SEARCH

SEARCH BY CITATION

Keywords:

  • corrosion;
  • HIPIMS/UBM;
  • multilayer coating;
  • sulphidation/oxidation;
  • TiAl

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

This paper discusses the sulphidation/oxidation protection offered by multilayer CrAlYN/CrN coatings etched by Cr, CrAl and Y ions, deposited by a combined high power impulse magnetron sputtering (HIPIMS)/unbalanced magnetron sputtering (UBM) technique on a Ti–45Al–8Nb alloy (at%). The test was performed at 750 °C in an environment of H2/H2S/H2O yielding low oxygen (10−18 Pa) and high sulphur (10−1 Pa) partial pressures for up to 1000 h. The results show that all the exposed materials underwent uneven degradation; some places developed a thin protective oxide scale (Al,Cr)2O3 with a tiny sulphur content, whilst others developed a porous non-protective TiO2 + Al2O3 scale as well with a tiny sulphur content.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

γ-TiAl-based intermetallic alloys possess the appropriate properties to potentially become structural materials, due to their low density, high specific strength and low weight compared to the currently used Ni-base super alloys [1]. However, the main concern regarding ordered γ-TiAl is its poor sulphidation/oxidation resistance at temperatures above 650 °C, and low ductility and toughness. Instead of a protective α-Al2O3 scale, ordered γ-TiAl alloy starts to develop mainly non-protective TiO2 with small amounts of β or γ-Al2O3 in both oxidising and oxidising/sulphidising environments [2-5]. This formed scale has a porous structure, where oxygen is likely to ingress, to deeper regions of the corroded material. Ideally, ordered γ-TiAl substrates need to be protected by a low-growth protective scale such as Al2O3.

The protectiveness of the coating deposited on the ordered γ-TiAl alloy mainly depends on: coating composition, service conditions, ambient atmosphere and operating temperature. Thus to overcome these problems, ordered γ-TiAl alloy needs to be protected by a coating with appropriate concentrations of the valuable elements: as Al or Cr to extend its lifetime. On the one hand Cr may undergo sulphidation in highly aggressive environments or volatile CrO3 can form in dry air, in wet air (steam) the formation of CrO2(OH)2 is likely to occur [6]. Thus the content of Cr needs to be well balanced, Cr can be beneficial, helps to develop Al2O3 phase [7], on the other hand Cr can seriously decrease the performance of the coating in long term exposures.

Although, some work related to the oxidation of nanostructured coatings deposited on the ordered γ-TiAl and stainless steels was presented previously [8, 9], high temperature sulphidation/oxidation behaviour of the ordered γ-TiAl-based intermetallics was not studied extensively. Some work regarding hot corrosion and sulphidation resistance of newly developed HIPIMS multilayered protective coatings based on CrAlYN/CrN was studied by Lasanta et al. [10].

In this study, CrAlYN/CrN based multilayer coatings were deposited on the ordered γ-TiAl substrate [Ti–45Al–8Nb (at%)], etched by Cr, CrAl and Y ions. The sulphidation/oxidation test was carried out in H2/H2S/H2O atmosphere of partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of O2 (pO2 = 10−18 Pa) at 750 °C for up to 1000 h.

2 Experimental procedure

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

2.1 Sulphidation/oxidation test

The coated samples and uncoated material (used here as a reference sample) were sealed in a silica tube within an electric furnace. After flushing with argon gas for 2 h, the silica tube was purged with pre-mixed 90%H2/10%H2S gas which passed through a gas flow rate gauge, then fluxed through a gas bubbler, containing de-ionised water in a bath maintained at a constant temperature of 23 °C. This latter step was designed to achieve a humidity level of 2.78% H2O within the mixture, with an oxygen partial pressure of 1.2 × 10−18 Pa at 750 °C. The ratio of H2/H2S was chosen so as to yield pS2 values of 10−1 Pa at 750 °C. The schematic diagram of the sulphidation test set-up is presented in Fig. 1; sample geometry used in this study is shown in Fig. 2. In this test 4 samples were used: (i) the uncoated alloy (Ti–45Al–8Nb) – reference sample, (ii) CrAlYN/CrN coated Ti–45Al–8Nb etched by Cr, (iii) CrAlYN/CrN coated Ti–45Al–8Nb etched by CrAl and (iv) CrAlYN/CrN coated Ti–45Al–8Nb etched by Y.

image

Figure 1. Sulphidation rig used for the experiments with high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C for 1000 h

Download figure to PowerPoint

image

Figure 2. Sample geometry used in this work for sulphidation in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressures of oxygen (pO2 = 10−18 Pa) at 750 °C for 1000 h

Download figure to PowerPoint

At the end of the experiment the furnace was allowed to cool naturally to room temperature. The corrosion kinetics was determined from all exposed materials in this study by discontinuous gravimetric method. All samples were investigated by Quanta 200 scanning electron microscope (SEM) using 10 mm working distance, high voltage (HV) = 20 kV in secondary electron mode (SE). The SEM was equipped with an energy dispersive X-ray analysis system (EDS) manufactured by OXFORD Instruments (INCA Energy). The phases formed on the surface during exposure were examined by Siemens 5000 X-ray diffractometer (XRD) using Cu Kα source radiation.

The cross section of the exposed samples after EDS and XRD analysis were prepared as follows: exposed samples were cut by a linear precision saw (Buehler ISOMET 5000). The sample was placed in a mounting press machine (Simplimet 2000) with appropriate amount of conductive resin. The sample after 10 min was withdrawn from the mounting press and ground by SiC paper (600, 1200 and 4000), then polished using 6 and 1 µm water-based diamond polishing paste (Buehler MetaDi). The sulphidation rig and the samples shape and dimensions are presented in Figs. 1 and 2, respectively.

The samples were suspended by platinum wire on a thin ceramic rod in order to avoid reaction with the SiO2 boat.

2.2 Deposition technique: HIPIMS/UBM

The investigated multilayer/super-lattice coatings were deposited by a combined high impulse power magnetron sputtering system HIPIMS and unbalanced magnetron sputtering system (UBM) in an industry sized HAUZER 1000-4 HTC coater. The standard coater had a chamber volume of 1 m3 and comprises four linear cathodes (target dimensions 600 × 200 mm2). Two of the magnetrons are operated in HIPIMS mode using HIPIMS power supplies (Hüttinger Elektronik, Warsaw, Poland), whereas the other two magnetrons were operated in UBM mode.

The system was equipped with a threefold planetary rotating substrate holder with a variable primary rotation frequency in the range of 1.5–8 rpm allowing uniform coating deposition. The coating deposition was performed at 450 °C. The substrate surface was etched before the coating was applied with different types of ions, CrAl, Y and Cr in order to enhance the adherence between layers. However only one of the samples were subjected to the HIPIMS Cr ion etch, i.e. the Cr etch sample – the other two had a CrAl ion etch from the CrAl target driven by the HIPIMS power supply and the other the Y ion etch from the CrAlY target driven by the HIPIMS power supply [8, 9].

To produce CrAlYN/CrN, two CrAl targets [Cr 40 (at%), Al 60 (at%), one Cr and one CrAlY (Cr 48 (at%), Al 48 (at%), Y 4 (at%)] were used.

A bias voltage of −1000 V was applied to the substrate during the etching step, further details of this procedure can be found elsewhere [11, 12]. Due to the high ionisation of the metal as well as gas fluxes when using HIPIMS the high currents are drawn by the substrate. This creates conditions for heavy arcing of the substrate, (pre-conditions intrinsic to HIPIMS). The solution was found by developing a dedicated HIPIMS bias power supply which can cope with the high currents, stabilise the applied voltage and suppress arcing due to the very fast arc detection and arc suppression unit of the power supply.

The HIPIMS treatment resulted in a very clean surface free of oxides, and a sharp interface, providing conditions for local epitaxial growth and high adhesion of the coating [13].

In the next steps a thin 0.4 µm CrAlN base layer was deposited followed by a 4.6 µm CrAlYN/CrN nanoscale multilayer structure deposited by standard UBM in a mixed N2–Ar atmosphere. The multilayer structure of the coating forms due to the sequential exposure of the substrate surface to the material fluxes produced by the four magnetrons during the rotation of the table. X-ray diffraction analyses were carried out on CrAlYN/CrN deposited in a wide range of bias voltages from −75 to −120 V. The analyses revealed that the coatings exhibit single phase NaCl FCC unit cell structure [8]. The average coating composition determined by SNMS depth profile analysis of 22.5% Cr, 30.2% Al, 0.3% Y and 50% N suggests that the Al/Cr ratio in the CrAlYN layer is far from the theoretically predicted maximum solubility of FCC AlN in FCC CrN of approximately 77 at% [14], which explains the overall FCC structure of the coating. Figure 3 shows SEM picture in SE mode and the concentration profiles of the coated γ-TiAl alloy performed on as received material.

image

Figure 3. EDS cross sectioned SEM image in SE mode and concentration profiles of the as received sample (CrAlYN/CrN)

Download figure to PowerPoint

3 Results

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

3.1 Kinetics

Figure 4 presents the weight gain per unit area (mg/cm2) versus time (h) data of the materials exposed in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C for 1000 h. In general, the mass gain exhibited by all three coating variants was found to be very low. However, while the Cr and Y etched materials gave similar mass gain characteristics, the CrAl etched sample showed the lowest mass gain from all the coated materials.

image

Figure 4. Mass gain change plot for CrAlYN/CrN coated Ti–45Al–8Nb etched by Y, CrAl and Cr after 1000 h sulphidation at 750 °C in atmosphere of pS2 = 10−1 Pa, pO2 = 10−18 Pa

Download figure to PowerPoint

The mass gain of the uncoated material reached 2.5 (mg/cm2), compared to the coated alloys which all showed a much lower weight gain ∼0.5 (mg/cm2) after 1000 h of exposure. Also, mass gain of the coated materials after 200 h of exposure remained constant to the end of the experiment (1000 h). Obtained results suggest that the sulphidation/oxidation of the studied alloys do not fit to the parabolic rate law. Generally all three materials coated by nanostructured coatings, showed enhanced corrosion resistance compared to the ordered uncoated γ-TiAl alloy.

Figure 5 shows EDS X-ray mapping of the uncoated sample (reference sample) after exposure at 750 °C in aggressive environment. It was observed, that the material, developed a thick scale consisting layers of TiO2 − Al2O3 + sulphur, sulphur was detected as well underneath the oxide scale, at the oxide–substrate interface.

image

Figure 5. EDS mapping of sulphidised ordered γ-TiAl uncoated alloy at 750 °C after 1000 h

Download figure to PowerPoint

The morphology and the scale structure formed on the exposed alloy is in good correlation with the results presented by other researches [15-17].

3.2 Surface morphologies

The SEM surface morphology of CrAlYN/CrN coated Ti45Al8Nb alloy etched by (A) Y, (B) CrAl and (C) Cr after 1000 h exposure to the H2/H2S/H2O atmosphere at 750 °C is given in Fig. 6. All three samples show an uneven surface morphology with isolated regions (free form oxide formation); some regions developed corrosion products. The EDS analysis in the corroded places confirmed formation of TiO2 oxide with a small amount of Al2O3 and sulphur.

image

Figure 6. Surface morphology of CrAlYN/CrN coated Ti–45Al–8Nb etched by (A) Y, (B) CrAl and (C) Cr after 1000 h sulphidation at 750 °C

Download figure to PowerPoint

The isolated regions of the exposed samples consisted mainly of the Al and Cr–oxides confirmed by EDS analysis, CrN and AlN phase was not detected by XRD performed from the surface of the exposed material, also XRD analyses was unable to detect aluminium or titanium sulphides. The results obtained in this research are confirmed by stability phase diagrams of Al[BOND]S[BOND]O and Ti[BOND]S[BOND]O (discussion section), where Al and Ti sulphides could not be developed, Al and Ti sulphides can be formed only when pO2 is lower than pO2 = 10−25 Pa. All the samples exposed to the sulphidized/oxidized environment at 750 °C showed a similar behaviour, the corroded materials developed two regions, which in this paper, will be called the ‘affected’ and the ‘unaffected’ region.

The affected region is a region where the coating cracked due to the mismatch of coefficient of thermal expansion (CET) between the coating and the substrate during cooling or heating processes. Due to the crack formation (white circles on a SEM image in SE mode and EDS X-ray mappings), development of a non-protective mixed scale consisting TiO2 + Al2O3 with a small amounts of sulphur was observed in CrAlYN/CrN coated Ti–45Al–8Nb etched by Y ions. In the other two materials investigated in this study, CrAlYN/CrN coated Ti–45Al–8Nb etched by CrAl ions, and etched by Cr ions; along Al2O3 with sulphur, TiO2 + Cr2O3 scale was also formed.

The unaffected region developed in areas where the coating and the substrate developed a protective scale, which comprised a thin and dense mixture of Al2O3 with small amounts of sulphur and Cr2O3 oxide. Based on stability diagram Al[BOND]S[BOND]O shown in the discussion section, presence of sulphur did not decrease the performance of the coating in the unaffected region due to the formation of A2O3(S4) phase instead of pure Al[BOND]S sulphide. The unaffected region did not crack due to the mismatch of CET between the coating and the substrate. Unaffected regions were detected in all the exposed samples.

3.3 Cross-sectioned investigations

3.3.1 Affected region

The cross-sectional SEM images in SE mode are presented in Figs. 7-9 and display the typical scale structure which developed in affected regions of the exposed materials.

image

Figure 7. SEM image in SE mode and EDS concentration profiles obtained from CrAlYN/CrN coated Ti–45Al–8Nb etched by Y ions after 1000 h of sulphidation/oxidation at 750 °C

Download figure to PowerPoint

image

Figure 8. SEM image in SE mode and EDS concentration profiles obtained from CrAlYN/CrN coated Ti–45Al–8Nb etched by CrAl ions after 1000 h of sulphidation/oxidation at 750 °C

Download figure to PowerPoint

image

Figure 9. SEM image in SE mode and EDS concentration profiles obtained from CrAlYN/CrN coated Ti45Al8Nb etched by Cr ions after 1000 h of sulphidation/oxidation at 750 °C

Download figure to PowerPoint

The top part of the scale in the sample etched by Y (Fig. 7) was composed of TiO2 with small amounts of Al2O3. Different structures were observed in the other two samples etched by CrAl and Cr (Figs. 8 and 9, respectively). In these samples, instead of a small amount of Al2O3 forming in the bottom layer of the TiO2 oxide scale, Cr2O3 formed.

Figure 7 reveals that, under the deposited coating a large region of internal sulphidation/oxidation developed. In the other two coated alloys, much smaller sulphidation/oxidation regions formed (Figs. 8 and 9, respectively). The size of this region depends on the crack dimensions, thus it can be suggested that sulphidation/oxidation region is proportional to the crack size; because more sulphur and oxygen can diffuse through a larger crack.

During exposure in the sulphidation/oxidation region, other phases including NbS, NbS2 and NbAl3 could form. In the region of internal sulphidation/oxidation (Fig. 7) sulphur and oxygen from ambient atmosphere diffused through the crack and reacted with Nb from the substrate. CrAlYN/CrN coating etched by CrAl formed TiAlN layer on the interface between the deposited coating and the substrate (Fig. 8).

In CrAlYN/CrN coating etched by Cr (Fig. 9) instead of a large sulphidation/oxidation region, a region with high Al content (∼65 at%) developed. This region is clearly visible in Fig. 9, region 3, this can suggest the formation of NbAl3 phase. The formation of NbAl3 was detected in all exposed materials by EDS X-ray mapping (Figs. 10-12), also all the exposed materials formed TiO2 and NbS2 phases (confirmed by Ti[BOND]S[BOND]O and Nb[BOND]S[BOND]O phase diagrams shown in the discussion section). The bright map of Nb and Ti (meaning a high concentration of the element) overlapped with a bright map of sulphur (high concentration of the element). These overlapped maps clearly suggest the formation of above mentioned phases. In contrast, darker area suggests a low concentration of the element. Similar to this work, other studies also reported the formation of TiO2, NbS2 and NbAl3 phases at the coating–alloy interface, or under the oxide scale of the uncoated alloy [16, 18].

image

Figure 10. Digimaps of CrAlYN/CrN coated Ti–45Al–8Nb etched by Y ions after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

image

Figure 11. Digimaps of CrAlYN/CrN coated Ti–45Al–8Nb etched by CrAl ions after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

image

Figure 12. Digimaps of CrAlYN/CrN coated Ti–45Al–8Nb etched by Cr ions after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

Figure 13 presents XRD spectra obtained from the exposed samples, XRD findings were confirmed by the calculations performed using FACTSAGE 6.2 software [19]. The formation mainly of the following phases: TiO2, Al2O3, Cr2O3 and NbAl3 were detected, no other phases were found through XRD investigations.

image

Figure 13. XRD analyses of CrAlYN/CrN coated Ti–45Al–8Nb etched by (A) Y, (B) CrAl and (C) Cr after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

Through the XRD analysis shown in Fig. 13 it is possible to suggest that the incorporation of Y ions as the etching agent plays an important role in the Al2O3 formation, as the peaks of Al2O3 are strong in CrAlYN/CrN coated Ti–45Al–8Nb etched by Y ions (Fig. 13A). In contrast, Al2O3 peaks became weaker when samples were etched by Cr or CrAl ions (Fig. 13B and C, respectively). The high temperature influence of Y ions was reported previously [20]. Incorporation of Y improves oxidation resistance, due to the formation of a dense and protective mixed α-(Al,Cr)2O3 scale retards further oxidation. Gil and coworkers [21] reported that implantation of Y ions enhance the selective oxidation of aluminium.

Rutile (TiO2) formation was also detected in all of the exposed samples: the strongest peaks of TiO2 are related to the Cr and CrAl etched samples (Fig. Fig. 13B and C, respectively), whereas Y showed relatively weaker peaks of TiO2 (Fig. Fig. 13A). During the XRD analyses NbAl3 phase was detected, the development of this phase suggest that Ti ions diffused outward from the substrate, while aluminium reacted with Nb and formed a thick and detectable layer of NbAl3 phase. The formation of NbAl3 was also reported by other researchers [16, 18].

3.3.2 Unaffected region

It was observed that the coating protected the substrate material from the corrosive atmosphere (pO2 = 10−18 Pa, pS2 = 10−1 Pa) until the onset of coating failure. Figure 14 shows the lack of TiO2 in the outer scale; the formation of dense (Al,Cr)2O3 oxide scale with small amounts of sulphur inhibited the outward diffusion of Ti ions and formation of TiO2. However, some microcracks developed during the exposure, which could become a nucleation point for the crack formation and TiO2 formation due to the outward diffusion of Ti from the substrate. Similar microcracks were found by Leyens et al. [22] on TiAlCrYN/CrN superlattice coatings in oxidation environment at 750 °C. The formation of singular oxide agglomerations was observed.

image

Figure 14. SEM image in SE mode and EDS of unaffected region of CrAlYN/CrN coated Ti–45Al–8Nb etched by Y after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

The EDS concentration profiles in Fig. 14 from the unaffected region show the formation of thin (1 μm) Cr and Al oxides with tiny amounts of sulphur on the outer scale. Underneath the Cr, Al oxide scale, the deposited coating was observed. Beneath the deposited coating, a region with a large amount of oxygen (25 at%) and sulphur (10 at%) with Al, Cr and Ti ions was formed. XRD analysis shown in Fig. 13A also reveals the formation of Al2O3 and Cr2O3, which supports EDS investigations. The results shown here are in good agreement with other studies, which reported the sulphidation behaviour of these type of coatings [23], in particular that Cr2O3 and Al2O3 prevent further sulphidation degradation of exposed materials. The EDS X-ray mapping in Fig. 15 performed on one of the exposed samples in an unaffected region clearly shows the formation of protective oxides.

image

Figure 15. EDS X-ray digimaps of the one of the exposed materials in unaffected region after 1000 h sulphidation/oxidation at 750 °C

Download figure to PowerPoint

It is important to note that the TiN phase was not detected underneath the deposited coating after exposure to sulphidized/oxidised atmosphere. The lack of TiN formation during the test shows that the deposited coating had high stability at 750 °C in a low partial pressure of oxygen (pO2 = 10−18 Pa). It was reported [24] that TiN may be formed by the decomposition of a (Ti,Nb)N phase in high Nb-containing alloys due to the high partial pressure of oxygen caused by the inward diffusion of oxygen. However here in this study it was observed that partial pressure of oxygen was too low to decompose (Ti,Nb)N phase. On the other hand, Hovsepian et al. [8] suggested that this TiN phase formed due to the inward diffusion of N2 from the base CrN and AlN phases into the substrate. In this study TiN was not detected in any regions, and it can be suggested that the formation of the TiN phase can be related to the inward diffusion of N2 from the decomposed CrN, AlN phases only in highly oxidising atmospheres where partial pressure of oxygen (pO2) is high enough.

4 Discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

The aim of this paper was to investigate, the sulphidation/oxidation resistance of the novel type of nanostructured multilayer CrAlYN/CrN coatings deposited on a Ti–45Al–8Nb alloy (pre-etched by Cr, CrAl and Y ions) at 750 °C in environment of low oxygen (10−18 Pa) and high sulphur (10−1 Pa) partial pressures for up to 1000 h. The degradation of ceramic coatings was associated with the development of two regions, called here ‘affected’ and ‘unaffected’. These regions are distributed randomly on the surface of the exposed material, discussion in this study is divided further on two sub-sections; Section 'Affected region' affected region, Section 'Unaffected region' unaffected region.

4.1 Affected region

The scale degradation as was mentioned in this work previously in the affected region, was related to the formation of non-protective and porous TiO2 oxide + Al2O3 with tiny amounts of sulphur in the places, where the coating cracked. Low value of Ti diffusion coefficient [17] led to the formation of TiO2 oxide on the top part of the scale. The degradation of the applied coatings can be also enhanced by the presence of hydrogen in the atmosphere (H2/H2S/H2O). Haddad and Eliezer [25], found that hydrogen can significantly increase the brittleness and change the microstructure of the γ-TiAl alloy by the formation of Ti hydrides (TiH). However more work needs to be done in order to rationale this effect. Each of the exposed coatings showed a lack of resistance to temperature changes: during discontinuous sulphidation/oxidation test, cracks formed in all exposed materials at 750 °C, in sulphidation/oxidation atmosphere. Similar results were obtained by Lasanta et al. [10]. The crack formation was the key point influencing the sulphidation/oxidation resistance of nanostructured ceramic coatings. Because the concentration of sulphur was not enormous in the formed oxide scale, thus the influence of sulphur in overall corrosion degradation can be neglected.

When the coating cracked mainly TiO2 and Al2O3 phases with tiny amounts of sulphur developed, beside TiO2 and Al2O3 according to Nb[BOND]S[BOND]O stability diagram shown in Fig. 17 also Nb2O5 should be formed (Al[BOND]S[BOND]O and T[BOND]S[BOND]O stability diagrams are presented in Figs. 16 and 18, respectively). The oxidation process started from the decomposition of H2O at 750 °C according to the following reaction:

  • display math(1)

The released oxygen from the reaction above, further react with Al, Ti and Nb in order to develop Al2O3, TiO2 and Nb2O5, respectively:

  • display math(2)
  • display math(3)
  • display math(4)

Gibbs free energy for the Al2O3, TiO2 and Nb2O5 can be calculated using following equations [26]:

  • display math(5)
  • display math(6)
  • display math(7)

The phases presented above are confirmed by stability phase diagrams for Al[BOND]S[BOND]O, Ti[BOND]S[BOND]O and Nb[BOND]S[BOND]O presented in Figs. 16-18, respectively. The stability phase diagrams were calculated using FACTSAGE 6.2 software [19] in log pO2 = 10−18 Pa and log pS2 = 10−1 Pa at 750 °C. The calculated values of Gibbs free energy formation from Equations (5)-(7) are shown in Table 1. The results show, that Nb2O5 should form first due to the lowest value of Gibbs free energy formation.

image

Figure 16. Phase stability diagram for Al[BOND]S[BOND]O in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C

Download figure to PowerPoint

image

Figure 17. Phase stability diagram for Nb[BOND]S[BOND]O in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C

Download figure to PowerPoint

image

Figure 18. Phase stability diagram for Ti[BOND]S[BOND]O in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C

Download figure to PowerPoint

Table 1. The Al2O3, TiO2 and Nb2O5 inline image formation values at 750 °C
inline image (kJ/mol)T = 750 °C
Al2O3−1349
TiO2−733
Nb2O5−1381

According to Gibbs free energy formation, phase stability diagrams for Al, Ti and Nb at 750 °C and activities of Al, Ti and Nb, the development of TiO2 and Al2O3 is predominant, and can be rationalised in the following way.

The development mainly of TiO2 and Al2O3 is due to the much higher concentration of Al and Ti in the coating and in the alloy than Nb. Due to the cracked coating the outward diffusion of Ti from the alloy was possible and formation of TiO2 was observed. Additionally the values of Al and Ti activities (Fig. 19) favour the formation of TiO2 and Al2O3.

image

Figure 19. Minimum activities of Al, Ti and Nb required for the formation of Al2O3 and TiO2 and Nb2O5 at 750 °C

Download figure to PowerPoint

Therefore the formation of Nb2O5 should not be expected due to the low concentration and low activity, but according to Nb[BOND]S[BOND]O phase stability diagram presented in Fig. 17, Nb2O5 in the present conditions (pO2 = 10−18 Pa and pS2 10−1 Pa at 750 °C) can form, when concentration of Nb in TiAl ordered alloys is much higher than 8 at%. The calculations of Ti, Al and Nb activities were performed as follows:

  • display math(8)
  • display math(9)
  • display math(10)

Figure 19 shows the results of these calculations. It is confirmed that the formation of Nb2O5 phase could not be achievable due to the highest value of the activity of Nb ions (aNb = 10−13) and low concentration in the alloy (8 at%), thus Nb is less active, than Al and Ti. The activity of Ti is slightly higher than that for Al (aAl ∼ 10−20 for Al, aTi ∼ 10−18 for Ti, respectively); finally Al2O3 should form first than TiO2. On the other hand Ti ions diffuse slightly faster than Al ions [17]:

  • display math
  • display math

Thus the formation of the oxide scale, in sulphidation/oxidation environment at 750 °C can be compared to the formation of the scale on the uncoated material, because the access of oxygen and sulphur from ambient atmosphere is not limited due to the cracking of the coating.

Thus when the coating cracked, transport of Ti ions and formation of TiO2 phase was slightly faster than the transport of Al ions and formation of Al2O3 phase. Depletion of Ti ions from the bulk material Ti–45Al–8Nb (at%) and increased activity of Al led to the formation of Al2O3.

It is worth mentioning, the role of Nb in the ordered γ-TiAl based alloys, the addition of Nb ions to the alloy is believed to decrease the concentration of oxygen vacancies in an oxide scale (TiO2), and slows down the diffusion transport through the oxide scale to the substrate. The high concentration of Nb near modified coating/substrate interface might reduce the outward diffusion transport of Ti and slow down the oxidation process. TiO2 is semiconductor n-type and contains the interstitials based on Ti and Ti existing together with doubly ionized oxygen vacancies [27]. Kekare and Aswath [28] indicated that the doubly ionised oxygen vacancies are responsible for the kinetic rate of growth of the TiO2 scale over Ti, thus, any dopant element in the titanium oxide scale (TiO2) that is able to minimize the concentration of these vacancies will reduce the oxidation rate. Nb atoms substitute the Ti site in the TiO2 lattice, reducing the number of interstitial oxygen ion vacancies in the oxide [29]. It is also important to note in this instance the number of phases (layers) developed during experiments is strongly dependant on the time of exposure. Increase of exposure time leads to an increase in the number and thickness of the layers.

It was observed that the depleted zone of Ti ions was filled by the oxygen and sulphur diffused inward, and formed the internal oxidation and sulphidation region, where mainly Al2O3 and NbS2, Nb2O5 formed, these results are confirmed by stability phase diagrams presented in Figs. 16 and 17, respectively. Additionally, due to the outward diffusion of Ti ions from the substrate, concentration and activity of Al increased, resulting in NbAl3 formation underneath the internal sulphidation/oxidation region.

The exposed samples, beside oxidation also underwent sulphidation process. According to the phase stability diagrams of Al[BOND]S[BOND]O and Nb[BOND]S[BOND]O in Figs. 16 and 17, beside NbS2 also Al2O3(S4) could form in this condition. The formation of sulphides is initiated by the decomposition of H2S compound according to the following reaction:

  • display math(11)

The released sulphur from decomposed H2S (reaction 11) and oxygen from decomposed H2O (reaction 1) reacts further with Al and Nb to form NbS2 and Al2O3(S4) sulphides, respectively, according to the following reactions:

  • display math(12)
  • display math(13)

Due to the sulphidation reaction and decomposition of H2O, H2 is released to the test environment. As a result the reaction between H2 and Ti which diffused outwards from the alloy to from a brittle titanium hydrate (TiH) is likely to occur:

  • display math(14)

Similarly to the results presented in this study, also Haddad and Eliezer [25] found the formation of TiH phase during sulphidation experiment.

4.2 Unaffected region

The discussion of the results achieved from the unaffected region is related mainly with the formation of Al2O3 and Cr2O3 phases; here sulphidation reactions are neglected, the formation of Al2O3 + (S4) phase as was mentioned previously did not change the performance of the unaffected region. This subsection focused on the formation of Al2O3 and Cr2O3. The oxidation of Al and Cr can be presented by the following way:

  • display math(15)
  • display math(16)

For both reactions it is possible to estimate Gibbs free energy of formation, according to the equations below:

  • display math(17)
  • display math(18)

where K is the equilibrium constant, R the gaseous constant (8314 J/mole K), T the temperature (K), and a is the activity. The equilibrium constants for reactions (17) and (18) of the formation of Cr2O3 and Al2O3 oxides can also be described as:

  • display math(19)
  • display math(20)

The calculated values of Gibbs free energy formation from Equations (19) and (20) are shown in Table 2. The minimum activities of Al, and Cr were calculated based on Equations (19) and (20) and are shown in Fig. 20. According to Cr[BOND]S[BOND]O phase stability diagram shown in Fig. 21, in the mixture of pO2 = 10−18 Pa and pS2 10−1 Pa, only Cr2O3 exclusively can be formed.

Table 2. The Al2O3 and Cr2O3 inline image formation values at 750 °C
inline image (kJ/mol)T = 750 °C
Al2O3−1349
Cr2O3−580
image

Figure 20. Minimum activities of Al and Cr required for the formation of Al2O3 and Cr2O3 at 750 °C

Download figure to PowerPoint

image

Figure 21. Phase stability diagram for Cr[BOND]S[BOND]O in high partial pressure of S2 (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa) at 750 °C

Download figure to PowerPoint

In order to form Cr-sulphide, partial pressure of oxygen should be decreased as low as log pO2 10−25 Pa, however in this study pO2 was much higher than 10−25 Pa, thus Cr-sulphide could not be formed.

Based on above calculations Al2O3 phase is more stable than Cr2O3, the conditions are also more favourable for the formation of Al2O3 than Cr2O3. The free energy formation at 750 °C for Al2O3 (−1349 kJ/mole) has a more negative value than that for Cr2O3 (−580 kJ/mole). Thus in the unaffected region mainly Al2O3 and Cr2O3 phases can form.

The values in Table 3 show that, Al diffuses slightly faster than Cr, thus Al is more favourable to form Al2O3 than Cr to form Cr2O3 phase. The average intrinsic diffusivities of Al and Cr were calculated by the inverse method (IM). In this method to calculate the intrinsic diffusivities, it is essential to use a concentration profile after an arbitrary time (concentration profiles of the element). The detailed description of the IM can be found elsewhere [30]. The inverse method was used previously by Datta and coworkers [31] and recently by Dudziak et al. [32] to calculate intrinsic diffusivities in CrAl2%YN coated γ-TiAl system. The intrinsic diffusivities are presented in Table 3.

Table 3. The average intrinsic diffusion coefficients of Al and Cr at 750 °C calculated by the IM [28]
Diffusion coefficient (cm2/s)
TemperatureDAlDCr
750 °C1.97 × 10−128.12 × 10−13

The formation of Al2O3 phase decreased the activity of Al in the deposited coating; this decrease led to the formation of Cr2O3 oxide underneath Al2O3 phase, due to the increase of Cr activity in the deposited CrAlYN/CrN coating. It was also observed that despite the formation of Al2O3 and Cr2O3 in the unaffected region, sulphur and oxygen diffused inward underneath the deposited coating probably through the microcracks, and these microcracks could form by the mismatch in CET, between the deposited coating and the substrate. Hovsepian et al. [8] demonstrated also that microcracks could form due to the formation of sulphides and oxides, and this formation increases the volume of expansion which is associated with a phase change. Here it can be suggested that these microcracks are a prelude to the development of larger cracks shown in this work. Thus, the degradation of nanostructured coatings can be related to several factors: mismatch of a CET of the coating and of the substrate (crack formation), formation of sulphides and oxides underneath the deposited nanostructured coating.

5 Conclusions

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

The aim of this study was to show the possibility of adopting nanomultilayer coatings on Ti–45Al–8Nb (at%) alloys in order to improve sulphidation/oxidation resistance at 750 °C in high partial pressure of sulphur (pS2 = 10−1 Pa) and low partial pressure of oxygen (pO2 = 10−18 Pa). The conclusions of this study are as follows:

  1. Multilayered nanostructured coatings CrAlYN/CrN underwent defect and crack formation at 750 °C, due to the mismatch of CET between the substrate and the deposited coating.
  2. All coatings exposed to sulphidation/oxidation atmosphere (pO2 = 10−18 Pa, pS2 = 10−1 Pa) at 750 °C for 1000 h showed the development of different regions undergoing various degrees of attack. Two regions named ‘unaffected’ and ‘affected’ were found.
  3. All exposed materials do not follow the parabolic rate law.
  4. The affected regions showed the formation of defects in the nitride coating CrAlYN/CrN. Due to the thermal exposure, the formation of non-protective TiO2 was observed.
  5. The unaffected region developed a dense (Al,Cr)2O3 oxide scale.

Acknowledgment

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References

The authors gratefully acknowledge the financial support from The European Union (INNOVATIAL-Project NMP3-CT-2005-515844).

References

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental procedure
  5. 3 Results
  6. 4 Discussion
  7. 5 Conclusions
  8. Acknowledgment
  9. References