 S/m in the range of 0.05 and 50.60 % fat and conductivities between 0.54 and 0.17 S/m (Figure 2-2 A), while the conductivity increased by 0.05 S/m for each 1% of increase in solids content in the range of 4.1 % and 9.0 % solids and 0.21 and 0.51 S/m for conductivity.
Mabrook and Petty (2003) reported milk conductivities of 0.54 S/m for 0.1% fat,
0.52 S/m for 1.6% fat, and 0.51 S/m for 3.6 % fat, which represented a slope of -0.010
S/m per % fat and a 0% fat value (intercept) of 0.54 S/m. Sobrino-L'pez and others
(2006) observed conductivity values of 0.60 S/m for 0% fat, 0.60 S/m for 1.5 % fat and,
0.55 S/m for 3.0% fat, respectively, or a slope of -0.018 S/m˙% fat, and intercept of 0.61
S/m. Compared to the results presented in this study the previous authors observed slightly higher conductivity values (i.e. 0.60 or 0.54 compared to 0.50 S/m for skim milk), and a higher decrease in conductivity with increasing fat content. This variation may be explained by the sensitivity of the methods used to determine conductivity and % fat
(not specified).
On average, the pH of all milk products investigated ranged from 6.71 and 6.98.
Milks with fat contents between 1.1 to 3.1% varied between 6.71 and 6.83, while the pH
()! ! ! of cream was between 6.84 and 6.92, and for non-fat milk products a pH from 6.82 to
6.98 was obtained. There were no differences (P%0.05) in pH between separation products (raw, cream and skim milk), nor within the different milk, cream, and non-fat milk products. In a comparable study, Michalac and others (2003) found no differences in electrical conductivity or pH of raw or UHT skim milk after PEF treatment (at 3.5 kV/mm for 47.4 !s using a flow rate of 3.6 L/h), or heat pasteurization (73 °C for 30 s), with pH values between 6.68 and 6.76. Saboya and Maubois (2000) obtained, after
CFMF treatment of debacterized skim milk with a pH 6.72, milk retentate and permeate with a pH of 6.73 and 6.69, respectively. Among the few studies on PEF treatment of cream no variation in pH was reported following exposure to PEF, which is in line with the findings of this study.
2.4.Conclusions
In the present study the most effective PEF treatment of milk inoculated with native bacteria led to comparable or slightly lower reductions in total mesophilic, psychrotroph and coliform counts than those obtained with HTST pasteurization.
Treatment efficacy of PEF against the microorganisms was generally enhanced when higher electric field strengths and energy densities were applied. Moreover, a lower milk fat content enabled higher bacterial inactivation by both PEF and HTST in general.
Combining CFMF and PEF in a hurdle strategy increased the antimicrobial effect to an equivalent level observed for HTST treatment and consequently, further in-depth research on CFMF/PEF processing of milk is recommended. For cream processing,
(*! ! !
PEF proved to be more effective for standardized than for homogenized cream. The conductivity of the PEF-treated milk was primarily affected by the solids content and secondarily by the fat content, while no variation in the pH of the milk products was observed during processing.
)+! ! !
3. Microbial inactivation and shelf life comparison of ‘cold’ hurdle processing with pulsed electric fields and cross-flow microfiltration, and conventional thermal pasteurisation in milk
3.1. Introduction
Conventional heat pasteurisation has been applied and acknowledged as the most effective processing method for the commercial production of microbiologically safe milk for nearly a century. With the emergence of novel technologies for milk processing such as microfiltration (Hoffmann and others 2006; Makardij and others
1999), pulsed electric fields (Smith and others 2002; Walkling-Ribeiro and others 2009), ultrasound (Noci and others 2009; Zenker and others 2003), bactofugation (De Noni and others 2007; Te Giffel and Van Der Horst 2004), and high hydrostatic pressure
(Datta and Deeth 1999; Guan and others 2005), the cornerstone has been laid to meet growing consumer demands for enhanced nutritiousness, sensory attributes, and shelf stability of milk. Due to its complex structure milk serves as an ideal growth medium for a wide variety of microorganisms, of which, coliform, psychrotrophic, and endospore- forming bacteria are of particular interest and importance from a food safety point of view (Christiansson and others 1989, Oliver and others 2005, Vasavada 1988) because of their respective potential pathogenicity, verifiable growth at refrigerated storage conditions, and distinct resistance to conventional milk processing conditions. The latter are limited by the susceptibility of milk to thermal stress, which affects product quality and consumer acceptability. Incorporation of CFMF in the milk manufacturing process has led to an improvement of the microbial and organoleptic quality of milk following the principle of liquid separation by membranes with a pore size in the range of 0.1 to 10
)"! ! !
!m (Brans and others 2004, Daufin and others 2001, Kaufmann and Kulozik 2006), which generally enable high flow rates, improved yield, reduced processing costs and higher quality compared to conventional heat treatments. Nonetheless, CFMF is combined with thermal pasteurisation (TP) for commercial scale production of milk due to current regulatory food safety requirements (Eberhard and Gallmann 2009, Lewis
2010).
Milk processing with PEF is based on a different mode of action known as electro-permeabilisation of microbial membranes (Wouters and others 2001a), bringing about loss of cell viability. Similar to CFMF, commercial milk processing with PEF is subject to legal restraints and therefore, additional TP treatment of milk is required
(European Union 1997; FDA 2009). Combining CFMF with PEF in a hurdle technology
(Leistner 2000) for ‘cold’ pasteurisation of milk has not been studied to date and may not only represent an alternative for the production of minimally processed milk of potentially better microbiological safety and shelf life as well as nutritional and sensory quality than heat pasteurised milk but it could also contribute to an increase in legislative acceptance of emerging food processing technologies in general.
An investigation of the efficacy of a novel nonthermal processing approach for milk with CFMF, PEF and their combinations (CFMF/PEF, PEF/CFMF) compared to that of conventional heat pasteurisation for the inactivation of native biota was the first objective of this study. Another aim was the evaluation of the microbiological shelf stability of milk preserved with the most effective ‘cold’ hurdle technology in comparison to TP-treated milk.
)#! ! !
3.2. Materials and methods!
3.2.1. Milk preparation and analysis of microbial inactivation
Separated raw bovine skim milk (0.08 % fat content, 50 °C) was obtained from a dairy processing plant (Gay Lea Foods, Guelph, ON, Canada) and subsequently 90% of the product was refrigerated while 10% of the raw milk was used for inocula. Raw milk provided for inocula was spiked with an additional portion (10% of the inoculum volume) of refrigerated whole bovine raw milk from another dairy (University of Guelph Dairy
Research Station, Elora, ON, Canada) in order to diversify the microbial population further and then incubated for 24 h at 25 °C to pro duce final inocula containing 1010
CFU of native biota per ml. Prior to processing, these inocula were added to the refrigerated raw skim milk for a 1:10 (v/v) dilution ratio, thereby, providing high initial counts that ensured accurate analysis of the inactivation of native microorganisms above the detection limit of the assays used for their enumeration (1.5 log10 CFU/ml) after hurdle processing. Additionally, inoculated raw milk was filtered through a cheese cloth (Softwipe, American Fiber and Finishing Inc., Albemarle, NC, USA) allowing removal of possible particles for uniform and reproducible treatment conditions.
)$! ! !
A B
Electrode Top View
Product outlet
+
PEF Cooling CFMF
Charge connectors C Lateral View
Insulating spacers Charged electrodes
Product Product inlet inlet CFMF PEF Cooling Ground electrode !
!
Figure 3-1. Schematic drawing of the pulsed electric fields (PEF) treatment chamber used for skim milk processing and sequence for combined processing of PEF with cross-flow microfiltration (CFMF)
!
A: Aerial view and side view of the pulsed electric fields (PEF) treatment chamber used for skim milk processing; B: schematic sequence for combined processing with PEF and followed by cross-flow microfiltration (CFMF), and C:reversed order (B) with CFMF followed by PEF
),! ! !
Aliquots of 100 !l from treated milk samples, or their ten-fold dilutions prepared with Ringers solution (BR0052G, Oxoid Ltd., Basingstoke, Hampshire, UK), were spread on plate count agar (PCA; 247940, BD Difco, Sparks, MD, USA) under aseptic conditions in a class 2 biosafety cabinet (Forma 4071-1200, Thermo Fisher Scientific
Inc., Newlington, NH, USA) for the analysis of total aerobic mesophilic counts (TAMC).
Agar plates were incubated at 30 °C for 72 h as des cribed in a study by Zárate and others (1997) before counting.
3.2.2.Nonthermal processing with PEF
For nonthermal processing of inoculated raw skim milk, a treatment chamber designed at the University of Guelph and the University of Waterloo was used with a custom-built pulsed electric fields (PEF) generating system (PPS 30, University of
Waterloo, Waterloo, ON, Canada). The horizontally-oriented treatment chamber (Figure
3-1) is equipped with 3 food grade stainless steel cylindrical electrodes fitted in threaded stainless steel cylinders, which, in turn, are screwed into insulating cylinders of thermo- resistant plastic at the top and at the bottom. A stainless steel pole (3.8 mm diameter) mounted vertically on a circular base plate functions as the grounding electrode, creating an average electrode gap of 2.5 mm to the encircling charged electrodes with a total treatment surface area of 49 cm2 per electrode pairing. The product entered and exited the PEF chamber through food grade stainless steel connectors installed on the bottom and top, respectively, and gaskets were placed before the thread sections to ensure leak tightness. The total product volume in the chamber was 600 ml with a treatment volume of 17 ml in the case that all 3 cylindrical electrodes were operated. A
)%! ! ! pulse generator unit consisting of 2 charging capacitors (each with a capacitance of
0.31 !F) and a thyratron switch (TM-27) was used for the generation of monopolar exponential decay pulses (with a pulse width of 1.5 !s), which were measured with a high voltage probe (P6015A, Tektronix Inc., Beaverton, OR, USA) and monitored on a
2-channel digital storage oscilloscope (TD 1012, Tektronix Inc., Beaverton, OR, USA). A peristaltic pump (Masterflex pump drive 7524-40 and pump head 77201-60, Cole
Parmer Instrument Co., Vernon Hills, IL, U SA) was used for passing inoculated milk at a flow rate of 2.4 L/h through peroxide-cured silicone precision tubing (Masterflex
L/S 16, Cole Parmer Instrument Co., Vernon Hills, IL, USA). The latter was connected by food grade stainless steel flow-through cells for temperature measurement before and after the processing equipment, at the inlet and outlet of the PEF chamber. The tubing was subsequently coiled and submerged in a cooling bath (NESLAB RTE 7,
Thermo Fisher Scientific Inc., Newlington, NH, USA) prior to the sampling point. The product processing temperature was measured and recorded with a portable wireless data-logger (OM-SQ2020-2F8, Omega, Stamford, CT, USA), which was connected to
T-type thermocouples (TMTSS-040G-6, Omega, Stamford, CT, USA), which were connected to the flow-through cells described above.
Milk processing with PEF was carried out at electric field strengths of 1.6, 2.0, 3.0 and 4.2 kV/mm for respective treatment times of 2105, 1454, 983, and 612 !s. These treatments conditions corresponded to specific energy densities of 407, 632, 668, and
815 kJ/L, respectively. Milk temperatures ranged from 16 °C at the PEF chamber inlet to
40, 45, 47 and 49 °C, depending on the energy densi ty, at the chamber outlet and the subsequent cooling step reduced the product temperature to 10 °C prior to sampling.
)&! ! !
Samples of processed milk were placed in a Styrofoam container filled with ice prior to microbiological analysis.
3.2.3. Nonthermal processing with CFMF
A 1.4 !m pore size microfiltration unit (MFS-1, Tetra Pak Filtration Systems,
Aarhus, Denmark) was applied for ‘cold’ pasteurisation of inoculated milk. Based on the commercial practice of CFMF application in the milk industry, a 1:10 ratio of retentate
(12 L/h) to permeate filtration membrane flow rate (120 L/h) was set for the milk processing with respective pressures of 90 and 20 kPa at the inlet and outlet of the membrane module. The active membrane surface area of the CFMF module was 0.2 m2 producing a transmembrane flux of 660 L/h m2. The CFMF milk processing temperature was constant at 35 °C and treated samples were cool ed prior to microbiological analysis.
3.2.4. Nonthermal hurdle processing
Both nonthermal technologies, PEF and CFMF, were combined for milk hurdle processing in two different sequences as shown in Figure 3-1. For PEF/CFMF processing, untreated milk was microfiltered under the conditions described above in subsection 2.3, and subsequently, PEF-treated and cooled prior to sampling using the conditions outlined in subsection 2.2. A reversed sequence for hurdle processing with
CFMF followed by PEF (CFMF/PEF) was also applied with the treatment conditions previously mentioned in subsections 2.2 and 2.3, respectively, in order to determine
)(! ! ! which configuration of the nonthermal technologies allowed the most effective ‘cold’ pasteurisation of milk.
3.2.5. Conventional heat pasteurisation
Milk was thermally pasteurised in a dual stage heat exchanger unit (UHT/HTST
Lab-25 EDH, Microthermics Inc., Raleigh, NC, USA). Inoculated milk entered the TP unit at 16 °C with a flow rate of 1.0 L/min and was heated to 75 °C, passing through a holding tube for 24 s, prior to cooling of the pasteurised product to 10 °C and subsequent sampling and analysis.
3.2.6. Microbiological shelf life
In order to provide optimum conditions for high shelf stability, separated skim milk was obtained from the same supplier as for the inactivation analysis (subsection
2.1), stored uninoculated for a short time under refrigeration prior to pre-heating (55 °C) for subsequent homogenization (NS2006H, GEA NiroSoavi S.p.A., Parma, Italy) at 50 and 200 bar for the first and second pressure stages, respectively. This was followed by
‘cold’ hurdle or heat preservation treatment of the homogenised product on the same day. The microbiological shelf life of homogenised milk treated with the most potent
PEF/CFMF approach (4.2 kV/mm, 612 !s, 815 kJ/L) or TP was investigated over a period of 35 days. Due to the high initial TAMC (FIL-IDF, 1974) of the homogenized raw milk (6.4 log10 CFU/ml) on the day of processing, which remained high following either of the milk preservation treatments (>3.0 log10 CFU/ml) and consequently, bringing about a very demanding processing scenario, microbiological shelf life of the processed
))! ! !
milk was considered to be expired after exceeding a limit of 5.0 log10 CFU/ml. This is in
line with reference microbial values for raw milk (Villar and others 1996). Milk samples
were taken in duplicate under sterile conditions before and after homogenization as well
as following the preservation treatments, and after refrigerated storage at 4±0.1°C for 0,
1, 7, 14, 21, 28, and 35 days. Extensive microbiological analyses were carried out
based on the spread-plating method for the determination of: a) TAMC on PCA as previously described in subsection 3.2.1; b) Total enteric bacteria counts (TEBC) on MacConkey agar (MCA; 211387, BD Difco,
Sparks, MD, USA), incubated at 35 °C for 24 h as de scribed in a study by Jocson and
others (1997); c) Total moulds and yeasts counts (TMYC) on potato dextrose agar (PDA; 254920, BD
Difco, Sparks, MD, USA), incubated at 35 °C for 48 h as described by the American
Public Health Association (1992); d) Total lactic acid bacteria counts (TLABC) on DeMan, Rogosa and Sharpe agar (MRS;
288210, BD Difco, Sparks, MD, USA), incubated at 30 °C for 48 h as described in a
study by Mennane and others (2007); e) Total staphylococci counts (TSC) on Baird Parker agar (BPA; 276840 + 277910, BD
Difco, Sparks, MD, USA), incubated at 37 °C for 24 h as described by Rajowkski and
others (1994).
In addition, milk samples were pour-plated with PCA for the analysis of: f) Total psychrotroph counts (TPC), incubated at 7 °C for 240 h as described in the
procedures of the FIL-IDF (1991);
)*! ! ! g) Total thermoduric psychrotroph counts (TTPC), incubated at 7 °C for 240 h based on
the procedural descriptions of FIL-IDF (1991) and Edgell and others (1950); h) Total thermoduric mesophilic counts (TTMC), incubated at 30 °C for 72 h as described
by Frank and others (1985); i) Total thermoduric thermophilic counts (TTTC), incubated at 55 °C for 24 h according to
the descriptions of the FIL-IDF (1991) and Murphy and others (1999).
Pour-plating was accomplished with liquid tempered PCA (18 ml portions, 45 °C),
which was poured on 1 ml of the neat or diluted milk sample, previously pipetted in a
sterile petri dish, and distributed by steady rotating and crosswise movements prior to
agar solidification. For thermoduric counts heat exposure of the neat and serial-diluted
milk samples placed in sterile test tubes in a water bath at 63 °C for 35 min preceded
the pour-plating.
3.2.7. Electrical conductivity and pH
Measurements of electrical conductivity and pH before and after milk processing
were carried out with a conductivity meter (Oaktron Con 11 Series, Thermo Fisher
Scientific Inc., Newlington, NH, USA) and a pH meter (Accumet AB15, Thermo Fisher
Scientific, Newlington, NH, USA), respectively.
3.2.8. Statistical analysis
Statistical analysis of experimental data was accomplished with Sigmastat
version 3.5 (Systat Software Inc., London, UK). For all treatments, data from three
different batches of milk (n = 3) were analysed using analysis of variance (ANOVA) with
*+! ! ! a level of statistical significance of + = 0.05. In the case of statistical differences, treatment means were compared with the Holm-Sidak test for multiple pairwise comparisons.
3.3. Results
3.3.1. Microbial inactivation with PEF, CFMF, CFMF/PEF, PEF/CFMF, and TP
Processing of inoculated skim milk with PEF, CFMF, TP, CFMF/PEF and
PEF/CFMF led to different reductions of native milk microorganisms as shown in Table
3-1. Application of a single processing technology for milk treatment led to the highest reduction for TP and CFMF (P%0.05); 4.6 and 3.7 log10 CFU/ml, respectively, while PEF led to lower microbial inactivation of 2.0, 2.1, 2.3 and 2.5 log10 CFU/ml for respective energy densities of 407, 632, 668 and 815 kJ/L compared to TP (P<0.05). Milk microflora reduction with PEF of 2.3 and 2.5 log10 CFU/ml similar to CFMF (P%0.05) was achieved exclusively with the two most potent PEF treatment conditions (668 kJ/L and
815 kJ/L, respectively). By contrast, a combination of CFMF and PEF for a hurdle treatment of milk led to microbial reductions ranging from 4.1 to 4.9 log10 CFU/ml, which was comparable to that of TP (P%0.05). An impact of increased electric field strength, treatment time and energy density was not observed when PEF or CFMF/PEF was used. However, a change of the hurdle processing sequence to PEF/CFMF enhanced the overall processing efficacy from 4.9 to 7.1 log10 CFU/ml (Table 3-1), thus, indicating an effect due to the change in PEF treatment parameters on the overall effectiveness of this hurdle processing approach). Moreover, the most effective PEF/CFMF treatment
*"! ! ! also achieved a significantly higher reduction of native milk biota than with CFMF, TP and CFMF/PEF (7.1 vs. 3.7, 4.6, and up to 4.8 log10 CFU/ml, respectively (P<0.05)).
Stand-alone PEF (up to 2.5 log10 CFU/ml) and single CFMF (3.7 log10 CFU/ml) processing accounted for an additive treatment effect of up to 6.2 log10 CFU/ml, consequently, the maximum inactivation of 7.1 log10 CFU/ml of native microorganisms in milk with PEF/CFMF hurdle processing marginally fell short of a synergistic treatment effect (P%0.05). No changes were observed between the pH of untreated milk (5.65) and PEF-treated (5.66), CFMF-treated (5.69) and TP-treated (5.74) milk (P%0.05). Initial electrical conductivity of milk (6.58 mS/cm) changed significantly after CFMF treatment
(2.31 mS/cm (P<0.05)) in contrast to PEF (6.50 mS/cm) and TP (6.65 mS/cm), which had no effect (P%0.05).
3.3.2.Microbial shelf stability of milk processed with PEF/CFMF and TP
The findings of the shelf life analysis for PEF/CFMF and TP-treated milk are summarised in Table 3-2. There was similar microbial shelf stability for ‘cold’ pasteurised milk using the most potent PEF/CFMF hurdle approach (4.2 kV/mm, 612
!s, 815 kJ/L) in comparison to thermally pasteurised milk. Due to a high initial TAMC of
6.4 log10 CFU/ml prior to homogenisation of the raw skim milk, which did not affect the
TAMC significantly (6.8 log10 CFU/ml (P%0.05)), a complete inactivation of the microbial population in milk was achieved neither by PEF/CFMF nor by TP, instead these treatments reduced microbial loads by 3.1 and 4.3 log10 CFU/ml, respectively. Similarly, initial TYMC and TLABC in untreated milk, 5.6 and 6.4 log10 CFU/ml, respectively, were reduced to 2.8 and 3.9 log10 CFU/ml following PEF/CFMF treatment and to 1.8 and 4.6
*#! ! !
log10 CFU/ml after TP, respectively. Inactivation below the microbiological detection limit
(1.5 log10 CFU/ml) was obtained for milk containing initial counts for TSC of 2.0 log10
CFU/ml, TPC of 1.7 log10 CFU/ml, and TTTP of 5.6 log10 CFU/ml, with both processing technologies. Raw milk TEBC of 3.8 log10 CFU/ml were reduced by TP below the detection limit while PEF/CFMF treatment led to a reduction in TEBC to 1.6 log10
CFU/ml. By contrast, an initial TTMC of 5.6 log10 CFU/ml was decreased below the detection limit after PEF/CFMF, whereas a reduction to 1.5 log10 CFU/ml was attained with TP. TTPC of untreated milk were originally found to be below the limit of detection.
In spite of slightly higher microbial reductions obtained directly after milk preservation with PEF/CFMF compared to TP, the microbiological shelf life of skim milk processed with either method was limited to 7 days of refrigerated storage.
*$! ! !
Table 3-1. Comparison of reductions in total mesophilic aerobic counts in skim milk obtained following pulsed electric fields (PEF), cross-flow microfiltration (CFMF), thermal pasteurisation (TP); combined CFMF/PEF, and PEF/CFMF treatment
Treatment Maximum Processing Microbial Reduction
Temperature (Log10 CFU/ml) (°C) 1 a PEF1.6//2105/407 40 2.0 1 a PEF2.0/1454/632 45 2.1 1 ab PEF3.0/983/668 47 2.3 1 ab PEF4.2/612/815 49 2.5 CFMF 35 3.7bc TP 75 4.6c 1 c CFMF/ PEF1.6//2105/407 40 4.1 1 c CFMF/ PEF2.0/1454/632 45 4.1 1 c CFMF/ PEF3.0/983/668 47 4.4 1 c CFMF/ PEF4.2/612/815 49 4.8 1 c PEF1.6//2105/407 /CFMF 40 4.9 1 cd PEF2.0/1454/632 /CFMF 45 5.3 1 cd PEF3.0/983/668 /CFMF 47 5.7 1 d PEF4.2/612/815 /CFMF 49 7.1 P2 - *** SEM3 - 0.26 a,b,c,d Different superscripted letters in the column indicate statistical significance between the two means. 1 Subscripted numbers: before “/”, describes the electric field strength [kV/mm] of PEF; between two “/”, specifies the treatment time [!s] of PEF; after “/”, denotes the specific energy density [kJ/L] induced by PEF. 2 P stands for the statistical probability.3 SEM abbreviates the standard error of the mean.*** refers to a statistical significance of P<0.001. Microfiltration was conducted using a 1.4 !m pore size and a 660 L/h m2 transmembrane flux, thermal pasteurisation at 75 °C x 24 s; combined MF/PEF, and PEF/MF treatment.
*,! ! !
For PEF/CFMF-treated milk, TAMC, TEBC, TMYC, and TPC exceeded the shelf life limit of 5.0 log10 CFU/ml after 7 storage days accounting for respectively 8.6, 7.6,
5.1, and 7.5 log10 CFU/ml on day 14 of the shelf life analysis. It was noticeable that the
TAMC, TEBC and TPC followed the same microbial growth pattern over the duration of the analysis (P%0.05) thus, suggesting that one dominant microorganism, which based on the observations of the TEBC was non-lactose fermenting and seemed to be able to grow rapidly under psychrotrophic conditions, accounted for the highest resistivity and could have been the overall limiting factor for the PEF/CFMF shelf stability. By comparison, shelf life of milk processed with TP expired when the TAMC passed over the limit after 7 days of storage and rose slightly to 5.2 log10 CFU/ml on day 14. Over the duration of the shelf life analysis TLABC, TSC, TTPC, TTMC and TTTC in
PEF/CFMF-processed milk did not go beyond the limit of microbial stability and safety and at the same time TP-treated milk remained also shelf stable under refrigeration for
35 days with regard to TEBC, TYMC, TLABC, TSC, TTPC, TTMC and TTTC.
The previous results for pH and electrical conductivity (subsection 3.1) were confirmed for the first part of the shelf life analysis (i.e. pre and post processing). In addition, stability of the pH (from 6.19 on day 0 to 6.29 on day 35) and of the conductivity (from
2.06 mS/cm on day 0 to 2.72 mS/cm on day 35) was observed in ‘cold’ preserved milk with PEF/CFMF over the duration of the shelf life analysis (P%0.05) as well as for pH
(from 6.94 on day 0 to 6.73 on day 35) and conductivity (from 5.50 mS/cm on day 0 to
5.52 mS/cm on day 35) in conventional, TP-treated milk.
*%! ! !
Table 3-2. Microbial counts in skim milk before and after pulsed electric fields (PEF) and cross-flow microfiltration (CFMF) hurdle treatment (PEF/CFMF) or thermal pasteurisation (TP) during storage
#$%&'(!)&*+! 0%.-$1(!2&)(! 67&%&$*!8.97%:! @A=?B=:! 2@:! )&,-./&.%$! 345! 3*.1;?)*5! 3*.1;?)*5! 3*.1;?)*5! /0123!2450678!! +! &9,:;+9(0?@734>! "! ! $9":;+9#! +! %9&:;+9#1>! "! ! #9%:;+9,12?@F3080887! +! #9+:;+9"F8@50150?@>! +! "9(:;+9#
*&! ! !
/0123!1@45=0ED578! +! B"9%! B"9%! B"9%! ?>F8@50150?@>! "! ! B"9%! B"9%! ! (! ! B"9%! $9(:;+9"0?@734>! "! ! ,9":;+9+! "! ! B"9%! #9*:;+9"
*(! ! !
3.4. Discussion
In the present study similar inactivation of native milk biota in bovine skim milk to that observed for TP was achieved when CFMF was applied individually or followed by
PEF. Research conducted on pre-filtered ovine (Beolchini and others 2004) and bovine milk (Beolchini and others 2005) microfiltered with 1.4 !m pore diameter CFMF system using an average permeate flux of approximately 240 L/h m2 suggested respective microbial reductions of more than 3.0 (at 40 °C) an d 2.0 (at 40 °C) to 3.0 (30 °C) log 10
CFU/ml, respectively, which is slightly lower than the findings of this study. A possible reason for the lower antimicrobial impact may have been the use of less sophisticated
CFMF equipment than the Tetra Pak unit used in our study. The latter features uniform transmembrane pressure (Bird 1996), which can be of considerable impact during processing. By contrast, Elwell and Barbano (2006) obtained results in agreement with those of the present study; reporting a mean reduction of total bacteria of 3.8 log10
CFU/ml in raw skim milk after CFMF processing with 1.4 !m pore size at 50 °C using permeate and retentate flow rates of 462 and 25 L/h, respectively. When the same researchers (Elwell and Barbano 2006) thermally pasteurised the CFMF permeate at 72
°C for 16 s following commercial practice, the bact erial reduction was increased by another 1.8 log10 CFU/ml, which overall is comparable to the microbial reduction for hurdle processed milk in this study but 1.5 log10 CFU/ml less than the inactivation achieved with our most effective nonthermal pasteurisation using PEF/CFMF at 4.2 kV/mm for 612 !s with 815 kJ/L. Possible explanations for an increased effectiveness of
PEF/CFMF over CFMF/PEF could be a PEF-induced change in the steric configuration
*)! ! ! of electroporated microorganisms (García and others 2007, Wang and others 2008) or an increased agglomeration between microorganisms or between microorganisms and milk components (Hodgson and others 1993, Jimenez-Lopez and others 2008, Zhao and others 2009) due to electrical charge polarisation after PEF and prior to CFMF. This may improve the filtration effect and, thus, enhance the overall process efficacy and microbial reduction in the treated product. Similar inactivation of native microbiota in skim milk by CFMF of 4.0 log10 CFU/ml and more was also reported in other studies
(Payfylias and others 1996, Guerra and others 1997, Fritsch and Moraru 2008). The majority of the findings on conventional heat pasteurisation of milk reported in the literature (Sepúlveda-Ahumada and others 2000, El-Hadi Sulieman 2007, and
Odriozola-Serrano and others 2006) were achieved at considerably lower initial TAMC
(4.4, 4.7, and 3.2 log10 CFU/ml, respectively) than those of the present study, therefore, resulting in lower respective microbial reductions of 0.6 (72 °C for 15 s), 1.9 (65 for 30 min), and approximately 2.2 (75 °C for 15 s) than t he 4.6 log10 CFU/ml obtained for TP in this study. Nonetheless, there are also studies quoting higher initial counts of 3.6 and
6.7 log10 CFU/ml, which correspond to comparable (Aaku and others 2004) and higher
(Walkling-Ribeiro and others 2009, using 72 °C for 26 s) reductions of TAMC in TP- treated milk. Nonthermal milk processing with PEF below temperatures of 50 °C has led to broadly comparable reductions in total counts of native microorganisms of up to 1.4,
2.0 and 2.2 log10 CFU/ml as reported by Floury and others (2006) (< 50 °C, 3.5 kV/mm,
188 !s), Fernández-Molina and others (2005) (<32 °C, 3.6 kV/mm, 84 !s), and Otunola and others (2008) (4.0 kV/mm, 360 !s), respectively, which is in accordance with the findings for the stand-alone PEF treatment of skim milk described in this study. Higher
**! ! !
and lower microbial inactivation of 5.4 and 1.2 log10 CFU/ml of native flora in milk PEF- treated under ‘cold’ processing conditions have also been reported by Li and others
(2003) (<50 °C, 4.1 kV/mm, 54 !s) and Sepúlveda-Ahumada and others (2000) (<30
°C, 3.5 kV/mm, 30 pulses), respectively. These diff erences are possibly the result of more or less effective PEF treatment conditions and PEF equipment, and variations in the microbial composition of the milk. Likewise, substantial volatility in the reduction of total aerobic counts in PEF-processed milk of 1.3, 2.4, 2.7, 3.0, and 6.0 log10 CFU/ml was obtained when PEF was applied alone, or in combination with a heat treatment, at moderate processing temperatures of 50 °C or above as indicated by results of different research groups, including Smith and others (2002), using 52 °C, 8.0 kV/mm, 100 !s;
Shamsi and others (2008), using 60 °C, 3.5 kV/mm, 1 9.6 !s; Michalac and others
(2003), using 52 °C, 3.5 kV/mm, 188 !s; Fernández-Molina and others (2005), using
PEF at 3.2 kV/mm for 84 !s and post-heating at 60 °C for 21 s; and Walkling -Ribeiro and others (2009), pre-heating at 50 °C for 60 s, a nd using PEF at 4.0 kV/mm for 60 !s; respectively. Exposing the microorganisms to moderate (%50 to 65 °C) instead of the
‘cold’ (<50 °C) PEF/CFMF milk hurdle pasteurisation used in this study (by either incorporating an additional heating step in the processing sequence or application of more powerful PEF settings that cause a higher concurrent dissipation of heat during product treatment) is likely to increase the overall effectiveness of the PEF/CFMF hurdle technology even further and achieve a synergistic effect. However, a decrease in nutritional and sensory quality of the processed milk may occur depending on the heat impact (De Luis and others 2009). However, the temperatures achieved by the PEF treatment could still have a considerably lower effect on the product quality due to lower
"++! ! ! temperature levels and much shorter treatment times in the !s range by comparison to those minimally required by law for TP (72 °C, 15 s ).
The effect of single TP (72 °C, 15 s) and in combin ation with CFMF (1.4 !m, 200
L/h m2) on the microbiological shelf stability of milk stored at 4 °C with a bacterial limit of
4.3 log10 CFU/ml was investigated by Elwell and Barbano (2006), who observed longer shelf life of 29 d for CFMF/TP than for TP-treated milk (15 d). Malmberg and Holm
(1998) also stated that addition of CFMF after conventional pasteurisation can potentially increase overall shelf life of milk by 2 to 3 times, from 6 to 8 days up to 16 to
21 days before expiry. The former shelf life range of conventionally pasteurised milk coincided with the shelf life of 7 days found for both TP and PEF/CFMF-treated milk in the present study, which, in turn, is broadly in line with the findings of Odriozola-Serrano and others (2006). The latter study described milk that was microbiologically stable for 5 days at 4 °C following TP and PEF treatment at 75 ° C, 15 s and 3.6 kV/mm for 1000 !s, respectively. Fernández-Molina and others (2005) generally showed a 7 day shelf life
(<5.0 log10 CFU/ml) of refrigerated skim milk after PEF treatment when using 2.8 to 3.6 kV/mm for 84 !s, which corresponds to the results of this study, but combination with subsequent heat treatments at 60 or 65 °C for 21 s led to a significant increase of product shelf stability of up to 30 days at 4 °C st orage. In more comprehensive studies by Walkling-Ribeiro and others (2009) and Sepulveda and others (2009) the impact of specific microbial groups on the microbial shelf life of milk at 4 °C was studied following treatments with combined heat (50 °C, 60 s) and PEF (<55 °C, 4.0 kV/mm, 60 !s) or TP
(72 °C, 26 s), and heat-generating PEF (65 °C, <10 s; 3.5 kV/mm, 13 !s) or TP (72 °C,
15 s), respectively. The first group of researchers (Walkling-Ribeiro and others, 2009)
"+"! ! ! found that heat- and PEF-treated milk was stable for the entire 21-day period of the microbial shelf life analysis in contrast to TP-treated milk, which expired after 14 days of storage. As the investigated microbial groups of lactobacilli, pseudomonads, coliforms and non-lactose fermenters, and staphylococci showed generally no significant or detectable growth after either preservation treatment, psychrotrophs other than pseudomonads and endo-sporeforming bacteria were suggested as comprising the milk shelf life limiting microbiota (e.g. Streptococcus, Arthrobacter, Corynebacterium,
Microbacterium, and Bacillus subspecies). The results for PEF preservation obtained by the other research group (Sepulveda and others, 2009) were different suggesting that the microbial shelf life limit of milk treated with heat-generating PEF was exceeded after
22 days while TP-treated milk was shelf stable for 44 days. The growth of mesophiles exceeding the limit after 44 days was identified as a shelf life limiting factor and not enterobacteria or psychrotrophs. The discrepancy in shelf stability of milk between the results obtained for PEF/CFMF and for TP in the present study and the findings for heat-assisted PEF or TP treatments described in the literature could be explained by the substantially higher initial load of microorganisms prior to milk processing and the difference in the composition of the microbial milk populations, which may also be a reason for the deviations with regard to the most recalcitrant groups of native microbiota limiting the overall shelf life of hurdle preserved or conventionally pasteurised milk. The decline in electrical conductivity following CFMF was attributed to a reduction in nutrient content (Brans and others, 2004; Guerra and others, 1997; Jimenez-Lopez and others,
2008).
"+#! ! !
Combining PEF and CFMF in a hurdle technology for ‘cold’ pasteurisation of milk proved to be more effective than conventional TP and even higher antimicrobial efficacy of PEF/CFMF for longer preservation of the microbial milk shelf life can be expected if the processing temperature is increased to a moderate level (%50 to 65 °C), which will have a greater impact on total aerobic counts, in particular, enteric bacteria, psychrotrophs and moulds and yeasts. Due to the promising results obtained in the present study and the high commercialization potential, further research on the optimization of this nonthermal hurdle technology with a focus on the specific requirements for different milk products (moderately heat-assisted PEF/CFMF processing for higher microbial safety and shelf stability vs. ‘cold’ PEF/CFMF processing for higher nutritional and sensory quality) is recommended. In addition, the interactions or mechanisms at the molecular level following PEF milk treatment, which were responsible for an increased antimicrobial efficacy of the PEF/CFMF over the
CFMF/PEF processing sequence, merit in-depth analysis.
"+$! ! !
4. Resistance of temperature, pH, osmotic and starvation stress induced
Escherichia coli O157:H7 to pulsed electric fields in milk
4.1. Introduction
The injury and the repair of microorganisms after inactivation with PEF has been studied to gain an insight into the shelf-life of products treated with this using technology
(Aronsson and others 2004, García and others 2006). Although similar studies have been carried out on other processing technologies (Hayman and others 2007, Teo and others 2001), few studies have examined the exposure of microorganisms to sub- optimal growth conditions and their subsequent resistance to PEF inactivation.
Exposure of microorganisms to challenging environmental conditions can trigger stress response mechanisms, including the expression of genes related to the production of shock proteins (Chung and others 2006, Hill and others 2002). A study by
Lou and Yousef (1996) showed that when Listeria monocytogenes was previously exposed to sub-lethal levels of heterologous stresses an increased resistance to heat was obtained. This phenomenom has been referred to as a cross-protective effect.
Cross-protection is similar to the expression of cross-resistance among chemicals, and in relation to food processing it has been reviewed by several authors (Rodríguez-Romo and Yousef 2005, Juneja and Novak 2003, Johnson 2003, Ravishankar and Juneja
2003).
When studying cross-protection, Rowe and Kirk (1999) emphasized the importance of analyzing the inactivation kinetics by means of accurate models that can be incorporated into an appropriate risk assessment. In an attempt to develop a
"+,! ! ! mechanistic approach to monitor sub-lethal injury Gawande and Griffiths (2005a) used constructs of Escherichia coli O157:H7 in which the promoters for genes encoding major stress response proteins were fused to a green fluorescent protein (GFP) gene.
They were then able to identify conditions that led to the expression of these stress response proteins and relate them to the ability of the organism to survive subsequent stresses (Gawande and Griffiths 2005b).
The dairy industry has incorporated several predictive models to describe microbial growth and inactivation (Griffiths 1994). In general, growth models have been evaluated for their ability to predict lag phase and growth rates based on viable counts and turbidimetric data (Dalgaard and others 1994; Dalgaard and Koutsoumanis 2001,
McKellar and Lu 2004b). The Baranyi model (Baranyi and others 1993; Baranyi and
Roberts 1994; Baranyi and others 1995) has been widely used because of the availability of virtual resources that ease the calculation of growth parameters, and has been proposed to describe inactivation kinetics (Ngadi and Dehghannya 2010). Lag phase and growth rates are the parameters most greatly affected by changes to temperature or pH in secondary models (Zwiettering and others 1991; Rosso and others
1995; Juneja and others 2007; Juneja and others 2009). Application of these models has led to innovations in predictive environmental and food microbiology, as well as quantitative risk assessment (McMeekin and others 2008).
Due to the non-linear nature of microbial inactivation, and the requirement of a multi-parametric approach for modelling pulsed electric field (PEF) treatments, the
Weibull distribution has been successfully applied to describe PEF inactivation (Rodrigo
"+%! ! ! and others 2003, García and others 2005), and further modifications of this model were implemented to enable a comparison of microbial resistance using a zPEF value (Álvarez and others 2003a, Álvarez and others 2003b, Rodríguez-Calleja and others 2006). With the development of the WeLL (Weibullian-log logistic) model microbial inactivation that allows the variation of the inactivation rate due to temperature and other lethal agents to be determined, the effect of dynamic conditions during microbial inactivation can now be predicted (Peleg and Cole 2000, Fernández and others 1999, Campanella and Peleg
2001, Peleg and Normand 2004).
The objective of this research was the mathematical analysis of the growth, stress response and inactivation of a biological indicator (E. coli O157:H7) by using predictive models. Escherichia coli O157:H7 was subjected to different stress factors, simulating characteristic dairy processing conditions, and subsequently exposed to
PEF. The data were analyzed using a variety of modelling techniques.
!
4.2. Materials and Methods
4.2.1. Microorganism
An E. coli O157:H7 ATCC 43888 (American Type Culture Collection, Manassas,
VA, USA) was chosen as the model microorganism due to the absence of the genes that encode shiga toxins I and II (Stx1-, Stx2-).
"+&! ! !
4.2.2. Growth medium preparation
Bacterial growth limits were evaluated in two broths: lactose broth (LACB;
DIFCOTM Lactose Broth, 211835, Becton Dickinson and Company, Sparks, MD, USA), and tryptic soy broth (TSB; BACTOTM Tryptic Soy Broth without dextrose, 286220) supplemented with 0.5% lactose (102137, D-(+)-lactose monohydrate, MP Biomedicals,
Solon, OH, USA) to make up TSBL.
The growth of the bacterium was also analyzed in modified TSBL. The medium pH of 7.0 was adjusted by using lactic acid (LA; A162, Fisher Scientific, Fairlawn, NJ,
USA) or acetic acid (AA; A35, Fisher Scientific, Fairlawn, NJ, USA) to final pH values of
6.0, 5.0, and 4.0, which was monitored with a pH meter (AB 15 Accumet Basic, Fisher
Scientific, Hampton, NH, USA). To determine the effect of water activity, sodium chloride (NaCl; S7653, Sigma-Aldrich Co., St. Louis, MO, USA) was added to final concentrations of 5.0%, 7.5%, or 10.0% weight by volume including the initial medium
NaCl content of 0.5%. The addition of NaCl resulted in water activity values of 0.999,
0.946, 0.920, and 0.894, respectively, as analyzed in a water activity meter (AquaLab model CX-2, Decagon, Pullman, WA, USA).
For the generation of the growth curves 1:100 of a stationary phase E. coli
O157:H7 population, incubated at 35 °C for 24 h, wa s used to inoculate LACB or TSBL placed in a 4 x 6 well flat bottom microplate (Corstar® 3526, Corning Incorporated, NY,
USA) at 2.0 ml per well, and incubated at temperatures of 5, 7, 12, 28, 35, 45, and 48
°C .
"+(! ! !
4.2.3. Construction of stress reporters
Plasmids containing the gfpuv reporter gene fused to the respective promoters uspA (universal stress protein A), rpoS (RNA polymerase sigma factor), and grpE
(glucose-regulated protein) were extracted from E. coli O157:H7 C9490 obtained from the culture collection of the Canadian Research Institute for Food Safety (strain ID: C
476). A QIA prep® spin miniprep kit 50 (27104, Qiagen, Germantown, MA, USA) was used to obtain plasmid concentrations between 70 and 150 ng/!L as estimated spectrophotometrically (Nanodrop 1000, Thermo Fisher Scientific Wilmington, DE,
USA). These plasmid preparations were kept frozen at -20 °C until used. Exponential phase E. coli O157:H7 (ATCC 4388) cells were harvested after 7 h of growth at 37 °C in
Luria-Bertani broth (LB; Miller, 244620, Becton Dickinson and Company, Sparks, MD,
USA), and prepared for electroporation by dilution in ice-cold distilled water containing
10% glycerol. The cell suspensions (25 !L) were combined with the plasmids (5 !L) in a
0.2 cm gap micro pulser cuvette (165-2086, Bio-Rad Laboratories, Hercules, CA, USA) and one exponential decay pulse of 2.5 V of 2480 ms (25 !F, 200 &) generated by a
Gene Pulser Xcell electroporation system (Bio-Rad Laboratories, Hercules, CA, USA) was applied as in the pre-set bacterial cell protocol (Bio-Rad, 2010). After stabilizing in
S.O.C. medium (15544-034, Invitrogen, Carlsband, CA, USA) the transformants were selected by growing the cells in LB broth supplemented with 50 !g/ml of ampicillin sodium salt (BP1760, Fisher Scientific, Fairlawn, NJ, USA).
All of the cells for biosensor analysis were grown at 35 °C for 24 hours to reach the stationary phase in LB. Starvation stress was induced after centrifuging (Avanti J-20
"+)! ! !
XPI centrifuge, Beckman Coulter, Brea, CA, USA) the cell suspension at 5000 × g-1 for
5 min at 4 °C, re-suspending it to 1:10 of its orig inal volume with filtered (0.22 !m pore size) distilled water, and incubating the bacteria at 25 °C for 100 hours. Subsequently, the samples were transferred to microplates and examined as detailed in Section 4.2.4.
To induce cold shock the cell suspension was cooled to 5 °C and 1.5 ml of the suspensions were transferred into micro-centrifuge tubes (05-408-129, Fisher Scientific
Company, Hampton, NH, USA) every 24 h, pelleted using a microcentrifuge
(Spectrafuge 16M, Labnet International, Woodbridge, NJ, USA) and re-suspended to
1:10 of its original volume with Ringers solution (BR0052, Oxoid Ltd. Basingstoke,
Hampshire, England) for a final volume of 150 !L per sample.
4.2.4. Monitoring of the physiological state
Absorbance and fluorescence were both measured with a 1420 Multilabel counter Victor3V (Perkin Elmer, Turku, Findland). Duplicate absorbance (450 nm) measurements were recorded after placing the cells in 24-well micro-plates, with 1.0 s of shaking in between readings. Fluorescence was measured in duplicate after placing the cells in an 8 x 12 well black with clear bottom assay plates (Corstar® 3603, Corning
Incorporated, NY, USA) at wavelengths of 485 nm for excitation and 535 nm for emission. For each stress response promoter::gfp construct relative fluorescence intensity (RFI) and response ratios (R) (derived from Zhang and Griffiths 2003;
Gawande and Griffiths 2005ab) were estimated following Eq. 4-1 and Eq. 4-2, respectively.
"+*! ! !
153 "653 /653 Eq. 4-1! ¬≠B√ƒ <¬≠B√ƒ <≈<∆M
1 "153 Q153 Eq. 4-2 î«´»… qî«´»… .î«´»…
4.2.5. PEF treatment
For PEF treatment 104 to 107 CFU/ml of microbial cells grown as described in
Section 4.2.2, and presented in Table 4-2, were used for inoculation (5 % of the total volume) of 2% micro-filtered milk (Lactantia P,r Filtre, Parmalat SpA, Montreal, QC,
Canada) and pumped (Masterflex pump drive 7524-40 and pump head 77201-60, Cole
Parmer Instrument Co., Vernon Hills, IL, USA) continuously through silicone tubing to a
PEF chamber. Before entering the chamber the inoculated milk passed through a stainless steel coil submerged in a water bath (Isotemp 10L, Fisher Scientific, Hampton,
NH, USA) to pre-heat the product to an initial temperature of 30 °C at the chamber entrance. Exponential decay shaped electric pulses were applied in a co-axial treatment chamber consisting of one vertical rod shaped ground electrode and one ring-shaped stainless steel charging electrode centered on the ground, and mounted in a circular ring-shaped metal segment. The latter was threaded into plastic segments for insulation. The PEF treatment chamber has been previously described and applied by
Walkling-Ribeiro and others (2010) but for the present study only one ring shaped electrode was charged. Between the electrodes a gap of 2.5 mm was maintained, and the effective electrode treatment area and volume were 49 cm2 and 3.52 cm3, respectively. The 1.5 !s width pulses were generated in a PEF unit designed at the
University of Waterloo (PPS 30, University of Waterloo, Waterloo, ON, Canada). The temperature of the product leaving the chamber outlet was between 32 to 57 °C and
""+! ! ! was immediately cooled to 10 °C by passing it throu gh coiled tubing submerged in a refrigerated water bath (NESLAB RTE-7, Thermo Scientific, Newington, NH, USA).
Treatment conditions comprised an electric field between 0.8 and 2.8 kV/mm, frequencies of 15 or 25 Hz and flow rates of 1.8 or 3.0 L/h. Treatment times were calculated using the product flow rate measured at the product outlet, which was adjusted between 10.8 and 1.8 L//h. Three setups treatment conditions were used: 1)
1.8 kV/mm and 20 Hz at 9.0, 7.5, 6.0, 4.5, 3.0 and 1.5 L/h; 2) 2) 1.8 kV/mm and 20 Hz at 10.8, 9.0, 7.2, 5.4, 3.6, and 1.8 L/h, and 3) 2.4 kV/mm and 20 Hz at 10.8, 9.0, 7.2,
5.4, 3.6, and 1.8 L/h. The specific energy was estimated based on the charging capacitance for exponential decay pulses as indicated by Zhang and others (1995).
4.2.6. Microbial counts
Serial dilutions of untreated and treated samples were made in Ringers solution, and 50 !L of the appropriate dilution were spread on the surface of milk agar plates
(CM0681, Oxoid Ltd. Basignstoke, Hampshire, England) using a Spiral Plater (Spiral
Btotech, Bethesda, MD, USA). The plates were incubated at 35 °C for 48h, and counts determined (Laird and others 2004).
4.2.7. Mathematical Analysis
Mathematical analyses were conducted using Mathematica® 7 version 7.0.1.0
(Wolfram Research, Inc. Champaign, IL, USA), and the DMFit Microsoft® Excel add-in
(Institute of Food Research, Norwich UK), and statistical analysis with R 2.12.0 (R
"""! ! !
Foundation for Statistical Computing, Vienna, Austria) in conjunction with R
Commander 1.6-1 (McMaster University, Burlington, ON, Canada).
Growth rates and lag phases were calculated using the Baranyi model (Baranyi and Roberts 1994) as explained by Eqs. 4-3 and 4-4, and correlated to pH and water activity (aw) with the Ratkowsky model (Eq. 4-5) and to temperature using the Cardinal model (Eqs. 4-6 to 4-8) because its goodness-of-fit and capabilities to estimate optimum growth conditions. Stress response was modeled using the Logistic (Eq. 4-9), Gompertz
(Eq. 4-10), and Weibull growth (Eq. 4-11) models.
The Weibull model was fitted using the natural logarithm of the surviving population (Couvert and others 2005; Eq. 4-12), and the mean treatment time was calculated (Eq. 4-13; Fernández and others 1999) as an indicator of microbial sensitivity or resistance based in the model distribution. The WeLL (Weibullian-log logisitc) model
(Eqs. 4-14 and 4-15) was applied to account for PEF generated heat (Corradini and
Peleg 2009, Peleg and Normand 2004, Campanella and Peleg 2001 (Table 4-1).
Several mathematical models can be used to describe inactivation non-linearity (Juneja and Marks 2003), in this study only the Weibull model was considered because of its flexibility to describe PEF inactivation, and the availability of studies that provide comparable data (Alvarez and others 2003, Rodríguez-Calleja and others 2006). The goodness-of-fit was compared based on the Akaike Information Criterion (AIC) and the
Bayesian Information Criterion (BIC) as described by Juneja and others (2007, 2009), and the means of the parameters for each treatment were compared using one or multi- way ANOVA and Tukey statistical tests.
""#! ! !
Table 4-1. Mathematical models used to analyze Escherichia coli O157:H7 during stress induced growth and its resistance to subsequent pulsed electric field processing in milk.
Microbial Model Equation Adapted from activity Ngadi and !"#$%& ' !"#( ) *+,-. / !$%&0 Eq. 4-3 Dehghannya (2010), Aronsson Growth 12$345&/6783)**9:;?@/A$B&C 1305 9:;D/$AEF>?@GAEFH&C0IJ Baranyi and others (2004), (time) McKellar and Lu :G$L/B&C !$%& ' % ) 2$35K& / 678*9 ) M(053 ) M(IJ Eq. 4-4 (2004b), Baranyi and Roberts (1994)
R Eq. 4-5 Zwiettering and Ratkowski +,-. ' N / $O 1 O,PQ& / $ST 1 ST,PQ& / *UV 1 UV 0 ,PQ others (1991)
Eq. 4-6 Rosso and others +,-. ' +WXB / Y$O& / Z$ST& / [$UV& (1998)
""#! Growth O]O (Environment) ,PQ Eq. 4-7 Y$O& '\O ]O]O ^ Cardinal ,PQ ,-. O,-. ] O
$O 1 O &R / $O 1 O & O ] O ] O _ ,PQ ,-. ,PQ ,-. *O 1 O 0 / 8*O 1 O 0 / *O 1 O 0 1 *O 1 O 0 / *O ) O 1 `O0I Eq. 4-9 WXB ,PQ WXB ,PQ WXB WXB ,-. WXB ,PQ
:Gb/$BGBc&C Eq. 4-11 McKellar and Lu Logistic a$%& ' a,PQ ) *$a,-. 1 a,PQ&5*3 ) 9 00 (2004b)
Eq. 4-12 McKellar and Lu eGf8g/*hihc0Ij Stress Gompertz a$%& ' a ) d$a 1 a & / 9 k (2004b) response ,PQ ,-. ,PQ
> Weibull a$%& ' a 1 2$a 1 a & / 9:G$l/B& CJ Eq. 4-13 López and others ,-. ,-. ,PQ (2004)
! ! !
!"#$%&'( ) *%&,+'- Eq. 4-14 Peleg and Penchina (2000), Mafart and Weibull Eq. 4-15 others (2002), & ) + / 0%1 2 345' . Fernández and Inactivation others (1999) WeLL ; (Weibullian !67#$%&'( ) *8%9' / : Eq. 4-16 Corradini and Peleg Log- (2009),Peleg and Eq. 4-17 Normand (2004) logistic) 8%9' ) !"<1 2 =#>/%?4?@'(A †a can be denoted as the time for the first decimal reduction if n = 1. ‡Curves can be described as convex with n>1 or concave with n<1.
""#!
! ! !
4.3. Results and discussion
4.3.1. Growth environment
Tryptic Soy Broth with added lactose (TSBL) allowed the growth to a higher concentration (ABS450) of E. coli O157:H7, and higher challenging temperature conditions (7 to 45 °C) compared to the lower nutri ent density medium LACB (10 to 45
°C). In Figure 4-1A the Cardinal model was used to describe the variation in the growth rates between 7 and 45 °C. The analysis indicated o ptimum growth temperatures at
31.7 and 27.5 °C, and maximum growth rates of 0.03 and 0.07 h-1 when the organism was grown in LACB (Adjusted-R2 of 0.912; standard error of fit of 0.003) or TSBL
(Adjusted-R2 0.980; standard error of fit of 0.004), respectively. When growth rates in media in which the pH was adjusted using AA or LA were analyzed (Figure 4-1B), the minimum pH for required growth was 5.2 (Adjusted-R2 of 0.992; standard error of fit of
0.011) and 4.8 (Adjusted-R2 0.988 and standard error of fit of 0.013), respectively. The minimum aw for growth was estimated to be 0.915 (Adjusted-R2 of 0.974 and standard error of fit of 0.016) as shown in Figure 4-1C. The presence of lactose has been shown to enable E. coli to grow at temperatures as high as 44.5 °C, and )-galactosidase production was increased in the presence of high NaCl (10%) cocentrations (Anderson and others 1979). Also TSB supplemented with 0.5% lactose was reported to increase the recovery of heat injured (at 55.5 °C for 50 min ) L. monocytogenes F5069, similar to the case of glucose and sucrose supplements (Busch and Donnelly 1992).
""%! ! !
0.10
0.08
TSBL LACB A 0.06
(1/s) Series3
max Series4 !
0.04 Growth rate,
0.02
0.00 01020304050 Temperature ("C)
0.10
0.08 (1/s)
max 0.06 LA ! AA LA Fitted
0.04 AA Fitted Growth rate, rate, Growth
0.02 B
0.00 3.0 4.0 5.0 6.0 7.0 8.0 pH
(Continued)
""&! ! !
% NaCl 10 86420 0.10
0.08 (1/s)
max 0.06 !
Awaw Aw fitted 0.04 Growth rate,
0.02 C
0.00 0.875 0.900 0.925 0.950 0.975 1.000 Water activity (a ) w !
Figure 4-1. Growth rates of Escherichia coli O157:H7 in various growth media, predicted by the Cardinal (A = temperature; B = acid) and Ratkowski (C = NaCl) models. LACB = lactose broth (gray circle), TSBL = tryptic soy broth with 0.5% added lactose (black circle), AA = acetic acid (black circles), LA = lactic acid (grey circles), NaCl = sodium chloride, aw = water activity (black circles).
""(! ! !
4.3.2. Stress response
The analysis of the trends pertaining to the response ratios (R) of the stress reporters construct’s showed similar goodness-of-fit for the three models and the three promoters, with the exception of a significantly higher AIC (P < 0.05) observed in the logistic model of the rpoS gene response (Table 4-2). Because AIC and BIC values close to zero indicate a good fit, the Weibull growth model may represent the best choice for future studies. Overall, high correlation coefficients (adjusted R2) were observed for all of the investigated models and treatments. In starved cells the predicted
Rmax for grpE had the lowest stress RR (P < 0.05), in contrast to cold shocked cells in which the lowest stress RR (P < 0.05) was observed for rpoS indicating that because its higher RR’s uspA may be the best promoter to monitor microbial stress (Table 4-2). The slower response to cold shock was reflected by lower reaction rates (r) or scale parameters (g) in the Weibull model. In Figure 4-2 mechanistic and probabilistic models are presented, which explain the response ratio of the uspA and rpoS gene promoters.
The general stress response gene regulator (rpoS) which encodes the sigma factor ("S) is the most studied with regard to stress response in Gram negative bacteria
(Chung and others 2006, Hill and others 2002). For example, Rowe and Kirk (1999) removed the rpoS gene from E. coli O157:H7 to create a mutant to study the effect of the exposure to acid conditions on the cross-protection to NaCl, and Leenanon and
Drake (2001) compared a normal E. coli O157:H7 with an rpoS mutant to study cross protection resistance to heat and freeze-thawing observing higher resistances in the cross protected normal E. coli O157:H7.
"")! ! !
2 Table 4-2. Models goodness-of-fit (AIC, BIC and adjusted R ), and coefficients (Rmax, r, ti, g, m), resulting from analysis of fluorescence response of Escherichia coli O157:H7 stress gene promoters after induced stress (starvation or cold-shock).
Biosensor Starvation Cold Shocked (stress gene Estimates reporter Gompertz Logistic Weibull Gompertz Logistic Weibull plasmid) !" !" ! # # # puspA::gfpUV! 9.9 ! 18.3 ! -2.8 ! -2.7 ! 2.3 ! -6.0 ! !" " !" # # # prpoS::gfpUV! AIC 19.1 ! 25.6 ! 4.2 ! 4.3 ! 4.9 ! 1.8 ! !" !" !" # # # pgrpE::gfpUV! 3.3 ! 3.4 ! 2.1 ! 7.4 ! 7.4 ! 6.5 ! $ $ $ % % % puspA::gfpUV! 12.4 ! 20.9 ! -0.3 ! -2.6 ! 2.4 ! -5.8 ! $ $ $ % % % prpoS::gfpUV! BIC 18.4 ! 28.1 ! 6.8 ! 4.6 ! 5.2 ! 2.2 ! $ $ $ % % % pgrpE::gfpUV! 4.6 ! 4.7 ! 3.5 ! 7.4 ! 7.4 ! 6.6 ! & & & ' ' ' puspA::gfpUV! 0.974 ! 0.974 ! 0.998 ! 0.994 ! 0.993 ! 0.992 ! ""#! 2 & & & ' ' ' prpoS::gfpUV! Adjusted R 0.995 ! 0.991 ! 0.998 ! 0.993 ! 0.994 ! 0.977 ! & & & ' ' ' pgrpE::gfpUV! 0.993 ! 0.986 ! 0.994 ! 0.989 ! 0.989 ! 0.993 ! ( ( ( ) ) ) puspA::gfpUV! 5.8 5.7 6.1 5.9! 5.4 7.7 ( ( ( * * * prpoS::gfpUV! Rmax 6.6 6.5 7.4 3.5! 3.4 3.2 + + + ) ) ) pgrpE::gfpUV! 2.2 2.1 2.2 5.2! 4.9 6.6 , , - . . / puspA::gfpUV! 0.123 ! 0.193 ! 17.1 0.017! 0.027 ! 206.1 ‡ , , - . . / prpoS::gfpUV! r (g ) 0.100 ! 0.147 ! 22.4 0.013! 0.023 ! 123.5 , , - . . / pgrpE::gfpUV! 0.603 ! 0.580 ! 5.3 0.015! 0.025 ! 225.9 0 0 1 2 2 3 puspA::gfpUV! 8.4 ! 11. 5 1.1 11.7 ! 12.5 0.1 ‡ 0 0 1 2 2 3 prpoS::gfpUV! ti (p ) 8.9 ! 12.8 0.9 1.5! 6.3 0.6 0 0 1 2 2 3 pgrpE::gfpUV! 3.6 ! 4.6 2.0 2.8! 1.2 1.0 ‡Weibull model coeffients. !,",#,$,%,&,',(,+,),*,,, -,., /, 0, 1,2,3 Greek letters indicate that means within inactivation parameters were statistically different within their group at P < 0.05 using the
Tukey procedure. Rmax = maximum response ratio, r = response rate, ti = time to response (equivalent to lag phase), AIC = Akaike Information Coefficient, BIC = Bayesian Information Coefficient.
! ! !
8
7
6
5 ) R 4
uspAuspA::gfpUV 3 rpoSrpoS::gfpUV
Responseratio ( Logistic: uspA 2 Logistic Logistic: rpoS Weibull: uspA 1 Weibull Weibull: rpoS A
0 0102030405060708090 time (h)
8
7
uspAuspA::gfpUV rpoSrpoS::gfp 6 UV Logistic: uspA Logistic Logistic: rpoS 5 Weibull: uspA Weibull ) Weibull: rpoS R 4
3 Responseratio ( 2
1 B
0 050100150200250300350 time (h) !
Figure 4-2. . Response ratios of Escherichia coli O157:H7 fluorescence gene promoters (grey circles = uspA::gfpUV; black circles = rpoS::gfpUV) fitted by mathematical models (regular line = Weibull; dashed line = logistic). A = Starvation, B = Cold Shock. See Eq. 1 and Eq. 2 for Response Ratio (R).
"#+! ! !
As previously observed by Zhang and Griffiths (2003), higher fluorescence units were obtained from the puspA::gfpuv reporter plasmid, with variations depending on the method applied to analyze fluorescence. While puspA::gfpuv may give higher response ratios, a conservative approach may be to analyze all the three gene reporters when studying new stresses. As observed by Gawande and others (2005ab) there is a variation in responses within the gene reporters with stresses, and also with time.
Hence, trends analysis may contribute to better understanding on simultaneous responses.
4.3.3.Cross-protection to physical hurdles
Bacterial cells were harvested after exposure to different environmental stress conditions (Table 4-3). E. coli cells grown in LACB resulted in similar mean times (tc) as determined with the Weibull model (P % 0.05). On the other hand, cells grown in TSBL varied in their resistance to PEF. The E. coli cells were significantly more susceptible to
PEF (P < 0.05) when grown at 7 °C compared to those exposed to acid (LA) and high
NaCl content (low aw). Bacteria grown in TSBL at 45 °C showed the highe st resistance to the PEF treatment (P < 0.05) (Table 4-3). Starved and cold shocked cells also showed similar tc values, which, as a result of a higher electric field, were two to five times lower than the previous treatments (PEF treatments 1 and 2).
"#"! ! !
Table 4-3. Growth and stress conditions of Escherichia coli O157:H7 and its relative resistance to pulsed electric fields (PEF).
Pre-processing conditions Physiological status PEF Processing Resistance
Tempe- Electric Medium time frequency tc rature Status at Method of Field inactivation monitoring (°C) (h) (kV/mm) (Hz) (!s) 10 100 99.5" Stationary LACB 25 25 Photometry 1.8 20 70.3" phase 40 25 87.3"
TSBL 7 150 97.4# "##! TSBL pH 5.0 (LA) 35 150 168.4$ Stationary Photometry 1.8 20 phase TSBL 7.5% NaCl 35 150 153.0$
TSBL 45 25 240.9%
LB† 35 20 Non-stressed 24.3&
5 50 40.3& LB Cold stressed Fluorometry 2.4 20 5 250 54.6& Starvation Water 25 100 stress 46.2&
",#,$,%,&Greek letters indicate means within inactivation parameters that are statistically different at P < 0.05 using the Tukey pairwise comparison test. †Non-shocked control. LACB = lactose broth, TSBL = tryptic soy broth supplemented with 0.5 % lactose, LA = lactic acid, LB = Luria-Bertani broth, Miller, tc = mean time calculated according to the model. †Absorbance values determined at 450 nm, fluorescence values determined using 485 nm 535 nm filters for excitation and emission respectively from the rpoS promoter gene reporter.!
! ! !
A detailed analysis of the inactivation trends resulted in downwards concavity for all the treatments; with similar magnitude in the scale and shape parameters of the
LACB treatments as for the cell exposed to stress (Table 4-4). In the TSBL grown cells the scale parameter (a) showed similar variation as the tc with a significantly lower value (P < 0.05) for E. coli cells grown at 7 °C and higher (P < 0.05) for thos e grown at
45 °C. By contrast, the shape parameter ( n) was found to be different in bacteria grown in acid environments compared to those grown at 45 °C (P < 0.05). Physiologically tc can be used as a global indicator of microbial health as it reflects the mean of the distribution of resistances in the population, in the other hand, the shape parameter indicates potential “shouldering” or “tailing”, which result from a genetically protected or mutant population to low or high intensity treatments, respectively (Juneja and others
1998, Hayman 2007). At a constant shape factor, the scale parameter (calculated using the appropriate model) has been incorporated in secondary models (z values) to reflect microbial resistance considering a broader temperature or pH range (Couvert and others 2005).
Power dissipation produces a joule heating effect (Heinz and others 2002), which can be called PEF dissipated heat. Analysis of the product temperature showed that the
PEF treatment of E. coli cells at 1.8 kV/mm and 20 Hz using PEF treatment 1 (Figure 4-
3A) resulted in a steeper increase of dissipated heat into the inoculated milk compared to treatment 2 (Figure 4-3B) with the highest temperature below 50 °C. In the case of the milk PEF treatment at 2.4 kV/mm and 20 Hz (Figure 4-3C) the increase in temperature was even higher with time, and exceeded 50 °C in approximately 100 !s.
"#$! ! !
Table 4-4. Weibull and Weibullian-log logistic (WeLL) inactivation parameters (n = scale parameter, a = shape
parameter, k = rate, Tc = onset temperature level of inactivation) describing Escherichia coli O157:H7 inactivation in milk by pulsed electric fields (PEF) processing.
PEF Weibull WeLL Medium Temperature Time Intensity parameters Parameters #" %&"" " (°C) (h) (kV/mm, Hz) !" $" (!s-1)" (°C) "
LACB 10 100 7.4" 106.4# 3.6$ 47.1%
LACB 25 25 1.8, 20 3.9" 77.4# 7.4$ 40.1%
" # $ % LACB 40 25 5.5 94.6 8.3 39.8
TSBL 7 150 2.4&' 109.2( 9.3) 47.3*
& + ) *
"#$! TSBL pH 5.0 (LA) 35 150 3.0 188.5 6.3 57.1 1.8, 20 TSBL 7.5% NaCl 35 150 2.4&' 172.1(+ 8.1) 54.6*
TSBL 45 25 1.2' 257.0, 10.2) 59.2*
LB†† 35 20 1.4! 26.0- 7.3. 28.6/
LB 5 50 3.1! 44.6- 2.6. 30.3/ 2.4, 20
LB 5 250 4.7! 59.8- 3.4. 31.9/
Water 25 100 4.3! 50.7- 1.6. 31.5/
",#,$,%,&, ', (, +,,,),*,!,-,., / † Superscripted symbols indicate means within inactivation parameters that are statistically different at P < 0.05 using the Tukey procedure. Water activity (aw) = 0.920. ††Non-shocked control. LACB= lactose broth, TSBL = tryptic soy broth supplemented with 0.5 % lactose, LA = lactic acid, LB = Luria-Bertani broth, Miller.
! ! !
time (!s) 0255075100125150 1.0 75
0.0 A 65
77 "'CC -1.0 2525 'C"C 55 Temperature ('C) Temperature Temperature ( Temperature 4040 'C"C -2.0 Joule7 'C heating
25 'C 45 " -3.0 40 'C C)
Joule Heating Microbialreduction (Log S(t)) 35 -4.0 Joule Heating
-5.0 25 0+#%%+(%"++"#%"%+18 36 56 74 92 110 Energy (kJ/l)
time (!s) 050100150200250 1 75
0 B 65 77 "'CC
-1 pH 5.0 7.5% Salt 55 Temperature ( Temperature ( Temperature Temperature ( Temperature ('C) Temperature 4545 'C"C -2 7Joule 'C heating 45 pH 5.0 " " C) C) " -3 40 'C C)
Microbialreduction (Log S(t)) 7.5% Salt 35 -4
-5 25 0 36 74 110 146 184 Energy (kJ/l)
(Continued)
"#%! ! !
time (!s) 020406080100 1 75
0 C 65
-1 Control 55 Temperature ( Temperature
Cold Shocked (50 h) ('C) Temperature Cold Shocked (250 h) -2 Starved JouleControl heating 45
CS 50 " -3 C) CS 250 Starved Microbialreduction (Log S(t)) Joule Heating 35 -4 JH
-5 25 0+#+,+&+)+"++26 52 78 104 130 Energy (kJ/l)
Figure 4-3. Inactivation trends of Escherichia coli O157:H7 by pulsed electric fields (PEF) processing after growth at challenging conditions and PEF treatment at 1.8 kV/mm and 20 Hz (A, B) or following environmental stress and PEF treatment at 2.4 kV/mm and 20 Hz (C). S = survival population, LACB = lactose broth, TSBL = tryptic soy broth with added 0.5% lactose, NaCl = sodium chloride, aw = water activity. Trends were calculated using the Weibull model, and specific energy based on capacitance.
"#&! ! !
Several studies have shown that the resistance of bacterial cells to stress is greater when they are in the stationary phase of growth. For example, Lou and Yousef
(1996) showed that the heat resistance of Listeria monocytogenes (measured as D56°C value) increased eight-fold from early logarithmic to early stationary phase. These observations also apply to PEF, Wouters and others (2001) demonstrated that L. plantarum cells exhibited higher sensitivity to PEF when they were at exponential phase rather than the stationary phase of growth. They attributed the effect to the elongation of the bacteria cells in preparation for division compared to stationary phase bacterial cells that have been adapted to survive stress conditions.
Growth temperature prior to hydrostatic high pressure (HHP) treatment was reported to affect the resistance of L. monocytogenes to inactivation (Bull and others
2005). Similarly, exposure of L. monocytogenes to low pH increased its resistance to subsequent heat treatment (Lou and Yousef 1996). Ryu and Beuchat (1998) studied the effect of acid adaptation, and shock of E. coli O157:H7 in TSB (with 1% added glucose), and observed a cross-protective effect when the organism was heat treated in apple cider, and orange juice. They reported that the acid adapted bacteria were almost twice as resistant as the acid shocked E. coli or the controls. Rowe and Kirk (1999) noticed higher heat resistance (D56°C ) after exposing E. coli O157:H7 for 1 h to acidified TSB
(with glucose). Leenanon and Drake (2001) observed that acid adapted E. coli O157:H7 were more heat resistant (D56°C ) than those subjected to heat shock exposure. With reference to PEF resistance, García and others (2005) analyzed the impact of acid shock on eight microorganisms in citrate-phosphate buffer pH 4.0 (2 h) and found an increase in resistance of the Gram negative organisms to PEF.
"#(! ! !
Starvation and cold shock have been also evaluated quantitatively for their effect on resistance. Lou and Yousef (1996) observed that the thermo-tolerance (D56°C ) of L. monocytogenes increased after starving the cells for up to 156 h. A similar trend of increased thermo-tolerance (D56°C ) was observed by Leenanon and Drake (2001) after starving E. coli O157:H7 at 37 °C for 48 h with rpoS negative mutan ts exhibiting a lower resistance. Moreover, Zhang and Griffiths (2003) suggested increased D52°C values of
E. coli O157:H7 as a result of expression of stress response genes at various temperatures. Leenanon and Drake (2001) reported that cold shock (at 5 °C for 1 week) led to a decrease of the thermo-tolerance (D56°C ) of E. coli cells except for the rpoS negative mutant.
When comparing all of the stresses applied in the present study it could have been expected that starvation will produce more resistant cells, as found by Lou and
Yousef (1996), and by Leenanon and Drake (2001). Direct comparison of the different environmental stresses indicated that starvation induced stress resulted in similar PEF resistance of E. coli O157:H7 than exposure to cold stress. However, the ability to survive PEF was lower for bacteria exposed to osmotic and acid stress compared to cells grown at higher growth temperatures were more resistant to PEF. The variance in the findings obtained in this study could be explained by the stress specificity, and the need of bacteria to maintain membrane fluidity for growth under challenging environmental conditions as well as to adapt to changes in their intracellular composition (Lado and others 2007, Chung and others 2001). Because the similarity in the mechanisms of action and resistance, the protective effect observed in E. coli cells grown at high temperatures on resistance to subsequent PEF treatment may be related
"#)! ! ! to membrane fluidity. Non-linear inactivation trends may result from microbial resistance highlighted by “lag” periods (Juneja and others 1998), and by “tailing” (Hayman and others 2007). “Tailing” as a consequence of PEF processing was reported by Alvarez and others (2003) for the inactivation of Salmonella enterica, and Rodríguez-Calleja and others (2006) with Staphylococcus aureus, and these results can be attributed to the temperature control (below 35 °C) both research gro ups applied during their treatments.
Predictive microbiology is evolving by incorporating the effect of the physiological characteristics on the inactivation kinetics under static and dynamic conditions including the modified Dabes kinetics (Vladramidis and others 2007), and the Weibullian-log logistic (Peleg and Penchina 2000, Corradini and Peleg 2009). Thus, it is important to point out that most of these variations in the progression of the kinetics occur at temperatures where metabolic reactions occur, and this may in turn, depend on both the microorganism and the food system investigated. Probabilistic modeling of growth/no growth interfaces represents an alternative for the analysis of environmental growth limitations (Palou and Malo 2005). The analysis of growth parameters have shown a dependence on environmental factors, leading to the development of Arrhenius-type kinetics, and secondary models like those applied in this study, which have the potential to be applied in the design of processes using the hurdle technology approach
(McMeekin and others 2000).
"#*! ! !
4.4.Conclusions
The analysis of E. coli O157:H7 growth dynamics in two growth mediums resulted in an enhanced ability to growth at challenging conditions when the nutrient density was higher. In addition, the growth of E. coli at a lower pH (4.8) in a weaker acid
(LA), and a minimum water activity of 0.915 were predicted linear using secondary models. The response ratios of the E. coli O157:H7 general stress reporters were successfully modeled using mechanistic and probabilistic models, providing advancement to predictive risk assessment. Also, the variation in resistance of cross protected E. coli cells was reflected as mean time values from the Weibull model, and onset temperatures for inactivation, at various PEF treatment conditions were estimated, providing an additional insight to the mathematical analysis of PEF inactivation of microbial cells.
"$+! ! !
5. General conclusion
As reviewed in the introductory section, non-thermal dairy processes continue to be developed with several purposes including the inactivation or removal of microorganisms to improve the safety of dairy products and their stability during storage. Cross-flow micro-filtration is becoming well established in the dairy industry with the objective of removing spores that cause cheese “late blowing” and to produce extended shelf-life (ESL) milk with enhanced quality attributes.
Because the regulations related to pasteurization are based on temperature and time combinations, the emergence of new processes is leading to the evaluation of alternatives to thermal processing because of their potential benefits on enhanced nutritional benefits, sensory attributes and shelf stability among others. For that reason the analysis of FSO’s is being considered as the criteria to determine the safety of products resulting from a line of processes.
Sustainable processing is an emerging trend driven by the public awareness of the environmental impact of human activities, and led by the dairy industry by analyzing the factors that affect the manufacture of its products. This movement continues to be incorporated in the development and commercialization of non-thermal processes.
Within the electro-magnetic processes PEF, which consists on pulsing an electric field to a food flowing through a treatment chamber, has demonstrated that can be applied for the inactivation of microorganisms in milk without damaging its macronutrients and micronutrients and quality aspects such as color, moisture.
"$"! ! !
Parallel to the development of new processes, their combination in sequence or as a unit has also been implemented commercially, in what has been named hurdle technology. The combination of non-thermal and thermal processes to achieve higher safety and stability levels has also become a movement within the new processes. In that regard, the flexibility on the application of PEF facilitates its potential to be combined with other technologies.
One of the drawbacks of the application of hurdle technology with the objective of minimal processing is the potential development of microbial stress responses that increase their resistance to decreased intensity levels of treatment, and or may cause cross-protection to heterologous stresses. Increasing research in the subject has developed several techniques that facilitate the monitoring of the physiological status of microbial cells when exposed to extreme environments. Combining these novel techniques with a more accurate quantitative analysis of the inactivation of the microbial population can result in a more accurate risk analysis of new processes.
Initial experiments (chapter 2) highlighted the various factors that affect the effectiveness of PEF for the inactivation of native microorganisms in milk. Statistical differences (P < 0.05) were observed on the reduction of mesophilic populations at fat contents ranging between 1.1 and 3.1 % at an intensity of 4.8 kV/mm and 918 µs, with higher reductions at the lower fat content. While there was no clear trend in the reduction of the coliform and psychrotroph populations between these fat levels the results obtained at the highest PEF intensity level were comparable to the ones obtained with HTST.
"$#! ! !
By increasing the PEF inlet temperature from 5.0-6.3 °C to 34.0-34.3 °C the PEF treatment time required to inactivate 3.0 or more Log CFU/ml of the mesophilic population and more than 3.8 log CFU/ml was decreased from 1530 µs to 510 µs at 4.0 kV/mm and from 459 µs to 230 µs at 5.6 kV/mm. And by applying CFMF prior to PEF the reduction was increased to more than 4.7 log CFU/ml in mesophilics, coliforms and psychrotrophs, which was similar to the CFMF/HTST combination. In addition, differences in cream (12%) homogenization were noticeable in the reduction of the coliform population when applying 4.0 kV/mm for 1530 µs suggesting better PEF inactivation effects in homogenized cream. This study concluded establishing the relation between conductivity and fat or solids content, which resulted in a decrease in conductivity of 0.07 S/m per 10% of fat content in non-homogenized products, and an increase in conductivity of 0.05 S/m per 1 % of solids in the non-fat products with correlations of 0.97 and 0.92 respectively.
The first study was complemented with an investigation of the combination of
PEF and CFMF in the alternate sequences. Higher microbial reductions were achieved by applying PEF prior CFMF (PEF/CFMF) with up to 7.1 log CFU/ml reductions at 4.2 kV/mm and 612 µs, which was higher compared to 4.8 log CFU/ml achieved in the traditional sequence (CFMF/PEF), and thermal pasteurization (75 °C x 24 s) where 4.6 log CFU/ml were achieved. After 7 days of the shelf-life study the microbial counts of
PEF/MF treated milk and TP were similar for some of the populations analyzed (total aerobic mesophiles, total moulds and yeasts, total lactic acid bacteria, total staphylococci), the total enteric bacteria counts and total psychrotrophs, were higher in
PEF/CFMF, while the total thermoduric psychrotrophs, total thermoduric mesophiles,
"$$! ! ! and total thermoduric thermophiles were higher in thermally pasteurized milk, leading to a conclusion that the kind of microorganisms inactivated by both technologies may differ resulting in different profiles of microorganisms that recover from inactivation.
A different approach was taken in the third study where the growth of Escherichia coli O157:H7 was monitored by means of absorbance in two mediums (LACB and
TSBL) resulting in an enhanced capability to growth at extreme temperature conditions in a medium with higher nutrient density (TSBL). Also, TSBL was modified to growth E. coli O157:H7 in acid and high salt concentrations. The Cardinal and Ratkowsky models were used to describe the microbial growth from optimum to extreme environments.
Additionally microbial constructs allowed monitoring the expression of stress proteins by means of fluorescence. The Gompertz, Logistic and Weibull models have similar goodness-of-fit in explaining stress response ratios. In general, the kinetics of inactivation by PEF explained by the weibull distribution showed downwards concavity as the temperature increased with treatment time. When PEF treated at the peak of their growth no differences in the weibull constants were found among cells grown in
LACB at three temperatures (10, 25 or 40 °C), while when grown in TSBL at 45 °C the cells more than two times as resistant compared to the cells grown in acid (lactic acid pH 5.0), high salt concentrations (7.5%) or at cold temperatures (7 °C). No differences were found in cells stressed by starvation for 100 hours (25 °C) or cold shocked at 5 °C for 50 or 250 hours. It was concluded that the cross-protective effect of heterologous stresses might be counteracted by the combined effect of electric fields and joule heating. The minimum temperatures for inactivation to occur predicted by the WeLL model varied depending on the setup applied, with no differences within the treatments.
"$,! ! !
Also, E. coli O157:H7 cells grown at higher temperatures may require an increase in treatment time to achieve similar reductions.
In general, this work addressed the various factors involved in the implementation of PEF as a microbial inactivation technique in milk, its combination with microfiltration, and explained the potential risks that need to be accounted when a stress adapted microorganism (E. coli O157:H7) may represent when exposed to PEF using advanced models.
5.1. Applications and future research!
By studying the survival of natural microflora passing through various processing steps of milk production critical factors were highlighted, which enable possible implementation of CFMF in conjunction with PEF in a milk production line as a possible alternative to conventional pasteurization methods. To implement of PEF in combination with other processes as CFMF and heat at the plant level more investigation in the temperature effect is needed; higher CFMF temperatures (around 50 °C) have been shown to increase the level of microbial reduction which implemented in would allow
PEF processing at higher flow rates. By improving the processing conditions there is a potential to increase the milk shelf life, creating a process to produce milk with a lower level of spores and thermodurics than traditionally pasteurized milk. Research at the protein and microbial level when reversing the sequence of CFMF and PEF processing would elucidate the molecular mechanisms that cause an enhanced microbial reduction.
In addition, the effect of homogenization on the microbial membrane integrity when
"$%! ! ! treating milk products of high fat level can reveal a different pattern of microbial inactivation.
Research on the effects of environmental stress response of a model microorganism as E. coli O157:H7 the observations can be extrapolated to similar bacteria. As the study of modified liquid media revealed a protective effect of the nutrients additional research in the effect of sugar content, and type (glucose, sucrose or lactose), may expose different patterns of microbial growth at extreme conditions (i.e. temperature and food ingredients like acids and salts) that can be further modeled. The construction of fluorescent stress reporters (biosensors) of other types of microorganisms (i.e. gram positives) may also reveal different trends. Also, further studies in the cross-protective effect of environmental hardening prior PEF processing at low temperatures may result in variation on microbial inactivation due to variation on the protective mechanisms that could be utilized for microbial selection. From a safe processing perspective the study of the inactivation of cross-protected bacteria by PEF processing at higher temperatures would help to establish the optimum conditions for use as a pasteurization process. Applying quantitative methods to optimize the PEF processing conditions where a desired microbial inactivation level is achieved at optimum temperature and flow conditions can strengthen this understanding.
"$&! ! !
6. References
Aaku EN, Collison EK, Gashe BA, Mpuchane S. 2004. Microbiological quality of milk
from two processing plants in Gaborone Botswana. Food Control. 15: 181-186.
Abee T, Wouters JA. 1999. Microbial stress response in minimal processing. Int J Food
Microbiol. 50: 65-91.
Advanced Electron Beams (AEB). 2009. AEB Blu sterilization for aseptic packaging:
ambient temperature, chemical free, water free sterilization with in-line electron
beams. Advanced Electron Beams. Wilmington, Massachusetts. Webminar.
Available at: http://www.aeb.com/downloads/
Advanced Electron Beams (AEB). 2009. The evolution of aseptic packaging: ambient
temperature, chemical free, water free sterilization with in-line electron beams.
Advanced Electron Beams. Wilmington, Massachusetts. Webminar. Available at:
http://www.aeb.com/downloads/
Alvarez I, Heinz V. 2007. Hurdle technology and the preservation of food by pulsed
electric fields. In: Food preservation by pulsed electric fields: from research to
application. Boca Raton, FL: CRC Press. p 167-177.
Álvarez I, Mañas P, Condón S, Raso J. 2003. Resistance variation of Salmonella
enterica Serovars to pulsed electric fields treatments. J Food Sci. 68 (7): 2316-
2320.
Alvarez I, Pagán R, Condón S, Raso J. 2003. The influence of process parameters for
the inactivation of Listeria monocytogenes by pulsed electric fields. Int J Food
Microbiol. 87: 87-95.
"$(! ! !
American Public Health Association. 1992. Compendium of Methods for the
Microbiological Examination of Foods, APHA Inc, Washington, DC, USA.
Amiali M, Ngadi MO, Raghavan VGS. 2006. Electrical conductivities of liquid egg
products and fruit juices exposed to high pulsed electric fields. Int J Food Prop. 9:
533-540.
Amiali M, Ngadi MO, Smith JP, Raghavan GSV. 2007. Synergistic effect of temperature
and pulsed electric field on inactivation of Escherichia coli O157:H7 and
Salmonella enteritidis in liquid egg yolk. J Food Eng. 79: 689-694.
Anderson IC, Rhodes M, Kator H. 1979. Sublethal stress in Escherichia coli: a function
of salinity. Appl Env Microbiol. 38 (6): 1147-1152.
Aronsson K, Borch E, Stenlöf B, Rönner U. 2004. Growth of pulsed electric field
exposed Escherichia coli in relation to inactivation and environmental factors. Int
J Food Microbiol. 93: 1-10.
Aronsson K, Rönner U, Borch E. 2005. Inactivation of Escherichia coli, Listeria innocua
and Saccharomyces cerevisiae in relation to membrane permeabilization and
subsequent leakage of intracellular compounds due to pulsed electric field
processing. Int J Food Microbiol. 99: 19-32.
Baranyi J, Pin C, Ross T. 1999. Validating and comparing predictive models. Int J Food
Microbiol. 48: 159-166.
Baranyi J, Pin C. 2004. Modeling the history effect on microbial growth and survival:
deterministic and stochastic approaches. In: McKellar R, Lu X. Modeling
microbial responses in food. Boca Raton, FL, USA: CRC Press. 285-301.
"$)! ! !
Baranyi J, Roberts TA, McClure P. 1993. A non-autonomous differential equation to
model bacterial growth. Food Microbiol. 10: 43-59.
Baranyi J, Roberts TA. 1994. A dynamic approach to predicting bacterial growth in food.
Int J Food Microbiol. 23: 277-294.
Baranyi J, Roberts TA. 1995. Mathematics of predictive food microbiology. Int J Food
Microbiol. 26: 199-218.
Barbosa-Cánovas GV, Altunakar B. 2006. Pulsed electric fields processing of foods: an
overview. In: Raso J, Heinz V. Pulsed electric fields technology for the food
industry: fundamentals and applications. New York, NY: Springer
Science+Business Media. 3-26.
Barbosa-Cánovas GV, Bermúdez-Aguirre D. 2010. Pasteurization of milk with pulsed
electric fields. In: Griffiths MW, editor. Improving the safety and quality of milk.
Vol 1: milk production and processing. Boca Raton FL: Woodhead Publishing
Ltd. 400-419.
Barbosa-Cánovas GV, Juliano P. 2008. Food sterilization by combining high pressure
and thermal energy. In: Gutiérrez-L'pez GF, Barbosa-Cánovas GV, Welti-
Chanes J, Parada-Arias E. Food engineering: integrated approaches. New York,
NY: Springer Science+Business. 9-46.
Barbosa-Cánovas GV, Swanson BG, San Martín MF, Harte F. 2005. Use of magnetic
fields as a nonthermal technology. In: Barbosa-Cánovas GV, Tapia MS, Cano
MP. Novel food processing technologies. Boca Raton, FL: CRC Press. 443-452.
Barbosa-Cánovas GV, Tapia MS, Cano MP. 2005. Novel food processing technologies.
Boca Raton, FL: CRC Press. 692 p.
"$*! ! !
Barer MR, Harwood CR. 1999. Bacterial viability and culturability. In: Poole RK, editor.
Advances in microbial physiology. Vol 41. London: Academic Press. 93-137.
Barnes G. 2009. Non-thermal pasteurization technology – is it ready for dairy beverage
processing?. Dairy Management. IDFA milk & cultured products seminar. March
31 to April 2. Personal communication.
Bazhal MI, Ngadi MO, Raghavan GSV, Smith JP. 2006. Inactivaion of Escherichia coli
O157:H7 in liquid whole egg using combined pulsed electric field and thermal
treatments. LWT Food Sci Technol. 39: 419-425.
Beolchini F, Cimini S, Mosca L, Veglio F, Barba D. 2005. Microfiltration of bovine and
ovine milk for the reduction of microbial content: effect of some operating
conditions on permeate flux and microbial reduction. Sep Sci Technol. 40: 757-
772.
Beolchini, F., Veglio, F., Barba, D., 2004. Microfiltration of bovine and ovine milk for the
reduction of microbial content in a tubular membrane: a preliminary investigation.
Desalin. 161: 251-258.
Bio-Rad. (2010). Gene pulser xcellTM electroporation system. Instruction manual. Bio-
Rad Laboratories, Inc. Hercules, CA, USA. 83 p. Available at: http://www.bio-
rad.com
Bird J. 1996. The application of membrane systems in the dairy industry. J Soc Dairy
Technol. 49: 16-23.
Bourdichon F, Uyttendaele M, Geeraerd A, Zwiettering M, Membré J. 2009. Application
of QMRA to go beyond safe harbors in thermal processes.Part 1: Introduction
and Framework. International Conference on Predictive Modeling in Foods.
",+! ! !
International Life Sciences Institute. Washington DC. 23 p. [Accessed 2010
August 28]. Available from:
http://www.icpmf.org/pp/S10_1_PM100_Application%20of%20QMRA%20to%20g
o%20beyond%20safe%20harbors_Part1.pdf
Brans G, Schroën CGPH, van der Sman RGM, Boom RM. 2004. Membrane
fractionation of milk: state of the art and challenges. J Membr Sci. 243: 263-272.
Brul S, Westerhoff HV. 2007. Systems biology and microbiology. In: Brul S, Van
Gerwen S, Zwietering M, editors. Modelling microorganisms in food. Boca Raton,
FL: CRC Press. 250-294.
Bull MK, Hayman MM, Stewart CM, Szabo EA, Knabel. 2005. Effect of prior growth
temperature, type of enrichment medium, and temperature and time of storage
on recovery of Listeria monocytogenes following high pressure processing of
milk. Int J Food Microbiol. 101: 53-61.
Busch SV, Donnelly CW. 1992. Development of a repair-enrichment broth for
resuscitation of heat-injured Listeria monocytogenes and Listeria innocua. Appl
Env Microbiol. 58 (1): 14-20.
Campanella OH, Peleg M. 2001. Theoretical comparison of a new and the traditional
method to calculate Clostridium botulinum survival during thermal inactivation. J
Sci Food Agric. 81: 1069-1076.
Canadian Food Inspection System (CFIS). 2005. National Dairy Code: production and
processing regulations. 4th Ed. [Accessed 2010 June 18]. Available from:
http://www.cfis.agr.ca/english/regcode/ndrc/amdmt_jul_and_oct2005/2005_ppr_e
","! ! !
Castro AJ, Swanson BG, Barbosa-Cánovas GV, Zhang QH. 2001. Pulsed electric field
modification of milk alkaline phosphatase activity. In: Barbosa-Cánovas GV,
Zhang QH, editors. Pulsed electric fields in food processing: fundamental aspects
and applications. Lancaster, PA: Technomic Publishing Co. p 83-103.
Center for Dairy Research (CDR). 2004. Membranes 101. Wisconsin Center for Dairy
Research. University of Wisconsin Madison. [Accessed 2010 December 12].
Available from:
http://www.cdr.wisc.edu/programs/dairyingredients/pdf/membranes_101.pdf
Cerf O. 1986. Introduction. In: International Dairy Federation (IDF). Monograph on
pasteurized milk. Bulletin of the International Dairy Federation no. 200. Brussels,
Belgium. 98 p.
Christiansson A, Naidu AS, Nilsson I, Wadström T, Pettersson, HE. 1989. Toxin
production by Bacillus cereus dairy isolates in milk at low temperatures. Applied
Environ Microbiol 55, 2595-2600.
Chung HJ, Bang W, Drake MA. 2006. Stress response of Escherichia coli. Compre Rev
Food Sci Food Saf. 5: 52-64.
Cole MB, Davies KW, Munro G, Holyoak CD, Kilsby DC. 1993. A vitalistic model to
describe the thermal inactivation of Listeria monocytogenes. J Ind Microbiol. 12:
232-239.
Corradini MG, Normand MD, Peleg M. 2010. Stochastic and deterministic model of
microbial heat inactivation. Conc Rev and Hypotheses Food Sci. 75 (2): R59-
R70.
",#! ! !
Corradini MG, Peleg M. 2009. Dynamic model of heat inactivation kinetics for bacterial
adaptation. Appl Env Microbiol. 75 (8): 2590-2597.
Couvert O, Gaillard S, Savy N, Mafart P, Leguérinel I. 2005. Survival curves of heated
bacterial spores: effect of environmental factors on Weibull parameters. Int J
Food Microbiol. 101: 73-81.
Dalgaard P, Koutsoumanis K. 2001. Comparison of maximum specific growth rates and
lag times estimated from absorbance and viable count data by different
mathematical models. J Microbiol Methods. 43: 183-196.
Dalgaard P, Ross T, Kamperman L, Neumeyer K, McMeekin TA. 1994. Estimation of
bacterial growth rates from turbidimetric and viable count data. Int J Food
Microbiol. 23: 391-404.
Datta N, Deeth HC. 1999. High pressure processing of milk and dairy products.
Australian J Dairy Technol. 54, 41-48.
Datta N, Deeth HC. 2007. UHT and Aseptic Processing of Milk and Milk Products. In:
Tewari G, Juneja VK. Advances in Thermal and Non-Thermal Food Preservation.
63-90. Blackwell Publishing: Oxford.
Daufin G, Escudier JP, Carrère H, Bérot S, Fillaudeau L, Decloux M. 2001. Recent and
emerging applications of membrane processes in the food and dairy industry.
Food Bioprod Process 79, 89-102.
Davidson PM, Roth LA, Gambrel-Lenarz SA. (2004). Coliform and other indicator
bacteria. In: Wehr HM, Frank JF, editors. Standard methods for the examination
of dairy products. 17th ed. Washington DC: American Public Health Association.
187-226.
",$! ! !
De Luis R, Arias O, Puértolas E, Benedé S, Sánchez L, Calvo M, Pérez MD. 2009.
Effect of high-intensity pulse electric fields on denaturation of bovine whey
proteins. Milchwissenschaft. 64: 422-426.
De Noni I, Pellegrino L, Cattaneo S, Resmini P. 2007. HPLC of proteose peptones for
evaluating ageing of packaged pasteurized milk. Int Dairy J. 17, 12-19.
Deeth HC, Datta N, Ross AIV, Dam XT. 2007. Pulsed electric field technology: effect on
milk and fruit juices. In: Tewari G, Juneja VK, editors. Advances in thermal and
non-thermal food preservation. Ames: Blackwell Publishing. 241-269.
Degen PJ, Kendall TA, Dehn J, inventors. Pall Corporation, assignee. 1995 Apr 9.
Production of sterile milk through dynamic microfiltration. US Patent 5,401,523.
DeHaan SWH. 2007. Circutry and pulse shapes in pulsed electric field treatment of
food. In: Lelieveld HLM, Notermans S, de Haan SWH, editors. Food preservation
by pulsed electric fields: from research to application. Boca Raton, FL: CRC
Press. p 43-69.
Deutsche Forschungsgemeinschaft (DFG). 2007. Thermal processing of food: potential
health benefits and risks Symposium. Weinheim: Wiley VCH. 283 p.
Donnelly CW, Baigent GJ, Briggs EH. 1988. Flow cytometry for automated analysis of
milk containing Listeria monocytogenes. J Assoc Off Anal Chem. 71 (3): 655-658.
Donnelly CW, Baigent GJ. 1986. Method for flow cytometric detection of Listeria
monocytogenes in milk. Appl Env Microbiol. 52 (4): 689-695.
Eberhard P, Gallmann PU. 2009. New developments in heating technology for food
preservation and safety. [Accessed 2010 April 10], Available from:
",,! ! !
http://www.milkproduction.com/Library/article_series/idf_fao_symp/New_develop
ments_in_heating_technology.htm
Edgell JW, Thomas LF, Clegg, LFL, Cuthbert WA. 1950. Thermoduric organisms in
milk. Part III. Provisional standard technique for the laboratory pasteurization test
for milk, and rinses and swabs of dairy equipment, with suggestions for
interpretation of results. J Appl Microbiol. 13: 132-134.
Elez-Martínez, Martín-Belloso O, Rodrigo D, Sampedro F. 2007. Impact of pulsed
electric fields on food enzymes and shelf-life. In: Lelieveld HLM, Notermans S, de
Haan SWH, editors. Food preservation by pulsed electric fields: from research to
application. Boca Raton, FL: CRC Press. 212-246.
El-Hadi Sulieman AM. 2007. Effect of pretreatment of milk quality characteristics of
Jibna-Beida (white cheese). Int J Food Eng. 3(4): Article 15.
El-Hag AH, Boggs S, Jayaram SH. 2006. Calculation of temperature rise due to the
application of pulsed electric fields in the context of food sterilization. Proceeding
of the ESA/ IEJ/ IEEE-IAS joint meeting, Berkeley, USA on June 6-9. 261-268.
Elwell MW, Barbano DM. 2006. Use of microfiltration to improve fluid milk quality. J
Dairy Sci. E20-E30.
European Union, 1997. Regulation (EC) No 258/97 of the European Parliament and of
the Council of 27 January 1997 concerning novel foods and novel food
ingredients. Official Journal of the European Communities No L43, 1-6.
Faraway JJ. 2002. Practical regression and Anova using R. Bath, UK: University of
Bath. 212 p. [Accessed 2010 Apr 2]. Available from: http://cran.r-
project.org/doc/contrib/Faraway-PRA.pdf
",%! ! !
Farber JM, Brown BE. 1990. Effect of prior heat shock on heat resistance of Listeria
monocytogenes in meat. Appl Environ Microbiol. 1584-1587.
Farid MM. inventor. University of New Zealand. Assignee. Pressure assisted thermal
sterilization or pasteurisation method and apparatus. 2009 Apr 10. US Patent
12/296,789.
Fernández A, Salmerón C, Fernández PS, Martínez A. 1999. Application of a frequency
distribution model to describe the thermal inactivation of two strains of Bacillus
cereus. Trends Food Sci & Technol. 10: 158-162.
Fernández-Molina JJ, Fernández-Gutiérrez SA, Altunakar B, Swanson BG, Barbosa-
Cánovas GV. 2005. The combined effect of pulsed electric fields and
conventional heating on the microbial quality and shelf life of skim milk. J Food
Process Preserv. 29: 390-406.
Floury J, Grosset N, Leconte N, Pasco M, Madec MN, Jeantet R. 2006. Continuous raw
skim milk processing by pulsed electric field at non-lethal temperature: effect on
microbial inactivation and functional properties. Lait. 86: 43-57.
Food and Drug Administration (FDA). 2000. Kinetics of microbial inactivation for
alternative food processing technologies. A report of the Institute of Food
Technologists for the Food and Drug Administration of the U. S. Department of
Health and Human Services. College Park, MD: CFSAN. Available at:
http://www.fda.gov/Food/ScienceResearch/Research
Areas/SafePracticesforFoodProcesses/ucm100158.htm
Food and Drug Administration (FDA). 2009. Grade “A” Pasteurized Milk Ordinance.
Silver Spring, MD: US Public Health Service/ US Food and Drug Administration.
",&! ! !
[Accessed 2010 May 15]. Available from:
http://www.fda.gov/downloads/Food/FoodSafety/Product-
SpecificInformation/MilkSafety/NationalConferenceonInterstateMilkShipmentsNCI
MSModelDocuments/UCM209789.pdf
FOSS. 1998. Reference Manual of Milkoscan FT 120 (Type 71200). Foss Electric A/S:
Hillerød, Denmark. 115 p.
Fox J. 2009. R Commander. A Basic-Statistics Graphical User Interface to R. Version
1.5-3. Hamilton, ON, Canada: McMaster University [Accessed 2010 Apr 2].
Available from: http://socserv.mcmaster.ca/jfox/Misc/Rcmdr
Fox MB. 2007. Microbial inactivation kinetics of pulsed electric field treatment. In:
Lelieveld HLM, Notermans S, de Haan SWH. Food preservation by pulsed
electric fields: from research to application. Boca Raton, FL: CRC Press. 127-
137.
Frank JF, Hankin L, Koburger JA, Marth EH. 1985. Test for groups of microorganisms.
In: G.H. Richardson, editor. Standard methods for the examination of dairy
products. American Public Health Association, Washington, DC. USA, 189-201.
Frank JF, Yousef AE. 2004. Tests for groups of microorganisms. In: Wehr HM, Frank
JF, editors. Standard methods for the examination of dairy products. 17th ed.
Washington DC: American Public Health Association. p 227-248.
Fritsch JA, Moraru CI. 2008. Development and optimization of a carbon dioxide-aided
cold microfiltration process for the physical removal of microorganisms and
somatic cells from skim milk. J Dairy Sci. 91: 3744-3760.
",(! ! !
Fung D. 2009. Rapid methods and automation in food microbiology: 25 years of
development and predictions. In: Barbosa-Cánovas G, Mortimer A, Lineback D,
Spiess W, Buckle K, Colonna, P. Global issues in food science and technology.
Burlington, MA: Academic Press. 165-176.
García D, Gómez N, Mañas P, Raso J, Pagán R. 2007. Pulsed electric fields cause
bacterial envelopes permeabilization depending on the treatment intensity, the
treatment medium pH and the microorganism investigated. International J Food
Microbiol. 113: 219-227.
García D, Gómez N, Raso J, Pagán R. 2005. Bacterial resistance after pulsed electric
fields depending on the treatment medium pH. Innov Food Sci Emerg Technol. 6:
388-395.
García D, Mañas P, Gómez N, Raso J, Pagán R. 2006. Biosynthetic requirements for
the repair of sublethal membrane damage in Escherichia coli cells after pulsed
electric fields. J Appl Microbiol. 100: 428-435.
Gawande PV, Griffiths MW. 2005. Effects of environmental stresses on the activities of
the uspA, grpE, and rpoS promoters of Escherichia coli O157:H7. Int. J Food
Microbiol. 99: 91-98.
Gawande PV, Griffiths MW. 2005. Growth history influences starvation-induced
expression of uspA, grpE, and rpoS and subsequent cryotolerance in Escherichia
coli O157:H7. J Food Protect. 68(6): 1154-1158.
Geciova J, Bury D, Jelen P. 2002. Methods for disruption of microbial cells for potential
use in the dairy industry – a review. Int Dairy J. 12: 541-553.
",)! ! !
Geeraerd AH, Valdramidis VP, Van Impe JF. 2005. GInaFiT, a freeware tool to assess
non-log-linear microbial survivor curves. Int J Food Microbiol. 102: 95-105.
Gésan-Guiziou G. 2010. Removal of bacteria, spores and somatic cells from milk by
centrifugation and microfiltration techniques. In: Griffiths MW, editor. Improving
the safety and quality of milk. Vol 1: milk production and processing. Boca Raton
FL: Woodhead Publishing Ltd. 349-372.
Geveke DJ. 2005. Non-thermal processing by radio frequency electric fields. In: Sun D,
editor. Emerging technologies for food processing. San Diego, CA: Elsevier
Academic Press. 307-322.
Grahl T and Märkl H. 1996. Killing of microorganisms by pulsed electric fields. Appl
Microbiol Biotechnol. 45: 148-157.
Griffiths MW. 1986. Use of milk enzymes as indices of heat treatment. J Food Protect.
49 (9): 696-705.
Griffiths MW. 1994. Predictive modeling: applications in the dairy industry. Int J Food
Microbiol. 23: 305-315.
Guan DS, Chen HQ, Hoover DG. 2005. Inactivation of Salmonella typhimurium DT 104
in UHT whole milk by high hydrostatic pressure. Int J Food Microbiol. 104, 145-
153.
Guerra A, Jonsson G, Rasmussen A, Waagner Nielssen E, Edelsten D. 1997. Low
cross-flow velocity microfiltration of skim milk for removal of bacterial spores. Int
Dairy J. 7: 849-861.
",*! ! !
Gunasekera TS, Attfield PV, Veal DA. 2000. A flow cytometry method for rapid
detection and enumeration of total bacteria in milk. Appl Env Microbiol. 66 (3):
1228-1232.
Gunasekera TS, Sørensen A, Attfield PV, Sørensen A, Veal DA. 2002. Inducible gene
expression by nonculturable bacteria in milk after pasteurization. Appl Env
Microbiol. 68 (4): 1988-1993.
Gunasekera TS, Veal DA, Attfield PV. 2003. Potential for broad applications of flow
cytometry and fluorescence techniques in microbiobiological and somatic cell
analyses of milk. Int J Food Microbiol. 85: 269-279.
Hanneman H, Robertson LJ. 2005. Heat recovery systems. In: International Dairy
Federation (IDF). Energy use in dairy processing. Bulletin of the International
Dairy Federation No. 401. Brussels, Belgium. p 32-48.
Hayes M, Boor K. 2001. Raw milk and fluid milk products. In: Marth EH, Steele JL.
Applied dairy microbiology. 2nd ed. New York, NY, USA: Marcel Dekker. p 59-76.
Hayman M. 2007. Factors that influence barotolerance of Listeria monocytogenes and
the mechanism of inactivation by high pressure processing. DPhil thesis. State
College: Pennsylvania State University Dublin. 199 p.
Hayman MM, Anantheswaran RC, Knabel SJ. 2007. The effects of growth temperature
and growth phase on the inactivation of Listeria monocytogenes in whole milk
subject to high pressure processing. Int J Food Microbiol. 115: 230-236.
Heinz V, Alvarez I, Angersbach A, Knorr D. 2002. Preservation of liquid foods by high
intensity pulsed electric fields- basic concepts for process design. Trends Food
Sci & Technol. 12:103-111.
"%+! ! !
Heinz V, Philips ST, Zenker M, Knorr D. 1999. Inactivation of bacillus subtilis by high
intensity pulsed electric fields under close to isothermal conditions. Food
Biotechnol. 13: 155-168.
Heinz V, Toepfl S, Knorr D. 2003. Impact of temperature on lethality and energy
efficiency of apple juice pasteurization by pulsed electric fields treatment.
Innovative Food Sci Emerg Tech. 4: 167-175.
Hening DR, Flowers R, Reiser R, Ryser ET. 2004. Pathogens in milk and milk products.
In: Wehr HM, Frank JF, editors. Standard methods for the examination of dairy
products. 17th ed. Washington DC: American Public Health Association. p 103-
152.
Hill C, Cotter PD, Sleator RD, Gahan CGM. 2002. Bacterial stress response in Listeria
monocytogenes: jumping the hurdles imposed by minimal processing. Int Dairy J.
12: 273-283.
Ho S, Mittal GS. 2000. High voltage pulsed electrical field for liquid food pasteurization.
Food Rev Int. 16(4): 395-434.
Hodgson PH, Leslie GL, Schneider RP, Fane AG, Fell CJD, Marshall KC. 1993. Cake
resistance and solute rejection in bacterial microfiltration: The role of the
extracellular matrix. J Membrane Sci. 79: 35-53.
Hoffman W, Kiesner C, Clawin-rädecker I, Martin d, Einhoff K, Lorenzen PC, Meisel H,
Hammer P, Suhren G, Teufel P. 2006. Processing of extended shelf life milk
using microfiltration. Int J Dairy Technol. (59) 4: 229-235.
Holm C, Jepsen L, Larsen M, Jespersen L. 2004. Predominant microflora of
downgraded Danish bulk tank milk. J Dairy Sci. 87: 1151-1157.
"%"! ! !
Holm C, Jespersen L. 2003. A flow-cytometric gram-staining technique for milk-
associated bacteria. Appl Env Microbiol. 69(5): 2857-2863.
Holm C, Mathiasen T, Jespersen L. 2004. A flow cytometric technique for quantification
and differentiation of bacteria in bulk tank milk. J Appl Microbiol. 97: 935-941.
Holm S, Elisabet IB, Holm EC, CF Holm. Inventors. Alfa-Laval Food and Dairy
Engineering AB. assignee. 1989 Jun 13. Method for producing milk with a
lowered bacterial content. US Patent 4,876,100.
Hoogland H, de Haan W. 2007. Economic aspects of pulsed electric field treatment of
food. In: Lelieveld HLM, Notermans S, de Haan SWH, editors. Food preservation
by pulsed electric fields: from research to application. Boca Raton, FL: CRC
Press. 257-265.
H*lsheger H, Potel J, Niemann EG. 1981. Kiling of bacteria with electric pulses of high
field strength. Radiat Environ Biophys. 20: 53-65.
Ibarz A, Barbosa-Cánovas GV. 2003. Unit operations in food engineering. Boca Raton,
FL, USA: CRC Press. 889 p.
Innovation Center for U.S. Dairy (ICUSD). 2008. U.S. dairy sustainability initiative: a
roadmap to reduce greenhouse gas emissions and increase business value.
[Accessed 2010 November 11]. Available from:
http://www.usdairy.com/Sustainability/Documents/PublicDocs/RoadmapToReduc
eGHGEmissions.pdf
Innovation Center for U.S. Dairy (ICUSD). 2009. Non-thermal ultraviolet processing.
Project Summary. [Accessed 2010 June 18]. Available from:
"%#! ! !
http://www.usdairy.com/sustainability/TheCommitment/Documents/NonThermalP
rocessingProjectSummary063009.pdf
International Commission for the Microbiological Specifications of Foods (ICMSF).
2002. Microbiological Testing in Food Safety Management. New York, NY:
Science+Business Media Inc. 362 p.
International Dairy Federation (FIL-IDF), 1974. Bacteriological quality of cooled bulk
milk. Document 83. International Dairy Federation, Brussels, Belgium.
International Dairy Federation (FIL-IDF). 1990. Bactoscan technique. In: International
Dairy Federation (FIL-IDF). Methods for assessing the bacteriological quality of
raw milk from the farm. Bulletin No. 256. FIL-IDF: Brussels, Belgium. 24-30.
International Dairy Federation (FIL-IDF). 2009. Environmental / Ecological impact of the
dairy sector: literature review on dairy products for an inventory of key issues, list
of environmental initiatives and influences on the dairy sector. International Dairy
Federation. Bulletin no. 436. 60 p.
International Dairy Federation (IDF). 1991. Liquid milk. Psychrotrophic microorganisms.
Colony count at 6.5 °C. Standard 101A. Internationa l Dairy Federation, Brussels,
Belgium.
International Dairy Federation (IDF). 2006. The world dairy situation 2006. Bulletin of
the International Dairy Federation no. 409. Brussels, Belgium. 89 p.
Jablasone J, Warriner K, Griffiths M. 2005. Interactions of Escherichia coli O157:H7,
Salmonella typhimurium and Listeria monocytogenes plants cultivated in a
gnobiotic system. Int J Food Microbiol. 99: 77-18.
"%$! ! !
Jaeger H, Schulz A, Karapetkov N, Knorr D. 2009. Protective effect of milk constituents
and sublethal injuries limiting process effectiveness during PEF inactivation of Lb.
rhamnosus. Int J Food Microbiol. 134: 154-161.
Jayaram S, Castle GSP, Margaritis A. 1991. Effects of high electric field pulses on
Lactobacillus brevis at elevated temperatures. IEEE. 674-681.
Jayaram SH. 2000. Sterilization of liquid foods by pulsed electric fields. IEEE Electr
Insul Mag. 16(6): 17-25.
Jelen P. 1992. Pressure-driven membrane processes: principles and definitions. In:
International Dairy Federation (IDF). New applications of membrane processes.
IDF special issue no. 9201. p 33-50..
Jimenez-Lopez AJE, Leconte N, Dehainault O, Geneste C, Fromont L, Gésan-Guiziou
G. 2008. Role of milk constituents on critical conditions and deposit structure in
skim milk microfiltration (0.1m). Sep Purif Technol. 61: 33-43.
Jocson MAL, Mason EO, Schanler RJ. 1997. The effects of nutrient fortification and
varying storage conditions on host defense properties of human milk. Pediatrics
100, 240-243.
Johnson EA. 2003. Microbial adaptation and survival in foods. In: Yousef AE, Juneja
VK, editors. Microbial stress adaptation and food safety. Boca Raton, FL: CRC
Press. p 75-103.
Juneja VK, Klein PG, Marmer BS. 1998. Heat shock and thermotolerance of
Escherichia coli O157:H7 in a model beef gravy system and ground beef. J Appl
Microbiol. 84: 677-694.
"%,! ! !
Juneja VK, Marks HM. 2003. Mathematical description of non-linear survival curves of
Listeria monocytogenes as determined in a beef gravy model system at 57.5 to
65°C. Innov Food Sci and Emerg Technol. 4: 307-317.
Juneja VK, Novak JS. 2003. Adaptation of foodborne pathogens to stress from
exposure to physical intervention strategies. In: Yousef AE, Juneja VK, editors.
Microbial stress adaptation and food safety. Boca Raton, FL: CRC Press. p 31-
53.
Juneja VK, Valenzuela Melendres M, Huang L, Gumudavelli V, Subbiah J, Thippareddi
H. 2007. Modeling the effect of temperature on growth of Salmonella in chicken.
Food Microbiol. 24: 328-335.
Juneja VK, Valenzuela Melendres M, Huang L, Subbiah J, Thippareddi H. 2009.
Mathematical modeling of growth of Salmonella in raw ground beef under
isothermal conditions from 10 to 45 °C. Int J Food Microbiol. 131: 106-111.
Kang D, Fung DYC. 1999. Thin agar layer method for recovery of heat-injured Listeria
monocytogenes. J Food Protect. 62 (11): 1346-1349.
Kaufmann V, Kulozik U. 2006. Kombination von Mikrofiltration und thermischen
Verfahren zur Haltbarkeitsverlängerung von Lebensmitteln. Chemie Ingenieur
Technik 78, 1647-1654.
Kessler HG. 2002. Food and bio process engineering - dairy technology. 5th ed.
Munich: Publishing House A Kessler. 681 p.
Knorr D, Heinz V, Luscher C. 2007. Thermal processing of foods: technological aspects.
In: Deutsche Forschungsgemeinschaft (DFG). Thermal processing of food:
potential health benefits and risks Symposium. Weinheim: Wiley VCH. p 17-25.
"%%! ! !
Kumar M, Hora R, Kostrzynska M, Warriner K. 2006. Mode of Salmonella and
Escherichia coli O157:H7 inactivation by a stabilized oxychloro-based sanitizer. J
Appl Microbiol. 102: 1427-1436.
Lado B, Yousef AE. 2007. Characteristics of Listeria monocytogenes important to food
processors. In: Ryser ET, Marth EH, editors. Listeria, Listeriosis, and Food
Safety. Third ed. Boca Raton, FL: CRC Press. p 157-213.
Laird DT, Gambrel-Lenarz SA, Scher FM, Graham TE, Reddy R. 2004. Microbiological
count methods. In: Wehr HM, Frank JF, editors. Standard methods for the
examination of dairy products. 17th ed. Washington DC: American Public Health
Association. p 153-186.
LASP. 2010. Electromagnetic Spectrum. University of Colorado Boulder. Available from:
http://lasp.colorado.edu/cassini/education/Electromagnetic Spectrum.htm
Lee S. 2004. Microbial safety of pickled fruits and vegetables and hurdle technology. Int.
J Food Safety. 4: 21-32.
Leenanon B, Drake MA. 2001. Acid stress, starvation, and cold stress after poststress
behavior of Escherichia coli O157:H7 and nonpathogenic Escherichia coli. J
Food Protect. 64 (7): 970-974.
Legan D, Vandeven M, Stewart C, Cole M. 2002. Modelling the growth, survival and
death of bacterial pathogens in foods. In: Blackburn CW. Foodborne pathogens –
hazards, risk analysis and control. 53-95.
Lei Y, Chen W, Mulchandani A. 2006. Microbial biosensors. Anal Chim Acta. 568: 200-
210.
"%&! ! !
Leistner L, Gould G. 2002. Hurdle technologies: combination treatments for food
stability, safety and quality. New York, NY: Kluver Academic/ Plenum Publishers.
194 p.
Leistner L, Morris GM. 1995. Food preservation by hurdle technology. Trends Food Sci
Technol. 6: 41-46.
Leistner, L., 2000. Basic aspects of food preservation by hurdle technology.
International Journal of Food Microbiology 55, 181-186.
Lewis M. 2010. Improving pasteurized and extended shelf-life milk, In: Griffiths MW,
editor. Improving the Safety and Quality of Milk: Volume One, Milk Production
and Processing. CRC Press, Boca Raton, FL, USA.
Li SQ, Zhang QH, Lee YZ, Pham TV. 2003. Effects of pulsed electric fields and thermal
processing on the stability of bovine immunoglobulin G (IgG) in enriched soymilk.
Journal of Food Sci. 68: 1201-1207.
López, S., Prieto, M., Dijkstra, J., Dhanoa, M.S., France, J. (2004). Statistical evaluation
of mathematical models for microbial growth. International Journal of Food
Microbiology. 96, 289-300.
Lou and Yousef. 1996. Resistance of Listeria monocytogenes to heat after adaptation to
environmental stresses. J Food Protect. 59 (5): 465-471.
Mabrook MF, Petty MC. 2003. Effect of composition on the electrical conductance of
milk. J Food Eng. 60: 321-325.
Madec MN, Mejean S, Maubois JL. 1992. Retention of Listeria and Salmonella cells
contaminating skim milk by tangential membrane microfiltration (“Bactocatch”
process). Lait. 72: 327-332.
"%(! ! !
Mafart P, Couvert O, Gaillard S, Leguerinel I. 2002. On calculating sterility in thermal
preservation methods: application of the weibull frequency distribution model. Int
J Food Microbiol. 72: 107-113.
Makardij A, Chen XD, Farid MM. 1999. Microfiltration and ultrafiltration of milk: some
aspects of fouling and cleaning. Food Bioprod Process 77, 107-113.
Malmberg R, Holm S. 1998. Low bacteria skim milk by microfiltration. N Eur Food Dairy
J. 54: 75-78.
Mañas P, Barsotti L, Cheftel JC. 2001. Microbial inactivation by pulsed electric fields in
a batch treatment chamber: effects of some electrical parameters and food
constituents. Innovat Food Sci and Emerging Technol. 2: 239-249.
Marcelo PA, Rizvi SSH. 2009. Applications of membrane technology in the dairy
industry. In: Pabby AK, Rizvi SSH, Sastre AM. Handbook of membrane
separations. Boca Raton, FL, USA: CRC Press. 636-669.
Marks BP. 2008. Status of microbial modeling in food process models. In: Datta A,
editor. Models for safety, quality, and competitiveness of the food processing
sector. Comp Rev Food Sci Food Saf. 7: 137-143.
Mastwijk HC, Gulfo-van Beusekom K, Pol-Hofstad IE, Schuten H, Boonman M, Bartels
PV. 2007. Definition and guidelines for reporting on pulsed electric field
experiments. In: Lelieveld HLM, Notermans S, de Haan SWH, editors. Food
preservation by pulsed electric fields: from research to application. Boca Raton,
FL: CRC Press. p 320-345.
"%)! ! !
Mathys A. 2008. Inactivation mechanism of geobacillus and bacillus spores during high
pressure thermal sterilization. DPhil thesis. Berlin: TU Berlin. 162 p. [Accessed
December 10 2010]. Available from: @11?IJJE4?0>719EE69E4J
McBee LE. 1996. Innovative methods of energy transfer. Poult Sci. 75: 1137-1140.
McKellar RC, Knight K. 2000. A combined discrete – continuous model describing the
lag phase of Listeria monocytogenes. Int J Food Microbiol. 54: 171-180.
McKellar RC, Lu X. 2004. Modeling microbial responses in food. Boca Raton, FL, USA:
CRC Press. 343 p.
McKellar RC, Lu X. 2004. Primary Models. In McKellar, R. C., Lu, X. Modeling microbial
responses in food (pp. 21-62). Boca Raton: CRC Press.
McKellar RC. 2003. Modelling the effectiveness of pasteurization. In: Smit G. Dairy
processing: improving quality. Cambridge, England: Woodhead Publishing Ltd.
104-129.
McMeekin T, Bowman J, McQuestin O, Mellefont L, Ross T, Tamplin M. 2008. The
future of predictive microbiology: strategic research, innovative applications and
great expectations. Int J Food Microbiol. 128: 2-9.
McMeekin TA, Presser K, Ratkowsky D, Ross T, Salter M, Tienungoon. 2000.
Quantifying the hurdle concept by modeling the bacterial growth/ no growth
interface. Int J Food Microbiol. 55: 93-98.
Membré J. Uyttendaele M, Bourdichon F, Geeraerd A, Zwiettering M. 2009. Application
of QMRA to go beyond safe harbors in thermal processes.Part 2: Introduction
and Framework. International Conference on Predictive Modeling in Foods.
International Life Sciences Institute. Washington DC. 23 p. [Accessed 2010
"%*! ! !
August 28]. Available from:
http://www.icpmf.org/pp/S10_2_PM120_1h20_Application%20of%20QMRA%20t
o%20go%20beyond%20safe%20harbors_Part2.pdf
Mendiburu F. 2009. Agricoale. Statistical procedures for agricultural research. Version
1.0-8. Lima, Peru: National University of Engineering in Lima-Peru [Accessed
2010 Apr 2]. Available from: http://tarwi.lamolina.edu.pe/~fmendiburu
Mennane, Z., Ouhssine, M., Khedid, K., Elyachioui, M., 2007. Hygienic quality of raw
cow’s milk feeding from domestic waste in two regions in Morocco. Int J Agric
Biol. 9, 46-48.
Michalac S, Alvarez V, Ji T, Zhang QH. 2003. Inactivation of selected microorganisms
and properties of pulsed electric field processed milk. J Food Process Preserv.
27: 137-151.
Miller EJ. 1986. Energy management in milk processing. In: Singh RP, editor. Energy in
Food Processing. Vol 1. Amsterdam, NL: Elsevier Science Publishers. 135-154.
Miller RB. 2005. Electronic irradiation of foods: an introduction to the technology. New
York, NY, USA: Springer Science+Business Media Inc. 295 p.
Muir DD, Tamime AY. 2001. Liquid milk. In: Tamime AY and Law BA, editors.
Mechanisation and automation in dairy technology. Sheffield: Sheffield Academic
Press. p 53-94.
Murano EA, Pierson MD. 1992. Effect of heat shock and growth atmosphere on the heat
resistance of Escherichia coli O157:H7. J Food Protect. 55(3): 171-175.
Murphy PM, Lynch D, Kelly PM. 1999. Growth of thermophilic spore forming bacilli in
milk during the manufacture of low heat powders. Int J Dairy Technol. 52: 45-50.
"&+! ! !
National Advisory Committee on Microbiological Criteria for Foods (NACMCF). 2006.
Requisite scientific parameters for establishing the equivalence of alternative
methods of pasteurization. J Food Prot. 69(5): 1190-1216.
Nebe-von Caron G, Stephens P, Bradley RA. 1998. Assessment of bacterial viability
status by flow cytometry and single cell sorting. J Appl Microbiol. 84: 988-998.
Needs E. 2001. High pressure processing of dairy products. In: Hendrickx MEG, Knorr
D, editors. Ultra high pressure treatments of foods. New York: Kluver
Academic/Plenum Publishes. p 269-296.
Ngadi M, Dehghannya J. 2010. Modeling microbial inactivation by a pulsed electric field.
In: Mohanned MF. 2010. Mathematical modeling of food processing (pp. 559-
573). Boca Raton, FL, USA: CRC Press.
Nicol RS, Kydas N, Robertson LJ. 2005. World standard energy practice. In: IDF.
Energy use in dairy processing. Bulletin of the International Dairy Federation No.
401. Brussels, Belgium. 3-9.
Nielsen WK. 2000. Membrane filtration and related molecular separation technologies.
Silkeborg, Denmark: APV Systems. 223 p.
Noci F, Walkling-Ribeiro M, Cronin DA, Morgan DJ, Lyng JG. 2009. Effect of
thermosonication, pulsed electric field and their combination on inactivation of
Listeria innocua in milk. Int Dairy J. 19: 30-35.
Norton T, Sun D. 2008. Recent advances in the use of high pressure as an effective
processing technique in the food industry. Food Bioprocess Technol. 1: 2-34.
"&"! ! !
Odriozola-Serrano I, Bendicho-Porta S, Martín-Belloso O. 2006. Comparative study on
shelf life of whole milk processed by high-intensity pulsed electric field or heat
treatment. Journal Dairy Sci. 89: 905-911.
Olesen N, Jensen F. 1989. Microfiltration. The influence of operation parameters on the
process. Milchwissenschaft. 44 (8): 476-479.
Oliver SP, Jayarao BM, Almeida RA. 2005. Foodborne pathogens in milk and the dairy
farm environment: food safety and public health implications. Foodborne Pathog
Dis. 2: 115-129.
Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA). 1990. Milk Act.
R.S.O. 1990. Chapter M. 12. [Accessed 2010 June 18]. Available from:
http://www.e-laws.gov.on.ca/html/statutes/english/elaws_statutes_90m12_e.htm
Ordoñez Pereda JA, Burgos González J, Raso Pueyo J, Lopez Buesa P, Cond'n Us'n
S, Sala Trepat FJ, inventors. Universidad de Zaragoza, asignee.1994 Feb 1.
Procedimiento para la destrucci'n de microorganismos y enzimas mediante la
aplicaci'n combinada de calor y ultrasonidos bajo presi'n:proceso MTS (Mano-
Termo-Sonicaci'n). Oficina Española de Patentes y Marcas No. 2 046 944.
Ormerod MG. 2008. Flow cytometry – a basic introduction – Wiki version. DeNovo
Software. http://flowbook-wiki.denovosoftware.com/
Orsat V, Raghavan GSV. 2005. Radio-frequency processing. In: Sun D, editor.
Emerging technologies for food processing. San Diego, CA: Elsevier Academic
Press. 445-468.
Otunola A, and El-Hag A, Jayaram S, Anderson WA. 2008. Effectiveness of pulsed
electric fields in controlling microbial growth in milk. Int J Food Eng. 4(7): 1-14.
"&#! ! !
Palmieri L, Cacae D. 2005. High intensity pulsed light technology. In: Sun D, editor.
Emerging technologies for food processing. San Diego, CA: Elsevier Academic
Press. p 279-306.
Palou E, López-Malo A. 2005. Growth/No-growth interface modeling and emerging
technologies. In: Barbosa-Cánovas GV, Tapia MS, Cano MP. Novel food
processing technologies. Boca Raton, FL: CRC Press. 629-651.
Payfylias I, Cheryan M, Mehaia MA, Saglam N. 1996. Microfiltration of milk with ceramic
membranes. Food Res Int. 29: 141-146.
Pedersen PJ. 1992. Microfiltration for the reduction of bacteria in milk and brine. In:
International Dairy Federation (IDF). New applications of membrane processes.
IDF special issue no. 9201. p 33-50.
Peleg M, Cole MB. 2000. Estimating the survival of Clostridium botulinum spores during
heat treatments. J Food Protect. 63 (2): 190-195.
Peleg M, Normand MD. 2004. Calculating microbial survival parameters and predicting
survival curves from non-isothermal inactivation data. Critical Rev Food Sci Nutr.
44: 409-418.
Peleg M, Penchina CM. 2000. Modeling microbial survival during exposure to a lethal
agent with varying intensity. Crit Rev Food Sci Nutr. 40 (2): 159-172.
Peleg M. 1995. A model of microbial survival after exposure to pulsed electric fields. J
Sci Food Agric. 67: 93-99.
Peleg M. 2006. Advanced quantitative microbiology for foods and biosystems. Boca
Raton, FL, USA: CRC Press. 417 p.
"&$! ! !
Pereira RN, Vicente AA. 2010. Novel technologies for milk processing. In: Engineering
aspects of milk and dairy products. Boca Raton, FL, USA: CRC Press. p 155-
174.
Picart L, Dumay E, Cheftel JC. 2002. Inactivation of Listeria innocua in dairy fluids by
pulsed electric fields: influence of electric parameters and food composition.
Innovat Food Sci and Emerging Technol. 3: 357-369.
Pillai SD, Braby LA. 2007. Electronic pasteurization. In: Tewari G, Juneja VK, editors.
Advances in thermal and non-thermal food preservation. Ames: Blackwell
Publishing. p. 195-202.
Rajowkski KT, Calderone SM, Jones E. 1994. Effect of polyphosphate and sodium
chloride on the growth of Listeria monocytogenes and Staphylococcus aureus in
ultra-high temperature milk. J Dairy Sci. 77: 1503-1508.
Raso J, Alvarez I, Condón S, Sala Trepat FJ. 2000. Predicting inactivation of
Salmonella senftenberg by pulsed electric fields. Innov Food Sci & Emerg
Technol. 21-29.
Raso J, Barbosa-Cánovas GV. 2003. Nonthermal preservation of foods using combined
processing technologies. Crit Rev Food Sci Nutr. 43(3): 265-285.
Ravishankar S, Juneja VK. 2003. Adaptation or resistance responses of
microorganisms to stresses in the food-processing environment. In: Yousef AE,
Juneja VK, editors. Microbial stress adaptation and food safety. Boca Raton, FL:
CRC Press. 105-158.
Ray B. 2004. Fundamental food microbiology. 2nd ed. Boca Raton, FL: CRC Press. 608
p.
"&,! ! !
Reina LD, Jin ZT, Zhang QH, Yousef AE. 1998. Inactivation of Listeria monocytogenes
in milk by pulsed electric fields. J Food Protect. 61(9): 1203-1206.
Rodrigo D, Barbosa-Cánovas GV, Martínez A, Rodrigo M. 2003. Weibull distribution
function based on an empirical mathematical model for inactivation of
Escherichia coli by pulsed electric fields. J Food Protect. 66 (6): 1007-1012.
Rodrigo D, Martínez A, Harte F, Barbosa-Cánovas GV, Rodrigo M. 2001. Study of
inactivation of Lactobacillus plantarum in orange-carrot juice by means of pulsed
electric fields: comparison of inactivation kinetics models. J Food Protect. 64 (2):
259-263.
Rodrigo D, Zúñiga M, Rivas A, Martínez A. 2007. Adaptation potential of
microorganisms treated by pulsed electric fields. In: Lelieveld HLM, Notermans S,
de Haan SWH. Food preservation by pulsed electric fields: from research to
application. Boca Raton, FL: CRC Press. 156-164.
Rodríguez-Calleja JM, Cebrián G, Condón S, Mañas P. 2006. Variation in resistance of
natural isolates of Staphylococcus aureus to heat, pulsed electric field and
ultrasound under pressure. J Appl Microbiol. 100: 1054-1062.
Rodríguez-Romo L, Yousef AE. 2005. Cross-protective effects of bacterial stress. In:
Griffiths MW. Understanding pathogen behaviour: virulence, stress response,
and resistance. Boca Raton, FL: CRC Press. 128-151.
Ross AIV, Griffiths MW, Mittal GS, Deeth HC. 2003. Combining nonthermal
technologies to control foodborne microorganisms. Int. J Food Microbiol. 89: 125-
138.
"&%! ! !
Ross T, Dalgaard P. 2004. Secondary models. In: McKellar, R. C., Lu, X. Modeling
microbial responses in food (pp. 63-150). Boca Raton: CRC Press.
Rosso L, Lobry JR, Bajard S, Flandrois JP. 1995. Convenient model to describe the
combined effects of temperature and pH on microbial growth. Appl Env Microbiol.
61 (2): 610-616.
Rowe MT, Kirk R. 1999. An investigation into the phenomenon of cross-protection in
Escherichia coli O157:H7. Food Microbiol. 16: 157-164.
Ruhlman KT, Jin ZT, Zhang QH. 2001. Physical properties of liquid foods for pulsed
electric field treatment. In: Barbosa-Cánovas GV, Zhang QH. Pulsed electric
fields in food processing: fundamental aspects and applications. Lancaster:
Technomic Publishing Company, Inc. p 45-56.
Ryu J, Beuchat LR. 1998. Influence of acid tolerance responses on survival, growth,
and thermal cross-protection of Escherichia coli O157:H7 in acidified media and
fruit juices. Int J Food Microbiol. 45: 185-193.
Saboya LV, Maubois J. 2000. Current developments of microfiltration technology in the
dairy industry. Lait. 80: 541-553.
Sampedro F, Rivas A, Rodrigo D, Martínez A, Rodrigo M. 2006. Effect of temperature
and substrate on PEF inactivation of Lactobacillus plantarum in an orange juice-
milk beverage. Eur Food Res Technol. 223: 30-34.
Sampedro F, Rivas A, Rodrigo D, Martinez A, Rodrigo M. 2007. Pulsed electric fields
inactivation of Lactobacillus plantarum in an orange juice–milk based beverage:
Effect of process parameters. J Food Eng. 80: 931-938.
"&&! ! !
Sampedro F, Rodrigo M, Martinez A, Rodrigo D, Barbosa-Canovas GV. 2005. Quality
and safety aspects of PEF application in milk and milk products. Crit Rev Food
Sci Nutr. 45(1): 25-47.
Sauer A, Moraru CI. 2009. Inactivation of Escherichia coli ATCC 25922 and Escherichia
coli O157:H7 in apple juice and Apple Cider, Using Pulsed Light Treatment. J
Food Protect. 72 (5): 937-944.
Saulis G, Wouters PC. 2007. Probable mechanisms of microorganism inactivation by
pulsed electric fields. In: Lelieveld HLM, Notermans S, de Haan SWH. Food
preservation by pulsed electric fields: from research to application. Boca Raton,
FL: CRC Press. p 138-155.
Sepúlveda DR, Góngora-Nieto MM, Guerrero JA, Barbosa-Cánovas GV. 2005.
Production of extended-shelf life milk by processing pasteurized milk with pulsed
electric fields. J Food Eng. 67: 81-86.
Sepúlveda DR, Góngora-Nieto MM, Guerrero JA, Barbosa-Cánovas GV. 2009. Shelf life
of whole milk processed by pulsed electric fields in combination with PEF-
generated heat. LWT Food Sci Technol. 42(3): 735-739.
Sepúlveda-Ahumada DR, Ortega-Rivas E, Barbosa-Cánovas, GV. 2000. Quality
aspects of cheddar cheese obtained with milk pasteurized by pulsed electric
fields. Food Bioprod Process. 78 (2): 65-71.
Shamsi, K., Versteeg, C., Sherkat, F., Wan, J., 2008. Alkaline phosphatase and
microbial inactivation by pulsed electric field in bovine milk. Innovative Food Sci
Emerg Technol. 9: 217-223.
"&(! ! !
Shapiro HM. 2003. Practical flow cytometry. 4th ed. New Jersey, NJ: John Wiley &
Sons. 681 p. Available from: http://www.coulterflow.com/bciflow/practical.php
Smiddy MA, Martin J, Huppertz T, Kelly AL. 2007. Microbial shelf-life of high-pressure-
homogenised milk. Int Dairy J. 17: 29-32.
Smith K, Mittal GS, Griffiths MW. 2002. Pasteurization of milk using pulsed electrical
field and antimicrobials. J Food Sci 67(6): 2304-2308.
Smith M. 2007. Regulatory acceptance. In: Lelieveld HLM, Notermans S, de Haan
SWH. Food preservation by pulsed electric fields: from research to application.
Boca Raton, FL: CRC Press. 352-357.
Sobrino-López A, Martín-Belloso O. 2008. Enhancing the lethal effect of high-intensity
pulsed electric field in milk by antimicrobial compounds as combined hurdles. J
Dairy Sci. 91: 1759-1768.
Sobrino-López A, Martín-Belloso O. 2009. Review: potential of high-intensity pulsed
electric field technology for milk processing. Food Eng Rev. Available online: DOI
10.1007/s12393-009-9011-7. 11 p.
Sobrino-López A, Raybaudi-Massilia R, Martín-Belloso O. 2006. High-intensity pulsed
electric field variables affecting Staphylococcus aureus inoculated in milk. J Dairy
Sci. 89: 3739-3748.
Southeastern Universities Research Association (SURA). 2005. Electromagnetic
Spectrum. Available at: http://www.electrical-res.com/EX/10-17-
19/SURA_Electromagnetic_Spectrum_Full_Chart.jpg
"&)! ! !
Stephens PJ, Joyson JA, Davies KW, Holbrok R, Lappin-Scott HM, Humprey TJ. 1997.
The use of an automated growth analyser to measure recovery times of single
heat-injured Salmonella cells. J Applied Microbiol. 83: 445-455.
Suhren G, Reichmuth J, Heeschen W. 1990. Bactoscan technique. In: International
Dairy Federation (FIL-IDF). Methods for assessing the bacteriological quality of
raw milk from the farm. Bulletin No. 256. FIL-IDF: Brussels, Belgium. 24-30.
Sumnu G, Sahin S. 2005. Recent developments in microwave heating. In: Emerging
technologies for food processing. San Diego, CA: Elsevier Academic Press. 419-
444.
Sun D. 2005. Emerging technologies for food processing. San Diego, CA, USA: Elsevier
Academic Press. 771.
Swarts AJ, Hastings JW, Roberts RF, von Holy A. 1998. Flow cytometry demonstrates
bacteriocin-induced injury to Listeria monocytogenes. Curr Microbiol. 36: 266-
270.
Te Giffel MC, Van Der Horst HC. 2004. Comparison between bactofugation and
microfiltration regarding efficiency of somatic cell and bacteria removal. Bulletin
IDF No. 389, 49-53.
Teixeira AA. 2007. Mechanistic models of microbial inactivation behaviour in foods. In:
Brul S, Van Gerwen S, Zwietering M. Modelling microorganisms in food.
Cambridge, England: Woodhead Publishing Ltd. 198-213.
Teo AY, Ziegler GR, Knabel SJ. 2001. Optimizing detection of heat-injured Listeria
monocytogenes in pasteurized milk. J Food Protect. 64 (7): 1000-1011.
"&*! ! !
Tewari G, Juneja VK. 2007. Advances in thermal and non-thermal food preservation.
Blackwell Publishing: Ames, IA, USA. 281 p.
Tewari G. 2007. Microwave and radio-frequency heating. In: Tewari G, Juneja VK,
editors. Advances in thermal and non-thermal food preservation. Ames, IA:
Blackwell Publishing. 91-98.
Todar K. 2008. Todar’s online textbook of bacteriology. University of Wisconsin-
Madison Department of Bacteriology. [Accessed 2010 May 15]. Available at:
http://www.textbookofbacteriology.net/
Toepfl S, Heinz V, Knorr D. 2007. High intensity pulsed electric fields applied for food
preservation. Chem Eng Process. 46: 537-546.
Toepfl S, Mathys A, Heinz V, Knorr D. 2006. Review: potential of high hydrostatic
pressure and pulsed electric fields for energy efficient and environmentally
friendly food processing. Food Rev Int. 22: 405-423.
Toepfl S. 2006. Pulsed electric fields (PEF) for permeabilization of cell membranes in
food- and bioprocessing – applications, process and equipment design and cost
analysis. [DPhil dissertation]. Berlin, Germany: Technical University of Berlin
[Accessed 2010 Apr 2]. Available from:
http://opus.kobv.de/tuberlin/volltexte/2006/1397/pdf/toepfl_stefan.pdf
Ulmer HM, Gänzle MG, Vogel RF. 2000. Effects of high pressure on survival and
metabolic activity of Lactobacillus plantarum TMW1.460. Appl Env Microbiol.
66(9): 3966-3973.
"(+! ! !
Ulmer HM, Heinz V, Gänzle MG, Knorr D, Vogel RF. 2002. Effects of pulsed electric
fields on inactivation and metabolic activity of Lactobacillus plantarum in model
beer. J Appl Microbiol. 93: 326-335.
Unal R, Yousef AE, Dunne CP. 2002. Spectrofluorimetric assessment of bacterial cell
membrane damage by pulsed electric field. Innov Food Sci Emerg. Technol. 3:
247-254.
United States Department of Agriculture (USDA). 2010. Processing intervention
technologies for enhancing the safety and security of fluid foods and beverages.
Food Safety and Intervention Technologies Unit Research Project. [Accessed
2010 June 18] Available from:
http://www.ars.usda.gov/research/projects/projects.htm?ACCN_NO=410488
Valdramidis VP, Geeraerd AH, Van Impe JF. 2007. Stress-adapatative responses by
heat under the microscope of predictive microbiology. J Appl Microbiol. 1922-
1930.
Van den Berg G. 2001. Semi-hard cheeses. In: Tamime AY, Law BA. Mechanisation
and automation in dairy technology. Sheffield: Sheffield Academic Press. p 225-
249.
Van den Berg RW, Hoogland BH, Lelieveld HLM, Van Schepdael L. 2001. High
pressure equipment designs for food processing applications. In: Hendrickx
MEG, Knorr D, editors. Ultra high pressure treatments of foods. New York: Kluver
Academic/Plenum Publishes. p 297-313.
Van den Bosch HFM. 2007. Chamber design and process conditions for pulsed electric
field treatment of food. In: Lelieveld HLM, Notermans S, de Haan SWH. Food
"("! ! !
preservation by pulsed electric fields: from research to application. Boca Raton,
FL: CRC Press. p 70-93.
Vasavada PC. 1988. Pathogenic bacteria in milk - a review. J Dairy Science. 71, 2809-
2816.
Vicente A, Castro IA. 2007. Novel thermal processing technologies. In: Advances in
thermal and non-thermal food preservation. Ames, IA: Blackwell Publishing. 99-
130.
Villar A, García JA, Iglesias L, García ML, Otero A. 1996. Application of principal
component analysis to the study of microbial populations in refrigerated raw milk
from farms. Int Dairy J. 6: 937-945.
Visvanathan C, Ben Aim R. 1989. Application of an electric field for the reduction of
particle and colloidal membrane fouling in crossflow microfiltration. Sep Sci
Technol. 24 (5 & 6): 383-398.
Wakeman RJ. 1998. Electrically enhanced microfiltration of albumin suspensions. Trans
IChemE. 76 (C): 53-59.
Walkling-Ribeiro M, Noci F, Cronin DA, Lyng JG, Morgan DJ. 2009. Antimicrobial effect
and shelf-life extension by combined thermal and pulsed electric field treatment
of milk. J Appl Microbiol. 106(1): 241-248.
Walkling-Ribeiro M, Noci F, Cronin DA, Lyng JG, Morgan DJ. 2010. Shelf life and
sensory attributes of a fruit smoothie-type beverage processed with moderate
heat and pulsed electric fields. LWT – Food Sci Technol. 43 (7): 1067-1073.
Walkling-Ribeiro M, Rodríguez-González O, Jayaram S, Griffiths MW. 2010. Microbial
inactivation and shelf life comparison of ‘cold’ hurdle processing with pulsed
"(#! ! !
electric fields and microfiltration, and conventional thermal pasteurisation in skim
milk. Int Food Microbiol (in press). DOI:10.1016/j.ijfoodmicro.2010.10.023.
Walkling-Ribeiro M. 2008. The impact of high voltage pulsed electric fields in hurdle
strategies on preservation and selected quality parameters of minimally
processed non-alcoholic beverages. DPhil thesis. Dublin: UCD Dublin. 210 p.
Walstra P, Wouters JTM, Geurts TJ. 2006. Dairy science and technology. Second
edition. Boca Raton: CRC Press. 782 p.
Wang Y, Hammes F, Düggelin M, Egli T. 2008. Influence of size, shape, and flexibility
on bacterial passage through micropore membrane filters. Environ Sci Technol.
42: 6749-6754.
Wardrop Engineering 1997. Guide to energy efficiency opportunities in the dairy
industry processing industry. Natural Resources Canada. Ottawa, Canada.
[Accessed 2010 May 15]. Available from:
http://oee.nrcan.gc.ca/Publications/industrial/M27-01-827E/index.cfm?attr=24
Wilmarante SK, Farid MM. 2008. Pressure assisted thermal sterilization. Food Bioprod
Process. 86(4): 312-316.
Wilson MJ, Baker R, inventors. Kal Kan Foods Inc, assignee. 2000 Dec 14. High
temperature/ultra-high pressure sterilization of foods. US Patent 6,086,936.
Wouters PC, Alvarez I, Raso J. 2001. Critical factors determining inactivation kinetics by
pulsed electric field food processing. Trends in Food Science and Technology 12,
112-121.
"($! ! !
Wouters PC, Bos AP, Ueckert J. 2001. Membrane permeabilization in relation to
inactivation kinetics of Lactobacillus species due to pulsed electric fields. Appl
Env Microbiol. 67 (7): 3092-3101.
Wu VCH, Fung DYC. 2001. Evaluation of thin agar layer method for recovery of heat-
injured foodborne pathogens. J Food Sci. 66 (4): 580-583.
Yamaguchi N, Sasada M, Yamanaka M, Nasu M. 2003. Rapid detection of respiring
Escherichia coli O157:H7 in apple juice, milk, and ground beef by flow cytometry.
Cytometry. 54A: 27-35.
Yousef AE, Courtney PD. 2003. Basics of stress adaptation and implications in new-
generation foods. In: Yousef AE, Juneja VK, editors. Microbial stress adaptation
and food safety. Boca Raton, FL: CRC Press. 105-158.
Zárate V, Belda F, Pérez C, Cardell E. 1997. Changes in the microbial flora of Tenerife
goats’ milk cheese during ripening. Int Dairy J. 7: 635-641.
Zenker M, Heinz V, Knorr D. 2003. Application of ultrasound-assisted thermal
processing for preservation and quality retention of liquid foods. J Food Protect.
66: 1642-1649.
Zhang Q, Barbosa-Cánovas CV, Swanson BG. 1995. Engineering aspects of pulsed
electric field pasteurization. J Food Eng. 25: 261-281.
Zhang Y, Griffiths MW. 2003. Induced expression of the heat shock protein genes uspA
and grpE during starvation at low temperatures and their influence on thermal
resistance of Escherichia coli O157:H7. J Food Protect. 66 (11): 2045-2050.
"(,! ! !
Zhao W, Yang RJ, Tang YL, Zhang WB, Hua X. 2009. Investigation of the protein-
protein aggregation of egg white proteins under pulsed electric fields. J Agric
Food Chem. 57: 3571-3577.
Zwietering MH, De Koos JT, Hasenack BE, De Wit JC, Van ‘t Riet K. 1991. Modeling of
bacterial growth as a function of temperature. Appl Env Microbiol. 57 (4): 1094-
1101.
"(%! ! !
7. Annex A1. Relative scale of energy sources at the electromagnetic spectrum.
1012 109 106 103 1 10-3 10-6 10-9 10-12 10-15
Energy 1024 1021 1018 1015 1012 109 106 103 1 (./)
Frequency Å nm %m mm m km (()*+!,-") 10-18 10-15 10-12 10-9 10-6 10-3 1 103 106 109
Wavelength X Thermal Hertzian (m) Electro- Cosmic Rays Radiation Waves Electric "#$! magnetic Rays Gamma Micro- Radio Power spectrum Rays wave Waves
0.1 1 10 100
Wavelength Thermal Visible Near Intermediate Far (%&) UV Radiation Light Infrared Infrared Infrared
400 500 600 700
Wavelength Visible Violet Indigo Blue Green Yellow Orange Red ('&) Light
Adapted from LASP 2010, SURA 2005, and McBee 1996.
! !