# Fileset

[M187429 _revised version_ver2.pdf](https://mdr.nims.go.jp/filesets/4b77bc30-3d3e-4943-9d1f-ce73789a64c0/download)

## Creator

[Chihiro Tabata](https://orcid.org/0000-0001-6597-4998), [Toshio Osada](https://orcid.org/0000-0003-1539-9264), [Takahito Ohmura](https://orcid.org/0000-0001-7528-566X), Tadaharu Yokokawa, [Kyoko Kawagishi](https://orcid.org/0000-0001-7652-9232), Shinsuke Suzuki

## Rights

[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Interfacial Strength Evaluation Between Sulfur-Segregated Al2O3 and Ni–Al Single Crystal Alloy Using Nanoindentation](https://mdr.nims.go.jp/datasets/d0a93e59-8190-4b96-a625-82f6f44cd35a)

## Fulltext

Interfacial Strength Evaluation between Sulfur-segregated Al2O3 and Ni-Al Single 1 Crystal Alloy Using Nanoindentation 2  3 Chihiro Tabata1,2,*, Toshio Osada1, Takahito Ohmura1, Tadaharu Yokokawa1, Kyoko Kawagishi1,2, and 4 Shinsuke Suzuki2,3,4 5  6 1. National Institute for Materials Science (NIMS), 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan 7 2. Department of Materials Science, Waseda University,3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, 8 Japan 9 3. Department of Applied Mechanics and Aerospace Engineering, Waseda University, 3-4-1 Okubo, 10 Shinjuku-ku, Tokyo 169-8555, Japan 11 4. Kagami Memorial Institute for Materials Science and Technology, Waseda University, 2-8-26 12 Nishiwaseda, Shinjuku-ku, Tokyo 169-0051, Japan 13 *Corresponding author: chihiro448@akane.waseda.jp (C. Tabata) 14   15 Keywords: Nanoindentation, Oxidation, Sulfur segregation, Interfacial strength, Ni-base single crystal 16 alloy 17  18 Abstract 19 Ni-base superalloys have excellent oxidation resistance, but impurity S drastically decreases their 20 properties. This is due to the segregation of S to the oxide/substrate interface, but direct and quantitative 21 measurements of the interfacial strength in relation to the S segregation level have not been widely 22 conducted. The objective of this research is to quantitatively analyze the interfacial strength between the 23 Al2O3 layer and Ni-base substrate interfacial strength, depending on the S segregation level, using 24 nanoindentation. Ni-9.8 wt.% Al alloys were prepared by melting the material using either an Al2O3 crucible 25 (high Sinterface alloy) or a CaO crucible (low Sinterface alloy). Nanoindentation tests using a 60 degree 26 pyramidal diamond indenter were conducted, and the cross-sections of both specimens exposed the (100) 27 plane. Indentation near the interface formed cracks at the boundary between the two layers, which can be 28 observed as pop-ins in the load-depth curves. The amount of load at the initial pop-in most likely represents 29 the interfacial strength between the Al2O3 layer and Ni-base substrate. A Weibull analysis of results showed 30 that suppression of the S segregation level increased the critical β scale parameter for crack formation by 31 650 μN. This suggests that we were able to successfully compare the effect of S segregation on the 32 interfacial strength between the Al2O3 layer and the Ni-base substrate quantitatively. 33  34 Introduction 35 about:blank Ni-base single-crystal superalloys are designed to have excellent oxidation resistance and mechanical 36 properties at high temperatures of about 1000 °C. To improve the thermal efficiency of jet engines and gas 37 turbines, single-crystal superalloys such as TMS-238, which has temperature capabilities higher than 38 1100 °C, has been created [1]. For these superalloys, cyclic oxidation resistance plays an important role. 39 However, impurities such as S, which enter the alloys from the fuel and the environment used, are known 40 to be detrimental to the high temperature oxidation, even at a ppm level [2-8]. This is most likely caused 41 by decrease in the adhesion of the interface between the Al2O3 oxide layer and the Ni-base substrate, due 42 to the S segregation at this interface. To avoid any S segregation, according to Smialek, the ideal S 43 concentration without the addition of reactive elements would be around 0.1 ppm; for wall thickness 44 relevant to turbine components, it needs to be <1 ppm to observe oxidation benefit [6].  45 Additionally, with the rise in the cost of rare metals, methods to reuse materials will be required, and 46 techniques to remove impurity S will be crucial. Utada et al. developed the direct recycling method which 47 uses CaO crucibles rather than the conventional Al2O3 crucibles to remelt the used turbine blades, made of 48 Ni-base superalloys [9]. This not only removes the S from the alloy, but also limits the S segregation level 49 at the Al2O3/alloy interface  by trapping S within the substrate by the formation of CaS, thus improving its 50 oxidation resistance [7,8]. Using this method, we were able to see improvement in oxidation resistance even 51 for alloys with S content higher than 1 ppm, due to the decrease in the S segregation level at the interface 52 observed using scanning transmission electron microscopy with energy-dispersive X-ray spectroscopy 53 (STEM-EDS) and 3D atom probe (3DAP) [7,8]. These results suggest that the critical factor is the S 54 segregation level at the interface, rather than the S concentration of the alloy. 55 Therefore, to determine the S segregation level acceptable for creating oxidation resistant alloys from both 56 ingots and reused materials, it is important to analyze the oxide/substrate interfacial strength quantitatively. 57 Several authors reported that calculations and simulations had made it possible to predict the differences in 58 interfacial strength caused by S segregation. For example, density functional theories have been used by 59 several researchers to determine the changes in adhesion of Al2O3 layer and Ni-base substrate by the 60 existence of S [10-12]. Other methods such as macroscopic atom models have also been used by Bennett 61 et al. to determine the adhesion of oxides and metals [13]. However, this has been difficult to quantify the 62 segregation level and the changes in interfacial strength, as well as proving these results experimentally. 63 Scratch or pull-off tests have been used previously to assess the adhesion of oxide layers [6,14-15]. Hou et 64 al. determined the S segregation level using Auger electron spectroscopy and compared the interfacial 65 strength using Vickers micro-indenters and tensile testers [4,16]. They have shown that the oxide/substrate 66 interface had weakened due to the existence of S. But these methods are difficult to use when trying to 67 compare the differences directly and quantitatively for the interfacial strengths caused by the S, specifically 68 at the Al2O3/alloy, due to the sizing and adherence of the interface and the formation of various layers of 69 oxides.  70 Therefore, the objective of this research is to incorporate nanoindentation to quantitatively clarify the 71 interfacial strength between the Al2O3 layer and the Ni-base substrate, depending on the S segregation level. 72 Several factors such as the placement of the indenters within the specimens as well as the reliability of the 73 tests were considered, and what had occurred during the nanoindentation tests was analyzed as well.  74  75 Experimental Methods 76  Ni-9.8 wt.% Al single crystal (SC) alloys, used in a previous study [8], were also used for this research to 77 simplify the interfacial structure by constraining the oxide species formed, while retaining similar γ/γ’ 78 structures as the Ni-base superalloys.  Alloys with different levels of S segregation at the interface were 79 created by melting the materials using an Al2O3 crucible (alloy (high Sinterface)), and a CaO crucible (alloy 80 (low Sinterface)). The alloys contained 2 ppm of S and <1 ppm of Ca for alloy (high Sinterface), and 3 ppm of S 81 and 6.2 ppm of Ca for alloy (low Sinterface), both of which were mirror polished and chemically analyzed 82 using glow discharge mass spectrometry [8]. Although the S contents were about the same, S segregation 83 level at the interface between Al2O3 and Ni-Al alloy was significantly higher for alloy (high Sinterface) 84 compared to alloy (low Sinterface) [8]. In the previous research the samples were oxidized at 1100°C for 1 h 85 and cooled in air, and the S segregation levels were measured using STEM-EDS, during which, the alloy 86 (high Sinterface) had around 2.0 at.% S segregation at the interface, while the alloy (low Sinterface) showed less 87 than 0.5 at.% S segregation. Note that how much S segregation is present at the scale-metal interface 88 depends on the S reservoir, which is proportional to metal thickness, and the time and temperature exposed, 89 all of which were consistent between the two types of samples. Separate samples created from the same 90 ingot were prepared for nanoindentation measurements. To match the conditions to the previous research 91 and to grow the Al2O3 scale and limit the oxide spallation as much as possible, the specimens were oxidized 92 at 1100 °C for 1 h in air and cooled off within the furnace at 10 °C/min. Each specimen was cut into a 93 cuboid with 5 mm in height and placed on an Al-based platform so that the Al2O3/alloy interface faced 94 upwards. The crystal orientations of the specimens were measured using the Laue X-ray backscattering 95 method (RIGAKU SA-HF3S) and adjusted so that all specimens faced the (100) plane during testing. The 96 cross-sections of each specimen were both mirror and vibration polished (Saphir Vibro, ATM).  97 Nanoindentation (HYSITRON, TI950 TriboIndenter) tests were conducted to measure the differences in 98 the interfacial strength of the alloys, using a diamond, 3-sided pyramid tip with an angle of 60 degrees. 99 Figure 1 shows the schematic diagram of the nanoindentation tests conducted for both alloy (high Sinterface) 100 and alloy (low Sinterface) in this research. The initial indent position was in the Al2O3 oxide layer, 0.5 μm 101 away from the interface of the protective oxide layer and Ni-base alloy for testing at the interface (Fig. 102 1(b)).  This is far enough from the interface so that the cracks forms between the two layers during testing, 103 but not too far so that the indenter does not fall within the Al2O3 layer. Only for the Al2O3 layer, initial 104 indent positions were set at around 1.0 nm away from the interface (Fig. 1(a)). Tests were also conducted 105 within the substrate as well (Fig. 2(c)). The distance was determined by the observation of the 106 oxide/substrate interface using atomic force microscopy, where the microscope was incorporated within the 107 TI950 TriboIndenter. The tests conducted near the interface were executed 17 times for each specimen, 108 where each test was conducted at 600 μN/s loading and unloading rate and 10 s holding time in between 109 (only for test conducted directly at the interface in the discussion section, it was conducted at 750 μN/s 110 loading and unloading rate and 10 s holding time in between). Scanning Electron Microscope (SEM) images 111 of each indentation marks were taken afterwards using field emission scanning electron microscope (FE-112 SEM, ZEISS GeminiSEM300). Electron backscatter diffraction (EBSD, ZEISS GeminiSEM300) maps 113 were taken at the interface of both specimens as well.  114  115  116 Figure 1 –  Schematic diagram of the nanoindentation test sample and the position of indentation for testing 117 at (a) Al2O3 layer, (b) interface, and (c) substrate. 118  119 Results 120 Figure 2 (a) shows an example of load (P)-penetration depth (h) curves of alloy (low Sinterface), tested within 121 the Al2O3 layer, near the Al2O3/alloy interface, and on the Ni-Al alloy. The white plots within the SEM 122 images are the locations of where the indenters were placed initially. For indentation near the interface, 123 deformation by the nanoindentor started off within the Al2O3 layer, resulting in the same slope as the curves 124 for the Al2O3 oxide layer. The large strain burst event, called a ‘pop-in’, is triggered by crack formation on 125 the interface as shown in the SEM image. This load is surmised to correspond to interfacial cracking of the 126 alloy, since there appears a significantly higher pop-in load for the indents within the Al2O3 layer. In this 127 study, the initial pop-in load (Pc) was determined to be an indicator of interfacial strength. Further, after the 128 large initial pop-in on interface, the indent tip moved from initial position to the alloy. Due to the tip 129 movement, the slope of P-h curve near the interface significantly decreased compared to before the pop-in 130 and exhibited a shallower slope than that on the alloy. Furthermore, the P-h curve included several small 131 pop-in events. Thus, the decreased slope includes both alloy deformation and crack propagations.  132 Figure 2 (b) shows the results for the same tests conducted on alloy (high Sinterface). The Pc taken from the 133 tests conducted within the Al2O3 was 5503 μN for alloy (high Sinterface) and 5671 μN for alloy (low Sinterface), 134 and the curves are similar in shape. The same shift in the P-h curve could be observed in both Figs. 2 (a) 135 and (b), showing that this test could be conducted for both alloys. The occurrence of pop-in was judged by 136 a rapid increase of penetration depth under nearly constant load. The rapid increase was judged by the 137 criterion that the increase of the penetration depth Δh from the previous plot was larger than 1.0 nm (3.0 138 nm only for Al2O3 layer).  This was done to eliminate the effect of small crack formations at the beginning 139 of the test, most likely effects from the porous Ni-Al-O layer formed nearby. 140  141  142 Figure 2 –  P-h curves of (a) Al2O3 layer (gray line on left), interface (blue), and alloy (gray line on the 143 right) for Ni-Al melted using a CaO crucible (low Sinterface). (b) Al2O3 layer (gray line on left), interface 144 (red), and alloy (gray line on right) for Ni-Al melted using an Al2O3 crucible (high Sinterface). 145  146 Figure 3 shows five P-h curves near the interface for (a) alloy (high Sinterface) and (b) alloy (low Sinterface), 147 and one of the SEIs after the indentation tests. The crack lengths associated with the indents were slightly 148 larger for alloy (high Sinterface) compared to alloy (low Sinterface), and all test results showed crack formation 149 at the Al2O3/alloy interface. This suggests that the method is reproducible for detecting interfacial cracking 150 events. Further, Pc for alloy (high Sinterface) is lower than that for alloy (low Sinterface). However, Pc for both 151 alloy (high Sinterface) and (low Sinterface) show large scatters because of the brittle fracture on the interface. 152 These results suggest that the Al2O3/alloy interfacial strength was lower for alloy (high Sinterface) than alloy 153 (low Sinterface).  154  155  156 Figure 3– SEIs of the indentation marks and load-depth curves for (a) Ni-Al melted using an Al2O3 157 crucible (alloy (high Sinterface)) and (b) Ni-Al melted using CaO crucible (alloy (low Sinterface)). Arrows 158 show where the initial pop-in had occurred.  159  160 Discussion 161  To statistically evaluate the differences in Pc for brittle interfacial cracking of the two alloys, Pc values 162 were evaluated based on two-parameter Weibull distributions as shown in Figure 4. Here, the median 163 ranking method;  𝐹(𝑃c) = 𝑖 − 0.3 𝑁 + 0.4⁄ , where i is the Pc order, and N is the number of tests, was used. 164 The Weibull equation is as follows:  165 ln ln11 − 𝐹(𝑃c)= 𝑚(ln 𝑃c − ln𝛽) (1) 166 where m is the Weibull parameter, and β is the scale parameter. The slope and intercept of approximated 167 linear lines shown for each alloy in Fig. 4 are used to determine m- and β-values. The red triangles represent 168 the results for alloy (high Sinterface), and the blue squares represent the results for alloy (low Sinterface). For 169 alloy (high Sinterface), 𝑚 = 3.48, and 𝛽 = 2248 μN, while for alloy (low Sinterface), 𝑚 = 6.39, and 𝛽 = 2898 170 μN. The values of the coefficient of determination were 0.958 and 0.989, respectively, indicating that both 171 linear lines are well-fitted. This implies that alloy (high Sinterface), which has higher S segregation at the 172 interface between Al2O3 layer and Ni-base substrate, shows significantly lower strength and larger scatter 173 than low S interface. Therefore, it can be said that alloy (low Sinterface) had 650 μN higher Pc compared to 174 alloy (high Sinterface), which also directly relates to the interfacial strength between the Al2O3 layer and Ni-175 base substrate, due to the suppression of S segregation. The m- and β-values consider the effect of not only 176 the interfacial strength but also the shape of the interface and the location of initial indent position as well. 177 However, the shape of interface and the locations of initial indent position were selected to be about the 178 same for each test. Therefore, the effects are most likely small and significant differences were observed in 179 the pop-in load (Pc), making this test method effective for differentiating interfacial strength, most likely 180 for superalloys as well. 181  182  183 Figure 4 – Weibull plots of the initial pop-in with Δh larger than 1.0 nm for Ni-Al alloys with different 184 interfacial S segregation level. The red triangles indicate the results for Ni-Al melted using an Al2O3 185 crucible (alloy (high Sinterface)), and the blue squares represents the results for Ni-Al melted using a CaO 186 crucible (alloy (low Sinterface)).  187  188 Next, the most likely process during this test was evaluated. When the indentation is conducted directly 189 at the interface, it proved difficult to form cracks between the two layers. Figure 5 shows the (a) SEM image 190 and (b) the P-h curve of when nanoindentation was conducted directly at the interface of the Al2O3 layer 191 and Ni-base substrate for the alloy (low Sinterface). Here, large crack formation can be seen within the Al2O3 192 layer; however, pop-in events cannot be observed in the P-h curve. It is evident that the indenter has slid 193 towards the substrate before creating a clear indent at the interface. This is most likely due to the difference 194 in the relative hardness of the two layers which were 13.7 GPa for Al2O3 and 3.64 GPa for the substrate, 195 respectively. The P-h curve is similar to that taken at the substrate in Figs. 2 (a) and (b), suggesting that the 196 test took place mostly in the substrate. From this result we can assume that the indenter has the tendency to 197 deviate towards the substrate.  198  199 lnln(11  )2.00.0-2.0-4.06.8 7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4ln(  )Initial pop-in load  ( N)Fractureprobability(%)1000 2000 3000 4000103050709095High Sinterface(Al2O3 crucible)Low Sinterface(CaO crucible)63.2% probability     N99 200 Figure 5 – Results of nanoindentation test conducted directly at the interface. (a) SEI of indentation mark 201 conducted directly at the interface for alloy (low Sinterface) with the white dotted lines representing the indent 202 perimeter and (b) the load-depth curve of this result.  203  204 Several factors may affect these nanoindentation test results. For example, the crystal orientations and 205 grain sizes of the Al2O3, as well as the thickness of these layers may alter the Pc values. To evaluate such 206 points, EBSD analyses were conducted at the interface of both alloy (high Sinterface) and alloy (low Sinterface). 207 Figure 6 shows the crystal orientations of the Al2O3 grains and the substrate of both alloy (high Sinterface) and 208 alloy (low Sinterface). The substrate, which is the bottom half of the figures, faces the (100) direction for both 209 alloys and are shown in red. The small grains shown in the top half of the images most likely represents the 210 Ni-Al-O layers. The Al2O3 grains, shown right on top of the substrate, are around 1 μm in width and between 211 0.4 and 0.7 μm height for both alloys, showing that the size of these grains is similar for both alloys. The 212 crystal orientations for these grains, however, varies greatly. But for this research, the same indentation 213 tests were conducted several times at different locations. By comparing the results using Weibull 214 distribution, the differences caused by the crystal orientations of Al2O3 are incorporated into the results. We 215 also assume that there may be a range of values with the increase in monolayers of S at the interface. 216 However, factors such as these Al2O3 crystal orientations will also affect the results, and currently it 217 is difficult to distinguish which factors affect the values of the interfacial strength the most. Therefore, 218 this is also something to consider for future works. 219  220  221 Figure 6 – EBSD crystal orientation maps of Al2O3 and Ni-base substrate taken at the interface of (a) alloy 222 (high Sinterface) and (b) alloy (low Sinterface). 223  224 Figure 7 shows the schematic diagram depicting what is most likely occurring during these 225 nanoindentation tests; (a) the schematic P-h curve, (b) the top view, and (c) the side view of the tests 226 conducted. (i) At the beginning of the test, indenter is placed on the oxide layer, 0.5 μm away from the 227 interface. (ii) The indenter penetrates within the oxide, creating stress regions shown in yellow. (iii) The 228 stress gradually increases as the indenter penetrates deeper within the sample. Before the indenter itself 229 reaches the interface, stress will be applied to the interface indirectly. This will most likely lead to crack 230 formations at the interface. (iv) Because the substrate is mostly γ phase due to the Al depletion by oxidation, 231 it easily deforms compared to the Al2O3 oxide layer, which is ceramic, and the indenter gradually slides 232 toward the interface. The indenter is pushed onto the interface, leading to the growth of the crack at the 233 interface. When the indent size increased enough in h and the indenter edge with high stress concentration 234 approached the interface, pop-in event is observed in the P-h curves. Afterwards, the indenter will continue 235 moving towards the substrate, leading to the large pile-up at the end.  236  237  238 Figure 7 – Schematic diagram of (a) load-depth curve, (b) top-view and (c) side-view of the nanoindentation 239 test conducted near the interface.  240  241 (a) High Sinterface (Al2O3 crucible)1  mAl2O3 grains (b) Low Sinterface (CaO crucible)1  mAl2O3 grains Oxide layerNi-basesubstrateOxide layerNi-basesubstrateAl2O3Ni Ni-Al-ONi-Al-OOxide Substrate Substratecrack Oxide Substrate i 0.5  mIndenterOxide SubstrateLarge pile-upslideIndenterIndenterIndentationmark iv SubstrateOxide  iv  b   op-view Oxide Substrate a  Schematic    curve Oxide SubstrateInitial set point      Initial pop-inCrack initiation Oxide Substrate0.5  mIndentationmarkInitial set pointInterface i  ii  iii  oad   N  enetration depth   nm  i  ii  iii  iv Initial pop-in c  Side-view InterfaceIndenter ii Oxide Indenter iii slideCrackConclusions 242 To conclude, the following have been clarified regarding the measurement of the interfacial strength 243 between the Al2O3 layer and Ni-base substrate of the two types of Ni-9.8 wt.% Al alloys; an alloy melted 244 using an Al2O3 crucible (high Sinterface) and an alloy melted using a CaO crucible (low Sinterface):  245 1. Nanoindentation tests were conducted so that the side of the triangular indenter was parallel to the 246 interface of the Al2O3 layer and Ni-base substrate. The indenter was placed on the Al2O3 layer, 0.5 247 μm away from the interface, since placing it directly at the interface causes the indenter to slide 248 towards the substrate before crack formation. The large difference in the hardness most likely leads 249 to this phenomenon.  250 2. Clear differences in the interfacial strength between alloys were observed using nanoindentation. 251 According to the P-h curves, alloy (high Sinterface), which has higher S segregation level at the interface 252 of the Al2O3 layer and Ni-base substrate, showed that the initial pop-in load, Pc, was lower compared 253 to alloy (low Sinterface). The tests conducted near the interface, within Al2O3, and on the substrate 254 showed that the pop-in seen in the P-h curves most likely represents the crack initiation at the interface.  255 3. The Weibull distributions showed that alloy (low Sinterface) had 650 μN higher median pop-in load than 256 alloy (high Sinterface), due to the suppression of the S segregation level. The fitted linear lines showed 257 high coefficient of determination values, indicating that the values are reliable. This method can 258 directly compare the interfacial strength quantitatively and is most likely suitable for the measurement 259 of complex superalloys as well.  260 We also plan on testing using TMS-238, a sixth generation Ni-base single crystal superalloy, since the S 261 segregation level has been measured using 3DAP and STEM-EDS in previous research [7]. Ideally, the 262 nanoindentation tests should also be conducted at elevated temperatures, to recreate the high temperature 263 conditions and the effects of cyclic oxidation. However, that will be for future work.  264   265 Acknowledgements 266 This research was financially supported by Grant-in-Aid for JSPS Fellows Grant Number 23KJ2024. This 267 paper is also based on results obtained from a project, JPNP21007, commissioned by the New Energy and 268 Industrial Technology Development Organization (NEDO).  The authors would like to express our gratitude 269 to Dr. Makoto Osawa for the helpful discussions. We also would like to thank Mr. Yuji Takata for the 270 preparation of the single crystal alloys, Mr. Takuma Kohata for the SEM observations, and Ms. Eri 271 Nakagawa for assisting the nanoindentation tests. 272  273 References  274 [1] Kawagishi K, Yeh AC, Yokokawa T, Kobayashi T, Koizumi Y, Harada H (2012) Development of an 275 oxidation-resistant high-strength sixth-generation single-crystal superalloy TMS-238, Superalloys 2012. 276 189-195. doi: 10.1002/9781118516430 277 [2] Smith MA, Frazier WE, Pregger BA (1995) Effect of sulfur on the cyclic oxidation behavior of a single 278 crystalline, nickel-base superalloy. Mater. Sci. Eng. A. 203:388-398. doi:10.1016/0921-5093(95)09819-4 279 [3] Molins R, Rouzou I, Hou P (2007) A TEM study of sulfur distribution in oxidized Ni40Al and its effect 280 on oxide growth and adherence. Mater. Sci. Eng. A. 454-455:80-88. doi:10.1016/j.msea.2006.11.015 281 [4] Hou PY, Izumi T, Gleeson B (2009) Sulfur Segregation at Al2O3/γ-Νi + γ′-Ni3Al Interfaces: Effects of 282 Pt, Cr and Hf Additions. Oxid. Met. 72:109-124. doi: 10.1007/s11085-009-9149-y 283 [5] Fedorova E, Monceau D, Oquab D (2010) Quantification of growth kinetics and adherence of oxide 284 scales formed on Ni-based superalloys at high temperature. Corros. Sci. 52:3932-3942. 285 doi:10.1016/j.corsci.2010.08.013 286 [6] Smialek JL (2022) Invited Review Paper in Commemoration of Over 50 Years of Oxidation of Metals: 287 Alumina Scale Adhesion Mechanisms: A Retrospective Assessment. Oxid. Met. 97:1-50. doi: 288 10.1007/s11085-021-10091-2 289 [7] Tabata C, Kawagishi K, Uzuhashi J, Ohkubo T, Hono K, Yokokawa T, Harada H, Suzuki S (2021) 290 Quantitative analysis of sulfur segregation at the oxide/substrate interface in Ni-base single crystal 291 superalloy. Scr. Mater. 194:113616. doi:10.1016/j.scriptamat.2020.11.003 292 [8] Tabata C, Kawagishi K, Uzuhashi J, Ohkubo T, Yokokawa T, Harada H, Suzuki S (2022) Unveiling the 293 Mechanism of Improved Oxidation Resistance for CaO Crucible Melting Using Ni-Al Alloy. Metall. Mater. 294 Trans. A. 53A:2452-2458. doi: 10.1007/s11661-022-06678-2 295 [9] Utada S, Joh Y, Osawa M, Yokokawa T, Sugiyama T, Kobayashi T, Kawagishi K, Suzuki S, Harada H 296 (2018) Creep Property and Phase Stability of Sulfur-Doped Ni-Base Single-Crystal Superalloys and 297 Effectiveness of CaO Desulfurization. Metall. Mater. Trans. A. 49A:4029-4041. doi: 10.1007/s11661-018-298 4710-4 299 [10] Jarvis EA, Carter EA (2002) The role of reactive elements in thermal barrier coatings. Comp. Sci. Eng. 300 4(2):33-41. doi: 10.1109/5992.988645 301 [11] Carling KM, Carter EA (2007) Effects of segregating elements on the adhesive strength and structure 302 of the α-Al2O3/β-NiAl interface. Acta Mater. 55(8):2791-2803. doi: 10.1016/j.actamat.2006.12.020 303 [12] Zhang Z, Zhang RF, Legut D, Li DQ, Zhang SH, Fu ZH, Guo HB (2016) Pinning effect of reactive 304 elements on adhesion energy and adhesive strength of incoherent Al2O3/NiAl interface. Phys. Chem. Chem. 305 Phys. 18:22864. doi:10.1039/c6cp03609k 306 [13] Bennett IJ, Kranenburg JM, Sloof WG (2005) Modeling the Influence of Reactive Elements on the 307 Work of Adhesion between Oxides and Metal Alloys. J. Am. Ceram. Soc. 88(8):2209-2216. 308 doi:10.1111/j.1551-2916.2005.00408.x 309 [14] Hou PY, Stringer J (1992) Oxide scale adhesion and impurity segregation at the scale/metal interface. 310 Oxid. Met. 38:323-345. doi: 10.1007/BF00665658 311 [15] Hou PY, Tolpygo VK (2007) Examination of the platinum effect on the oxidation behavior of nickel-312 aluminide coatings. Surf. Coat. Tech. 202(4-7):623-627. doi:10.1016/j.surfcoat.2007.06.013 313 [16] Hou PY, Priimak K (2005) Interfacial Segregation, Pore Formation, and Scale Adhesion on NiAl 314 Alloys. Oxid. Met. 63:113-130. doi: doi.org/10.1007/s11085-005-1954-3 315