NO + CO Reaction on Polycrystalline Palladium Foils with Specific Surface Crystallographic Orientation: A New Approach to Develop Foil Catalysts Based on Texture Control

Abstract

Catalytic properties of the metal foils were tuned by the specific surface crystallographic orientation control. The NO + CO reactivity of polycrystalline Pd foils with (101) or (211) orientation, characterized by using electron backscattered diffraction (EBSD), was investigated. The (211)-oriented Pd foils exhibited much greater activity compared with (101) orientated Pd foils, which is in good agreement with the previous structure sensitivity studies on NO + CO reactions over Pd. Hence, the present study introduces the surface crystallographic orientation control of foil catalysts as a new development strategy for the unique catalysts in the material gap.

1. Introduction

A “material gap” exists between practical catalysts (highly dispersed nanoparticles) and model catalysts (well-defined surfaces of a single crystal) in catalyst chemistry. Thus, research into bulk materials as a new type of catalyst, such as polycrystalline foils, powders, porous materials, etc. has been reported recently and has the potential to provide unique results. Metal foil catalysts, especially used with effective honeycomb or coil structure, are one of the unique catalysts in the material gap.^1–5)

The present study describes the new approach of foil catalyst development: “texture control”. Metal polycrystalline foils are known to exhibit “texture” with specific crystallographic orientation, depending on pretreatments such as rolling or annealing. Because most catalytic reactions are structure sensitive (i.e., surface crystallographic planes of the catalyst metal could have a large effect on catalytic reactivity) as shown by numerous model catalyst studies, and thus the texture control would possibly be an effective method to tune the properties of foil catalysts. In catalyst chemistry, however, the idea of texture control is unfamiliar possibly because nanoparticle catalysts have long been the focus of interest in this field. Even the latest reports of foil catalysts^1–5) do not have texture control as a main topic while texture control has been successfully applied in other areas of chemistry; for example, texture-controlled Cu substrates for chemical vapor deposition (CVD) have been reported.^6,7)

Indeed, in the early report by Miyao et al., a relation between CO hydrogenation reactivity and the texture of Pt, Ni, and Pd foil catalysts was investigated.⁸⁾ However, the method they used to investigate surface crystallographic orientation was not optimal. They determined the orientation of the foils by X-ray diffraction (XRD), which is the same method used for powder samples. However, the information was obtained mainly from the bulk rather than the surface because of the deep penetration depth of the X-rays (0.1∼10 µm; depending on diffraction angle). In addition, in principle, XRD peaks from several planes cannot be observed due to deconstructive interference. For example, for an FCC structure, the planes with mixed odd and even indices do not produce diffraction peaks. These difficulties make quantitative evaluation of orientation using XRD difficult.

Therefore, an alternative characterization technique for surface crystallographic orientation is necessary to develop texture-controlled foil catalysts. Electron backscattered diffraction (EBSD) is a good choice. The EBSD system usually is attached to scanning electron microscopy (SEM), which can detect the crystallographic orientation for each point scanned by the electron probe.⁹⁾ Advantages of EBSD include the ability: (i) to obtain crystal orientation data with locational information over a wide range and (ii) to detect every crystallographic orientation to an accuracy of ∼1°.¹⁰⁾ It should be noted that EBSD is not so sensitive to the exact surface structure as surface science techniques like low energy electron diffraction (LEED), reflection high energy electron diffraction (RHEED) or scanning tunneling microscope (STM), etc., but it limits analysis, at least, near the surface (∼50 nm). Even though Miyao et al. also employed EBSD,⁸⁾ it was not the main technique, and only a narrow area range against the grain size was scanned.

To demonstrate this novel approach, the NO + CO reaction over Pd was selected as an example. The reaction is represented by eq. (1), although it may coincide with formation of the byproduct, N₂O, as shown in eq. (2).   

\begin{equation} \text{2NO} + \text{2CO} \to \text{N$_{2}$} + \text{2CO$_{2}$} \end{equation} (1)

  

\begin{equation} \text{2NO} + \text{CO} \to \text{N$_{2}$O} + \text{CO$_{2}$} \end{equation} (2)

Pd is known to catalyze these reactions, and numerous studies have been reported,^11–17) including theoretical calculations.^18–20) Model catalyst studies also have been conducted for this reaction system,^21–32) and this catalysis is known to be structure sensitive.

In summary, this study investigated the relation between NO + CO reactivity and surface crystallographic orientation for polycrystalline Pd foils to demonstrate a new approach for foil catalyst development by texture control.

2. Experimental Procedure

2.1 Sample preparation

Pure Pd rolled foils (0.05 mm thick) were purchased from Nilaco Co. Ltd. (Tokyo, Japan). First, the foils were cut into 2 mm × 5 mm pieces. The side of 2 mm and the side of 5 mm correspond to the transverse direction (TD) and rolling direction (RD) of the foil, respectively. Next, two different types of pretreatments were conducted for the foil pieces in order to change the surface crystallographic orientation (Fig. 1). The external force on samples such as bending or cutting was carefully avoided after the initial cut to minimize the influence on texture.

Fig. 1

Pretreatment profiles for (a) (101)-oriented Pd foils and (b) (211)-oriented Pd foils. The heating rate was 10°C/min for each heating process.

Annealing at high temperature (1100°C) can flatten the sample surface, provide recrystallization grains with low defect density, and result in recrystallization texture. Accuracy of EBSD analyses requires a planer surface of the sample,¹⁰⁾ and thus a fully flattened surface is preferred. In addition, rolled foils store excess energy mainly as dislocations, which might affect catalytic performance. However, annealing at high temperature can release this energy, and let grains with low defect density grow by: (i) release of the stored energy typically with the rearrangement of dislocations (recovery), (ii) grain growth driven by stored energy (recrystallization), and (iii) growth of the mean grain size driven by grain boundary energy (grain coarsening).³³⁾ Depending on initial (deformed) states, annealed foils exhibit texture with specific surface crystallographic orientation. In the experiments, the pre-annealing state was modified by catalysis (Fig. 1(b)) although this is not commonly used method for texture control. The recrystallization texture formed at temperatures (1100°C) higher than that used for catalytic tests (500°C) is preferable for eliminating the influence on texture due to temperature alone during the tests.

Field emission scanning electron spectroscopy (FESEM) was conducted using a Hitachi SU-6600 instrument, and the surface crystallographic orientation of Pd was determined using an EBSD system (Oxford, channel 5) at 20.0 keV acceleration voltage. The dataset obtained was processed using the software installed in the EBSD system. The obtained data can be displayed in two ways. First, it can be mapped in accordance with its location (IPF mapping). Second, the distribution density of the surface crystallographic orientations determined by EBSD analysis can be summarized in inverse pole figure (IPF) using color contour and lines. The IPF for the FCC structure is displayed in Fig. 2, along with illustrations of the representative planes drawn by VESTA.³⁴⁾ The EBSD results in Fig. 3 confirmed that (101)-oriented foils (a) and (211)-oriented foils (b) were produced through simple annealing (Fig. 1(a)) and by a combination of catalysis and annealing (Fig. 1(b)), respectively. It is interesting that the relatively high distribution density was obtained for the crystallographic orientation roughly around $[\bar{1}11]$ zone axis. The crystallographic orientation with respect to RD is also displayed in Fig. A1(a) and (b), and both as-prepared samples showed [111] orientation to RD. Though there are few literatures about pure Pd foil texture, general tendency of the crystallographic orientation is similar to the recent report by Murase et al., in which the IPF mapping of polished surface of the pure cold-rolled Pd sheet showed the crystallographic planes such as (211) and (110) with respect to the normal direction.³⁵⁾

Fig. 2

Inverse pole figure for the FCC structure with cross-sectional illustrations by VESTA³⁴⁾ for representative crystal planes.

Fig. 3

SEM images, IPF mappings, and distributions of crystallographic planes with respect to surface normal direction for as-prepared (a) (101)-oriented foil and (b) (211)-oriented foil. The probe size was set to 5 µm and 92–96% of the scanned area was indexed. For IPF mappings, noise reduction was conducted by software, and grain boundaries (>10°) are displayed as black lines. (RD and TD represent the rolling direction and transverse direction, respectively.)

2.2 Catalytic tests

Catalytic tests were conducted using a fixed-bed flow reactor. Twenty foil pieces were loaded into a 4-mm i.d. quartz tube. Twenty pieces were the maximum amount of sample foils to load one layer of foils without any external force. The RD of the foils (the side of 5 mm) was set parallel to the gas flow in this study, which provides similar condition to the previous foil catalyst studies.^3–5) Temperature was monitored by a thermocouple and controlled by a vertical furnace. Prior to the test, temperature was increased to 500°C at a heating rate of 10°C/min under He flow (30 mL/min). Immediately after the temperature reached 500°C, the flow gas was changed to reactant gas, 0.5% NO + 0.5% CO + He (50 mL/min), and the test initiated. Product gas was analyzed using gas chromatography (Shimadzu, GC-14B) with a thermal conductivity detector (TCD). An MS-5A column was used for NO, CO, and N₂, and a PQ column was used for CO₂ and N₂O.

The CO₂ formation rate and the N₂ selectivity were determined from the results of gas chromatography. The CO₂ formation rate ($R_{\text{CO}_{2}}$) was calculated from the CO₂ concentration ($C_{\text{CO}_{2}}$) and the flow rate of reactant gas (v ∼ 50 mL/min = 0.83 mL/s) using eq. (3).   

\begin{equation} R_{\text{CO${_{2}}$}}\ [\text{mol${\cdot}$s$^{-1}$}] = C_{\text{CO${_{2}}$}}\ [\%]\times \frac{v\ [\text{mL${\cdot}$s$^{-1}$}]}{22.4 \times 10^{3}\ [\text{mL${\cdot}$mol$^{-1}$}]} \end{equation} (3)

The areal reaction rate for CO₂ formation ($r_{\text{CO}_{2}}$) was calculated from the BET specific surface area (A_BET), the weight of the sample foils (W), and the CO₂ formation rate ($R_{\text{CO}_{2}}$) using eq. (4).   

\begin{equation} r_{\text{CO${_{2}}$}}\ [\text{mol${\cdot}$s$^{-1}{\cdot}$m$^{-2}$}] = \frac{R_{\text{CO${_{2}}$}}\ [\text{mol${\cdot}$s$^{-1}$}]}{A_{\text{BET}}\ [\text{m$^{2}{\cdot}$g$^{-1}$}] \times W\ [\text{g}]} \end{equation} (4)

Selectivity to N₂ ($S_{\text{N}_{2}}$) was calculated from the N₂ and N₂O concentrations in the product gas, $C_{\text{N}_{2}}$ and $C_{\text{N}_{2}\text{O}}$, using eq. (5).   

\begin{equation} S_{\text{N${_{2}}$}}\ [\%] = \frac{C_{\text{N${_2}$}}\ [\%]}{C_{\text{N${_{2}}$}}\ [\%] + C_{\text{N${_{2}}$O}}\ [\%]} \times 100 \end{equation} (5)

2.3 Characterization

Before and after the catalytic test, the specific surface area of the samples was calculated using the BET method from the results of Kr physisorption experiments at 77 K by BELSORP-max. Pretreatment at 150°C for 1 h under vacuum was conducted prior to the measurement.

The surface crystallographic orientation of the samples was evaluated using SEM-EBSD. In addition to evaluating the prepared foils (Fig. 3), the same EBSD method was applied to analyze the surface of the foils after the catalytic test.

Hard X-ray photoelectron emission spectroscopy (HAXPES) was performed to study the chemical state of Pd. The excitation source (5.95 keV) was generated by SPring-8 (BL 15XU). Base pressure was kept under 2 × 10⁻⁷ Pa during the measurements.

X-ray diffraction was done to verify the phase and estimate surface orientation. The 2θ range between 20 and 130° was scanned using a Rigaku Ultima IV-TS instrument equipped with Cu Kα radiation. Peaks from Cu Kα 2 were eliminated by the software.

3. Results and Discussions

3.1 Orientation change during the sample preparation procedure

The texture change of the Pd foils during the pretreatments was investigated, and the results are summarized in the appendix (A1). The difference in surface crystallographic orientation depending on the pretreatments seems to be related to the texture modification by combination of hydrogen absorption and catalysis. First, hydrogen pretreatment (250°C, 1 h) induced (100)-oriented hydride phase texture as indicated by XRD analysis in Fig. A2(c) and EBSD analysis in Fig. A3(b). Next, catalysis pretreatment (NO + CO, 500°C, 6 h) decomposed the hydride (see Fig. A2(d)) with keeping the (100) orientation as shown in EBSD results in Fig. A3(c). Interestingly, both hydrogen pretreatment and catalysis pretreatment were necessary to obtain (100)-oriented metal Pd foils; for example, applying He pretreatment (500°C, 1 h) instead of catalysis pretreatment did not produce (100)-oriented foils. Though the detailed mechanism cannot be clarified yet, such modified texture might affect the formation the (211)-oriented texture, which is different from the (101)-oriented one after just simply annealed.

3.2 Catalytic reactivity

The time dependence of CO₂ formation and N₂ selectivity is displayed in Fig. 4. During the test, (211)-oriented foils produced a greater rate of CO₂ formation compared with (101)-oriented foils. Values for CO₂ formation rate gradually increased for both samples, and were nearly stable in the late stage of the test (about t > 90 min), implying the catalytic reaction reached a steady state. To quantitatively evaluate the activity of the samples, “initial” and “final” areal reaction rates of CO₂ formation were calculated from the first and last measured CO₂ formation rate, respectively (Table 1). The values of mass of sample foils are also shown. The values of BET specific surface area before and after the catalytic test were applied for “initial” and “final” values, respectively. It should be noted that the obtained values of total BET surface area are quite (ca. 12∼25 times) larger than the estimated value for the foils with completely flat surface. The discrepancy is possibly because of the surface roughness of the sample foils. Although the foils were sufficiently flat to obtain EBSD maps, the trace of rolling can be still observed in SEM images even after annealing at 1100°C (Fig. 3), and such roughness would affect the BET surface area.

Fig. 4

Time dependence of (a) CO₂ formation rate and (b) N₂ selectivity at 500°C.

Table 1 Results of catalytic tests at 500°C.

The “initial” measurements indicated that the areal reaction rates of the (211)-oriented foils were about 10-fold greater than those of the (101)-oriented foils, indicating that (211) orientation provides much greater activity than (101) orientation. In contrast, selectivity toward N₂ was higher for the (101)-oriented foils than for the (211)-oriented foils.

The results of “final” measurements for (101)-oriented foils showed that areal CO₂ formation rates slightly increased compared with the “initial” value; however, N₂ selectivity decreased. The “final” N₂ selectivity of (101) foils was similar to that of the “initial” N₂ selectivity values of (211) foils.

In summary, the order of activity was: (211) foil (initial) ≫ (101) foil (final) > (101) foil (initial); and for N₂ selectivity was: (101) foil (initial) > (211) foil (initial) ∼ (101) foil (final).

3.3 Characterization

3.3.1 SEM-EBSD

The results of SEM-EBSD analysis for the foils after the catalytic tests are displayed in Fig. 5.

Fig. 5

SEM images, IPF mappings, and distributions of crystallographic planes with respect to surface normal direction for both (a) (101)-oriented foil and (b) (211)-oriented foil after the catalytic test. The probe size was set to 5 µm and 92–96% of the scanned area was indexed. For IPF mappings, noise reduction was achieved using software, and grain boundaries (>10°) are displayed as black lines. (RD and TD represent the rolling direction and transverse direction, respectively.)

First, orientation of the (101)-oriented foil changed to (311) after the catalytic test, yet an area corresponding to the (101) plane remained (Fig. 5(a)). Supposing that this orientation change slightly increased the areal reaction rates, as shown in Table 1, the (311) plane would be more active than (101) plane but not as active as the (211) plane. In addition, the (311) plane would have lower N₂ selectivity than the (101) plane because the N₂ selectivity of (101)-oriented foils decreased during the catalytic test.

Next, orientation of the (211)-oriented foil was strengthened after the catalytic test (Fig. 5(b)). The high activity of these foils is likely related to the (211) plane since CO₂ formation rate of the (211)-oriented foils increased during the catalytic test, as did the degree of (211) orientation after the test in comparison to that before the test.

Another significant finding was the growth of the grains of both samples throughout the catalytic test (see appendix (A2)). As shown in the grain size statistics in Fig. A4, this grain growth was much clearer for the (101)-oriented foils. Usually, the grains do not grow during heating at 500°C after annealing at 1100°C for 2 h. These results indicate that the catalytic reaction itself could change the texture (both orientation and grain size) of the foil catalyst. It was implied that the planes with higher activity tend to grow after the catalysis. The detailed mechanism of such texture modification was not elucidated, but the surface strain modification by adsorbates might be involved.

3.3.2 HAXPES and XRD

Figure 6 shows HAXPES spectra for the samples. Compared with the reference spectrum of pure Pd metal, none of the spectra contained the chemical shift of the Pd 3d peaks. The chemical state of Pd near the surface should be metallic. Wide scan and narrow scan for C1s spectrum were also conducted, and the results are summarized in appendix (A3). All the peaks in the wide scan results, except the small peak originating from C1s, (Fig. A5) can be attributed to Pd. Indeed, the peak area of the C1s band estimated from the narrow scan results (Fig. A6) was quite low (approximately 3% of that of Pd3d_5/2) for all the samples, ensuring that the contamination was avoided through the experiments. These results confirm that the foils were composed mainly of metallic Pd, with only small amounts of other compounds such as oxides. Therefore, the difference in catalytic performance (Table 1) was determined by surface crystallographic orientation rather than the chemical state of Pd.

Fig. 6

HAXPES spectra for (a) pure Pd as a reference, (b) as-prepared (101)-oriented foil, (c) (101)-oriented foil after the catalytic test, (d) as-prepared (211)-oriented foil, and (e) (211)-oriented foil after the catalytic test.

Figure 7 shows XRD patterns of the samples. In agreement with the HAXPES results, only diffraction peaks from metallic Pd (FCC) were observed. For each sample, the patterns before and after the catalytic test were similar. The peak intensity ratio depended on the surface crystallographic orientation. The XRD patterns of the (101)-oriented foils had extraordinarily strong 220 peaks while the intensity of the 111 and 200 peaks was weak. In contrast, intensity of the 220 peaks of (211)-oriented foils was weaker than those from the other planes when compared with the (101)-oriented foils. This tendency is consistent with the characterization obtained by EBSD (Fig. 3 and Fig. 5), but the 211 peaks were not observed due to deconstructive interference. Hence, a detailed evaluation by XRD could not be carried out. However, the results in Fig. 7 confirm that the surface crystallographic orientation of the two types of the foils was different.

Fig. 7

XRD patterns for (a) as-prepared (101)-oriented foil, (b) (101)-oriented foil after the catalytic test, (c) as prepared (211)-oriented foil, and (d) (211)-oriented foil after the catalytic test.

3.4 Discussion

The present results were compared with the previously published literature concerning the structure sensitivity NO + CO reactivity over Pd.

Goodman’s group^21–24) reported that Pd(111) is 5-fold more active than the more open (100) and (101) planes though it should be noted that their studies were conducted in lower temperature range than 500°C. They reported that the rate-determining step was nitrogen desorption. The presence of strongly bonded nitrogen species on the less active surface and their removal is thought to play a significant role. Therefore, they concluded that the close-packed structure of Pd(111), on which nitrogen species may not be stabilized, would be much more active compared to that of the edged structure of Pd(101).

Although the foil catalysts in the present study are not much well-defined surface (in terms of the contributions from grain boundaries, various crystallographic planes, cutting edges or side, etc.), the present results can be, basically, explained by the structure sensitivity reported by Goodman’s group.^21–24) The relatively low activity of the (101)-oriented foils is consistent with Goodman’s group’s results, and the high activity of (211)-oriented foils can be attributed to the effect of (111) terraces in the (211) plane. Although the (311) plane is not mentioned in their study, (311) would likely have a trend similar to that of the (101) plane due to the similar “edged” structures (Fig. 2). Thus, (101)-oriented foils in the late stage of the catalytic test, which might be interpreted as (311)-oriented foils (according to Fig. 5), showed relatively low activity. The slightly greater activity of the (311) orientation compared to that of the (101) orientation may correspond to the (100) and (111) facets of the (311) plane (Fig. 2).

Selectivity toward N₂ also can be related to the orientation. Goodman’ group^23–25) reported that Pd(101) possessed greater selectivity (∼60%) than Pd(111) (<20%). Thus, the less active, the more selective toward N₂ it tends to be. Goodman’s group proposed that more nitrogen adatoms on the less active surface resulted in greater N₂ selectivity, increasing the probability of N₂ formation (reactions between adatoms).

This relation between selectivity and activity also can be seen in the results of the “initial” measurements shown in Table 1. However, selectivity values were greater in the present study than in the reports of Goodman’s group. even though our higher catalysis temperature is expected to result in smaller N₂ selectivity according to their reports. The (211)-oriented foils, especially, exhibited much greater selectivity (53.0%) than expected from the result of Pd(111), which might be explained by considering the step site of the (211) plane. Hammer¹⁸⁾ calculated the energy diagram of NO + CO reaction on Pd for the (111), (100), edged (311), and stepped (211) planes, and concluded that the (211) step was the most active for NO dissociation and N₂ formation. Thus, the foils are highly active due to the (111) facet of the (211) plane, and simultaneously relatively N₂ selective due to the (211) step. In contrast, N₂ selectivity of the (101)-oriented foils (73.5%) was slightly greater than expected from the result of Pd(101), yet could be interpreted as similar. Selectivity of the (101)-oriented foils decreased as they became (311) orientated; then, (311) might be less N₂ selective. Selectivity of the (311) plane has not been reported, but the present results correspond to the typical relation between selectivity and activity; i.e., the (311) plane was more active and less N₂ selective than the (101) plane. The similar N₂ selectivity values of the (211) orientation foils to the (311) orientation foils may be related to the (111) or (100) facets.

In conclusion, the present study demonstrates that the catalytic properties of the Pd foils with specific surface orientation is in good agreement with the structure sensitive reaction properties (activity and selectivity) of the Pd single crystal surfaces reported in the previous UHV studies. Thus, the texture-controlled polycrystalline foils, which had been lying in the material gap, could be a unique catalyst taking advantage of the structure sensitivity.

4. Conclusion

Polycrystalline Pd foils with (101) and (211) surface orientations were characterized by EBSD. The NO + CO reactivity of these foils was determined by the orientation, with the (211) orientation much more active than the (101) orientation. Upon an orientation change of (101)-oriented foils to (311)-oriented foils after the catalytic test, the following order of surface crystallographic orientation for activity, (211) ≫ (311) > (101), and N₂ selectivity, (101) > (211) ∼ (311), was implied. These results are consistent with the previously reported structure sensitivity. This work demonstrates the utility of texture control for polycrystalline foils as novel development methodology for unique catalyst in the material gap.

Acknowledgments

This work was financially supported by Grants-in-Aid for Scientific Research [(B) 18H01783 and (B) 19H02452] from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan and “Dynamic Alliance for Open Innovation Bridging Human, Environment and Materials”. The HAXPES experiments were performed at the BL15XU beamline of SPring-8 with the approval of the National Institute for Materials Science (NIMS) under the “Nanotechnology Platform” of the MEXT, Japan (Proposal No. 2019A4906, 2019B4907).

REFERENCES

• 1)   Y.  Xu,  T.  Hirano,  H.  Kunieda,  Y.  Hara and  Y.  Miyata: Catal. Sci. Technol. 10 (2020) 2004–2019.
• 2)   Y.  Xu,  T.  Harimoto,  L.  Wang,  T.  Hirano,  H.  Kunieda,  Y.  Hara and  Y.  Miyata: Chem. Eng. Process. - Process Intensif. 129 (2018) 63–70.
• 3)   T.  Hirano and  Y.  Xu: Int. J. Hydrogen Energ. 42 (2017) 30621–30629.
• 4)   Y.  Hiramitsu,  M.  Demura,  Y.  Xu,  M.  Yoshida and  T.  Hirano: Appl. Catal. A Gen. 507 (2015) 162–168.
• 5)   H.  Yoshida,  R.  Kakei,  Y.  Kuzuhara,  S.  Misumi and  M.  Machida: Catal. Today 332 (2019) 245–250.
• 6)   Y.P.  Hsieh,  D.R.  Chen,  W.Y.  Chiang,  K.J.  Chen and  M.  Hofmann: RSC Adv. 7 (2017) 3736–3740.
• 7)   Z.R.  Robinson,  P.  Tyagi,  T.M.  Murray,  C.A.  Ventrice,  S.  Chen,  A.  Munson,  C.W.  Magnuson and  R.S.  Ruoff: J. Vac. Sci. Technol. A 30 (2012) 011401.
• 8)   T.  Miyao,  Y.  Kawashima,  Y.  Sawaguri and  S.  Naito: Sekiyu Gakkaishi (J. Jpn. Petrol. Inst.) 45 (2002) 382–388.
• 9)   D.J.  Dingley and  V.  Randle: J. Mater. Sci. 27 (1992) 4545–4566.
• 10)   F.J.  Humphreys: Scr. Mater. 51 (2004) 771–776.
• 11)   X.  Chen,  Y.  Lyu,  U.  Nwabara and  J.W.  Schwank: Catal. Today 323 (2019) 148–158.
• 12)   Y.  Cao,  R.  Ran,  X.  Wu and  D.  Weng: Appl. Catal. A Gen. 501 (2015) 17–26.
• 13)   J.  Cortés,  E.  Valencia,  G.  Aguila,  E.  Orellana and  P.  Araya: Catal. Lett. 126 (2008) 63–71.
• 14)   R.  Di Monte,  P.  Fornasiero,  J.  Kašpar,  P.  Rumori,  G.  Gubitosa and  M.  Graziani: Appl. Catal. B 24 (2000) 157–167.
• 15)   K.  Almusaiteer and  S.S.C.  Chuang: J. Catal. 184 (1999) 189–201.
• 16)   A.  Obuchi,  S.  Naito,  T.  Onishi and  K.  Tamaru: Surf. Sci. 122 (1982) 235–255.
• 17)   H.U.  Shin,  D.  Lolla,  Z.  Nikolov and  G.G.  Chase: J. Ind. Eng. Chem. 33 (2016) 91–98.
• 18)   B.  Hammer: J. Catal. 199 (2001) 171–176.
• 19)   L.  Zhang,  I.A.W.  Filot,  Y.Q.  Su,  J.X.  Liu and  E.J.M.  Hensen: J. Catal. 363 (2018) 154–163.
• 20)   H.  Gao: Appl. Catal. A Gen. 529 (2017) 156–166.
• 21)   E.  Ozensoy and  D.W.  Goodman: Phys. Chem. Chem. Phys. 6 (2004) 3765–3778.
• 22)   D.R.  Rainer,  S.M.  Vesecky,  M.  Koranne,  W.S.  Oh and  D.W.  Goodman: J. Catal. 167 (1997) 234–241.
• 23)   S.M.  Vesecky,  D.R.  Rainer and  D.W.  Goodman: J. Vac. Sci. Technol. A 14 (1996) 1457–1463.
• 24)   S.M.  Vesecky,  P.  Chen,  X.  Xu and  D.W.  Goodman: J. Vac. Sci. Technol. A 13 (1995) 1539–1543.
• 25)   X.  Xu,  P.  Chen and  D.W.  Goodman: J. Phys. Chem. 98 (1994) 9242–9246.
• 26)   X.  Xu and  D.W.  Goodman: Catal. Lett. 24 (1994) 31–35.
• 27)   K.  Nakao,  S.I.  Ito,  K.  Tomishige and  K.  Kunimori: J. Phys. Chem. B 109 (2005) 17579–17586.
• 28)   K.  Nakao,  H.  Hayashi,  H.  Uetsuka,  S.I.  Ito,  H.  Onishi,  K.  Tomishige and  K.  Kunimori: Catal. Lett. 85 (2003) 213–216.
• 29)   G.  Prévot,  O.  Meerson,  L.  Piccolo and  C.R.  Henry: J. Phys. Condens. Matter 14 (2002) 4251–4269.
• 30)   L.  Piccolo and  C.R.  Henry: J. Mol. Catal. A Chem. 167 (2001) 181–190.
• 31)   L.  Piccolo and  C.R.  Henry: Appl. Surf. Sci. 162–163 (2000) 670–678.
• 32)   F.  Gao,  Y.  Wang and  D.W.  Goodman: J. Catal. 268 (2009) 115–121.
• 33)   R.D.  Doherty,  D.A.  Hughes,  F.J.  Humphreys,  J.J.  Jonas,  D.  Juul Jensen,  M.E.  Kassner,  W.E.  King,  T.R.  McNelley,  H.J.  McQueen and  A.D.  Rollett: Mater. Sci. Eng. A 238 (1997) 219–274.
• 34)   K.  Momma and  F.  Izumi: J. Appl. Crystallogr. 44 (2011) 1272–1276.
• 35)   Y.  Murase,  N.  Miyauchi,  A.  Itakura and  H.  Katayama: Mater. Trans. JIM 62 (2021) 41–47.

Appendix

A1. Orientation change during the sample preparation procedure

It is implied that (101) orientation of the simply annealed foils might be related to the plane slipping during the rolling. Based on the EBSD analysis in Fig. 3(a) and Fig. A1(a), the texture of the (101)-oriented foil can be described accurately as ⟨111⟩{101} texture. Since the slip system of FCC consists of the slip planes of {111} and the slip direction of ⟨101⟩, some grains would probably be induced the orientation related to this slip system during the rolling, and then, the annealing might let these grains grow. However, the EBSD analysis of the as-rolled (as-received) Pd foil is not shown mainly because the roughness of the surface made it difficult to obtain the clear data. Besides, the recrystallization texture formation behavior is, in general, still not fully understood. Then, elucidating the exact mechanism was difficult at this time.

Fig. A1

IPF mappings with respect to rolling direction for as-prepared (a) (101)-oriented foil and (b) (211)-oriented foil, and (c) (101)-oriented foil and (d) (211)-oriented foil after the catalytic test. The same data sets as Fig. 3 are used for (a) and (b), and as Fig. 5 for (c) and (d).

The XRD patterns for the samples after each treatment are shown in Fig. A2. The results for the foils after such as H₂ treatment & He treatment (e), H₂ treatment & He treatment & NO + CO treatment (f), and just NO + CO treatment (g), are included as references. In this section, XRD was performed just to verify the phase not to evaluate the orientation. The dotted lines in Fig. A2 correspond to the diffractions from Pd (FCC), and hence, the foils except right after the H₂ treatment (Fig. A2(c)) mainly consists of metallic Pd. The deviated peaks in Fig. A2(c) can be attributed to hydride (PdH_x), whose structure is also FCC and lattice constant is larger than that of pure Pd. These diffraction peaks from hydride disappeared when the heat treatment at 500°C was applied in both NO + CO (d) and He (e) flow, implying the hydride decomposition.

Fig. A2

The XRD pattern for (a) as-rolled (as-received) foil and foils after various pretreatments ((b) annealing, (c) H₂ treatment, (d) H₂ treatment & NO + CO, (e) H₂ treatment & He treatment, (f) H₂ treatment & He treatment & NO + CO, (g) NO + CO, and (h) H₂ treatment & NO + CO & annealing).

Figure A3 shows the results of EBSD analysis for the foils after each pretreatment step. The results for some reference sample (e, f) are also displayed. The distributions of crystallographic planes are displayed by color contour and lines. The foils after H₂ treatment (b) exhibited (100)-orientated grains of hydride. This (100) orientation was maintained after the NO + CO treatment (c) though the hydride disappeared during this treatment as shown in Fig. A2. It is interesting that when hydride decomposition was conducted in He flow, (100) orientation was not maintained, and instead, weak (101) orientation was observed after the following NO + CO treatment (f). Therefore, to obtain (100)-oriented Pd foil, NO + CO reaction right after H₂ treatment was required in our experiments. The most likely possibility is that the (100)-oriented hydride grains were formed during H₂ treatment because of the lattice expansion and/or the phase transition, and this (100)-orientation was pinned during the NO + CO treatment possibly due to the interaction with adsorbates. At any rate, the foil surface after H₂ treatment & NO + CO treatment was modified, which resulted in the different crystallographic orientation after the annealing at 1100°C compared with the simply annealed foil. The exact reason of (211) preference in such modified surface is not clear, but flat structure of (100) plane might relate to (211) orientation because (211) can be described as the flattest plane in $[\bar{1}11]$ zone axis.

Fig. A3

Distributions of crystallographic planes determined by EBSD analysis for surface of foils after various pretreatments ((a) annealing, (b) H₂ treatment, (c) H₂ treatment & NO + CO and (d) H₂ treatment & NO + CO & annealing, (e) NO + CO, (f) H₂ treatment & He treatment & NO + CO). 253 × 190 points were scanned with probe size of 1 µm for (b), (c), (d), (e), (f) or 5 µm for (a), (d).

A2. Grain size statistics

The histogram analysis of the grain size was conducted by the software installed in the EBSD system. The results are shown in Fig. A4. The data sets used for the analysis of the as-prepared sample (before the catalytic test) and of the samples after the catalytic test are same as Fig. 3 and Fig. 5, respectively.

Fig. A4

Grain size statistics for (a) (101)-oriented foil and (b) (211)-oriented foil before and after the catalytic test.

A3. Supplementary data of HAXPES

Wide scan and narrow scan for C1s spectrum were also conducted for the same sample as Pd3d analysis in Fig. 6. The results of the wide scan analysis are summarized in Fig. A5. Quite close peak positions of O1s (ca. 531 eV) and Pd3p_3/2 (ca. 533 eV) should be noted, but all the peaks with sufficient intensities can be attributed to Pd. The narrow scan data for C1s displayed in Fig. A6 was used to estimate the peak area. Shirley type background was used for the calculation. The calculated values of the C1s/Pd3d_5/2 peak area ratio were 0.029, 0.025, 0.032, and 0.025 for the as-prepared (101) oriented foil, (101) foil after the catalytic test, as-prepared (211) oriented foil, and (211) oriented foil after the catalytic test, respectively.

Fig. A5

Wide scan analysis for (a) as-prepared (101)-oriented foil, (b) (101)-oriented foil after the catalytic test, (c) as-prepared (211)-oriented foil, and (d) (211)-oriented foil after the catalytic test.

Fig. A6

C1s spectra for (a) as-prepared (101)-oriented foil, (b) (101)-oriented foil after the catalytic test, (c) as-prepared (211)-oriented foil, and (d) (211)-oriented foil after the catalytic test.