International Journal of Food Microbiology 173 (2014) 105–111

Contents lists available at ScienceDirect

International Journal of Food Microbiology

journal homepage: www.elsevier.com/locate/ijfoodmicro

Pulsed electric field processing of different fruit juices: Impact of pH and

temperature on inactivation of spoilage and pathogenic micro-organisms
R.A.H. Timmermans a,⁎, M.N. Nierop Groot a, A.L. Nederhoff a, M.A.J.S. van Boekel b, A.M. Matser a, H.C. Mastwijk a
a
Wageningen UR Food & Biobased Research, P.O. Box 17, 6700 AA Wageningen, The Netherlands
b
Food Quality & Design Group, Wageningen University, P.O. Box 8129, 6700 EV Wageningen, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Pulsed electrical field (PEF) technology can be used for the inactivation of micro-organisms and therefore for
Received 4 September 2013 preservation of food products. It is a mild technology compared to thermal pasteurization because a lower tem-
Received in revised form 23 December 2013 perature is used during processing, leading to a better retention of the quality. In this study, pathogenic and spoil-
Accepted 23 December 2013
age micro-organisms relevant in refrigerated fruit juices were studied to determine the impact of process
Available online 31 December 2013
parameters and juice composition on the effectiveness of the PEF process to inactivate the micro-organisms. Ex-
Keywords:
periments were performed using a continuous-flow PEF system at an electrical field strength of 20 kV/cm with
Pulsed electric field (PEF) variable frequencies to evaluate the inactivation of Salmonella Panama, Escherichia coli, Listeria monocytogenes
Inactivation kinetics and Saccharomyces cerevisiae in apple, orange and watermelon juices.
Fruit juice Kinetic data showed that under the same conditions, S. cerevisiae was the most sensitive micro-organism, follow-
Weibull ed by S. Panama and E. coli, which displayed comparable inactivation kinetics. L. monocytogenes was the most re-
sistant micro-organism towards the treatment conditions tested. A synergistic effect between temperature and
electric pulses was observed at inlet temperatures above 35 °C, hence less energy for inactivation was required
at higher temperatures. Different juice matrices resulted in a different degree of inactivation, predominantly de-
termined by pH. The survival curves were nonlinear and could satisfactorily be modeled with the Weibull model.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction result of electroporation, an enhanced permeability across the mem-

brane occurs (electropermeabilization) that, depending on the applied
There is a demand for mild preservation processes that can enhance electric field, leads to cell death (irreversible, complete inactivation)
the shelf life of high quality foods, without affecting the quality and safe- or resealing of the cell to its initial viable state (reversible, sub-lethal
ty of the food products or using chemical preservatives. Pulsed electric or partial inactivation). For PEF processing, strong electric fields
field (PEF) processing is one of these mild preservation techniques (5–50 kV/cm) are applied to ensure irreversible effects, leading eventu-
that has been investigated during the last decades as an alternative for ally to cell death.
thermal processing without compromising sensorial and nutritional A considerable number of studies have been conducted to determine
properties of food (Hartyáni et al., 2011; Vervoort et al., 2011). the inactivation kinetics of micro-organisms by PEF in liquid food prod-
PEF effectiveness in the inactivation of vegetative bacterial cells, ucts. Comparison of the published experimental data is not straightfor-
yeast and molds, has been demonstrated, however, microbial spores ward, since the degree of microbial inactivation is strongly dependent
are resistant to PEF treatment (Grahl and Markl, 1996; Van Heesch on the type and design of the equipment, pulse shape, and intensity of
et al., 2000; Wouters et al., 2001). Therefore, PEF should be considered the pulses in terms of field strength, energy and number of pulses ap-
as a pasteurization method, and fruit juices are a suitable candidate plied (Heinz et al., 2003). Moreover, inactivation is influenced by the
product, since the acid conditions of these products control the germi- properties and type of (food) matrix, and by the potential addition of
nation of microbial spores (Raso and Barbosa-Cánovas, 2003). preservatives (García et al., 2005; Gurtler et al., 2011; Saldaña et al.,
The mechanism of PEF processing to inactivate micro-organisms is 2011).
not fully elucidated, but could be explained as a combination of electro- A large number of studies have been carried out in buffer systems in
poration and electropermeabilization (Teissié et al., 2005; Weaver and order to gain insight into the mechanism of inactivation and process
Chizmadzhev, 1996; Wouters et al., 2001). Induction of an (external) conditions required for microbial inactivation using varying conditions,
electric field on the membrane of a micro-organism can lead to local in- including electrical conductivity, pH, water activity and ionic composi-
stabilities, resulting in pores in the membrane (electroporation). As a tion (Álvarez et al., 2003b, 2003c; Aronsson and Rönner, 2001;
Aronsson et al., 2005; Gómez et al., 2005a, 2005b; Toepfl et al., 2007;
⁎ Corresponding author. Tel.: +31 317481305; fax: +31 317483011. Wouters et al., 1999). Inactivation studies of micro-organisms in fruit
E-mail address: [email protected] (R.A.H. Timmermans). juices using PEF have been described for a number of studies (Elez-

0168-1605/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ijfoodmicro.2013.12.022
106 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111

Martínez et al., 2004a, 2004b; Gómez et al., 2005a, 2005b; Gurtler et al., Table 2
2010; Heinz et al., 2003; Mosqueda-Melgar et al., 2007; Rodrigo et al., pH and conductivity of juice matrices after inoculation with micro-organisms.

2001, 2003; Saldaña et al., 2011; Qin et al., 1995), however only a single Matrix pH Conductivity [S/m]
micro-organism or one type of juice was selected to evaluate the effect
Apple juice 3.5 0.26
of PEF. Here, we describe a comparative analysis including species Orange juice 3.7 0.38
representing spoilage and pathogenic micro-organisms in varying Watermelon juice 5.3 0.30
juices using treatment conditions based on an existing industrial Watermelon juice 3.6 0.35
scale system for fresh fruit juice processing with a capacity of
1500 L/h and an electrical field strength of 20 kV/cm.
For industrial implementation, it is of paramount importance to de- 2.3. Micro-organisms and culture conditions
termine and validate effective process conditions and constraints in real
food systems with pathogens or realistic surrogates to minimize micro- Pathogenic and spoilage micro-organisms were selected based on
bial hazards. Furthermore, it is important to test this on a representative their association with and prevalence in fruit juices, their ability to sur-
PEF system, with the same configuration as the industrial system, as the vive and to grow in the juice matrices selected in this study and if
configuration of the system has a considerable effect on the effective- known, their resistance towards PEF, as this is inherent to species and
ness of inactivation. strain level (Gurtler et al., 2010). The micro-organisms were tested in
The objectives of this study were to investigate inactivation of spoil- fresh NFC juices and FC juices with variation in pH, and evaluated
age and pathogenic micro-organisms in fruit juice, to assess the impact after 1, 4 and 7 days of incubation. Comparable results of survival
of process parameters and juice composition on the effectiveness of the were found between both types of juices. The mentioned criteria lead
PEF process, and to find a suitable model to predict the minimum re- to the choice for the micro-organisms shown in Table 1. No survival
quired process conditions. (i.e., decline in number of micro-organisms) was observed in high acidic
juices (pH 1.8), survival (i.e., a stable number of micro-organisms) was
observed in acid juices (pH 3.6) and growth (i.e., an increase in number
2. Material and methods of micro-organisms) was observed for all micro-organisms in weak acid
juices (pH 5.3).
2.1. Selection of juice products Fresh cultures of the yeast strain Saccharomyces cerevisiae (Table 1)
were prepared by plating from frozen stocks on glucose-peptone-
Fruit juices used in this study were fresh, non-heated not-from- yeast (GPY) agar plates, containing 40 g glucose, 5 g peptone, 5 g
concentrate (NFC) juices delivered by a commercial producer of fruit yeast extract and 15 g agar per 1 L distilled water. Plates were incubated
juices directly after production (orange juice, orange-strawberry juice, overnight at 25 °C. One single colony isolate was used for inoculation of
tropical juice) or delivered frozen at −18 °C (lime juice). Watermelon a 50 mL flask containing 10 mL GPY broth and cells were cultivated for
juice used to evaluate the survival of relevant micro-organisms was ex- 24 h at 25 °C in a shaking air incubator (Innova, 180 rpm). Of this over-
tracted directly from watermelons obtained from a local supermarket, night culture, 0.1 mL was used for inoculation of 9.9 mL fresh GPY me-
using a table juice extractor type HR 1861 (Philips) to obtain watermel- dium supplemented with 1% of glucose in a 50 mL flask and incubated
on juice. One single batch of 30 L juice was squeezed, mixed and frozen for exactly 24 h at 20 °C and 180 rpm.
in batches of 1 L. Prior to an experiment, the watermelon juice was Fresh cultures of Salmonella enterica serotype Panama and
thawed and prepared similar as the other juices. Since the matrix com- Escherichia coli strains (Table 1) were prepared from frozen stock cul-
position of NFC juices is subject to seasonal variation, apple juice tures that were plated on TSB (Oxoid) agar plates and incubated over-
(Appelsientje Goudappel) and orange juice (Minute Maid) from con- night at 37 °C. Listeria monocytogenes strain (Table 1) was cultivated
centrate (FC) were used for inactivation kinetics studies to ensure re- on BHI (Oxoid) agar plates and incubated overnight at 30 °C. A single
producibility of experiments over time. colony isolate was used to inoculate a 50 mL flask with 10 mL TSB
(S. Panama, E. coli) or BHI broth (L. monocytogenes) and cultivated for
24 h at 20 °C in an Innova shaking incubator (180 rpm). Of this culture,
2.2. Preparation and analysis of juices 0.1 mL was used to inoculate 9.9 mL fresh TSB or BHI broth supple-
mented with 1% glucose (50 mL flask) and incubated for 24 h at 20 °C
For all experiments, juices were pasteurized by heating them in glass and 180 rpm.
bottles filled with 1000 mL of fruit juice for 45 min in a water bath at
85 °C. For inactivation kinetics, orange juice and watermelon juice 2.4. Determination of PEF inactivation kinetics
were sieved after pasteurization using a sterile sieve (0.225 mm pore
size) to remove pulp, while apple juice did not contain any pulp. Inactivation kinetics of selected spoilage and pathogenic micro-
Electric conductivity (Hi 933000, Hanna instruments) and pH organisms associated with fruit juice were determined to identify opti-
(Metrohm 744, Metrohm, Herisa, Switzerland) were measured of each mal PEF conditions for inactivation of the target organisms in apple juice
juice after inoculation with micro-organisms, and shown in Table 2. and orange juice (FC) and watermelon juice (NFC). Inactivation kinetics
For some experiments, the pH of watermelon juice was changed were determined for single strain pathogens (or a pathogen surrogate)
from 5.3 to 3.6 by adding HCl. or spoilage micro-organisms rather than for a cocktail of strains.

Table 1
Bacterial strains and yeast strains used in this study.

Species/strains used Culture collection Source of isolation Reference

Saccharomyces cerevisiae CBS 1544 Fermenting fruit juice isolate Zhang et al. (1994)
Salmonella Panama 10908 NA Human outbreak isolate Noël et al. (2010)
Escherichia coli ATCC 35218 PEF resistant surrogate for E. coli O157:H7 Gurtler et al. (2010)
Listeria monocytogenes Scott A Clinical isolate Fleming et al. (1985)

CBS: Centraal Bureau voor Schimmelcultures (Fungal Biodiversity Centre, Utrecht, The Netherlands).
ATCC: American Type Culture Collection, USA.
NA: not applicable.
 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111 107

Overnight grown cells of S. cerevisiae (GPY), S. Panama and E. coli (Tektronix Inc., Beaverton, USA). Energy balance was made up and correct
(TSA) or L. monocytogenes (BHI) were pelleted by centrifugation for at least 90%, according to Mastwijk et al. (2007).
(4000 rpm, 5 min) at 20 °C, washed with 100 mL PSDF and resuspend- Specific energy used for each condition was calculated according to
ed in 100 mL juice. Cells were diluted (1/10 ratio) in 900 mL juice in a Eq. (1), where w is the specific energy (kJ/kg), Tout is the outlet temper-
glass bottle to reach approximately 107 cfu/mL (S. cerevisiae) or ature (°C), Tin is the inlet temperature (°C) and cp is the specific heat ca-
108 cfu/mL (E. coli, S. Panama, L. monocytogenes). pacity, which is 3.8 kJ/kg·K for fruit juice of 10–12° Brix.
Two individual one liter bottles were inoculated with the selected
microbial strain and used to perform duplicate PEF experiments. w ¼ ðTout –Tin Þ  cp ð1Þ
All samples, prior to or after PEF treatment were immediately placed
on ice after collection. The number of viable cells was determined by
plating 100 μL of serially diluted PEF-treated juice in sterile peptone 2.6. Experimental design and statistical analysis
physiological salt diluent (PSDF) in duplicate on suitable agar plates
supplemented with 0.1% sodium pyruvate to enhance outgrowth of Data was collected for each species in two independent PEF experi-
sub-lethally damaged cells (McDonald et al., 1983; Sharma et al., ments, with cleaning and sterilization between the two experiments.
2005). Surviving cells were enumerated after 5–7 days at 25 °C Different inactivation models were evaluated (linearized first order
(S. cerevisiae), at 37 °C (S. Panama, E. coli) and 30 °C (L. monocytogenes). model, exponential decay, sigmoidal and Weibull). Based on the good-
ness of the fit, evaluated by calculating R2 and RMSE values, it was
2.5. PEF processing found that the empirical model of Weibull was the one that fitted best
to all data.
The PEF system used was a continuous-flow system, where the con- A mathematical model based on the Weibull distribution was there-
figuration was a downscaled copy from the pilot-scale PEF equipment fore used to fit the survival data (log10 survival fraction vs. specific ener-
described by Mastwijk, (2006) and Timmermans et al. (2011) and gy) of all micro-organisms, as shown in Eq. (2) where N is the number of
from an industrial scale system of 1500 L/h capacity. Specific attention micro-organisms that survived the treatment, N0 is the initial number of
was paid to design criteria, to guarantee the homogeneity of the electri- microbial population, w is the specific energy used (kJ/kg) and α and β
cal field when downscaling the treatment device. Higher field strengths are the two parameters of the distribution; α is called the scale param-
than 20 kV/cm will lead to arcing and the formation of toxic substances eter (a characteristic of the specific energy used) and β is the shape pa-
(Mastwijk, 2006), and are therefore not used in this study. rameter (Van Boekel, 2002).
The inoculated juice was pumped by means of a peristaltic pump
N wβ
(Masterflex L/S pump, Cole-Parmer) through a 6 mm (disposable) Log ¼ ð2Þ
silicone hose (Masterflex 6424-16, Cole-Parmer) at a flow rate of N0 α
14–16 mL/min. Before PEF treatment, the juice was preheated, by
heating in continuous flow through a 1 m × 6 mm diameter SS316 Parameters were estimated via nonlinear regression using least-
heat spiral that was immersed in a water bath at 43 ± 1 °C. Next, the squares. Individual results were modeled using Gnu-plot.
liquid was pumped through a co-linear PEF treatment device, compris-
ing two vertically positioned treatment chambers with each a diameter 3. Results and discussion
of 1.0 mm and a gap of 2.0 mm, involving a residence time of 13.5 μs in
the treatment chambers at a flow rate of 14 mL/min. The temperature 3.1. Influence of inlet temperature prior to PEF treatment
at the outlet was measured using a 0.3 mm thermocouple type K
(TM-914C, Lutron, Taiwan), and yielded 36 ± 1 °C when juice was Initial inactivation experiments were performed in apple juice at an
preheated and PEF was turned off. Subsequently, the juice was pumped inlet temperature of 20 °C and an electric field strength of 20 kV/cm.
through a cooling spiral of 2 mm diameter and 500 cm length cooled on Under the tested conditions, inactivation of S. Panama was virtually ab-
melting ice. After cooling, the samples of the liquid were collected at the sent up to an input of specific energy of 60 kJ/kg. Further increase of
exit under aseptic conditions. The throughput of the juice in the system specific energy with a concomitant temperature rise did inactivate
was periodically measured using a digital scale analytical balance (Sar- cells. To substantiate whether the inactivation resulted from the higher
torius, Gottingen, Germany) by recording the weight of the collected specific energy input or that cells were more susceptible to PEF at ele-
juice. vated temperature, the experiment was repeated using an inlet temper-
Prior to the experiment, all parts that came into contact with the ature of 36 °C. This experiment confirmed that at 36 °C S. Panama cells
juice were sterilized by moist heat in an autoclave for 20 min at were inactivated at lower specific energy input (Fig. 1A). Similar find-
120 °C. The system was started using 1 L of pasteurized juice to fill the ings were obtained for S. cerevisiae, where the inactivation was en-
system and to obtain stationary processing conditions of flow in the hanced by an elevated inlet temperature from 20 °C to 36 °C (Fig. 1B).
range of 14–16 mL/min. A sample of this start-up liquid was taken at Data showed that the initial temperature had a strong influence on
the exit as a negative control. After this, the pasteurized juice at the the specific energy required to inactivate the micro-organisms, with
inlet was replaced by a bottle of inoculated juice. The start over occurred less specific energy required to obtain a similar level of inactivation at
at the most intense PEF conditions (i.e. highest repetition rate). A sta- elevated temperatures. Similar observations have been described for
tionary state was reached before the inoculated juice reached the PEF Listeria innocua and E. coli (Heinz et al., 2003; Toepfl et al., 2007;
treatment chamber. The PEF treated juice was collected every 5 min Wouters et al., 1999).
after the repetition frequency was set to lower levels. The collected sam- The effect of the starting temperature on the efficacy of the PEF
ples were treated at the following repetition frequencies and number of treatment may be explained by the temperature dependent character-
pulses are indicated when a flow rate of 14 mL/min was used: 964 Hz istics of the membrane of the micro-organisms. Phase transitions of
(13 pulses), 785 Hz (10.6 pulses), 650 Hz (8.8 pulses), 560 Hz (7.5 the phospholipids from gel to liquid-crystalline phase are temperature
pulses), 390 Hz (5.3 pulses), 270 Hz (3.6 pulses), 220 Hz (3.0 pulses), related, which affects the stability of the cell membrane at higher tem-
180 Hz (2.4 pulses), 140 Hz (1.9 pulses), and 120 Hz (1.6 pulses) and fi- peratures (Stanley and Parkin, 1991). The critical membrane break-
nalized by sampling of a positive control (no treatment). down potential decreases when the temperature of the solution
Monopolar pulses of 2.0 μs duration at a field strength of 20 kV/cm increases (Coster and Zimmermann, 1975), and consequently electro-
were used. Pulse waveform, voltage, and intensity in the treatment cham- poration occurs at lower external electrical fields. This suggests, togeth-
bers were recorded with a digital oscilloscope Tektronix TDS3052B er with data shown in Fig. 1, a synergistic effect between temperature
108 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111

0 Temperature [°C]
26°C 30°C
-1
35°C
A 0
36 41 46 51 56 61
40°C
-2 -1 A
log (N/N0) [-]

43°C 46°C 49°C

log (N/N0) [-]

-3 -2

-4 48°C -3
52°C -4
-5
-5
-6
56°C 57°C -6
-7
0 20 40 60 80 100 -7
0 20 40 60 80 100
Specific energy [kJ/kg]
Specific energy [kJ/kg]

0
25°C Temperature [°C]
-1 38°C 29°C B 36 41 46 51 56 61
0
-2
log (N/N0) [-]

40°C 32°C
-1 B
-3 35°C

log (N/N0) [-]

41°C -2
-4 -3
37°C
-5 42°C -4

-6 -5
-6
-7
0 20 40 60 80 100 -7
Specific energy [kJ/kg] 0 20 40 60 80 100
Specific energy [kJ/kg]
Fig. 1. Reduction of viable counts of A) S. Panama added to apple juice after PEF treatment,
preheated at 20 °C (◊) and 36 °C (♦) and B) S. cerevisiae precultured in apple juice after
Temperature [°C]
PEF treatment, preheated at 20 °C (□) and 36 °C (■).
36 41 46 51 56 61
0
and electrical pulses, implying that every additional PEF pulse is more -1 C
effective than the previous one: the temperature increases as a result
log (N/N0) [-]

-2
of pulsing, leading to a weaker cell and making it more vulnerable to-
-3
wards the next PEF pulse.
-4
For further experiments, an inlet temperature of 36 °C was chosen.
At inlet temperatures exceeding 36 °C, the outlet temperature reaches -5
critical levels where product quality was compromised. -6
-7
3.2. PEF effectiveness towards different microbial species in diverse juice 0 20 40 60 80 100
matrices Specific energy [kJ/kg]

The inactivation of S. cerevisiae, S. Panama, E. coli and L. monocytogenes Fig. 2. Reduction of viable counts of S. Panama (◊), E. coli (▲), L. monocytogenes (○) and
S. cerevisiae (□) added to apple juice, pH 3.5 (A), orange juice, pH 3.7 (B), and watermelon
by PEF treatment was investigated in apple juice (Fig. 2A), orange juice
juice, pH 5.3 (C) after PEF treatment, preheated at 36 °C.
(Fig. 2B) and watermelon juice (Fig. 2C). All species were susceptible
to PEF treatment in apple juice, and inactivation curves obtained
showed a non-linear response (Fig. 2A). With the exception of characteristic dimensions (Heinz et al., 2001). As L. monocytogenes is
L. monocytogenes, similar findings were obtained for orange juice much smaller (short rods, 0.4–0.5 μm × 0.5–2 μm) than S. Panama
(Fig. 2B), but a reduced susceptibility was pronounced in watermelon (straight rods, 0.7–1.5 μm × 2–5 μm) and E. coli (straight rods, 1.1–
juice (Fig. 2C). PEF sensitivity followed the order S. cerevisiae N 1.5 μm × 2.0–6.0 μm), it costs more energy to inactivate this micro-
S. Panama N E. coli N L. monocytogenes, meaning that the energy ex- organism, hence it requires less energy to inactivate the large
pense to inactivate the bacteria is higher than for yeast. Therefore, pro- S. cerevisiae (ellipsoidal shape, 3–15 μm × 2–8 μm) (characteristic di-
cess validation using yeast species can lead to an overestimation of mensions taken from Bergey (1986)). A second argument is that
effectiveness of the process conditions. Validation designs that include L. monocytogenes is a Gram-positive bacterium with a cell membrane
(pathogenic) bacteria on identical small scale equipment are therefore structure different from the other tested bacteria that are Gram-
preferred. negative. Research of Hülsheger et al. (1983), and Toepfl et al. (2007),
The reduced sensitivity of L. monocytogenes towards electric pulses also described that this could have an influence on PEF sensitivity.
compared to that of other micro-organisms might be explained by two The reduced sensitivity of L. monocytogenes towards PEF was also
factors. First, the size and shape of the micro-organisms affect the re- found by Gómez et al. (2005a), and Saldaña et al. (2010), where at
quired electric field to lethally damage the cells, where at smaller cell field strengths up to 20 kV/cm inactivation was virtually absent. In-
sizes, a lower membrane potential is induced by an external field, leading crease of the field strength up to 35 kV/cm significantly improved sen-
to a higher microbial resistance to the treatment (Álvarez et al., 2006; sitivity of L. monocytogenes towards PEF.
Hülsheger et al., 1983; Zimmermann et al., 1974). Moreover, the shape The high resistance of L. monocytogenes in orange juice shown in this
of the micro-organism has an influence on the membrane potential, study is remarkable, showing virtually no inactivation, where for apple
where a rod-shaped cell requires an electric field more than five times juice, 3.5 log inactivation at a specific energy input of 87 kJ/kg was
stronger than that required by a spherical shaped cell with the same reached. There was a slight difference in the pH of apple juice
 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111 109

Temperature [°C] Protons may enter the cell by passive influx or via proton dependent
36 41 46 51 56 61 transporters. Reduction of the intracellular pH thereby affects the bio-
0
chemical processes in the cell including transport processes over the
-1 membrane, redox state, and enzyme activities (Cotter and Hill, 2003).
log (N/N0) [-]

-2 Similar observations in pH difference were found for Gram-positive

-3 and Gram-negative micro-organisms in buffer solutions (Álvarez et al.,
-4
2002; García et al., 2005; Geveke and Kozempel, 2003; Gómez et al.,
2005a, 2005b; Saldaña et al., 2010; Wouters et al., 1999).
-5
-6 3.3. Mathematical modeling
-7
0 20 40 60 80 100
Microbial inactivation data for the different juice matrices were used
Specific energy [kJ/kg] to construct a predictive model for an optimal PEF process. The empirical
model of Weibull (Eq. (2)) was fitted to all inactivation data obtained
Fig. 3. Reduction of viable counts of S. Panama (◊) and S. cerevisiae (□) in watermelon juice
after addition of HCl to pH 3.6 and PEF treatment, preheated at 36 °C.
from plotting logarithmic inactivation of the survival fraction against spe-
cific energy. The estimated parameters α and β obtained from the differ-
ent micro-organisms in the tested juices are shown in Table 3. The
goodness of the fit was evaluated by calculating R2 and RMSE values.
(pH 3.5) and orange juice (pH 3.7) which could have influence on the The determination coefficient R2 for each model of micro-organism in
observed difference in sensitivity. Similar findings were reported previ- varying medium was higher than 0.93, which means that less than 7%
ously for heat treated E. coli O157:H7 in apple and orange juices with a of the total response variation remained unexplained by the Weibull
higher resistance in orange juice (Mazzotta, 2001). The reduced sensi- equation. The values for parameter RMSE were in the range of 0.01 to
tivity of L. monocytogenes was pronounced even more in the watermel- 0.56, and can be assumed close to the observed data (Saldaña et al., 2011).
on matrix (pH 5.3) (Fig. 2C), and a remarkably higher resistance was Another test for judging the applicability of the Weibull model is by
observed for the other tested species as well. To determine if this higher plotting the observed data to the model calculations, as shown in Fig. 4.
PEF resistance of the micro-organisms in watermelon juice could (in The difference between a point in the graph and the line of equivalence
part) be explained by a pH effect or, alternatively, induced by the pres- is a measure of the inaccuracy of the corresponding estimation. Since
ence of other components in watermelon juice, experiments were car- the data points of all micro-organisms are randomly distributed above
ried out where the pH of watermelon juice was decreased with HCl and below the equivalence line, no systematic tendency is found, and
from pH 5.3 to 3.6. The results, presented in Fig. 3, show an enhanced therefore the Weibull model studied was satisfactory in terms of de-
performance of PEF inactivation of micro-organisms when the pH was scribing all data being analyzed.
reduced. The shape of the inactivation curve of micro-organisms in wa- As an indication for the quality of the estimated model, the statistical
termelon juice with an adapted pH is comparable to the curves shown correlation between the parameters α and β (expressed in the correla-
of apple and orange juices, which suggest that the observed differences tion coefficient) has been calculated. If this correlation coefficient is
were caused by pH effects. N0.99, it signals that the parameter estimates could not be well estimat-
It is assumed, that the role of pH of the medium in the inactivation of ed in the regression procedure (van Boekel., 2002). The results show
micro-organisms is related to the fact that most micro-organisms main- that there is no such problem with the current dataset. Modeling of
tain the cytoplasmic pH near neutrality (Corlett and Brown, 1980). The the data where the pH of watermelon juice was decreased showed an
pH of the medium affects the effectivity of weak organic acids, present increase in the statistical correlation.
in fruits, by influencing the ratio of non-dissociated and dissociated or- The shape factor β in Table 3 indicates that the survival curves of
ganic acids. A low pH favors formation of non-dissociated organic acids, S. Panama, E. coli, S. cerevisiae and L. monocytogenes fitted with the
the form that may pass the cell membrane and enter the cytoplasm. In- Weibull model were all concave downward (β N 1), which was also ob-
side the cell it can dissociate in the cytoplasm and lead to acidification. vious from Figs. 1, 2 and 3. Downward concavity (β N 1) indicates that

Table 3
α and β values estimated from the fitting of the mathematical model based on the Weibull distribution to experimental data for different micro-organisms in varying juices and the cal-
culated specific energy based necessary for a 5 log reduction.

Micro-organism Medium pH α (kJ/kg) (95% CI)a β (95% CI)a R2b RMSEc Correlation Calculated specific
between α and β energy necessary for 5 log
reduction [kJ/kg]d

Salmonella Panama Apple juice 3.5 20.52 (18.39–22.65) 1.40 (1.29–1.51) 0.96 0.19 0.98 65
Orange juice 3.7 23.63 (21.79–25.47) 1.38 (1.25–1.51) 0.99 0.20 0.96 76
Watermelon juice 5.3 70.15 (67.36–72.95) 2.60 (1.86–3.34) 0.96 0.10 0.39 130
Watermelon juice 3.6 41.12 (38.76–43.47) 3.05 (2.37–3.73) 0.99 0.19 0.86 70
E. coli Apple juice 3.5 15.15 (8.65–21.65) 1.08 (0.72–1.44) 0.93 0.56 0.97 67
Orange juice 3.7 16.39 (13.16–19.62) 1.27 (0.99–1.55) 0.97 0.26 0.95 58
Watermelon juice 5.3 71.29 (65.89–76.68) 1.58 (1.16–2.01) 0.94 0.09 0.62 197
Saccharomyces Apple juice 3.5 29.56 (24.36–34.75) 2.19 (1.64–2.74) 0.96 0.53 0.96 62
cerevisiae Orange juice 3.7 24.85 (22.63–27.08) 1.90 (1.71–2.09) 0.99 0.24 0.98 60
Watermelon juice 5.3 51.28 (48.52–54.04) 2.92 (2.35–3.50) 0.97 0.16 0.89 89
Watermelon juice 3.6 38.43 (33.78–43.07) 3.22 (2.27–4.17) 0.93 0.44 0.95 63
Listeria Apple juice 3.5 28.88 (27.66–30.09) 1.18 (1.09–1.27) 0.97 0.05 0.89 113
monocytogenes Orange juice 3.7 88.46 (73.05–103.87) 1.51 (1.05–1.97) 0.91 0.07 0.91 257
Watermelon juice 5.3 187.07 (146.81–227.33) 1.61 (1.31–1.92) 0.97 0.01 0.98 508
a
95% CI: confidence interval.
b
R2: determination coefficient.
c
RMSE: root mean square error.
d
Calculation based on estimated α and β parameters and the Weibull model (Eq. (2)).
110 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111

-7 4
Calculated data (Log N/N0)
-6
3.5 A
3

β-parameter
-5
2.5
-4 2
1.5
-3
1
-2
0.5
-1 0
3 3.5 4 4.5 5 5.5
0 pH
0 -1 -2 -3 -4 -5 -6 -7
Observed data (Log N/N0)
200
175 B

α-parameter (kJ/kg)
Fig. 4. Plot of observed values of microbial inactivation of S. Panama (◊), E. coli (▲),
L. monocytogenes (○) and S. cerevisiae (□) in apple, orange, watermelon and adapted wa- 150
termelon juices vs. calculated values using the Weibull equation, with the line y = 1
representing a perfect fit. 125
100
remaining cells become increasingly damaged, whereas upward 75
concavity (β b 1) indicates that remaining cells have the ability to 50
adapt to the applied stress (van Boekel, 2002). Therefore, concave 25
downward (β N 1) survival curves of all micro-organisms tested 0
in this research can be interpreted as evidence that the microbial 3 3.5 4 4.5 5 5.5
cells show the tendency to become weaker when specific energy pH
and temperature increase, indicating that accumulated damage
Fig. 5. Graph of A) shape parameter β and B) scale parameter α, both with 95% confidence
due to synergy occurs. This synergistic effect between temperature
interval as function of pH for inactivation experiments of S. Panama (◊), E. coli (△),
and pulses is in line with the results of the experiments on the in- L. monocytogenes (○) and S. cerevisiae (□) in fruit juices, where filled symbols are
fluence of inlet temperature on PEF inactivation, as described in representing acidified watermelon juice.
Section 3.1. Contradicting, in literature concave upward (β b 1)
survival curves of E. coli (Álvarez et al., 2003a; Rodrigo et al.,
2003; Saldaña et al., 2010), Lactobacillus plantarum (Gómez et al., The dependence of the parameters α and β on the pH of the different
2005b; Rodrigo et al., 2001), Yersinia enterocolitica (Álvarez et al., juices is shown in Fig. 5. Although limited experiments with a variable
2003c), S. enterica serovars (Álvarez et al., 2003d; Saldaña et al., pH were carried out, it can be concluded from Fig. 5A that the β param-
2010) and L. monocytogenes by Álvarez et al. (2003b) and Gómez eter does not depend on pH in a systematic way. Contrary, the α param-
et al. (2005a) were reported. It is difficult to compare these studies eter, depicted in Fig. 5B, seems to be dependent on the pH of the fruit
with our results, since different PEF equipment with varying treatment juice for all micro-organisms tested. In low-acid fruit juices, more spe-
chambers and settings were used. All these studies with an upward con- cific energy is needed to inactivate micro-organisms than in more acid
cavity had in common that they operated at a maximum temperature of fruit juices. The effect of acidification of the low-acid watermelon juice
35 °C and used square-wave pulses. If we consider the data up to 35 °C, of pH 5.3 to pH 3.6 is indicated with the filled symbols in Fig. 5, showing
we observe the same findings: survival of S. Panama up to 35 °C as no effect on the β parameter, but a reduction of the α parameter. Similar
shown in Fig. 1A (open symbols) gave comparable inactivation curves results for the shape independency and scale dependency towards pH
as described by the abovementioned studies with concave upward pa- were found for PEF treatment of L. monocytogenes and L. plantarum in
rameters (α = 124.95, β = 0.34, R2: 0.96, RMSE: 0.05), and when media with varying pH (Gómez et al., 2005a, 2005b).
more pulses were given, PEF showed to be more effective in inactivation,
leading to concave downward parameters when Weibull fit was per- 4. Conclusion
formed over data up to 49 °C (α = 50.43, β = 1.56, R2: 0.92, RMSE:
0.32). This suggests that the synergistic effect between temperature and PEF processing conditions for fruit juices were assessed and inactiva-
pulses is apparent at temperatures above 35 °C, and was therefore not tion of micro-organisms was dependent on pH, the type of microbial
observed in the other studies. species and inlet temperature of the matrix.
Evaluation of the Weibull parameters α and β in the aforementioned A synergistic effect between temperature and PEF treatment was
studies with respect to the variable electric field strengths, showed no demonstrated and suggests that optimization of the PEF conditions to
influence on the β parameters when higher or lower electric field reduce the energy input should aim for processing at higher inlet tem-
strengths were used. Nevertheless, increase of the electric field strength perature to allow more effective inactivation per pulse.
greatly affects the α parameter, showing a more efficient PEF process, The diversity in PEF resistance across the different microbial species
what is expressed by a lower energy use or shorter treatment time. shows the importance to validate industrial processes with relevant
When the parameters α and β are compared for all different micro- micro-organisms (spoilage and pathogens) for the food products. Test-
organisms, it can be seen that the β-value for the S. cerevisiae dataset ing of pathogens has to be done in the food matrix desired to be PEF pas-
was higher compared to that for other micro-organisms, confirming teurized, as intrinsic factors such as pH and conductivity, influence the
that S. cerevisiae is more susceptible to the PEF treatment than other amount of energy required to reach the required reduction of micro-
micro-organisms. To facilitate the comparison among micro-organisms organisms.
and matrices the energy needed to inactivate 5 log cycles is calculated.
Based on the estimated α and β parameters, a calculation is made to es-
References
timate the amount of specific energy necessary to have a 5 log reduction
of a certain micro-organism in a specific fruit juice (Table 3). Most energy
Álvarez, I., Pagán, R., Raso, J., Condón, S., 2002. Environmental factors influencing the in-
is necessary to inactivate the micro-organism L. monocytogenes in water- activation of Listeria monocytogenes by pulsed electric fields. Lett. Appl. Microbiol.
melon juice. 35, 489–493.
 R.A.H. Timmermans et al. / International Journal of Food Microbiology 173 (2014) 105–111 111

Álvarez, I., Virto, R., Raso, J., Condón, S., 2003a. Comparing predicting models for Preservation by Pulsed Electric Fields: From Research to Application. Woodhead Pub-
Escherichia coli inactivation by pulsed electric fields. Innov. Food Sci. Emerg. Technol. lishing, Cambridge UK.
4, 195–202. Mazzotta, A.S., 2001. Thermal inactivation of stationary-phase and acid-adapted
Álvarez, I., Pagán, R., Condón, S., Raso, J., 2003b. The influence of process parameters for Escherichia coli O157:H7, Salmonella, and Listeria monocytogenes in fruit juices.
the inactivation of Listeria monocytogenes by pulsed electric fields. Int. J. Food J. Food Prot. 64, 315–320.
Microbiol. 87, 87–95. McDonald, L.C., Hackney, R.C., Ray, B., 1983. Enhanced recovery of injured Escherichia coli
Álvarez, I., Raso, J., Sala, F.J., Condón, S., 2003c. Inactivation of Yersinia enterocolitica by by compounds that degrade hydrogen peroxide or block its formation. Appl. Environ.
pulsed electric fields. Food Microbiol. 20, 691–700. Microbiol. 45, 360–365.
Álvarez, I., Mañas, P., Condón, S., Raso, J., 2003d. Resistance variation of Salmonella enterica Mosqueda-Melgar, J., Raybaudi-Massilia, R.M., Martín-Belloso, O., 2007. Influence of treat-
serovars to pulsed electric fields treatment. J. Food Sci. 68, 2316–2320. ment time and pulse frequency on Salmonella Enteritidis, Escherichia coli and Listeria
Álvarez, I., Condón, S., Raso, J., 2006. Microbial inactivation by pulsed electric fields. In: monocytogenes populations inoculated in melon and watermelon juices treated by
Raso, J., Heinz, V. (Eds.), Pulsed Electric Field Technology for the Food Industry — Fun- pulsed electric fields. Int. J. Food Microbiol. 117, 192–200.
damentals and Applications. Springer, New York. Noël, H., Hofhuis, A., De Jonge, R., Heuvelink, A.E., De Jong, A., Heck, M.E.O.C., De Jager, C.,
Aronsson, K., Rönner, U., 2001. Influence of pH, water activity and temperature on the in- Van Pelt, W., 2010. Consumption of fresh fruit juice: how a healthy food practice
activation of Escherichia coli and Saccharomyces cerevisiae by pulsed electric fields. caused a national outbreak of Salmonella Panama gastroenteritis. Foodborne Pathog.
Innov. Food Sci. Emerg. Technol. 2, 105–112. Dis. 7, 1–7.
Aronsson, K., Rönner, U., Borch, E., 2005. Inactivation of Escherichia coli, Listeria innocua Qin, B.-L., Chang, F.-J., Barbosa-Cánovas, G.V., Swanson, B.G., 1995. Nonthermal inactiva-
and Saccharomyces cerevisiae in relation to membrane permeabilization and subse- tion of Saccharomyces cerevisiae in apple juice using pulsed electric fields. LWT
quent leakage of intracellular compounds due to pulsed electric field processing. Food Sci. Technol. 28, 564–568.
Int. J. Food Microbiol. 99, 19–32. Raso, J., Barbosa-Cánovas, G.V., 2003. Nonthermal preservation of foods using combined
Bergey, 1986. Manual of Systematic Bacteriology. Williams and Wilkins, Baltimore. processing techniques. Crit. Rev. Food Sci. Nutr. 43, 265–285.
Corlett, D.A., Brown, M.H., 1980. pH and acidity. In: Silliker, J.H., Baird-Parker, A.C., Brylan, Rodrigo, D., Martínez, A., Harte, F., Barbosa-Cánovas, G.V., Rodrigo, M., 2001. Study of in-
F.L., Christian, J.H.B., Clark, D.S., Olson, J.C., Roberts, T.A. (Eds.), Microbial Ecology of activation of Lactobacillus plantarum in orange-carrot juice by means of pulsed elec-
Foods. Academia Press, New York, pp. 92–111. tric fields: comparison of inactivation kinetics models. J. Food Prot. 64, 259–263.
Coster, H.G.L., Zimmermann, U., 1975. The mechanism of electrical breakdown in the Rodrigo, D., Barbosa-Cánovas, G.V., Martínez, A., Rodrigo, M., 2003. Weibull distribution
membranes of Valonia utricularis. J. Membr. Biol. 22, 73–90. function based on an empirical mathematical model for inactivation of Escherichia
Cotter, P.D., Hill, C., 2003. Surviving the acid test: responses of gram-positive bacteria to coli by pulsed electric fields. J. Food Prot. 66, 1007–1012.
low pH. Microbiol. Mol. Biol. Rev. 67, 429–453. Saldaña, G., Puértolas, E., Condón, S., Álvarez, I., Raso, J., 2010. Modeling inactivation kinet-
Elez-Martínez, P., Escolà-Hernández, J., Soliva-Fortuny, R.C., Martín-Belloso, O., 2004a. In- ics and occurrence of sublethal injury of a pulsed electric field-resistant strain of
activation of Saccharomyces cerevisiae suspended in orange juice using high-intensity Escherichia coli and Salmonella Typhimurium in media of different pH. Innov. Food
pulsed electric fields. J. Food Prot. 67, 2596–2602. Sci. Emerg. Technol. 11, 290–298.
Elez-Martínez, P., Escolà-Hernández, J., Soliva-Fortuny, R.C., Martín-Belloso, O., 2004b. In- Saldaña, G., Puértolas, E., Monfort, S., Raso, J., Álvarez, I., 2011. Defining treatment condi-
activation of Lactobacillus brevis in orange juice by high-intensity pulsed electric tions for pulsed electric field pasteurization of apple juice. Int. J. Food Microbiol. 151,
fields. Food Microbiol. 22, 311–319. 29–35.
Fleming, D.W., Cochi, S.L., MacDonald, K.L., Brondum, J., Hayes, P.S., Plikaytis, B.D., Holmes, Sharma, M., Adler, B.B., Harrison, M.D., Beuchat, L.R., 2005. Thermal tolerance of acid-
M.B., Audurier, A., Broome, C.V., Reingold, A.L., 1985. Pasteurized milk as a vehicle of adapted and unadapted Salmonella, Escherichia coli O157:H7, and Listeria
infection in an outbreak of listeriosis. N. Engl. J. Med. 312, 404–407. monocytogenes in cantaloupe juice and watermelon juice. Lett. Appl. Microbiol. 41,
García, D., Gómez, N., Raso, J., Pagán, R., 2005. Bacterial resistance after pulsed electric fields 448–453.
depending on the treatment medium pH. Innov. Food Sci. Emerg. Technol. 6, 388–395. Stanley, D.W., Parkin, K.L., 1991. Biological membrane deterioration and associated qual-
Geveke, D.J., Kozempel, M.F., 2003. Pulsed electric field effects on bacteria and yeast cells. ity losses in food tissues. Crit. Rev. Food Sci. Nutr. 30 (5), 487–553.
J. Food Process. Preserv. 27, 65–72. Teissié, J., Golzio, M., Rols, M.P., 2005. Mechanisms of cell membrane electropermeabilization:
Gómez, N., García, D., Álvarez, I., Condón, S., Raso, J., 2005a. Modelling inactivation of a minireview of our present (lack of ?) knowledge. Biochim. Biophys. Acta Gen. Subj.
Listeria monocytogenes by pulsed electric fields in media of different pH. Int. J. Food 1724, 270–280.
Microbiol. 103, 199–206. Timmermans, R.A.H., Mastwijk, H.C., Knol, J.J., Quataert, M.C.J., Vervoort, L., Van der
Gómez, N., García, D., Álvarez, I., Raso, J., Condón, S., 2005b. A model describing the kinet- Plancken, I., Hendrickx, M.E., Matser, A.M., 2011. Comparing equivalent thermal,
ics of inactivation of Lactobacillus plantarum in a buffer system of different pH and in high pressure and pulsed electric field processing for mild pasteurization of orange
orange an apple juice. J. Food Eng. 70, 7–14. juice. Part I: impact on overall quality attributes. Innov. Food Sci. Emerg. Technol.
Grahl, T., Markl, H., 1996. Killing of microorganisms by pulsed electric fields. Appl. 12, 235–243.
Microbiol. Biotechnol. 45, 148–157. Toepfl, S., Heinz, V., Knorr, D., 2007. High intensity pulsed electric fields applied for food
Gurtler, J.B., Rivera, R.B., Zhang, H.Q., Geveke, D.J., 2010. Selection of surrogate bacteria in preservation. Chem. Eng. Process. 46, 537–546.
place of E. coli O157:H7 and Salmonella Typhimurium for pulsed electric field treat- van Boekel, M.A.J.S., 2002. On the use of the Weibull model to describe thermal inactiva-
ment of orange juice. Int. J. Food Microbiol. 139, 1–8. tion of microbial vegetative cells. Int. J. Food Microbiol. 74, 139–159.
Gurtler, J.B., Bailey, R.B., Geveke, D.J., Zhang, H.Q., 2011. Pulsed electric field inactivation of van Heesch, E.J.M., Pemen, A.J.M., Huijbrechts, P.A.H.J., Van der Laan, P.C.T., Ptasinski, K.J.,
E. coli O157:H7 and non-pathogenic surrogate E. coli in strawberry juice as influenced Zanstra, G.J., De Jong, P., 2000. A fast pulsed power source applied to treatment of
by sodium benzoate, potassium sorbate, and citric acid. Food Control 22, 1689–1694. conducting liquids and air. IEEE Trans. Plasma Sci. 28, 137–143.
Hartyáni, P., Dalmadi, I., Cserhalmi, Z., Kántor, D.B., Tóth-Markus, M., Sass-Kiss, A., 2011. Vervoort, L., Van der Plancken, I., Grauwet, T., Timmermans, R.A.H., Mastwijk, H.C., Matser,
Physical–chemical and sensory properties of pulsed electric field and high hydrostatic A.M., Hendrickx, M.E., Van Loey, A., 2011. Comparing equivalent thermal, high pres-
pressure treated citrus juices. Innov. Food Sci. Emerg. Technol. 3, 255–260. sure and pulsed electric field processes for mild pasteurization of orange juice. Part
Heinz, V., Alvarez, I., Angersbach, A., Knorr, D., 2001. Preservation of liquid foods by high II: impact on specific chemical and biochemical quality parameters. Innov. Food Sci.
intensity pulsed electric fields — basic concepts for process design. Trends Food Sci. Emerg. Technol. 12, 466–477.
Technol. 12, 103–111. Weaver, J.C., Chizmadzhev, Y.A., 1996. Theory of electroporation: a review. Bioelectrochem.
Heinz, V., Toepfl, S., Knorr, D., 2003. Impact of temperature on lethality and energy effi- Bioenerg. 41, 135–160.
ciency of apple juice pasteurization by pulsed electric fields treatment. Innov. Food Wouters, P.C., Dutreux, N., Smelt, J.P.P.M., Lelieveld, H.L.M., 1999. Effects of pulsed electric
Sci. Emerg. Technol. 4, 167–175. fields on inactivation kinetics of Listeria innocua. Appl. Environ. Microbiol. 65,
Hülsheger, H., Potel, J., Niemann, E.G., 1983. Electric field effects on bacteria and yeast 5364–5371.
cells. Radiat. Environ. Biophys. 22, 149–162. Wouters, P.C., Álvarez, I., Raso, J., 2001. Critical factors determining inactivation kinetics
Mastwijk, H.C., 2006. Pulsed Power systems for application of pulsed electric fields in the by pulsed electric field processing. Trends Food Sci. Technol. 12, 112–121.
food industry. In: Raso, J., Heinz, V. (Eds.), Pulsed Electric Field Technology for the Zhang, Q., Monsalve-González, A., Qin, B.-L., Barbosa-Cánovas, G.V., Swanson, B.G., 1994.
Food Industry — Fundamentals and Applications. Springer, New York. Inactivation of Saccharomyces cerevisiae in apple juice by square-wave and exponen-
Mastwijk, H.C., Gulfo-van Beusekom, K., Pol-Hofstad, I.E., Schuten, H., Boonman, M., tial decay pulsed electric fields. J. Food Process Eng. 17, 469–478.
Bartels, P.V., 2007. Definitions and guidelines for reporting on pulsed electric field ex- Zimmermann, U., Pilwat, G., Riemann, F., 1974. Dielectric breakdown of cell membranes.
periments. In: Lelieveld, H.L.M., Notermans, S., de Haan, S.W.H. (Eds.), Food Biophys. J. 14, 881–899.