Surface & Coatings Technology 312 (2017) 91–100

Contents lists available at ScienceDirect

Surface & Coatings Technology

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

Physical and chemical effects of the hydrogen irradiation on nitrided

titanium surfaces
Gelson B. de Souza a,⁎, Silvio H. Gonsalves a, Karen C. Ribeiro a, Dair G. Ditzel a, Mario Ueda b, Wido H. Schreiner c
a
UEPG - Universidade Estadual de Ponta Grossa, Departamento de Física, Av. Gen. Carlos Cavalcanti 4748, 84.030-900 Ponta Grossa, PR, Brazil
b
INPE - Instituto Nacional de Pesquisas Espaciais, Laboratório Associado de Plasma, 12227-010 São José dos Campos, SP, Brazil
c
UFPR - Universidade Federal do Paraná, Departamento de Física, Centro Politécnico 81531-980, Curitiba, PR, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Many applications aiming the use of titanium and titanium-based materials involve the surface interaction with
Received 6 April 2016 protons or hydrogen isotopes, such as in chemical and nuclear power plants, marine and aerospace environ-
Revised 30 August 2016 ments. Because of the high affinity of titanium for hydrogen and the consequent deleterious effects, the study
Accepted in revised form 1 October 2016
of surface interactions contributes to the understanding of the protection mechanisms, such as the H-diffusion
Available online 4 October 2016
barrier provided by nitriding, as well as the fine surface tailoring achieved by hydrogenation. Nitrided titanium
Keywords:
surfaces were produced via plasma nitriding (PN), carried out at 400 °C and 600 °C. Afterwards, titanium and ni-
Titanium trided titanium were submitted to the hydrogen plasma immersion ion implantation (H-PIII), using ion energies
Hydrogen from 1.2 to 5.0 keV. The hydrogen modifications imposed were restricted to the nanometer depth range, causing
Plasma nitriding no significant variations in hardness and elastic modulus (measured in depths larger than 70 nm) or in the crys-
Plasma immersion ion implantation talline structure, as inferred from nanoindentantion and grazing incidence X-ray diffraction, respectively. From
Denitriding the X-ray photoelectron and micro-Raman spectroscopies, the hydrogen irradiation was found to cause
denitriding on the nitrided titanium, changing the TiN stoichiometry as a consequence of conjoined physical
and chemical effects on the surface atoms. The atomic N/Ti ratio changed from 0.9 (not hydrogenated) to 0.6
(5 keV) up to ~2 nm depth. The Ti/N decrement and the resulting TiO2 surface layer growth followed approxi-
mately linear correlations with the implantation energies. Additional investigations with Fourier-transform in-
frared spectroscopy disclosed vibrational bands featured by a strong line at 668 cm−1 (possibly OTi(OH)2) in
samples submitted independently to PN and H-PIII, suggesting that different mechanisms of hydroxyl absorption
on titanium took place on those surfaces.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction special attention, due to the well-known embrittlement of the bulk ma-
terial, mediated by the precipitation of titanium hydrides [5–7]. Thus,
Varied technological areas have drawn attention to titanium, titani- the use of titanium-based materials in hydrogen-rich environments,
um-based compounds and titanium-based alloys, substantiated by the which might be found in chemical plants and marine and aerospace ap-
superior properties of these materials or even due to their unique re- plications, faces restrictions.
sponses to physical and chemical conditions. For instance, photocatalyt- Another strategic area of research demanding special materials is the
ic properties of titanium oxides and oxinitrides can be used in super fusion energy production. Titanium alloys are candidates to be used as
hydrophilic windows, air purification devices, and for degradation of or- raw materials in the reactor first wall and other structures in the
ganic molecules on their surfaces [1]. In a different approach, titanium- power plant, because of their suitable mechanical properties, low ex-
graphene composites were investigated to be used as hydrogen storage pansion coefficient, resistance to high heating fluxes and fast induction
means in fuel cells, due to the large H-affinity of these materials [2]. It of radioactive decay [8,9]. In these situations, strong interaction of high-
was reported that the surface bioactivity (i.e., induction of the bone ly energetic particles from the plasma with the reactor walls occurs, no-
healing and growth) can be attained in titanium prosthesis directly tably proton, deuterium and tritium. Moreover, a product from the last
through high energy hydrogen implantation [3,4]. In the situations two isotopes reaction is 14.1 MeV neutrons [10], which eventually de-
above, titanium interacts with hydrogen in varied degrees of severity. cays into a proton, an electron and an antineutrino. Fast neutrons are
Those circumstances where the hydrogen influx is significant demand also obtained in vacuum neutron tubes and gas filled neutron tubes, in
which tritium is stored in titanium targets [11].
⁎ Corresponding author. The bulk mechanical properties of hydrogen-containing titanium
E-mail addresses: [email protected], [email protected] (G.B. de Souza). have been extensively studied [5–7], motivated by the technological

http://dx.doi.org/10.1016/j.surfcoat.2016.10.005
0257-8972/© 2016 Elsevier B.V. All rights reserved.
92 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100

challenge to extend this material applicability to many environments. pulse generator, as detailed in reference [16]. The base pressure was in
Regarding titanium surfaces, the hydrogen diffusion, trapping and barri- the order of 10− 3 Pa, whereas the treatment pressure was 3 ×
er effect were investigated using H-ion implantation in surfaces con- 10−1 Pa. Pulse width, frequency and current density were 30 μs,
taining different dopants such as C, N and O [12,13]. In a previous 300 Hz and 2 mA/cm2, respectively. Samples were hydrogenated for
work [14], the surface ductile to brittle transition induced by the hy- 15 min under maximum pulse voltages 2.5 kV and 5 kV. An additional
dride precipitation was investigated at nanoscale level by tribological hydrogen implantation at 1.2 kV was prepared to support the Raman
and mechanical tests, in titanium surfaces submitted to the glow dis- analysis, as discussed in Section 3.2.2. The treatment temperatures
charge hydrogenation. The same hydrogen processing carried out on ni- were inferred with an optical pyrometer. Because of the device lower
trided titanium revealed some singular phenomena. The hydrogen end of the scale, it was possible to assert that treatment temperatures
penetration at shallow depths into the nitride barrier caused escalation were lower than 250 °C.
of the surface brittleness due to imposed competitive tensile stresses. Table 1 shows the sample nomenclature and treatment parameters
Moreover, XPS chemical analysis strongly suggested the denitriding of for all the employed combinations of PN and H-PIII. It is worth mention-
the surface, i.e., suppression of Ti\\N bonds, as well as changes in the ing that approximate ion energies could be ascribed to each of the PIII
surface oxidation states. high voltages employed, as summarized in Table 1. This is appropriate
The work reported here investigated further the surface effects of since a fair approximation for energies of the incoming ions on the top
the hydrogen irradiation, focusing on the structural and chemical surfaces can be the potential energy transferred to them through the
changes produced at near surfaces. Titanium and nitrided titanium sam- plasma sheath. By considering a collisionless situation, the energy con-
ples were subjected to controlled and energetic H irradiation, attained servation gives [17]
trough the plasma immersion ion implantation (PIII) technique. The
findings are presented in three sections. The crystalline structure and 1

2
mv2 ðxÞ−Es ¼ −eΦðxÞ ð1Þ
mechanical properties are discussed in Section 3.1, which allowed the
inference of depth range of the H-induced modifications. The next two
sections analyze the denitriding effect (Section 3.2) and the surface hy- where v(x) is the ion velocity in the sheath, Es is the ion energy when
droxylation (Section 3.3), disclosed through different spectroscopic penetrating the plasma and sheath interface, and m and e are the ion
methods. mass and charge, respectively. Φ(x) is the sheath voltage, which drops
to the maximum negative value in the sample holder (Φcathode). Assum-
2. Materials and methods ing that the ion kinetic and potential energies when it is in the bulk plas-
ma (in the order of eV) are much smaller than the cathode potential
2.1. Surface preparation (keV), the equation above can be reduced to

Commercially pure titanium samples (grade 2) were ground with 1


2
mv2max ¼ −eΦcathode ð2Þ
sandpaper and polished with diamond paste and colloidal silica to the
mirror-like surface finishing. Subsequently, samples were cleaned
with successive ultrasound baths of acetone, isopropyl alcohol and dis- where the left hand side is the maximum ion energy.
tilled water. The surfaces were initially treated through the convention- Of course, in a realistic situation where momentum transfer events
al plasma nitriding (PN) method, and then submitted to hydrogenation. to the neutral or ionic species take place during the ion time of flight
The plasma nitriding (PN) was carried out in a DC regime with 60% through the sheath, the above assumption can be taken just as a
N2 40% H2 atmosphere using 99.9995% purity gases for 3 h. The treat- rough estimate for the implantation energy. It is worth considering
ment temperatures were 400 °C and 600 °C. The chamber's base pres- that H+ is the dominant species in the hydrogen plasma at the
sure was 3 × 10− 2 Pa. After the working gas mixture inlet, the employed PIII working pressures, followed by smaller fractions of H+ 2
pressures were 470 Pa (400 °C) and 570 Pa (600 °C). Voltage and cur- and H+ 3 [18,19]. Thus, for referencing purposes, Table 1 also presents
rent were adjusted in order to set the treatment temperatures. The the projected ranges into TiN (density 5.4 g/cm3, angle of ion incidence
choice of parameters followed results reported in a previous paper [15]. 0°) for the H+ implantation ion energies, as calculated by the SRIM sim-
The surface hydrogenation of nitrided titanium samples was ulation method [20]. The computations simulated the implantation
attained by means of the hydrogen plasma immersion ion implantation time employed in the study, sampling 105 ions. The input data were as
(H-PIII), using pure H2 gas. Titanium samples, not nitrided, were also follows: displacement energies of 25 eV and 28 eV for Ti and N, respec-
submitted to H-PIII for comparison purposes. The system consisted of tively [21]; cohesive energies of 3.1 eV for both elements [22]. Individual
a DC plasma source, a heated filament producing electron shower at binding energies were employed: 4.89 eV for Ti and 2.00 eV for N, in ac-
the vicinity of the sample holder, a vacuum system and a high voltage cordance with [23,24].

Table 1
Sample nomenclature and variable treatment parameters of titanium samples submitted to plasma nitriding (PN) and hydrogen plasma immersion ion implantation (H-PIII). Estimated
ion energies, projected ranges and sputtering yields ratios are presented for each of the H-PIII treatments. The last two columns correspond to experimental values from the XPS (Fig. 4 and
Eq. (10)) and micro-Raman (Fig. 8 and Eq. (11)) spectroscopies, respectively.
 
Nomenclature Temperature in PN (°C) Ion energy in the H-PIII (keV)a Projected ion range (nm)b Sputtering yields ratiob Y N Y Ti ln IIlay
ox
¼ ln ITiO
ITiO2
Atomic N Ti
2
þITiNx Oy þI TiN

ref – – – – – –
ref/2.5 – 2.5 21 – – –
ref/5.0 – 5.0 40 – – –
nit400 400 – – – – –
nit400/2.5 400 2.5 21 – – –
nit400/5.0 400 5.0 40 – – –
nit600 600 – – – −2.8 0.94
nit600/1.2 600 1.2 11 2.0 – 0.83
nit600/2.5 600 2.5 21 2.3 −1.8 0.80
nit600/5.0 600 5.0 40 1.9 −0.7 0.63
a
Representative values, numerically equal to the PIII pulse maximum voltage. See text in Section 2.1 for details.
b
Projected ranges and sputtering yields were calculated by the SRIM simulation method [20], considering the H+ implantation on TiN.
 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100 93

The fluence (or dose) D was inferred by the treatment and material 3. Results
parameters [25]:
3.1. Micrometer-range analysis

N τf τfL Results in this section, namely GI-XRD and instrumented indenta-

D¼ ¼ qions ¼j ; ð3Þ
A Ze Zeð1 þ γ Þ tion, allowed the inference of depth range of the H-PIII changes caused
to titanium and nitrided titanium surfaces.
The structural, morphological and tribo-mechanical characterization
where N is the amount of implanted ions, A the implanted area, qions the of the nitrided titanium (nit400 and nit600 samples) was discussed
implanted ion charge, j the current density, and τ the total treatment elsewhere [15], therefore, this report focuses on the hydrogenation ef-
time. f and L are the pulse frequency and width, respectively. As afore- fects produced on the samples. Fig. 1 shows the grazing incidence X-
mentioned, the average atomic number is Z≅1 [18,19]. The secondary ray diffractograms of the nitrided samples and those hydrogenated
electrons emission coefficient (γ) is ~0.8 for 2–5 keV H+ over Ti-based with 2.5 and 5 keV. The untreated titanium presented only α-Ti peaks,
targets [26,27]. It is worth discussing the value of the pulse width, whose diffractogram showed no alterations after neither the 2.5 kV
since the real pulse shape differs from the ideal rectangular one. A (Fig. 1) nor the 5 kV hydrogenation. Thus, differently from the case
raise and a fall periods intercalate the flat-top portion of the pulse. where there is intense hydrogen inlet on titanium [5,6,14,31], Ti-H
With the employed pulser facility, the raise time was of few microsec- phases such as TiH and TiH2 were not identified after H-PIII. Main con-
onds, delivering low ion current to the sample. In the fall time, voltage tributions of these compounds to the diffractograms were expected in
drops up to zero in a long “tail”. However, the plasma sheath is depleted the range 34° N 2θ N 44° [according to PDF cards no. 9-371 (TiH2), 40-
of its initial ion population in the fall time, so that the average ion cur- 1244 (γ-TiH), and 44-1217 (δ-TiH)]. When compared to other methods,
rent is also low. As the ion energy rapidly drops in this portion of the the present H-PIII can be considered a “soft” hydrogenation, featured by
pulse, implantation is no longer the dominant effect [28]. Therefore, a relatively low total implantation fluence (1016 ions/cm2), as estimated
the most significant part of hydrogen implantation occurs during the in Section 2.1. On the other hand, titanium hydrides may have precipi-
flat-top portion of the pulse, which is 30 μs long, under the maximum tated on titanium at very shallow depths, not observable by GI-XRD
implantation energies as summarized in Table 1. With these assump- but observed by micro-Raman analysis, as seen in Section 3.2.2.
tions, Eq. (3) led to the rough estimate of 1016 ions/cm2 delivered to Sample plasma nitrided at 400 °C presented nitrides only at shallow
the surface during each of the treatments. depths and formation of nitrogen in solid solution – Ti(N), as previously
reported [15]. After hydrogenation, Ti-H phases were not identified ei-
ther. Likewise, diffractograms of surfaces nitrided at 600 °C showed no
2.2. Surfaces characterization evident formation of hydrides after H-PIII. Such results indicated that
crystalline Ti-H compounds, if present, were in amounts smaller than
The crystalline structure was investigated through grazing incidence the detection limit of the GI-XRD with the employed parameters, as
X-ray diffraction – GI-XRD (Rigaku Ultima IV) with 3° fixed incidence discussed above for untreated titanium. Actually, this is an unlikely
angle, in the range from 20° to 60°, using Cu Kα radiation under 40 kV event for the nitrided surfaces. The hydrogen barrier effect given by
and 30 mA. The scan velocity was 0.3 °/min.
Hardness and elastic modulus profiles were measured by instru-
mented indentation (MTS NanoIndenter XP) following the Oliver and
Pharr [29] method. Surfaces were tested in 20 different sites using cy-
cling loads from 1 to 50 mN. The indentation imprints produced with
the Berkovich-type diamond tip were examined through field emission
scanning electron microscopy – FEG-SEM (Tescan Mira 3), operating at
20 kV.
The X-ray photoelectron spectroscopy – XPS measurements
(VGMicrotech ESCA 3000) were employed to investigate chemical fea-
tures of the nitrided and hydrogenated samples, using Al-Kα radiation
(1486.7 eV). The spectra were recorded at a 45° take off angle. The
band-pass energies were 50 and 20 eV for the survey and high resolu-
tion spectra, respectively. The overall spectral resolution was 0.8 eV.
The C1s binding energy at 284.5 eV was taken as reference for correction
of the elemental peaks. The Shirley method [30] was employed for the
background subtraction, whereas the spectra were deconvoluted by a
multi-peak Gaussian fit using the software Fityk [30].
The Fourier-transform infrared spectroscopy – FTIR (Shimadzu IR
Prestige) provided adjuvant tests for the chemical characterization of
the modified samples. The spectra were acquired in the range 400–
4000 cm− 1 with a He-Ne laser using the reflection mode, with 2
cm−1 resolution and 128 scans. Backgrounds were subtracted with re-
spect to a mirror-like polished untreated titanium surface.
Micro-Raman spectra (Bruker Senterra) were obtained from differ-
ent surface regions using a solid state laser with λ = 532 nm. The radi-
ation power was 20 mW with an estimated 2 μm spot size diameter.
Each spectrum was collected with 5 co-additions of a 30 s exposure,
with 5 cm−1 spectral resolution. The general spectra were resolved Fig. 1. Grazing-angle X-ray diffractograms of plasma nitrided at 400 °C (nit400) and 600 °C
(nit600) titanium samples, before and after the H-PIII at 2.5 kV and 5 kV. The untreated
through multi-peaks Gaussian-Lorentzian fit using the Fityk software
surface (ref) hydrogenated at 2.5 kV is also shown. Titanium hydride contributions, if
[30] to quantify the N/Ti ratio on the nitrided and hydrogenated visualized, would lay approximately in positions 36°, 41° and 43°. α = α-Ti; ε = Ti2N;
surfaces. * = Ti(N) = nitrogen in solid solution.
94 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100

titanium nitride layers is well known [12,13], limiting the hydrogena- similar features as those hydrogenated with 5 kV. The ref and nit400
tion effects to shallow depths at the surfaces. Intense plasma hydroge- samples disclosed a ductile-like deformation, featured by the plastic dis-
nation (glow discharge during 3 h) was reported to produce hydrogen placement of material under normal loading. On the other hand, several
bubbles on the nitrided surfaces but prevented the precipitation of Ti- cracks were observed in the nit600 sample, a condition in which the ni-
H compounds [14]. Bubbles were not visualized through FEG-SEM in- triding effects were more effective (see Fig. 1). In both cases, titanium
spection in the present case. Despite the higher ion energy in the H- and nitrided titanium, the imprint morphologies did not present strik-
PIII as compared to a glow discharge process, the relatively low-fluence ing differences caused by the subsequent hydrogenation, differently
implantation may also explain the absence of visible morphology from changes observed after hydrogen bombardment by glow dis-
changes. charge [14]. As previously reported, the plasma-based hydrogenation
Fig. 2 shows typical indentation imprints produced with the pyrami- promoted a ductile to brittle-like transition in titanium (related to the
dal Berkovich-type indenter, for the untreated (ref) and nitrided sur- precipitation of hydrides), whereas surfaces containing nitride precipi-
faces (nit400 and nit600), along with images of tests carried out after tates showed increase in brittleness [14]. In Fig. 2, it is possible to ob-
H-PIII at 5 kV. Indentations on samples treated at 2.5 kV H-PIII presented serve a discrete increase in cracking in the nit600/5.0 sample, laying at

Fig. 2. In the left side column, secondary-electrons FEG-SEM images of typical indentation imprints produced with 400 mN applied load, for the untreated (ref) and plasma nitrided at
400 °C (nit400) and 600 °C (nit600) titanium samples. Micrographs of the corresponding surfaces submitted to hydrogen implantation at 5 kV are shown in the right column. Larger
frames are magnifications of imprints shown in the insets.
 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100 95

the indentation edges inside the imprint. The implanted hydrogen, even
lodged at shallow depths, produced additional stresses and point de-
fects to the nitrided titanium surface. Therefore, the H-induced embrit-
tlement was more intense in those samples (the nit600) containing
significant amount of residual stresses [32].
As shown in Fig. 3, the surface plasma nitrided at 400 °C were harder
than the untreated ones at shallow depths, whereas hardness profiles
significantly increased for 600 °C treatments. In the latter, at 100 nm, av-
erage values were threefold higher than the substrate. Elastic modulus
for the same treatment increased about 40%. More detailed discussion
on the nitrided titanium surfaces were presented in [15]. After hydroge-
nation at 5 kV (presenting the larger projected range - Table 1), the
hardness and elastic modulus profiles were statistically similar to
those obtained before the H irradiation. Thus, lattice defects induced
by H-PIII did not produce significant changes in mechanical properties
of the surfaces under study, untreated and nitrided, in the analyzed
range of depth (N 70 nm).
The results presented in this section indicate that modifications
caused by the H-PIII, using the conditions summarized in Table 1,
seemed to be limited to short-range depths. Noticeable differences
emerged from the spectroscopy methods presented in the following
sections. Fig. 4. XPS Ti2p spectra obtained from plasma nitrided surfaces at 600 °C (nit600), before
and after the H-PIII at 2.5 kV (nit600/2.5) and 5 kV (nit600/5.0). The doublet spectra were
3.2. The surface “denitriding” deconvoluted into characteristic contributions assigned to TiN, TiNxOy and TiO2. The
spectra shift with the implantation energy can be ascribed to the increase in the TiO2
contribution and the concomitant reduction in the TiN peak intensity.
The near surface analysis through spectroscopy methods disclosed
changes produced by H-PIII in their constitution and chemical features.
This section discusses the XPS and micro-Raman findings, whereas the increase in the TiO2 peaks intensities, and the concomitant reduction
complement FTIR results are presented in Section 3.3. in the TiN and TiNxOy contributions. Similar trends were observed in
the N1s and O1s XPS spectra, shown in Fig 5a and b, respectively,
3.2.1. XPS which allowed comparison of the involved titanium compounds in
The doublet in the XPS Ti2p spectrum of the untreated sample pre- pairs. The N1s peaks were found at energies corresponding to TiN and
sented peaks consistent with TiO2. This result is in accordance with TiNxOy [37], whose shapes were similar among the nitrided and nitrided
the literature [14,33–36], meaning that the titanium dioxide layer ex- plus hydrogenated samples. The shoulder at ~399.5 eV can be ascribed
tends at least to the XPS depth of analysis. Fig. 4 presents XPS spectra to molecular absorbed N and/or N-O/N-C contaminants [38]. On the
corresponding to the Ti2p doublet for selected treated samples. Unlike other hand, the TiO2 peak in the O1s spectra, with center in 529.9 eV
the untreated titanium, the plasma nitrided samples at 600 °C [37], increased in intensity after H-PIII, whereas the TiOH (at 531.5
(nit600) were resolved by considering different titanium chemical envi- and 533.3 eV) and TiNxOy (~ 532 eV) peaks [14,38] clearly decreased.
ronments, in line with the NIST database [37] and related literature [14, Concerning the hydroxyl groups, their associated energy in the Ti2p
33–36], whose values for the Ti2p3/2 orbital were 458.4 eV (TiO2) and high resolution spectrum depends on its nature and, therefore, the sur-
455.1 eV (TiN). Likewise, an additional peak in 457.5 eV ascribed to face treatment method employed [39–41]. In this study, OH– bonds pos-
TiNxOy was added to the XPS envelopes [14,33–36,38]. After the surface sibly had minor contributions in the Ti4+ peaks (that is, close to the fully
irradiation (with 2.5 kV), and then increasing the H-PIII energy (5. 0 kV), oxidized state TiO2, Fig. 4), as discussed in the FTIR results, Section 3.3.
the overall Ti2p spectra shifted to higher binding energies due to the

Fig. 5. (a) XPS N1s and (b) O1s spectra obtained on plasma nitrided surfaces at 600 °C
Fig. 3. Elastic modulus and hardness profiles obtained for the untreated (ref) and plasma (nit600), before and after the H-PIII at 2.5 kV (nit600/2.5) and 5 kV (nit600/5.0). Dashed
nitrided at 400 °C (nit400) and 600 °C (nit600) titanium samples, before and after the H- lines are guide for the eyes, indicating expected positions for the labelled bonds. In (a),
PIII at 5 kV. Profiles for the corresponding samples submitted to H-PIII at 2.5 kV, not shown the shoulders at about 399 eV can be ascribed to N-contaminants or absorbed molecular
for the sake of simplicity, where statistically similar to their pristine surfaces. nitrogen.
96 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100

The XPS findings were consistent with two phenomena associated where I is the intensity emerging normally to the surface, x the depth
with the hydrogen irradiation of the nitrided titanium surfaces, namely position, and λ is the material inelastic mean free path (IMFP), a param-
the increase in the oxide surface thickness, as well as changes on its hy- eter associated with the surface sensitivity in electron spectroscopies
droxylation state. The latter is treated in the next section. The following [43,44]. Photoelectrons originated in the base of the TiO2 layer (with
paragraphs discuss further the intensity variations in the oxide, nitride thickness d) and reaching the spectrometer detector, positioned at an
and oxinitride peaks, as seen in Fig. 4. angle θ with respect to the surface, must cross a distance d/ sinθ, as rep-
The presence of TiO2 on nitrided titanium, as observed here and re- resented in the inset of Fig. 6. The attenuation experienced by photo-
ported by other authors [14,33–36], is compatible with TiNx species un- electrons from deeper regions depends on the specific IMPF found in
dergoing oxidation in oxidant environments. Piscanec et al. [36] each of the crossed layers. The IMPF relates with material properties
investigated such mechanism by XPS and first-principles molecular dy- such as the volumetric density ρ, molecular mass M, number of valence
namics, proposing that the resulting surface comprises an oxynitride electrons per atom (or per molecule) Nv, and the band gap energy Eg.
TiNxOy region, intercalating TiO2 and TiN layers, as sketched in the IMPF values were calculated for a series of elements and compounds
inset of Fig. 6. Some level of mixture among these phases may occur at by Tanuma, Powell and Penn [45,46], and can be obtained through the
the surface, but the stratification model was assumed in the present software QUASES IMPF [47] following their so-called TPP-2 M equation.
case. This is a reasonable choice, based on findings from previous results Considering Ti2p photoelectrons with energies [42]
[14]. Nit600-like samples exposed to the glow discharge hydrogenation
for 3 h, with much higher fluence than the present H-PIII, changed from E ¼ Eγ −BE−φspec ≈ 1000 eV ð5Þ
the typical TiN-TiNxOy-TiO2 XPS spectra to only TiO2 peaks, although the
thicker nitride layer beneath was preserved. Instead of a well-defined
boundary between oxides and nitrides, they interface inwardly in the (where Eγ is the photon energy, BE the binding energy, and φspec the
TiNxOy region. Here, samples were exposed to air after nitriding, or spectrometer work function), the calculated TiO2 IMPF was
may be oxidized due to residual oxygen present in the plasma chamber.
In spite of the XRD results (Fig. 1), it is important to mention that an ef- λTiO2 ≈ 2:1 nm ð6Þ
fective δ-TiN layer with few nanometers can be assumed on these plas-
ma nitrided samples, as previously shown [15]. Thus, the variations in
The TiN IMPF (ρ = 5.2 g/cm3, M = 61.9 g/mol, Nv = 7) demands
TiO2 peak intensities with increasing H-PIII energy, as observed in the
knowledge of the band gap energy. In spite of the predominant covalent
Ti2p XPS spectra (Fig. 4), correlated with the increasing outermost
character of TiN, Allmaier et al. [48] described this compound as a mate-
oxide layer thickness. However, it is important to get insights on how in-
rial in the metal-insulator transition, presenting a “paseudogap” at the
tensity and thickness correlate. This requires some assumptions, as
Fermi level, whose minimum is at the center of the Brillouin zone. By
follows.
considering the minimal value Eg ≈ 1 eV [48,49], the calculated λTiN
(i) Only those photoelectrons arising from the surface, which did was approximately 1.9 nm. Because of the similar values of λTiN and
not experience inelastic collisions (such as plasmon excitation, λTiO2, the TiNxOy IMPF is presumably close to them. Such values are co-
electron-hole pair formation, and vibrational excitation), con- herent with typical IMPF reported in the literature [42,45,46]. This
tribute to the characteristic peak formation [42]. means that the Ti2p photoelectron attenuation when passing through
(ii) The photoelectron intensity I0, produced in inner photon-solid the different layers shown in the inset of Fig. 6 correlates roughly to
interactions, is attenuated when passing through the material ac- the same IMPF. Thus, from Eq (4), the ratio between the TiO2 peak in-
cording to the Beer-Lambert law, given by [42,43] tensities (the oxide) and those concerning the completely analyzed
depth (TiO2, TiNxOy and TiN – the layer) can be written as
 
 x d
= x
I¼
x
I 0 ∫ x21 exp − dx; ð4Þ I 0;ox ∫ 0 sinθ exp − dx
λ Iox λox
¼   ; ð7Þ
Ilay ∞ x
I 0;lay ∫ 0 exp − dx
λlay

where the denominator on the right side was integrated for an “infinite-
ly thick” solid [44,50]. Considering an average value λ as representative
for the comparable λox and λlay,
  
Iox I 0;ox d
¼ 1− exp − ; ð8Þ
Ilay I 0;lay λ sinθ

So that the film thickness is

 
I 0;lay Iox
d ¼ −λ sinθ ln 1− : ð9Þ
I 0;ox Ilay

(In the equation above, the content in parenthesis is less than one).
Fadley et al. [50] initially presented such an expression through a la-
bored approach.
Fig. 6. Correlations of the H-PIII ion energies with features of the nitrided (at 600 °C) plus
hydrogenated sample surfaces. The outermost TiO2 oxide layer thickness correlates with As Iox and Ilay are experimental measurements, inferred from the
the natural logarithm values through Eq. (10) represented by open symbols and spectra of Fig. 4, I0,ox and I0,lay (intensities for infinitely thick materials
addressed to the left side axis. In the analyzed XPS sampling depth, the higher such emerging normally to the surface) are unknown material parameters.
quantity is, the thicker the film thicknesses are. The N/Ti ratios (close symbols addressed Theoretical intensities can be calculated, however, they depend on
to the right side axis) were inferred from micro-Raman analysis by quantifying optical
and acoustical Raman scattering, following Eq. (11). The inset represents the surface
many variables such as the density of atoms in crystal planes, the atomic
stratification produced by nitriding, whereas arrows indicate photoelectron fluxes level cross section, the IMFP, the probability of no loss escape, and the
toward the spectrometer detector. instrumental detection efficiency [42,50]. To avoid unnecessary
 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100 97

inaccuracies, and because I0,ox/I0,lay is a constant parameter, only quali- hydride scatterings [55,56]. In this situation, Raman spectroscopy is
tative comparisons were made. For this, Eq. (8) can be rewritten as more sensitive than XRD (Fig. 1), where such phases were not identi-
   fied. Differently, on the nitrided surfaces, the H-barrier effect prevented
Iox I0;ox d hydrides precipitation, as discussed in Section 3.1. Spectra of samples
ln ¼ ln þ ln 1− exp − ð10Þ
Ilay I0;lay λ sinθ nitrided at 400 °C (nit400) and 600 °C (nit600) were compatible with
the fcc δ-TiN, in accordance with the literature [52–54] and previous re-
The XPS sampling depth is assumed to be less than the IMPF (~30% sults [15]. Nitrides were formed on the 400 °C case only at shallow
shorter than λ) [42]. In such range (d ~ 1.4 nm), it is reasonable to con- depths, that is the reason why they were not identified by XRD [15].
sider a few atomic layers TiO2-TiNxOy over TiN. Untreated titanium sur- After hydrogenation, the 400 °C spectra (shown in Fig. 7b) were clearly
faces present a ~ 2 nm thick native TiO2 layer, which is amorphous or modified with the increasing implantation energy. The relative intensi-
nanocrystalline [51]. Oxides from the same nature grew thinner than ties of ~530 cm−1 peaks reduced with respect to the 200 cm−1 peaks.
that over nitrided surfaces [36,38], therefore consistent with spectra The same trend was observed in surfaces nitrided at 600 °C, Fig. 7c.
from Fig. 4. This phenomenon is consistent with changes in the TiN stoichiome-
In the same probing range, for d up to ~1.4 nm, Eq. (10) increased in try. Fig. 8 shows the micro-Raman spectra for the entire 600 °C nitriding
a linear-like fashion. Fig. 6 presents the ln IIlay
ox
values with respect to the batch. The peaks labelled as TA and LA correspond to transversal and
longitudinal acoustical Raman modes, respectively, ascribed to the tita-
H-PIII energies, as inferred from Fig. 4, and summarized in Table 1. Be-
nium fcc-sublattice. Likewise, TO and LO correspond to transversal and
cause the TiO2 thickness increased with ln IIlayox
, and the XPS sampling longitudinal optical Raman modes, ascribed to the N fcc-sublattice.
depth in the three cases was essentially the same (due to the similar Peaks at 420 cm−1 are second-order 2TA scattering [52–54]. The small-
IMPF values), it was concluded that the film grew with the hydrogen ir- er peaks at ~138 and ~184 cm−1 correspond to titanium oxides [57]. It
radiation energy, pushing the TiN interface deeper into the solid. Aiming is well accepted that the ratio between integrated areas of first-order
at investigating further this phenomenon, samples were analyzed by optical and acoustical modes can be used to estimate the N/Ti atomic
micro-Raman spectroscopy, a suitable technique to evaluate TiN layers ratio in TiN [52–54], that is,
at shallow depths.
N I TO þ ILO
∝ : ð11Þ
3.2.2. Micro-Raman Ti ITA þ ILA
As previously mentioned, titanium nitride is a metal-like material,
even though it is less conductive than pure metals [48]. Notwithstand- The N/Ti values are summarized in Table 1, and correlated with the
ing the visible light reflexivity, Raman spectroscopy is a well- H-PIII energy in Fig. 6. The sample only nitrided (nit600) was not fully
established technique for phase analysis in TiN. In short, First-order stoichiometric, that is, Ti/N ≠ 1. Even so, the higher the implantation en-
Raman scattering in the NaCl-type structure of δ-TiN is not entirely for- ergy was, the more reductions in the nitride stoichiometries occurred,
bidden due to intrinsic nitrogen point defects, which reduce the crystal featured by the lack of nitrogen in the TiN structure.
effective symmetry. More detailed discussions were presented in refer-
ences [15,52–54]. As a result, the N-sublattice produce phonons in the
optical mode, whereas the titanium sublattice originates scattering in
the acoustical mode [52–54]. In addition, the sampling depth of the
Raman spectroscopy in nitrided titanium surfaces is ~25 nm, as demon-
strated elsewhere [15], which is one order of magnitude larger than that
of XPS, Eq. 6.
Fig. 7 shows micro-Raman spectra of titanium and nitrided titanium,
obtained before and after the H-PIII. The untreated sample (ref) pre-
sented no Raman scattering, as expected. After hydrogenation using
2.5 and 5 kV acceleration voltages (Fig. 7a), peaks at 270, 800 and
1050 cm−1 were visible, which can be possibly attributed to titanium

Fig. 8. Raman spectra of titanium plasma nitrided surfaces at 600 °C, before (nit600) and
after H-PIII at 1.2 kV (nit600/1.2), 2.5 kV (nit600/2.5) and 5 kV (nit600/5.0). TA/LA =
Fig. 7. (a) Raman spectra of titanium before (ref) and after H-PIII at 2.5 kV (ref/2.5) and first order transversal and longitudinal acoustical scattering modes. TO/LO: first order
5 kV (ref/5.0). In (b) and (c), spectra obtained on plasma nitrided surfaces at 400 °C transversal and longitudinal optical scattering modes. Peaks at 420 cm−1 are second-
(nit400) and 600 °C (nit600), before and after the H-PIII at 2.5 kV and 5 kV. order 2TA scattering. The peaks at ~138 and ~184 cm−1 correspond to titanium oxides.
98 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100

In Fig. 6, it is possible to observe that the reduction in the TiN stoichi-

ometry was accompanied by the increase in the TiO2 layer, suggesting
that N bonds were replaced by O bonds.

The denitriding mechanism assisted by the energetic hydrogen irra-

diation is likely to be a conjunction of physical and chemical effects. The
calculated electronic stopping power values for the used energy range
were one order of magnitude larger than the nuclear stopping powers,
in eV/nm [20]. Thermal spikes, electronic excitations and electron ejec-
tions from the target atoms can act as catalyst for N-chemical reactions
with the environment; the role of the nitrogen sputtering is discussed in
the paragraphs below. For instance, the denitriding in steels involve for-
mation of gaseous nitrogen (N2) or ammonia (NH3) [58]. According to
Porte et al. [59], the removal of nitrogen states (p-type) in the TiN
band structure leads to the electron redistribution to higher unoccupied
(d-type) states. Moreover, the ionic character of the Ti\\N bonds de-
creases and the covalent one increases, since metallic atoms surround-
ing N vacancies regain some of their electron densities. The eventual
oxidation leads to new redistribution of charges and energy states
with Ti [60] or Ti and N [1].
Table 1 shows that the linear correlation of the calculated ion ranges
and energies is straightforward. However, the electronic collisions by
energetic H-ions do not denitride the whole volume of implantation,
since the projected hydrogen ion range, listed in Table 1, is much deeper
than the oxide layer thickness (which in turn is less than the XPS sam-
pling depth, equation (6)). The implantation energy, say, at half the
projected ion range, is still in the order of keV, much higher than the
TiN cohesive energy or NH3/N2 enthalpies, of few eV. Thus, the
degassing and the nitride oxidation, directly correlated with the
denitriding, are depth-limited effects in TiN.
In order to gain more insights in this phenomenon, additional SRIM
Fig. 9. FTIR spectra of titanium (ref), titanium plasma nitrided at 400 °C (nit400) and
[20] computations were carried out for H+ ions implanted on a TiN target. 600 °C (nit600), and the corresponding samples after H-PIII at 2.5 kV and 5 kV. The
The required input data are detailed in Section 2.1. The average number narrow peak at 668 cm−1 is possibly ascribed to the Ti-OH stretching in OTi(OH)2 [63].
of Frenkel pairs produced per implanted ion increased with beam energy, The inset highlights the region delimited by a rectangle, involving the nit600 batch
from 1.3 (1.2 keV) to 3.8 (5.0 keV). Concomitantly, the maximum of samples.

the generated defects distribution moved inwards the surface, as expect-

ed, deeper than the oxide layer thickness inferred in Section 3.2.1 (oxide demanding little activation energy. In their report, the effect of O and
b2 nm). The estimated sputtering yields (Y = (mean number of emitted H isotopes in the peaks positions were also calculated.
atoms)/(number of incident particles)) [61] were particularly interesting. In the XPS O1s measurements, Fig. 5b, the Ti_OH (at 531.5 eV) cor-
Values were ~10−2 atom/ion for N and Ti, varying slightly with the inci- respond to physically adsorbed water (free molecules). The chemically
dent ion energy. However, as shown in Table 1, YN was nearly twofold the adsorbed water Ti-OH (at 533.3 eV), existing as hydroxyl groups on
YTi for all the analyzed H+ energies. Such preferential sputtering of nitro- the top of oxides [65], decreased significantly after the H-PIII. Therefore,
gen atoms seems to play an important role in denitriding, in conjunction the O_Ti–(OH)2 here identified contributed with the XPS O1s spectra
with the N-H chemical affinity, as discussed above. Since titanium is high- in a different range, possibly in the Ti4+ range.
ly reactive with oxygen [36,60], TiO2 is formed with residual O2+ inside Fig. 9 revealed that (i) the OTi(OH)2 was identified on titanium after
the plasma chamber, or right after air exposure. hydroxylation, however, (ii) such band was already present on the ni-
trided surfaces before hydroxylation, (iii) whose absorbance intensities
3.3. Hydroxylation reduced after H-PIII. From the first statement, it can be concluded that
the hydrogen irradiation indeed promoted chemical changes on titani-
Fig. 9 shows FTIR spectra of hydrogenated and nitrided + hydroge- um. However, the mechanisms through which the hydroxyl bands cor-
nated titanium samples. All the treated samples present noticeable dif- related with the energetic hydrogen bombardment were not clear and
ferences with respect to the untreated one (ref), marked by the narrow demand further investigations. Some possible hypotheses arise from
line at ~668 cm−1 (indicated by an asterisk). Other smaller absorption the literature, as follows.
peaks also appeared in 418, 445, 618 and 720 cm−1, as labelled in the On the titanium samples. The hydrogen bombardment of titanium
inset of Fig. 9. Some of these bands seemed to be disclosed in other oxides promotes its reduction via hydroxylation, as proposed by several
FTIR reports on TiN [62], although with lower intensity. By considering authors [3,4,41]. A simultaneous reconstruction of the oxide layer dur-
the literature, the appearance of these bands can be ascribed to oxida- ing H-PIII is possible, since titanium hydrides, disclosed by the Raman
tion states of titanium [39], and/or the formation of hydroxyl groups, analyses (Fig. 7a), can undergo oxidation [66] with the residual oxygen
“stuck” on the top of bridging oxygen atoms [36,60]. Some of these in the chamber.
bands can be TiO2 modes, particularly those lying at ~ 418 cm−1 [40] On the nitrided samples. A different line of reasoning suggests the for-
and ~440 cm−1 [63], as well as the broad band at ~950 cm−1 [64]. How- mation of hydroxyls after air exposure of the nitrided titanium. The
ever, in accordance with the work by Shao et al. [63], the strongest peak presence of oxinitrides at the surfaces may promote a photocatalytic ef-
at ~668 cm−1 and other absorbance bands found after the surface treat- fect (the decomposition of water) more significantly than the pure TiO2,
ments on titanium are consistent with OTi(OH)2 molecule vibrational as demonstrated by Asahi et al. [1]. In this scenario, a subsequent H-PIII
modes. The authors [63] observed that OTi(OH)2 is formed spontane- could lead to desorption of OH radicals as water molecules and/or re-
ously during laser ablation annealing, via reactions of TiO2 and water, duction in the photocatalysis [41].
 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100 99

4. Conclusions [12] K. Neu, H. Baumann, N. Angert, D. Riick, K. Bethge, Gettering of hydrogen in titanium
caused by nitrogen implantation, Nucl. Inst. Methods Phys. Res. B 89 (1994) 379–381.
[13] M. Soltani-farshi, H. Baumann, D. Ru, E. Richter, U. Kreissig, K. Bethge, Content of hy-
Titanium and plasma nitrided titanium were submitted to hydrogen drogen in boron-, carbon-, nitrogen-, oxygen-, fluorine-, and neon-implanted titani-
irradiation by means of plasma immersion ion implantation (H-PIII), um, Surf. Coat. Technol. 104 (1998) 299–303.
[14] G.B. de Souza, C.E. Foerster, C.M. Lepienski, N.K. Kuromoto, S.L.R. Da Silva, W.H.
using ion energies from 1.2 to 5 keV. After hydrogenation, additional Schreiner, Structural, chemical and tribo-mechanical surface features of Ti and ni-
phases were not identified by GI-XRD, whereas surfaces nitrided at trided Ti submitted to hydrogen low energy implantation, Mater. Chem. Phys. 124
600 °C presented a slight increase in brittleness under indentation. On (2010) 443–452, http://dx.doi.org/10.1016/j.matchemphys.2010.06.064.
[15] G.B. de Souza, B.A. da Silva, G. Steudel, S.H. Gonsalves, C.E. Foerster, C.M. Lepienski,
the other hand, analyses carried out at the nanometer range using spec- Structural and tribo-mechanical characterization of nitrogen plasma treated titani-
troscopy techniques disclosed energy dependent effects on titanium um for bone implants, Surf. Coat. Technol. 256 (2014) 30–36, http://dx.doi.org/10.
and nitrided titanium, as follows. 1016/j.surfcoat.2013.12.009.
[16] M. Ueda, L.A. Berni, J.O. Rossi, J.J. Barroso, G.F. Gomes, A.F. Beloto, Plasma immersion
ion implantation experiments at the Instituto Nacional de Pesquisas Espaciais
– Nitrided titanium surfaces underwent denitriding, up to ~ 2 nm (INPE), Brazil, Surf. Coat. Technol. 136 (2001) 28–31.
depth, featured by the removal of nitrogen atoms from TiN and the [17] M.A. Lieberman, Fundamentals of plasmas and sheaths, in: A. Anders (Ed.), Handb.
subsequent increase in the surface oxide thickness, in the presence Plasma Immers. Ion Implant. Depos. John Wiley & Sons, Inc. 2000, pp. 29–123.
[18] O. Fukumasa, R. Itatani, S. Saeki, Numerical simulation of hydrogen ion species in
of oxygen. As inferred by XPS and Micro-Raman spectroscopies, thes teady-state plasma of a low-pressure ion source, J. Phys. D. Appl. Phys. 18
the higher the implantation energy was, the less the atomic N/Ti (1985) 2433–2449.
ratio and the thicker the TiO2 layer were, both parameters following [19] O.D. Cortazar, A. Megia-Macias, O. Tarvainen, T. Kalvas, H. Koivisto, Correlations be-
tween density distributions, optical spectra, and ion species in a hydrogen plasma,
an approximately linear correlation.
Rev. Sci. Instrum. 87 (2016) 02A704, http://dx.doi.org/10.1063/1.4931720.
– The denitriding is likely to be a conjunction of physical effects pro- [20] J.F. Ziegler, J.P. Biersack, M.D. Ziegler, The Stopping and Range of Ions in Matter,
vided by the energetic ion beams, as the preferential nitrogen http://www.srim.org/2013.
sputtering from the surface, with N-H chemical reactions in the [21] J.X. Xue, G.J. Zhang, F.F. Xu, H. Bin Zhang, X.G. Wang, S.M. Peng, et al., Lattice expan-
sion and microstructure evaluation of Ar ion-irradiated titanium nitride, Nucl. In-
chamber environment. struments Methods Phys. Res. Sect. B Beam Interact. Mater. Atoms 308 (2013)
– The hydroxylation states of surface oxides changed after the titani- 62–67, http://dx.doi.org/10.1016/j.nimb.2013.05.011.
um nitriding or the H-PIII. FTIR lines ascribed to OTi(OH)2 vibrational [22] Y.M. Kim, B.J. Lee, Modified embedded-atom method interatomic potentials for the
Ti-C and Ti-N binary systems, Acta Mater. 56 (2008) 3481–3489, http://dx.doi.org/
states (strongest line at 668 cm−1) were identified in samples sub- 10.1016/j.actamat.2008.03.027.
mitted independently to both methods, suggesting that different [23] Y. Miyagawa, S. Nakao, M. Ikeyama, K. Saitoh, S. Miyagawa, High fluence implanta-
mechanisms of hydroxyl absorption on titanium took place on tion of nitrogen into titanium: fluence dependence of sputtering yield, retained
fluence and nitrogen depth profile, Nucl. Instruments Methods Phys. Res. Sect. B
those surfaces. Beam Interact. Mater. Atoms 121 (1997) 340–344.
[24] E. Franke, H. Neumann, M. Zeuner, W. Frank, F. Bigl, Particle energy and angle distri-
butions in ion beam sputtering, Surf. Coat. Technol. 97 (1997) 90–96, http://dx.doi.
org/10.1016/S0257-8972(97)00304-6.
Acknowledgements [25] D.M. Goebel, R.J. Adler, D.F. Beals, W.A. Reass, Pulser Technology, in: A. Anders (Ed.),
Handb. Plasma Immers. Ion Implant. Depos. John Wiley & Sons, Inc., New York 2000,
pp. 467–513.
The authors are grateful to Professor C. M. Lepienski (UFPR, Brazil) [26] L.N. Large, The effect of absorbed gases on proton induced electron emission from
for the use of nanoindenter, and the Institutional Laboratory C- titanium, Proc. Phys. Soc. 81 (1963) 175–180.
[27] L.N. Large, W.S. Whitlock, Secondary electron emission from clean metal surfaces
LABMU/UEPG for the use of the characterization facilities. We acknowl-
bombarded by fast hydrogen ions, Proc. Phys. Soc. 79 (1962) 148, http://dx.doi.
edge the Brazilian Funding Agencies CNPq (grant no. 477973/2011-6) org/10.1088/0370-1328/79/1/319.
and Fundação Araucária (grants no. 363/2012 and 608/2014) for the fi- [28] C.M.B.P. Wood, D.J. Rej, A. Anders, I.G. Brown, R.J. Faehl, S.M. Malik, Fundamentals of
nancial support. plasma immersion ion implantation and deposition, in: A. Anders (Ed.), Handb.
Plasma Immers. Ion Implant. Depos. John Wiley & Sons, Inc., New York, 2000.
[29] W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus by instru-
References mented indentation: advances in understanding and refinements to methodology,
J. Mater. Res. 19 (2004) 3–20http://www.mrs.org/publications/jmr/jmra/2004/jan/
[1] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light photocatalysis in ni- 0002.html.
trogen-doped titanium oxides, Science 293 (2001) 269–271, http://dx.doi.org/10. [30] M. Wojdyr, Fityk - A Curve Fitting and Data Analysis Program, http://fityk.nieto.pl/
1126/science.1061051 ((80-.)). 2010.
[2] S. Chu, L. Hu, X. Hu, M. Yang, J. Deng, Titanium-embedded graphene as high-capacity [31] T.B. Massalski (Ed.), Binary alloys phase diagrams, 2nd ed.ASM International, Mate-
hydrogen-storage media, Int. J. Hydrog. Energy 36 (2011) 12324–12328, http://dx. rials Park, OH, 1990.
doi.org/10.1016/j.ijhydene.2011.07.015. [32] A. Fleszar, J. Mizera, R.Y. Fillit, T. Wierzchoń, Influence of the residual stresses on the
[3] X. Liu, X. Zhao, R.K.Y. Fu, J.P.Y. Ho, C. Ding, P.K. Chu, Plasma-treated nanostructured corrosion and wear resistance of nitrided layers produced isothermally and cyclical-
TiO(2) surface supporting biomimetic growth of apatite, Biomaterials 26 (2005) ly on Ti-1Al-1Mn titanium alloy under glow discharge conditions, Vacuum 57
6143–6150, http://dx.doi.org/10.1016/j.biomaterials.2005.04.035. (2000) 405–411, http://dx.doi.org/10.1016/S0042-207X(00)00234-7.
[4] Y. Xie, X. Liu, A. Huang, C. Ding, P.K. Chu, Improvement of surface bioactivity on ti- [33] N. Jiang, H. Zhang, S. Bao, Y. Shen, Z. Zhou, XPS study for reactively sputtered titani-
tanium by water and hydrogen plasma immersion ion implantation, Biomaterials um nitride thin films deposited under different substrate bias, Phys. B Condens.
26 (2005) 6129–6135, http://dx.doi.org/10.1016/j.biomaterials.2005.03.032. Matter 352 (2004) 118–126, http://dx.doi.org/10.1016/j.physb.2004.07.001.
[5] C.Q. Chen, S.X. Li, H. Zheng, L.B. Wang, K. Lu, An investigation on structure, deforma- [34] V. Fouquet, L. Pichon, A. Straboni, M. Drouet, Nitridation of Ti6Al4V by PBII: study of
tion and fracture of hydrides in titanium with a large range of hydrogen contents, the nitrogen diffusion and of the nitride growth mechanism, Surf. Coat. Technol. 186
Acta Mater. 52 (2004) 3697–3706, http://dx.doi.org/10.1016/j.actamat.2004.04.024. (2004) 34–39, http://dx.doi.org/10.1016/j.surfcoat.2004.04.006.
[6] L. Yan, S. Ramamurthy, J.J. Noël, D.W. Shoesmith, Hydrogen absorption into alpha ti- [35] H.C. Man, Z.D. Cui, X.J. Yang, Analysis of laser gas nitrided titanium by X-ray photo-
tanium in acidic solutions, Electrochim. Acta 52 (2006) 1169–1181, http://dx.doi. electron spectroscopy, Appl. Surf. Sci. 199 (2002) 293–302.
org/10.1016/j.electacta.2006.07.017. [36] S. Piscanec, L.C. Ciacchi, E. Vesselli, G. Comelli, O. Sbaizero, S. Meriani, et al., Bioactiv-
[7] Y. Lu, P. Zhang, First-principles study of temperature-dependent diffusion coeffi- ity of TiN-coated titanium implants, Acta Mater. 52 (2004) 1237–1245, http://dx.
cients: hydrogen, deuterium, and tritium in a-Ti, J. Appl. Phys. 113 (2013) 193502, doi.org/10.1016/j.actamat.2003.11.020.
http://dx.doi.org/10.1063/1.4805362. [37] NIST, X-ray photoelectron spectroscopy database, Natl. Inst. Stand. Technol. (2016)
[8] P. Marmy, T. Leguey, I. Belianov, M. Victoria, Tensile and fatigue properties of two ti- http://srdata.nist.gov/xps/ (accessed January 31, 2016).
tanium alloys as candidate materials for fusion reactors, J. Nucl. Mater. 283–287 [38] M. Braic, M. Balaceanu, A. Vladescu, A. Kiss, V. Braic, G. Epurescu, et al., Preparation
(2000) 602–606. and characterization of titanium oxy-nitride thin films, Appl. Surf. Sci. 253 (2007)
[9] P. Groot, J.G. Van der Laan, M. Mack, M. Dvorak, P. Huber, Plasma-sprayed titanium 8210–8214, http://dx.doi.org/10.1016/j.apsusc.2007.02.179.
carbide coatings for first-wall applications in fusion devices, J. Nucl. Mater. 179–181 [39] G. Lu, S.L. Bernasek, J. Schwartz, Oxidation of a polycrystalline titanium surface by
(1991) 370–374, http://dx.doi.org/10.1016/0022-3115(91)90102-D. oxygen and water, Surf. Sci. 458 (2000) 80–90http://linkinghub.elsevier.com/re-
[10] A.K. Suri, N. Krishnamurthy, I.S. Batra, Materials issues in fusion reactors, J. Phys. trieve/pii/S0039602800004209.
Conf. Ser. 208 (2010) 012001, http://dx.doi.org/10.1088/1742-6596/208/1/012001. [40] P.M. Kumar, S. Badrinarayanan, M. Sastry, Nanocrystalline TiO2 studied by optical,
[11] A.M. Zakharov, O.A. Dvoichenkova, A.E. Evsin, Modification of surface oxide layers of FTIR and X-ray photoelectron spectroscopy: correlation to presence of surface
titanium targets for increasing lifetime of neutron tubes, Phys. At. Nucl. 78 (2015) states, Thin Solid Films 358 (2000) 122–130, http://dx.doi.org/10.1016/S0040-
1643–1645, http://dx.doi.org/10.1134/S106377881514015X. 6090(99)00722-1.
100 G.B. de Souza et al. / Surface & Coatings Technology 312 (2017) 91–100

[41] A.F. Carley, P.R. Chalker, J.C. Riviere, M.W. Roberts, The identification and character- [54] M.A.Z. Vasconcellos, R. Hinrichs, C.S. Javorsky, G. Giuriatti, J.A.T. Borges da Costa,
isation of mixed oxidation states at oxidised titanium surfaces by analysis of X-ray Micro-Raman characterization of plasma nitrided Ti6Al4V-ELI, Surf. Coat. Technol.
photoelectron spectra, J. Chem. Soc. Faraday Trans. Phys. Chem. Condens. Phases 202 (2007) 275–279, http://dx.doi.org/10.1016/j.surfcoat.2007.05.038.
83 (1987) 351–370, http://dx.doi.org/10.1039/f19878300351. [55] J. Íñiguez, W. Zhou, T. Yildirim, Vibrational properties of TiHn complexes adsorbed
[42] D.P. Woodruff, T.A. Delchar, Electron spectroscopies, Mod. Tech. Surf. Sci. 2nd on carbon nanostructures, Chem. Phys. Lett. 444 (2007) 140–144, http://dx.doi.
ed.Cambridge Univ. Press, Cambridge, UK 1999, pp. 105–264. org/10.1016/j.cplett.2007.06.133.
[43] I.S. Tilinin, A. Jablonski, W.S.M. Werner, Quantitative surface analysis by Auger and [56] T.K.A. Hoang, L. Morris, D. Reed, D. Book, M.L. Trudeau, D.M. Antonelli, Observation
X-ray photoelectron spectroscopy, Prog. Surf. Sci. 52 (1996) 193–335. of TiH5 and TiH7 in bulk-phase TiH3 gels for Kubas-type hydrogen storage, Chem.
[44] C.J. Powell, A. Jablonski, Electron effective attenuation lengths for applications in Mater. 25 (2013) 4765–4771, http://dx.doi.org/10.1021/cm402853k.
Auger electron spectroscopy and X-ray photoelectron spectroscopy, Surf. Interface [57] X. Wang, J. Shen, Q. Pan, Raman spectroscopy of sol-gel derived titanium oxide thin
Anal. 33 (2002) 211–229, http://dx.doi.org/10.1002/sia.1204. films, J. Raman Spectrosc. 42 (2011) 1578–1582, http://dx.doi.org/10.1002/jrs.2899.
[45] S. Tanuma, C.J. Powell, D.R. Penn, Calculations of electron inelastic mean free paths. [58] H. Dong, S-phase surface engineering of Fe–Cr, Co–Cr and Ni–Cr alloys, Int. Mater.
III Data for 15 inorganic compounds over the 50–2000 eV range, Surf. Interface Anal. Rev. 55 (2010) 65–98, http://dx.doi.org/10.1179/095066009X12572530170589.
17 (1991) 927–939, http://dx.doi.org/10.1002/sia.1997. [59] L. Porte, L. Roux, J. Hanus, Vacancy effects in the X-ray photoelectron spectra of TiNx,
[46] S. Tanuma, C.J. Powell, D.R. Penn, Calculations of electron inelastic mean free paths. Phys. Rev. B Condens. Matter Mater. Phys. 28 (1983) 3214–3224, http://dx.doi.org/
VIII. Data for 15 elemental solids over the 50–2000 eV range, Surf. Interface Anal. 36 10.1017/CBO9781107415324.004.
(2004) 1–14, http://dx.doi.org/10.1002/sia.1997. [60] U. Diebold, The surface science of titanium dioxide, Surf. Sci. Rep. 48 (2003) 53–229,
[47] S. Tougaard, QUASES-IMPF-TPP2M, http://www.quases.com/products/quases-imfp- http://dx.doi.org/10.1016/S0167-5729(02)00100-0.
tpp2m/2010. [61] M. Nastasi, W. Möller, W. Ensinger, Ion implantation and thin film deposition, in: A.
[48] H. Allmaier, L. Chioncel, E. Arrigoni, Titanium nitride: a correlated metal at the Anders (Ed.), Handb. Plasma Immers. Ion Implant. Depos. John Wiley & Sons, Inc.,
threshold of a Mott transition, Phys. Rev. B Condens. Matter Mater. Phys. 79 New York 2000, pp. 125–240.
(2009) 235126, http://dx.doi.org/10.1103/PhysRevB.79.235126. [62] L.I. Wei, C. Jun-fang, Growth of TiN films at low temperature, Appl. Surf. Sci. 253
[49] P. Patsalas, S. Logothetidis, Optical, electronic, and transport properties of nanocrys- (2007) 7019–7023, http://dx.doi.org/10.1016/j.apsusc.2007.02.028.
talline titanium nitride thin films, J. Appl. Phys. 90 (2001) 4725, http://dx.doi.org/10. [63] L. Shao, L. Zhang, M. Chen, H. Lu, M. Zhou, Reactions of titanium oxides with water
1063/1.1403677. molecules. A matrix isolation FTIR and density functional study, Chem. Phys. Lett.
[50] C.S. Fadley, R.J. Baird, W. Siekhaus, T. Novakov, S.A.L. Bergstrom, Surface analysis and 343 (2001) 178–184.
angular distributions in X-ray photoelectron spectroscopy, J. Electron Spectrosc. [64] X. Wang, J. Shen, Q. Pan, Raman spectroscopy of sol–gel derived titanium oxide thin
Relat. Phenom. 4 (1974) 93–137, http://dx.doi.org/10.4028/www.scientific.net/ films, J. Raman Spectrosc. 42 (2011) 1578–1582, http://dx.doi.org/10.1002/jrs.2899.
AMR.748.79. [65] H. Tamura, K. Mita, A. Tanaka, M. Ito, Mechanism of hydroxylation of metal oxide
[51] S.K. Yen, Retardation effects of thermally grown oxide films on the hydrogen em- surfaces, J. Colloid Interface Sci. 243 (2001) 202–207, http://dx.doi.org/10.1006/
brittlement of commercial pure titanium, Corros. Sci. 41 (1999) 2031–2051http:// jcis.2001.7864.
www.sciencedirect.com/science/article/pii/S0010938X99000220. [66] W. Lisowski, A.H.J. Van Den Berg, D. Leonard, H.J. Mathieu, Characterization of tita-
[52] Y.H. Cheng, B.K. Tay, S.P. Lau, H. Kupfer, F. Richter, Substrate bias dependence of nium hydride films covered by nanoscale evaporated Au layers: ToF-SIMS, XPS
Raman spectra for TiN films deposited by filtered cathodic vacuum arc, J. Appl. and AES depth profile analysis, Surf. Interface Anal. 29 (2000) 292–297.
Phys. 92 (2002) 1845, http://dx.doi.org/10.1063/1.1491588.
[53] W. Spengler, R. Kaiser, H. Bilz, Ressonant Raman scattering in a superconducting
transition metal compound - TiN, Solid State Commun. 17 (1975) 19–22.