Journal of The Electrochemical
Society
     
OPEN ACCESS
Nanostructured Coating for Aluminum Alloys Used in Aerospace
Applications
To cite this article: Maido Merisalu et al 2022 J. Electrochem. Soc. 169 071503
 
View the article online for updates and enhancements.
This content was downloaded from IP address 85.254.75.54 on 24/08/2022 at 16:26
Nanostructured Coating for Aluminum Alloys Used in Aerospace
Applications
Maido Merisalu,1 Lauri Aarik,1 Helle-Mai Piirsoo,1 Jekaterina Kozlova,1 Aivar Tarre,1
Roberts Zabels,2 Johanna Wessing,3 Abel Brieva,3 and Väino Sammelselg1,z
1University of Tartu, Institute of Physics, 50411 Tartu, Estonia
2University of Latvia, Institute of Solid State Physics, LV-1063 Riga, Latvia
3European Space Agency, European Space Research and Technology Centre, 2201 AZ Noordwijk, Netherlands
A thin industrial corrosion-protection nanostructured coating for the Al alloy AA2024-T3 is demonstrated. The coating is prepared in a
two-step process utilizing hard anodizing as a pre-treatment, followed by sealing and coating by atomic layer deposition (ALD). In the
first step, anodizing in sulfuric acid at a low temperature converts the alloy surface into a low-porosity anodic oxide. In the second step,
the pores are sealed and coated by low-temperature ALD using different metal oxides. The resulting nanostructured ceramic coatings
are thoroughly characterized by cross-sectioning using a focused ion beam, followed by scanning electron microscopy, transmission
electron microscopy, X-ray microanalysis, and nanoindentation and are tested via linear sweep voltammetry, electrochemical
impedance spectroscopy, salt spray, and energetic atomic oxygen flow. The best thin corrosion protection coating, made by anodizing
at 20 V, 1 °C and sealing and coating with amorphous Al2O3/TiO2 nanolaminate, exhibits no signs of corrosion after a 1000 h ISO
9227 salt spray test and demonstrates a maximum surface hardness of 5.5 GPa. The same coating also suffers negligible damage in an
atomic oxygen test, which is comparable to 1 year of exposure to space in low Earth orbit.
© 2022 The Author(s). Published on behalf of The Electrochemical Society by IOP Publishing Limited. This is an open access
article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/
by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/
1945-7111/ac7bb2]
Manuscript submitted April 30, 2022; revised manuscript received June 22, 2022. Published July 1, 2022.
Supplementary material for this article is available online
Aluminum alloys of the AA2000 series, such as AA2024-T3, are
favored for automobile, drone, and aerospace applications because
of their light weight, high strength, and adequate workability.1–5 The
latter alloy has an elemental composition, phase structure, and
vulnerability to localized pitting corrosion that are very similar to
those of AA2219-T851, which is used for structural components of
the International Space Station, outer-space rovers, and satellites.6,7
The use of lightweight materials for the manufacture of structural
parts for the applications mentioned above minimizes the consump-
tion of fuel or energy stored in batteries, which expands the
operation time of vehicles, reduces the carbon footprint from the
use of fossil-based fuels, and minimizes the emission of greenhouse
gases. The latter applies to both fossil fuels and hydrogen, which
contribute to global warming in different ways. For instance, CO2
and water vapor are both potent greenhouse gases (GHGs) and are
formed from the combustion of fossil fuels and hydrogen, respec-
tively. Furthermore, hydrogen is mostly produced via steam methane
reforming, which is cost-effective but generates significant CO2
emissions.8 Aluminum alloys are also widely used for satellite
construction for several reasons. The light weight of the alloys
minimizes the total mass of the satellite and, therefore, the cost of
shipping it into orbit.9 The low atomic number of aluminum makes it
also a good radiation shielding material because its interaction with
high-energy cosmic radiation results in less harmful secondary
electromagnetic radiation than higher atomic number materials
such as steels.10 Furthermore, aluminum alloys exhibit excellent
electrical conductivity, which is useful for thermal regulation, as
well as for mitigating the charging of surfaces due to cosmic
radiation.
The benefits of using aluminum alloys are clear; however, these
metals suffer from distinct problems in their application environ-
ments. For instance, the presence of humidity and chloride ions in
terrestrial applications and even aboard space stations may cause
galvanic corrosion, which is promoted by other passive metals in
electrical contact with the primary aluminum component as well as
by different types of near-surface intermetallic particles.11–14 In
space applications, however, connected metal parts may get stuck
owing to cold welding, which is particularly problematic for
components that need to retain their mobility (e.g., mechanical
scanning mechanisms).15 The problem of cold welding may be
further enhanced by fretting, which is caused by vibration during
launch or by movement of antennas in space. A possible example of
this phenomenon is a bolt that became stuck on the Hubble Space
Telescope, hindering the maintenance of the expensive instrument.16
However, in this case, corrosion caused by atomic oxygen17,18 may
have also played a role, causing corrosion products to build up
around the threads of the bolt over time.
The aforementioned problems with aluminum alloys can be
mitigated with different types of functional coatings depending on
the dimensional tolerance, shape, and purpose of the component. In
particular, for coating substrates with sophisticated three-dimensional
shapes, there are currently three distinct methods that cover different
thickness ranges: anodizing,19,20 plasma electrolytic oxidation
(PEO),15,21 and atomic layer deposition.22–28 During anodization,
the top layer of the aluminum alloy is electrochemically converted
into porous anodic aluminum oxide (AAO), which normally has a
thickness ranging from tens of micrometers to several millimeters.19,20
After anodizing, the AAO layer needs to be sealed to gain improved
resistance against corrosion and wear or to alter other properties of the
surface, such as its optical properties. The PEO process is performed
in mildly alkaline electrolytes and at higher potentials than anodizing,
which causes discharge at the alloy-electrolyte interface.15,21 The high
temperature at discharge causes the obtained oxide layer to be partially
crystalline, which increases its overall hardness. Optimized PEO
processes can therefore be used to create oxide layers that are tens of
microns thick and outperform the coatings obtained via anodization,
particularly at the lower end of the thickness range. Overall, both
anodizing and PEO can be used to protect aluminum components
against wear and corrosion, but also to prevent cold welding of parts
in space. In contrast to anodizing and PEO, ALD allows the
preparation of the thinnest protective coatings, which are as a rule
⩽100 nm thick, see overviews in earlier works.23,27 Unfortunately,
these coatings provide only limited corrosion protection.23,24 Even the
best performing nanolaminates containing Al2O3 and TiO2 layers, and
being deposited directly onto the alloy surface, fail at characteristic
sites, where the failure mechanism is most likely related to near-
surface intermetallic particles.12,27 In our previous studies, we also
showed that electrochemical pre-treatments enhance the performance
of coatings made by ALD and potentially even allow the preparationzE-mail: vaino.sammelselg@ut.ee
Journal of The Electrochemical Society, 2022 169 071503
of a new type of high-performance nanostructured coating by sealing
the AAO layer with different ceramic materials.25,27,28 If the approach
can be scaled up in a practical manner, then such nanostructured
coatings may be able to occupy the 0.1–5 μm thickness range, which
is particularly lucrative for precision components in high-performance
applications, where excellent resistance to corrosion and wear are
required.
The aforementioned problems with aluminum alloys can be
mitigated with different types of functional coatings depending on
the dimensional tolerance, shape, and purpose of the component. In
particular, for coating substrates with sophisticated three-dimen-
sional shapes, there are currently three distinct methods that cover
different thickness ranges: anodizing,19,20 plasma electrolytic oxida-
tion (PEO),15,21 and atomic layer deposition.22–28 During anodiza-
tion, the top layer of the aluminum alloy is electrochemically
converted into porous anodic aluminum oxide, which normally has
a thickness ranging from tens of micrometers to several
millimeters.19,20 After anodizing, the AAO layer needs to be sealed
to gain improved resistance against corrosion and wear or to alter
other properties of the surface, such as its optical properties. The
PEO process is performed in mildly alkaline electrolytes and at
higher potentials than anodizing, which causes discharge at the
alloy-electrolyte interface.15,21 The high temperature at discharge
causes the obtained oxide layer to be partially crystalline, which
increases its overall hardness. Optimized PEO processes can there-
fore be used to create oxide layers that are tens of microns thick and
outperform the coatings obtained via anodization, particularly at the
lower end of the thickness range. Overall, both anodizing and PEO
can be used to protect aluminum components against wear and
corrosion, but also to prevent cold welding of parts in space. In
contrast to anodizing and PEO, ALD allows the preparation of the
thinnest protective coatings, which are as a rule ⩽100 nm thick, see
overviews in earlier works.23,27 Unfortunately, these coatings
provide only limited corrosion protection.23,24 Even the best
performing nanolaminates containing Al2O3 and TiO2 layers, and
being deposited directly onto the alloy surface, fail at characteristic
sites, where the failure mechanism is most likely related to near-
surface intermetallic particles.12,27 In our previous studies, we also
showed that electrochemical pre-treatments enhance the perfor-
mance of coatings made by ALD and potentially even allow the
preparation of a new type of high-performance nanostructured
coating by sealing the AAO layer with different ceramic
materials.25,27,28 If the approach can be scaled up in a practical
manner, then such nanostructured coatings may be able to occupy
the 0.1–5 μm thickness range, which is particularly lucrative for
precision components in high-performance applications, where
excellent resistance to corrosion and wear are required.
Experimental
Pre-treatment of the substrates.—AA2024-T3 (Al alloy) alclad
plates (Goodfellow) with dimensions of 20 × 20 mm2 (smaller) and
40 × 110 mm2 (larger) samples were used in most of the
experiments, with the Al cladding mechanically removed from the
original 3 mm sheet by machine milling. The thickness of the plates
after milling was 2.6 ± 0.2 mm. The standard composition of the
alloy5 is listed in Table SI of the Supplementary Material. A
reproducible clean alloy surface was achieved by polishing with
abrasive paper (P240, Al2O3 grains, Mirka). To study the scalability
of the technology for preparation of nanostructured coatings for
industrial applications on arbitrarily shaped high-precision compo-
nents, Al alloy substrates were prepared for satellites ESTCube-2
and WISA Woodsat. On the former satellite, a cover panel was
prepared by machine milling from an Al–Mg alloy sheet containing
2.9% Mg. The components for WISA Woodsat were 3D printed with
an industrial metal 3D printer EOS M290 (EOS GmbH) from Al
alloy powder, EOS Aluminum AlSi10Mg (EOS GmbH)). Prior to
coating, all loose particles and organic contamination were removed
via a standard pre-treatment, by first rinsing with deionized water
and then by three solvents in an ultrasonic bath: 3 min in toluene
(purity 99.5%, Reahim), 3 min in acetone (purity 99.5%, Sigma-
Aldrich), and finally 3 min in isopropanol (purity 99.5%, Sigma-
Aldrich). The 3D printed substrates received an additional chemical
treatment for removing impurities and near-surface metal oxides,
which consisted of a) immersion in 10% NaOH solution (purity
⩾98%, Sigma-Aldrich) for 1 min, b) rinsing with deionized water, c)
immersion in concentrated HNO3 (purity 65%, Sigma-Aldrich) for
1 min, and d) rinsing with deionized water. The latter treatment was
necessary because mechanical polishing could not be used for the
high-precision components.
Anodizing of the substrates.—The anodizing process was carried
out in a two-electrode setup with a PS 8360-10 DT 1 kW power
supply (Elektro-Automatick GmbH), using a stainless-steel bath as
the cathode and the Al alloy substrates as the anode. A 15% sulfuric
acid solution was used as the anodizing electrolyte, and its
temperature was maintained at 1 °C ± 0.5 °C by surrounding the
stainless-steel anodizing bath with an additional ice bath. A 15%
sulfuric acid solution was prepared from concentrated sulfuric acid
(SA; 95%–97%, Honeywell, Fluka). For the potentiostatic anodizing
process, the voltage was set at 10 V for thinner and 20 V for thicker
AAO, and the maximum current density at the beginning of both
processes was set at 12.5 mA cm−2.
Sealing and coating of anodized substrates by ALD.—ALD was
used to seal and coat the nano-scaled pores of anodic aluminum
oxide with different metal oxides, such as TiO2, Al2O3, an
Al2O3-TiO2 mixture, or an Al2O3/TiO2 nanolaminate. Ceramic films
with target thicknesses of 50 and 110 nm were deposited on the
smaller substrates in a low-pressure flow-type reactor29 at 125 °C by
using TiCl4 (Aldrich, purity 99.9%) and Al(CH3)3 (purity 98%,
Strem Chemicals) as metal precursors and water as the source of
oxygen. Larger substrates were coated using the same precursors and
deposition temperatures in a commercial Picosun® R200 (Picosun
OY) ALD reactor.27 In the nanolaminate, the bottom layer consisted
of 20 nm Al2O3, followed by 10 nm TiO2 and 10 nm Al2O3 layers in
turn until the total target thickness of 50 or 110 nm was reached,
with TiO2 always as the top layer. The substrates, which were sealed
by ALD with a 50 nm material, were used for thorough studies. For
the salt-spray test, the thickness of the sealing material was increased
to 110 nm. A 50 nm thick Al2O3-TiO2 mixture layer was grown with
250 ALD supercycles as described earlier.28 Each successive
supercycle consisted of one full ALD cycle for the deposition of an
Al2O3 sublayer and two full ALD cycles for the deposition of a TiO2
sublayer, thus the mixture was built by successive one alumina
sublayer following two titania sublayers. This was performed to
compensate for two times lower growth rate of titania per full
growth cycle in comparison with the growth rate of alumina.30 As
shown in our earlier work the mixture layer deposited by exact the
same conditions onto Si substrate has thickness of 52 nm, deter-
mined via X-ray reflection, and has following composition: Ti =
33.3, Al = 25.3, O = 40.0, and residual Cl = 1.4 mass%, determined
via thin film X-ray fluorescence analysis.28 Based on the analysis
data the Al2O3-TiO2 mixture has the oxides ratio of 1:1.5.
Surface and coating characterization.—HR-SEM and local
analyses were performed using a Helios NanoLab 600 (FEI)
equipped with an energy-dispersive X-ray spectrometry (EDX)
analyzer INCA Energy 350 (Oxford Instruments) and focused ion
beam (FIB) device. HR-SEM surface and cross-sectional studies
were performed using a 10 kV accelerating voltage for the primary
electrons. For the SEM-EDX studies, an accelerating voltage of
5‒30 kV was used. The cross-sections and (S)TEM lamellae were
prepared as follows. Prior to cross-sectioning, the surface of the
sample was locally coated with a ∼1 μm thick protective Pt layer
using first a 10 kV electron beam and then a 30 kV focused Ga ion
beam. Both beams scanned a predetermined area and decomposed an
organometallic platinum compound, which was sprayed onto the
Journal of The Electrochemical Society, 2022 169 071503
surface with a needle. Subsequently, FIB milling was performed
through the Pt-layer using a high voltage (30 kV) and an intense
(21 nA) focused ion beam directed at the surface at 90 degree angle.
The final cleaning of the cross sections was carried out at the same
acceleration voltage using a beam current of 9 nA. The milling and
cleaning processes were monitored using ion beam generated
secondary electrons. The (S)TEM lamellae were cut out using an
accelerating voltage of 30 kV and a beam current of 9 nA. The
lamellae were thinned to electron transparency using the same
accelerating voltage, gradually reducing the beam current and
periodically checking the transparency of the lamellae with the
electron beam and a STEM retractable detector. The lamellae were
then polished by gradually reducing the acceleration voltage and the
beam current in order to avoid amorphization of the lamellae surface
layers caused by the ion beam. The final polishing was done with
2 kV and 3 pA ion beam. The HR-STEM and -EDX studies were
performed with a Titan 200 (FEI) analytical probe-corrected high-
resolution electron microscope, equipped with a ChemiSTEM 4
SuperX SDD EDX system (FEI/Bruker), using primary electrons
with acceleration voltage of 200 kV. The hardness of the coatings
was measured with a Nano Indenter G200 (Agilent Technologies) in
the continuous stiffness measurement (CSM) mode. The plots were
Figure 1. SEM images of the AAO layer obtained by potentiostatic anodizing at 20 V, 1 °C depicting the most abundant shallow craters from the removal of
micrometric IMP-s (a), a FIB made cross-section of a sealed AAO with ALD TiO2 layer, which visualize well the 3D structure of branched nanopores (b), top
view of a cracked crater (c), FIB-made cross-section of a cracked crater (d); a cross-section of another cracked crater sealed with TiO2 using ALD to demonstrate
the efficiency of sealing (e); image (f) to show the right part of image (e) at higher magnification. In images (e) and (f), the regions of metal substrate, sealed
nanoporous AAO layer and partially removed IMP are labelled I, II and III, respectively. Arrows indicate cracks, and label IV indicates the gap left by the IMP
during the anodizing process and subsequently closed by ALD.
Journal of The Electrochemical Society, 2022 169 071503
obtained by averaging the results of 20 nanoindentation tests for
each sample. Hardness and Young’s modulus values were deter-
mined at their maximum values in the corresponding plots.
Testing of the coatings.—The corrosion resistance of coated and
uncoated Al alloy substrates was studied by linear sweep voltam-
metry (LSV) and electrochemical impedance spectroscopy (EIS),
and tested via immersion in an unmixed salt solution, neutral salt
spray, and flux of atomic oxygen. The immersion tests were carried
out in a naturally aerated neutral (pH ∼ 6.5) 0.5 M NaCl (purity
⩾99.8%, Sigma-Aldrich) solution at room temperature (22 °C ±
2 °C), using smaller and larger Al-alloy substrates. Preliminary tests
with smaller plates took place for up to 1 year, and tests with larger
plates took place for up to 1000 h. Salt spray tests for up to 1000 h
also took place and were performed in a ClimaCORR® 400-FL
chamber (VLM GmbH) at 35 °C ± 2 °C using a 5% NaCl
(SaliCORR) solution at pH 6.5–7.0, according to ISO 9227:2012.
The samples were photographed before, periodically during, and
after the immersion and salt-spray tests. Electrochemical corrosion
studies were carried out in a PTC1TM Paint Test Cell (Gamry) using
a potentiostat Reference 600 (Gamry), a saturated calomel reference
electrode (SCE), and a Pt wire as the counter electrode. The
electrolyte used in the electrochemical tests was a naturally aerated
0.5 M NaCl (⩾99.8%, Sigma-Aldrich) solution, and the experiments
were performed at room temperature, 23 °C ± 1 °C. A reproducible
surface area of 1 cm2 was achieved on the coated and uncoated
smaller substrates using 1 cm2 electrochemical sample masks
(PortHolesTM, Gamry). For LSV testing at a scan rate of 1 mV s−1,
the initial and end potentials were −1 V and +1 V vs SCE,
respectively. All potentials in this work were compared with SCE,
unless otherwise mentioned. The current density (j1V) was measured
at 1 V to evaluate the performance of the coatings at anodic
potentials. The EIS measurements were performed twice during
24 h immersion in 0.5 M NaCl at open circuit potential (OCP). The
first measurement was carried out at the start of immersion after the
sample had stabilized for 30 min, and the second measurement after
24 h of immersion. The measurements were performed in the
frequency range 10−2‒106 Hz, with an AC perturbation amplitude
of 10 mV (RMS). Echem AnalystTM (Gamry) software was used to
interpret the EIS data using equivalent circuit models.
To evaluate the performance of the nanostructured coating in
space, the coated substrate was tested by exposure to a flux of atomic
oxygen (ATOX test) at a low earth orbit simulation facility
(LEOX).31 In the test, the atomic oxygen flow was 2.7 × 1021 atoms
cm−2, which is comparable to 1.1 years of exposure to a direct flux
of atomic oxygen in low Earth orbit at an altitude of 400 km, the
same as for the International Space Station.
Results
Studies of anodized substrates (see also Supplementary Material)
have shown that in order to achieve complete protection against
corrosion, it is necessary to seal all the nano-scaled pores in the
AAO layer as well as the cracks in it, and any leftover intermetallic
particle (IMP) left inside particle craters. For this purpose, ALD was
used to deposit 50‒110 nm of Al2O3, TiO2, Al2O3-TiO2 mixture, or
Al2O3/TiO2 nanolaminate coatings on anodized substrates using
appropriately chosen ALD parameters.27 The temperature window
for coating AA2024-T3 alloy is actually quite narrow. The alloy is
artificially aged at 180 °C in the thermal treatment process.
Therefore, 125 °C is a temperature at which the alloy can be
processed for several hours without changing its mechanical proper-
ties. On the other hand, alumina and titania grown at this
temperature by thermal ALD are amorphous.26,27 Amorphous phase
is preferred for corrosion protection coating because the material do
not contain crystalline grains and defective interfaces between them
that allow fast diffusion of corrosive species through the coating.
The effect of potentiostatic anodizing on the Al-alloy.—The pre-
treatment of the Al alloy was performed by potentiostatic anodization at
20 V, 1 °C because it can be easily performed with a low-cost anodizing
setup while also having a good control over the temperature of the
electrolyte by using an external ice bath. Anodizing at temperatures
lower than room temperature would also result in denser and harder
AAO, which is desired in practical applications.19,20
High-resolution scanning electron microscopy (HR-SEM) images
of the AAO shown in Fig. 1a reveal shallow craters that were likely
left behind by the removal of the most abundant Al2CuMg IMPs.
Similar craters are seen by the use of potentiodynamic anodizing in
our previous paper.27 Furthermore, the focused ion beam made cross
section depicted in Fig. 1b shows that the AAO obtained with
potentiostatic anodizing is denser than that obtained with potentio-
dynamic anodizing, see latter in Fig. 7e, in the earlier paper.27
However, current study also revealed a new type of cracked crater in
AAO (Fig. 1c), which was not observed in the potentiodynamic
anodizing process. There were only a few such cracked craters, but
surface studies suggest that it is a recurring phenomenon.
Surprisingly, the edges of the cracked craters protrude outwards.
We believe that such features could have been created by the
removal of partially exposed IMPs that were deeper in the alloy.
Figure 2. SEM images of the FIB made cross-sections of Al-alloy substrates anodized at 10 V, 1 °C (a)–(d) and 20 V, 1 °C (e)–(h), having the AAO layer sealed
by ALD with 50 nm TiO2 (a), (e), Al2O3 (b), (f), Al2O3-TiO2 mixture (c), (g) or Al2O3/TiO2 nanolaminate (d), (h).
Journal of The Electrochemical Society, 2022 169 071503
The rapid removal of IMPs during the initial stages of anodizing
would take place through the narrow path on the top, which could
become clogged and result in a buildup of pressure and eventually
rupture. Further cracking would also be promoted by the mechanical
stresses created by the expansion of the material during anodization,
as aluminum oxide takes up more space than the original metal. It is
also important to mention that the cracks penetrate deeply into the
AAO layer and may even reach the metal (Fig. 1d). This would be
problematic in practical applications because such cracks would also
provide a pathway for corrosive species to the metal substrate. To
confirm the presence of such dangerous pathways, we used electro-
plating to visualize the defects, as in a previous work.32 Study of the
AAO layer with Ag electrodeposition is discussed in the
Supplementary Material in detail. Figures 1e and 1f show a cross-
section of another cracked crater coated with ALD deposited TiO2
nanolayer, see the darker stripe around the crater and cracks. The
images demonstrate the conformal coating of the crater wall and
successful sealing of tip part of the cracks (marked with arrows), and
the nanochannel connecting the partially removed IMP with the
crater (IV). Thus, all potentially corrodible areas are well sealed. At
the bottom of the crater, in the center, it is seen a mound formed by
Pt that moved through the crater opening during the deposition of the
Pt-layer by FIB. For this reason, the TiO2 ALD coating is under the
Pt mound.
Thorough study of the AAO layer sealing with varied ALD top
coatings.—The purpose of this thorough study was to determine the
best possible material for sealing the AAO layer obtained by
potentiostatic anodizing to create an efficient nanostructured coating.
For this purpose, ALD was used to grow Al2O3, TiO2, an
Al2O3-TiO2 mixture, or Al2O3/TiO2 nanolaminate films on anodized
substrates using appropriately chosen ALD pulse times.27 In these
studies, a target thickness of 50 nm was deemed to be sufficient in
the ALD process to seal the nanoscale pores, as observed in Fig. 1b.
SEM-FIB studies of nanostructured coatings.—SEM-FIB studies
revealed that the total thickness of the coatings was within the
targeted ⩽5 μm range (Fig. 2). As expected, the thickness of the
coating was mainly determined by the potential used in the
anodization process. In particular, anodizing at 10 V, 1 °C resulted
in a coating thickness of ∼1 μm (Figs. 2a–2d). In comparison,
anodizing at 20 V, 1 °C provided coatings that had a thickness of
2–4 μm (Figs. 2e–2h). In addition, the pore diameter and overall
volume in a unit are larger in the latter case, as can be seen by
comparing the cross-sectional images sealed by TiO2 AAO in
Figs. 2a and 2e (note that image a has 1.5 × higher magnification
than image e). The SEM-FIB studies show that by using appro-
priately chosen ALD pulse times, it is possible to efficiently seal the
nanoscale pores in the AAO layer with various ceramic materials
such as TiO2 (Figs. 2a, 2e), Al2O3 (Figs. 2b, 2f), Al2O3-TiO2
mixture (Figs. 2c, 2g), and Al2O3/TiO2 nanolaminate (Figs. 2d, 2h).
It should be noted that the first layer of the nanolaminate consists of
20 nm of Al2O3, which seals most of the nanoscale pores in the AAO
layer. Consequently, in the case of the nanolaminate, the amount of
TiO2 in the sealed AAO layer is small, as it can only be deposited
into wider pores, cracks, and craters that are not completely sealed
by the first 20 nm of Al2O3. The sealing of nanoscale pores also has
an impact on the overall conductivity of the entire nanostructured
coating. For instance, sealing with TiO2 results in a more conductive
coating, which makes it easier to study via HR-SEM as it does not
suffer from extensive charging, which has a negative impact on the
imaging process for visualization of the internal nanoscale structure
(Fig. 2a). In contrast, sealing the AAO with Al2O3 (Fig. 2b) or
Al2O3/TiO2 nanolaminate (Fig. 2d) results in a more dielectric
coating, which is observed in the SEM studies. However, the top
TiO2 layer in the nanolaminate provides near-surface conductivity
owing to its semiconducting properties (Fig. 2d). Using the
Al2O3-TiO2 mixture to seal the AAO (Fig. 2c) results in a coating
that has better overall conductivity than that obtained by sealing with
Al2O3 (Fig. 2b). Based on these observations, it is clear that the
electrical conductivity of nanostructured coatings can be designed to
satisfy the requirements of specific applications. Dielectric coatings
can be made by sealing and coating AAO with Al2O3 whereas
semiconductor coatings can be made by sealing and coating AAO
Figure 3. The results of LSV corrosion tests performed in naturally aerated 0.5 M NaCl solution for the Al-alloy substrates that were coated using ALD with
50 nm TiO2 (a), Al2O3 (b), Al2O3-TiO2 mixture (c) or Al2O3/TiO2 nanolaminate (d), showing the polarization curves for the: polished sample (1), polished
samples coated by ALD (2), samples that were anodized at 10 V, 1 °C, and coated by ALD (3), samples that were anodized at 20 V, 1 °C, and coated by ALD (4).
Journal of The Electrochemical Society, 2022 169 071503
Table I. Data obtained via LSV and nanoindentation studies on bare and coated substrates with different pre-treatments and 50 nm ceramic films grown by ALD.
Substrate pre-treatment ALD coating Pitting potential Epit, V Current density j1V, A cm
−2 Coating efficiency, CE (Eq. 1) Hmax, GPa Emax, GPa
Standard — −0.59 0.20a) — 3.0 ± 1.2 100 ± 13
Standard TiO2 −0.53 0.16 0.20 3.3 ± 0.9 96 ± 23
Standard Al2O3 −0.48 6.2 × 10
−2 0.68 4.1 ± 1.3 93. ± 23
Standard Mixtureb) −0.57 0.16 0.20 3.3 ± 1.4 98 ± 34
Standard Laminatec) −0.54 6.4 × 10−2 0.68 3.5 ± 0.8 96 ± 10
10 V anod. — −0.56 0.13 0.33 3.0 ± 1.3 88 ± 22
10 V anod. TiO2 −0.45 1.5 × 10
−4 1.00 5.5 ± 1.8 102 ± 18
10 V anod. Al2O3 N/A 1.8 × 10
−11 1.00 7.2 ± 1.8 121 ± 26
10 V anod. Mixture −0.46 3.5 × 10−3 0.98 5.2 ± 2.2 100 ± 37
10 V anod. Laminate 0.35 1.7 × 10−5 1.00 6.5 ± 1.3 111 ± 17
20 V anod. — −0.52 0.14 0.30 2.3 ± 0.5 75 ± 11
20 V anod. TiO2 N/A 1.3 × 10
−7 1.00 4.5 ± 1.2 95 ± 20
20 V anod. Al2O3 0.75 6.6 × 10
−9 1.00 5.2 ± 1.6 95 ± 25
20 V anod. Mixture −0.27 1.6 × 10−3 0.99 3.0 ± 0.9 70 ± 8
20 V anod. Laminate N/A 6.5 × 10−10 1.00 5.5 ± 1.3 99 ± 19
a) jmax = 20nA. b) – Al2O3–TiO2. c) – Al2O3/TiO2.
Journal
of
T
he
E
lectrochem
ical
Society,
2022
169
071503
with TiO2. Alternatively, a coating can be made dielectric in the
depth direction by sealing the AAO with Al2O3 or semiconducting
on the top surface by applying a top layer that includes TiO2.
Linear sweep voltammetry study of nanostructured coatings.—
Corrosion testing by linear sweep voltammetry was carried out on
polished and anodized substrates sealed and coated by ALD with
50 nm films (Fig. 3, Table I). In comparison with the bare polished
sample, all the coated substrates exhibited lower current values
across the entire scan range, which indicates an improvement in
corrosion resistance. However, a more detailed comparison can be
made by observing the pitting potential (Ecor) and current density
(j1V) values at 1 V. The latter can also be used to calculate the short-
term coating efficiency (CE), similarly to our previous study:28
Figure 4. Surface hardness, H, vs indentation depth, h, curves of anodized at 10 V, 1 °C (a) and 20 V, 1 °C (b) samples. The curves 1 belonging to just-polished
substrate and curves 2 to anodized samples, other curves to anodized and ALD sealed and coated samples, using 50 nm of TiO2 (curve 3), Al2O3 (curve 4),
Al2O3-TiO2 mixture (curve 5) or Al2O3/TiO2 nanolaminate (curve 6).
Figure 5. EIS study results: Bode plots for the substrates with nanostructured coatings prepared by anodizing at 20 V, 1 °C and sealed by ALD with 50 nm (a)
and 110 nm nanolaminate (b). In Bode plots the markers show experimental and linear modelled values. An equivalent circuit model (ECM) given in (c) was
used for modelling the plot curves in (b).
Table II. Calculated ECM (Fig. 5c) variables for Al alloy substrates with nanostructured coating prepared in this paper (Fig. 5b) and in Ref. 27
before and after 24 h immersion in salt solution.
EMC variables
Coatings
Anodized at 20 V, 1 °C plus 110 nm nanolaminate Anodized at 10 V, 22 °C plus100 nm laminate27
Immersion time, h 0.5 24 0.5 24
Fixed Rsoln, Ω cm
2 10.0 10.0 10.0 10.0
Rc, Ω cm
2 2.6 × 109 2.4 × 109 1.2 × 109 1.3 × 109
Cc, F cm
−2 sn−1 3.7 × 10−9 3.7 × 10−9 7.2 × 10−9 7.2 × 10−9
n 0.96 0.96 0.97a) 0.97a)
a) n = n2.
Journal of The Electrochemical Society, 2022 169 071503
= − ( / ) [ ]CE 1 j j . 11V max
According to the measured polarization curves depicted in Fig. 3 and
summarized data given in Table I, no pitting corrosion was observed
for the substrates that were, first, anodized at 20 V, 1 °C and sealed
with TiO2 (Fig. 3a, curve 4), second, anodized at 10 V, 1 °C and
sealed with Al2O3 (Fig. 3b, curve 3), and third, anodized at 20 V, 1 °
C and sealed with Al2O3/TiO2 nanolaminate (Fig. 3d, curve 4).
These results indicate that these coatings were defect-free and that
the corrosive species could not come into direct contact with the
metal substrate. In the case of sealing the AAO with TiO2, the
measured current densities across the entire scan range were notably
higher than those of the samples where the AAO was sealed by
Al2O3 or by Al2O3/TiO2 nanolaminate. This can be explained by the
higher conductivity of amorphous TiO2 in comparison with Al2O3,
which was also utilized in our first study to visualize the pores in the
AAO layer.27 Photographs of the substrates after the LSV tests
(Fig. S5 (available online at stacks.iop.org/JES/169/071503/mmedia)
in the Supplementary Material) are in good agreement with the
measured polarization curves depicted in Fig. 3.
Hardness of nanostructured coatings.—Nanoindentation studies
showed that the measured hardness (H) and Young’s modulus (E)
values depend on the surface pre-treatment, as well as on the
material grown by ALD (Fig. 4 and Table I).
The coatings grown by ALD directly onto the substrate slightly
increased the surface hardness value in comparison with the bare
substrate (Table I). However, the differences observed in the
measurements were negligible, except for the alumina film, which
was somewhat harder than the substrate. It should be noted that the
hardness of the coatings grown by ALD was studied separately in
our previous study, where the materials were grown on Si
substrates.28 Comparing the hardness values of the same coatings
deposited on a relatively soft Al alloy (H = 3 GPa; Table I) and on a
hard Si substrate (H = 13 GPa),28 the former values are considerably
lower than the latter. It seems that during the nanoindentation of thin
coatings on a soft substrate, the ceramic coating is cracked and
penetrated by the tip at the very beginning of the indentation.
Therefore, the deformed substrate has a greater impact on the
measured H values than the coatings do. However, the situation is
different for substrates with thicker non-sealed AAO layers. For
instance, the ∼1 μm thick AAO obtained by anodizing 10 V, 1 °C
exhibited a hardness similar to that measured for the substrate.
However, the H-curve in Fig. 4a (curve 2) shows that the non-elastic
deformation starts at the very beginning of the indentation process.
This is much more obvious for the AAO obtained at 20 V, 1 °C,
which had a thickness of ∼2 μm (Fig. 4b, curve 2). This behavior
can be explained by the structure of AAO, in which the pores have a
nanometric diameter and the separating ceramic walls have a similar
thickness. It is also noted that thicker AAO exhibited lower hardness
values than those of thinner AAO. As shown in Figs. 2a and 2e, the
thicker AAO has a somewhat lower density owing to wider pores
and larger volume in a unit, which explains its lower hardness in
both the non-sealed and sealed coatings (Fig. 4 and Table I).
Sealing AAO by ALD with different materials resulted in
nanostructured coatings that were mostly harder than the Al alloy
(Fig. 4 and Table I). This can be explained by filling of empty pores
Figure 6. Photos of 40 × 110 mm2 AA2024-T3 samples after 1000 h ISO 9227 salt spray test at 35 °C depicting a just-polished sample (a), sample just-anodized
at 20 V, 1 °C (b), sample with deposits onto the polished substrate of 50 nm nanolaminate (c) and 110 nm nanolaminate (d), sample anodized at 20 V, 1 °C and
sealed with 50 nm nanolaminate (e) and 110 nm nanolaminate (f), sample anodized at 20 V, 1 °C following hydrothermal sealing (g), sample anodized at 20 V,
23 °C and sealed with Rust Stop paint (h).
Journal of The Electrochemical Society, 2022 169 071503
in the hard AAO layer with ceramic material, which increases the
overall density of the coating. Furthermore, the materials used for
sealing also have distinct hardness, which contributes to the
mechanical properties of the nanostructured coating. The hardest
coatings were obtained using Al2O3 or Al2O3/TiO2 nanolaminate to
seal the AAO (Fig. 4, curves 4, 6 and Table I). This is not surprising,
as Al2O3 and Al2O3/TiO2 nanolaminate films grown by ALD on Si
substrates also exhibit the highest hardness values.28
Furthermore, the denser AAO obtained by anodizing at 10 V,
1 °C also provided harder nanostructured coatings than using 20 V,
1 °C anodizing for pre-treatment (Fig. 4, curves 3–6 and Table I).
The lowest hardness was observed for the nanostructured coating,
which was prepared by anodizing at 20 V, 1 °C and sealed with the
Al2O3-TiO2 mixture (Fig. 4b, curve 5, and Table I).
Characterization and testing of best-performing nanostructured
coating.—The studies described above showed that only three
recipes resulted in a nanostructured coating that did not exhibit
pitting corrosion in the linear sweep voltammetry (LSV) tests and
that using 20 V for anodizing generally results in better short-term
corrosion protection coatings (Fig. 3). On the other hand, we have to
consider the fact that aluminum oxide is converted into aluminum
hydroxide by staying in a moist environment or aqueous media for a
long time.28 Therefore, pure alumina coatings are impractical.
Furthermore, from these three coatings, the highest surface hardness
was observed for the substrate that was anodized at 20 V, 1 °C and
then sealed by ALD with 50 nm nanolaminate (Fig. 4). Based on
these results, we further studied and optimized the best-performing
coating recipe.
EIS study of nanostructured coating.—The purpose of the
electrochemical impedance spectrometry study depicted in Fig. 5
was to evaluate the stability of the best-performing nanostructured
coating before and after 24 h immersion in a neutral 0.5 M NaCl
Figure 7. STEM study results of nanostructured coating, depicting a) bright field and b) high-angle annular dark field (HAADF) images made of a lamella
prepared from the coating by FIB. STEM-EDX analysis results of the same lamella show the distribution of Al (c), Ti (d), O (e), Cu (f), Mn (g), and S (h) in the
top part of the nanostructured coating.
Journal of The Electrochemical Society, 2022 169 071503
solution. The EIS studies were performed with substrates that were
obtained by anodizing at 20 V, 1 °C and sealed by ALD with a
50 nm nanolaminate (Fig. 5a) and with a 110 nm nanolaminate
(Fig. 5b). The EIS measurements were done after 0.5 h stabilization
at OCP (denoted as “0 h”) and after 24 h immersion (denoted as
“24 h”) in 0.5 M NaCl solution.
Figure 8. Photos, SEM and STEM images of an as-prepared and ATOX-tested nanostructured coating, made by anodizing at 20 V, 1 °C following enhancement
by ALD by adding a 110 nm thick nanolaminate: a) photo of the substrate before ATOX test and b) after ATOX test, c) HR-SEM image of the nanostructured
coating before ATOX test and d) after ATOX test, e) STEM high-angle annular dark field image of the lamella of the nanostructured coating after ATOX test and
f) with overlay of the Ti distribution map measured by EDX. A local Pt mask was used for surface protection during preparation of the lamella by FIB.
Journal of The Electrochemical Society, 2022 169 071503
In the equivalent circuit model depicted in Fig. 5c, R.E. is the
reference electrode, W.E. is the working electrode (the Al alloy),
Rsoln is the electrolyte resistance, Rc is the coating resistance, Ø
marks constant phase element (CPE). The constant phase element
considers the coating capacitance (Cc), and exponent n is used in the
calculation of the CPE. The exponent indicates the deviation of the
insulator of a capacitor from an ideal dielectric behavior such that
n = 1 for an ideal capacitor and n = 0 for an ideal resistor.33
EIS study results (Fig. 5) showed that the nanostructured coating
prepared by sealing with 50 nm of nanolaminate was unstable, and at
the end of the 24 h test the impedance ∣Z∣ had considerably decreased
at lower frequencies f and no longer exhibited linear behavior
(Fig. 5a). Furthermore, the θ-curves, where θ is phase angle
depending on f, changed in shape and were near the 90° level only
in a narrow high-frequency range. In contrast, the nanostructured
coating formed on the Al alloy substrate by anodizing at 20 V, 1 °C
and sealed by ALD with 110 nm nanolaminate performed well in the
EIS study, as the ∣Z∣- and θ-curves did not change during the 24 h
immersion experiment (Fig. 5b). Furthermore, the impedance curve
exhibited an almost linear behavior over the entire frequency range
of 2 × 10−2‒105 Hz. The measured θ- curve was also near the 90°
level in a broad frequency range. To model the behavior of the better
(second) coating depicted in Fig. 5b, a simpler ECM (Fig. 5c) with a
single time constant was found to be adequate. As can be seen from
Table II, the Rc, Cc, and n values modelled for the second coating did
not change after the 24 h immersion test, which confirms that the
coating is chemically stable and free of major defects. Similar
behavior and modelled ECM values were observed for a nanos-
tructured coating in our previous paper,27 where potentiodynamic
anodizing in a three-electrode setup was used for pretreatment prior
to sealing by ALD with 100 nm nanolaminate.
Salt spray testing of nanostructured coatings.—Salt spray testing
was performed on eight substrates with different pre-treatments and
coatings. The first substrate was polished and left bare (Fig. 6a). The
second one was polished and anodized at 20 V, 1 °C, but the AAO
was not sealed (Fig. 6b). The third and fourth substrates were
polished and coated by ALD at 50 nm (Fig. 6c) and 110 nm
(Fig. 6d), respectively. The fifth and sixth substrates had a
nanostructured coating, which was prepared by anodizing the
substrates first at 20 V, 1 °C and then sealed by ALD with 50 nm
(Fig. 6e) and 110 nm (Fig. 6f) nanolaminate. For comparison, the
seventh substrate (Fig. 6g) was anodized at 20 V, 1 °C, and then
hydrothermally sealed by immersing the sample in boiling deionized
water for 10 min The eighth substrate (Fig. 6h) was anodized in 15%
sulfuric acid at 20 V, 23 °C, which was followed by sealing the pores
with a commercial Rust Stop spray paint.34
Photos taken after the 1000 h salt spray test show severe damage
to both polished and anodized substrates (Figs. 6a, 6b), where the
latter exhibit slightly better performance. The polished substrates,
which were coated with 50 and 110 nm nanolaminate, remained less
damaged but had numerous small corrosion pits (Figs. 6c, 6d).
Although the difference in performance between these two samples
was not large, the polished Al alloy plate with a 110 nm coating had
a smaller number of corrosion pits. In contrast, the anodized
substrates that were sealed by ALD with a 50 nm (Fig. 6e) and a
110 nm nanolaminate (Fig. 6f) exhibited exceptional performance,
with only two corrosion pits on the former sample (marked with red
circles) and none on the latter sample. The Al alloy plate, which was
anodized and then sealed by hydrothermal treatment (Fig. 6g),
showed similar performance to that of the polished substrates, which
were coated with nanolaminate, but had a much higher number of
corrosion pits. Finally, the sample anodized at room temperature,
which was sealed with commercial Rust Stop paint, remained
unharmed during the salt spray test (Fig. 6h). The latter coating
had a total thickness of 28–33 μm; thus, it was not a thin coating.
The testing results show that the performance of the nanostructured
coating developed in this work is comparable to that of commercial
counterparts 10 times thicker, as well as with the thin nanostructured
coating investigated in our first paper, where potentiodynamic
anodization at room temperature was used for substrate
pre-treatment.27
STEM study of nanostructured coating.—A lamella of the
nanostructured coating was prepared by FIB for scanning transmis-
sion electron microscopy (STEM) studies from the sample that was
prepared by anodizing the alloy at 20 V, 1 °C, followed by sealing
with 110 nm nanolaminate by ALD. The bright field STEM image of
Figure 9. Use of nanolaminate and nanostructured coatings in space technology. Photos of satellite components (a)–(e) with nanolaminate (a) and
nanostructured coating (b)–(e) that will be used on WISA Woodsat (a), (c)–(e) and ESTCube-2 (b). The surface of the nanostructured coating on 3D printed
aluminum is depicted in an SEM image (f).
Journal of The Electrochemical Society, 2022 169 071503
the lamella depicted in Fig. 7a shows individual layers of nanola-
minate on the top part of the coating and a sealed AAO layer
underneath it. However, a high-angle annular dark-field image of the
same region depicted in Fig. 7b clearly shows the presence of a
nanoscale structure. It should be noted that the lamella in the studied
region has a uniform thickness, and the anodic aluminum oxide and
atomic layer deposited aluminum oxide have the same or similar
chemical composition. Further analysis of the lamella by mapping
the distribution of elements with energy dispersive X-ray spectro-
meter (Figs. 7c–7h) shows that the AAO layer was depleted of Cu
(Fig. 7f) during anodizing, which is in agreement with the SEM-
EDX study depicted in Table SII in the Supplementary Material. The
elemental maps also show the presence of sulfur in AAO owing to
the use of sulfuric acid for anodizing the alloy (Table SII).
According to the STEM-EDX study (Fig. 7h), sulfur is distributed
uniformly in the AAO. However, it is no longer part of any reactive
compound. After the end of the anodizing residual acid left in the
pores reacts with the metal and/or oxide creating Al2(SO4)3, thus it is
not generating further corrosion. The STEM-EDX study also shows
a region with a higher Mn content (Fig. 7g), which likely originates
from an IMP that was not completely removed. However, the mean
concentration of Mn in the coating measured in SEM-EDX studies is
negligible (Table SII). Finally, the distribution of Ti (Fig. 7d)
confirms that the nanostructured coating was prepared as intended,
resulting in an AAO sealed mainly by Al2O3 and having a top layer
of nanolaminate.
ATOX testing of nanostructured coating.—The best-performing
coating, which was prepared by anodizing at 20 V, 1 °C and sealed
by ALD with 110 nm nanolaminate, was tested with a flux of
energetic atomic oxygen at the LEOX facility of the European Space
Research and Technology Centre. Visual observation of the sample
before (Fig. 8a) and after (Fig. 8b) testing with atomic oxygen
showed no obvious changes in the ceramic coating. However, the
blue marker stripe seen in Fig. 8a was completely removed
compared to the photograph taken after the test (Fig. 8b). The slight
color difference of the samples was caused by different lighting
conditions when taking the photos. Therefore, visual observation
showed that exposure to atomic oxygen dramatically damaged a thin
organic film, the stripe of a blue marker,35 but the condition of the
ceramic coating could not be evaluated.
A detailed HR-SEM study of the surface of an untested (Fig. 8c)
and tested substrate (Fig. 8d) revealed that exposure to a flow of
atomic oxygen caused a change in the surface morphology only at
the nanoscale. Small grooves with a depth of ⩽10 nm were formed
on the surface of the tested substrate (Fig. 8d). HR-SEM studies also
revealed that both the untested (Fig. 8c) and tested (Fig. 8d)
substrates were sufficiently conductive for characterization by
SEM. However, the surface of the tested substrate (Fig. 8d) could
be slightly less conductive, which would enhance the image contrast,
making it easier to characterize by SEM. Because the only
conductive component in the coating is amorphous TiO2 prepared
by ALD, it was necessary to study the change in detail.
Further HR-TEM analysis of a lamella prepared by SEM-FIB
from the tested sample (Figs. 8e, 8f) did not show any changes in the
top TiO2 layer of the nanolaminate (Fig. 8f). Thus, the precise cause
of the change in the surface morphology of the tested sample is not
clear. One possibility is the partial crystallization of the top TiO2
layer due to exposure to energetic atomic oxygen; however, TEM
studies did not reveal nanocrystallites in the surface layer.
Alternatively, the original TiO2 layer on top of the nanolaminate
may be non-uniform in terms of mechanical properties, which could
cause energetic atomic oxygen to sputter away the softer part of the
titania topmost layer; however, both of these hypotheses are difficult
to study in the TEM image because the upper part of the top layer
was masked by platinum particles during the lamella preparation.
Because the coating suffered only minimal changes during the
test with atomic oxygen while retaining its conductivity, it is likely
that the developed nanostructured coating is suitable for practical
applications in low Earth orbit.
Development and application of protective coatings for small
satellites.—Functional coatings of different thicknesses can be
produced for aluminum alloys to significantly improve their perfor-
mance. The developed coatings are particularly useful for satellite
parts as they can be applied on substrates with complicated three-
dimensional shapes without having a significant impact on their
dimensions and weight. These coatings can be prepared by ALD and
appropriate pre-treatments, including also conventional potentio-
static anodizing. Some examples of how these different types of
coatings can be used are shown in Fig. 9.
On ESTCube-2, the nanostructured coating was applied to an
aluminum 65 × 41 × 3 mm3 cover panel (Fig. 9b, site I) of the
materials testing module, which exposes the tested materials through
holes (e.g., sites II and III) to atomic oxygen at a low Earth orbit. In
this module, site II was used to test the nanostructured coating on a
0.1 mm foil, which was compared with the performance of an
uncoated foil at site III. Visual inspection of the coated cover panel
did not reveal any imperfections, and the sample looked similar to
the AA2024-T3 sample with a nanostructured coating in the salt
spray test (Fig. 6f).
On WISA Woodsat, a nanostructured coating was used on over
50 aluminum parts fabricated by 3D printing. These parts include
panels (Fig. 9c), the main frame structure of the cube satellite
(Fig. 9d), and parts of the mechanism (Fig. 9e) that extract the
camera from the satellite to capture photos of its external surface
while in orbit. In the extraction mechanism, a coating is used to
mitigate cold welding in space. Assembly of the coated parts did not
result in any loss of mobility. An additional SEM study of the
surface showed that the coating was defect-free and followed the
surface morphology of the 3D printed substrate (Fig. 9f).
Discussion
Choosing ALD process temperature and coating materials for
protection of the Al alloy.—The list of materials that could be used
as an ALD protective coating for the Al-alloy is quite limited.
Firstly, the temperature at which the coating is prepared must be
relatively low, < 150 °C not to decrease mechanical properties of the
alloy.1–4,36 Secondly, it is preferred that the corrosion protection
coating is insulating and amorphous. Thirdly, since the metal
precursor molecules must diffuse rapidly in and out of the AAO
pores, the molecules need to be as small as possible. Fourthly, the
precursors must be sufficiently reactive at low deposition tempera-
tures but must not initiate corrosion of the metal substrate at the
beginning of the process. Fifthly, for sealing the narrow and long
pores of AAO and coating the samples with complex 3D geometry,
it is not possible to increase the rate of the surface chemical reactions
in ALD using plasma or other physical excitants. Sixthly, the cost of
the precursors must be relatively low for the technology to be
commercialize. All these conditions narrow the range of useful
precursors/materials. We are currently not aware of any studies that
use ALD coatings of materials other than alumina and/or titania to
more or less successfully protect the Al2024 alloy. The Al2O3/TiO2
nanolaminates, that were first proposed by Matero et al. in 1999 for
the protection of a steel,37 and later investigated by other researchers
on Al alloy substrates,23,27,28 are also one of the best candidates used
for sealing the AAO when preparing the new nanostructured
coatings (Figs. 2, 3 and Table I). Because the benefits of Al2O3 and
TiO2 compensate the weaknesses of one another. For instance,
Al2O3 would normally gradually dissolve in aqueous medium but
this is prevented by layers of TiO2, which is chemically more
stable.26 Laminates and mixtures of ALD alumina and tantalum (V)
oxide have also been used for the protection of a steel,38 but
tantalum (V) oxide is not as good an ion barrier as alumina and its
precursor materials are more expensive than TMA. In this study,
Journal of The Electrochemical Society, 2022 169 071503
alumina and titania were chosen because the materials and their
precursors used fulfill all listed above conditions. Additionally,
because alumina is a good insulator, ion barrier and suits chemically
well with AAO, and titania grown at low temperature has a certain
conductivity. Therefore, the conductivity of the coatings can be
varied, which is important in many aerospace applications. The
metal precursors, TMA and TiCl4 react well with H2O also at low
temperature, 125 °C,27,39 and are relatively cheap. Finally, the
laminate materials, alumina and titania are composed of light
elements that will not emit significant hard X-ray fluorescence
radiation due to cosmic particle/radiation bombardment in space (see
also S3 in the Supplementary Material), so there is no additional
threat to electronic components that could be shielded only by
relatively thin Al-alloy walls.
Additional aspects in the preparation of nanostructured coat-
ings for the Al alloy.—As expected, the properties of materials
grown by ALD seem to carry over to the nanostructured coatings
and affect their corrosion resistance, conductivity and hardness. For
instance, in terms of corrosion resistance, the nanostructured coating
with 110 nm nanolaminate (Figs. 5, 6) proposed in this study, is far
superior to conventional ALD coatings on the Al alloy.23,24 The
highest hardness values were observed also for nanostructured
coatings that were made by AAO sealing with Al2O3 (Fig. 4, curve
4) or Al2O3/TiO2 nanolaminate (Fig. 4, curve 6), which is in good
agreement with a previous study,28 where the two aforementioned
materials grown by ALD on Si substrate were harder than TiO2 and
Al2O3-TiO2 mixture.
The nanostructured coating with ALD nanolaminate (Figs. 5, 6)
developed in this study also seems to outperform other single and
double-layered coatings prepared with the help of ALD by other
research groups.23,24,37,38,40 For instance, a double-layered coating
made by combining sol-gel with ALD was far superior to only ALD
coating but exhibited a dramatic rise in current densities already at
mild anodic potentials in the LSV test.23 A better performance was
observed for a coating made by combining PVD with ALD, where
the pitting potential was around 0.3 V.40 In a later study performed
by Härkönen et al., the best performing PVD + ALD coating
exhibited notable signs of corrosion after 672 h neutral salt spray
test.38 In comparison, the nanostructured coating with nanolaminate
investigated in this study has no signs of corrosion after 1000 h salt
spray test (Fig. 6f) and exhibits no pitting in an LSV experiment up
to 1 V (Fig. 3d, curve 4). Furthermore, the use of electrochemical
pre-treatment and ALD allow applying the nanostructured coating
on substrates that have a sophisticated three-dimensional shape,
internal cavities or threads. As a rule, this is not possible for sol-gel
(+ ALD) and PVD (+ ALD) coatings.
When comparing our nanostructured coatings with the earlier
double deposition coatings obtained by the ALD method mentioned
above, it is important to highlight the difference in the structure and,
as a rule, in the thickness of these coatings. The earlier works were
dealing with ALD thin films that were exploited for sealing defects
in coatings made by other techniques (PVD, CVD, sol-gel method,
etc).23,39 In the electrochemical tests, the defects can be described as
pores,24 but no microscopic study has yet confirmed this. Thus, the
defects could also be cracks due to mechanical stresses in relatively
thick coatings that are more or less strait (Figs. 1d–1f) or growth
defects due to any peculiarity of the substrate surface, etc. In current
study, we first create an AAO via anodizing. The layer is chemically
bonded to the substrate that gives best adhesion at the metal∣oxide
interface; the process also cleans and homogenizes chemically the
surface. The AAO branching nanopores (Figs. 1b, 2a, 2e) are fully
sealed, first, with the same Al-oxide and if having larger diameters
then are following covered with other metal oxide layers, using for
all ALD process (Figs. 1e, 1f)—this is principal difference between
the thin nanostructured coating and mentioned above defect-sealed
double coatings, which are as a rule thick coatings.
A crucial part of preparing nanostructured coatings is also
anodizing, which creates the porous AAO layer. The latter usually
determines the final coating thickness, which depends on the
application and is discussed in more detail in the next section.
Based on our previous study27 and the findings in this study we saw
that potentiodynamic anodizing at room temperature (∼22 °C) or
potentiostatic anodizing at low temperature (1 °C) (Fig. 1), both
result in well performing nanostructured coatings, although the
radius of the AAO pores somewhat differs, but the branching
structure remains. These results show that the proposed method
for the preparation of nanostructured coatings can be quite versatile
in terms of anodization and mainly depends on the ALD process
used for the sealing. The ALD parameters has to be chosen carefully
to maintain ALD mode growth, otherwise the branched pores could
be easily clogged before the filling. Thus, an important aspect of
anodizing is ensuring the reproducibility of the process, which
requires a good control over the temperature of the electrolyte. We
found that this could be achieved easily for small parts with an
external ice bath around the anodizing bath. However, the most
appropriate temperature control method is ultimately chosen by the
user and depends on the application.
Choosing the thickness of the nanostructured protective coating
depending on the application.—The thickness range of effective
ALD based protective coatings for vulnerable to corrosion Al alloy
AA2024 has to be discussed for two types of coatings. The first type
are coatings that are deposited directly to the alloy substrate that has
been polished and/or solvent cleaned, or on the alloy surface that is
additionally electrochemically pre-treated at low potential. These
coatings are as a rule <0.3 μm thick and we classify them as ultra-
thin coatings. The second type are the coatings that were deposited
onto the anodized alloy substrate. These coatings have the thickness
0.5–3 μm and we call them as thin coatings.
The use of ultra-thin nanostructured coatings.—The corrosion
resistance of high precision and/or 3D printed aluminum alloy
components can be significantly enhanced by coating them with the
Al2O3/TiO2 nanolaminate via ALD. A 50–150 nm nanolaminate can
be applied on the substrates in less than 6 h in our laboratory ALD
reactor, but in a much shorter time in some new type reactors,41,42
making it practical for industrial use in the near future. Generally,
thorough cleaning with a mechanical surface finish and organic
solvents is sufficient. It is important to note that in moist chloride-
containing environments, the performance of the nanolaminate
depends to some degree on the condition of the initial surface,
exhibiting a coating efficiency between 0.35 in an earlier28 and 0.68
in this study, see the latter in Table I. One possible reason for that
could be different time interval between the substrate cleaning and
ALD deposition, because of relatively fast start of local corrosion
near the IMPs that are reaching the alloy surface.28 The low potential
electrochemical pretreatment can partly solve the problem.
Therefore, electrochemical treatment at low potentials can improve
the coating efficiency to 0.81.28 However, it is not possible to
completely remove large IMPs that are several micrometers in size.
For that reason, nanostructured coatings deposited directly onto the
∼20 nm thick oxide layer produced by the electrochemical pretreat-
ment could fail in some of these locations, if working in harsh
corrosive environments. At the same time, a rule of thumb is still
that the smaller are the IMPs, the lower is the risk that the protective
coating may fail—this is also a signal to the alloy manufacturers on
how to improve the alloy.
Otherwise, the nanolaminate is chemically stable in water28 and
suffers little damage from the direct flux of atomic oxygen (Fig. 8),
which makes the coating a good choice for protecting materials
against corrosion in terrestrial and space applications. An example of
using the nanolaminate is shown in Fig. 9a, where the high-precision
aluminum substrate is only a few millimeters in size and in direct
contact with another metal after assembly, which increases the risk
of galvanic corrosion due to potential difference. Because of the
small size of such substrates, hundreds of them can be coated
simultaneously during the ALD process in a commercial reactor,
Journal of The Electrochemical Society, 2022 169 071503
which greatly improves their performance at a very low increase in
cost per item.
The use of thin nanostructured coatings.—If the nanolaminate is
insufficient for high-performance applications, and the sample
dimension tolerances are not so critical, then a sub-micrometric
nanostructured coating can be considered, which is mechanically
more durable and provides excellent protection against corrosion.
Such a coating can be achieved using a carefully controlled
electrochemical pre-treatment at anodic potentials to create a sub-
micrometric porous AAO layer, which is then sealed by ALD with
the nanolaminate, see cross-sections in Figs. 1b,1d, and in Fig. 7c in
an earlier study.27 This process may require more stringent pretreat-
ment checks, but the resulting coating is defect-free and has an
efficiency of 1.0. Which is particularly useful for mitigating galvanic
corrosion on high-precision components, where the dimensional
tolerances are less than a micron. It is worth mentioning that
anodizing increases the sample dimensions by about 50% of the
AAO thickness.2
In applications where the dimensional tolerance is 1–5 μm or
greater, nanostructured coatings can be prepared using conventional
potentiostatic anodizing for pre-treatment prior to sealing by ALD
with a suitable material. In this study, we demonstrated how
anodizing at 10 V, 1 °C, and 20 V, 1 °C can be used to prepare
nanostructured coatings with thicknesses ∼1 and 2–4 μm, respec-
tively (Fig. 2). This anodizing process also benefits from a simple
method of temperature control, where an external ice bath was used
to keep the electrolyte around 1 °C and ensure stable parameters
during anodization of several batches of substrates. The choice of
material used for sealing by ALD allows the modification of the
properties of the nanostructured coating. For instance, insulating
coatings are achieved by sealing AAO with Al2O3, which is
beneficial for creating a dielectric barrier or simply a hard surface.
Alternatively, sealing the nanoscale pores of AAO with Al2O3 and
applying a nanolaminate on top will result in a nanostructured
coating, which benefits from its dielectric properties and hardness
while also providing additional chemical stability for use in moist
environments, and a semiconductive surface to mitigate charging in
space. If it is necessary to create a more conductive coating overall,
then the nanoscale pores in the AAO can be sealed by ALD with
TiO2 instead of Al2O3. The only possible disadvantage of using
potentiostatic anodization at low temperatures for pre-treatment may
be the creation of cracks in the AAO layer and limited dissolution of
the IMP leftovers. In our study, we mitigated this possible issue by
sealing the AAO with 20 nm of Al2O3 and then coating the entire
surface with an additional 90 nm of nanolaminate. This type of
nanostructured coating was also used to cover the satellite parts
depicted in Figs. 9b–9e for different purposes.
Acknowledgments
The authors would like to acknowledge the Estonian Ministry of
Education and Research by granting the projects IUT2–24,
TLTFY14054T, PSG448, PRG4, SLTFY16134T and by the EU
through the European Regional Development Fund under project
TK141 (2014-2020.4.01.15-00). The atomic oxygen testing was
performed in the framework of the “Announcement of opportunity
for atomic oxygen in the ESTEC Materials and Electrical
Components Laboratory/ESA-TECQE-AO-013375),” through a col-
laboration with Picosun Oy. The authors also thank Dr. Elo Kibena-
Põldsepp for the electrodeposition of Ag onto the anodized
substrates.
ORCID
Lauri Aarik https://orcid.org/0000-0002-7785-1351
Jekaterina Kozlova https://orcid.org/0000-0002-7775-1246
Aivar Tarre https://orcid.org/0000-0003-1518-2989
Väino Sammelselg https://orcid.org/0000-0002-3598-9184
References
1. C. R. Brooks, Heat Treatment, Structure and Properties of Nonferrous Alloys
(American Society for Metals, Metals Park, Ohio) p. 121 (1982).
2. J. R. Davis, Aluminum and Aluminum Alloys (ASM International, Materials Park,
OH, USA) ASM Specialty Handbook., (1993), and Davis J.R., Corrosion of
Aluminum and Aluminum Alloys (ASM International, Materials Park, OH, USA)
ASM Specialty Handbook. (1999); and Kaufman G., Corrosion of Aluminum and
Aluminum Alloys. In: Corrosion: Materials, Cramer, S.D. and Covino B.S. Jr.
(Eds.) (ASM International, OH, USA), 95 (2005).
3. K. H. Matucha, “Structure and properties of nonferrous alloys.” Mater. Sci.
Technol. (Weinheim Publisher, New York, NY) 8, 213 (1996).
4. G. E. Totten and D. Scott MacKenzie, Handbook of Aluminum, volume 2, Alloy
production and Materials Manufacturing (CRC Press, New York, NY) (2003).
5. [dataset] Material Property Data (ASM MatWeb), http://asm.matweb.com/search/
SpecificMaterial.asp?bassnum=MA2024T3, and (http://matweb.com/search/DataSheet.
aspx?MatGUID=781ce4adb30c4d548320b0ab262a5d28); last visited in June 2022.
6. B. D. Dunn, The Corrosion Properties of Spacelab Structural Alloy Aluminum
2219-T851 (ESA STR-212). (ESA Communication Production Office, ESTEC,
Noordwijk, The Netherlands) (1984), http://esmat.esa.int/Publications/Published_papers/
ESA_STR-212.pdf; last visited in June 2022.
7. A. M. Pereira, G. Pimenta, and B. D. Dunn, Assessment of Chemical Conversion
Coatings for the Protection of Aluminum Alloys A Comparison of Alodine 1200 with
Chromium-Free Conversion Coatings (ESA-STM-276, 2008). (ESA Communication
Production Office, ESTEC, Noordwijk, The Netherlands).
8. EU hydrogen policy. Hydrogen as an energy carrier for a climate-neutral economy,
https://europarl.europa.eu/RegData/etudes/BRIE/2021/689332/EPRS_BRI(2021)
689332_EN.pdf; last visited in June 2022.
9. S. Kramer and D. Mosher. Here’s how much money it actually costs to launch stuff
into space. Tech Insider, https://businessinsider.com/spacex-rocket-cargo-price-by-
weight-2016-6; last visited June 2022.
10. O. Zeynali, D. Masti, and S. Gandomkar, Adv. in Appl. Sci. Research, 3, 446 (2012),
(https://www.primescholars.com/articles/shielding-protection-of-electronic-cir-
cuits-against-radiation-effects-of-spacehigh-energy-particles.pdf); last visited in
June 2022.
11. A. Boag, A. E. Hughes, N. C. Wilson, A. Torpy, C. M. MacRae, A. M. Glenn, and
T. H. Muster, Corros. Sci., 51, 1565 (2009).
12. A. Boag, A. E. Hughes, A. M. Glenn, T. H. Muster, and D. McCulloch, Corros.
Sci., 53, 17 (2011).
13. A. E. Hughes, A. Boag, A. M. Glenn, D. McCulloch, T. J. Muster, C. Ryan, C. Luo,
X. Zhou, and G. E. Thompson, Corros. Sci., 53, 27 (2011).
14. A. M. Glenn, T. H. Muster, C. Luo, X. Zhou, G. E. Thompson, A. Boag, and A.
E. Hughes, Corros. Sci., 53, 40 (2011).
15. A. Merstallinger, M. Sales, E. Semerad, and B. D. Dunn, “Assessment of Cold
Welding Between Seperable Contact Surfaces Due to Impact and Fretting Under
Vacuum..” Materials Science (European Space Agency, Noordwijk, Netherlands)
European Space Agency Scientific and Technical Reports, STM-279 (2009).
16. S. Borenstein, (2022), Phys.org, https://phys.org/news/2009-05-stuck-dead-battery-
bedevil-hubble.html Stuck bolt, dead battery bedevil Hubble repairs; last visited in
June 2022.
17. A. de Rooij, “Corrosion in Space.” Encyclopedia of Aerospace Engineerings,
ed. R. Blockley and W. Shyy (Wiley, https://onlinelibrary.wiley.com/) (2010),
https://onlinelibrary.wiley.com/doi/abs/10.1002/9780470686652.eae242http://
esmat.esa.int/publications/published_papers/corrosion_in_space.pdf; last visited in
June 2022..
18. K. K. de Groh, B. A. Banks, J. A. Dever, D. A. Jaworske, S. K. Miller, E.
A. Sechkar, and S. R. Panko, NASA Glenn Research Center’s Materials
International Space Station Experiments (MISSE 1–7). NASA/TM–2008-215482,
published under NASA Glenn Research Center’s Materials International Space
Station Experiments (MISSE 1-7) - NASA Technical Reports Server (NTRS):
Microsoft Word - E-16690TM.doc (nasa.gov); last visited in June 2022.
19. D. R. Gabe, Met. Finish., 100, 52 (2002).
20. A. D. Juhl, Met. Finish., 108, 20 (2010).
21. S. Shresta and B. D. Dunn, “Advanced plasma electrolytic oxidation treatment for
protection of light weight materials and structures in a space environment.”
Advanced Surface Treatment. (European Space Agency, Noordwijk, Netherlands)
(2022), can be found under (http://esmat.esa.int/Publications/Published_papers/
Keronite2007paper.pdf); last visited in May 2022.
22. M. Ritala and M. Leskelä, “Atomic layer deposition.” Handbook of Thin Films, 1,
103 (2002).
23. E. Marin, A. Lanzutti, F. Andreatta, M. Lekka, L. Guzman, and L. Fedrizzi, Corros.
Rev., 29, 191 (2011).
24. S. E. Potts et al., J. Electrochem. Soc., 158, C132 (2011).
25. V. Sammelselg, L. Aarik, and M. Merisalu, WO 2014102758 A1 (2012), EP 2 938
758 B1 (2016), US 9,834,849 B2 (2017), JP 63336477 B2 (2018).
26. V. Sammelselg, I. Netšipailo, A. Aidla, A. Tarre, L. Aarik, J. Asari, P. Ritslaid, and
J. Aarik, Thin Solid Films, 542, 219 (2013).
27. M. Merisalu, L. Aarik, J. Kozlova, H. Mändar, A. Tarre, and V. Sammelselg, Surf.
and Coat. Technology, 411, 126993 (2021).
28. M. Merisalu, L. Aarik, J. Kozlova, H. Mändar, A. Tarre, H. M. Piirsoo, and
V. Sammelselg, Surf. and Coat. Technology, 435, 128240 (2022).
29. J. Aarik, A. Aidla, A. Kasikov, H. Mändar, R. Rammula, and V. Sammelselg, App.
Surf. Sci., 252, 5723 (2006).
30. R. Ali, M. R. Saleem, P. Pääkkönen, and S. Honkanen, Nanomaterials, 5, 792
(2015).
Journal of The Electrochemical Society, 2022 169 071503
31. Testing of materials with atomic oxygen at the Low Earth Orbit Facility (LEOX),
https://esa.int/Enabling_Support/Space_Engineering_Technology/Atomic_oxygen_
generator_simulates_fire_in_the_sky; last visited in June 2022.
32. Y. Zhang, J. A. Bertrand, R. Yang, S. M. George, and Y. C. Lee, Thin Solid Films,
517, 3269 (2009).
33. P. Zoltowski, J. Electroanal. Chem., 443, 149 (1998).
34. Rust Stop paint, https://motipdupli.com/en/products/dupli-color/rust-protection/
rust-stop/ipg-1270/tm-1270.html; last visited May 2022..
35. CompoundChem, Chemical composition of highlighter colors, can be found
under; last visited in June 2022, https://compoundchem.com/2015/01/22/high-
lighters.
36. Data from a private industrial partner.
37. R. Matero, M. Ritala, M. Leskelä, T. Salo, J. Aromaa, and O. Forsen, J. Phys. IV
France, 9, 493 (1999).
38. E. Härkönen et al., ACS Appl. Mater. Interfaces, 6, 1893 (2014).
39. S. M. George, Chem. Rev., 110, 111 (2010).
40. E. Marin, L. Guzman, A. Lanzutti, L. Fedrizzi, and M. Saikkonen, Electrochem.
Commun., 10, 2060 (2009).
41. P. Poodt, A. Lankhorst, F. Roozeboom, K. Spee, D. Maas, and A. Vermeer, Adv.
Mater., 22, 3564 (2010).
42. D. Muñoz-Rojas, T. Maindron, A. Esteve, F. Piallat, J. C. S. Kools, and
J.-M. Decams, Mater. Today Chem., 12, 96 (2019).
Journal of The Electrochemical Society, 2022 169 071503