Materials Chemistry and Physics 257 (2021) 123703

Contents lists available at ScienceDirect

Materials Chemistry and Physics

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

Characterization of the nitrided γ-Ti-46Al–2Nb and

γ-Ti-46Al–2Nb-0.7Cr-0.3Si intermetallic alloys
M.N. Mathabathe a, b, **, A.S. Bolokang a, c, *, G. Govender a, C.W. Siyasiya b, R.J. Mostert b
a
Council for Scientific Industrial Research, Materials Science and Manufacturing, Light Metals, Meiring Naude Road, P O Box 395, Pretoria, South Africa
b
Department of Material Science and Metallurgical Engineering, Faculty of Engineering, Built Environment and Information Technology, University of Pretoria, South
Africa
c
Department of Physics, University of the Western Cape, Private Bag X 17, Bellville 7535, South Africa

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Annealed Ti–46Al–2Nb and

Ti–46Al–2Nb-0.7Cr-0.3Si induced γ-TiAl
and α2-Ti3Al.
• The β and Ti5Si3 phases were formed in
Ti–46Al–2Nb alloy and Ti–46Al–2Nb-
0.7Cr-0.3Si alloys.
• Cyclic oxidation of nitrided
Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-
0.3Si was improved.
• Thermal properties of Ti–46Al–2Nb-
0.7Cr-0.3Si were better than
Ti–46Al–2Nb alloy.
• TEM analysis confirmed partial
amorphization in Ti–46Al–2Nb.

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

Keywords: The thermal, surface properties and microstructures of the nitrided γ-Ti-46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si
Nitridation intermetallic alloys were investigated. These vacuum-melted alloys were annealed at 1100 ◦ C to stabilize the
Cyclic oxidation (α2+γ) lamellae structure. Moreover, the annealed alloys were further annealed at 900 ◦ C under nitrogen gas
Intermetallics
(nitridation). The alloys microstructures and nitride surface coatings were characterized. The Ti–46Al–2Nb alloy
XPS
was comprised of γ-TiAl, α2-Ti3Al and β-phases after the annealing, while the Ti–46Al–2Nb-0.7Cr-0.7Si alloy
revealed the Ti5Si3, γ-TiAl and α2-Ti3Al phases. Thin nitride layers formed on both alloys. After cyclic oxidation,
the nitrided alloys exhibited surface spallation at different rates. The X-ray photoelectron spectroscopy (XPS)
revealed the presence of mixed oxide layer of Al2O3, Cr2O3 and SiO2 near the Ti–46Al–2Nb-0.7Cr-0.7Si surface.

* Corresponding author. Council for Scientific Industrial Research, Materials Science and Manufacturing, Light Metals, Meiring Naude Road, P O Box 395, Pretoria,
South Africa.
** Corresponding author. Department of Material Science and Metallurgical Engineering, Faculty of Engineering, Built Environment and Information Technology,
University of Pretoria, South Africa.
E-mail addresses: nmathabathe@csir.co.za (M.N. Mathabathe), sbolokang@csir.co.za (A.S. Bolokang).

https://doi.org/10.1016/j.matchemphys.2020.123703
Received 23 December 2019; Received in revised form 28 April 2020; Accepted 15 August 2020
Available online 31 August 2020
0254-0584/© 2020 Elsevier B.V. All rights reserved.
M.N. Mathabathe et al.
Fig. 1. Schematic illustration of the thermal treatment process of the vacuum melted Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys.
1. Introduction
The titanium-aluminium (Ti–Al) binary alloys are composed of
intermetallic phases such as α-(Ti3Al), (TiAl), TiAl2 and TiAl3 depend­
ing on the alloying and heat treatment employed. In particular, the
γ-TiAl alloys have attracted research interest in the structural engi­
neering applications. Production of these γ-TiAl based intermetallic al­
loys are associated with machining difficulties. These problems led to
extensive research aiming to modify the γ-TiAl intermetallic alloys by
alloying and developments of new ternary, quaternary or quinary
compounds. The alloying elements improves the high-temperature
applicability of Ti–Al alloys especially those developed for aerospace
and automotive industries [1]. The ease to deform and machine the
γ-TiAl intermetallic alloys is enhanced by producing a class of beta
(β)/gamma (γ) alloys with superior properties [2]. Since these alloys
operates at high temperatures, their surface thermal oxidation proper­
ties are crucial. Alternatively, the surface properties of the γ-TiAl
intermetallic alloys are improved by coating the alloy with a
high-temperature resistant layer achievable by nitrogen surface diffu­
sion (nitridation) [3,4]. These surface modifications are commercial for
Ti alloys such as Ti–6Al–4V [5] and γ-TiAl alloys [6–8]. Nonetheless,
Zhao et al. [7], revealed the benefits and limitations of the nitrided
quinary Ti–47Al–2Nb–2Cr–0.2Si intermetallic alloy. It has emerged that
the non-nitrided Ti–47Al–2Nb–2Cr–0.2Si alloys displayed better
oxidation resistance than the nitrided alloys in the range of 800–1000 ◦ C
due to the dominant development of Al2O3 on the outer layer of the
scales [7]. It was noticed that the nitrided alloy exhibited increased
oxidising rate with prolongation of nitridation time at 800 ◦ C but
showed signs of better oxidation resistance at more severe oxidising
conditions when nitrided at 940 ◦ C for 50 h [7]. This alloy disobeyed the
parabolic law at 900 and 1000 ◦ C, which is an indication of the variation
in the uniformity of scale [7]. Furthermore, the Al-rich composition was
detected below the nitride/sub-layer interface whereby TiN and Ti2AlN,
ƞ-Ti3N2-x and Ti2N and Al5Ti2 intermetallic phases were detected [8].
Interventions were made to develop a model reference to the develop­
ment of a TiN scale on a γ-TiAl substrate [9]. This model revealed that
2
Materials Chemistry and Physics 257 (2021) 123703
the applied stress induces large residual stress in the nitride scale owing
to the mismatch of creep rates between the metal and the nitride scale.
The large stress influences the diffusional flux of nitrogen vacancies,
leading to deviation from the parabolic growth law [9]. The aim of the
current work is to investigate the cyclic oxidation properties of the
nitrided γ-Ti-46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si intermetallic al­
loys developed by vacuum melting. Structural analysis and surface layer
characterization are discussed in detail.
2. Materials and experimental procedure
The nominal composition of the γ-based alloys employed in the study
was Ti–48Al–2Nb, and Ti–48Al–2Nb-0.7Cr-0.7Si (at. %). Vacuum arc
melted buttons of 50 g each were initially produced from a proportional
mixture of high purity Ti, Al, Nb, Cr and Si, blended and cold-pressed,
respectively. The melting and casting of the alloys were done accord­
ing to the previous study [10]. Schematic illustration of the experi­
mental procedure is displayed in Fig. 1. The vacuum melted alloys were
homogenized at 1100 ◦ C under Ar atmosphere for 3.6 ks at Ar flow rate
of 1 L per minute (1LPM) before nitridation and cyclic oxidation ex­
periments. These Ti–48Al–2Nb and Ti–48Al–2Nb-0.7Cr-0.7Si alloys
were annealed at 900 ◦ C in a flowing nitrogen gas environment (Nitri­
dation). Other samples were characterized while others were annealed
in air at 900 ◦ C for 900 h for cyclic oxidation test. Nine cycles were
conducted (each cycle represents 100 h) at 900 ◦ C with subsequent
cooling at ambient temperature and returned to the furnace after
recording their mass. For both Ti–48Al–2Nb and
Ti–48Al–2Nb-0.7Cr-0.7Si, one sample for each alloy was used until the 9
cycles were completed. The elaborate prefabrication methodology is
outlined in Refs. [10,11]. Microstructural examinations were adminis­
tered using conventional methods for γ-TiAl based intermetallic alloys,
on as-stabilized, pre-nitridated and cyclically oxidized specimens. The
microstructure and the morphology of the phase structures were
determined by the scanning electron microscope (SEM) which is fur­
nished with energy dispersion spectroscopy (EDS) for microanalysis.
The X-ray photoelectron spectroscopy (XPS) was employed to determine
M.N. Mathabathe et al. Materials Chemistry and Physics 257 (2021) 123703

Table 1 3.1. Annealing of γ-TiAl based alloys

Final chemical composition of the vacuum melted Ti–48Al–2Nb, and
Ti–48Al–2Nb-0.7Cr-0.7Si intermetallic alloys in (at. %). The SEM image of a stabilized ternary γ-Ti-46Al–2Nb alloy is shown
Alloys Final melted alloy composition (at. %) References in Fig. 2a. Dominating phases on the microstructure are γ-(TiAl) and
Elements Ti Al Nb Cr Si
α2-(Ti3Al). The EDS analysis reveal similarities in Table 2 (1100 ◦ C-
Ternary 52.38 45.82 1.8 [11] annealed Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys) and the as-
Quinary 51.68 45.54 1.92 0.73 0.27 [13] cast composition in Table 1, which signify alloy homogeneity. Addi­
tionally, residual β-phase was detected as indicated by the double red
arrow. These phase is enriched with beta-stabilizer Nb in γ-Ti-46Al–2Nb
the varying composition of oxide scales. Atomic force microscope (AFM)
alloy [14]. A microstructure of the Ti–46Al–2Nb-0.7Cr-0.7Si alloy is
was utilized to evaluate the effect of surface roughness and wear on the
shown in Fig. 2b comprising of γ-(TiAl), α2-(Ti3Al) phases and Ti5Si3
nitride/oxidized alloys. For transmission electron microscopy (TEM)
precipitates. Similar to the γ-Ti-46Al–2Nb, the Ti–46Al–2Nb-0.7Cr-0.7Si
analysis, a small spatula tip of the oxidized sample was dispersed in
alloy has lamellar structure with different grain orientations attributed
ethanol to yield a lightly dispersed solution and ultrasonicated for ± 30
to alloying elements and heat treatment. Fig. 2c and d shows the
min at ambient temperature. The copper grid coated with carbon was
as-nitrided Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.7Si intermetallic
dipped into the dispersed solution and left to dry overnight. Observa­
alloys which were processed at 900 ◦ C before cyclic oxidation tests. The
tions were conducted in a JEOL JEM 2100 high-resolution electron
surface nitride layers of both alloys were very thin. Upon nitridation,
transmission microscopy (HRTEM) operating at 200 kV. The macro and
precipitation of phases was observed. EDS chemical composition of
micro-hardness profile values were obtained using a Vickers hardness
tester according to ASTM standard E384-11. Applied loads of 20 Kgf and
500 g were conducted, as reported by Ref. [11,12]. Density measure­ Table 2
ments of the as-stabilized annealed, nitrided and cyclically oxidized EDS spot analysis of the 1100 ◦ C-annealed Ti–46Al–2Nb, and Ti–46Al–2Nb-
specimens were obtained using a pycnometer according to ASTM D854. 0.7Cr-0.3Si alloys (at. %).
Alloy Ti Al Nb Cr Si
3. Results and discussion
TiAlNb 1. 53.55 ± 44.34 ± 2.11 ± – –
0.18 0.08 0.03
The final chemical composition of the as-cast vacuum melted 2. 50.04 ± 48.27 ± 1.69 ±
Ti–48Al–2Nb, and Ti–48Al–2Nb-0.7Cr-0.7Si intermetallic alloys are 0.06 0.10 0.14
shown in Table 1. These results are in agreement with the previous 3. 50.06 ± 48.28 ± 1.66 ±
0.21 0.13 0.05
reporting in Refs. [11,13]. The difference between nominal and final
TiAlNbCrSi 1. 50.92 ± 43.99 ± 1.92 ± 0.57 ± 2.60 ±
composition is due to varying vapor pressures of different components in 0.20 0.17 0.14 0.06 0.21
the alloys, and volatile Al that suffers a burning loss [13]. 2. 50.50 ± 45.79 ± 2.15 ± 0.60 ± 0.96 ±
0.18 0.15 0.09 0.04 0.13

Fig. 2. SEM-SEI micrographs of the as-stabilized γ-TiAl (a) Ternary Ti–46Al–2Nb (at. %), (b) Quinary Ti–46Al–2Nb-0.7Cr-0.7Si (at. %) alloys with their corre­
sponding nitrided images are shown in Fig. 2c and d, respectively.

3
M.N. Mathabathe et al. Materials Chemistry and Physics 257 (2021) 123703

Table 3
The precipitated particles beneath the nitride layer (a) β-phase and (b) Ti5Si3 of the Ti–46Al–2Nb, and
Ti–46Al–2Nb-0.7Cr-0.3Si alloys, respectively.

3.2. Cyclic oxidation behavior of the γ-TiAl alloys

Fig. 3 is the cyclic oxidation curves demonstrating the Ti–46Al–2Nb

Fig. 3. Cyclic oxidation conducted in air at 900 ◦ C for 900 h on the pre-nitrided
Ti–46Al–2Nb and the Ti–46Al–2Nb-0.7Cr-0.3Si alloys. Each cycle represents
100 h.
phases (Fig. 2c and d) circled point 1 is given in Table 3 for both the
ternary and quinary alloy. The composition differ for γ-Ti-46Al–2Nb and
Ti–46Al–2Nb-0.7Cr-0.7Si alloys.
and Ti–46Al–2Nb-0.7Cr-0.3Si alloy behavior at high temperature. The
cyclic oxidation trends of the nitrided alloys compared with those
unnitrided conducted revealed different thermal properties. The unni­
trided Ti–48Al–2Nb alloy initially show a sharp negative mass loss
during the first 100 h cycle. It steadily continues to show a decrease-
increase in mass until the 900 h cycle. This mass loss is indicative of
rapid oxide growth shown by initial positive weight gain and spallation
on the sample surface. Upon extensive exposure times, erratic specimen
weight gain and loss was observed throughout oxidation. It became
stable between the 900–1000 h. In contrast, the nitrided curve of the
Ti–46Al–2Nb alloy shows improved oxidation resistance due to inhibi­
tion of the fast oxide growth during the initial stages of cyclic oxidation.
It implies that the nitride coating played a significant role in prohibiting
the rapid development of the oxide layer despite both samples revealing
a mass loss. The unnitrided Ti–46Al–2Nb-0.7Cr-0.3Si alloy illustrates a
narrow initial negative mass loss compared to the Ti–46Al–2Nb alloy in
the initial 100 h cycle. On the other hand, the nitrided Ti–46Al–2Nb-
0.7Cr-0.3Si alloy shows a steady increase in mass. The show better
performance than the as-cast alloy counterpart. Despite the negative
mass, Ti–46Al–2Nb-0.7Cr-0.3Si alloy performed better than all alloys
until 900 ◦ C. The Ti–46Al–2Nb-0.7Cr-0.3Si alloy is composed of Ti5Si3
phase attributable to better thermal properties. Accordingly, at tem­
peratures slightly below 900 ◦ C, a mixed crystalline TiO2 (rutile) and

Fig. 4. SEM-SEI images and EDS analysis of the surface scales of (a, c) Ti–46Al–2Nb and (b, d) Ti–46Al–2Nb-0.7Cr-0.3Si intermetallic alloys after cyclic oxidation.

4
M.N. Mathabathe et al. Materials Chemistry and Physics 257 (2021) 123703

Fig. 5. AFM surface topography images of the (a–c) Ti–46Al–2Nb and (d–f) Ti–46Al–2Nb-0.7Cr-0.3Si intermetallic alloys, respectively, after cyclic oxidation.

SiO2 scale formed. This scale grows inwardly by diffusion of oxygen presence Ti5Si3 precipitates in Ti–46Al–2Nb-0.7Cr-0.3Si alloy was
along grain boundaries or through oxygen vacancies in rutile structure enhanced by (N) surface diffusion. Furthermore, the
[15]. Moreover, a linear mass gain rate of 0.02 mg/cm2/hr controlled Ti–46Al–2Nb-0.7Cr-0.3Si alloy showed a marginal positive mass gain
growth rate and inner mixed scale of TiOx or TiN and Si might have throughout the 600–900 h range cycle, compared to the Ti–46Al–2Nb
formed [16]. It implies that the oxidation resistance contributed by the alloy at these intervals. This behavior is attributed to Cr contributing to

Fig. 6. XPS survey scan of the (a) Ti46Al–2Nb intermetallic alloy after cyclic oxidation at 900 ◦ C for 900 h, and the corresponding experimental and fitted curves for
normalized (b) Al 2p (c) N 1s, (d) Ti 2p, and (e) O 1s XPS spectra.

5
M.N. Mathabathe et al.
Fig. 7. XPS survey scan of the (a) nitrided Ti46Al–2Nb-0.7Cr-0.3Si intermetallic alloy after cyclic oxidation at 900 ◦ C for 900 h, and the corresponding experimental
and fitted curves for normalized (b) C 1s, (c) Al 2p, (d) Ti 2p, (e) O 1s and.(f) Si 2p.
the oxidation properties [17].
Although nitridation was induced in Ti–46Al–2Nb and Ti–46Al–2Nb-
0.7Cr-0.3Si alloys, it should be noted that static laboratory air is
comprised of a large quantity of nitrogen. Both Ti and Al have higher
affinity for O than N. Therefore, N occupy O vacancies due to structural
instability created by vibrations caused by high temperature on oxides
and forms the titanium oxynitride (TiO2-xNx). The oxynitride layer for­
mation reduces further oxygen diffusion into the surface [11]. The linear
oxidation at high temperature is time-dependent and yields the change
in the weight as described by Eq. (1):
Δm = KL t
Δm is the change in weight monitored with time (t) with a positive kL
when weight is gained [18].
3.3. Characterization of oxide layers
The SEM images of the Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si
intermetallic alloys are shown in Fig. 4a and b. The black, red, yellow,
purple and blue arrows illustrate the EDS composition of the passive
scales. The Ti–46Al–2Nb scale in Fig. 4a confirms the formation of
multiple layers as well as spallation. Dark outer surface scale (black
arrow) peels off, leaving the underneath white scale exposed (red
arrow). The formation of Al2O3 is thermodynamically more feasible than
the formation of TiO2 at 900 ◦ C [11]. The same principle applies to the
Ti–46Al–2Nb-0.7Cr-0.3Si alloy in Fig. 4b. However, due to prolonged
exposure in the air at elevated temperatures, γ-alumina phase absorbs
the nitrogen ions to form a non-stoichiometric AlON compound rich in
defect structures (purple arrow). SEM-SEI images of the Ti–46Al–2Nb
and Ti–46Al–2Nb-0.7Cr-0.3Si surface oxides are displayed in Fig. 4c and
d. The oxide layer of Ti–46Al–2Nb alloy in Fig. 4c shows a
fibrous-morphology with interconnected particles. It is comprised of
valleys and multiple layers formed during oxidation. The oxide film of
the cyclically oxidized Ti–46Al–2Nb-0.7Cr-0.3Si in Fig. 4d revealed a
mixture of a rods and fibrous type structures.
AFM surface topography of the oxide films shown in Fig. 5a–f reveals
Materials Chemistry and Physics 257 (2021) 123703
multiple slip-stepped surface structure comprising different orienta­
tions. The deformed surface layers of the γ-TiAl based alloys are due to
cyclic oxidation. Fig. 5a and b illustrates a slipping pattern with steps
along a hill-valley groove in the Ti–46Al–2Nb surface. These are the
alternating layers of the oxide/oxynitride composition. Fig. 5c clearly
shows slip steps of thin layers. The slip steps are due to deformation
instead of crystallographic planes observed after spalling of the surface
scale. The spectral RMS amplitude of the Ti–46Al–2Nb oxide layer is
106 nm (inset of Fig. 5c) and a maximum fringe height of about 240 nm.
AFM surface topography of the Ti–46Al–2Nb-0.7Cr-0.3Si oxide films is
shown in Fig. 5d–f. On the contrary, the Ti–46Al–2Nb-0.7Cr-0.3Si oxide
(1)
film is comprised of a slip-step pattern of intermixed scale layers of
different morphology (Fig. 5d and e). The 3D topography in Fig. 5e has
peripheral slip fringe with a maximum fringe height of 198 nm. The
spectral RMS amplitude of the Ti–46Al–2Nb-0.7Cr-0.3Si alloy is 70.2 nm
(inset of Fig. 5f) smoother than the Ti–46Al–2Nb surface film. This
behavior is attributed to intermixed layers of oxides. This further sug­
gests that there may be more passive layer formation, which may enable
ease in the restriction of oxygen diffusion in the Ti–46Al–2Nb-0.7Cr-
0.3Si alloy in comparison to the Ti–46Al–2Nb alloy.
3.4. XPS spectra analysis
3.4.1. Ternary Ti–46Al–2Nb alloy
Fig. 6a shows a fully scanned spectrum of the x-ray photoelectron
spectroscopy (XPS) chemical composition of the Ti46Al–2Nb alloy. It
confirms the presence of Ti, Nb, Al, O, N, and C. The detected carbon is
associated with the carbon tape used during the sample preparation
analysis. In addition, the Nb3d and 4p XPS spectrum was observed at
209.8 eV and 50 eV, corresponding to the Nb5+ 3d3/2 and Nb5+ 3d5/2
spin-orbit peaks of the Nb2O5 phase, respectively [19,20]. Fig. 6b–e
displays the XPS spectra of the Ti2p core-level of the nitrided
Ti46Al–2Nb after cyclic oxidation with Fig. 6b showing the peak of
75.38 eV which is associated with Al–O (α-Al2O3) bonds [21]. Another
peak of 397.3 eV shown in Fig. 6c is related to N–Ti–O (TiO2-xNx) bonds
[20–23]. This suggests that more oxygen vacancies are produced during
6
M.N. Mathabathe et al. Materials Chemistry and Physics 257 (2021) 123703

Fig. 8. TEM images of the γ-Ti-46Al–2Nb alloy shown in (a) with a corresponding insert indicated by a zoomed-out area marked with a red circle, (b) elemental
mapping viz. (N, O, Ti) of the coating, (c) the high-resolution faceted morphology indicated by the red circle in Fig. 8a (d) with the corresponding selected area
electron diffraction (SAED). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

annealing for N atoms to occupy the interstitial position in TiO2. Partial
Ti–O bonds on the TiO2 surface break due to annealing and oxygen is
released from the TiO2 lattice creating vacancies for N to occupy. Fig. 6d
shows the peaks centered at around 459.8 eV and 465.5 eV assigned to
Ti4+(2p3/2), and Ti4+(2p1/2) for TiO2 layer [24], while Fig. 6e presents
the O1s core level spectra of the TiO2. Each O1s signal of the TiO2 scale
can be fitted into two symmetric peaks. The peaks located at 531.59 and
533.31 eV corresponds to the oxygen vacancy-Ti4+ and oxygen
vacancy-Ti3+.
3.4.2. Quinary Ti–46Al–2Nb-0.7Cr-0.3Si alloy
Fig. 7 illustrates the XPS high-resolution surface spectra of the cyclic
oxidized Ti46Al–2Nb-0.7Cr-0.3S intermetallic alloy at 900 ◦ C for 900 h.
The Cr 2p spectrum in Fig. 7a is characterized by the Cr 2p 3/2 binding
energy at 576.8 eV. According to Chang et al., [25] this peak is associ­
ated with the presence of Cr2O3. Simialr to Ti–46Al–2Nb alloy, the
detected carbon in Fig. 7b is associated with the carbon tape used during
sample preparation. The peak at 75.38 eV in Fig. 7c is associated with
Al–O (α-Al2O3) bonds [21]. Fig. 7d represents the chemical composition
of Ti, which yielded a pair of spin-orbit viz. (Ti 2p3/2 and Ti 2p1/2) in
the Ti 2p spectrum [21]. The spectrum consisted of broad peaks
considering that the Ti-bonds occurred at the same time. According to
Badini et al. [21], the Ti–O bonds in TiOx, Ti2O3 and TiO2 show bind
energy (BE) that increases with titanium oxide state viz. Ti 2+, Ti 3+ and
Ti 4+. The TiOx is a non-stoichiometric compound with composition
ranging from TiO0.7 to TiO1.3. The TiO2 spin-orbit doublet showed
binding energies of 458.8 eV (Ti 2p3/2) and 464.6 eV (Ti 2p1/2). On the
other hand, the XPS peaks at 457.0 eV and 463.1 eV are attributed to
Ti2O3 and TiOx and previously found at 455.3 eV and 461.5 eV [21]. The
O 1s XPS depth profile in Fig. 7e shows that the oxygen is present in
Al2O3 at the binding energy of 532.0 eV near the surface of the oxide
layer. Moreover, minor SiO2 at 532.9 eV, and Cr2O3 at the binding en­
ergy of 530.9 eV were located near the surface of the oxide [25]. Fig. 7f
shows Si peaks appearing in the spectrum and a small fraction of SiO2
near the surface from the Si 2p peak at the binding energy of 103.0 eV
[25]. Therefore, a mixed oxide layer including Al2O3, Cr2O3 and SiO2
was found near the surface [25].
Fig. 8a shows the TEM image with corresponding zoomed out area

7
M.N. Mathabathe et al.
Fig. 9. TEM images of the cyclically oxidized γ-Ti-46Al–2Nb-0.7Cr-0.3Si alloy at 900 ◦ C for 900 h. The corresponding insert of a zoomed-out particle is shown in (a),
the elemental mapping of (N, O, Al, Si, Ti, Cr, Cu, Nb) of the interface domain between the coating and the alloy (b), the high-resolution TEM indicating a faceted
particle (c) with the corresponding SAED (d).
indicating a γ-Ti-46Al–2Nb alloy faceted morphology. The micrograph
reveals the particles agglomerated on the carbon film and abroad the
margin of the holes within the film. This makes the process of attaining
high-resolution image complex because particles of interest are often
unnoticeable by the overlapping particles [26]. The EDS results in
Figs. 8b, 9b and 10b was done to analyse the interfacial reaction be­
tween the oxide film and the alloy after cyclic oxidation at 900 ◦ C for
100 h. The TEM results justifies the oxide XPS spectra analysis. The
elemental mapping in Fig. 8b confirm that the scale is an oxynitride. The
scanning transmission electron microscope (STEM) did not pick-up Nb
nor Al. Besides, the HRTEM Fig. 8c revealed evidence of surface reform
on a single-crystal faceted particle with the d-spacing of 0.3259 nm. The
corresponding SAED Fig. 8d suggests that the overall oxide layer is
crystalline, despite the weak brightness and intensity of the other
polymorphic rings due to partial amorphous structure of the oxide film.
Fig. 9a shows the TEM image of the cyclically oxidized γ-Ti-
46Al–2Nb-0.7Cr-0.3Si alloy with the complementary zoomed-out par­
ticle insert. The particle was characterized by STEM-EDS, which is
represented by elemental mapping in Fig. 9b. The elemental particles
8
Materials Chemistry and Physics 257 (2021) 123703
assume a regular morphology. Fig. 9c illustrates the HRTEM of the
faceted particle. The red arrow designates the ordered lattice fringes.
The diffraction pattern in Fig. 8d was taken from the dark ordered
particles in Fig. 8c. The SAED validated the crystalline nature of the
particles. Similar findings by Rozita et al. [26], revealed facets that
terminate with both (111) and (100) planes confirmed by partially
disordered sub-surface visible in Fig. 9c. The phase indexed by (300) and
(113) planes belongs to the alumina phase. The measured (100) plane
d-value is shown in the lattice fringe.
Fig. 10 shows the density graphs of the annealed, nitrided and cyclic
oxidized Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si intermetallic al­
loys. The annealed Ti–46Al–2Nb alloy has a density of 3.954 g/cm3
while nitridation at 900 ◦ C it decreased to 3.909 g/cm3. A slight increase
in the density of Ti–46Al–2Nb alloy to 3.912 g/cm3 was recorded after
cyclic oxidation attributable to oxide, nitride and oxynitride scale for­
mations. This behavior is agreement with the slight mass gain shown in
Fig. 3. The 3.946 g/cm3 and 3.919 g/cm3 densities were measured for
the annealed and cyclically oxidized Ti–46Al–2Nb-0.7Cr-0.3Si inter­
metallic alloys, respectively. The cyclically oxidized Ti–46Al–2Nb-
M.N. Mathabathe et al.
Fig. 10. Density graphs of the annealed, nitrided and cyclic oxidized
Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si intermetallic alloys.
Fig. 11. Micro-hardness profile of the Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-
0.3Si alloys.
0.7Cr-0.3Si alloy also showed a slight rise in density compared to the
nitrided sample. Moreover, the density variation between the nitrided
and cyclic oxidized alloys is very negligible in comparison to the nitrided
and cyclic oxidized Ti–46Al–2Nb alloys. This behavior implies that the
Ti–46Al–2Nb-0.7Cr-0.3Si alloy has shown better stability after cyclic
oxidation than the Ti–46Al–2Nb alloy where spallation was profound in
Fig. 3 between 700 and 900 h.
3.5. Macro and micro hardness profile of the γ-TiAl based alloys
Fig. 10. Macro-hardness profile of the γ-TiAl based alloys.
Fig. 10 illustrates the macro-Vickers hardness conducted on the
cross-sections of the Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys
in annealed, nitrided and cyclically oxidized samples. The annealed
samples had high hardness values when compared to all the other
samples. Moreover, the bulk hardness results show that Ti–46Al–2Nb
alloy is low in hardness. The nitrided Ti–46Al–2Nb-0.7Cr-0.3Si alloy
hardness is also lower than that of the annealed sample. On the contrary,
cyclic oxidation improved the hardness of the nitrided Ti–46Al–2Nb but
softened that of the Ti–46Al–2Nb-0.7Cr-0.3Si alloy. Chang et al. [25]
reported comparable results and found that the coexisting of Al2O3 and
SiO2 near the surface retards the diffusion of oxygen into the
Cr0.38Al0.56Si0.06N coating, hence resulted in the improved oxidation
behavior. Furthermore, in case of the Ti–46Al–2Nb-0.7Cr-0.3Si alloy, it
is believed that the Al may have been incorporated into the CrN
9
Materials Chemistry and Physics 257 (2021) 123703
Fig 12. Micro-hardness profile of the γ-TiAl alloys.
structure (B1-face centered cubic (FCC), which decreases the grain size,
by modifying its preferred crystalline orientation from (200) to (111)
[27], and inhibit the formation of hexagonal chromium nitride (Cr2N)
phase [23,24]. The previous study on the Ti–46Al–2Nb and
Ti–46Al–2Nb-0.7Cr-0.3Si alloys in as-cast and cyclically oxidized con­
ditions without nitridation treatment indicated a slight increase in
hardness [11] in comparison with the current findings. This behavior
was attributed to the brittleness prompted by the adsorption of inter­
mixed oxides on the surface of the alloys [11], contrary to the nitrided
scales that induced softness in the alloys in the current study. The
hardness values measured after Ti–46Al–2Nb alloy cyclic oxidation
without nitridation was 262 HV while that of Ti–46Al–2Nb-0.7Cr-0.3Si
alloy was 298HV [11], which is a decrease of 1.9%, and 8.7%, respec­
tively, to the nitridated alloys.
Fig. 11 illustrates the micro-hardness profile of the nitrided
Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys. The average profile
evaluation was conducted on the nitride surfaces and core of the sam­
ples. Generally, the hardness measurements obtained near the nitrided
surfaces are slightly higher than the core matrix. This is attributed to the
difference in the cooling effect between the core and the case of the
sample. Upon annealing, more Al was dissolved in the core structure due
to its greater mobility than Ti. The effect of vacancies within the
structure promotes steadfast oxide growth throughout the cavities [28],
and accordingly the reduction in hardness in the core of the alloys is due
to adsorption of oxides on the surface matrix causing deterioration/­
brittleness of the structure [11]. Abd El-Rahman et al. [29], reported
that surface hardness primarily depends on the heat treatment temper­
ature. Moreover, in nitrided samples, the surface enhancement is
attributed to the formation of the nitride layer on the surface. The sur­
face hardness of the Ti–46Al–2Nb-0.7Cr-0.3Si alloy is greater than
Ti–46Al–2Nb alloy in a nitrided condition because the alloy is comprised
of the Cr contribution to the protective scale as indicated by the
composition in Figs. 7a and 9b. The presence of Cr2N in the γ-TiAl based
alloys display superior hardness although inclined to oxidation damage
resulting in degradation of the thermal stability of the scale [28].
Accordingly, the variation in surface hardness of the alloys is due to the
relative substrate effect on the treated layer; along with small variations
in the microstructure as a consequence of different alloying elements
[25,26,30]. The Al and Si metals in Ti–46Al–2Nb-0.7Cr-0.3Si alloy hard
coatings have improved oxidation resistance due to the formation of
protective oxides on the surface layer and hindered ionic diffusion.
Similar findings by Menand et al. [31] showed that high O and C affinity
to γ-TiAl based alloys leads to an increase in the hardness values.(see fig
12)
4. Conclusion
The thermal properties, macro and micro-hardness, and micro­
structure of the vacuum melted, annealed and cyclically oxidized
M.N. Mathabathe et al.
Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys were studied. The
following conclusions are drawn:
• The annealing of Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si alloys
has stabilized γ-TiAl +α2-Ti3Al phases. Moreover, the residual
β-phase in Ti–46Al–2Nb, and the Ti5Si3 precipitates in Ti–46Al–2Nb-
0.7Cr-0.3Si alloys have emerged.
• Due to surface nitridation, the oxidation properties of Ti–46Al–2Nb
and Ti–46Al–2Nb-0.7Cr-0.3Si alloys slightly improved but suffered
the oxide spallation.
• The density of Ti–46Al–2Nb and Ti–46Al–2Nb-0.7Cr-0.3Si interme­
tallic alloys were decreased due to surface oxides, nitride and oxy­
nitride formations.
• TEM analysis confirmed partial amorphization in the Ti–46Al–2Nb
alloy and crystalline in the Ti–46Al–2Nb-0.7Cr-0.3Si alloy.
• XPS chemical composition confirmed the presence of mixed oxide
layer of Al2O3, Cr2O3 and SiO2 near the Ti–46Al–2Nb-0.7Cr-0.7Si
surface.
CRediT authorship contribution statement
M.N. Mathabathe: Formal analysis, Writing - original draft, con­
ducted the experiments and prepared the manuscript. A.S. Bolokang:
Formal analysis, Writing - original draft, conducted the experiments and
prepared the manuscript. C.W. Siyasiya: Formal analysis, Writing -
original draft, conducted the TEM and XPS discussions. R.J. Mostert:
Formal analysis, Writing - original draft, conducted the TEM and XPS
discussions. All authors contributed in the analysis of the data, editing
and commented on the manuscript.
Declaration of competing interest
The authors, M. N. Mathabathe, A. S. Bolokang, G. Govender, C.W.
Siyasiya, R.J. Mostert declare that they have no known competing
financial interests or personal relationships that could have appeared to
influence the work reported in this paper. There is no conflict of interests
for this article.
Acknowledgement
This work was supported by the Department of Science and Inno­
vation (DSI), Republic of South Africa through Grant number :
HCARD08.
References
[1] C.G. McKamey, S.H. Whang, C. Liu, Microstructural characterization of a γ-TiAl-Ni
alloy produced by rapid solidification techniques, Scripta Metall. 32 (3) (1995)
383–388.
[2] D.J. Kim, D.Y. Seo, X. Huang, Q. Yang, Y.W. Kim, Cyclic oxidation behavior of a
beta gamma powder metallurgy TiAl-4Nb-3Mn alloy coated with a NiCrAlY
coating, Surf. Coating. Technol. 206 (13) (2012) 3048–3054.
[3] A.S. Bolokang, D.E. Motaung, C.J. Arendse, T.F.G. Muller, Formation of the
metastable FCC phase by ball milling and annealing of titanium-stearic acid
powder, Adv. Powder Technol. 26 (2) (2015) 632–639.
[4] M. Song, M. Xiang, Y. Yang, Q. Zhu, C. Hu, P. Lv, Synthesis of stoichiometric TiN
from TiH2 powder and its nitridation mechanism, Ceram. Int. 44 (14) (2018)
16947–16952.
10
Materials Chemistry and Physics 257 (2021) 123703
[5] W. Zhu, D. Shi, Z. Zhu, J. Sun, Effect of the surface oxidization and nitridation on
the normal spectral emissivity of titanium alloys Ti-6Al-4V at 800-1100 K at a
wavelength of 1.5 μm, Infrared Phys. Technol. 76 (2016) 200–205.
[6] K. Sopunna, T. Thongtem, M. McNallan, S. Thongtem, Surface modification of the
γ-TiAl alloys by the nitridation, Surf. Sci. 566–568 (2004) 810–815.
[7] B. Zhao, J. Wu, J. Sun, B. Tu, F. Wang, Oxidation kinetics of the nitrided TiAl-based
alloys, Mater. Lett. 56 (4) (2002) 533–538.
[8] B. Zhao, J. Sun, J.S. Wu, Z.X. Yuan, Gas nitriding behavior of TiAl based alloys in
an ammonia atmosphere, Scripta Mater. 46 (2002) 581–586.
[9] A.M. Limarga, D.S. Wilkinson, A model for the effect of creep deformation and
intrinsic growth stress on oxide/nitride scale growth rates with application to the
nitridation of γ-TiAl, Mater. Sci. Eng. 415 (1–2) (2006) 94–103A.
[10] M.N. Mathabathe, A.S. Bolokang, G. Govender, C.W. Siyasiya, R.J. Mostert, Cold-
pressing and vacuum arc melting of γ-TiAl based alloys, Adv. Powder Technol. 30
(12) (2019) 2925–2939.
[11] M.N. Mathabathe, A.S. Bolokang, G. Govender, R.J. Mostert, C.W. Siyasiya, The
vacuum melted ɣ -TiAl ( Nb, Cr, Si) -doped alloys and their cyclic oxidation
properties, Vacuum 154 (2018) 82–89.
[12] M.N. Mathabathe, S. Govender, A.S. Bolokang, R.J. Mostert, C.W. Siyasiya, Phase
transformation and microstructural control of the alpha-solidifying ɣ-Ti-45Al-2Nb-
0.7Cr-0.3Si intermetallic alloy, J. Alloys Compd. 757 (2018) 8–15.
[13] M.N. Mathabathe, A.S. Bolokang, G. Govender, R.J. Mostert, C.W. Siyasiya,
Structure-property orientation relationship of a γ/α2/Ti5Si3 in as-cast Ti-45Al-2Nb-
0.7Cr-0.3Si intermetallic alloy, J. Alloys Compd. 765 (2018) 690–699.
[14] C. Yang, D. Hu, A. Huang, M. Dixon, Solidification and grain refinement in
Ti45Al2Mn2Nb1B subjected to fast cooling, Intermetallics 32 (2013) 64–71.
[15] J.J. Williams, Structure and High-Temperature Properties of Ti5Si3 with Interstitial
Additions, PhD thesis, 1999, p. 2494.
[16] A. Abba, A. Galerie, M. Caillet, High-temperature oxidation of titanium silicide
coatings on titanium, Oxid. Metals 17 (1) (1982) 43–54.
[17] C.M. Austin, T.J. Kelly, Gas turbine engine implementation of gamma titanium
aluminide, Superalloys 1996 (1996) 539–543.
[18] D.R. Askeland, The Science and Engineering of Materials, Springer US, 1996.
[19] K. Li, Y. Li, X. Huang, D. Gibson, Y. Zheng, J. Liu, L. Sun, Y.Q. Fu, Surface
microstructures and corrosion resistance of Ni-Ti-Nb shape memory thin films,
Appl. Surf. Sci. 414 (2017) 63–67.
[20] Q. Li, L. Zhang, X. Chen, D. Wei, P. Zhang, Y. Chen, X. Xu, Characterization of
plasma rotating electrode atomized Nb-Ti based alloy powder, Met. Powder Rep
(2019) 1–10, https://doi.org/10.1016/j.mprp.2019.04.064.
[21] C. Badini, S.M. Deambrosis, O. Ostrovskaya, V. Zin, E. Padovano, E. Miorin, Cyclic
oxidation in burner rig of TiAlN coating deposited on Ti-48Al-2Cr- 2Nb by reactive
HiPIMS, Ceram. Int. 43 (7) (2017) 5417–5426.
[22] A.S. Bolokang, Z.P. Tshabalala, G.F. Malgas, I. Kortidis, H.C. Swart, D.E. Motaung,
Room temperature ferromagnetism and CH4 gas sensing of titanium oxynitride
induced by milling and annealing, Mater. Chem. Phys. 193 (2017) 512–523.
[23] J. Sjölén, L. Karlsson, S. Braun, R. Murdey, A. Hörling, L. Hultman, Structure and
mechanical properties of arc evaporated Ti-Al–O–N thin films, Surf. Coating.
Technol. 201 (2007) 6392–6403.
[24] A.S. Bolokang, F.R. Cummings, B.P. Dhonge, H.M.I. Abdallah, T. Moyo, H.C. Swart,
C.J. Arendse, T.F.G. Muller, D.E. Motaung, Characteristics of the mechanical
milling on the room temperature ferromagnetism and sensing properties of TiO2
nanoparticles, Appl. Surf. Sci. 331 (2015) 362–372.
[25] Y.Y. Chang, C.P. Chang, D.Y. Wang, S.M. Yang, W. Wu, High temperature oxidation
resistance of CrAlSiN coatings synthesized by a cathodic arc deposition process,
J. Alloys Compd. 461 (1–2) (2008) 336–341.
[26] Y. Rozita, R. Brydson, A.J. Scott, An investigation of commercial gamma-Al2O3
nanoparticles, J. Phys. Conf. Ser. 241 (2010), 012096.
[27] E. Mohammadpour, Z.-T. Jiang, M. Altarawneh, N. Mondinos, M.M. Rahman, H.
N. Lim, N.M. Huang, Z. Xie, Z.-F. Zhou, B.Z. Dlugogorski, Experimental and
predicted mechanical properties of Cr1− xAIxN thin films, at high temperatures,
incorporating in situ synchrotron radiation X-ray diffraction and computational
modelling, RSC Adv. 7 (36) (2017) 22094–22104.
[28] M.N. Mathabathe, G. Govender, C.W. Siyasiya, R.J. Mostert, A.S. Bolokang, Surface
characterization of the cyclically oxidized γ-Ti-48Al-2Nb-0.7Cr alloy after
nitridation, Mater. Char. 154 (2019) 94–102.
[29] A.M. Abd El-Rahman, M.F. Maitz, M.A. Kassem, F.M. El-Hossary, F. Prokert,
H. Reuther, M.T. Pham, E. Richter, Surface improvement and biocompatibility of
TiAl24Nb10 intermetallic alloy using rf plasma nitriding, Appl. Surf. Sci. 253 (23)
(2007) 9067–9072.
[30] A.I.P. Nwobu, R.D. Rawlings, D.R.F. West, Nitride formation in titanium based
substrates during laser surface melting in nitrogen-argon atmospheres, Acta Mater.
47 (2) (1999) 631–643A.
[31] A. Menand, A. Huguet, A. Nerac-Partaix, Interstitial solubility in γ and α2 phases of
TiAl-based alloys, Acta Mater. 44 (1996) 4729–4737.