Surface & Coatings Technology 432 (2022) 128066

Contents lists available at ScienceDirect

Surface & Coatings Technology

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

Hot corrosion behavior of dense CYSZ/YSZ bilayer coatings deposited by

atmospheric plasma spray in Na2SO4 + V2O5 molten salts
Jhonattan de la Roche a, *, Juan Manuel Alvarado-Orozco b, Pablo Andrés Gómez a, c,
Irene García Cano d, Sergi Dosta d, Alejandro Toro a
a
Tribology and Surfaces Group, Universidad Nacional de Colombia, Calle 75 # 79A-51, C.P. 050034 Medellín, Colombia
b
Centro de Ingeniería y Desarrollo industrial- CIDESI, Av. Playa Pie de la Cuesta No. 702, Desarrollo San Pablo, C.P. 76125 Querétaro, Mexico
c
Empresas Públicas de Medellín – EPM, Medellín, Colombia
d
CPT Thermal Spray Center, Materials Engineering, Departament d'Enginyeria Química I Metallúrgica, Universitat de Barcelona, Martí i Franquès 1, E-08028 Barcelona,
Spain

A R T I C L E I N F O A B S T R A C T

Keywords: In this work, the hot corrosion (HC) resistance of bilayer thermal barrier coatings (TBC) architectures composed
TBCs of dense Ceria-Yttria Stabilized Zirconia (CYSZ) and Yttria-Stabilized Zirconia (YSZ) deposited using the At­
YSZ mospheric Plasma Spray (APS) technique was evaluated. Two kinds of HC tests were performed using different
CYSZ
salt concentrations and thermal cycling (adding fresh salt between cycles) with a constant salt mixture (Na2SO4-
Hot corrosion
Air plasma spray
V2O5) and constant testing temperature (900 ◦ C). The results showed that the HC mechanism changes with the
testing procedure, and the outer dense CYSZ layer acts as a barrier to protect the inner layers against HC
degradation. However, the formation of vertical cracks during the coating deposition process affected the
corrosion resistance of the TBC systems.

1. Introduction
Thermal barrier coatings (TBCs) are used to protect parts exposed to
the highest operating temperatures in industrial gas turbines (IGTs),
allowing operating temperatures above the inherent capacity of the
metallic substrates and thus, enhancing turbine efficiency during oper­
ation. The parts to be protected are mainly those located on the hot gas
path, including the combustor chamber, liners, first stage stationary and
rotating blades, and seals [1–4].
The Top Coat (TC) layer most used as a thermal insulator is Yttria-
Stabilized Zirconia containing 7–8 wt% of yttria in solid solution
(7YSZ). This amount of yttria stabilizes the non-transformable tetrag­
onal phase (t′ ) of zirconia at room temperature, meaning it does not
transform to the monoclinic phase during cooling and is more stable
[1,5]. This phase possesses a high thermal expansion coefficient
(10.5–11.5 × 10− 6 K− 1), high toughness, and low thermal conductivity
(1.2–1.8 Wm− 1 K− 1) [6–10] when compared to the monoclinic
structure.
When an IGT operates with low-quality liquid fuels containing high
amounts of impurities such as sodium (Na), vanadium (V), and sulfur
(S), these might form salts with strong acid-base properties during
combustion, e.g., sodium sulfate (Na2SO4), vanadium pentoxide (V2O5),
and sodium vanadate (NaVO3) [11]. These molten salts infiltrate
through any TBC defects such as pores and cracks, and react mainly with
the Y2O3 phase, promoting the destabilization of the t′ -phase [12–14].
This failure mechanism is known as salt-induced hot corrosion (HC)
attack, the main reactions between the molten salts and YSZ being as
follows [13]:
Na2 SO4 + V2 O5 ⇌2NaVO3 + SO3 (1)
t′ − ZrO2 (Y2 O3 ) + V2 O5 ⇌m − ZrO2 + 2YVO4 (2)
t′ − ZrO2 (Y2 O3 ) + 2NaVO3 ⇌m − ZrO2 + 2YVO4 + Na2 O (3)
Eqs. (1)–(3) describe reactions with a Lewis acid-base nature and the
fluxing (dissolution) of the 7YSZ in the molten salt. These synergistic
interactions generate depletion of the Y2O3 phase content in the 7YSZ
solid solution, promoting the destabilization of t′ -YSZ to form m-ZrO2;
this phase transformation involves a 3–5% volumetric expansion and
formation of yttrium orthovanadate (YVO4) along the infiltration path

* Corresponding author.
E-mail addresses: jdey@unal.edu.co (J. de la Roche), juan.alvarado@cidesi.edu.mx (J.M. Alvarado-Orozco), pablo.gomez@epm.com.co (P.A. Gómez), igcano@
cptub.eu (I.G. Cano), sdosta@cptub.eu (S. Dosta), aotoro@unal.edu.co (A. Toro).

https://doi.org/10.1016/j.surfcoat.2021.128066
Received 13 November 2021; Received in revised form 21 December 2021; Accepted 23 December 2021
Available online 4 January 2022
0257-8972/© 2022 Elsevier B.V. All rights reserved.
J. de la Roche et al. Surface & Coatings Technology 432 (2022) 128066

Fig. 1. D-CZ/YSZ coating systems deposited by Air Plasma Spray (APS).

600–850 μm), cleaned with isopropyl alcohol in an ultrasonic bath for 3

Table 1
min, and dried with cold air. The samples were then pre-heated at
Deposition condition of the NiCoCrAlY, L-YSZ, YSZ, and D-CZ layers.
approximately 160 ◦ C before spraying each layer. The spray parameters
Parameters Power Spray Ar gas Secondary Spray of each layer had previously been optimized, and the best conditions are
(kW) distance flow gas (NLPM) rate
summarized in Table 1.
(cm) (NLPM) (kg/s)

Layer NiCoCrAlY 36 11.3 80 70 (He) 0.025 2.2. Hot corrosion tests

L-YSZ 48 10.5 50 7 (H2) 0.02
YSZ 40 10.5 50 7 (H2) 0.02
D-CZ 50 9 50 5 (H2) 0.02
through the TBC thickness generating stress, cracking and subsequent
delamination of the YSZ layer [14–16].
One way to counter the accelerated degradation generated by HC
attack is to design multimaterial, multilayered TBC architectures that
can slow down the infiltration of the salts and promote their self-healing
during service. These TBC architectures might include bilayer struc­
tures, densified vertical cracking overlays, and functionally graded
structures [17–21]. Laser glazing treatments have also been employed to
modify the TC surface [22–24], and alternative materials with higher HC
resistance to molten salts like La2Zr2O7 [25], Sm2Zr2O7 [26], Gd2Zr2O7
[27–29], and ZrO2 stabilized and co-stabilized with In2O3, Sc2O3 or
Ta2O5 [16,30–33] have been tested.
Ceria-Yttria Stabilized Zirconia (CYSZ) is a good TC candidate for
TBC systems due to its higher thermal expansion coefficient and good
thermal shock resistance, lower thermal conductivity, and greater sta­
bility in comparison to 7YSZ [9]. Moreover, CYSZ has presented
outstanding resistance to salt-induced attack [34–38] because the
cerium oxide (CeO2) is a t′ -phase stabilizer of zirconia and exhibits
better resistance to salt-induced HC due to its higher acidity compared to
Y2O3, which reduces the reactivity of the TC with the molten salts [39].
This work studies four different TBC architectures composed of dense
CYSZ (D-CZ) and porous 7YSZ layers, exposed to two different HC
conditions. The degradation of the coatings due to HC was investigated
and correlated with the crystal structure, chemical and morphological
changes observed in the samples.
2. Materials and methods
2.1. Coatings deposition
A Sinplex pro-gun (Oerlikon Metco, USA) was used to deposit the D-
CZ/YSZ systems onto Inconel 625 disks (Ø25.4 × 5 mm) by Atmospheric
Plasma Spray (APS) in the facilities of Empresas Públicas de Medellín,
Colombia (EPM). Fig. 1 illustrates the four multilayered TBC architec­
tures deposited, in which the thickness of the outer D-CZ layer was
designed to be 50, 100 or 150 μm, maintaining a total thickness of the TC
of 300 μm. All Top Coat (TC) architectures had a high-toughness, Low-
Porosity YSZ (L-YSZ) layer with a thickness of about 50 μm between the
bond coat (BC) and the porous YSZ layer, which has proven effective at
to enhancing the thermomechanical resistance of the system [40–42].
Amdry 386-2.5, Metco 204NS-G, and 205NS (Oerlikon Metco, USA)
powders were used as feedstock material for the NiCoCrAlY, BC, and YSZ
and CYSZ as the TC layers, respectively. Before the spraying process,
Inconel 625 substrates were blasted with Al2O3 particles (F24 grain size:
The HC tests were performed using high-purity Na2SO4 (JT Baker)
and V2O5 (Aldrich chemistry) powders in a mixture of 32 wt% of Na2SO4
and 68 wt% V2O5 at the CIDESI facilities in Queretaro, Mexico. The
mixture was spread over the surface of the samples, avoiding the edges.
The D-CZ/YSZ TBC architectures were evaluated following the meth­
odology reported in [11,42] in triplicate for each test, using two
different types of HC tests as described below.
HC-Concentration test: The salt deposit concentration was set to 1, 3,
and 5 wt% of the total mass of the TC. Samples were covered with the
salt deposit, heated up to 900 ◦ C, and held for 10 h in air atmosphere
inside the oven.
HC-Cyclic test: The samples were exposed to a salt deposit with a
concentration of 1 wt% and then treated at 900 ◦ C for 10 h to
complete one cycle. Then, the samples were cooled down to room
temperature, and a fresh salt mixture was added (1 wt%) over the
sample and exposed to the next cycle. The process was repeated until
the tetragonal-phase percentage decreased to 10 wt% or less.
2.3. Coating characterization
Microstructural characterization of the CYSZ/YSZ TBC architectures
was conducted at the Tribology and Surfaces laboratory of Universidad
Nacional de Colombia in Medellín, Colombia, and the CPT Thermal
Spray Center in Barcelona, Spain, according to ASTM E1920 for sample
preparation and ASTM2109 for porosity quantification, using an eclipse
LV1000 (Nikon, Japan) optical microscope (LOM) equipped with a
Nikon DS-2Mv digital camera. Morphological characterization was
carried out using a JSM-5910LV (JEOL, Japan) scanning electron mi­
croscope (SEM). Chemical characterization was carried out using Energy
Dispersive Spectroscopy (EDS) with a Pentafet 7324 detector (Oxford
Instruments, UK), operating at 133 eV (nominal area 10mm2) with an
ATW2 window.
A Smart Lab (Rigaku, Japan) X-Ray diffractometer (XRD) was used
for the structural characterization of the samples before and after the HC
attack in the CIDESI, Queretaro, Mexico facilities. Cu-kα radiation (λ =
1.5406 Å), Bragg Brentano geometry and measurement range from 20 to
80◦ with steps of 0.01◦ were the parameters used to collect the dif­
fractograms. Rietveld analysis was carried out with the GSAS software,
using Chebyshev polynomials with eight terms and the Pseudo-Voigt
and asymmetry function to fit the background and the peak profiles,
respectively, obtaining Chi2 values below 3.3. The following databases
were used for quantitative analyses:

2
J. de la Roche et al. Surface & Coatings Technology 432 (2022) 128066

Table 2 3. Results
Thickness of the layers of the TBC systems studied.
System NiCoCrAlY L-YSZ YSZ D-CZ Total ceramic 3.1. Microstructural characterization of D-CZ/YSZ systems
(μm) (μm) (μm) (μm) thickness (μm)

YSZ 120 ± 12 46 ± 6 277 ± – 324 ± 6 Table 2 shows the thickness of the APS-deposited YSZ and D-CZ
6 layers, which were intended to be as close as possible to the architec­
CZ50 210 ± 52 ± 7 309 ± 8 tures defined in Fig. 1. The thickness of the BC and L-YSZ layers was kept
12 constant for all TBC architectures and is shown in Table 2. The average
CZ100 151 99 ± 8 297 ± 6
porosity of YSZ layers of the different architectures was 15.9 ± 1.6% that
±
5
CZ150 101 ± 159 ± 9 307 ± 7 to the L-YSZ layers was 6.0 ± 1.3%. The coatings included globular pores
8 with sizes between 1 and 6 μm (fine), 6–16 μm (medium), and larger
than 16 μm (large) [43], and a high density of inter-splat and intra-splat
cracks, including high inter-connectivity between them, was observed,
• YSZ tetragonal phase with 8 mol% YO1.5 of yttria content: t′ -YSZ
as shown in Fig. 2.
(ICOD 01-082-1241)
The average porosity of the outer D-CZ layer of the three CYSZ sys­
• YSZ tetragonal phase with 4mol% YO1.5 of yttria content: t-YSZ
tems was 5.1 ± 1.2%, with a high density of fine globular pores and
(ICOD 01-078-1808)
intra-splat cracks. This layer also presented vertical cracks with a density
• CYSZ tetragonal phase: t′ -CYSZ (ICSD 98-008-6607)
of 4.3 cracks/mm for the CZ50 sample, 3.9 cracks/mm for the CZ100
• Zirconia monoclinic phase: m-ZrO2 (ICOD 01-078-0047)
sample, and 3.2 cracks/mm for the CZ150 sample. During the pre-
• Cerium oxide phase: CeO2 (ICSD 98-018-0851)
heating and spraying processes, heat transfer led to differential ther­
• Yttrium orthovanadate phase: YVO4 (ICOD 01-074-11-72)
mal expansion of the D-CZ and YSZ layers due to their dissimilar thermal
expansion coefficients [9,44]. This differential of thermal expansion
Structural analysis of the TC was performed using a μ-Raman spec­
leads to biaxial stresses in the D-CZ layer, which, together with the splat
troscope (Horiba LabRAM HR Evolution, Japan) equipped with an
boundaries, constitute preferential paths for vertical crack growth with
Olympus microscope using a 50× objective and a 532 nm Ar laser in the
further temperature changes [45,46].
facilities of the Universidad Nacional Medellín, Colombia. Spectra were
taken every 50 μm until reaching the BC range from 120 to 200 cm− 1
3.2. Hot corrosion resistance
with an acquisition time of 5 s.

3.2.1. Analysis of microstructural changes in the TBC systems

Fig. 3 compares the X-ray diffractograms of the multilayer systems
before and after the HC attack under different conditions.

Fig. 2. Microstructure of the cross-sectional views of TBC architectures: (a) YSZ, (b) CZ50, (c) CZ100, and (d) CZ150. The samples were labeled as a function of the
thickness increase of the D-CZ layer.

3
J. de la Roche et al.
Fig. 3. XRD results of the YSZ and D-CZ/YSZ systems after HC attack. (a) HC-Concentration tests, (b) HC-Cycles tests.
A reduction in the intensity of the (101) and (011) main peaks cor­
responding to the tetragonal phases (2θ = 30.1◦ for YSZ and 2θ = 29.9◦
for CYSZ), combined with an increase in the intensity of the YVO4, CeO2,
and m-ZrO2 peaks were observed after HC attack. In the HC-
concentration tests, an increase in salt concentration accelerated the
tetragonal-to-monoclinic phase transformation, as demonstrated by the
high intensity of the m-ZrO2 main peaks (2θ = 28◦ and 31.5◦ ), as well as
the emergence of new peaks in this phase at 2θ = 24◦ and 24.4◦ . These
transformations occurred because more salt was available to interact
with the stabilizing compounds (Y2O3 and CeO2), which increased the
4
Surface & Coatings Technology 432 (2022) 128066
amount of HC products as demonstrated by the higher intensity of the
YVO4 (2θ = 24.9◦ ) and the CeO2 (2θ = 28.5◦ ) peaks when the salt
concentration was increased.
HC-Cycles tests also showed a progressive decrease in the intensity of
peaks the YSZ and CYSZ tetragonal phases (Fig. 3b), combined with an
increase in the intensities of the peaks of the YVO4, CeO2, and m-ZrO2
phases as a function of the number of cycles. HC-Cycles tests showed
more extensive destabilization of the YSZ system than the CYSZ archi­
tectures after each cycle, indicating the higher resistance of the D-CZ
layer against the HC attack.
J. de la Roche et al.
Fig. 4. Amount of tetragonal phase in the different TBC systems after HC attack. HC-Concentration tests (left), and HC-Cycles tests (right).
CeO2 is considered a precipitation byproduct that results from the so-
called mineralization process described by Jones [14], according to
which exposure to molten salts contributes to the formation of the stable
phases predicted by the ZrO2-CeO2 equilibrium diagram at the tem­
perature of interest, i.e., monoclinic zirconia and free CeO2 particles for
the 700-900 ◦ C range. This destabilization of tetragonal zirconia was
previously reported by Park et al. [34] in CYSZ coatings exposed to a
V2O5-Na2SO4 salt mixture. Based on the experimental evidence
observed in this work and the existing literature [13,14,47–49], a
summary of the expected reactions occurring in zirconia in the presence
of molten vanadium-rich salts is as follows:
t′ − ZrO2 (Y2 O3 .CeO2 )(s)+V2 O5 (l)⇌m− ZrO2 (s)+2YVO4 (s)+CeO2 (s)
(4)
t′ − ZrO2 (Y2 O3 .CeO2 ) (s) + 2NaVO3 (l)⇌m − ZrO2 (s) + 2YVO4 (s)
(5)
+ CeO2 (s) + Na2 O (s)
Reidy and Jones reported that the formation of CeVO4 requires an
increase in the chemical activity of V2O5 in the melt for the chemical
reaction described by Eq. (6) to occur [47]. However, in this work, no
HC products associated with cerium including cerium orthovanadate
(CeVO4), were found after the XRD analyses of the samples studied.
Fig. 5. Amount of YVO4 and CeO2 phases after HC attack. (a) HC-Concentration tests, (b) HC-Cycles tests.
5
Surface & Coatings Technology 432 (2022) 128066
2CeO2 (s) + V2 O5 (l)⇌m − ZrO2 (s) + 2CeVO4 (s) + ½ O2 (g) (6)
Previous results by the authors [11] showed that in a Na2SO4-V2O5
salt mixture, the V2O5 activity is proportional to SO3 partial pressure,
which is low when the experiment is performed in air. Moreover, a
higher concentration of Na2SO4 decreases the melt's acidity. The high
acidity of ceria compared with yttria [13,48] favors the formation of
YVO4 in CYSZ systems, which generates a preference of reactions of the
molten salts with the co-stabilizing compound. Therefore, the HC attack
mechanism of the CYSZ layers is dominated by the mineralization (CeO2
precipitation) and the formation of YVO4.
Fig. 4 presents the tetragonal phase content after HC tests for the TBC
systems studied. A reduction in the content of the tetragonal phases (t′ -
YSZ, t-YSZ, and t′ -CYSZ) with the increase in salt concentration and
cycles is shown. In YSZ samples, the t′ -YSZ phase disappeared when the
salt concentration used was 5 wt% in HC-Concentration tests or after 4
cycles in the HC-cycle tests. The protection provided by the D-CZ layer is
evident since, for any testing condition, the amount of tetragonal phase
in CYSZ samples was always higher than in the YSZ samples.
Fig. 5 shows the results of phase quantification of the HC products.
The amount of HC products increased with salt concentration (Fig. 6a),
obtaining values of up to 50 wt% for the YSZ sample, consistent with the
results shown in Figs. 3 and 4. Fig. 5b shows that the amount of YVO4
J. de la Roche et al. Surface & Coatings Technology 432 (2022) 128066

Fig. 6. Morphological analysis of the samples after HC-Concentration tests. (a) YSZ sample tested with a salt concentration of 3 wt%, (b) CZ50 sample tested with a
salt concentration of 3 wt%, (c) YSZ sample tested with a salt concentration of 5 wt%, (b) CZ50 sample tested with a salt concentration of 5 wt%, (e) EDS spectra of
CYSZ system tested with a salt concentration of 5 wt%.

increased linearly with the number of cycles for the YSZ system, whereas
in the CYSZ systems, the CeO2 remained stable with the number of cycles
with a slight downward trend.
To summarize the above results, Table 3 shows the phase quantifi­
cation of the systems studied after the HC tests (HC-Concentration and
HC-Cycles tests).
3.2.2. Analysis of corroded surfaces and HC products
3.2.2.1. HC-Concentration tests. The surface of the YSZ and CYSZ TBC
architectures submitted to HC-concentration tests is presented in Fig. 6.
The YSZ system (Fig. 6a) exhibited an evident TC peel-off at the edges,
not observed in CYSZ systems (Figs. 6b). In the YSZ system, the HC
products showed rod-like and needle-like morphologies characteristic of
YVO4 crystals. On the other hand, the surface of CYSZ systems (Fig. 6b)
evidenced the presence of rods held together with cubic and cuboid
crystals and “free” cubic and semi-cubic crystals.
Fig. 6c and d show the surface view of the YSZ and CYSZ TBC ar­
chitectures tested with 5 wt% of salt concentration. It is important to
note that similar results were observed for all CYSZ architectures. The D-
CZ layer was delaminated above 20% in the CYSZ architectures, whereas
only coating spallation and delamination at the samples' edges were
observed in the YSZ architecture. This salt concentration produced se­
vere HC attack on the tested TBC architectures, generating failure at

6
J. de la Roche et al. Surface & Coatings Technology 432 (2022) 128066

Table 3 to analyze the coating's failure. Fig. 7 shows the surface of the D-CZ/YSZ
Results of phase quantifications after HC tests. interface and the corresponding EDS analyses.
Test Sample Phase amounts after the test (wt%) The analysis revealed the characteristic morphologies of the HC
products, i.e., cuboidal crystals (Fig. 7a) of the CeO2 phase and typical
HC-Concentration YSZ t′ -YSZ: 0
t-YSZ: 2.69 ± 1.17 rod-like morphology of the YVO4 phase (Fig. 7b). These results are
YVO4: 46.04 ± 3.63 consistent with the results of studies by Jones [14], Daroonpavar et al.
CY50 t′ -CYSZ:12.16 ± 7.77 [15], Park et al. [34], and Levi et al. [50], among others, in which the
YVO4: 33.71 ± 0.27 corrosion products formed at the coating's defects, decreasing the
CeO2: 20.39 ± 7.55
CZ100 t′ -CYSZ: 5.17 ± 1.17
compliance and generating mechanical stresses that promote cracking
YVO4: 33.31 ± 5.11 and delamination of the coating.
CeO2: 24.69 ± 6.14
CZ150 t′ -CYSZ: 8.60 ± 1.63 3.2.2.2. HC-Cycles tests. Fig. 8 shows the HC products at the surface of
YVO4: 44.72 ± 0.37
CeO2: 44.72 ± 0.37
the multilayer systems after the HC-cycles tests. No visible coating
HC-Cycles YSZ t′ -YSZ: 0 spallation was observed in the samples after 5 cycles, even at a salt
t-YSZ: 4.74 ± 2.48 concentration of 5 wt% and once the complete destabilization of the
YVO4: 29.48 ± 3.27 tetragonal phases had occurred. SEM images of the YSZ architecture
CZ50 t′ -CYSZ: 9.28 ± 1.83
revealed the presence of needle-type YVO4 crystals (Fig. 8a). A com­
YVO4:19.89 ± 3.75
CeO2: 11.51 ± 2.19 parison of the results with those corresponding to the HC-Concentration
CZ100 t′ -CYSZ: 8.56 ± 2.29 test (see Fig. 6), when the samples (cycles and concentration) were
YVO4: 29.66 ± 5.90 exposed to the same salt concentration (5 wt%), showed that YVO4
CeO2: 8.16 ± 1.96 crystals growing under cyclic conditions are more numerous and have a
CZ150 t′ -CYSZ: 8.14 ± 1.96
YVO4: 26.27 ± 3.97
larger length-to-width ratio than those growing under isothermal con­
CeO2: 7.31 ± 2.15 ditions. A similar situation was observed in CYSZ systems, in which the
HC products have needle-like morphology growing uniformly across the
entire surface (Fig. 8b) and some cubic and semi-cubic CeO2 crystals,
both interfaces and within the layers that comprise the coating system. confirming the results of the XRD analyses (Fig. 5b). These results sug­
The YSZ system had a YVO4 phase predominantly with rod-like gest that changing HC corrosion tests' parameters promotes different
morphology and needle-like particles (Fig. 6c). growth kinetics of the corrosion products, affecting their morphology
On the other hand, the HC products for all CYSZ systems (Fig. 6d) and size.
exhibited a rod-like morphology (YVO4 phase) mixed with cubic and
semi-cubic crystals (CeO2 phase) that were larger compared with the 3.2.3. Analysis of L-YSZ/YSZ interface
samples tested at 3 wt% salt concentration, which correlates well with Fig. 9 shows the BC/YSZ interface after HC-concentration tests with
the trend observed after the quantitative Rietveld analysis presented in 5 wt% salt concentration (Fig. 9a) and five HC-cycles tests (Fig. 9b). In
Fig. 5a. The elemental analysis shown in Fig. 6e confirmed the presence both HC tests, the infiltration of the molten salt through the TC was
of V and Y in the rod-like particles and Ce and O2 in the cubic and semi- evident. This accelerated BC degradation in different areas, resulting in
cubic crystals. The back-side of the spalled D-CZ layers was also studied heterogeneous growth of the TGO along the interface. It also resulted in

Fig. 7. Hot-corrosion products found in the back-side of the detached D-CZ layer. (a) CeO2, (b) YVO4.

7
J. de la Roche et al. Surface & Coatings Technology 432 (2022) 128066

Fig. 8. Surface characterization of bilayer systems after HC-Cycles test (a) YSZ sample, (b) CZ100 sample, (c) CZ100 sample with higher magnification.

Fig. 9. Molten salt infiltration in CYSZ systems during HC tests. (a) HC-concentration test, 5 wt% salt concentration (b) HC-Cycles test, 5 HC cycles.

8
J. de la Roche et al.
Fig. 10. YSZ, CZ100, and CZ150 systems after HC-cycles tests. (a) Cross-section LOM, (b) Raman spectra of the cross-section of YSZ, CYSZ100, and CYSZ150 systems
after 5 cycles.
Al depletion in the BC at the BC/YSZ interface and, therefore, reduced
the amount of β precipitates when a thicker TGO was formed. The EDS
spectra show the presence of vanadium in the TGO, which promotes
coating failure. This result confirms that the D-CZ layer did not offer
effective protection against molten salt penetration under the aggressive
HC test conditions studied in this work due to the presence of vertical
cracks that allow the penetration of molten salts.
4. Discussion
The results presented above demonstrate that 5 wt% salt concen­
tration (HC-Concentration test) and 5 cycles (HC-Cycles) were the most
aggressive HC conditions tested in this work, as the X-ray diffractograms
showed greater phase destabilization of the tetragonal phases (t′ -YSZ, t-
YSZ, and t′ -CYSZ). These test conditions also generated more HC
9
Surface & Coatings Technology 432 (2022) 128066
products, as confirmed by the phase quantification results and the SEM
examination of the corroded surfaces. However, after 5 cycles, no
coating detachment was observed in the CYSZ systems, as shown in
Fig. 10a. After the HC-Cycles test, the microstructure observed in the
bilayer systems revealed areas where the YSZ system exhibited severe
cracking without delamination or spallation and reduced porosity
throughout the thickness, which was also presented in the YSZ coatings
of CYSZ systems. It was found that the vertical cracks in D-CZ layers
were infiltrated with products of the corrosion reactions without
delamination. The change in the bilayer microstructure is produced by
the fluxing (dissolution) of the coating due to its reaction with the salts
[12] and solidification of the molten salts, and formation of HC products
in the coating defects [15].
The μ-Raman analysis of the YSZ system in the cross-section shows
the characteristic bands of monoclinic and yttrium orthovanadate
J. de la Roche et al.
Fig. 11. (a) Cross-section of YSZ and CYSZ100 systems after HC test, (b) Evidence of impregnation of vertical cracks after HC test in CYSZ systems.
phases at the surface and throughout the thickness of the coating
(Fig. 10b). This analysis is consistent with the X-ray diffraction results,
where the YSZ exhibited an amount of the tetragonal phase of less than
10% (Fig. 4). The intensities decrease close to the BC, where the vi­
bration at 148 cm− 1 of the tetragonal phase is observed [51,52] due to
the higher density of the L-YSZ layer. Despite the infiltration of salts, the
dissolution reactions were less aggressive in such a dense layer. The
CYSZ systems presented higher phase destabilization at the surface due
to the dissolution produced by the molten salts (as inferred from the XRD
results). However, no phase transformation was observed inside the D-
CZ layer in any samples studied since they presented peaks corre­
sponding only to the tetragonal phase (t′ -CYSZ). Moreover, the YSZ
layers of the CYSZ systems did not present peaks corresponding to HC
products at 158 cm− 1 [51] as observed in the YSZ system, and CZ100
and CZ150 systems showed vibrations in the tetragonal phase (t′ -YSZ),
indicating that the D-CZ layers offered protection against attack by
molten salts despite the infiltration of molten salt, because the dissolu­
tion and formation of HC products were less aggressive.
Cross-sections shown in Fig. 11a confirm the occurrence of delami­
nation and spallation at the surface of the systems tested under 5 wt%
HC concentration (as observed in surface photos of Fig. 6c and d). All
samples showed several microstructure changes, even in areas where the
TBC system did not fail. Delamination in the YSZ system was observed
near the L-YSZ/YSZ interface. On the contrary, the bilayer systems
presented delamination in the D-CZ layer and infiltration of vertical
cracks with HC products, as shown in Fig. 11b. The EDS spectrum
confirmed the presence of Y and V in the gray areas within the vertical
cracks, indicating a preference to form YVO4 crystals in the D-CZ layer.
These cracks allowed the penetration of the molten salts through the
coatings.
The HC-Cycles test was less aggressive than the HC-concentration
test for the same salt concentration. The “stagnation effect” of the HC
10
Surface & Coatings Technology 432 (2022) 128066
attack in the TBC system could be responsible for this behavior, given
that similar results have been observed in HC tests in NiCoCrAlY alloys
[53]. Replenishing with fresh salt every cycle during the test produces
changes in the molten salts' surface and chemistry deposited in the
coating, possibly decreasing the HC reaction's kinetics. This effect was
demostrated by the difference in the morphology of the YVO4 crystals
(Figs. 7 and 9), which are highly susceptible to the precursors' compo­
sition [54]. Moreover, the infiltration of cracks and pores with the early
HC products (with high melting points) slows down the intake of new
molten salt in the subsequent cycles, avoiding new reactions and
generating more HC products within the coating.
5. Conclusions
This work showed that for CYSZ/YSZ TBC systems, the predominant
HC mechanisms in the presence of combustion gases produced with low-
quality liquid fuel are dissolution (fluxing) and mineralization, which
produce YVO4 and CeO2 crystals together with the stabilization of
monoclinic zirconia. A CYSZ layer with thickness between 100 and 150
μm applied onto a YSZ TBC system acts as a barrier to molten salts'
attack, increasing the HC resistance of the TBC system. However, ver­
tical cracks formed during the coating deposition process reduce the HC
resistance of the TBC system.
HC test results showed that tetragonal-to-monoclinic phase trans­
formation and the formation of HC products at interfaces and defects
have a high influence on the spallation and delamination of the coatings
of the TBC system. These results were verified experimentally with the
X-ray diffraction and Raman spectroscopy results and the examination
of the corrosion products found in the detached layers' back-side.
The choice of HC test has relevant effects on the magnitude of
damage in the TBC system, presenting different mechanisms of coating's
degradation. Isothermal exposure to a fixed amount of salt (HC-
J. de la Roche et al.
Concentration tests) is more aggressive than the progressive addition of
the salt (HC-Cycles tests) because the latter does not allow the system to
stabilize, retarding the reaction kinetics.
CRediT authorship contribution statement
Hot corrosion behavior of dense CYSZ/YSZ bilayer coatings depos­
ited by atmospheric plasma spray in Na2SO4 + V2O5 molten salts was a
result of the doctoral thesis of the Jhonattan de la Roche, having a
collaboration of the industry (Empresas Publicas de Medellín), CIDESI in
Queretaro, Mexico, and CPT in Barcelona, Spain.
The contribution of the authors in the manuscript is shown below:
• Jhonattan de la Roche: Conceptualization, Methodology, Formal
analysis, Investigation, Writing - original draft,
• Juan Manuel Alvarado - Orozco: Conceptualization, Methodology,
Resources, Writing - review & editing, Visualization,
• Pablo Andrés Gómez: Resources, Writing - review & editing, Su­
pervision, Funding acquisition.
• Irene García Cano: Resources, Writing - review & editing,
Supervision.
• Sergi Dosta: Resources, Writing - review & editing, Supervision.
• Alejandro Toro: Conceptualization, Methodology, Resources,
Writing - review & editing, Project administration, Funding
acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgments
The authors express gratitude to Empresas Públicas de Medellín-EPM
to the financial support. The authors also thank CONACYT (FORDECYT
277265 project), CONMAD Consortium, and Thermal Spray Center-CPT
of the University of Barcelona for technical and financial support
(2017SGR 01777 project).
References
[1] N.P. Padture, M. Gell, E.H. Jordan, Thermal barrier coatings for gas-turbine engine
applications, Science 296 (2002) 280–284, https://doi.org/10.1126/
science.1068609.
[2] D.R. Clarke, S.R. Phillpot, Thermal barrier coating materials, Mater. Today 8
(2005) 22–29, https://doi.org/10.1016/S1369-7021(05)70934-2.
[3] C.U. Hardwicke, Y.C. Lau, Advances in thermal spray coatings for gas turbines and
energy generation: a review, J. Therm. Spray Technol. 22 (2013) 564–576, https://
doi.org/10.1007/s11666-013-9904-0.
[4] D.R. Clarke, M. Oechsner, N.P. Padture, Thermal-barrier coatings for more efficient
gas-turbine engines, MRS Bull. 37 (2012) 891–898, https://doi.org/10.1557/
mrs.2012.232.
[5] N.P. Padture, Advanced structural ceramics in aerospace propulsion, Nat. Mater.
15 (2016) 804–809, https://doi.org/10.1038/nmat4687.
[6] R. Darolia, Thermal barrier coatings technology: critical review, progress update,
remaining challenges and prospects, Int. Mater. Rev. 58 (2013) 315–348, https://
doi.org/10.1179/1743280413Y.0000000019.
[7] L.B. Chen, Yttria-stabilized zirconia thermal barrier coatings — a review, Surf. Rev.
Lett. 13 (2006) 535–544, https://doi.org/10.1142/S0218625X06008670.
[8] A.G. Evans, D.R. Clarke, C.G. Levi, The influence of oxides on the performance of
advanced gas turbines, J. Eur. Ceram. Soc. 28 (2008) 1405–1419, https://doi.org/
10.1016/j.jeurceramsoc.2007.12.023.
[9] X.Q. Cao, R. Vassen, D. Stoever, Ceramic materials for thermal barrier coatings,
J. Eur. Ceram. Soc. 24 (2004) 1–10, https://doi.org/10.1016/S0955-2219(03)
00129-8.
[10] D. Song, T. Song, U. Paik, G. Lyu, Y.G. Jung, Hot corrosion behavior in thermal
barrier coatings with heterogeneous splat boundary, Corros. Sci. 163 (2020),
108225, https://doi.org/10.1016/j.corsci.2019.108225.
[11] J. De la Roche, J.M. Alvarado-Orozco, A. Toro, Hot corrosion mechanism of yttria-
stabilized zirconia powder in the presence of molten Na2SO4 + V2O5 salts, Rare
Metals (2020), https://doi.org/10.1007/s12598-020-01388-3.
11
Surface & Coatings Technology 432 (2022) 128066
[12] N. Reddy, A.S. Gandhi, Molten salt attack on t’ yttria-stabilised zirconia by
dissolution and precipitation, J. Eur. Ceram. Soc. 33 (2013) 1867–1874, https://
doi.org/10.1016/j.jeurceramsoc.2013.02.022.
[13] R.L. Jones, C.E. Williams, S.R. Jones, Reaction of vanadium compounds with
ceramic oxides, J. Electrochem. Soc. 133 (1986) 227–230, https://doi.org/
10.1149/1.2108530.
[14] R.L. Jones, Some aspects of the hot corrosion of thermal barrier coatings, J. Therm.
Spray Technol. 6 (1997) 77–84, https://doi.org/10.1007/BF02646315.
[15] M. Daroonparvar, M.A.M. Yajid, N.M. Yusof, H.R. Bakhsheshi-Rad, E. Hamzah,
M. Nazoktabar, Investigation of three steps of hot corrosion process in Y2O3
stabilized ZrO2 coatings including nano zones, J. Rare Earths 32 (2014) 989–1002,
https://doi.org/10.1016/S1002-0721(14)60173-3.
[16] C. Chen, T. Liang, Y. Guo, X. Chen, Q. Man, X. Zhang, J. Zeng, V. Ji, Effect of
Scandia content on the hot corrosion behavior of Sc2O3 and Y2O3 co-doped ZrO2
in Na2SO4 + V2O5 molten salts at 1000 ◦ C, Corros. Sci. 158 (2019), https://doi.
org/10.1016/j.corsci.2019.108094.
[17] T. Liu, S.W. Yao, L.S. Wang, G.J. Yang, C.X. Li, C.J. Li, Plasma-sprayed thermal
barrier coatings with enhanced splat bonding for CMAS and corrosion protection,
J. Therm. Spray Technol. 25 (2016) 213–221, https://doi.org/10.1007/s11666-
015-0345-9.
[18] P. Carpio, A. Borrell, M.D. Salvador, A. Gómez, E. Martínez, E. Sánchez,
Microstructure and mechanical properties of plasma spraying coatings from YSZ
feedstocks comprising nano- and submicron-sized particles, Ceram. Int. 41 (2015)
4108–4117, https://doi.org/10.1016/j.ceramint.2014.11.106.
[19] M. Mohammadi, A. Kobayashi, S. Javadpour, S.A.J. Jahromi, Evaluation of hot
corrosion behaviors of Al2O3-YSZ composite TBC on gradient MCrAlY coatings in
the presence of Na2SO4-NaVO3 salt, Vacuum 167 (2019) 547–553.
[20] D. Song, T. Song, U. Paik, G. Lyu, J. Kim, Y.-G. Jung, Improvement in hot corrosion
resistance and chemical stability of YSZ by introducing a Lewis neutral layer on
thermal barrier coatings, Corros. Sci. (2020), 108776, https://doi.org/10.1016/j.
corsci.2020.108776.
[21] L. Guo, H. Xin, Z. Zhang, F. Ye, Z. Yan, Preparation of (Gd0.9Sc0.1)2Zr2O7/YSZ
thermal barrier coatings and their corrosion resistance to V2O5 molten salt, Surf.
Coat. Technol. 389 (2020), 125677, https://doi.org/10.1016/j.
surfcoat.2020.125677.
[22] C. Batista, A. Portinha, R.M. Ribeiro, V. Teixeira, C.R. Oliveira, Evaluation of laser-
glazed plasma-sprayed thermal barrier coatings under high temperature exposure
to molten salts, Surf. Coatings Technol. 200 (2006) 6783–6791, https://doi.org/
10.1016/j.surfcoat.2005.10.011.
[23] A.H. Pakseresht, A. Kimiayi, M. Alizadeh, H. Nuranian, A. Faeghinia,
Microstructural study and hot corrosion behavior of bimodal thermal barrier
coatings under laser heat treatment, Ceram. Int. 46 (2020) 19217–19227, https://
doi.org/10.1016/j.ceramint.2020.04.259.
[24] L. Guo, H. Xin, Z. Zhang, X. Zhang, F. Ye, Microstructure modification of Y2O3
stabilized ZrO2 thermal barrier coatings by laser glazing and the effects on the hot
corrosion resistance, J. Adv. Ceram. 9 (2020) 232–242, https://doi.org/10.1007/
s40145-020-0363-z.
[25] S.M. Naga, M. Awaad, H.F. El-Maghraby, A.M. Hassan, M. Elhoriny, A. Killinger,
R. Gadow, Effect of La2Zr2O7 coat on the hot corrosion of multi-layer thermal
barrier coatings, Mater. Des. 102 (2016) 1–7, https://doi.org/10.1016/j.
matdes.2016.03.133.
[26] L. Guo, M. Li, F. Ye, Comparison of hot corrosion resistance of Sm2Zr2O7 and
(Sm0.5Sc0.5)2Zr2O7 ceramics in Na2SO4+V2O5 molten salt, Ceram. Int. 42
(2016) 13849–13854, https://doi.org/10.1016/j.ceramint.2016.05.190.
[27] Y. Ozgurluk, K.M. Doleker, A.C. Karaoglanli, Hot corrosion behavior of YSZ,
Gd2Zr2O7 and YSZ/Gd2Zr2O7 thermal barrier coatings exposed to molten sulfate
and vanadate salt, Appl. Surf. Sci. 438 (2018) 96–113, https://doi.org/10.1016/j.
apsusc.2017.09.047.
[28] M. Bahamirian, S.M.M. Hadavi, M. Farvizi, M.R. Rahimipour, A. Keyvani,
Enhancement of hot corrosion resistance of thermal barrier coatings by using
nanostructured Gd2Zr2O7 coating, Surf. Coat. Technol. 360 (2019) 1–12, https://
doi.org/10.1016/j.surfcoat.2018.12.113.
[29] M.H. Habibi, L. Wang, S.M. Guo, Evolution of hot corrosion resistance of YSZ,
Gd2Zr2O7, and Gd2Zr2O7+YSZ composite thermal barrier coatings in Na2SO4+
V2O 5 at 1050??C, J. Eur. Ceram. Soc. 32 (2012) 1635–1642, https://doi.org/
10.1016/j.jeurceramsoc.2012.01.006.
[30] M.H. Habibi, S. Yang, S.M. Guo, Phase stability and hot corrosion behavior of
ZrO2-Ta2O5 compound in Na2SO4-V2O5 mixtures at elevated temperatures,
Ceram. Int. 40 (2014) 4077–4083, https://doi.org/10.1016/j.
ceramint.2013.08.062.
[31] M.H. Habibi, L. Wang, J. Liang, S.M. Guo, An investigation on hot corrosion
behavior of YSZ-Ta2O5 in Na2SO4+V2O5 salt at 1100◦ C, Corros. Sci. 75 (2013)
409–414, https://doi.org/10.1016/j.corsci.2013.06.025.
[32] R.L. Jones, D. Mess, India as a hot corrosion-resistant stabilizer for zirconia, J. Am.
Ceram. Soc. 75 (1992) 1818–1821, https://doi.org/10.1111/j.1151-2916.1992.
tb07202.x.
[33] H.Fei Liu, X. Xiong, X. bin Li, Y.Lei Wang, Hot corrosion behavior of Sc2O3-Y2O3-
ZrO2thermal barrier coatings in presence of Na2SO4+V2O5molten salt, Corros.
Sci. 85 (2014) 87–93, https://doi.org/10.1016/j.corsci.2014.04.001.
[34] S.Y. Park, J.H. Kim, M.C. Kim, H.S. Song, C.G. Park, Microscopic observation of
degradation behavior in yttria and ceria stabilized zirconia thermal barrier
coatings under hot corrosion, Surf. Coat. Technol. 190 (2005) 357–365, https://
doi.org/10.1016/j.surfcoat.2004.04.065.
[35] M.H. Habibi, S.M. Guo, The hot corrosion behavior of plasma sprayed zirconia
coatings stabilized with yttria, ceria, and titania in sodium sulfate and vanadium
oxide, Mater. Corros. 66 (2015) 270–277.
J. de la Roche et al.
[36] M. Hajizadeh-oghaz, R.S. Razavi, Z. Valefi, Na2SO4 and V2O5 molten salts
corrosion resistance of plasma-sprayed nanostructured ceria and yttria co-
stabilized zirconia thermal barrier coatings, Ceram. Int. 42 (2016) 5433–5446,
https://doi.org/10.1016/j.ceramint.2015.12.085.
[37] H. Edris, R.S. Razavi, A. Ghasemi, Z. Valefi, H. Jamali, Plasma-sprayed
nanostructured Scandia-yttria and ceria-yttria codoped zirconia coatings:
microstructure, bonding strength and thermal insulation properties, Ceram. Int. 44
(2018) 12042–12047, https://doi.org/10.1016/j.ceramint.2018.03.204.
[38] B. Mukherjee, A. Islam, K.K. Pandey, O.S. Asiq Rahman, R. Kumar, A.K. Keshri,
Impermeable CeO2 overlay for the protection of plasma sprayed YSZ thermal
barrier coating from molten sulfate-vanadate salts, Surf. Coat. Technol. 358 (2019)
235–246, https://doi.org/10.1016/j.surfcoat.2018.11.048.
[39] Y.S. Hwang, R.A. Rapp, Thermochemistry and solubilities of oxides in sodium
sulfate-vanadate solutions, Corrosion 45 (1989) 933–937, https://doi.org/
10.5006/1.3585003.
[40] V. Viswanathan, G. Dwivedi, S. Sampath, Multilayer, multimaterial thermal barrier
coating systems: design, synthesis, and performance assessment, J. Am. Ceram.
Soc. 98 (2015) 1769–1777, https://doi.org/10.1111/jace.13563.
[41] V. Viswanathan, G. Dwivedi, S. Sampath, Engineered multilayer thermal barrier
coatings for enhanced durability and functional performance, J. Am. Ceram. Soc.
97 (2014) 2770–2778, https://doi.org/10.1111/jace.13033.
[42] J. de la Roche, P.A. Gómez, J.M. Alvarado-Orozco, A. Toro, Hot corrosion and
thermal shock resistance of dense-CYSZ/YSZ bilayer thermal barrier coatings
systems applied onto Ni-base superalloy, J. Eur. Ceram. Soc. (2020) 1, https://doi.
org/10.1016/j.jeurceramsoc.2020.07.004.
[43] L. Tobón, C. Barrios, D. Zambrano, A. Toro, Análisis morfológico de la porosidad de
un sistema de barrera térmica sometido a cargas térmicas constantes, Rev. Colomb.
Mater. (2014) 35–41.
[44] A. Kulkarni, A. Vaidya, A. Goland, S. Sampath, H. Herman, Processing effects on
porosity-property correlations in plasma sprayed yttria-stabilized zirconia
coatings, Mater. Sci. Eng. A 359 (2003) 100–111.
12
Surface & Coatings Technology 432 (2022) 128066
[45] M. Karger, R. Vaßen, D. Stöver, Atmospheric plasma sprayed thermal barrier
coatings with high segmentation crack densities: spraying process, microstructure
and thermal cycling behavior, Surf. Coat. Technol. 206 (2011) 16–23, https://doi.
org/10.1016/j.surfcoat.2011.06.032.
[46] M. Madhwal, E.H. Jordan, M. Gell, Failure mechanisms of dense vertically-cracked
thermal barrier coatings, Mater. Sci. Eng. A 384 (2004) 151–161.
[47] R.F. Reidy, R.L. Jones, Thermogravimetric analysis of the reaction of CeO2 with
the NaVO3 - SO3 system, J. Electrochem. Soc. 142 (1995) 1353–1357, https://doi.
org/10.1149/1.2044177.
[48] Y.S. Zhang, R.A. Rapp, Solubilities of CeO2, HfO2 and Y2O3 in fused Na2SO4•30
mol% NaVO3 and Ce02 in pure Na2SO4 at 900 ◦ C, Corrosion 43 (1987) 348–352,
https://doi.org/10.5006/1.3583869.
[49] R.L. Jones, C.E. Williams, Ceramic Oxide Reactions With V2O5 and SO3, 1985.
[50] C.G. Levi, J.W. Hutchinson, M.-H. Vidal-Sétif, C.A. Johnson, Environmental
degradation of thermal-barrier coatings by molten deposits, MRS Bull. 37 (2012)
932–941.
[51] M.C. Mayoral, J.M. Andrés, M.T. Bona, V. Higuera, F.J. Belzunce, Yttria stabilized
zirconia corrosion destabilization followed by Raman mapping, Surf. Coat.
Technol. 202 (2008) 5210–5216.
[52] V. Lughi, D.R. Clarke, Transformation of electron-beam physical vapor-deposited 8
wt% yttria-stabilized zirconia thermal barrier coatings, J. Am. Ceram. Soc. 88
(2005) 2552–2558, https://doi.org/10.1111/j.1551-2916.2005.00452.x.
[53] C. Leyens, I.G. Wright, B.A. Pint, Effect of experimental procedures on the cyclic,
hot-corrosion behavior of NiCoCrAlY-type bondcoat alloys, Oxid. Met. 54 (2000)
255–276, https://doi.org/10.1023/a:1004698211940.
[54] S. Erdei, F.W. Ainger, Crystal growth of YVO4 using the LHPG technique, J. Cryst.
Growth 128 (1993) 1025–1030, https://doi.org/10.1016/S0022-0248(07)80091-
1.