ToF-SIMS study of the behavior of thermally oxidized films formed on nickel-based 690 alloy in high-temperature water

Abstract

Alloy 690 is a high-chromium nickel alloy widely used for steam generator tubes in pressurized water reactors. We have applied an optimized thermal oxidation treatment to alloy 690 consisting of 4 h at 500 °C (0.6 at% O₂ in Ar) to make an improved diffusion barrier layer limiting the release of Ni during processing. The efficiency of such a treatment was tested by immersion in simulated pressurized water reactor primary water for 9 weeks. This work also explores the use of time-of-flight secondary ion mass spectrometry (ToF-SIMS) to characterize the oxidized films formed after different treatments. Reference spectra were acquired by analyzing thermally oxidized Ni, Cr, and Ni-30Cr samples as well as hydroxides and spinel powders [Ni(OH)₂ and NiFe₂O₄]. ToF-SIMS spectra of model oxides have allowed the selection of a set of ionic intensity ratios to be used for data processing of the spectra and profiles, obtained after the different oxidation steps applied on alloy 690. Thus, the following information was obtained: (i) the oxide layer structure is triplex; (ii) an intermediate layer is composed of Cr₂O₃, on which surface coverage depends on treatment; (iii) the outer layer is Ni- and Fe-rich, whereas the inner layer is most probably constituted by a NiCr spinel oxide; and (iv) the total oxide layer grown during thermal oxidation is thinner than the one directly formed in primary water. If similar in thickness, the thermally grown oxide layer has a different composition once exposed to primary water. This ToF-SIMS strategy was very fruitful in such a surface tube characterization. Copyright © 2012 John Wiley & Sons, Ltd.

Introduction

Nickel-based alloy 690 is now widely used as a structural material for steam generator (SG) tubes in pressurized water reactors (PWRs) of nuclear plants. During processing, oxide scales develop on SG tubes: it is composed of a Cr-rich inner layer and a Ni-rich external layer.[1-7] The release of Ni in primary water (named PW) leads to the increase of radioactivity in the reactor. For safety reasons, it is important to limit this activity. It is known that the Cr-rich layer acts as a diffusion barrier and limits the release of nickel.[8] To improve the properties of such diffusion barrier layers, we have applied an optimized thermal oxidation treatment[9] to commercial alloy 690 samples, 4 h at 500 °C under a flux of gaseous mixture (0.6 at% O₂ in Ar), before immersion for 9 weeks in simulated high-temperature and high-pressure water.

The objectives of this work were to chemically characterize the composition of the thermally treated (TT) oxide compared with the passive film formed in PW, by collecting the molecular chemical signatures of model oxides (including pure oxides, hydroxides, and mixed oxides) using time-of-flight secondary ion mass spectrometry (ToF-SIMS). In a second step, the behavior of the TT oxide was tested after 9 weeks in conditions simulating PWRs, from the chemical characterization of the composition of the oxide film (TT + PW) to a careful examination of the total oxide thickness determined from the calibrated ToF-SIMS depth profiles.

Experimental

Surface analysis

Time-of-flight secondary ion mass spectrometry data were acquired using a TOF.SIMS V spectrometer (ION-TOF GmbH, Muenster, Germany). The analysis chamber was maintained at less than 5 × 10⁻⁷ Pa under operational conditions. The total primary ion flux was less than 10¹² ions × cm⁻² ensuring static conditions. A pulsed 25 keV Bi⁺ primary ion source at a current of 1.3 pA (high current bunched mode), rastered over a scan area of 100 µm × 100 µm was used as the analysis beam. ToF-SIMS depth profiles were measured with the instrument working in the dual-beam mode. The sputtering was performed using a 0.5 keV (30 nA) or 2 keV (80 nA) Cs⁺ ion beam, rastered over an area of 300 µm × 300 µm. Both ion beams were impinging the sample surface forming a 45° angle with the surface normal and were aligned in such a way that the analyzed ions were taken from the center of the sputtered crater. The oxide sputtering rate was established by combining a series of external measurements of crater depths, obtained with a mechanical profilometer (Dektak 150, Veeco, Veeco-Instrument Europe, Dourdan, France), after depth profiling for a certain time, in the same energy and current conditions. Assuming a constant rate, whatever the oxide composition (as the oxide density is always very similar), the resulting value is 0.032 ± 0.003 nm s⁻¹.

Data acquisition and processing analyses were performed using the commercial IonSpec program. The exact mass values of at least five known species, from H⁻, C⁻, C₂⁻, C₃⁻, and Cl⁻, were used for calibration of the data acquired in the negative ion mode, in which the best information was obtained with regard to our oxidized systems. The mass resolution, M/ΔM, was >10 000 for the low mass range (<100).

Materials and corrosion

Materials

Tubes (45 mm long, 22 mm in diameter, thickness of 1.3 mm) of commercial alloy 690 [Ni − 30Cr − 11Fe (wt. %)] were used, their main impurities being lower than 0.25 wt.% (Si, Ti, and Al). Half tubes were cut into coupons of 12 × 50 mm² for analysis of their inner surface.

Ni and Cr were provided by Goodfellow (Goodfellow SARL, Lille, France) (purity of 99.99%). Ni-30Cr alloy was provided by Imphy (Ugine-SavoieImphy, France) (purity of 99.50%). These samples were mechanically polished to a 1-µm diamond finish. Ni, Cr, and Ni-30Cr were thermally oxidized in air at 900 °C for 4 h. Powders of Ni(OH)₂ (purity of 99.999%) and NiFe₂O₄ (purity of 98%) were provided by Alfa-Aesar (Alfa Aesar France, Schiltigheim, France).

PWRs simulating conditions: corrosion loop

Corrosion tests were conducted in a recirculating autoclave (10 l/h) at 155 bar and 325 °C for 9 weeks. The aqueous solution contained 2 ppm of lithium and 1000 ppm of boron. A hydrogen overpressure of 0.3 bar was maintained to ensure a dissolved H₂ concentration of 25 to 30 cm³·kg⁻¹ and a low oxygen content of less than 5 ppb.

Results and discussion

Model oxide characterization

The surface of the TT Cr consists in a 150-nm-thick continuous Cr₂O₃ film (as calculated from the sputtering speed), covered by an outer Cr(OH)₃ layer.[10] From the ToF-SIMS depth profiles, it was possible to calculate a CrO₂⁻/CrO⁻ ratio of 1.2 in the Cr₂O₃ film. The external Cr(OH)₃ thin layer was characterized by both an intense OH⁻ signal and a higher CrO₂⁻/CrO⁻ ratio (≈1.7). In addition to the CrO⁻ or CrO₂⁻ ions, CrO₃⁻ and CrO₄⁻ ions presented the same intensity variations until the interface with bulk Cr.

The analyses performed on thermally oxidized Ni presented very similar results: the NiO continuous 130-nm-thick thin film (as determined from the sputtering speed) was nicely profiled by NiO⁻, NiO₂⁻, or NiO₃⁻ ions. In the oxide film, the NiO₂⁻/NiO⁻ ratio equals unity. There was no NiO₄⁻ signal.

The thermal treatment was applied on Ni-30Cr to form NiCr₂O_4.[11, 12] According to the ToF-SIMS depth profile (Fig. 1), the oxidized surface of TT Ni-30Cr can be described by up to three oxide layers (total oxide thickness >500 nm). The outermost layer is composed of a mixture of NiO and Cr₂O₃. An intermediate layer can be described by the NiCrO⁻ signal and by a NiO₂⁻/NiO⁻ ratio that is equivalent to 4.0 × 10⁻¹, significantly different from the value in pure NiO (~1.0). The third Cr₂O₃-rich layer is characterized by a CrO₂⁻/CrO⁻ ratio of 1.9, also different from the value in pure Cr₂O₃. On the basis of other published data,[11, 12] we can consider that we have prepared a layer of pure NiCr₂O₄ at the interface with the Cr-rich oxide layer. In this chromite layer, the NiCrO⁻/CrO⁻ ratio is equivalent to 8.0 × 10⁻³, NiCrO₂⁻/CrO₂⁻ is constant and is equivalent to 6.0 × 10⁻³. The CrO₂⁻ ion has its maximum intensity in this layer.

Figure 1. ToF-SIMS depth profile (negative mode) of Ni-30Cr thermally treated in air for 4 h at 900 °C (Cs⁺ 2 keV, current 80 nA).

Additional reference data on mixed oxides have been obtained from the analysis of the NiFe₂O₄ powder: the NiFeO_x⁻ and NiFe₂O_x⁻ signals were observed in the ToF-SIMS spectrum (negative mode), confirming that these mixed ions are markers of mixed oxides (or spinels), in agreement with Aubriet et al.[13] The ionic NiFeO⁻/NiO⁻ ratio is equivalent to 0.053 in the powder (Table 1). The analysis of bulk Ni(OH)₂ powder gave us a new series of ions to be distinguished both from the ionic response of the NiO thin film (which may be hydroxylated in its outermost layer) and from the external NiO-rich layer of the TT Ni-30Cr, which may also be hydroxylated. Characteristic ions are NiOH⁻, NiOH₂⁻, NiO₂H⁻ (the more intense signal), and NiO₂H₂⁻, the intensity ratio of NiO₂H⁻/NiO⁻ allows us to discriminate between hydroxylated NiO, Ni(OH)₂, and hydroxylated NiCr₂O₄ (Table 1).

Table 1. ToF-SIMS ionic ratios obtained on model oxides

Species	Material	Analysis conditions	Mean ratio (standard deviation)
TT, thermal treatment (see Experimental).
Cr2O3	TT pure Cr	Depth profiling (0.5 keV, 30 nA)	CrO2^−/CrO^− = 1.0 (2 × 10^−2)
		Depth profiling (2 keV, 80 nA)	CrO2^−/CrO^− = 1.2 (7 × 10^−3)
NiO	TT pure Ni	Depth profiling (0.5 keV, 30 nA)	NiO2^−/NiO^− = 6.4 × 10^−1 (2 × 10^−2)
		Depth profiling (2 keV, 80 nA)	NiO2^−/NiO^− = 1.2 (3 × 10^−2)
			NiO2H^−/NiO^− = 0.13
NiCr2O4	TT Ni-30Cr	Depth profiling (2 keV, 80 nA)	CrO2^−/CrO^− = 2.6 (2 × 10^−1)
			NiO2^−/NiO^− = 4.0 × 10^−1 (1 × 10^−1)
			NiCrO^−/NiO = 1.2 × 10^−2 (4 × 10^−3)
			NiCrO^−/CrO^− = 8.0 × 10^−3 (1 × 10^−1)
Ni(OH)2	Powder	Surface spectra (no sputtering)	NiO2^−/NiO^− = 6.8 × 10^−1 (standard deviation not defined)
			NiO2H^−/NiO^− = 3.2 (standard deviation not defined)
NiFe2O4	Powder	Surface spectra (no sputtering)	NiO2^−/NiO^− = 2.2 (standard deviation not defined)
			NiFeO^−/NiO^− = 1.1 × 10^−1 (standard deviation not defined)
			NiFeO^−/FeO^− = 6.0 × 10^−2 (standard deviation not defined)

Application to industrial SG tubes

Figure 2 presents the ToF-SIMS depth profile of a TT SG tube, sputtered at 0.5 keV (30 nA) as a function of the sample thickness. The profile presents the ions of interest as determined from the previous reference analysis: CrO⁻, CrO₂⁻, NiO⁻, FeO⁻, as well as a set of the more intense mixed ions, NiCrO⁻, NiFeO⁻, and CrFeO⁻. It is possible to distinguish three different parts in the oxide layer. The outer part would be composed of a mixed Ni and Fe-rich oxide layer, the intermediate layer would be chromium oxide–rich, the inner layer, mainly characterized by an intense signal of NiCrO^-, would correspond to a spinel-rich NiCr₂O₄ portion. It would be located at the interface with the alloy.

Figure 2. ToF-SIMS depth profile (negative mode) of an SG tube (alloy 690) thermally treated in air for 4 h at 500 °C (Cs⁺ 0.5 keV, current 30 nA).

We have brought special attention to the intermediate Cr₂O₃-rich layer chemical composition, which would be the main protection against Ni diffusion in the case of the two other treatments: the SG tube directly immersed in PW and the TT SG tube immersed in PW (TT + PW) for 9 weeks in both cases. The depth profiles obtained with the PW and TT + PW samples (data not shown) present a similar three-layered structure. Figure 3 presents the CrO₂⁻/CrO⁻ and NiCrO⁻/NiO⁻ ratios as a function of thickness for TT, PW, and TT + PW samples. For TT, the CrO₂⁻/CrO⁻ ratio is approximately 0.9, significantly lower than in the case of pure Cr₂O₃ (Table 1). This difference is tentatively interpreted by the loss of the continuous character of the Cr₂O₃ layer in favor of an island-like distribution, which was recently described after immersion in PW.[3, 12] This hypothesis still has to be confirmed, taking into account the effect of the energy of the sputtering ions on the ionic ratios.

Figure 3. CrO₂⁻/CrO⁻ and NiCrO⁻/NiO⁻ ratios for the three oxidation treatments applied to SG tubes (alloy 690).

Another hypothesis is that it corresponds to a mixed chromium oxide compound, but not to pure NiCr₂O₄, as, in this case, the CrO₂⁻/CrO⁻ ratio of 1.5 is by far larger than 0.9 (Table 1). For the PW sample, the same ionic ratio is higher, in satisfactory agreement with pure Cr₂O₃. Besides these results, the TT oxide is thinner than the PW: 27 nm versus 35 nm for the total oxide thickness, respectively. At this point, we can conclude that the two processes do not lead to the formation of the same intermediate oxide layer.

For TT + PW, the CrO₂⁻/CrO⁻ ratio of the intermediate layer increases from the TT sample (0.9) to the TT + PW sample (1.2; Fig. 3) as a function of sputtering time, the total oxide thickness being in the same range as for the TT sample (25 nm). For the value of 1.2, as in the case of PW, we tend to suggest that it corresponds to a continuous Cr₂O₃ layer, formed in PW. We propose that the TT oxide layer is not stable in composition in PW conditions, but is stable in thickness, thus protecting the metallic bulk. The chemical modifications, occurring at constant total oxide thickness, tend to result in the same composition as the PW layer.

Regarding the inner layer, taking into account a NiCrO⁻/NiO⁻ ratio reaching a value of approximately 9 × 10⁻² for all treatments (Fig. 3), we propose that it is systematically composed of a chromite spinel. The NiCrO⁻/NiO⁻ ratio increases gradually in the inner layer from 5 × 10⁻² to 9 × 10⁻²: we attribute these changes either to a change in the spinel stoichiometry or to different mixtures between the spinel and the pure oxides (NiO or Cr₂O₃). The three-layer structure has also been proposed by Zhiming et al.[7] after immersion in PW but not with the same oxide order (the chromite being located in the intermediate position). The ToF-SIMS approach seems to bring a new and important analytical insight for the establishment of the chemical composition of the oxide, the properties of which are of major interest.

Conclusion

This work explored the use of ToF-SIMS to characterize oxidized films formed after different treatments on alloy 690, a high-chromium nickel alloy widely used for SG tubes in PWRs. Reference spectra of thermally oxidized Ni, Cr, and Ni-30Cr samples were obtained using ToF-SIMS analysis in the negative mode, as well as from reference spectra of hydroxides and spinel powders [Ni(OH)₂ and NiFe₂O₄]. From these spectra, the intensity ratios of the ions of interest were calculated. On the basis of this data processing, the spectra and profiles obtained after different oxidation steps applied on alloy 690 gave the following information: (i) the oxide layer structure is triplex; (ii) the outer layer is Ni- and Fe-rich, whereas the intermediate layer is Cr₂O₃-rich, the surface coverage of which depends on the treatments; (iii) the inner layer is most probably constituted by a Ni-Cr spinel oxide; and (iv) the total oxide layer grown during thermal oxidation is thinner than the one directly formed in the PW. The thermally grown oxide layer has revealed a satisfactory resistance to further corrosion: its chemical composition is different once exposed to PW but its total thickness remains constant at approximately 25 nm.

Acknowledgements

Financial support by Région Ile-de-France (SESAME program) is acknowledged.