close

Вход

Забыли?

вход по аккаунту

?

Corrosion mechanism of electrodes in ohmic cooking.

код для вставкиСкачать
ASIA-PACIFIC JOURNAL OF CHEMICAL ENGINEERING
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
Published online 17 August 2007 in Wiley InterScience
(www.interscience.wiley.com) DOI:10.1002/apj.086
Research Article
Corrosion mechanism of electrodes in ohmic cooking
Zhenmin Liu,* Sisira Jayasinghe, Wei Gao and Mohammed M. Farid
The Department of Chemical and Materials Engineering, The University of Auckland, Auckland, New Zealand
Received 5 December 2006; Revised 26 February 2007; Accepted 27 February 2007
ABSTRACT: Different types of bulk and coating materials, including Ag, TiN, ZrN, Ni(P) and Ni(P)-PTFE, have been
used as the electrodes for combined ohmic and plate-cooking of solid food (meat patties). The plates of the grill are
electrically heated, but are also connected to a power supply to act as two electrodes, which pass an electrical current
through the meat patty placed between them. It was found that complex corrosion phenomena of different levels took
place on most of the metallic and coated electrodes because of the strong electrochemical and Faradaic reactions.
Scanning electron microscopy (SEM) and EDX were used to analyze the corrosion mechanism. On the basis of the
experimental results, a physical model is introduced to explain the corrosion mechanism under the ohmic cooking
conditions. Suggestions for electrodes selection and proper operating conditions are also recommended.  2007 Curtin
University of Technology and John Wiley & Sons, Ltd.
KEYWORDS: ohmic cooking; arc; plasma physics; electrochemical reaction; faradaic reaction
INTRODUCTION
In ohmic heating, the electrical energy provided to the
heating cell is ideally used only for heat generation
to be delivered homogeneously and completely within
the materials (Lima et al ., 2003). For liquid heating
this offers a considerable advantage over conventional
methods of heating such as those using plate or tube
heat exchangers. In heat exchangers the surface of
the plate or tube is usually heated to a temperature
higher than that of the bulk of the liquid, which
induces fouling on the surface of the exchanger and also
undesirable temperature distribution within the liquid
(Ozkan et al ., 2004).
Ohmic heating and cooking therefore has attracted
interest for industrial applications and food processing.
However, the energy must be transferred through the
electrode/solution interface with possible Faradaic reactions, which may result in partial electrolysis of the
solution and electrode corrosion via direct metal corrosion or via the electrochemical generation of corroding chemicals (acido-basic gradients, metallic ion ligands decreasing the metal oxidation overpotential, etc.)
(Amatore et al ., 1998). At low-frequency (50–60 Hz)
alternating currents, corrosion of electrodes and apparent electrolysis of the heating medium have been
observed with almost all types of electrodes (Samaranayake and Sastry, 2005). This problem becomes
*Correspondence to: Zhenmin Liu, The Department of Chemical
and Materials Engineering, The University of Auckland, Auckland
1001, New Zealand. E-mail: zliu032@ec.auckland.ac.nz
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
particularly crucial in food processing, where biological molecules naturally present in innocuous forms in
food may result in poisonous or carcinogenic matters,
or bring bad taste or coloration into the processed food
through their electrochemical reduction or oxidation. In
food applications, any electrochemical reaction at electrode/solution interface for food is not desirable even at
low level.
Electrode surfaces in ohmic heating can be regarded
as a ‘junction’ between a solid-state conductor (i.e.
current feeder) and a liquid-state or semi-solid state
conductor (i.e. heating medium cell). They play a
vital role by conveying the current uniformly into the
heating medium. In electrochemistry, it is known that
both the physical and chemical properties of electrodes
(specifically, the electrode surfaces) have an influence
on electrochemical processes at the electrode/solution
interfaces. With some electrodes, a particular electrochemical reaction may occur slowly or not at all; but
with another type of electrode, the same reaction may
be faster under the same conditions. Such information about electrodes under ohmic heating conditions
is, therefore, important to avoid or inhibit the electrochemical reactions by choosing appropriate electrode
materials.
So far, most researchers agree that the electrolysis of
solutions or electrode corrosion can be reduced by using
sufficiently large frequencies (Amatore et al ., 1998;
Jayasinghe, 2004). More specifically, Amatore et al .
(1998) proposed that the high frequency can impede
full charging of the electrode double layer (sheath). In
other words, the potential difference across each double
488
Z. LIU ET AL.
Asia-Pacific Journal of Chemical Engineering
layer (sheath) is always maintained under any threshold
value corresponding to the onset of Faradaic reactions.
However, Samaranayake and Sastry (2005) proposed
that the electrode corrosion mechanism is dominated by
electrochemical reactions, although the electrode double
layer exists. The overall electrolysis reaction and anodic
half-reaction of the metallic electrode (M) corrosion are
as follows (Soffer and Folman, 1972; Yerokhin et al .,
1999; Samaranayake and Sastry, 2005):
2H2 O(liq) ⇔ 2H2(g) + O2(g)
(1)
−
M(s) ⇔ Mn+
(aq) + ne , where n = 1, 2, 3 . . . . . . (2)
However, either of the above theories cannot singly
explain the corrosion nature of ohmic heating, especially with low frequency AC current (50–60 Hz). To
our knowledge, the effect of Faradaic reaction on the
ohmic cooking was not studied and discussed. The
objective of this study was to investigate the corrosion
behaviour of several electrode materials to understand
the corrosion mechanism with the help of arc (plasma)
physics theory, which may lead to the design of efficient
and reliable heating cells.
EXPERIMENTAL
A Breville (Model No. SG 600B) sandwich plate griller
was modified to allow for combined and conventional
ohmic heating (Farid, 2001; Ozkan et al ., 2004). The
upper and lower heating plates were connected to
an AC power supply and modified so that electrode
plates of different materials could be attached. It was
modified to use a voltage (50 V) lower than the main
supply of 230 V. The electric current passed through
the circuit is relatively high (∼15 A), with 50 Hz
sinusoidal alternating mode. Silver, TiN, ZrN, Ni(P) and
Ni(P)-PTFE were used as electrodes and investigated
in this paper. The frozen meat patties were cooked
for about 3 min under the above conditions. Detailed
experimental set-up and materials used can be found
elsewhere (Jayasinghe, 2004).
The radius of the meat patties is about 5.5 cm, and
its moisture content is around 74%. The actual heated
area is about 94 cm2 , so the peak electric current
density is around 1600 A/m2 . Figure 1 shows the center
temperature of the meat patty and the current passing
through it as heating progresses while the voltage
applied was kept constant during most of the heating
time. The observed low rate of change of temperature
with time in Fig. 1 is due to thawing at the early stage
of cooking and evaporation during the latter part of
cooking. In fact, the center temperature of the patties
remains at 373 K (100 ◦ C) owing to the high latent
heat of vaporisation of water Figure 1 also shows a
continuous increase in current up to almost 20 A. The
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
Figure 1. Temperature, voltage and current as functions
of time during the combined ohmic and plate cooking of
7 mm thick meat patty (50 V).
figure also shows how the current increases as the
meat patty thaws, reaching a maximum and then falling
to zero as meat electrical conductivity diminishes due
to moisture loss. All the measured temperature–and
current–time curves have almost same value and trend
with the different electrode materials.
A small piece of the sample from the area where the
corrosion reaction took place after ohmic cooking was
cut and prepared for the scanning electron microscopy
(SEM) and microanalysis (EDX). Samples were cleaned
using a detergent and dried before microscopy observation. The corrosion attacks on all electrodes were
visible, as it had left permanent discolored patches or
pits.
RESULTS AND DISCUSSIONS
The Ag plates used as electrodes for ohmic heating
showed the least corrosion attack when inspected with
the naked eye. However, the SEM photograph of
Ag sample shown in Fig. 2 demonstrates that it has
also been subjected to corrosion. The dark and white
areas in these photographs are carbonized and certain
Ag compound enriched layers, respectively, which
were confirmed by means of compositional analysis
shown in Fig. 4(a). In particular, it can be seen from
Fig. 2(b) that the severe corroded areas are composed
of superfine particles, which look as the results of
rapid melting and/or solidification. Those discrete dark
carbonized areas should come from the heated food,
but look abnormal as compared to the appearance
of conventional thermal conduction cooking since the
heating time is no more than 3 min, and it was very
hard to be washed away after cooking. The possible
mechanism will be discussed in detail later.
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
DOI: 10.1002/apj
Asia-Pacific Journal of Chemical Engineering
CORROSION OF ELECTRODES IN OHMIC COOKING
Figure 2. SEM image of corroded Ag electrode: (a) low magnification; (b) high magnification.
Figure 3. SEM image of corroded area: (a) Ni(P); (b) Ni(P)-PTFE.
Figure 3 shows that the cracks have developed around
the contact surface areas of the Ni(P) and Ni(P)-PTFE
coatings with the meat patties. The cracks on the Ni(P)PTFE coating surface looks even worse, which is probably due to the large difference in thermal expansion
coefficients of the coating, PTFE particles and substrate. This appearance strongly suggests that thermal
shocks happened as a result of Faradaic reaction during the ohmic cooking. Furthermore, the abnormal dark
carbonized spots and severe discolored patches on the
Ni(P) surface were confirmed by the EDX results in
Fig. 4(b), implying that the corrosion mechanism is
a combination of electrochemical and Faradaic reactions. In contrast with Fig. 4(a), there is much less
oxygen element detected, which indicates that Ag is
more oxidation resistant or electrochemically stable.
This implies that the role of electrochemical reaction is
different on different electrodes because of their different surface nature. However, it was reported that even
Pt, the most electrochemical stable metal, also could
be corroded (Samaranayake and Sastry, 2005), which
further confirms that the Faradaic process is independent of the chemical nature of electrode surface (Soffer
and Folman, 1972). The above discussion indicates that
the effect of the Faradaic reaction during the ohmic
cooking on the electrode corrosion is significant.
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
Similarly, carbonization and oxidation reactions were
confirmed by means of EDX, but slightly different corrosion appearance on TiN and ZrN coated electrodes
are shown in Fig. 5. Being ceramic coatings, TiN and
ZrN have relative low conductivity among the electrodes used but high electrochemical stability. Therefore, the corrosion attack in Fig. 5 should be caused
mainly by Faradaic reaction. Interestingly, those dispersed craters or pits look almost exactly same as the
results of spark erosion (Prashad, 2001). Furthermore,
on the basis of the mechanisms of plasma electrolysis,
spark erosion and arc (plasma) ignition (field emission)
(Hoyaux, 1968; Yerokhin et al ., 1999; Prashad, 2001),
the effect of Faradaic reaction during ohmic heating are
technically different, but, in fact, physically identical.
The typical current–voltage plots in Fig. 6 represent
the metal–electrolytic (liquid solution and semi-solid
meat) systems with underlying gas liberation on either
the anode or cathode surface as discussed above.
Though there are some differences between the results
in the literature (Yerokhin et al ., 1999) and the curve
we measured, because of the different states of the
electrolyte involved and the voltage set at no more than
50 V during the ohmic heating, essentially the principle
is same under 50 V. At a relatively low voltage the
kinetics of the electrode processes for both systems
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
DOI: 10.1002/apj
489
490
Z. LIU ET AL.
Asia-Pacific Journal of Chemical Engineering
Figure 4. EDX results of corroded areas: (a) Silver electrode; (b) Ni(P).
Figure 5. SEM image of corroded area on: (a) TiN; (b) ZrN.
Figure 6. Currrent–voltage sketch for the processes of
discharge phenomena developed in the near electrode area
from liquid (Yerokhin et al., 1999) and semi-solid (meat)
electrolyte.
conform to Faraday’s laws and the current–voltage
characteristics of the cell vary according to Ohm’s law.
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
Thus, an increase in voltage leads to a proportional rise
in the current (region 0–U1 ). In the region U1 –U2 , a
potential rise leads to a current oscillation accompanied
by luminescence. The current rise is limited by the
gaseous reaction products (O2 or H2 ) over the electrode
surface from a partial action of electrolysis. In areas
where the electrode remains in contact with the liquid
and conductive solid food, however, the current density
continues to rise, causing local boiling (ebullition)
of the electrolyte adjacent to the electrode. Upon
progression to point U2 , the electrode is enshrouded
by the continuous gaseous vapor plasma envelop of
low electrical conductivity. Almost all the voltage drop
across the cell is now in this thin and near-electrode
region. The electric field strength E within this region
therefore reaches a value of 106 –108 V/m, which is
sufficient for initiation of the ionization processes (also
called field emission) in the vapor envelope (Hoyaux,
1968).
In our experiment, field emission is also reached at
some limited spots in the whole surface of a gap, causing a spark. In other words, some ‘weak’ points represented by surface irregularities or irregular contact with
the meat patties are dominant in appraising the efficiency of this mechanism. An increase of the local field
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
DOI: 10.1002/apj
Asia-Pacific Journal of Chemical Engineering
CORROSION OF ELECTRODES IN OHMIC COOKING
by one order of magnitude seems quite normal. Furthermore, it was reported that the voltage required for such
an ionization can be as low as 15 V (Hoyaux, 1968).
Consequently, the ionization phenomenon appears initially as a rapid sparking in scattered gaseous bubbles
and then may generate local ultra-hot points. Owing to
the hydrodynamic stabilization of the vapor envelop in
the region U2 –U3 , the current drops, and beyond the
point U3 , the glow discharge transforms into intensive
arcing accompanied by a characteristic low-frequency
acoustic emission (Yerokhin et al ., 1999). This is basically consistent with our experimental observation as
shown in Figs 1 and 6, in which the current drops from
about 15 A to near zero uncontrollably owing to moisture loss after the surface crust of the meat patties was
formed in a few minutes, whereas some arc discharge
noise could still be heard. In practice, a number of the
above electrode electrolysis processes may occur concurrently over adjacent areas of the electrode surface.
This was confirmed by the consistent noise discharge
during the ohmic cooking of meat patties.
Estimations of the discharge temperature vary widely
in theory and practical environments, from 800–3000 K
to 3000–6000 K or even 10 000–20 000 K (Yerokhin
et al ., 1999). These differences may be attributed to the
complex structure of the discharge channel, in which
a hot core (of 6800–9500 K) and cold circumferential
area (of 1600–2000 K) have been identified by spectral
studies. The diameter of the discharge channel and
the thickness of the heat-affected zone are estimated
to be 1–10 µm and 5–50 µm, respectively (Yerokhin
et al ., 1999), which also has a good agreement with the
corrosion attack appearance from Figs 2–5.
For practical systems the current–voltage regions
corresponding to the conditions of plasma electrolysis
can be determined by the critical voltage point Ui
indicated in Fig. 6. The critical values of Ui can often
be quite accurately estimated on the basis of theoretical
or empirical expressions. Thus, U1 can be assessed
against the critical field strength for breakdown across a
gaseous envelope, as follows from the theory of impact
ionization (Yerokhin et al ., 1999):
U1 = bp ln
α
ap
(3)
Here a and b are constants, p (Pa) is the vapor
pressure and α is the impact ionization coefficient of
the vapor species. From Eqn (3) the breakdown voltage
U1 of a vapor bubble in water can be estimated to be
∼40 V (Yerokhin et al ., 1999), which lies close to our
experimental voltage (50 V).
Region U2 –U3 can be estimated from the critical
value of Joule heat density Wc for conversion from
vapor bubbles to continuous film boiling (Yerokhin
et al ., 1999):
(4)
jU2−3 = Wc
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
Here j is the current density. Assume that Wc
for aqueous electrolytes is close to the critical density of heat flow for a water boiling process (i.e.
∼8 × 105 W/m2 ) and that j can vary in the range
4–10 kA/m2 . As compared to our experimental electric current density of 1.6 kA/m2 , it is reasonable to
expect that an increase of the local field by one order
of magnitude to the average field should be quite normal for heating semi-solid meat patties. Therefore, such
current density in our experimental conditions is sufficient for initiation and maintenance of local arc or spark.
This theory can explain the above corrosion phenomena, which particularly happened on the Ni(P), TiN and
ZrN coated electrodes.
As regards to the effect of the frequency on the
Faradaic reaction, it may be concluded that a low frequency such as 50–60 Hz provides enough time for
full charging of the electrode double layer (sheath)
in this particular electrode/electrolyte system. Consequently, the potential difference across each double
layer is also easy to overcome any threshold value
corresponding to the onset of the Faradaic reactions:
whereas certain high-frequency current does not provide enough time for full charging of the sheath on the
anodic electrode surface during the half period of alternating time (Amatore et al ., 1998) and then the anodic
electrode soon changes into cathodic protective one.
Research is in progress to gain better understanding of
the effect of the frequency on Faradaic reaction or spark
erosion in ohmic heating.
CONCLUSIONS AND RECOMMENDATIONS
On the basis of the experimental work, it can be concluded that corrosion of the electrodes during the combined ohmic and plate cooking is a complex result of
electrochemical and Faradaic reactions. More specifically, the arc/spark erosion as a result of Faradaic
process plays a dominant role on TiN and ZrN coated
electrodes, and is less affected by the nature of electrode
surface.
There are three factors that are directly related to
the ohmic cooking electrode corrosion: the conductivity
of the electrode, the electrochemical stability and the
spark or plasma generation. Highly conductive materials
or coatings with high electrochemical stability such as
graphite (C), Pt, Au and Ag are good candidates. On
the basis of the discussion above, there are two ways
that may eliminate electrode corrosion: one is using a
sufficiently high frequency alternating current to avoid
the spark or arc ignition, and the other is to eliminate
the gaseous liberation on the surface of anodic electrode
if the frequency is not high enough. This will suppress
the ionization burst or sparks.
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
DOI: 10.1002/apj
491
492
Z. LIU ET AL.
Asia-Pacific Journal of Chemical Engineering
REFERENCES
Amatore C, Berthou M, Hebert S. Fundamental principles of
electrochemical ohmic heating of solutions. J. Electroanal. Chem.
1998; 457(1–2): 191–203.
Farid MM. New methods and apparatus for cooking. #PCT/NZ02/
00108. The University of Auckland: New Zealand, 2001.
Hoyaux MF. Arc Physics, Applied Physics and Engineering Series.
Vol. 8 . Springer- Verlag New York Inc: New York, 1968; 305.
Jayasinghe S. Ohmic cooking of hamburger patties. Chemical and
Materials Engineering. The University of Auckland: Auckland,
2004; 115.
Lima M, Zhong T, Lakkakula NR. Ohmic Heating: A Value-added
Food Processing Tool. Department of Food Science, LSU
AgCenter: Baton Rouge, LA, 2003.
 2007 Curtin University of Technology and John Wiley & Sons, Ltd.
Ozkan N, Ho I, Farid M. Combined ohmic and plate heating of
hamburger patties: quality of cooked patties. J. Food Eng. 2004;
63(2): 141–145.
Prashad H. Appearance of craters on track surface of rolling element
bearings by spark erosion. Tribol. Int. 2001; 34(1): 39–47.
Samaranayake CP, Sastry SK. Electrode and pH effects on
electrochemical reactions during ohmic heating. J. Electroanal.
Chem. 2005; 577(1): 125–135.
Soffer A, Folman M. The electrical double layer of high surface
porous on carbon electrode. J. Electroanal. Chem. 1972; 38(1):
25–43.
Yerokhin AL, Nie X, Leyland A, Matthews A, Dowey SJ. Plasma
electrolysis for surface engineering. Surf. Coat. Technol. 1999;
122(2–3): 73–93.
Asia-Pac. J. Chem. Eng. 2007; 2: 487–492
DOI: 10.1002/apj
Документ
Категория
Без категории
Просмотров
1
Размер файла
583 Кб
Теги
electrode, cooking, mechanism, corrosion, ohmic
1/--страниц
Пожаловаться на содержимое документа