Synchrotron-based high-resolution photoemission spectroscopy study of ZIRLO cladding with H2O adsorption: Coverage and temperature dependence

The coverage and temperature dependence of ZIRLO cladding with H2O adsorption are studied using synchrotron-based high-resolution photoemission spectroscopy (HRPES). Based on the analytical results of the Zr 3d, O 1 s, C 1 s, and Sn 3d HRPES profiles prior to H2O adsorption, we determine the surface compositions of O2−, hydroxyl OH−, chemisorbed H2O, zirconium carbide, adventitious carbon, Sn metal, and SnO2 in ZIRLO. When ZIRLO is exposed to H2O molecules, the relative proportion of zirconium metal decreases, whereas that of the total zirconium oxides increases, suggesting the reaction between H2O and the zirconium metal in ZIRLO. On annealing a sample with 1000 L H2O on ZIRLO at 300 °C, Zr2O3 and ZrO2 decompose, and oxygen diffuses into the bulk, thereby reducing the oxidation states of zirconium on the surface. Moreover, at this temperature, the excess H2O molecules on ZIRLO are thoroughly desorbed and tin element is diffused into the bulk in ZIRLO.

a highly surface-sensitive signal from ZIRLO based on the universal curve 19 . We determined that even before the deposition of H 2 O, the Zr 3d HRPES profile obtained from ZIRLO exhibited several peaks related to the Zr 0 , Zr + , Zr 2+ , Zr 3+ , and Zr 4+ oxidation states, indicating that the surface of ZIRLO comprises zirconium metal and zirconium oxides. After the exposure of ZIRLO to H 2 O, the relative quantity of zirconium metal decreased, whereas the total zirconium oxide quantity including Zr + , Zr 2+ , Zr 3+ , and Zr 4+ relatively increased. Sequentially, we annealed a sample with H 2 O adsorbed on ZIRLO to check the temperature dependence. On annealing up to 100 °C, there were no significant changes, whereas after annealing beyond 300 °C, the relative proportion of zirconium metal increased, and the relative quantities of the total zirconium oxides reduced with the decrease in Zr 3+ and Zr 4+ because of the decomposition of Zr 2 O 3 and ZrO 2 accompanied by oxygen diffusion into the bulk, in agreement with literature 10,20 . To the best of our knowledge, the coverage and temperature dependence of ZIRLO cladding with H 2 O adsorption have not been systematically studied using synchrotron-based HRPES.

Materials and methods
To prepare a sample for the synchrotron-based HRPES experiments, ZIRLO (Westinghouse Electric Co.) with an area of approximately 10 × 10 mm 2 was polished with 3000 grit SiC paper, and rinsed with deionized water. This ZIRLO sample contained 1.27 wt% niobium, 1.13 wt% tin, and 0.11 wt% iron with very small quantities of oxygen, carbon, nitrogen, and hydrogen as reported elsewhere 3 . Deionized water (H 2 O) was prepared using a Milli-Q water purification system (Millipore), and was further purified through several freeze-pump-thaw cycles to remove all the dissolved gases prior to deposition. The amount of H 2 O exposure is expressed in Langmuir (L) which is calculated as the product of the H 2 O molecule pressure and the exposure time (1 L = 10 −6 Torr•s).
HRPES data were obtained using the 8A2 beamline at the Pohang accelerator laboratory (PAL). In an ultra-high vacuum (UHV) chamber in which the HRPES system was installed, ZIRLO was further cleaned through five cycles of sputtering with 0.5 keV Ar + ions for 1 h at 380 °C, followed by annealing at 600 °C for 30 min. Although we attempted to completely eliminate the C 1 s peak in the ZIRLO cladding, it continued to remain after five cycles of sputtering and annealing, similar to that reported in a previous study on ZIRCALOY-4 cladding 10,21 . Therefore, this cycle of sputtering and annealing was stopped when the intensity of the C 1 s peak no longer varied. All the core-level spectra of the sample with H 2 O adsorbed on ZIRLO were recorded at room temperature using a high-performance electron analyzer (SCIENTA2002, Sienta-Omicron) at photon energies of 400, 630, and 710 eV for Zr 3d, and 710 eV for O 1 s, C 1 s, and Sn 3d, where the total spectral resolution at each photon energy was 0.15 eV, determined by measuring Au Fermi-edge. HRPES data were obtained with a pass energy of 20 eV for Zr 3d and 50 eV for O 1 s, C 1 s, and Sn 3d at an energy step of 0.05 eV. The binding energies of the core-level spectra were relatively calibrated with respect to that of the 4f 7/2 HRPES spectrum (84.0 eV) of clean Au for the same photon energy. The base pressure of the UHV chamber was maintained below 5.0 × 10 −10 Torr. All the spectra were measured in the normal-emission mode, and analyzed using a standard nonlinear least squares fitting procedure with Voigt functions 22 . Previous study reveals that tin element alone was detected by X-ray photoelectron spectroscopy (XPS), among the minor alloying elements present in ZIRCALOY-4 cladding 10 . When we performed preliminary experiments using commercial XPS (VG Scientific ESCALAB 220i-XL), the Fe 2p and Nb 3d peaks were not observed in ZIRLO; moreover, the Nb 3d signal was not detected in our HRPES spectrum. Therefore, to enhance the surface sensitivity without considering the detection of the Fe 2p signal, we lowered the photon energy as much as possible, resulting in a photon energy of 710 eV. Furthermore, as the primary analysis in this study involves the Zr 3d spectra, we reduced the photon energy up to 400 eV to obtain the most sensitive Zr 3d HRPES profiles. Fig. 1a displays the Zr 3d spectra acquired at photon energies of 400, 630, and 710 eV, respectively. The Zr 3d profiles obtained at photon energies of 630 and 710 eV were similar; however, that obtained at 400 eV was different. The electron inelastic-mean-free-paths (IMFPs) at photon energies of 400, 630, and 710 eV for Zr 3d, and 710 eV for C 1 s and Sn 3d were calculated to be 7.1, 11.4, and 12.8 Å, and 14.2 and 8.3 Å, respectively, using the NIST electron inelastic-mean-free-path database (Version 1.2) with an algorithm developed by Tanuma, Powell, and Penn 23,24 . In addition, the IMFP at photon energy of 710 eV for O 1 s was approximately estimated as 7 Å using the universal curve 19 . In particular, the IMFPs for Zr 3d indicates that the spectrum at a photon energy of 400 eV contains more surface information compared to the others. Therefore, only the Zr 3d spectrum measured at a photon energy of 400 eV will be considered henceforth. The detailed analysis of the Zr 3d spectra with the peak fitting results is presented in Fig. 2.  [25][26][27] . The C 1 s HRPES profile is depicted in Fig. 1c, where two distinct peaks can be observed at 281.5 eV and 284.6 eV. We assigned the peak at 281.5 eV to zirconium carbide produced by the reaction between zirconium metal and the adsorbed hydrocarbon; this value is similar to the reported binding energy of 281.6-282.0 eV in previous investigations 10, 28 . In addition, the peak at 284.6 eV was assigned to adventitious carbon composed of various hydrocarbon species based on the previously reported value (284.6 eV) 29 . As shown in Fig. 1c, the full width at half maximum (FWHM) of adventitious carbon at 284.6 eV is broader compared to that of zirconium carbide at 281.5 eV, which may be due to the existence of various types of hydrocarbons. In the Sn 3d HRPES profile (Fig. 1d), two types of Sn 3d 5/2 peaks at 484.1 and 486.0 eV, and 3d 3/2 signals at 492.5 and 494.4 eV, respectively, appeared with binding energy separation of 8.4 eV between them. In general, the binding energy of Sn 3d 5/2 is used to analyze the Sn 3d HRPES profile. It is known that the binding energies of Sn metal and the SnO 2 features in ZIRCALOY-4 cladding occur between 483.9-484.7 eV and 486.0-487.2 eV, respectively 10,18 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Therefore, we assigned the two Sn 3d 5/2 peaks at 484.1 and 486.0 eV to Sn metal and the SnO 2 compositions, respectively. Thus, by analyzing the HRPES spectra (Fig. 1)  The fact that such features appeared despite our cleaning process suggests that they may be due to the intrinsic oxygen and carbon in ZIRLO, which was also observed in prior XPS studies on ZIRCALOY-4 cladding 10,21 . Fig. 2a depicts the Zr 3d HRPES profile, shown in Fig. 1a, with the peak fitting results. Because the binding energy of Zr 3d 5/2 is typically utilized to analyze the Zr 3d HRPES profile, we explain the zirconium species using the binding energy of Zr 3d 5/2 . In Fig. 2a 4 ) because its reported binding energy (183.6 eV) was similar to our value 33,34 . Previously, through the analysis of the C 1 s HRPES profile in Fig. 1c, we had established the existence of zirconium carbide in ZIRLO. According to literature, the binding energy of zirconium carbide in the Zr 3d spectrum is 179.2 eV 10,28 , which is almost the same as that of zirconium metal (179.3 eV). Although the zirconium carbide component should be considered when the Zr 3d HRPES profile is analyzed, we could not confirm whether this peak was due to zirconium metal or zirconium carbide, in agreement with the previous report 10 . Hence, we had to unavoidably ascribe the peak at 179.3 eV to zirconium metal.

Results and discussion
It has been previously revealed that the order of the surface components in intact ZIRCALOY-4 cladding are as follows: The uppermost hydrocarbon, zirconium hydroxide, zirconium dioxide, zirconium suboxides, and zirconium bulk layers 10 . In addition, zirconium carbide is expected to exist in the hydrocarbon layer. As a result, based on the analysis of the Zr 3d, O 1 s, C 1 s, and Sn 3d HRPES spectra, along with the prior report, we concluded that the surface species in ZIRLO cladding could be the same as previously reported. Additionally, we propose that Sn metal and SnO 2 compositions could exist within the zirconium bulk layer, and that the chemisorbed H 2 O feature could be present in the uppermost layer.
To confirm the change in the relative proportions of the zirconium features depending on the H 2 O coverage and annealing temperature, we calculated the ratio of each peak area obtained from the peak fitting results because the ratio of each peak integral in the Zr 3d HRPES spectra corresponds to their relative proportion ( Table 1). As the proportion of zirconium hydroxide is negligible, we consider the other populations, herein. As shown in Fig. 2 and Table 1, the relative proportion of zirconium metal is the largest, and gradually decreases in the following order Zr + , Zr 2+ , Zr 3+ , and Zr 4+ , in accordance with their relative trends reported in literature 1 . After the exposure of ZIRLO to 100 L and 1000 L H 2 O, the relative proportion of zirconium metal decreased whereas those of the zirconium oxides including the Zr + , Zr 2+ , Zr 3+ , and Zr 4+ features mostly increased ( Fig. 2 and Table 1). This indicates that when H 2 O is adsorbed on ZIRLO, H 2 O and zirconium metal in ZIRLO react each other, relatively decreasing and increasing the metal population and the total quantity of zirconium oxides, respectively, in agreement with previous research on H 2 O adsorbed on pure zirconium and ZIRCALOY-2 samples 1 . This phenomenon can be explained by the dissociation of H 2 O into the adsorbed oxygen and molecular hydrogen gas on ZIRLO at room temperature as reported in the prior study of water molecule on Zr(0001) 35 . We performed annealing experiments on the 1000 L H 2 O system adsorbed on ZIRLO. As shown in Fig. 2d, the Zr 3d HRPES profile after annealing at 100 °C for 30 min is similar to that before annealing, indicating that any detectable change did not occur due to annealing at this temperature. However, when temperatures of 300 °C and 500 °C were applied for 30 min, the relative percentage of the Zr 0 , Zr + , and Zr 2+ valence states increased, whereas those of the Zr 3+ and Zr 4+ oxidation states decreased (Fig. 2e,f). According to literature, on annealing at 200 °C or more, the oxidation states of zirconium are converted to lower oxidation states because of the decomposition of the Zr 2 O 3 and ZrO 2 compositions, and the depopulation of oxygen in the surface region accompanied by oxygen diffusion into the bulk 10,20 . Therefore, we concluded that the decomposition of Zr 2 O 3 and ZrO 2 and the diffusion of oxygen into the bulk lead to the reduction of the oxidation states of zirconium at 300 °C. Fig. 3 displays the coverage and temperature dependence of the O 1 s and Sn 3d HRPES profiles. As shown in Fig. 3a www.nature.com/scientificreports www.nature.com/scientificreports/ increases. This increase is attributed to the excess H 2 O molecules on ZIRLO, which could be the remaining quantity after sufficient reaction with zirconium metal. After annealing at 100 °C for 30 min, the profile of the O 1 s HRPES spectrum remined unchanged, in accordance with the analytical result of the Zr 3d HRPES spectrum at this temperature. When the sample was annealed at 300 °C and 500 °C for 30 min, the enhanced peaks returned to the state before the adsorption of H 2 O on ZIRLO, indicating that the excess H 2 O molecules were completely desorbed at 300 °C. In the Sn 3d HRPES profiles (Fig. 3b), the peaks related to the Sn metal and SnO 2 compositions gradually disappear due to the deposition of 100 L and 1000 L H 2 O on ZIRLO. As the adsorption of H 2 O on ZIRLO increases the surface thickness, these signals may be reduced because the probing depth for ZIRLO itself relatively becomes shallow. After annealing at 100 °C for 30 min, only the peak at 486.0 eV related to the SnO 2 configuration remained indistinctly in the Sn 3d HRPES spectrum. Furthermore, after annealing at 300 °C for 30 min, it completely vanished instead of recovery, despite the desorption of H 2 O molecules at this temperature. Therefore, we infer that SnO 2 composition decomposed and that Sn metal migrated into zirconium bulk at 300 °C, sequentially.

conclusions
In this study, we investigated the coverage and annealing-temperature dependence of the ZIRLO cladding with H 2 O adsorption using synchrotron-based HRPES. Through the analysis of the Zr 3d, O 1 s, C 1 s, and Sn 3d HRPES profiles obtained from ZIRLO before H 2 O exposure, we confirmed the existence of O 2− , hydroxyl OH − , chemisorbed H 2 O, zirconium carbide, adventitious carbon, Sn metal, and SnO 2 species in ZIRLO. After the deposition of H 2 O on ZIRLO, the relative proportion of zirconium metal decreased, whereas that of the total zirconium oxides increased, indicating that H 2 O reacted with zirconium metal in ZIRLO. On annealing a sample with 1000 L H 2 O adsorbed on ZIRLO at 300 °C, the decomposition of Zr 2 O 3 and ZrO 2 as well as the diffusion of oxygen into the bulk occurred. Furthermore, we determined that the excess H 2 O molecules were completely desorbed and that on SnO 2 decomposition, the tin element diffused into the zirconium bulk in ZIRLO at that temperature.