ArticlePDF Available

Ultra-high oxidation resistance of nano-structured thin films

  • Plansee Composite Materials GmbH
  • Oerlikon Surface Solutions

Abstract and Figures

Diffusion driven high-temperature oxidation is one of the most important failure mechanisms of protective thin films in industrial applications. Within this study, we investigated the diffusion of oxygen at 800 to 1100 °C through nano-laminated crystalline Ti-Al-N and amorphous Mo-Si-B based multilayer coatings. The most prominent oxygen diffusion pathways, and hence the weakest points for oxidation, were identified by combining ¹⁸O tracer diffusion and atom probe tomography. An oxygen inward diffusion along column boundaries within Ti-Al-N layers in front of a visually prevalent oxidation front could be proven, highlighting the importance of these fast diffusion pathways. Furthermore, the amorphous Mo-Si-B layers act as barriers and therefore mitigate the migration of oxygen by accumulating reactive O species at a nanoscale range. Preventing oxygen diffusion along column boundaries – through the implementation of amorphous interlayers – lead to paralinear oxidation behavior and stable scales even after 7 h at 1100 °C. Our results provide a detailed insight on the importance of morphological features such as grain and column boundaries during high-temperature oxidation of protective thin films, in addition to their chemistry.
Content may be subject to copyright.
Ultra-high oxidation resistance of nano-structured thin lms
E. Aschauer
,T. Wojcik
, P. Polcik
, O. Hunold
,M. Arndt
, V. Dalbauer
, P.H. Mayrhofer
P. Felfer
, H. Riedl
Christian Doppler Laboratory for Surface Engineering of high-performance Components, TU, Wien, Austria
Institute of Materials Science and Technology, TU Wien, 1060 Wien, Austria
Plansee Composite Materials GmbH, 86983 Lechbruck am See, Germany
Oerlikon Balzers, Oerlikon Surface Solutions AG, 9496 Balzers, Liechtenstein
Department of Materials Science, Friedrich-Alexander-Universität Erlangen-Nürnberg, 91058 Erlangen, Germany
Ultra high oxidation resistance
Blocking fast diffusion pathways
through amorphous interlayers
O tracer diffusion for highly resolved
diffusion pathway analysis using atom
probe tomography
Identication of fast diffusion pathways
in fcc-(Ti,Al)N protective coatings
Atom probe tomography of nano-scaled
(Ti,Al)N/Mo-Si-B multilayer coating.
abstractarticle info
Article history:
Received 22 September 2020
Received in revised form 13 January 2021
Accepted 15 January 2021
Available online 19 January 2021
Oxidation Mechanism
Fast Diffusion Pathways
Tracer Diffusion
Diffusion driven high-temperature oxidation is one of the most important failure mechanisms of protective thin
lms in industrial applications. Within this study, we investigated the diffusion of oxygen at 800 to 1100 °C
through nano-laminated crystalline Ti-Al-N and amorphous Mo-Si-B based multilayer coatings. The most prom-
inent oxygen diffusion pathways, and hence the weakest points for oxidation, were identied by combining
tracer diffusion and atom probe tomography.An oxygen inward diffusion along column boundaries within Ti-Al-
N layers in front of a visually prevalent oxidation front could be proven, highlighting the importance of these fast
diffusion pathways. Furthermore,the amorphous Mo-Si-Blayers act as barriers and therefore mitigate the migra-
tion of oxygen by accumulating reactive O species at a nanoscale range. Preventing oxygen diffusion along col-
umn boundaries through the implementation of amorphous interlayers lead to paralinear oxidation
behavior and stable scales even after 7 h at 1100 °C. Our results provide a detailed insight on the importance
of morphological features such as grain and column boundaries during high-temperature oxidation of protective
thin lms, in addition to their chemistry.
© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
1. Introduction
withinmaterials and responsible for specicph enomena. Some prominent
examples are recovery, recrystallization, precipitation hardening, as well
as phase transformations in general, but also complex procedures such as
spinodal decomposition, nucleation and growth, or oxidation [16].
Diffusionisdeterminedby parameterssuch as gradientsin concentra-
tion (usedfor empirical Fick´srst and second laws [2],although it is the
chemical potential gradient that counts), electric elds, magnetic elds,
elastic strain elds, or morphology (by providing fast diffusion path-
ways). Especially, oxide layer formation on metal surfaces is a textbook
example for applying Ficksrst law, describing the diffusion-controlled
Materials and Design 201 (2021) 109499
Corresponding author.
E-mail address: (E. Aschauer).
0264-1275/© 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (
Contents lists available at ScienceDirect
Materials and Design
journal homepage:
scalegrowth. Tammann[7] mentioned thefamous parabolic growthrate
for scaleson metals alreadyin 1920. Later,the parabolic ratewas derived
from Fick´s law by Pilling and Bedworth [8] parabolic law. Considering a
simplied model(all metal oxides are of ionic nature), the transport ki-
netic andhence oxide scalegrowth are dominated by carrying metalcat-
ions or vacancies, electrons and holes, as well as oxygen anions (O
[9,10]. If the formed oxide is non-volatile and effectively resistant to the
transport of thesespecic species,the ongoing scale growth is restrained
by the slowest transport process with respect to the reaction zone. A.
Atkinson thoroughly summarized the fundamental mechanisms during
oxide lm growth of metal surfaces at elevated temperatures and
highlighted the importance of the prevalent microstructure [11]. In par-
ticular,thediffusionalong so-calledshort-circuitpaths e.g.grainbound-
aries, dislocations, or voids(extended defects) is most signicant if the
temperature,T, issufcientlybelowthe meltingtemperature,T
ing T < 0.6·T
[12,13]. Especially, the diffusion along grain boundaries
was investigated in the past using atom probe tomography (APT) and
tracer diffusion, highlightingtheir importance to describe the migration
ofcertainspeciesthougha material[1416].In particular,the accelerated
diffusion of oxygen along grain boundaries was shown for catalytic sur-
faces of Ir-Ru-O based materials [17]. This prerequisite is especially
true duringoxidation of Al and Cr containing protective coatings such as
N forming Al
and Cr
scales, respectively
[18,19]. Caplan and Sprouleemphasized the important character of oxy-
geninward diffusionalonggrainboundaries[20].Nevertheless,the inter-
play of the prevalent phases and amount of grain boundary interior with
respectto the oxidation resistance ofa protectivecoating is stilldiscussed
[2123].However, such coatings are typically applied in temperature re-
gimes between 700 and 1200 °C,whereas in the so-called high tempera-
ture area (900 °C) the kinetics are highly accelerated and, hence, the
oxidation protection is insufcient [24]. In addition, considering three
stages [24] during oxidation (i) transient stage, initial oxide formation
of the available metals within a compound, (ii) steady state stage,
diffusion controlled oxide scale growth, and (iii) breakaway oxidation,
uncontrolledoxide formationof non-continuous scales anin-depth un-
derstanding of the steady state regime is decisive for designing novel ox-
idation resistant coating materials. Especially, the interplay between
stress formation (growth and thermal stresses) and prevailing diffusion
processes (chemical depletion zones, similar to the Kirkendall effect
[25]),but also theinuenceof voids and cavitiesalong fastdiffusionpath-
ways still lacks experimental proof [9,26,27]. Here, the concept of oxide
formation within a growing scale is relatively unexplored, and the local
oxidationofgrain-boundaryinteriorwithinthin lmmaterialsexhibiting
grain sizes in the range of 10 to 50 nm depicts still uncharted territory
Therefore, we introduced a specic, nanostructured thin lm to de-
scribe oxidation mechanisms along fast diffusion pathways (mainly col-
umn boundaries within thin lms), see the schematics in Fig. 1aandb.
The used multilayer coating is composed of highly crystalline layers,
obtaining a well-known columnar morphology based on competitive
growth during synthesis [29], as well as an amorphous laminate, only
exhibiting very small nanocrystalline domains, as depicted in Fig. 1b.
This concept features several positive aspects not only for a retarded
scale growth itself, but also for a subsequent structural and analytical
characterization of the morphology using high-resolution techniques.
In principle, an array of crystalline and amorphous like layers depicts a
strong barrier for diffusion driven processes, such asoxide scale growth,
due to the annihilation of fast diffusion pathways [3032]see Fig. 1b.
We therefore used our recently developed Ti-Al-N/Mo-Si-B multilayer
coating (comprising alternating arc evaporated face centered cubic
(fcc) structured Ti
N and magnetron sputtered amorphous
layers see Fig. 1 [33]), which was proven to have un-
precedented oxidation resistance [3337]. In various isothermal but
also dynamic oxidation treatments, the selected Ti-Al-N/Mo-Si-B multi-
layer coatings resist temperatures up to 1115 °C, before a complete
breakdown emerges [38] (depicting a relatively high value with respect
to the chemistry and coatingthickness [39]). Furthermore, the recurring
nano-laminated architecture allows for excellent reference points dur-
ing characterization, especially concerning the local grain boundary in-
terior and thermal or chemical driven phase formations.
To trace oxygen diffusion pathways in such a nanostructuredcoating
material, we initially used two different methods, each well-established
for further characterization procedures, but rarely combined in such a
manner. First of all,
O tracer diffusion, which is a well-known tool
for describing diffusion pathways in oxide based materials such as pe-
rovskite typed electrolytes in fuel cells [4043]. Through a combined
conventional (
O) as well as tracer (
O) oxidation treatment, we
preconditioned our growing scale, and consequently stain the predom-
inant diffusion pathways.
In further consequence, to gain a sufcient spatial resolved chemical
distribution (in the range of 1 nm or smaller), we used APT for a quali-
tative as well as quantitative analysis. Our present work provides an
in-depth analysis of oxygen diffusion in a nanostructured Ti-Al-N
based multilayer system by combining
O tracer diffusion with APT,
highlighting the importance of an effective control and minimization
of oxygen diffusion.
2. Materials and Methods
The multi-layered Ti-Al-N/Mo-Si-B coatings were deposited in an in-
dustrial scale Oerlikon Balzers INNOVA deposition system onto (100)
oriented silicon, (11; 02) oriented sapphire, and polycrystalline Al
substrates as well as on low alloyed steel foil (which was subsequently
dissolved in hydro chloric acid in order to obtain free-standing, ake-
like coating material). The lms were grown using powder metallurgi-
cally manufactured Ti
and Mo
targets (Plansee
Composite Materials GmbH). The alternating Ti-Al-N/Mo-Si-B architec-
ture was realized by covering the respective sources with shutters. The
N layers were arc evaporated in a nitrogen containing atmo-
sphere (1000 sccm N
, purity 5.0) at a bias potential of -65 V, whereas
the Mo-Si-B layers were sputter deposited in owing argon atmosphere
(500 sccm Ar, purity 5.0) at -40 V, respectively. Additionally, we also
prepared an improved multilayer (with an optimized bilayer period),
using a bias potential of -20 V and a reduced cathode current on the
Ti-Al targets of 50 A during the deposition of the Mo-Si-B layers while
keeping the other deposition parameters constant.
The substrates were carried by a two-fold rotating carousel (ca. 1.6
rpm) at a substrate temperature of 500 °C. Based on this process, a
nanolayered architecture with a total bilayer period (Λ) of roughly
130 nm could be realized, comprising 16 bilayers with λ
100 nm and λ
=30nmsee also Fig. 1c to e. Further details on
the coating synthesis can be found in [33,3538].
The transport mechanisms of oxygen during high temperature oxi-
dation in the nanostructured Ti-Al-N/Mo-Si-B multilayer system was
analyzed by pre-annealing the coating in pure
O for 30 min in order
to form an adherent oxide scale and subsequently aging in
60 min at 900 °C. For cross-sectional bright-eld transmission electron
microscope (BF-TEM) and high-resolution TEM (HR-TEM) investiga-
tions, a 200 kV FEI TECNAI F20 S-TWIN TEM was used. For the cross-
sectional TEM analysis, the coated Si substrate (as deposited state)
was mechanically polished down to < 30 μm and subsequently ion
beam milled to electron transparency, using a Gatan PIPS machine. In
addition,the crystalstructure was analyzed on a relatively small length
scale using selected-area electron diffraction (SAED) and fast Fourier
transformation (FFT). APT samples were extracted by a focused ion
beam (FIB) lift-out in the as-deposited and oxidized state. However,
the as-deposited APT specimens were obtained by using a keyhole tech-
nique [44], which places the analysis axis in the growth direction ofthe
lm. In contrast, due to weak lm/substrate adhesion of the oxidized
lms whichwere deposited on sapphire, the oxidized samples were ex-
tracted perpendicular to the growth direction. This also allowed us to
capture a larger area of the respective layers in the sample volume. To
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
investigate the diffusion of
O underneath the prevalent oxide scale for
tracing the preferred pathways, three samples were taken at various
depths (see also graphical abstract). All sample sharpening operations
were undertaken at 8 keV ion beam energy, resulting in Ga damage
free datasets ([45], the Ga containing regions were discarded). The
APT analysis was carried out on a Cameca LEAP 4000X HR in pulsed
laser mode. This instrument is equipped with a 355 nm UV laser with
a spot size of ~2 μm and a reection lens resulting in a detection ef-
ciency of ~37 %. The experiments were done with a laser pulse energy
of 50 pJ at a target evaporation rate of 1 %.
During data set analysis, specicdifculties arise from the fact that
some of the occurring ionic species share the same mass-to-charge
ratio (m/c) [46]. For the consideration of the oxidation, this is most no-
tably the main naturally occurring isotope of oxygen,
O, which e.g.
overlaps with Ti
. Therefore, a conclusive analysis of the presence of
O before oxidation in
O is not possible.
also shares its m/c
with 1H
at 18 D, which can be present in small amounts as a
contaminant during APT experiments from the sample transfer in the
air. We therefore compared the presence of the 18 D peak in the data
set from the as-deposited lms to the oxidized material. This peak
yield less than 0.02 at% in the Mo-Si-B layer in any of the as-deposited
samples and is signicantly higher than > 0.04 at% in the oxidized
state at all height positions within the coating.
Finally, we derived specic strategies in order to improve the
proposed coating architecture based on the ndings of this study. The
improved multilayer system was than investigated with respect to its
oxidation resistance by analyzing free-stranding coating material
using a Netzsch STA 449 F1 Jupiter differential scanning calorimeter, es-
pecially focusing on the mass evolution due to oxidation. The samples
were isothermally exposed to a mixed atmosphere containing synthetic
air and He at temperatures of 800, 900, and 1000 °C for 7 h. An exclu-
sively isothermal oxidation treatment was guaranteed by keeping the
heating segment in inert He atmosphere. Additionally, we also oxidized
selected multilayers under the same conditions on hard substrates
(polycrystalline Al
platelets) at 1000 and 1100 °C.
3. Results and Discussion
3.1. Structure, morphology, and chemical constitution
The coatingconsists of 16 bilayers in total with a uniform bilayer pe-
riod of Λ=130nm,seeFig. 1c and d. This architecture is divergent to
the optimized coating thickness with respect to the oxidation resistance
based on our former studies as well as the mechanical properties
(consisting of 31 nm thin Ti-Al-N layers and 6 nm thin Mo-Si-B layers,
[35]). Still, this coating architecture provides a stable and adherent
oxide scale at 900 °C after 90 min,whereas Ti-Al-N of comparable thick-
ness and chemistry is already fully oxidized after a similar oxidation
treatment [24]. However, applying a signicantly larger bilayer period
tunes the architecture towards an optimized system for a detailed anal-
ysis of diffusion pathways by combining tracer diffusion experiments
with APT analysis.
Fig. 1. Schematic description of the multilayer system, consistingof fcc-structured,crystalline Ti-Al-N layers as well as amorphousMo-Si-B layers (a and b). The cross-sectional bright-eld
TEM (c and d) and HR-TEM (e) micrographs show the alternating stacked Ti-Al-N and Mo-Si-B layers in the as-deposited state, and reveal the V-shaped, crystalline nature of the Ti-Al-N
layers, highlighted with dashed lines in (d). The inset in (d) gives the area for the HR-TEM image in (e).
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
The Ti-Al-N layers crystallize in competitive growth as vertical,
V-shaped columnar grains, as it is typically observed for PVD deposited
thin lms (highlighted by dashed lines in Fig. 1d) [47]. For thismultilayer
system, the morphology is strongly inuenced by the incorporated Mo-
Si-B layers, which force the Ti-Al-N to renucleate after each bilayer, see
HR-TEM image in Fig. 1e. This limits on one hand, the crystallite size of
the Ti-Al-N to the respective layer thickness (Fig. 1d) and suppresses, on
the other hand, the evolution of continuous grain boundaries, ranging
from the substrate to the top of the coating. An effective interruption of
these structural defects depicts a crucial condition for a retarded kinetic
of fast-diffusion pathway dominated oxidation processes. In contrast,
the morphology of the Mo-Si-B layers appears mostly featureless, with
no pronounced, oriented growth direction (Fig. 1d), nor dened long-
range order (Fig. 1e), suggesting a nanocrystalline or even amorphous
state. However, the repeatedly occurring, layered structured inhomoge-
neitieswithin the Mo-Si-B layersare also indicated in Fig. 1d. These inho-
mogeneities are caused by the specicdeposition process design.During
the deposition sequence of the Mo-Si-B layer, the Ti-Alcathodes are kept
running behind the shutters (and vice versa for the Ti-Al-N layers), in
order to guarantee a stable synthesis process. Consequently, the
double-fold substrate rotation also passesthe shielded cathodes/targets,
leading to this cross-deposition pattern (for more details see [33,38]).
This effect is muchmore pronouncedfor the Mo-Si-Blayers, since thede-
position rate of the four Ti-Al cathodes is signicantly higher than that of
the two Mo-Si-B targets. Furthermore, the investigated coating system
also exhibits incorporated macro particles (often referred to as droplets),
which originate from the cathode duringarc evaporation [48,49], see top
area in Fig. 1c. However, usually the metallic droplets are also overgrown
by the following layers, which helps to inhibit the oxygeninward move-
ment alongthe coating material/defect interface. The elemental analysis
and density evaluation of the APT data set within one Ti-Al-N layer sug-
gestsfor an evenlydistributed atomiccomposition,Fig. 2a. Consequently,
we expect the material to be present as a homogeneous solid solution.
The evaluation of a cross-contamination of the Ti-Al-N layerwith impuri-
ties originating from the Mo-Si-B compound target is based on the
consideration of Mo, as Si and N share the same mass to charge ratio at
14 and 28 D [50].Following this approach, we couldnd only a negligible
contamination of Mo within the Ti-Al-N layer. A cut of the mass spectra,
highlighting the presence of
O vs.
of the as-deposited state can
be found in Fig. 4a (however, please note that the authors focused here
on the quantication of the most important species rather than on a
fully indexed mass spectra). Furthermore, a detailed analysis of the
Ti-Al-N layer using HR-TEM (Fig.2b) also conrms the nucleation of the
Ti-Al-N crystallites on top of the incorporated Mo-Si-B layers. The corre-
sponding FFT pattern (Fig. 2c) clearly reveals a single phasedcrystalline
fcc structure (see marked area in Fig. 2b),showing a preferred 200 orien-
tation in growth direction with no indications of the wurtzite AlN phase.
A detailedelementalcluster analysis of the APT data set of a Mo-Si-B
layer (Fig. 2d) indicates an alternating occurrence of B and Al (and also
Ti, not shown in the graph), which conrms the cross-deposition of Ti
and Al during the synthesis of the Mo-Si-B layers. Additionally, a de-
tailed HR-TEM analysis of a specic Mo-Si-B layer again reveals an
amorphous-like appearance, showing no long-range order or crystalline
clusters in (e).
This is also conrmed by the corresponding, blurred FFT pattern
(Fig. 2f), which is a strong proof for an amorphous material (see marked
area in Fig. 2e).
In summary,this detailed morphological and structuralanalysis con-
rms the proposed coating architecture and is the prerequisite for a
proper interpretation of the oxygen tracer experiments. In addition, to
rule out any thermally driven inter-diffusion between Mo-Si-B and
Ti-Al-N, vacuum annealing experiments combined with HR-TEM and
X-ray nano-diffraction proved an architectural integrity up to at least
1000 °C [51].
O tracer experiments
In order to identify the most prominent diffusion pathways of oxy-
gen during oxidation and hence the weakest points of such a columnar
coating structure, we oxidized coated sapphire platelets at 900 °C in
Fig. 2. DetailedAPT analysis of a specicTi-Al-N (a) and Mo-Si-B layer (d) in theas-deposited state,showing the elemental distribution of N and Ti as wellas Al and B, respectively.The HR-
TEM micrographs and the corresponding FFT of a crystalline, fcc structured Ti-Al-N layer (b and c) and of anamorphous Mo-Si-Blayer (e and f) reveal thecrystal structure of the two base
systems. For preparational reasons, the analyzed sample volumes originate from different areas.
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
O and subsequently
O atmosphere for 30 and 60 min, respec-
tively. An overview of the formed oxide scale and a visually unaffected
coating structure is given in Fig. 3a. The above described exposure
leads to an oxidation of 0.65 μm out of 2.2 μm in total of the as-
deposited coating thickness. The still intact coating morphology is cov-
ered by spacious Ti- and Al containing oxides, forming characteristic
needle-shaped crystallites, which partially appear slightly undersense.
Nevertheless, within this porous oxide scale, a kind of layered structure
is recognizable see Fig. 3a and c. Based on this visual inspection, the
question arises if the prevalent oxygen partial pressure, either based
O, already causes a pre-oxidation or degradation of the
grain boundaries underneath, being visually unaffected through the ox-
idation. Therefore, the oxygen distribution (especially the
O species)
within the visually intact coating material was determined by preparing
three APT specimens, extracted parallel to the coating/substrate inter-
face, as illustrated in Fig. 3b. This was done by ion milling a free-
standing slice, containing substrate, intact coating and oxide scale, see
also Fig. 3c. An APT specimen ready for measurement, highlighting
the parallel orientation of the layers with respect to the W post is
shown in Fig. 3d. In Fig. 3e-g, a detailed chemical overview of the
Ti-Al-N/Mo-Si-B interfaces is given, asobtained from the three different
specimens, extracted from the oxide scale near region (Fig. 3e), within
the middle regime of the unaffected coating (Fig. 3f), as well as from
the substrate near area, Fig. 3g.
The mass spectra of the as deposited as well as oxidized states are
presented in Fig. 4.Fig. 4a gives the mass-to-charge-state ratio of the
Λ= 130 nm in the as deposited state corresponding to theresults pre-
sented in Fig. 2 whereas Fig. 4b to d correspond to theoxidized states
as shown in Fig. 3e to g, respectively. To highlight the presence of
the oxidized state, but alsoto verify thatthe amount of
in the as de-
posited state is minor, we focused on the region between 8 to 20 Da. The
amount of species arriving at 18 Daduring APT is signicantly higher for
theoxidizedstates,whereastheintensities at 16 Da stays ratherconstant,
compare Fig. 4a to d. The tip which was prepared nearest tothe visually
oxidized area (porous scale), see Fig. 3b and e, obtains a 5 times higher
amount of species at 18 Da compared to the as deposited sample. The
ratio of
is highest for the oxidized, surface-near sample
(red tip),4.50, and decreases with progressing depth of the sample posi-
tion within visually unaffected coatingto 2.96 forthe middle (green tip)
and 2.30 for the bottom tip (blue one, substrate-near position), respec-
tively. In the as deposited state the ratio between
is 0.80.
In summary, the tips of the middle and bottom section see Fig. 4 cand
d corresponding to Fig. 3f and g still have signicant higher amounts
of speciesat 18 Da. Nevertheless, all datasets in Fig. 3etogoftherespec-
tive specimens are cropped in the same way, that a full Ti-Al-N/Mo-Si-B
bilayer is displayed, where the
O atom distribution (blue colored dots,
left side) is compared to the B atom distribution (yellow colored, right
side).Based on this selection,an interdiffusion of the two layered coating
materials as well as preferreddiffusion pathways are distinguishable.For
all threespecimens, the B atomsallow for a sharp distinction between Ti-
Al-N(nearlyB free,leftside) andMo-Si-Blayers(Bcontaining, right side).
In addition, for all specimens, the
O amount
Fig. 3. APT analysis of tracer diffusion of
O in the layered system after subsequent oxidation at 900°C in pure
O for 30 and 60 min, respectively. From theoxidized coating (a),
three APT samples were extracted at various distances to the surfaces (b). Afterwards,they were lifted out and attachedto a support structure(c) and subsequently prepared for the APT
measurement (d). The elemental distribution of
O (blue) and B (yellow) for the surface near region (closest to the visual scale), the middle of the coating and towards the substrate is
presentedin (e), (f), and (g).
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
is signicantly higher within the Mo-Si-B layers than within the
Ti-Al-N layers, see Fig. 3e-g. This suggests a very fast oxygen inward dif-
fusion through Ti-Al-N layers and simultaneously for a high oxygen af-
nity of the Mo-Si-B layers.
Furthermore, a quantitative analysis of the data sets indicates a
decreasing oxygen content from the oxide scale near region to the sub-
strate/coating interface near specimen. Considering the surprisingly
high permeability of Ti-Al-N for
O, the retarding effect of Mo-Si-B on
the oxygen diffusion is evident. However, a more detailed investigation
on the clustering behavior of
O within the Mo-Si-B layers shows
oxygen free areas besides a pathway network (Fig. 5a and b). This sug-
gests a preferential oxygen diffusion along triple junctions(indicated by
green arrows in Fig. 3e to g) between precipitated phases most likely
thermally driven. As shown in our previous study of thermally treated
and oxidized Ti-Al-N/Mo-Si-B coatings using
spatiallyresolved nanobeam X-ray diffraction in transmission geom-
etry [51], the intact coating material underneath the formed oxide scale
does not show any formed oxides, but an evolution of intermetallic
and T
based phases. The crystallization of the as-
deposited amorphous Mo-Si-B is furthermore accompanied by the
Fig. 4. Cut out between 8 to 20 Da of the mass to charge-state ratio of the APT tip of the as deposited coating (a) having Λ= 130 nm (presented in Fig. 1 and 2), and the corresponding
oxidizedAPT tips see also Fig. 3etogin (b) to (d), respectively.To clarify the increasing amount of
, the ranged peaksare highlighted in blue andgreen. After
oxidizing the samples in
O at 900 °C for 60 min (30 min peroxidation in
O), a clear increase at 18 Da is recognizable.
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
decomposition of fcc Ti-Al-N into TiN and hexagonal, wurtzite struc-
tured AlN [52]. Consequently, the thermal load on the coating material
leads to a crystallization of the initially amorphous Mo-Si-B interlayers,
but no pronounced oxide formation. The resulting grain boundary
phase between the precipitated grains generates a continuous network
of triple junctions, which functions as a diffusion pathway for the in-
coming oxygen species.
In order to analyze the movement of
O through the Mo-Si-B layer
during oxidation, we performed detailed clustering analyses of the
specimen extracted from the region directly underneath the oxide
scale. In Fig. 5a, the B and
O concentrations are compared. In both g-
urative representations, B and
O depleted sites share the same areas,
associable with precipitated T
grains. Simultaneously, the
probed volume also exhibits areas with signicantly higher B content,
possibly representing T
crystallites. In between, the
bution forms a dense network with high concentrations next to small
O depleted areas. This suggests for a diffusion of
O between these
precipitates. In Fig. 5b, the elemental distributions of Si and
O are presented, as viewed from a different perspective (about 45°
rotated to the MoSiB layer as depicted in Fig. 5agreen schematic).
Here, strong clustering of Si atoms is evident, indicating again the pre-
cipitation of T
and T
based phases. The areas with
very high Si concentrations are mostly free of
O, which preferably oc-
cupies sites in between. This proves
O diffusion along formed grain
boundaries between thecrystallized T
and T
phases. For a quantitative
evaluation of the
O enriched clusters between the crystalline phases,
we exclusively looked at the area as marked in Fig. 5b and c and calcu-
lated a radial concentration prole with its center in the middle of the
cluster. The
O content is highest in the cluster center (~4 at%) and rap-
idly decreases to ~0.1 % after 2 nm, see Fig. 5c. In Fig. 5d, the atomic con-
tents of B, Si, Mo, and Al are shown, calculated for the same volume
element as in Fig. 5c. The prole of Al and Si follow the trend of
also showing a maximum in the center of ~20 and 38 at%, respectively.
After ~2 nm, the contents of Al (~7 at%) and Si (~20 at%) decrease signif-
icantly. The contents of B and Ti are almost constant at ~19 and ~10 at%,
respectively. In contrast, the Mo fraction is low at high
O containing
sites and steeply increases after passing the 2 nm barrier.
Additionally, we determined the enrichment of the grain boundary
phase with
O, B, Si, Mo, Al, and Ti. Therefore, we compared the average
elementalconcentration within the 2 nm range with theelemental con-
centration of the region of interest, see Fig. 5d. In summary, the data
clearly indicates an increased concentration of
O and Al along formed
grain boundaries as well as a small enrichment of Ti and Si within the
phases arising. Therefore, we expect that Al, originating from the
cross-deposition during the synthesis process (or also from diffusing
Al after the spinodal decomposition of the Ti-Al-N), do not form a
solid solution with the Mo-rich intermetallic phases and, therefore, is
preferably populating the grain boundary phase. Also, the uniform dis-
tribution of Ti within the analyzed volume suggests for the formation
of a T
and T
solid solutions, since Ti is
known as a phase stabilizer for the T
and T
phase [53].
3.3. Improved coating design
These results allow to derive very specicstrategiesinorderto
improve the oxidation resistance and mechanical properties of the
Ti-Al-N/Mo-Si-B multilayer system. As we found during the APT inves-
tigations, the oxygen migrates preferably along triple junctions between
and T
precipitates within the Mo-Si-B layers. Consequently, control-
ling the crystal structure as well as the phase composition and morphol-
ogy in the as-deposited state and upon thermal load potentially
mitigates the oxygen mobility.
Fig. 5. Quantitative analysis of the tip extracted out of the oxide scale near region after subsequent oxidation at 900 °C in
O and
O for 30 and 60 min, respectively. The B and
concentrations are presented in (a) and the
O and Si in (b). The chemical environment of an
O cluster (as marked in (b)) was analyzed by constructing a spherical volume
element. The corresponding elemental proles are given in (c) for
O and (d) for B, Si, Mo, Al, and Ti. By comparing the elemental composition within the 2 nm range with the
average composition in the region of interest, we calculated the enrichment of the cluster with
O, Si, Mo, B, Al, and Ti (e).
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
Our recent investigations on homogeneously grown magnetron
sputtered Mo
thin lms with x = 0.05 and 0.10 at% Ti
showed the best oxidation resistance when applying a bias potential
of -20 V and using the Mo
target composition (while keep-
ing the other parameters as described in this manuscript). This setup
leads to an as-deposited amorphous state and also to the formation of
the T
and T
phase upon vacuum annealing. However, we also found
that, with an increasing bias potential, the crystalline character of
these Mo-Si-B layers increases, apparently reducing the oxidation resis-
tance. Additionally, we could resume that oxygen shows a strong asso-
ciation to Al when migrating through the Mo-Si-B layers, originating
from the specic process design. An explanation for this behavior can
be given by thermodynamics, as Al has one of the lowest Gibbs free en-
ergy for the formation of oxide based compounds [54]. Since the process
stability requires to run all targets permanently a continuous re-
igniting of the arc evaporated Ti-Al cathodes potentially cause process
instabilities the multilayer arrangement was realized by covering
the respective targets with shutters. However, a signicant reduction
of the cathode current potentially reduces the cross-deposition of Al
and Ti within the Mo-Si-B layers.
Based on all these ndings, the architectural design was adapted ac-
cording to ourprevious studies [33,35,38], so that the Mo-Si-B layers are
as thin as possible while a complete interruption of the re-nucleating
Ti-Al-N laminates is still assured. This guaranties excellent mechanical
properties in combination with an effective interruption of continuous
grain boundaries, ranging from the coating/substrate interface to the
surface. Therefore, we deposited an improved multilayer with an opti-
mized bilayer period of Λ~ 32 nm with λ
~ 26 nm and λ
6 nm and a total number of 80 bilayers. An inverted high angle annular
dark eld (HAADF) micrograph of the optimized multilayer is displayed
in Fig. 6a. Here, a 111 diffraction spot was exclusively surveyed with the
selected area electron diffraction (SAED) aperture in order to highlight
the growth mode of the Ti-Al-N grains. The image clearly suggests an
effective interruption of the Ti-Al-N grains and a consistent limitation
of the grain size to the Ti-Al-N layer thickness, assuring the functionality
of the proposed multilayer design. The respective HR-TEM image,
Fig. 6b, again shows the 26 nm thin Ti-Al-N layers and 6 nm thin
Mo-Si-B layers. The micrograph clearly indicates the re-nucleation of
the Ti-Al-N grain as well as their intermittent growth mode, exemplary
highlighted by the dashed line for a single Ti-Al-N grain, see Fig. 6b. Ad-
ditionally, we investigated the crystal structure by means of FFT (not
shown in the graph), nding the Ti-Al-N to be single phased fcc as
well as the Mo-Si-B to be amorphous.
A detailed 3Dchemical analysis of the Ti-Al-N and the Mo-Si-B layer
with respect to the cross-contamination of the optimized multilayer
system wasdone by means of APT. The reconstructed APT specimen, fo-
cusing on the distribution of B as well as 2D elemental maps of Al, Ti, and
B (representing the same region of interest) is shown in Fig. 6c. The data
set indicates very sharp interfaces as well as a constant bilayer period
across the whole length of the tip.
Furthermore, we calculated a line scan across two bilayers, only
considering B and Ti in order to avoid peak overlaps (e.g. Si and N
at 14 D and 28 Dalton as well as BO
and Al at 27 D), see Fig. 6d. Con-
sequently, only the relative elemental composition is shown. Even
though there are still contaminations visible in the Mo-Si-B layer,
the cross-contamination could be clearly reduced (compare Fig. 2
and ref. [33]). The highly concentrated B in the Mo-Si-B layer
shows almost no cross-deposition in the Ti-Al-N layer, being again
representative for very sharp interfaces.
3.4. Oxidation resistance and kinetics
In order to prove the effectiveness of the derived strategies, we in-
vestigated the oxidation performance using thermo-gravimetry in an
oxidative environment (all on free-standing coating material). There-
fore, we compared the multilayer initially investigated in this study
Fig. 6. Inverted HAADF micrograph (a) and HR-TEM image (b) as well as detailed APT investigations (c) of the optimized multilayer system (Λ= 32 nm) in the as-deposited state,
highlighting the total interruption of continuous column boundaries, ranging from the substrate/coating interface to the surface and showing only minor cross-deposition of Ti and Al
within the Mo-Si-Blayer. (d) Shows an EDX line scan across two bilayers by displaying the relative atomic composition of B and Ti, showing the reduced cross-contamination during
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
(Λ= 130 nm) as well as the coating as described in our previous studies
= 6 nm, syn-
thesized with -65 V bias and no reduced cathode current during the Mo-
Si-B synthesis) with the abovementioned optimized version (Λ=32
nm, -20 V bias potential and a reduced cathode current of 50 A during
the Mo-Si-B synthesis). The thermo-gravimetric mass change curves
after isothermal oxidation at 800, 900 and 1000 °C for 7 h are shown
in Fig.s 6a-c, respectively.
After oxidation at 800 °C, Fig. 7a, the multilayer with a bilayer period
of Λ= 130 nm shows a rapid mass gain of ~20 %, followed by a signi-
cant and steady mass loss. A decrease in mass in this temperature range
can be assigned to the formation of volatile MoO
species, as this mech-
anism was described in detail for Mo-Si-B based alloys [52]andinour
previous work based on detailed XRD studies of an oxidized Ti-Al-N/
Mo-Si-B multilayer system [38]. For the other investigated coatings
(Λ=37nmandΛ= 32 nm), the curves are also characterized by an ini-
tial mass gain, but following a paralinear behavior (obtaining a linear
and a subsequent parabolic part), being characteristic for stable scales.
However, the mass gain in this paralinear regime is more pronounced
for the optimized multilayer (Λ= 32 nm) plus 10 % mass gain com-
pared to 5 % for the on e with Λ= 37 nm. Nevertheless, the steadyprole
indicates a strong inuence of the bilayer period and is representative
for a stable oxidation performance, following a paralinear rate law.
At 900 °C, Fig. 7b, the multilayered coating with a bilayer period of
Λ= 130 nm oxidizes immediately after introducing synthetic air and
subsequently experiences a rapid mass loss. The pronounced spike in
the mass gain curve for this bilayer period is possibly related to a
rapid evaporation of the relatively thick Mo-Si-B layers, before the for-
mation of a dense oxide scale can occur. Here, also the kinetic difference
between free-standing coating materials and lms on hard substrates in
oxidative environments is evident, as this coating formed an adherent
scale in the combined
O heat treatments used for the APT in-
vestigations after overall 90 min at 900 °C see Fig. 3a. The two opti-
mized versions multilayers with Λ=37nmandΛ=32nm,
respectively exhibit a very similar paralinear behavior at 900 °C.
Here, no signicant differences in mass gains were detectable between
the two versions.
At 1000 °C, Fig. 7c, the situation changes drastically. The Λ=130nm
multilayer oxidizes immediately, also showing a pronounced spike,
comparable to 900 °C, whereas the Λ= 37 nm multilayer shows a
peak in massgain after 3 h followed by a pronounced decrease, beingin-
dicative for a volatile removal of the formed oxide. Hence, the coating
obtaining inhomogeneities within the amorphous Mo-Si-B diffusion
barrier fails, as short circuit diffusion along the abovementioned Al dec-
orated triple junctions is predominant. In contrast, the perfectly shaped
multilayer system (Λ= 32 nm) still established a dense adherent scale,
exhibiting a paralinear oxide growth modewith an overall mass gain of
only ~8 %.
The oxidation performance was additionally investigated by oxidiz-
ing the discussed multilayer systems, applying the identical environ-
mental conditions as for the thermo-gravimetric analysis. In Fig. 8a,
the λ= 130 nm multilayer is shown after annealing at 900 °C for 1.5
h. If compare this to the Λ= 37 nm multilayer, also oxidized at 900 °C
but for 7 h, a clear enhancement due to the changed architecture is ob-
vious. However, as we also intended to optimize the Ti-Al-N/Mo-Si-B
multilayer regarding the deposition parameters and the chemical integ-
rity, we also oxidized the Λ=37nmandΛ= 32 nm multilayer coating
at 1100 °C for 7 h, see Fig. 8b
and c, respectively. While the Λ=37nm
coating is already fully oxidized after this harsh treatment, the Λ=
32 nm multilayer still shows 4.3 μm remaining coating thickness (out
of 6.0), proving the efciency of the undertaken measures especially,
applying amorphous diffusion barriers to deaccelerate oxygen inward
diffusion along preferred diffusion pathways. These results also nicely
conrm the assessment based on the presented thermo-gravimetric
In order toquantify the oxidation kinetics, we also calculated the oxi-
dation rateconstants for the obtained mass evolution according to refer-
ences [24,55].As the oxide formation either follows a linear or paralinear
rate law,the thermo-gravimetriccurves were divided into Regime 1 (lin-
ear behavior, k
)) and Regime2 (parabolicbehavior, k
ForthecoatinghavingabilayerperiodofΛ= 130nm,weobservedan ex-
clusively linear behavior, followed by a signicant mass lossfor all oxida-
tion temperatures (see Fig. 7a-c). Hence, this bilayer period exhibits the
highest k
values of around 5 · 10
between 800 and 1000 °C.
The multilayers having a bilayer period of Λ=37nmandΛ=32nm
reveal both a paralinear character, therefore k
and k
acoording to Re-
gime 1 and 2 were estimated, respectively. The full dataset of the oxida-
tion rateconstants is given inTable 1. In general, the oxidation kineticsof
both coatings seems to be similar, but the coatingwith Λ=32nm(with-
out any inhomogeneities within the amorphous diffusion barriers) ex-
hibits by far thelowest linear rate constants (e.g. k
value of 8.9 · 10
at 900 °C). Consequently, the oxide growth mode changes instantly to
parabolicbehaviorleading to the formationof dense and adherentscales.
In comparison to previous oxidation studies, e.g. monolithic (Ti,Al,Ta)N
obtaining a value of 2.40 · 10
at 950 °C, the kinetics seems to be
clearlyretarded[24].Furthermore,otherstudies showedthat monolithic
Ti-Al-N and alloyed Ti-Al-X-N failed latest after 5 h at 950 °C [56].
Previousstudies on Mo-Si-B based alloys as bulk material suggest for
an excellent oxidation resistance in the high temperature range. How-
ever, at lower temperatures (< 900-1000 °C), the material is prone to
Fig. 7. Thermo-gravimetric mass-change curves duringisothermallyoxidation in synthetic air at (a) 800, (b) 900,and (c) 1000 °C of Λ= 130nm (blue lines with strokes), Λ=37nm(cyan
lines with crosses), and Λ= 32 nm (yellow lines with squares) multilayer coatings, respectively.
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
the formation of volatile MoO
and B
surface evaporation before a
stable oxidation state can be reached [57]. For the Λ= 130 nm coating
at 900 and 1000 °C, the thickness of the Mo-Si-B layer (being 30 nm)
does not allow for the formation of a dense and stable scale, as it can
be seen in Fig. 3a. However, for smaller bilayer periods, the thin Mo-
Si-B layer and its retarding effect on the oxygen inward diffusion into
the Ti-Al-N layer allow for the formation of a dense scale, leading to a
signicantly improved oxidation resistance and lower oxidation rate
4. Conclusion
Our work proves a signicant oxygen diffusion ahead of a visually
and chemically recognizable oxide scale within investigated Ti-Al-N/
Mo-Si-B multilayers. Moreover, within these highly crystalline Ti-Al-N
layers, nearly no
O species could be found, emphasizing grain bound-
ary diffusion being decisive overbulk diffusion (within the prevalent co-
lumnar grains) for oxidation processes. We could also show the strong
tendency of Mo-Si-B to form oxygen enriched areas, representing a
sponge-like behavior within these nanostructured layers, consequently
slowing down the oxygen inward diffusion.Based on these ndings, we
derived a concept based on minimized oxygen mobility during high-
temperature oxidation. The resulting multilayer yields a stable oxida-
tion state up to 1100 °C for 7 h, obtaining a remaining coating thickness
of 4.3 μmoutof6.0μm in the as-deposited state. In comparison, mono-
lithic Ti-Al-N as well as alloyed Ti-Al-X-N coatings fail latest at 950 °C
after 5 h. The detailed investigations on diffusion pathways during
high temperature oxidation emphasize the importance of fast diffusion
pathwayssuch as grain and column boundaries in addition to the chem-
istry. Hence, the morphology and architecture of coating materials is de-
cisive for oxidation kinetics.
Data availability
The data that support the ndings of this study are available from
the corresponding author upon reasonable request.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inu-
ence the work reported in this paper.
The nancial support by the Austrian Federal Ministry for Digital and
Economic Affairs and the National Foundation for Research, Technology
and Development is gratefully acknowledged (Christian Doppler Labo-
ratory Surface Engineering of high-performance Components). We
also thank for the nancial support of Plansee SE, Plansee Composite
Materials GmbH, and Oerlikon Balzers, Oerlikon Surface Solutions AG.
Part of this project has received funding from the European Research
Council (ERC) under the European Union's Horizon 2020 research and
innovation program under grant agreement No 805065. In addition,
we want to thank the X-ray center (XRC) of TU Wien for beam time as
well as the electron microscopy center - USTEM TU Wien - for using
the SEM and TEM facilities.
[1] G. Gottstein, Physikalische Grundlagen der Materialkunde, Springer, Berlin Heidel-
berg, Berlin, Heidelberg, 2007.
[2] A. Fick, Ueber Diffusion, Ann. Der Phys. Und Chemie. 170 (1855) 5986, https://doi.
[3] G.D. Preston, Structure of Age-Hardened Aluminium-Copper Alloys, Nature. 142
(1938) 570,
[4] W.D. Kingery, Densication during Sintering in the Presence of a Liquid Phase, I.
Theory, J. Appl. Phys 30 (1959) 301306,
[5] J.W. Cahn, On spinodal decomposition, Acta Metall,
43449102981&outputType=export 1961. (Accessed 6 April 2012).
[6] C.V. Ramana, S. Utsunomiya, R.C. Ewing, U. Becker, V.V. Atuchin, V.S. Aliev, et al.,
Spectroscopic ellipsometry characterization of the optical properties and thermal
Fig. 8. Cross-sectional SEM micrographs of the investigatedmultilayersafter oxidationon Al
substratematerials.(a) and (b
) showingthe Λ= 130 nm (1.5 h) and Λ=37nmmultilayer
(7 h), respectively. The Λ=37nmandΛ= 32 nm multilayer oxidized at 1100 °C for 7 h are shown in(b
) and (c), respectively.
Table 1
Rate constants k
obtained from the DSC mass change curves for 800, 900, and 1100 °C.
Coating # k
·t), k
; [1/s]
800 °C 900 °C 1000 °C
Λ= 130 nm 3.3 · 10
-6.5 · 10
-5.6 · 10
Λ= 37 nm 9.2 · 10
2.5 · 10
7.4 · 10
1.6 · 10
9.8 · 10
Λ= 32 nm 1.3 · 10
4.4 · 10
8.9 · 10
1.9 · 10
4.7 · 10
After an initial linear mass gain, these coating show a signicant mass loss and
therefore do not allow for a determination of the parabolic rate constant k
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
stability of ZrO2 lms made by ion-beam assisted deposition, Appl. Phys. Lett. 92
(2008), 011917,
[7] G. Tammann, Über Anlauffarben von Metallen, Zeitschrift Für Anorg. Und Allg.
Chemie. 111 (1920) 7889,
[8] N. Pilling, R.J. Bedworth, The Oxidaiton of Metals at High Temperatures, Inst. Met. 29
(1923) 529.
[9] G.H. Meier, F.S. Pettit, N. Birks, Interactive mechanisms in the high-temperature ox-
idation of metals, in: High-Temperature Oxid, Elsevier, Sulphidation Process., 1990
[10] P.H. Mayrhofer, R. Rachbauer, D. Holec, F. Rovere, J.M. Schneider, Protective Transi-
tion Metal Nitride Coatings, in: Compr, Elsevier, Mater. Process., 2014 355388,
[11] A. At kinson, Transportprocesses during the growth of oxide lms at elevated tempera-
ture, Rev. Mod. Phys.57 (1985) 437470,
[12] A. Atkinson, R.I. Taylor, The diffusion of Ni in the bulk and along dislocations
in NiO single crystals, Philos. Mag. A. 39 (1979) 581595,
[13] G. Ben Abderrazik, Growth mechanism of Al2O3 scales developed on Fe Cr Al alloys,
Solid State Ionics. 22 (1987) 285294,
[14] S.Piazolo,A. La Fontaine,P.Trimby,S.Harley, L.Yang, R.Armstrong, et al.,Deformation-
induced trace element redistribution in zircon revealed using atom probe tomogra-
phy, Nat. Commun.7 (2016) 17,
[15] M. Vaidya, K.G. Pradeep, B.S. Murty, G. Wilde, S.V. Divinski, Radioactive isotopes re-
veal a non sluggish kinetics of grain boundary diffusion in high entropy alloys, Sci.
Rep. 7 (2017) 12293,
[16] J.T. Cukjati, R.F. Cooper, S.W. Parman, N. Zhao, A.J. Akey, F.A.T.P. Laiginhas, Differ-
ences in chemical thickness of grain and phase boundaries: an atom probe tomog-
raphy study of experimentally deformed wehrlite, Phys. Chem. Miner. 46 (2019)
[17] K. Schweinar, R.L. Nicholls, C.R. Rajamathi, P. Zeller, M. Amati, L. Gregoratti, et al.,
Probing catalytic surfaces by correlative scanning photoemission electron micros-
copy and atom probe tomography, J. Mater. Chem. A. 8 (2020) 388400, https://
[18] O. Madelung,U. Rössler, M. Schulz, Chromium sesquioxide (Cr2O3): thermal expan-
sion, density, me lting point, Non-Tetrahedrally Bond. Bin. Compd. II, Springer-
Verlag, Berlin/Heidelberg 2020, pp. 12,
[19] N. Ishizawa, T. Miyata, I. Minato, F. Marumo, S.Iwai, A structural investigation of α-
Al 2 O 3 at 2170 K, Acta Crystallogr. Sect. B Struct. Crystallogr. Cryst. Chem. 36
(1980) 228230,
[20] D. Caplan, G.I. Sproule, Effect of oxide grain structure on the high-temperature oxi-
dation of Cr, Oxid. Met. 9 (1975) 459472,
[21] F. Vaz, L. Rebouta, M. Andritschky,M.F. da Silva, J.C. Soares, Thermal oxidation of Ti1
xAlxN coatings in air, J. Eur. Ceram. Soc. 17 (1997) 19711977,
[22] L. Chen, J. Paulitsch, Y. Du, P.H. Mayrhofer, Thermal stabilityand oxidation resistance
of TiAlN coatings, Surf. Coatings Technol. 206 (2012) 29 542960,
[23] G. Greczynski, L. Hultman, M. Odén, X-ray photoelectron spectroscopy studies of
Ti1-Al N (0 x0.83) high-temperature oxidation: The crucial role of Al concentra-
tion, Surf. Coatings Technol 374 (2019) 923934,
[24] R. Hollerweger, H. Riedl, J. Paulitsch, M. Arndt, R. Rachbauer,P. Polcik, et al., Origin of
high temperature oxidation resistance of Ti-Al-Ta-N coatings, Surf. Coatings Technol
257 (2014)
[25] F. Seitz, On the porosity observed in the Kirkendall effect, Acta Metall. 1 (1953)
[26] F.H. Stott, The Oxidation of Alumina-Forming Alloys, Mater. Sci. Forum. 251254
(1997) 1932,
[27] F.H. Stott, G.C. Wood, Growth and adhesion of oxide scales on Al2O3-forming alloys
and coatings, Mater. Sci. Eng. 87 (1987) 267274,
[28] T. Li, O. Kasian, S. Cherevko, S. Zhang, S. Geiger, C. Scheu, et al., Atomic-scale insights
into surface species of electrocatalysts in three dimensions, Nat. Catal. 1 (2018)
[29] I. Petrov, P.B. Barna, L. Hultman, J.E. Greene, Microstructural evolution during lm
growth, J. Vac. Sci. Technol. A Vacuum, Surfaces, Film 21 (2003) 117S128,
[30] M.-H. Tsai, C.-W. Wang, C.-H. Lai, J.-W. Yeh, J.-Y. Gan, Thermally stable amorphous
(AlMoNbSiTaTiVZr)50N50 nitride lm as diffusion barrier in copper metallization,
Appl. Phys. Lett. 92 (2008), 052109,
[31] M. Stueber, H. Holleck, H. Leiste, K. Seemann, S. Ulrich, C. Ziebert, Concepts for the
design of advanced nanoscale PVD multilayer protective thin lms, J. Alloys
Compd. 483 (2009) 321333,
[32] M. Stueber, C. Ziebert, H. Leiste, S. Ulrich, C. Sanz, E. Fuentes, et al., WEAR STUDIES
[33] E. Aschauer, S. Sackl, T. Schachinger, T. Wojcik, H. Bolvardi, M. Arndt, et al., Nano-
structural investigation of Ti-Al-N/Mo-Si-B multilayer coatings: A comparative
study by APT and HR-TEM, Vacuum. 157 (2018) 173179,
[34] P.V. Kiryukhantsev-Korneev, I.V. Iatsyuk, N.V. Shvindina, E.A. Levashov, D.V.
Shtansky, Comparative investigationof structure, mechanical properties, and oxida-
tion resistance of Mo-Si-B and Mo-Al-Si-B coatings, Corros. Sci. 123 (2017)
[35] H. Riedl, E. Aschauer, C.M. Koller, P. Polcik, M. Arndt, P.H. Mayrhofer, Ti-Al-N/Mo-Si-
B multilayers: An architectural arrangement for high temperature oxidation resis-
tant hard coatings, Surf. Coatings Technol 328 (2017)
[36] H. Riedl, C.M. Koller, A. Limbeck, J. Kalaš, P. Polcik, P.H. Mayrhofer, Oxidation behav-
ior and tribological properties of multilayered Ti-Al-N/Mo-Si-B thin lms, J. Vac. Sci.
Technol. A Vacuum, Surfaces, Film 33 (2015) 05E129,
[37] H. Riedl, A. Vieweg, A. Limbeck, J. Kalaš, M. Arndt, P. Polcik, et al., Thermal stability
and mechanical properties of boron enhanced Mo-Si coatings, Surf. Coatings
Technol. 280 (2015) 282290,
[38] E. Aschauer, S. Sackl, T. Schachinger, H. Bolvardi, M. Arndt, P. Polcik, et al., Atomic
scale investigations of thermally treated nano-structured Ti-Al-N/Mo-Si-B multi-
layers, Surf. Coatings Technol. 349 (2018) 480487,
[39] D.J. Tallman, B. Anasori,M.W. Barsoum, A Critical Reviewof the Oxidation of Ti2 AlC,
Ti 3 AlC 2 and Cr 2 AlC inAir, Mater. Res. Lett. 1 (2013) 115125,
[40] M. Gerstl, T. Frömling, A. Schintlmeister, H. Hutter, J. Fleig, Measurement of 18O
tracer diffusion coefcients in thin yttria stabilized zirconia lms, Solid State Ionics.
184 (2011) 2326,
[41] G. Bakradze, L.P.H. Jeurgens, T. Acartürk, U. Starke, E.J. Mittemeijer,Atomic transport
mechanisms in thin oxide lms grown on zirconium by thermal oxidation, as-
derived from 18O-tracer experiments, Acta Mater. 59 (2011) 74987507, https://
[42] T.M. Huber, E. Navickas, K. Sasaki, B. Yildiz, H. Tuller, G. Friedbacher, et al., Experi-
mental Design for Voltage Driven Tracer Incorporation and Diffusion Studies on
Oxide Thin Film Electrodes, J. Electrochem. Soc. 164 (2017) F809F814, https://
[43] M.P. Brady, A.V. Ievlev, M. Fayek, D.N. Leonard, M.G. Frith, H.M. Meyer, et al., Rapid
Diffusion and Nanosegregation ofHydrogen inMagnesium Alloys from Exposure to
Water, ACS Appl. Mater. Interfaces. 9 (2017) 3812538134,
[44] M.K. Miller, K.F. Russell, Atom probe specimen preparation with a dual beam
SEM/FIB miller, Ultramicroscopy. 107 (2007) 761766,
[45] J.E. Halpin, R.W.H. Webster, H. Gardner, M.P. Moody, P.A.J. Bagot, D.A. MacLaren, An
in-situ approach for preparing atom probe tomography specimens by xenon
plasma-focussed ion beam, Ultramicroscopy. 202 (2019) 121127,
[46] A. Huguet, A. Menand, Atom-probe determination of interstitial element concentra-
tion in two-phase and single-phase TiAl-based alloys, Appl. Surf. Sci. 7677 (1994)
[47] A. Anders, A structure zone diagram including plasma-based deposition and ion
etching, Thin Solid Films. 518 (2010) 40874090,
[48] J.E. Daalder, Erosion and the origin of charged and neutral species in vacuum arcs
Erosion and the origin of charged and neutral species in vacuum arcs, J. Appl.
Phys. 8 (1975) 1647.
[49] C.M. Koller, J. Ramm, S. Kolozsvári, J. Paulitsch, P.H. Mayrhofer, Role of droplets and
iron on the phase formation of arc evaporated AlCr-oxide coatings, Surf. Coatings
Technol. 276 (2015) 735742,
[50] D.L.J. Engberg, L.J.S. Johnson, J. Jensen, M. Thuvander, L. Hultman, Resolving mass
spectral overlaps in atom probe tomography by isotopic substitutions case of
N, Ultramicroscopy. 184 (2018) 5160,
2017.08.00 4.
[51] E. Aschauer,M. Bartosik, H. Bolvardi, M. Arndt, P. Polcik,A. Davydok, et al., Strain and
stress analyses on thermally annealed Ti-Al-N/Mo-Si-B multilayer coatings by syn-
chrotron X-ray diffraction, Surf. Coatings Technol. 361 (2019) 364370, https://
[52] P.H. Mayrhofer, A. Hörling, L. Karlsson, J. Sjölén, T. Larsson, C. Mitterer, et al., Self-
organized nanostructures in the Ti-Al-N system, Appl. Phys. Lett. 83 (2003)
[53] R. Sakidja, J.H. Perepezko, S. Kim, N. Sekido, Phase stability and structural defects in
high-temperature Mo Si B alloys, Acta Mater. 56 (2008) 52235244, https://doi.
[54] H.J.T. Ellingham, Transactions and communications, J. Soc. Chem. Ind. 63 (1944)
[55] S. Mrowec, A. Stokłosa, Calculations of parabolic rate constants for metal oxidation,
Oxid. Met. 8 (1974) 379391,
[56] H. Asanuma, F.F. Klimashin, P. Polcik, S. Kolozsvári, H. Riedl, P.H. Mayrhofer, Impact
of lanthanum and boron on the growth, thermomechanical properties and oxida-
tion resistance of TiAlN thin lms, Thin Solid Films 688 (2019)
[57] T.A. Parthasarathy, M.G. Mendiratta, D.M. Dimiduk, Oxidation mechanisms in Mo-
reinforced Mo
Si alloys, Acta Mater. 50 (2002) 18571868, https://
E. Aschauer, T. Wojcik, P. Polcik et al. Materials and Design 201 (2021) 109499
... The formation of Si rich oxide scales is preferred by fast track diffusion processes, e.g. along grain or column boundaries [25]. According to the ternary Hf-Si-B phase diagram at 1300 • C [26], HfB 2 exhibits a very low solubility for Si, hence a phase separation after thermal impacts is expected. ...
Full-text available
Within physical vapor deposited Hf-Si-B2±z thin films, selective diffusion-driven oxidation of Si is identified to cause outstanding oxidation resistance at temperatures up to 1500 °C. After 60 h at 1200 °C, the initially 2.47 µm thin Hf0.20Si0.23B0.57 thin film exhibits a dense oxide scale of only 1.56 µm. The thermally induced decomposition of metastable Hf-Si-B2±z leads not only to the formation of Si precipitates within the remaining thin film (related to a non-homogenous Si distribution after the deposition) but also to pure Si layers on top and bottom of the Hf-Si-B2±z coatings next to the excellent adherend SiO2 based scales.
... A similar effect has also been reported in literature for bulk Ti and Zr, which featured a high oxygen solubility [34]. Despite exhibiting a parabolic oxidation rate, the rate comprises a combination of two processes: Oxide scale growth and oxygen dissolution into the metal [35,36]. ...
Full-text available
The influence of the non-metal species on the oxidation resistance of transition metal ceramic based thin films is still unclear. For this purpose, we thoroughly investigated the oxide scale formation of a metal (Hf), carbide (HfC0.96), nitride (HfN1.5), and boride (HfB2.3) coating grown by physical vapor deposition. The non-metal species decisively affect the onset temperature of oxidation, ranging between 550 °C for HfC0.96 to 840 °C for HfN1.5. HfB2.3 and HfN1.5 obtain the slowest oxide scale kinetic following a parabolic law with kp values of 4.97∙10⁻¹⁰ and 5.66∙10⁻¹¹ kg² m⁻⁴ s⁻¹ at 840 °C, respectively. A characteristic feature for the oxide scale on Hf coatings, is a columnar morphology and a substantial oxygen inward diffusion. HfC0.96 reveals an ineffective oxycarbide based scale, whereas HfN1.5 features a scale with globular HfO2 grains. HfB2.3 exhibits a layered scale with a porous boron rich region on top, followed by a highly dense and crystalline HfO2 beneath. Furthermore, HfB2.3 presents a hardness of 47.7 ± 2.7 GPa next to an exceptional low inward diffusion of oxygen during oxidation. This study showcases the strong influence of the non-metallic bonding partner despite the same metallic basis, as well as the huge potential for HfB2 based coatings also for oxidative environments.
... The inclusion of Al was accompanied by a slight reduction in hardness from 42 GPa in the TiB 2.4 thin film, to 39 GPa for the Ti 0.68 Al 0.32 B 1.35 thin film. Reduction in oxidation rate has also been observed in WB x through alloying with Ta [21], with many recent studies expanding the understanding of the oxidation kinetics in hard coatings [22][23][24]. ...
Full-text available
Transition-metal diboride thin films, which have high melting points, excellent hardness, and good chemical and thermal conductivity, severely suffer from rapid oxidation in air. Here, we explore the influence of varying B content and resulting nanostructure change on the oxidation properties of TiBx thin films, with x = 1.43, 2.20, and 2.70. Results show that all as-deposited layers have columnar structure. The column boundaries of as-deposited TiB2.20 and TiB2.70 films grown by direct current magnetron sputtering (DCMS) are B-rich, while the as-deposited TiB1.43 films grown by high-power impulse magnetron sputtering (HiPIMS) show no apparent grain boundary phases and contain Ti-rich planar defects. The oxidation rate of TiB1.43 air-annealed at 400 oC up to 48 h is significantly lower than that of TiB2.20 and TiB2.70 films. The oxidation rate of TiB1.43, TiB2.20, and TiB2.70 films was measured at 2.9±1.5, 7.1±1.0, and 20.0±5.0 nm/h, respectively, with no spallation of even as thick oxide scales as 0.5 μm in any of the films. The improved oxidation resistance can be explained by the absence of B-rich tissue phase at the column boundaries of understoichiometric TiBx films, a phase that interlaces the nanocolumnar TiB2 structures in the corresponding overstoichiometric films. An easy oxidation pathway is thus eliminated.
AlTiN coatings with a NaCl structure have onset point of oxidation around 850°C. To withstand oxidation above 850°C, we studied the influence of Cr addition on the structure, mechanical properties and oxidation resistance of different AlTiCrN coatings.The structural evolution during the annealing of the AlTiCrN coating was revealing that all coatings kept the NaCl structure up to 950°C. The oxidized coatings shown that the addition of Cr promoted the formation of a TiO2 layer over a gradient CrAl2O3 layer.This last layer is very rich in aluminium under the TiO2 layer and richer in chromium at the nitride-oxide interphase.
γ-TiAl alloys have gained much attention due to their low density, excellent fatigue resistance, and high specific strength at elevated temperatures. Throughout this study, we showcase physical vapor deposited Al-rich γ-TiAl coatings as a considerable solution to extend the durability and longevity of γ-TiAl bulk materials in high temperature regimes (> 750 °C). In detail, the oxidation resistance is enhanced up to at least 850 °C for 1000 h, whilst exhibiting a maximum oxygen affected area of 8.4 µm (16 µm coating). Based on various oxidation treatments (800, 850, and 900 °C), we estimated a mean parabolic rate constant of about 1.03∙10⁻¹⁰ cm²/h and a logarithmic oxygen inward diffusion rate of approximately 1.20∙10⁻⁴ cm/h. The corresponding activation energies validate the considerably retarded oxide scale kinetics – particularly, for the thermally grown oxides. Furthermore, the thermal exposure introduced excellent intermixing between the Al-rich γ-TiAl coating and γ-TiAl bulk material, resulting in a coalescence of the coating-substrate interface.
Full-text available
The chemical composition and the electronic state of the surface of alloys or mixed oxides with enhanced electrocatalytic properties are usually heterogeneous at the nanoscale. The non-uniform distribution of the potential across their surface affects both activity and stability. Studying such heterogeneities at the relevant length scale is crucial for understanding the relationships between structure and catalytic behaviour. Here, we demonstrate an experimental approach combining scanning photoemission electron microscopy and atom probe tomography performed at identical locations to characterise the surface’s structure and oxidation states, and the chemical composition of the surface and sub-surface regions. Showcased on an Ir-Ru thermally grown oxide, an efficient catalyst for the anodic oxygen evolution reaction, the complementary techniques yield consistent results in terms of the determined surface oxidation states and local oxide stoichiometry. Significant chemical heterogeneities in the sputter-deposited Ir-Ru alloy thin films govern the oxide’s chemistry, observed after thermal oxidation both laterally and vertically. While the oxide grains have a composition of Ir0.94Ru0.06O2, the composition in the grain boundary region varies from Ir0.70Ru0.30O2 to Ir0.40Ru0.60O2 and eventually to Ir0.75Ru0.25O2 from the top surface into the depth. The influence of such compositional non-uniformities on the catalytic performance of the material are discussed, along with possible engineering levers for the synthesis of more stable and reactive mixed oxides. The proposed method provides a framework for investigating materials of interest in the field of electrocatalysis and beyond.
Full-text available
The resistance to high-temperature oxidation of Ti1-xAlxN films determines performance in numerous applications including coated cutting tools. Here, we present a comprehensive study covering Ti1-xAlxN films with 0 ≤ x ≤ 0.83 annealed in air for 1 h at temperatures Ta ranging from 500 to 800 °C. Layers are grown by the combination of high-power impulse and dc magnetron sputtering (HiPIMS/DCMS) in Ar/N2 atmospheres. We use X-ray photoelectron spectroscopy to study the evolution of surface chemistry and to reconstruct elemental distribution profiles. No dependence of oxidation process on the phase content, average grain size, or preferred orientation could be confirmed, to the accuracy offered by the employed X-ray diffraction techniques. Instead, our results show that, under the applied test conditions, the Ti1-xAlxN oxidation scenario depends on both x and Ta. The common notion of double-layer Al2O3/TiO2 oxide formation is valid only in a limited region of the x-Ta parameter space (Type-1 oxidation). Outside this range, a mixed and non-conformal Al2O3-TiO2 layer forms, characterized by larger oxide thickness (Type-2 oxidation). The clear distinction between different Ti1-xAlxN oxidation scenarios revealed here is essential for numerous applications that can benefit from optimizing the Al content, while targeting a given operational temperature range.
Full-text available
The chemical widths of grain and phase boundaries in deformed wehrlite (olivine–clinopyroxene; Ol–Cpx) aggregates are characterized by atom probe tomography (APT), a (laser-assisted) field evaporation technique employing time-of-flight mass spectrometry. The wehrlite was deformed to high finite strain in diffusion creep: The effective viscosity measured for the wehrlite is approximately an order of magnitude lower than that of either end-member aggregate; further, phase ordering, in which the spatial density of Ol–Cpx phase boundaries increases with accumulated strain, characterizes the deformation (Zhao et al. in Earth Planet Sci Lett 517:83–94, 2019). The mechanical results imply that the transport properties of the phase boundaries—dictated by their structure and composition—differ from those of grain boundaries. Our APT data show that, indeed, the chemical widths of crystalline Ol–Cpx phase boundaries—3.1–6.6 nm, depending on the element used for their characterization—are up to a factor of two greater than the chemical widths of crystalline Ol–Ol and Cpx–Cpx grain boundaries. Careful statistical analyses of the APT data reveal that the near-boundary compositional profiles of the presented Ol–Cpx phase boundary are consistent with—indeed, evidence for—the rheological model in which diffusion creep is rate-limited by the (mechanism-required) interfacial reactions at the Ol–Cpx phase boundaries. Such an analysis is unavailable by current electron beam/X-ray spectrometry approaches, which have not the requisite spatial precision. APT application to nanometer-scale problems in silicate petrology is challenging, particularly because signal overlap is caused by the evaporation of polyion species. We carefully outline the procedures used to acquire and discriminate the data in order to address the challenges of signal overlap.
Full-text available
A method for the rapid preparation of atom probe tomography (APT) needles using a xenon plasma-focussed ion beam (FIB) instrument is presented and demonstrated on a test sample of Ti-6Al-4V alloy. The method requires significantly less operator input than the standard lift-out protocol, is site-specific and produces needles with minimal ion-beam damage; electron microscopy indicated the needle's surface amorphised/oxidised region to be less than 2 nm thick. The resulting needles were routinely analysable by APT, confirming the expected microstructure and showing negligible Xe contamination.
Full-text available
The topmost atomic layers of electrocatalysts determine the mechanism and kinetics of reactions in many important industrial processes, such as water splitting, chlor-electrolysis or fuel cells. Optimizing the performance of electrocatalysts requires a detailed understanding of surface-state changes during the catalytic process, ideally at the atomic scale. Here, we use atom probe tomography to reveal the three-dimensional structure of the first few atomic layers of electrochemically grown iridium oxide, an efficient electrocatalyst for the oxygen evolution reaction. We unveil the formation of confined, non-stoichiometric Ir–O species during oxygen evolution. These species gradually transform to IrO2, providing improved stability but also a decrease in activity. Additionally, electrochemical growth of oxide in deuterated solutions allowed us to trace hydroxy-groups and water molecules present in the regions of the oxide layer that are favourable for the oxygen evolution and iridium dissolution reactions. Overall, we demonstrate how tomography with near-atomic resolution advances the understanding of complex relationships between surface structure, surface state and function in electrocatalysis. Morphological changes in catalyst structure are known to occur during electrocatalysis, and understanding such changes is important to gain insight into the catalytic process. Now, in the case of iridium oxide, these surface changes are probed in atomic-scale detail during the oxygen evolution reaction, and correlated with activity and stability.
State-of-the-art machining applications require new developments in the field of protective coatings, to guarantee for excellent dimensional accuracy and effective processes. Recently, we showed an enormous enhancement in deposition rate, mechanical strength, and thermal stability of Ti–Al–N when using Ce-alloyed Ti–Al composite targets. Within this study, we further follow the concept of alloying lanthanides to nitride-based coatings, by sputtering Ti 0.5 Al 0.5 composite targets alloyed with 2 mol% LaB 6 in mixed Ar/N 2 atmospheres. This allowed us to produce coatings that outperform not only Ti–Al–N (significantly) but also Ce-alloyed Ti–Al–N coatings. Compared with coatings prepared from non-alloyed Ti 0.5 Al 0.5 composite targets (Ti 0.43 Al 0.57 N), we achieved an impressive increase in deposition rate from 77 to 139 nm/min. Furthermore, we observed an increase in hardness from 34 to 40 GPa (on polished polycrystalline Al 2 O 3 substrates) in the as-deposited state, accompanied by a delayed spinodal decomposition, and an additional TiB 2 precipitation, allowing for still 40 GPa after annealing at 1100 °C. Also the oxidation resistance could be significantly improved through the addition of LaB 6 . After exposure to ambient air at 850 °C for 10 h or 950 °C for 1 h, the oxide scale thickness is only 500 nm for the LaB 6 alloyed coatings, whereas the Ti 0.43 Al 0.57 N counterpart was already fully oxidized after 10 h at 850 °C or 1 h at 950 °C.
In order to analyse the main failure mechanism of multilayered coating material in oxidative environments, we separately investigated the cross-sectional strain/stress evolution induced by thermal loads and oxidation for the Ti-Al-N/Mo-Si-B model system. The bilayer period (Λ) of the crystalline Ti-Al-N/amorphous Mo-Si-B layers was varied between 26, 130, 240, and 1085 nm. The stress state was characterised by synchrotron X-ray nano-diffraction, using monochromatic X-ray radiation with a beam size of around 200 × 300 nm². This allows for analysing the spatially resolved strain/stress evolution of the as-deposited state as well as after thermally treated coatings – either 1 h annealed in vacuum or ambient air at 900 °C. For small bilayer periods, the alteration of face centred cubic Ti-Al-N by amorphous Mo-Si-B layers effectively reduces the as-deposited compressive strain profile along the growth direction. Furthermore, for Λ ≤ 130 nm, the decomposition of Ti-Al-N to form hexagonal structured AlN as well as the crystallisation of the Mo-rich layers towards the intermetallics Mo5SiB2 and Mo5Si3 is significantly delayed. After oxidation, the oxide scale grows in the low-compressive stress regime, while the intact multilayer shows similar microstructural changes as the vacuum annealed coatings.
Nano-composites often represent a challenge for spatially resolved analysing methods. The overlap of specific signals, for example in the mass-to-charge spectrum during atom probe tomography (APT) analysis or the intersection of electron energy loss edges in electron energy loss spectroscopy (EELS), leads to highly challenging results. Therefore, a complementary use of different techniques is essential to fully characterise a given system. Within this work, we studied our Ti-Al-N/Mo-Si-B multilayer (comprising alternating 31 nm thin arc evaporated fcc-Ti0.57Al0.43N layers and amorphous 6 nm thin Mo0.58Si0.28B0.14 layers), by high-resolution transmission electron microscopy, EELS as well as APT and completed the results by comprehensive X-ray diffraction measurements. We focused on the microstructure and crystallographic evolution during thermal loading - up to 1400 °C - and hence a detailed analytical description applying a wide set of high resolution techniques. When exposed to high temperatures, the as-deposited amorphous MoSiB layers form the oxidation resistant, intermetallic phases T1-Mo5Si3 and T2-Mo5SiB2. The layered arrangement between Ti-Al-N and Mo-Si-B allows to postpone the formation of the comparative soft, hexagonal wurtzite type AlN (due to the decomposition of fcc-Ti1-xAlxN layers) up to 1200 °C, which is by ∼200–300 °C above the typical formation temperature of w-AlN in homogeneously grown single phase fcc-Ti1-xAlxN thin films.
Abstract Mo-Si-B alloys came into the focus of research activities as an extremely creep resistant material for high temperature oxidative environments. Here, we studied the mechanical, microstructural, and chemical changes of physical vapour deposited protective Ti-Al-N/Mo-Si-B multilayer coatings during annealing in inert and oxidative atmosphere (up to 1600 °C). We focused our investigation on morphological and structural changes of the coating architecture, by applying a set of high-resolution techniques (e.g., high-resolution transmission electron microscopy as well as atom probe tomography), in order to separate temperature or oxygen-induced effects on the protective behaviour. The 31 nm thick Ti-Al-N layers crystallise in the preferred face centred cubic structure, whereas the alternating Mo-Si-B layers are X-ray amorphous. Their thermomechanical properties, with a hardness of at least 31 GPa of the film up to vacuum annealing temperatures of 1200 °C, are clearly superior to homogeneously grown Ti-Al-N coatings. Applying X-ray powder diffraction, we observed that annealing in inert atmosphere leads to the formation of the Mo-rich intermetallics T1-Mo5Si3 (A15, cP8,Cr3 Si-prototype) and T2-Mo5SiB2 (D8l, tl32,Cr5B3-prototype), accompanied by the spinodal decomposition of Ti-Al-N. In contrast, exposing this multi-layered coating to oxygen causes the formation of a rutile-TiO2 and α-Al2O3 oxide scale. The presence of Mo-Si-B is evident up to temperatures of 900 °C, leading to a retarded oxidation process, where only 200 nm out of the 3.0 μm Ti-Al-N/Mo-Si-B multilayer is oxidised.
Die modernen Konzepte der physikalischen Metallkunde sind gleichermaßen grundlegend für das Verständnis aller nichtmetallischen Werkstoffe. Deswegen und der wachsenden Bedeutung der Verbundwerkstoffe wegen liegt es nahe, die klassisch nach den drei Werkstoffen Metall, Keramik und Kunststoff differenzierten Wissensgebiete unter der verbindenden Bezeichnung Werkstoffwissenschaft gemeinsam abzuhandeln. Von dieser Feststellung ausgehend will dieses Lehrbuch zwar zunächst in die Allgemeine Metallkunde einführen, damit und darüber hinaus aber auch die Grundlagen für die gesamten Werkstoffwissenschaften legen. Im Mittelpunkt steht dabei fraglos der naturwissenschaftliche Aspekt der Materialkunde, ohne daß deswegen aber ihr ingenieurwissenschaftlicher Anteil vernachlässigt wurde. Dieses Konzept wird auch in dieser inzwischen dritten Auflage erfolgreich umgesetzt. Modernen Entwicklungen wurde vor allem durch wesentliche Erweiterungen über quantitative Kristallographie Rechnung getragen.