Corrosion Science 158 (2019) 108077

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

Microstructure, corrosion resistance and oxidation behavior of Cr-coatings T

on Zircaloy-4 prepared by vacuum arc plasma deposition
Tianguo Weia, ,1, Ruiqian Zhanga,1, Hongyan Yanga, Hong Liua, Shaoyu Qiua, Yu Wanga,
⁎

Peinan Dua, Kun Hea, Xiaogang Hub, Chuang Dongb

a
Science and Technology on Reactor Fuel and Materials Laboratory, Nuclear Power Institute of China, Chengdu, 610213, China
b
Institute of Materials Science and Technology, Dalian University of Technology, Dalian, 116086, China

ARTICLE INFO ABSTRACT

Keywords: Cr-coated Zircaloy-4 samples are prepared by vacuum arc plasma deposition and their microstructure, corrosion
Accident tolerance fuel cladding resistance and oxidation behavior are investigated. The as-received coating is dense with no pores or defects
Cr coating observed. The corrosion rate of Cr coating is significantly lower than Zircaloy-4 under the condition simulating
Zirconium alloys the coolant of both PWRs and BWRs. Cr coating can prevent zirconium from harsh and fatal oxidation in air.
Vacuum arc plasma deposition
Bubbles are found on the surface of the coated samples after oxidizing at 1200°C. These bubbles are within the
Corrosion
Zr-substrate, locating at the interface of α Zr (O) layer and prior β Zr.
Oxidation

1. Introduction temperature steam under accident condition, while have minimal effect
on the thermo-mechanical behavior of the cladding, and do not notably
The concepts of accident tolerance fuel (ATF) were proposed since change the core physics in the reactor [23]. Furthermore, due to the
2011 when the Fukushima nuclear accident happened in Japan. poor thermal conductivity of zirconia which grows on Zr-based alloys
Compared with the current UO2-Zr system, the ATF is expected to en- during operation, adequate coatings also have the potential to enhance
hance the safety of the fuel assemblies beyond design basis events the heat transfer characteristic of the cladding [24].
(DBAs), while maintain or improve the performance of fuel assemblies The potential advantages of coating technology give rise to the
under normal operational condition [1–3]. The ATF concepts involve prevalent studies of coated Zr-based alloys in recent years. It is believed
fuel pallet and fuel cladding. For the fuel pallet, improved thermal that the materials used as ATF coating need to contain at least one of
conductivity and enhanced retention of fission products is suggested the elements Cr, Al, or Si, because only the oxides of these elements
[4–7]. High uranium density fuel is also desired if non-Zr-alloy cladding exhibit high temperature steam oxidation resistance [1,21]. Thereby,
is used [8,9]. For the fuel cladding, improvement of oxidation re- coatings investigated for ATF cladding concept include metallic Cr
sistance and mechanical strength under accident condition is mainly [13,14,25], FeCrAl [25–27], CrN [12,28], TiAlN [29], Ti2AlC [30],
considered [10,11]. In respect that the release of radio nuclides to the Ti3SiC2 [31], et al. Some researchers also study the multi-layers coating
public was caused by hydrogen explosion in Fukushima accident [1], with different components, such as Cr/CrN [12] and TiAlN/TiN [32].
lower hydrogen generation during corrosion and oxidation of the K.A. Terrani suggests that [1]: besides the bonding performance, the
cladding is also desired. coatings which are worth for further development as coated Zr-based
The ATF cladding developments mainly consist of coated Zr-based ATF cladding have to meet the following requirements: corrosion re-
cladding [12–14], iron-based (mainly refers to FeCrAl system) cladding sistance in LWR coolant enrolments, neutron irradiation stability, and
[15–17] and SiCf/SiC cladding [18–21]. Besides, Mo-Zr cladding is also oxidation resistance to high temperature steam. The comprehensive
studied [22], but not so prevalent as the formers. Within these above assessment implies that the metallic Cr coating is the most promising
claddings, coated Zr-based cladding is an evolutionary approach by technology for further development, because only chromia is stable in
adoption of a protective coating at the surface of Zr-based alloys and it the LWR coolant environment while silica and alumina tend to dissolve
is deemed to be the most economic technology [1]. Adequate thin [1]. CrN and FeCrAl are also not deemed useful ATF coatings compared
coatings can prevent the fast oxidation of Zr-based alloy in high with Cr. The former shows no improvement in high temperature

⁎
Corresponding author.
E-mail address: [email protected] (T. Wei).
1
Both authors contribute equally to this work and should be considered as the first authors.

https://doi.org/10.1016/j.corsci.2019.06.029
Received 2 November 2018; Received in revised form 26 June 2019; Accepted 30 June 2019
Available online 09 July 2019
0010-938X/ © 2019 Elsevier Ltd. All rights reserved.
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 1. SEM images of coated Zircaloy-4 (a) (b) surface morphologies, (c) cross-sectional morphology, (d) is the spectrogram of the EDS mapping line in (c).

oxidation [28] and the latter will form a eutectic with Zr at tempera- with chromium ion bombardment for 5 min at a voltage of −1000 V.
tures < 1200°C [27]. The deposition process lasts for 8 h. The coated samples are then cooled
Various deposition technologies have been studied for Cr-coated Zr- naturally in the furnace. It takes about 1 h for the chamber temperature
based alloys [2,12–14,25]. Hyun-Gil Kim et al. used a 3D laser coating to decrease from 400°C to below 100°C during the cooling process.
technology to prepared Cr-coated Zircaloy-4 tubes [2,13]. The Cr layer, Finally, the coated samples are extracted when the chamber tempera-
although with a rough surface (can be easily controlled by grinding ture is below 60°C. The overall thickness of the coating is about 20 μm.
process [2]), not only exhibited excellent oxidation resistance than Corrosion tests are performed in autoclave simulating the water
uncoated tube at 1200°C, but also exhibited well adherence to the chemistry of PWRs and BWRs. The simulated PWR test is performed in a
Zirloy-4 surface in both ring tensile test and hoop direction deformation static autoclave in deaerated water with the temperature at 360°C and
test. Yiding Wang et al. [25] deposited Cr layer and FeCrAl alloys on the pressure at 18.6 MPa, containing 200 mg/kg H3BO4 and 1.2 mg/kg
Zircaloy-4 using atmospheric plasma technology. In their study, Cr LiOH. The simulated BWR test is performed in another autoclave con-
coating had superior oxidation resistance, whereas FeCrAl coating de- nected to a recirculation loop, with the temperature at 360 ℃ and the
graded due to inter-diffusion. Martin ševeček et al. [14] coated Zir- pressure at about 18.6 MPa as well. The water is deionized water con-
caloy-4 with pure Cr metal using the cold spray (CS) technique. They taining no H3BO4or LiOH. However, the dissolved oxygen (DO) level in
concluded that the CS technique, which is more cost efficient due to the water is maintained at about 100 ppb (part per billion) in mass,
lower energy consumption and high deposition rates, had high poten- controlled by a water chemistry system in the recirculation loop.
tial benefits but required further optimization and out-of-pile and in- Oxidation tests are carried out in a furnace in atmospheric environment
pile testing. In this study, Cr coating is prepared on Zircaloy-4 by va- at 800°C, 1000°C and 1200°C, respectively. In each test, an uncoated
cuum arc plasma deposition method. The microstructural morpholo- Zircaloy-4 sample is also put in the furnace for comparation. The
gies, corrosion behavior and oxidation behaviors of the coating are temperature increase rate is about 10°C /min during heating. The
investigated, using uncoated Zircaloy-4 for comparation. samples are then furnace cooling after oxidation. The surface and cross-
section morphologies of the corroded samples and oxidized samples are
2. Experimental observed by SEM and TEM. The TEM samples are prepared by FIB
sectioning. XRD is used to analyze the phase transformation of the
Chromium coatings are deposited on Zircaloy-4 plates by vacuum coating after oxidation.
arc evaporation with application of straight filters to separate the
macro-particles from the plasma. The Cr cathodes with its purity > 3. Results
99.95% is 160 mm in diameter. The substrates are as-received Zircaloy-
4 plates with a size of 30 mm × 20 mm × 1.4 mm. The substrate sam- 3.1. Microstructural characteristics
ples are ultrasonically cleaned in a hot alkaline bath for 10 min and
then in acetone solution for another 10 min, and air dried finally. After The microstructural characteristics of the deposited Cr coatings can
cleaning they are fixed to a turntable with a revolution speed of 10 rmp be seen in Fig. 1(a) and (b). The coatings are silvery white in color and
and a rotation speed of 40 rmp. The minimum substrate-cathode dis- dense without any microspores or defects. The surface of the coating is
tance is about 200 mm during deposition. relatively smooth with its roughness ranged from 280 to 330 nm.
The vacuum chamber was evacuated to a pressure of 5 × 10−3 Pa Fig. 1(c) shows the cross-sectional SEM image of the coating. The Cr
initially and then heated to 400°C. Deposition of Cr coating is per- layer is well adherent to the Zirlaloy-4 substrate. No pores or micro-
formed with the bias voltage at −120 V and the arc current at 150 A. cracks are observed at the interface. The coating is about 20 μm in
Before deposition the surface of the substrates are subjected to cleaning thickness, with the deviation less than 10%. The EDS line analysis

2
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 2. IPF coloring images by EBSD of (a) coated sample and (b) uncoated sample, (c) is the pole figure of the coating layer in (a).

shows that there is a transitional layer beneath the coating, as seen in weight gain of the uncoated Zircaloy-4 is about 28.7 mg/dm2 and that
Fig. 1(d). In the transitional layer, the content of Cr decreases inwards of the coated samples is only about 4.1 mg/dm2 (Fig. 3(a)). For the DO
while the content of Zr changes in the reverse direction. This transi- containing water, after 3000 h exposure, the average weight gain of the
tional layer is thought to act as the bonding layer between the Cr uncoated Zircaloy-4 is about 33 mg/dm2 and that of the coated samples
coating and Zircaloy-4 substrate. is only about 3.5 mg/dm2 (Fig. 3(b)). Comparing the weight gain data
Detailed information of the microstructure is obtained by EBSD under the two water chemistry conditions, it reveals that Zircaloy-4
analysis. Fig. 2 (a) and (b) show the grain map of the coated sample and samples exhibit slight higher corrosion rate in DO containing water,
uncoated sample, respectively. Obvious growth of grains in the sub- while the coated samples exhibit higher corrosion rate in H3BO3-LiOH
strate is observed after 8 h deposition. The grain growth is caused by containing water. This means that presence of DO, which can enhance
the absorption of fine grains into the around larger grains. However, the corrosion of zirconium alloys [33,34], has no enhancement effect on
unlike the abnormal growth of the crystals during high temperature the corrosion of Cr coating. Conversely, perhaps tiny quantity of DO in
annealing, the grains in the coated Zircaloy-4 plates still remain small, the water is favorable for Cr coating to forming a dense passive film,
with their average diameter at about 3.9 μm. Fig. 2(a) also shows the and prevent the Cr coating from corrosion in the later stage of the test.
columnar crystals of the Cr layer. These columnar crystals are of variant The appearance of the coated samples at different intervals of the
size and tend to grow with the increase of the coating thickness. The corrosion tests can be seen in Fig. 4(a) and (b). Some difference is seen
minimum width of the columnar crystal is less than 1 μm, while the on the colors of the samples. The sample in H3BO3-LiOH containing
maximum width of the columnar crystal can reach to 4 μm. Pole figures water becomes light golden color firstly, and then become dark golden
of the Cr layer imply the presence of texture, with the {100} planes with the increase of exposure time. The sample in DO containing water
along the normal direction of Zircloy-4 plates, as seen in Fig. 2(c). tends to show a bluish-purple color and then blue-back. Both the golden
color and blue color suggest that the thickness of the oxide layer formed
stays on a very limited level [35]. Fig. 4(c) shows the image of the
3.2. Corrosion tests coated sample corroded in H3BO3-LiOH containing water for 1000 h, it
is seen that no evidence of solution or separation is observed on the Cr
Two different water chemistry conditions are considered for the coating during corrosion, but the grain boundaries of Cr coating faded
corrosion tests. One is carried out in a static autoclave with the addition out gradually. And the samples in DO-containing water exhibit the
of H3BO3 and LiOH to the water to roughly simulate the chemistry of same phenomenon. This phenomenon is perhaps caused by the pre-
the coolant in PWRs. The concentration of H3BO3 and LiOH is 200 mg/ ferential corrosion of the grain boundaries of Cr coating, which is ob-
kg and 1.2 mg/kg, respectively. The other is carried out in an autoclave served clearly at the surface of the 168 h exposure sample, as seen in
attached to a recirculation loop, which is used to control the DO level at Fig. 4(d). Fig. 4(e) and (f) show the SEM images of the cross section of
about 100 ppb in the deaerated water. The water chemistry here is to the samples after corrosion in H3BO3-LiOH containing and DO con-
simulate the effect of DO on the corrosion behavior of claddings in taining water for 3000 h exposure, respectively. It is seen that the Cr
BWRs. coating remains well adherent to the underlying Zircaloy-4 substrate
Weight gains vs. exposure time of coated and uncoated Zircaloy-4 after long time exposure and no evidence of ZrO2 is observed in the
plates in the two water chemistry conditions are plotted in Fig. 3. For substrate, indicating that Zircaloy-4 substrate is prevented totally from
the H3BO3-LiOH containing water, after 3000 h exposure, the average

3
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 3. Weight gains vs. exposure time of coated and uncoated samples in 360 ℃/18.6 MPa water with addition of (a) H3BO3 and LiOH, (b) DO.

corrosion by the external Cr coating at 360°C. However, no visible outer illustrated in Fig. 5. It is seen that Cr coating can effectively prevent the
Cr2O3 oxide scale is observed in Fig. 4(e) and (f) due to its very limited Zr-substrate from harsh oxidation, especially at 1000°C and 1200. In the
thickness. 1200℃/1 h test, small bubbles are observed at the surface of the Cr
coating (as seen in Fig. 5(a)), but the coated sample still remains its
integrety and original shape. Compared with the coated sample, the
3.3. Oxidation tests uncoated sample undergoes harsh and degraded oxidation at 1000°C
and 1200°C. The uncoated sample bumps due to degraded oxidation in
Oxidation tests are carried out in a furnace with the increase rate of 1000°C /1 h test, with the oxide film cracking up at the surface, and
the temperature at about 10°C/min. Both coated sample and uncoated even breaking up to small spallation fragments in 1200°C /1 h test, as
sample are then kept at 800°C, 1000°C and 1200°C for 1 h, respectively, seen in Fig. 5(b) and (c).
and then furnace cooling to room temperature. The surface appearances From Fig. 5, it is also seen that the color of the coated sample
of the coated and uncoated samples after the oxidation tests are

Fig. 4. Optical micrographs and SEM images of coated samples after corrosion in (a) (c) (d) and (e) H3BO3-LiOH containing water, and in (b) (f) DO containing water.

4
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 5. Surface appearance of coated and un-

coated samples after oxidation tests (a) 800°C/
1 h, (b) 1000°C/ 1 h and (c) 1200°C/ 1 h.

changes with the increase of oxidation temperature. It is green after

800°C/1 h test, dark green after 1000℃/1 h test, and black after
1200°C/1 h test. The change of oxide color is possibly caused by the
subsequent reaction of Cr2O3 with O2 above 900°C [36]:
Cr2O3 + O2 CrO3 . This reaction is thought to produce pores at the
surface of the chromia film. The porous chromia will absorb visible
lights, so that the color of the chromia turns to black finally.
X-ray analysis is performed on the surface of the original coated
sample and the oxidized coated samples, and the results are illustrated
in Fig. 6. It is shown that Cr2O3 is formed at the surface of the Cr
coating after an hour oxidation test. The signal of the Cr spectrum re-
mains strong for the 800°C/ 1 h sample, decreases sharply for the
1000°C/ 1 h sample, and vanishes for the 1200°C/ 1 h sample. These
results indicate a significant increase of the thickness of Cr2O3 layer
with the oxidation temperature.
Fig. 7(a)˜(d) show the SEM images of the coated samples after
Fig. 6. XRD spectra of coated samples before and after oxidation tests.
oxidation tests. It is seen that the surface of the 800℃/1 h sample is

Fig. 7. Surface morphologies of coated samples after oxidation tests (a) 800°C/1 h, (b) 1000°C/1 h, (c) 1200°C/1 h and (e) 1200°C/3 h, (d) is the magnified image of
region A in (c), (f) is the magnified image of region B in (e).

5
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 8. Cross-sectional morphologies of coated samples after oxdiation tests (a) 800°C/1 h, (b) 1000°C/1 h, (c) 1200 ℃/1 h and (d) 1200°C/3 h, (e) is the spectrogram
of the EDS mapping linge in (c), (f) is the spectrogram of the EDS mapping linge in (d).

relatively more dense than that of the 1000°C/1 h sample. For the in the Cr2O3 layer. For the 1200°C/1 h sample, most small bubbles are
1200℃/1 h sample, two kinds of bubbles are observed, as seen in observed at the interface of the chromia layer and the Cr layer. For the
Fig. 7(c) and (d). The larger bubbles are mainly circular and about 1200℃/3 h sample, with the thickening of the Cr2O3 layer, the small
hundreds of micros in diameter. The smaller bubbles are not regular, bubbles are only observed within the Cr2O3 layer. Rough measurements
may be circular or worm-shape, and only tens of microns in diameter. show that the Cr2O3 layer of the 1200℃/1 h sample is about 7–10 μm
An additional oxidation test is carried out to study the development of and that of the 1200℃/3 h sample is more than 12 μm. Combining the
these bubbles at 1200°C for 3 h. The results show that with the increase results exhibited in Fig. 7, we can conclude here that the small bubbles
of oxidation time, all the smaller bubbles become worm-shape, as seen begin to form when the Cr2O3 layer reaches a specific thickness.
in Fig. 7(e). The magnified image shows that the grain size of the worm- Obvious layered structure is observed for the samples oxidized at
shape bubbles is obviously smaller than the grain size at their sur- 1200℃, as seen in Fig. 8(c) and (d). The outmost layer is the Cr2O3
rounding regions, as seen in Fig. 7 (f). Cracks are often observed on the layer, followed by the Cr layered with soluted O, the Cr-Zr diffusion
surface of the samples in both 1200°C/1 h and 1200°C/3 h tests, in- layer, and the substrate finally. Linear EDS mappings are performed in
dicating the failure of Cr2O3 layer at this temperature. Fig. 8(c) and (d), and the results are shown in Fig. 8 (e) and (f), re-
The cross-sectional images of the coated samples after oxidation spectively. It is seen that the width of the diffusion layer has grown
tests are exhibited in Fig. 8. For the 800°C/1 h sample, a rough mea- after high temperature exposure (compared with Fig. 1(d)). Moreover, a
surement shows the thickness of the Cr2O3 layer is about 400 nm. Small composition step with relative stable content of Cr and Zr is observed
pores are observed at the interface of the Cr2O3 layer and Cr layer. For within the Cr-Zr diffusion layer. The appearance of such composition
the 1000°C/1 h sample, the thickness of the Cr2O3 layer is about 2 μm. It step indicates the possible formation of Cr-Zr intermetallic phase, which
is observed that some fragments break off from the surface of the Cr2O3 will be discussed later.
layer to the resin during the sample preparation, implying the porous
nature of the Cr2O3 layer formed at higher temperature. Obvious 4. Discussion
changes are found for the 1200℃/1 h sample and 1200°C/3 h sample
with the appearance of large bubbles and small bubbles, corresponding 4.1. Microstructural characteristics
to the surface morphologies shown in Fig. 7. It is seen that the large
bubbles are formed in the substrate while the small bubbles are formed The vacuum arc plasma deposition method for coating preparation

6
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 9. Grain boundaries of (a) coated sample and (b) uncoated sample, (c) and (d) shows the distribution of the grain diameter in (a) and (b), respectively.

needs a long exposure time to get a thick coating layer. Thus, the d = W
x M+ y O
evolution of the microstructure of the substrate after long deposition y O (1)
time should be considered. In this study, about 8 h of deposition is
needed for the Cr coating to reach 20 μm in thickness, and growth of the where d is the thickness of the oxide scale, is the theoretic density of
grains is observed in the substrate after coating deposition. The grain oxide, W is weight gain per unit area, x and y are the coefficients in
boundaries images of the coated sample and uncoated sample (corre- the metal oxide Mx Oy , in which M and O are referred to the relative
sponding to Fig. 2(a) and (b)) are shown in Fig. 9(a) and (b). As men- molecular mass of metal element and oxygen, respectively.
tioned before, the growth of the grains is mainly caused by the ab- By Eq. (1), the relation between the thickness of the oxide scale and
sorption of fine grains into their adjacent larger grains, but not like the the weight gain per unit area is: 1 μm is equal to about 16.5 mg/dm2 for
abnormal grains growth of zirconium at higher temperature. The dis- Cr2O3 and about 14.7 mg/dm2 for ZrO2. Take the results in DO con-
tributions of the grain size in the coated sample and the uncoated taining water for example, after 3000 h exposure, the average thickness
sample are illustrated in Fig. 9 (c) and (d), respectively. From the re- of oxide film of the uncoated samples is about 2.2 μm, while that of the
sults, it is seen that although decided grain growth happens in the coated samples is only 0.2 μm. The value of the latter is about one order
substrate during the deposition of Cr coating, the grain size in the of magnitude less than the value of the former. It is has been proved
coated sample still remains small at the acceptable level. microscopically that for ZrO2, which grows on the surface of zirconium
alloy during corrosion, the relation is that 1 μm ZrO2 is about 15 mg/
dm2 in weight gain [37]. However, it is notable in this study that the
4.2. Corrosion tests calculated thickness of Cr2O3 scale is higher than the result from TEM
observation (Fig. 10(b)). The reason is not clear. Perhaps limited de-
The corrosion tests show that the corrosion rate of Cr coating is position of corrosion products from the autoclave in wall or sample
much slower than that of Zircaloy-4 alloy in the simulated coolant of bracket (made by stainless steel) may bring visible increase to the
both PWRs and BWRs. The thickness of the oxide scale of the corroded weight gain of coated sample due to its very low value. Whatever, it is
samples is very limited even after 3000 h exposure. Fig. 10(a) and (b) obvious that the corrosion rate of Cr is much lower than that of Zr in
show the TEM images of the corroded samples in H3BO3-LiOH con- both H3BO3-LiOH containing water and DO containing water at 360 ℃,
taining water and DO containing water after 3000 h respectively. It is implying the excellent protective nature of the Cr coating at the surface
seen that the oxide scale in both samples is 50˜100 nm in thickness. The of Zircaloy-4.
oxide scale is very protective with no pore or crack observed within it It is well known that there is a transition on the curve of the cor-
or at the oxide / Cr interface. EDS analysis is also performed in both the rosion kinetics of zirconium alloys. The transition occurs when the
oxide scale and the underlying Cr layer, and it is seen that the signal of thickness of the oxide film reaches to about 2˜3 μm [37–39]. After
O is much stronger in the former than in the latter. Fig. 10(c) presents transition, the corrosion rate of zirconium alloys will increase sharply
the High Resolution TEM (HRTEM) image along the interface of the and follows a linear kinetics, which calls post-transition corrosion. In
corroded sample in H3BO3-LiOH containing water for 3000 h exposure. this study, the corrosion kinetics of Zircaloy-4 alloys still remains in the
The associated FFT shows that the oxide scale is Cr2O3 with HCP pre-transition stage due to the limited exposure time. If the post-tran-
structure and the underlying is Cr layer with BCC structure. sition corrosion is considered, the protective effect of Cr coating on the
Generally, the thickness of the oxide scale could be calculated from corrosion of Zircaloy-4 will be much more significant. Undoubtedly, Cr
the weight gain of the sample. Assuming that the oxides on the surface coated zirconium alloys, which is developed as ATF cladding, also have
of Zircaloy-4 and Cr coating are pure ZrO2 and Cr2O3 in theoretic the potential to extend burn-up of fuel assemblies under normal op-
densities, the relation could be expressed as follows: eration condition. Besides, application of Cr coating on zirconium alloys

7
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 10. (a) Cross sectionals TEM image of the sample in H3BO3-LiOH containing water for 3000 h exposure (a) Cross sectionals TEM image of the sample in DO
containing water for 3000 h exposure (c) HRTEM image along the interface and associated FFT of the sample in H3BO3-LiOH containing water for 3000 h exposure.

cladding can solve the problem of nodular corrosion in BWRs, which is observed in the Cr2O3 layer, as seen in the magnified regions in Fig. 8(c)
caused by the presence of DO in the coolant [37,39]. and (d). It is conferred that the porous Cr2O3 is caused by the eva-
poration of gaseous CrO3, which is produced by the reaction of Cr2O3
with O2 at above 900°C [36]. This reaction will degrade the protective
4.3. Oxidation tests
nature of Cr2O3 layer at high temperature. Early research reported that
the evaporation of CrO3 is slight at 1000°C, but become severe with the
Oxidation tests indicate that Cr coating can prevent Zircaloy-4 from
further increase of temperature [40]. That is why very different oxi-
oxidation effectively in this study. However, it is seen that the protec-
dation rate is found on the 1000°C/1 h sample and 1200°C /1 h sample
tive nature of Cr2O3 will degrade with the increase of oxidation tem-
in this study. Moreover, formation of small bubbles is only observed in
perature. Especially for the samples oxidized at 1200 ℃, pores are often

8
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 11. Cross-sectional morphology of the coated sample after oxidized at 1200 for 3 h and images of the on-site elemental area mapping.

Table 1 surface of the coating after oxidizing at 1200°C. The cross sectional
Chemical compositions of the EDS point analysis in Fig. 11 (at. %). images show that these bubbles are formed in the substrate, but not at
Element Position
the coating-substrate interface, suggesting good adherence of the Cr
coating to Zircaloy-4 substrate during oxidation tests. The good ad-
1 2 3 4 5 herence may be attributed to the formation of the diffusion layer, which
acts as a good bonding layer between the coating and Zr substrate, as
O 65.32 7.64 5.73 0.00 23.54
seen in Fig. 11. As mentioned before in Fig. 8(e) and (f), the con-
Cr 34.68 92.07 93.18 64.05 0.26
Zr 0.00 0.29 1.08 35.95 76.20 centration of Cr and Zr in the diffusion layer is relatively stable, im-
plying a possible new phase is formed. Semi-quantitative analysis of the
composition of this layer is performed by EDS using point-mode. The
result shows that the content of Cr and Zr in this diffusion layer is 64.05
at.% and 39.05 at.%, respectively. This composition is in the range of
ZrCr2 intermetallic phase in Zr-Cr phase diagram [41]. Thus, it is
concluded that ZrCr2 phase is formed in the diffusion layer during the
oxidation test. An interesting phenomenon is that not only Cr diffuses
into Zr to form ZrCr2; Zr can also diffuse reversely into the Cr through
the grain boundaries of the coating, as seen in Fig. 11.
From the elemental EDS mapping images in Fig. 11 and the results
of semi-quantitative composition analysis illustrated in Table 1, it is
seen that oxygen can diffusion through the Cr coating layer into the Zr-
substrate. The oxygen content in the substrate beneath the diffusion
layer is more than 20 at.% at point 5 in Fig. 11. Increase of oxygen
content in the outer layer of the substrate will cause the phase trans-
formation of Zr substrate from β phase to α phase during oxidizing, as
seen in the Zr-O phase diagram Fig. 12. The outer α phase layer is
stabilized by oxygen and it is expressed as Zr (O). The Zr (O) layer is
often seen in the matrix of zirconium alloys after oxidation [42,43].
Fig. 13(a) exhibits the morphology of a bubble in the 1200 ℃/3 h
sample. Another image in the same field of Fig. 13(a) with lower
Fig. 12. Zr-O phase diagram.
brightness and higher contrast is shown in Fig. 13(b), in which the clear
boundaries of α Zr (O) layer is observed in the substrate. It is decided
the Cr2O3 layer of the samples oxidized at 1200°C, and some decohesion that the bubble is formed at the interface of the α Zr (O) layer and the
is occurring with some small bubbles, as seen in Fig. 8(c). This phe- beneath prior β Zr. The prior β Zr is referred to the part which trans-
nomenon also leads to the degradation of the stability and density of the form to β phase during high temperature exposure. Semi-quantative
Cr2O3 layer. composition analysis shows that the solution content of O in the α Zr
As mentioned before, large circular bubbles are observed on the (O) layer is significantly higher than in prior β Zr, and the results are

9
T. Wei, et al. Corrosion Science 158 (2019) 108077

Fig. 13. Cross-sectional morphologies of coated samples after oxidized at 1200°C for (a) (b) 3 h and (c) (d) 1 h.

Table 2 illustrated in Table 2. Clear α Zr (O) layer is also observed in the

Chemical compositions of the EDS point analysis in Fig. 13(b) (at. %). 1200°C/1 h sample (Fig. 13 (c) and (d)), but much thinner than that in
Element Position
the 1200°C /3 h sample due to shorter exposure time. Statistical ob-
servations of the cross-sectional morphologies of the bubbles show that
1 2 3 4 5 the distances from the center of the bubbles to the coating-substrate
interface in 1200°C/3 h sample is further than in 1200℃/1 h sample,
O 21.27 20.37 13.96 10.64 18.98
Cr 0.17 0.16 9.37 0.46 0.00
proving again that the bubbles are formed at the interface between α Zr
Zr 78.56 79.47 76.77 88.90 81.02 (O) layer and prior β Zr.
The evolution of the coated sample during oxidation could be de-
scribed in Fig. 14. Before oxidation, the coating is pure Cr and the

Fig. 14. Schematic for the microstructural evolution of the coated sample during oxidation at 1200°C.

10
T. Wei, et al. Corrosion Science 158 (2019) 108077

Zircaloy-4 is in α phase. A diffusion layer is observed beneath the Cr temperature during oxidation.
coating. This diffusion layer is formed during the deposition process (5) Smaller bubbles and larger bubbles are observed at the surface of
and it acts as the bonding layer of the Cr coating and Zircaloy-4 sub- the coated samples after the oxidation tests at 1200°C. The smaller
strate. When the sample is exposed to high temperature (ex. 1200°C), bubbles are within the Cr2O3 layer while the larger bubbles are
Zircaloy-4 substrate transforms to β phase, and Cr2O3 is formed on the within the Zircaloy-4 substrate. The locations of these bubbles in-
surface of the coating due to oxidation. Since the temperature is above dicate indirectly the good adherence of Cr coating to Zircaloy-4
900°C, CrO3 is produced due to the further oxidation of Cr2O3 with O2. substrate even at 1200°C. The small bubbles are thought to begin to
Evaporation of CrO3 degrades the protective nature of Cr2O3 to some form when the Cr2O3 layer grows to a specific thickness. The larger
extent. Moreover, some oxygen can diffuse into the Cr coating, and then bubbles locate at the interface of α Zr (O) layer and the prior β Zr in
diffuses subsequently from the coating to the beneath Zr substrate, as Zircaloy-4 substrate. The α Zr (O) layer is caused by phase trans-
the schematic illustrated in Fig. 14(b). With the increase of oxygen formation from β Zr at high temperature due to ingress of oxygen
content, phase transformation of Zr from β to α phase occurs in the through the coating.
outer layer of the substrate, so that an oxygen-stabilized α Zr(O) layer is
formed. Simultaneously, the diffusion layer between the coating and Acknowledgements
substrate widens due to enhanced diffusion, and the intermetallic phase
ZrCr2 is formed within this layer. Financial support for this work is provided by the National Natural
From the crystal lattice parameters of Zr, it is calculated that a Science Fundation of China (No. 51801195). The authors are grateful to
volume expansion at about 4.28% will happen when Zr transforms from Professor Lefu Zhang in Shanghai Jiaotong University who provide his
β phase to α phase. The volume expansion will cause cracks at the in- great help in the corrosion tests.
terface of α (Zr) layer and β Zr. Initial site for a bubble will be formed in
the regions where such cracks connect together, as seen the in References
Fig. 14(c). It is not sure in this study whether the bubbles are formed in
the isothermal heating process or in the cooling process. Here we [1] K.A. Terrani, Accident tolerant fuel cladding development: promise, status, and
suggest that the bubbles are formed in the cooling process. If the bub- challenges, J. Nucl. Mater. 501 (2018) 13–30.
[2] H.G. Kim, J.H. Yang, W.J. Kim, Y.H. Koo, Development status of accident-tolerant
bles are formed in the isothermal heating process, the distance from the fuel for light water reactors in Korea, Nucl. Eng. Technol. 48 (2016) 1–15.
center of the bubble to the coating-substrate interface should be the [3] M. Wagih, B. Spencer, J. Hales, K. Shirvan, Fuel performance of chromium-coated
same for 1200°C/1 h sample and 1200°C /3 h sample, and continuous zirconium alloy and silicon carbide accident tolerant fuel claddings, Ann. Nucl.
Energy 120 (2018) 304–318.
growth of α (Zr) layer would be observed around or beneath the bubble, [4] S.J. Zinkle, K.A. Terrani, J.C. Gehin, L.J. Ott, L.L. Snead, Accident tolerant fuels for
but such assumed phenomenon is not observed. Therefore, we conclude LWRs: a perspective, J. Nucl. Mater. 448 (2014) 374–379.
that bubbles are formed during cooling of the sample. However, we still [5] D.J. Kim, Y.W. Rhee, J.H. Kim, K.S. Kim, J.S. Oh, J.H. Yang, Y.H. Koo, K.W. Song,
Fabrication of micro-cell UO2-Mo pellet with enhanced thermal conductivity, J.
insist that phase transformation of β Zr to α Zr (O) will cause cracks Nucl. Mater. 462 (2015) 289–295.
within the α Zr (O) layer, as seen in Fig. 13(b) and (d). It is well known [6] J.H. Yang, D.J. Kim, K.S. Kim, Y.H. Koo, UO2-UN composites with enhanced ur-
that the oxygen content is an important parameter of the mechanical anium density and thermal conductivity, J. Nucl. Mater. 465 (2015) 509–515.
[7] J.H. Yang, K.S. Kim, D.J. Kim, J.H. Kim, J.S. Oh, Y.W. Rhee, Y.H. Koo, Micro-cell
properties of zirconium alloys [37,44–47]. The oxygen stabilized α Zr
UO2 pellets for enhanced accident tolerant fuel, Top Fuel, American Nuclear
(O) phase has much higher hardness than the prior β phase [46]. The α Society, Charlotte, USA, 2013 Sep. 15-19, 2013.
Zr (O) zone is brittle while the prior β zone remains its ductility [8] K.A. Terrani, S.J. Zinkel, L.L. Snead, Advanced oxidation resistance iron-based al-
[46,47].Such distinct mismatch of hardness and ductility of the two loys for LWR fuel cladding, J. Nucl. Mater. 448 (2013) 420–435.
[9] Y. Yamamoto, B.A. Pint, K.A. Terrani, K.G. Field, Y. Yang, L.L. Snead, Development
phases makes the phase interface the weak site for the bubble formation and property evaluation of nuclear grade wrought FeCrAl fuel cladding for light
in Zr-substrate during cooling. water reactors, J. Nucl. Mater. 467 (2015) 703–716.
[10] R.B. Rebak, Alloy selection for accident tolerant fuel cladding in commercial light
water reactors, Metall. Mater. Trans. E 2 (2015) 197–207.
5. Conclusions [11] R.B. Rebak, P.L. Andresen, Y.J. Kim, E.J. Dolley, Characterization of Advanced
Steels As Accident Tolerant Fuel Cladding for Light Water Reactors, (2014) IAEA-
Vacuum arc plasma deposition method is used to prepare Cr coating TECDOC-1797.
[12] A.S. Kuprin, V.A. Belous, V.N. Voyevodin, V.V. Bryk, R.L. Vasilenko,
on the surface of Zircaloy-4 alloy. The microstructure, corrosion re- V.D. Ovcharenko, E.N. Reshetnyak, G.N. Tolmachova, P.N. V’yugov, Vacuum-arc
sistance and oxidation behaviors of these Cr-coated samples are studied chromium-based coatings for protection of zirconium alloys from the high-tem-
and the results are summarized as follows: perature oxidation in air, J. Nucl. Mater. 465 (2015) 400–406.
[13] H.G. Kim, I.H. Kim, Y.-I. Jung, D.J. Park, J.Y. Park, Y.H. Koo, Adhesion property
and high-temperature oxidation behavior of Cr coated Zircaloy-4 cladding tube
(1) The as-received coating is dense without any pores or defects and it prepared by 3D laser coating, J. Nucl. Mater. 465 (2015) 531–539.
is well adherent to the Zircaloy-4 substrate. A diffusion layer is [14] M. Ševeček, A. Gurgen, A. Seshadri, Y. Che, M. Wagih, B. Phillips, V. Champagne,
K. Shirvan, Development of Cr cold sprayecoated fuel cladding with enhanced ac-
observed beneath the coating, and it acts as the bonding layer be-
cident tolerance, Nucl. Eng. Technol. 50 (2018) 229–236.
tween the coating and the substrate. [15] K.G. Field, S.A. Briggs, K. Sridharan, R.H. Howard, Y. Yamamoto, Mechanical
(2) Grain growth is found in the substrate of Zircaloy-4 alloy after properties of neutron-irradiated model and commercial FeCrAl alloys, J. Nucl.
deposition of coating at 400°C for 8 h, but the grain size still remain Mater. 489 (2017) 118–128.
[16] S. Ukai, N. Oono, S. Ohtsuka, T. Kaito, K. Sakamoto, T. Torimaru, et al.,
small at the acceptable level. Development of FeCrAl-ODS steels for ATF cladding, Proc. TopFuel 2016, (2016),
(3) Corrosion tests show that Cr coating corrodes much slower than pp. 681–689.
Zircaloy-4 alloy in both H3BO3-LiOH containing water and DO [17] B.A. Pint, Performance of FeCrAl for accident-tolerant fuel cladding in high tem-
perature steam, Corros. Rev. 35 (2017) 167–175.
containing water at 360 ℃. The thickness of the Cr2O3 oxide scale is [18] H.K. Naofiimi Nakazato, Y. Kohno, A. Kohyamal, SiC/SiC fuel cladding by NITE
very limited, only about 50˜100 nm after 3000 h exposure. No process for innovative LWR-Cladding forming process development, Ceram.
evidence of solution or separation is observed on the surface of the Environ. Energy Appl. II Ceram. Trans. 246 (2014) 109.
[19] K. Shimoda, T. Hinoki, H. Kishimoto, A. Kohyama, Enchanced hightemperature
corroded sample and no pore or crack is observed in the Cr2O3 performances of SiC/SiC composites by high densification and crystalline structure,
oxide scale, implying the very stable nature of Cr coating in the Compos. Sci. Technol. 71 (2011) 326–332.
simulated coolant of both PWRs and BWRs. [20] C.P. Deck, G.M. Jacobsen, J. Sheeder, O. Gutierrez, J. Zhang, J. Stone, et al.,
Characterization of SiC/SiC composites for accident tolerant fuel cladding, J. Nucl.
(4) Oxidation tests show that Cr coating can prevent Zircaloy-4 from
Mater. 466 (2015) 667–681.
harsh and fatal oxidation effectively at high temperature. [21] B. Cheng, Y.J. Kim, P. Chou, Improving accident tolerance of nuclear fuel with
Examinations of the thickness of Cr2O3 layer indicate that the oxi- coated Mo-alloy cladding, Nucl. Eng. Technol. 48 (2016) 16–25.
[22] I. Younker, M. Fratoni, Neutronic evaluation of coating and cladding materials for
dation rate of Cr coating increases significantly with the exposure

11
T. Wei, et al. Corrosion Science 158 (2019) 108077

accident tolerant fuels, Biol. Sci. 88 (2016) 10–18. on oxidation and hydrogen pick up behaviour-Zircaloy vs Zr-Nb alloys, Nucl. Eng.
[23] D.H. Kam, J.H. Lee, T. Lee, Y.H. Jeong, Critical heat flux for SiC-and Cr-coated Des. 240 (2010) 985–994.
plates under atmospheric condition, Ann. Nucl. Energy 76 (2015) 335–342. [35] J. Bischoff, C. Delafoy, P. Barberis, D. Perche, B. Buerin, J.-C. Brachet, Development
[24] B.A. Pint, K.A. Terrani, Y. Yamamoto, L.L. Snead, Material Selection for accident of Cr-coated zirconium cladding for enhanced accident tolerance, Proceedings of
tolerant fuel cladding, Metall. Mater. Trans. E 2 (3) (2015) 190–196. Top Fuel Conference (2016).
[25] Yiding Wang, Wancheng Zhou, Qinlong Wen, Xingcui Ruan, Fa Luo, Guanghai Bai, [36] H.C. Graham, H.H. Davis, Oxidation / vaporization kinetics of Cr2O3, J. Am. Ceram.
Yuchang Qing, Dongmei Zhu, Zhibin Huang, Yanwei Zhang, Tong Liu, Rui Li, Soc. 54 (1971) 89–93.
Behavior of plasma sprayed Cr coatings and FeCrAl coatings on Zr fuel cladding [37] IAEA-TECDOD-996, 1998.
under loss-of-coolant accident conditions, Surf. Coat. Technol. 344 (2018) 141–148. [38] N. Ni, S. Lozano-Perez, J.M. Sykes, G.D.W. Smith, C.R.M. Grovenor, Focused ion
[26] W.C. Zhong, P.A. Mouche, X.H. Han, B.J. Heuser, K.K. Mandapaka, G.S. Was, beam sectioning for the 3D characterization of cracking in oxide scales formed on
Performance of iron-chromium-aluminum alloy surface coatings on Zircaloy 2 commercial ZIRLO™ alloys during corrosion in high temperature pressurized water,
under high-temperature steam and normal BWR operating Conditions, J. Nucl. Corros. Sci. 53 (2011) 4073–4083.
Mater. 470 (2016) 327–338. [39] B. Cox, Some thoughts on the mechanisms of in reactor corrosion of zirconium
[27] K.A. Terrani, C.M. Parish, D. Shin, B.A. Pint, Protection of zirconium by alumina- alloys, J. Nucl. Mater. 336 (2005) 331–368.
and chromia-forming iron alloys under high-temperature steam exposure, J. Nucl. [40] C.S. Tedmon, The effect of oxide volatilization on the oxidation kinetics of Cr and
Mater. 438 (2013) 64–71. Fe–Cr alloys, J. Electrochem. Soc. 113 (1966) 766–768.
[28] K. Terrani, B.A. Pint, L.L. Snead, Y. Yamamoto, High-temperature steam oxidation [41] T.B. Massalski, H. Okamoto, P.R. Subramanian, L. Kacprzak, Binary Alloy Phase
of accident tolerant fuel cladding candidate materials, High Temperature Steam Diagrams, second ed., ASM Publication, Ohio, 1996.
Oxid. Accid. Toler. Fuel Cladding Candidate Mater, (2014), pp. 169–176 IAEA- [42] M. Steinbrück, J. Birchley, A.V. Boldyrev, A.V. Goryachev, M. Grosse, T.J. Haste,
TECDOC-1797. Z. Hózer, A.E. Kisselev, V.I. Nalivaev, V.P. Semishkin, L. Sepold, J. Stuckert, N. Vér,
[29] E. Alat, A.T. Motta, R.J. Comstock, J.M. Partezana, D.E. Wolfe, Ceramic coating for M.S. Veshchunov, High-temperature oxidation and quench behavior of Zircaloy-4
corrosion (c3) resistance of nuclear fuel cladding, Surf. Coat. Technol. 281 (2015) and E110 cladding alloys, Prog. Nucl. Energy 52 (2010) 19–36.
133–143. [43] A. Michau, F. Maury, F. Schuster, F. Lomello, J.-C. Brachet, E. Rouesne, M. Le Saux,
[30] C.C. Tang, M. Steinbrueck, M. Stueber, M. Grosse, X.J. Yu, S. Ulnch, H.J. Seifeit, R. Boichot, M. Pons, High-temperature oxidation resistance of chromium-based
Deposition, characterization and high temperature steam oxidation behavior of coatings deposited by DLI-MOCVD for enhanced protection of the inner surface of
single-phase Ti2AlC-coated Zircaloy-4, Corros. Sci. 135 (2018) 87–98. long tubes, Surf. Coat. Technol. 349 (2018) 1048–1057.
[31] K. Lambrinou, M. Verwerft, J. Vleugels, A. Weisenburger, C. Lorrette, Y. de Carlan, [44] C.M. Lee, H.J. Lee, H.G. Kim, Y.K. Mok, Y.S. Han, D.S. Sohn, Effect of pre-oxidation
et al., Innovative accident-tolerant fuel cladding materials: the H2020 IL temperature on the ductility of a zirconium alloy cladding under simulated accident
TROVATORE perspective, Water React. Fuel Perform. Meet. Jeju Island, Korea, conditions, J. Nucl. Mater. 515 (2019) 80–90.
(2017). [45] H.M. Chung, Fuel behavior under loss-of-coolant accident situations, Nucl. Eng.
[32] E. Alat, A.T. Motta, R.J. Comstock, J.M. Partezana, D.E. Wolfe, Multilayer (TiN, Technol. 37 (2005) 327–362.
TiAlN) ceramic coatings for nuclear fuel cladding, J. Nucl. Phys. Mater. Sci. Radiat. [46] M. Billone, Y. Yan, T. Burstseva, R. Daum, Cladding Embrittlement During
Appl. 478 (2016) 236–244. Postulated Loss-of-Coolant Accidents, U.S. Nuclear Regulatory Commission, 2008
[33] B. Cox, F. Garzarolli, P. Rudling, Corrosion of Zr-Nb alloys, ZIRAT-9 NUREG/CR-6967 ANL-07/04.
SPECIALTOPICS REPORT, Advanced Nuclear Technology International, Ekbaken [47] A. Strasser, F. Garzarolli, P. Rudling, ‘Processes Going on in Nonfailed Rod During
33, SE-735, 35, SURAHAMMAR, Sweden, 2004 September. Accident Conditions (LOCA and RIA) Volume II’, ANT International, Mölnlycke,
[34] M.K. Kumar, S. Aggarwal, V. Kain, T. Saario, M. Bojinov, Effect of dissolved oxygen Sweden, 2010.

12