J O U R N A L O F M A T E R I A L S S C I E N C E 3 6 (2 0 0 1 ) 2547 – 2555

Corrosion behaviour of some stainless steel

alloys in molten alkali carbonates (I)
S. A. SALIH, A. N. EL-MASRI, A. M. BARAKA
Chemistry Department, Faculty of Science, Cairo University, Giza, Egypt
E-mail: [email protected]

In the present work it is aimed to study the corrosion behaviour of two types of stainless
steel alloys (one ferritic and two austenitic) in molten Li2 CO3 - Na2 CO3 - K2 CO3 mixture. This
mixture is of interest in corrosion studies because of its low melting point (397◦ C) and good
electrical properties. In this investigation the following techniques of measurements are
used: (i) open circuit-potential, (ii) galvanic current, (iii) impedance, (iv) atomic absorption
spectroscopy for the determination of the amount of metals dissolved in the melt
(v) corrosion tests, carried out on the oxide scales formed during the oxidation of stainless
steel alloys in carbonate melt. In this melt the electrode Ag/AgCl was used as a reference
electrode. In molten carbonates, the oxide ions originate by self-dissociation according to
the equilibrium CO2− 3 ↔ CO2 + O . The oxide ions, O , and carbonate ions, CO3 , play an
2− 2− 2−

important role in the oxidation process of these alloys and their passivation in the
carbonate melt. As previously mentioned in references it can be assumed that the oxide
scales formed on the alloy surface consist mainly of LiCrO2 and LiFeO2 . The cathodic path
of the corrosion process may be the reduction of CO2 and/or CO2− 3 . The resistance of alloys
against corrosion in melt increases with the increase of temperature. This may be due to
the increase of concentration of O2− and CO2 , enhancing both the anodic and cathodic
reactions. The activation energy was calculated and found to be 91.496, 23.412 and
37.956 kJ/mol for the alloys 1, 2 and 3 respectively. The above mentioned techniques of
measurements showed that the oxide scales of the austenitic stainless steel alloys (2, 3) are
more passive and protective than of ferritic stainless steel alloy (1). This means that the
resistance against corrosion, in the carbonate melts, of austenitic stainless steel alloys is
higher than that of ferritic one. ° C 2001 Kluwer Academic Publishers

1. Introduction This form of attack is thought to be due to “fluxing”

Interest in use of fused salts in industrial processes is wherein the normally protective oxide scale on an al-
continually increasing and these media are gradually loy destroyed by dissolution-precipitation mechanism.
becoming accepted as a normal field of chemical engi- Reaction between alloys and molten carbonates are
neering. technologically important in view of the long term sta-
Molten salts in particular alkali metal halides, ni- bility and compatibility of materials associated with
trates, sulphates and carbonates possess properties molten carbonate fuel cells.
which make them desirable in the development of effi- The molten carbonate fuel cells has been continues to
cient methods of materials processing and energy pro- be under intensive investigation and development. The
duction. corrosion behaviour of stainless steel and other alloys
Molten alkali carbonates, in view of fuel cell appli- were the subjected to study in molten alkali carbonate
cation, have drawn the interest of scientists for thirty by many investigators [1–13]. Most of these studies
years. More recently, these media appeared to have pos- were carried out under the conditions of operation of
sible developments for recovery of hydrogen, methane molten carbonate fuel cells.
or other hydrocarbons from natural biocompounds. In this article it is aimed to study the corrosion be-
In these organic processes, some potentially corrosive haviour of two types of stainless steel alloys in ternary
compounds such as oxygen or water are formed. Li2 CO3 - Na2 CO3 - K2 CO3 molten mixture. This mix-
Material degradation owing to corrosion is an im- ture is of interest in corrosion studies because of its low
portant issue virtually all energy producing systems, melting point (397◦ C) and good electrical properties. In
including the advanced molten carbonate fuel cells now this investigation the following techniques of measure-
under development. Current collectors and other metal- ments are used: (i) open circuit-potential, (ii) galvanic
lic components of the fuel cell may be rapidly degraded current, (iii) impedance, (iv) atomic absorption spec-
at high temperatures if molten salt contacts the surface. troscopy for the determination of the amount of metals

0022–2461 °
C 2001 Kluwer Academic Publishers 2547
dissolved in the melt (v) corrosion tests, carried out on spectrophotometer (Perkin-Elmer Apparatus Model 23,
the oxide scales formed during the oxidation of stain- Germany).
less steel alloys in carbonate melt.

3. Results and discussion

2. Experimental 3.1. Open circuit potential
In the present investigation of the corrosion behaviour measurements (OCP)
of one ferritic (15.05% Cr) and two austenitic stain- The variation of the corrosion potential, E corr , of the
less steel alloys (17.9% Cr, 7.08% Ni and 20.45% Cr, stainless steel alloys (no. 1, 2, 3), under open circuit
8.3% Ni) was studied in molten Li2 CO3 - Na2 CO3 - conditions in molten Li2 CO3 - Na2 CO3 - K2 CO3 mix-
K2 CO3 mixture (43.5 : 31.5 : 25 mole % respectively). ture, was followed as a function of time till steady state
The calculated quantities of carbonates were mixed to- values were established. The curves of Fig. 1a–c show,
gether in a special container and placed in an electri- respectively, the behaviour of the stainless steel alloys
cally heated furnace and melted at 550◦ C. The last trace at temperatures 475, 500, 525, and 550◦ C. It could be
of water removed by bubbling through the melt, pure seen that the corrosion potential of the alloys firstly
dry CO2 gas for a period of 2 hours. The mixture thus shifts in the positive direction reaching maximum. Af-
prepared was left to cool in a dry atmosphere and so- ter this the potential starts to move in the active direction
lidified mass was quickly crushed and kept in a closed till attaining its steady state.
desiccator till required. In each experiment 50.0 g of The results of open circuit potential measurements
the ternary mixture was used. for the three stainless steel alloys are depicted in
The electrochemical cell used in this study consisted Fig. 2a–c as E corr versus logarithm of time at different
of an outer cylindrical closed end alumina crucible temperatures. It is clear from these plots that, before
(500 ml capacity) in which a smaller Pyrex glass tube attaining the steady state potential, the obtained lines
(100 ml capacity) was placed. The latter contained the contain two segments. The slopes of these segments
electrolyte and electrodes. The cell was rendered gas are determined, kf , k−f which are taken as a measure of
tight by means of O-ring seal between the outer alu- the rate of the formation and dissolution processes of
mina crucible and a stainless steel header from which
the electrodes tubes were fixed.
The cell and its contents was heated to the required
temperature throughout the period of the experimental
in an electric vertical tube furnace. The temperature of
the furnace was regulated by means of a veriable trans-
former and followed by means of a Ni/Ni-Cr thermo-
couple and a temperature indicator. In all experiments,
the desired temperature was precisely adjusted in a limit
of ±2◦ C. The thermocouple was contained in a special
tube to separate it from the melt.
Stainless steel electrodes were cut from ferritic (alloy
no.1) and austenitic (alloy no. 2, 3) stainless steel sheets
of thickness 1 mm. These electrodes were of dimen-
sions 1 × 1 cm2 with a side arm of about 20 cm long
and 0.2 cm width, used for electrical connection. This
arm is coated by a pyrex glass tube of suitable diameter
which fitted to the arm at the upper end by an adhe-
sive. In each experiment a new electrode was used. Just
before of its immersion in the melt, the electrode was
abraded with emery papers of different grades and de-
greased with ether.
In this melt the electrode Ag/AgCl was used as a ref-
erence electrode. The reversibility of this electrode and
its stability to function as a reference one was estab-
lished by many authors [14–17].
The potential of working elctrode, relative to the
reference electrode, was measured on a D.C. Micro-
voltmeter (type TM8 Level Electronic LTD, England).
The galvanic current in microamperes was measured
on digitalmultimeter (Model 1008, Kyoritsu Electri-
cal Instruments work, LTD Japan). The impedance
measurements are carried out using Impedance Mea-
Figure 1 Variation of corrosion potential, E corr of the stainless steel
surements System (1M6 Zahner Electrik, Meßtechnik, alloys with time in pure carbonate melt at different temperatures.
Germany). The atomic adsorption measurements E1, 475◦ C.E2, 500◦ C.E3, 525◦ C.E4, 550◦ C. (a) alloy 1. (b) alloy 2.
are carried out using flame-atomic absorption (c) alloy 3.

2548
T A B L E I Data of the open circuit potential measurements for the
stainless steel alloys at different temperatures

kf k−f
Temp. E imm Es E s − E imm ts mV mV
Alloy ◦ C (mV) (mV) (mV) min decade−1 decade−1

st.st-1 475 −1210 −1200 10 80 33.892 107.92

500 −1250 −1230 20 90 86.416 87.782
525 −1275 −1249 26 40 124.83 108.49
550 −1310 −1277 33 15 129.75 74.867
st.st-2 475 −1490 −1419 71 40 151.65 43.356
500 −1520 −1438 82 25 202.59 42.164
525 −1580 −1462 118 30 195.13 55.974
550 −1595 −1467 128 50 223.74 45.04
st.st-3 475 −1466 −1406 60 35 152.5 64.234
500 −1489 −1408 81 40 167.92 67.788
525 −1540 −1412 128 45 219.36 39.131
550 −1578 −1432 146 45 256.55 36.286

Figure 3 (a): kf , k−f vs. Cr (wt%) in the alloy at 550◦ C. (b): (E s − E imm )
vs. Cr (wt%) in the alloy 550◦ C.

content of the stainless steel alloys at a temperature of

550◦ C.
Fig. 4a and b represent, respectively, the variation of
kf , k−f and (E s - E imm ), as a function of iron content of
the stainless steel alloys at a temperature of 550◦ C.
Fig. 5a and b represent the variation of the amount
of Cr and Fe (wt %) dissolved in melt as a function of
Cr and Fe contents of the stainless steel alloys, respec-
tively, at a temperature of 550◦ C.
Figure 2 Variation of corrosion potential, E corr , of the stainless steel In molten carbonate the oxide ions originate by self-
alloys with log (time) in pure carbonate melt at different temperatures.
dissociation according to the equilibrium:
E1, 475◦ C. E2, 500◦ C. E3, 525◦ C. E4, 550◦ C, (a), alloy 1, (b) alloy 2,
(c) alloy 3.
3 ↔ CO2 + O
CO2− 2−
(1)
the oxide scale, respectively. The value of these rates are
different and depend on the type of stainless steel al- this reaction is responsible for the presence of oxide
loy and temperature of the melt. Table I includes the ions in the carbonate melt. In Lux–Flood acid–base
values of E imm (potential at the moment of immer- properties, CO2 is the acid and O2− the base. It can
sion in melt), E s (steady state potential), (E s - E imm ), be assumed that the oxide ions, O2− , and carbonate
kf and k−f for the three stainless steel alloys at different ions, CO2−3 , play the important role in the oxidation
temperatures. process of the stainless steel alloys in the carbonate
Fig. 3a and b represent, respectively, the variation melt. Thus the reactions leading to the oxidation of
of (E s - E imm ), kf and k−f as a function of chromium different metals in the alloys may be represented by the
2549
 Figure 5 (a): Fe, Cr (wt %) in the melt vs. Cr (wt %) in the alloy at
Figure 4 (a): kf , k−f vs. Fe % in the alloy at 550◦ C. (b): (E s − E imm ) vs.
550◦ C. (b): Fe, Cr (wt %) in the melt vs. Fe (wt %) in the alloy at 550◦ C.
Fe % in the alloy at 550◦ C.

The possible cathodic reactions in the molten carbonate

following equations: may be [18–20].

M + n/2 O2− ↔ MOn/2 − ne (2) 3CO2 + 4e → C + 2CO2− (9)

3

3 + 4e → C + 3O
CO2− 2−
and/or (10)

M + n/2 CO2− and/or

3 ↔ MOn/2 + n/2 CO2 + ne (3)
2CO2 + 2e → CO + CO2−
3 (11)
in the case of carbonate melts and the undertest stain-
less steel alloys, it can be proposed, as in the previous CO2−
3 + 2e → CO + 2O 2−
(12)
works of Vossen et al. [18, 19], that the oxidation and
passivation of the alloys may proceed according to the According to the postulate of the electrochemical the-
following paths: ory, the anodic process may be metal dissolution and/or
anodic barrier layer formation. Barrier layer may be
Cr + Li+ + 2CO2−
3 → LiCrO2 + 2CO2 + 3e (4) formed by dissolution-precipitation mechanism and/or
solid state mechanism. For many passive metals, it was
Fe + CO2−
3 → FeOss + CO2 + 2e (5) found that the thickness of oxide scale, x, varies lin-
Fe + Li+ + 2CO2−
3 → LiFeO2ss + 2CO2 + 3e (6)
early with the electrode potential either under polariza-
tion [21] or under open circuit conditions [22–24], i.e
dEcorr/dx = const. The constant is actually the electric
A cubic solid solution (ss) FeO and LiFeO2 is formed.
field strength, H, i.e. E corr αx and the equation relating
However, FeO is stable only above 570◦ C, therefore
E corr and time, t, can be written as:
FeO may decompose to Fe3 O4 and, apparently Fe2 O3
at temperatures lower than 570◦ C as:
E corr = k1 + k2 log(t + t ◦ )− (13)
3FeO + O 2−
= Fe3 O4 + 2e (7)
where k1 , k2 and t ◦ are constants. The constant k2 is
2Fe3 O4 + O 2−
= 3Fe2 O3 + 2e (8) taken as a measure for the rate of oxide scale thickening
2550
 ferritic stainless steel alloy (1). For any alloy, the values
of k−f decrease with the increase of temperature. Also, it
can be concluded that the resistance against corrosion
of the undertest stainless steel alloys increases in the
order: alloy 1 < alloy2 ≈ alloy3.
Vossen et al. [12] studied the corrosion behaviour of
five commercially available alloys including stainless
steel alloys in molten alkali carbonates and reported
the following results. The oxide scale formed on stain-
less steel alloys is probably a mixture of porous Li-Fe-
oxide, Li-Cr-oxide and nickel islands. The corrosion
resistance of the alloys increases with the increase of
Cr content of the alloy as a result of formation of less
porous oxide scale on the metallic surface or forma-
tion of (lithium) chromium oxide scale underneath the
porous oxide scales. These assumptions are in good
agreement with obtained results in the present inves-
tigation. The increase of resistance against corrosion
of the undertest stainless steel alloys with increase of
temperature may be due to increase of the degree of
self dissociation of carbonate with the increase of tem-
perature. With increasing temperature the equilibrium
CO2−3 ↔ CO2 + O
2−
is shifted to right with the forma-
tion of more CO2 and O2− . The increase of O2− and
CO2 concentration in the melt enhances both, the pre-
viously mentioned, anodic and cathodic reactions. This
state gives rise to the formation of more thick and pro-
tective scales on the electrode surface.

3.2. Galvanic current measurements

Figure 6 Variation of log k2 with the reciprocal of the absolute temper- This type of measurements was previously used by
ature for stainless steel alloys in the pure carbonate melt. (a) alloy 1.
(b) alloy 2. (c) alloy 3.
Benzo and Takeo [26] for studying the oxidation be-
haviour of stainless steel alloys in alkali nitrate melt.
They measured the galvanic current between the stain-
per decade of time and has the meaning of kf (previously less steel and Pt electrode, till attaining the steady state.
mentioned). Similarly in this study trials were made to use the gal-
In the early stages of oxidation the increase of cor- vanic current measurements to study the corrosion be-
rosion potential in the positive direction is attributed to haviour of the undertest alloys in molten carbonate
covering the metal surface with a barrier film [25]. The baths. In these measurements the galvanic current be-
experimental results show that this film grows at a rate, tween the working electrode and Pt electrode was fol-
magnitude of which depends on the temperature of the lowed as a function of time, till attaining the steady-
metal and type of alloy. The value of k2 increases with state.
increase of temperature from 475◦ C to 550◦ C. Also, the Fig. 7 represents the galvanic current-time plots of
value of k2 for stainless steel alloys 2 and 3 are higher the three stainless steel alloys (no. 1, 2, 3) immersed
than those for stainless steel alloys 1. in carbonate melt at temperature of 550◦ C. It is clear
Fig. 6a–c represent the Arrhenius plots of log
k2 versus 1/T for the stainless steel alloys, respec-
tively. The calculated values of activation energy are:
91.496 kJ/mol for alloy 1, 23.412 kJ/mol for alloy 2
and 37.956 kJ/mol for alloy 3. The low values of acti-
vation energy for the alloys 2, 3 may signify diffusion-
controlled processes occurring in the oxide matrix.
On the other hand the higher values of activation en-
ergy for alloy (1) may signify activation-controlled
processes.
From the results listed in Table I and depicted in
Figs 3–5, it is clear that the values of kf , E s − E imm
are higher for the austenitic stainless steel alloys (2, 3)
than those for the ferritic stainless steel alloy (1). For
any alloy, these values increases with the increase of
temperature. On the other hand the values of k−f and
Fe% and Cr% dissolved in the melt are lower for the Figure 7 Variation of galvanic current with the time for stainless steel
austenitic stainless steel alloys (2, 3) than those for the alloys (1, 2, 3) in pure carbonate melt at 550◦ C.

2551
T A B L E I I Data of galvanic current measurements for the stainless
steel alloys at 550◦ C

Temp. i imm is i imm − i s ts

Alloys ◦C µA µA µA min

st.st-1 550 1074 886 188 15

st.st-2 550 1140 564 567 50
st.st-3 550 1145 590 555 40

from these plots that, on immersion of the electrode

in the carbonate melt, the galvanic current greatly de-
creases with time reaching minimum values. Afterthis
the galvanic current starts to increase reaching max-
imum and then decrease till steady state values. The
values of the galvanic current at the moment of elec-
trode immersion in the melt, i imm , and at the steady (a)
state, i s and the difference (i imm − i s ) are deduced from
the plots of Fig. 7 and listed in Table II. If we roughly
consider the value of (i imm − i s ) as a measure of thick-
ness and protective properties of oxide scales, the re-
sistance against corrosion for the undertest stainless
steel alloys is alloy 1 < alloy2 ≈ alloy3. These results
are in a good agreement with open circuit potential
measurements.

3.3. Impedance measurements

In this part the impedance measurements, were used to
study the corrosion and oxidation behaviour of the un-
dertest stainleess steel alloys (1, 2, 3) in molten carbon-
ate mixture, at temperature of 550◦ C. In principle, the
method, involves direct measurements of impedance of
the electrochemical system in frequency domain from
0.1 to 105 Hz. The electrode impedance, Z , and the
(b)
phase shift (θ ) are presented in Bode and Nyquist for-
mats. These experiments are carried out at steady state
under open circuit conditions.
Fig. 8a and b represent Bode plots and Fig. 8c rep-
resent the Nyquist plot for stainless steel alloys (1,
2, 3). Inspection of the Bode plots indicates that, in
most cases, the impedance, Z , remains unchanged in
the high frequency region. In the lower frequency re-
gion the impedance starts to increase with the decrease
of frequency till reaching the end of frequency limit
(10−1 Hz). The corresponding phase plots have the
same trend, except that (θ) has maximum value in
the frequency region from 0.5 to 0.2 Hz depending
on the type of the stainless steel alloy. The Nyquist
plots may be regarded as a plot of a semicircle and not
complete semicircle the length of this part increases
according to the order: alloy 1 < alloy 2 < alloy3.
These results and shapes of the plots of Fig. 8a–c are
similar to those reported by Walter [27] in his review
of impedance plot methods used for corrosion perfor-
mance analysis of painted metals and also those ob-
tained on iron [28] and nickel [29]. Walter [27] reported
(c)
that in many cases the equivalent circuit (Randell cell)
Figure 8 (a): Bode plot (impedance - frequency) in pure carbonate melt
will not be adequate as a model for the painted metal
at 500◦ C. ——–alloy 1, ..........alloy 2, - - - - - - - - - alloy 3. (b): Bode plot
solution interface. One modification takes into account, (phase angle - frequency) in pure carbonate melt at 500◦ C. ——–alloy 1,
diffusion processes within pores in the paint film, which ..........alloy 2, - - - - - - - - - alloy 3. (c): Nyquist impedance plot in pure
are modelled by the inclusion of a Warburg or pseudo carbonate melt at 500◦ C. ——alloy 1, ........alloy 2, - - - - - - - alloy 3.

2552
 which leads to Equation 15.
µ ¶
σ −1/2
log Z = log √ log f (15)
π
this predicts that a Bode plot of log Z versus solely
of Warburg impedance, Z , will be characterized by a
lower frequency slope of − 12 and an intercept on the
√
Z w axis at F = 1 Hz of σ/ π .
Estimates of σ can be obtained by finding a region
of the Nyquist complex plane plot at low frequencies,
where the diffusion tail is inclined at angle of 45◦ to the
axis Z 0 . Within this region σ can be caluclated from
Equation 16 which is obtained from Equation 14.

Z 00 = σ ω−1/2
Z 00
σ = −1/2 = Z 00 ω1/2 (16)
ω
Figure 9 Equivalent electrical circuit, for stainless steel alloys.
calculation of σ values can be made from log Z Bode
plots by finding a region at low frequencies where the
slope is − 12 . By back extrapolation, the value of the Z
√
impedance, Z w , placed in series with R3 , where Z w is can be read at frequency of 1 Hz and equated to σ/ π
defined according to Equation 14. in order to obtain σ .
The values of: 1) - Rcorr are calculated from Bode
plots at frequency 0.1 Hz by subtracting the value
Z w = σ ω−1/2 (1 − j). (14) of Z (R0 ) at high frequency from the value of Z
(R0 +√Rcorr ) and from00Nyquist plot using the modulus:
Z = Z 02 + Z 002 = Rcorr .
where σ : Warburg impedance coefficient (ohm1/2 ), 2)- σ (Warburg impedance coefficient) are calculated
−1
ω: angular
√ frequency = 2π f (rad s ), j: imaginary from Bode plots by calculation
unit = −1. √ of Z w at 1 Hz and by
using Equation 15 (σ1 = Z w π ), and from Nyquist
In our study the equivalent electrical circuit can be plots at f = 0.1 Hz using equation, σ2 = Z 00 ω1/2 , where
shown in Fig. 9. It must be mentioned that in Nyquist ω = 2π f (rad s−1 ).
plot at σ = 0 a semicircle can be obtained, but at higher 3)- the phase angle (θ ) at f = 0.1 Hz, are deduced from
σ values a diffusion tail begins to appear at low fre- Bode plots. These values are calculated for the stainless
quencies. When values of σ are about equal to R2 , this steel alloys (1, 2, 3) and listed in Table III. Inspection of
diffusion tail begins to overlap the semicircle and sub- these results reveals that: 1)- the values of Rcorr calcu-
tends an angle of 45◦ to the Z 0 axis. As σ increases, still lated either from Bode plots or Nyquist plots are near to
further, this overlap becomes increasingly more sever, each other. 2)- the values of Rcorr increases according
but the diffusion tail eventually becomes inclined at an the following order st. st alloy 1 < st. st alloy 2 < st. st
angle of 45◦ to Z 0 at low frequencies. Finally when σ alloy 3. This may be attributed to the increase of Cr %
is significantly greater than R2 (indicating that the dif- in the alloys and also to the presence of nickel in the
fusion processes are occurring slower than the metal austenitic stainless steel alloys (2, 3). 3)- the values of σ
charge-transfer reaction), the diffusion tail completely calculated either from Bode plots (at f = 1 Hz) or from
distorts the semicircle shape. In Bode plots for the cir- Nyquist plots ( f = 0.1 Hz) are near to each other. These
cuit, when there is no diffusion impedance, σ = 0 and values increase according to the above mentioned or-
the lower frequency portion of the log Z is a line hor- der. 4)- the phase angle θ at 0.1 Hz increases according
izontal to the frequency axis indicating a purely resis- to the above mentioned order.
tive value of modulus Z , equivalent to (R0 + R1 + R2 ), Fig. 10a and b represent respectively the plots of Rcorr
whilst the phase angle approaches zero. As the value and σ versus Fe% and Cr% of stainless steel alloys.
of σ increases, the log Z line at low frequencies is no
longer horizontal but curves upward, becoming steeper
and approaching a slop of − 12 at high value of σ . The T A B L E I I I Data of impedance measurements for the stainless steel
corresponding low frequency portion of the angle plot alloys at 550◦ C
show values of θ changing from zero towards 45◦ as σ
Alloy 1 Alloy 2 Alloy 3
increases. If the value of σ increases still further, be-
coming significantly greater than R2 , the slope of − 12 0
Rcorr (Ä) 3.62 11.4 15.21
00 (Ä)
is approached at low frequencies. From the definition Rcorr 3.8 11.6 15.4
of Z w in Equation 14, the modulus Z w (Ä) 1.5 4 5
σ1 (Äs1/2 ) 2.66 7.09 8.86
σ2 (Äs1/2 ) 2.1 6.7 9.11
p √ θ(degree) 36 44 46
Z= Z 02 + Z 002 = 2σ ω−1/2
2553
Figure 10 Variation of Rcorr and σ with Fe% (a) and Cr% (b) in the Figure 12 (a): Potential time curves for the stainless steel alloy 1 in 0.2 M
stainless steel alloys immersed in the pure carbonate melt at 500◦ C. HCl after oxidation in pure carbonate melt, under open circuit condition.
——¦ Rcorr —— ° σ. E2, 500◦ C, E4, 550◦ C. (b): Potential time curves for the stainless steel
alloy 2 in 0.2 M HCl after oxidation in pure carbonate melt under open
circuit conditions. E2, 500◦ C, E4, 550◦ C. (c): Potential time curves for
the stainless steel alloy 3 in 0.2 M HCl after oxidation in pure carbonate
melt under open circuit condition. E2, 500◦ C. E4, 550◦ C.

the surface of the three stainless steel alloys after their

immersion in carbonate melt. In these experiments, the
preoxidized electrodes were immersed in 0.2 M HCl
solution and their potential, measured against saturated
calomel electrode, was followed as a function of time
till reaching steady state, under the open circuit condi-
tions at ambient temperature (around 22◦ C).
Fig. 12a–c represent, respectively, the potential-time
curves for the three stainless steel alloys (1, 2, 3), im-
mersed in 0.2 M HCl solution after their oxidation un-
der open circuit conditions in pure carbonate melt at
temperatures of 500 and 550◦ C. It is clear that, on im-
mersion of the preoxidized electrodes in HCl solution,
their potential greatly and then slowly shifts to more
negative direction till attaining the steady state. The
time required to reach this steady state, ts , depends on
the temperature of oxidation and the type of stainless
steel alloy.
Figure 11 Variation of phase angle θ with Fe% (a) and Cr% (b) in the If we roughly consider the immersion potential, E imm
stainless steel alloys immersed in pure carbonate melt at 500◦ C. as the starting potential of the preoxidized electrode
and the steady state potential, E s , as the end potential,
While Fig. 11a and b represent the plots of phase angle the value of dissolution potential range (E imm − E s )
(θ) versus Fe% and Cr% of the alloys. The results of can be calculated. The time required for the shift of
table (3) and of Fig. 10a and b are in good agreement potential from E imm to E s , (time of dissolution) was
with those previously obtained by open circuit potential recorded, ts ,. The rate of dissolution of oxide scales
measurements. can be roughly calculated by dividing the value of dis-
solution potential range, (E imm − E s ) by the time of
3.4. Corrosion tests on the oxide scales dissolution ts ,. The obtained values of these calcula-
In this series of experiments trails are made to evaluate tions are listed in Table IV. Taking in consideration the
the corrosion resistance of the oxide scales formed on undertest stainless steel alloys and the values of their
2554
T A B L E I V Data of the corrosion tests on the oxide scales of the 4. R . A . D O N A D O , L . G . M A R I A N O W S K I , H . C . M A R U
stainless steel alloys and J . R . S E L M A N , ibid. 131(11) (1984) 2535.
5. Idem., ibid. 131(11) (1984) 2541.
Temp. E imm Es E imm − E s ts rate 6. O . P . P E N Y A G I N A , I . N . O Z E R Y A N A Y A , N . D .
Alloy (◦ C) (mV) (mV) (mV) (min) (mV/min) S H A M A N O V A and B . B . A N T O N O V , Tr. Inst. Electrochim.,
Ural Nauchn. Tsetr, Akad. Nauk SSSR 26 (1978) 48.
No. 1 500 −18 −544 536 15 35.73 7. R . T . C O Y L E , T . M . T H O M A S and G . Y . L A I , in High
500 −200 −530 330 3 110 Temp. Corros. Energy Syst. Proc. Symp. 672, 1984, edited by M. F.
No. 2 500 15 −269 284 24 11.83 Rothman (Metall. Soc. AIME. Werrendal, PA 1985).
550 132 −240 372 75 4.96 8. K . N A K A G A W A , T . I S O Z A K I , S . K I H A R A and B .
No. 3 500 180 −205 385 105 3.67 G I J U S T U , 36(7) (1987) 438.
550 168 −283 451 100 4.51 9. H . KIYOSHI, Y. TAKATOSHI, Y. TAKEHIKO,
F . Y U T A K A and O . K E I Z O U , Zairyo to Kankyo 40(2)
(1991) 130.
10. N . T A T S U O , Y . K O H I C H I and U . I S A M U , in Proc. Elec-
rates of dissolution, we can conclude that the oxide trochem. Soc. 1993, 93 (Proceedings of the Third International Sym-
scales formed on the surface of the austenitic stainless posium on Carbonate Fuel Cell Technology, 1993, p. 264).
steel alloys (2, 3) are more resistance to dissolution than 11. H . Y O K O K A W A , N . S A K A I , T . K A W A D A and M .
those formed on the ferritic stainless steel alloy (1). The D O K I Y A , J. Electrochem. Soc. 140(12) (1993) 3565.
values of E imm for austenitic electrodes (2, 3) are more 12. J . P . T . V O S S E N , L . P L O M P , J . H . W . DE W I T and
G . R I E T V E L D , ibid. 142(10) (1995) 3327.
positive than that of the ferritic electrode, indicating 13. C . G . L E E , H . N A K A N O , T . N I S H I N A , I . U C H I D A and
that the oxide scales of the austenitic alloys are more S . K U R O E , ibid. 145(8) (1998) 2747.
passive than those of ferritic alloys. The values of E s for 14. S . S E N D R O F F and A . B R E N N E R , ibid. 101 (1954) 31.
austenitic electrodes are more positive than that for the 15. J . O . M . B C K R I S , G . J . H I L L S , D . I N M A N and
ferritic electrode. This indicates that the oxide scales L . Y O U N G , J. Sci. Inster. 33 (1956) 438.
16. S . N . F L E N G A S and E . K . R I D E A L , Proc. Roy. Soc. A233
of the austenitic alloys are more resistant to dissolu- (1956) 443.
tion in HCl acid solution than those of the ferritic alloy. 17. S . N . F L E N G A S and T . R . I N G R A H A M , Canad. J. Chem. 35
Also, it seems that the oxide scales of austenitic alloys (1957) 1139.
are multilayered. The outer layer of these scales is less 18. J . P . T . V O S S E N , A . H . H . J A N S S E N and J . H . W . DE
protective and tends to dissolve in the acid, while the W I T . J. Electrochem. Soc. 143(1) (1996) 58.
19. J . P . T . V O S S E N , R . C . M A K K U S and J . H . W . DE W I T ,
inner layer od scales has high resistance to dissolution ibid. 143(1) (1996) 66.
in the present concentration of the acid solution. 20. M . D . I N G R A M , B . B A R O N and J . J A N Z , Electrochim. Acta
11 (1966) 1629.
Nomenclature 21. L . Y O U N G , “Anodic Oxidation Films” (Academic Press, London,
New York, 1961) p. 11257.
E imm : Immersion potential 22. Z . S Z L A R S K A - S M I A L O W S K A and R . W . S T E E H L E , J.
Es: Steady state potential. Electrochem. Soc. 12 (1974) 1393.
ts : Time needed to reach steady state values. 23. A . M . B E K H E E T , M . M . H E F N Y , A . A . M A Z H A R and
kf : Rate of formation of oxide scale M . S . E L - B A S I O U N Y , Ann. Chem. (Rome) 73 (1983) 63.
k−f : Rate of dissolution of oxide scale 24. M . S . E L - B A S I O U N Y , A . M . E L - K O T and M . M . H E F N Y ,
Corrosion 36 (1980) 244.
i Imm : Immersion galvanic current 25. U . R . E V A N S “The Corrosion and Oxidation of Metals” (Edward
is: Steady state current. Arnold, London, England, 1960) p. 898.
26. T . B U N Z O and O . T A K E O , Hyomen Gijutu 43(3) (1992) 233.
27. Q . W . W A L T E R , Corros. Sc. 26(9) (1986) 681.
References 28. M . A Z Z I and J . J . R A M E A U , ibid. 30 (1990) 439.
1. J . M . K I N G , J R ., Final Report, NASA CR, 134599, FCR-0237 29. J . P . T . V O S S E N , P . C . H . A M E N T and J . H . W . DE
W I T , J. Electrochem. Soc. 143(7) (1996) 2272.
(1977).
2. Quarterly Status Report for period Jan. 1–March 31, 1977, contract.
No.31-109-38-3952, Institute of Gas technology, Chicago, I11.
3. R . B . S W A R O O P , J . W . S I M and K . K I N O S H I T A , J. Elec- Received 12 November 1999
trochem. Soc. 125(11) (1978) 1799. and accepted 3 August 2000

2555