Imaging and chemically probing catalytic processes using field emission techniques: a study of NO hydrogenation on Pd and Pd–Au catalysts

Cédric Barroo *ab, Matthieu Moors a and Thierry Visart de Bocarmé *ab
aChemical Physics of Materials and Catalysis, Université libre de Bruxelles, CP243, 1050 Brussels, Belgium. E-mail: cbarroo@ulb.ac.be; tvisart@ulb.ac.be
bInterdisciplinary Center for Nonlinear Phenomena and Complex Systems (CENOLI), Université libre de Bruxelles, 1050 Brussels, Belgium

Received 17th May 2017 , Accepted 17th August 2017

First published on 18th August 2017


Nitric oxide hydrogenation is investigated on palladium and gold–palladium alloy crystallites, i.e. the extremity of sharp tip samples aimed at modelling a single catalytic grain. Field ion microscopy and field emission microscopy are used to monitor adsorption and reaction in real time. One-dimensional atom probe and atom probe tomography are used on the same samples to unravel the surface composition of the adsorbed layers and the composition of the very first atomic layers of the Pd–Au surface. At constant NO pressure and at 450 K, the surface composition of the adsorbed layer on Pd samples shows a strong hysteresis behavior when H2 gas is varied. Under oxidizing conditions, N2O is formed via the occurrence of surface (NO)2 dimers. In the presence of Pd–Au alloys, the NO–H2 interaction comprises a simple NO dissociation causing the formation of surface NO2 species. On Pd–Au tip samples, atom probe tomography proves the occurrence of significant surface enrichment of palladium atoms in the presence of NO gas, but it is not sufficient to drift the behavior of the surface to that of pure palladium. Accordingly, external control parameters could be changed to tune the surface composition of Pd–Au catalysts and thus their activity and/or selectivity.


A Introduction

In the frame of automotive pollution control by catalysis, palladium (Pd) is used due to its high thermal stability and favourable oxidation activity at low temperatures1,2 as compared to other platinum group metals (PGMs). Within recent years, there has been growing interest in the use of Au and Au-based materials for catalytic applications.3–5 Alloying gold with less noble transition metals seems to be very promising for fine tuning the catalytic performance of well-defined reactions.6 The control of the catalytic activity and selectivity of gold-based materials can be conducted using different compositions or composition gradients so as to obtain synergistic effects in the alloy as compared to isolated metals.7,8 In the case of the Pd–Au alloy, different catalytic reactions have been investigated, such as the combustion of methane,9 the selective oxidation of alcohols,10,11 and the water-gas shift reaction.12 Both Pd and Au can be used for deNOx reactions, and using Pd–Au alloys for the selective catalytic reduction of NOx would allow extending the activity of these catalysts. To obtain a better understanding of the catalytic processes occurring at the surface of the catalysts, field emission techniques are particularly suited. Indeed, these techniques use samples prepared as sharp tips, the size and morphology of which are close to the size and shape of a single catalytic nanoparticle. These techniques enable monitoring of the catalytic reaction, in real time, during the ongoing processes to probe the chemical composition of the adsorbate layer and eventually to determine the surface composition before and after reaction, down to the nanoscale. This manuscript focusses on the processes occurring during NO adsorption and NO hydrogenation on Pd and Pd–Au samples. The alloy composition Pd–38 at% Au was chosen for a composition close to that of Pd–Au catalysts used for selective NOx reduction.13,14 A combination of field emission microscopy, field ion microscopy, one-dimensional atom probe (1DAP) and atom probe tomography (APT) allows obtaining a rather complete picture of the system, from both the catalytic reaction and catalytic material viewpoints.

B Experimental

B1 Basic principle of the techniques

Field emission techniques include a series of imaging and tomographic techniques using metallic samples prepared as sharp tips. Among these techniques, field emission microscopy (FEM), field ion microscopy (FIM), one-dimensional atom probe, also referred to as pulsed field desorption mass spectrometry (PFDMS), and atom probe tomography (APT) are used in this study. FEM is based on the emission of electrons from a tip-sample when the latter is negatively charged as compared to a detector screen. The intensity of the electron emission depends on the distribution of the local electric field, as well as on the local work function, which in turn depends on the crystallographic orientation of the exposed facets, on temperature and on the nature of possible adsorbate species. Field emission is governed by the Fowler–Nordheim equation.15,16 FEM can be used not only under vacuum conditions, but also in reactive gas mixture environments. Currently, this method is mostly used to study adsorption properties, or the occurrence of specific dynamics on catalyst samples.17–21 Field ion microscopy is based on the ionisation of an imaging gas at the surface of the tip-sample when the sample is positively charged as compared to the detector screen. The ionisation probability depends on the local curvature of the surface and on the ionisation potential of the imaging gas. A noble gas is typically used at low temperatures (≈ 50 K) to image surface reconstructions,22–25 or reactive gases can be used at higher temperatures (up to ≈ 850 K) to image catalytic behaviours.26–28 One type of combination between a FIM device and a time-of-flight mass spectrometer is referred to as the 1D atom probe. For this case, a probe-hole is drilled into the detector, directly followed by a time-of-flight mass spectrometer. This technique is used during the ongoing process, i.e. in the presence of reactive gases in the reaction chamber. By using the combination of a static field and a pulsed field (of various amplitudes and frequencies), it is possible to desorb the adsorbate layer and, with pulses of higher amplitudes, the atoms on the surface of the sample. Changing the pulse frequency opens a possibility to investigate not only the nature of adsorbates and intermediates but also their evolution with time.29–31 More precisely, atom probe techniques are based on the field evaporation phenomenon where a high electric field is applied to the tip-sample so as to field evaporate the sample as ions, in a controlled way, atom by atom, atomic layer after atomic layer. By using a detector sensitive to the position of impact, oxidation and surface segregation processes can be studied.32,33 A more detailed description of the principle of these techniques can be found elsewhere.30,34–37

B2 Sample preparation

Samples suitable for FEM/FIM/1DAP/3DAP analysis are prepared as sharp tips. The tip extremity typically has a radius of curvature in the 10–50 nm size range. Pd and Pd–Au samples are electrochemically etched from a high purity wire in a 10–20% aqueous solution of KCN, with an applied potential of 3 VDC.24,38,39 After introduction in the microscope, cycles of thermal annealing, field evaporation and Ne+-ion sputtering are applied to the sample to obtain a clean and atomically smooth sample for imaging. The sample is first characterised by FEM and FIM at low temperatures before further physicochemical treatments.

C Results and discussion

The first part of this section presents the atomic-scale characterisation of Pd and Pd–Au samples by FIM. Sections C2 and C3 then describe the behaviour occurring during the NO hydrogenation on pure Pd, previously studied by both FEM (C2) and FIM (C3). Section C4 is dedicated to the observation of surface reconstruction of a Pd–Au sample under NO gas exposure, using FIM and FEM, as well as local chemical probing during NO exposure. Sections C5 and C6 finally discuss the presence of surface segregation and present a titration wave in the Pd–Au system.

C1 Sample characterisation

As mentioned, the first step towards the analysis of dynamic behaviour is the characterisation of the sample. A field ion micrograph of a clean (111)-oriented Pd tip is reproduced in Fig. 1a, where the atomic resolution capabilities of the FIM technique at cryogenic temperatures can be observed. The radius of curvature of this sample is estimated to be 15 nm. Fig. 1b presents a FIM micrograph of a clean (001)-oriented Pd–38 at% Au tip. Both Pd and Au crystallise in the same face-centered cubic lattice with cell parameters of 389.07 pm and 407.82 pm, respectively. Even though the presence of two different metals may induce an imperfect FIM image as compared to pure metals, due to the lattice parameter differences, it is still possible to discern facets of different crystallographic orientations with very good accuracy. Under these imaging conditions and after field evaporation at low temperatures, the composition of the surface before any reaction is identical to that of the bulk phase, a fact that might no longer be true under reactive conditions.
image file: c7cy00994a-f1.tif
Fig. 1 Field ion microscopy (FIM) characterisation of (left) a (111)-oriented Pd tip and (right) a (001)-oriented Pd–38 at% Au tip before any reaction. Atomic resolution is achieved and the main Miller indices are presented. Imaging conditions: T = 60 K, PNe = 10−3 Pa, and F ∼ 35 V nm−1.

C2 NO + H2 reaction on Pd: FEM studies

NO hydrogenation was previously studied on Pd tip-samples by both FEM and FIM techniques. This section reports the previously observed results obtained in FEM mode by Cobden et al.;40 it was shown that the adsorption of hydrogen in the 300–500 K temperature range induces an increase in the work function, Δφ > 0, with a maximum of 0.21 eV.40 Also, the adsorption of NO at 300 K induces a Δφ > 0 of 0.46 eV, but the FEM pattern remains the same as the clean sample, a sign of molecular adsorption of NO. However, above 400 K, the FEM pattern evolves with increasing electron emission, from {310} and {311} facets, and with decreasing emission on {110} facets. It was supposed that the changes in the FEM pattern result from the dissociative adsorption of NO gas.40Fig. 2a plots the variation in the work function Δφ during the NO + H2 reaction on Pd at 450 K. At this temperature, there is no variation of the work function for pure NO adsorption. Therefore, 450 K was chosen for the NO + H2 reaction to help in the identification of the species that cause the changes of the work function. When both NO and H2 gases are present in the system, the FEM pattern is similar to a clean Pd surface, as can be seen in (1.). When the supply of H2 is switched off, Δφ decreases and the pattern evolves towards a configuration encountered when pure NO is dosed at 575 K. The O(ads) originating from the dissociative adsorption of NO can be removed with H2 at T > 440 K.41 Therefore, it is supposed that when H2 is present, the reaction with O(ads) to form H2O(ads) is fast enough to hinder the accumulation of O(ads) species. In contrast, when H2 is absent, the O(ads) species builds up on the surface and induces a decrease in emission.40,42
image file: c7cy00994a-f2.tif
Fig. 2 Investigations of the NO + H2 reaction on Pd tips: (a) variations in the work function during FEM imaging of a Pd tip-sample during NO + H2 interaction (1.) and when the hydrogen gas supply is stopped (2.). The corresponding FEM micrographs are also presented. Imaging conditions: T = 450 K, PNO = 8 × 10−7 Pa, and PH2 = 8 × 10−6 Pa (when switched on); (b) mass spectrum obtained during FIM imaging of the NO + H2 reaction, the NO-side of the hysteresis, highlighting the presence of Pd-oxides and formation of (NO)2 dimers. The probed region of the (001)-oriented tip is represented by the red circle. Imaging conditions: T = 450 K, PNO = 10−3 Pa, and F ∼ 9 V nm−1; (c) mass spectrum during FIM imaging of the NO + H2 reaction, the H2-side of the hysteresis, highlighting the presence of a metallic surface. The probed region is represented by the red circle. Imaging conditions: T = 450 K, PNO = 10−3 Pa, and F ∼ 9 V nm−1. Panel (a) – reprinted from ref. 40, Copyright 1999, with permission from Elsevier.

C3 NO + H2 reaction on Pd: FIM and 1DAP studies

In FIM mode, NO hydrogenation was studied in the 400–600 K temperature range and shows a strong hysteresis effect.39 A kinetic phase diagram can be established. To achieve this, the sample is heated up to a targeted temperature, NO gas is introduced in the microscope, at a pressure in the 10−4–10−2 Pa range, and an electric field is applied, sufficient for ionisation of NO species to occur (∼9 V nm−1). Hydrogen gas is then introduced into the system, and two regions can be observed: the first one, observed at low PH2, referred to as the ‘NO-side’ (Fig. 2b), and the second one, observed at high PH2, referred to as the ‘H2-side’ (Fig. 2c). A hysteresis loop is defined, meaning that the transition from the ‘NO-side’ to the ‘H2-side’ depends on whether the hydrogen partial pressure is increased or decreased at constant NO partial pressure. If pressures are kept constant, the system remains stable, i.e. no transition between sides can be observed.39 However, when the hydrogen pressure is increased, abrupt transformation of the FIM pattern from bright to dark occurs. Complementary 1DAP measurements were performed when the system lies in one of the two branches of the hysteresis loop. The probed region is defined by the red circle in Fig. 2b and c and corresponds to a few hundreds of surface atoms. On the ‘NO-side’, we can observe a bright pattern without any correlation with the underlying structure. 1DAP measurements show that NO+ is a major species and NO thus plays the role of an imaging gas. The presence of (NO)2+ and N2O+ peaks suggests a dimerization process:
 
NO(g) + S ⇌ NO(ads)(1)
 
2NO(ads) ⇌ (NO)2(ads)(2)
 
(NO)2(ads) ⇌ N2O(ads) + O(ads)(3)
where ‘S’ stands for an empty surface site. Results indicate NO adsorption via the nitrogen atom and N–N bond formation between two adjacent species on the surface. These measurements also prove that Pd undergoes oxidation, with either the presence of adsorbed species O(ads) or subsurface species O(sub), coming from the dissociation of NO:
 
NO(ads) + S ⇌ N(ads) + O(ads)(4)
 
O(ads) + # ⇌ O(sub) + S(5)
where ‘#’ stands for an empty subsurface site.

Subsurface oxygen species have been observed in the past on Pd and were connected to the oscillation mechanism observed during CO oxidation on Pd(110).43

On the ‘H2-side’ however, the FIM pattern is dark. It is known that the O-covered surface should be brighter than the H-covered surface using NO as an imaging gas.34,44 The absence of N2O+ in the 1DAP spectra can be observed, as well as the presence of metallic palladium Pd+, a sign that the surface of the catalyst is not oxidised under high H2 partial pressures. At high H2 pressure, hydrogen atoms are known to diffuse in subsurface positions:

 
H2(g) + 2S ⇌ 2H(ads)(6)
 
H(ads) + # ⇌ H(sub) + S(7)
and it is supposed that the reverse process is inhibited at lower H2 pressures (NO-side) since Pd surface oxides reduce the mobility of the dissolved H from the bulk to the tip apex.39,45 It has to be noted that on both NO- and H2-sides, only low amounts of water are detected, probably due to contamination rather than the reaction. Also, no NHx species were observed, probably due to the low partial pressure of H2. The recombination of atomic N into N2(g) cannot be directly proven by a signal of N2+ at m/z = 28 because of the fast thermal desorption of N2 before its ionization at the surface. In agreement with the observations by FIM, the study of NO + H2 reaction on the extended Pd(111) facet also presents hysteresis phenomena at the rate of NH3 and H2O formation,2 where H2O is mainly formed by the reaction between O(ads) and H(sub).46,47

C4 NO adsorption on the Pd–Au surface: surface reconstruction and chemical probing

After the description of the Pd system by FEM and FIM, we turn the discussion towards the processes occurring on a Pd–Au alloy. The exposure of a Pd–38 at% Au tip sample to NO gas induces a severe morphological change that can be observed in Fig. 3a and b by FIM imaging at 400 K with NO as the imaging gas. The system evolves from a rather homogeneous bright FIM pattern, similar to the one depicted in Fig. 2b, to a FIM pattern presenting a 4-fold symmetry, with regions of high brightness intensity and regions of lower brightness intensity. Since NO acts as the imaging gas, any variations in the FIM brightness pattern must be attributed to variations of the probability of NO ionisation over the surface of the sample, which can be due to the changes of the surface composition and/or of the surface morphology. FEM imaging at a somewhat lower temperature is presented in Fig. 3c, together with the supposed indexing. The {012} facets appear brighter in FEM, due to their low work function, and are thus assigned to the bright regions.
image file: c7cy00994a-f3.tif
Fig. 3 FIM imaging of a (001)-oriented Pd–38 at% Au sample (a) before and (b) after reconstruction due to the exposure to NO for 1 h. Imaging conditions: T = 400 K, PNO = 10−3 Pa, and F ∼ 12 V nm−1; (c) FEM imaging of the same sample after reconstruction, obtained at 300 K. The circular dark region corresponds to the “probe-hole” for 1DAP experiments.

Early works highlighted the presence of surface reconstruction of Pd field emitter tips due to the formation of palladium hydride (PdHx) particles.48–50 Furthermore, on pure Pd tips, it was observed that a strong faceting occurs after annealing the tip at 500 K in the presence of a sub-monolayer amount of oxygen.24 The main observations are: the {100} planes broke up into small {100}, {112} and {012} facets; and only {111} and enlarged {011} facets are present at the surface of the reconstructed Pd sample. This is in agreement with the calculated equilibrium-shape of Pd in the presence of small amounts of O(ads).51 Except for the imaging part, experiments were performed under field-free conditions, and complementary analysis proved that thermally-induced faceting is absent up to an annealing temperature of 750 K. Therefore, reconstructions observed on Pd are triggered by the presence of O(ads) under the reported conditions within the timescale of the experiments. These reconstructions typically take a few tens of minutes to develop.

The tip evolves towards its equilibrium shape, which depends on the temperature, field and pressure conditions. In the case of Pd–Au samples, morphological reconstructions may explain the observed FIM and FEM patterns. The presence of {012} facets on Pd–Au is in line with reconstructed Pd tips. Dense facets tend to extend to the detriment of more open facets, and {113} facets are here supposed to extend. Also, a morphological reconstruction from a quasi-hemispherical shape to a polyhedral shape generates the presence of edges where the local electric field is higher, therefore inducing a higher probability of field ionisation or field emission at these specific edge-regions. Due to the dissociation of NO (as it will be described below), the presence of subsurface O(sub) species can't be discarded, and further characterisation at low temperature is required to obtain a full picture of the Pd–Au surface reconstruction.

The chemical analysis of the adsorbate layer conducted during the interaction of NO gas with a Pd–Au sample is presented in Fig. 4. The results were obtained by applying a static field (for imaging purposes) FS, on which a pulsed field FP was added so as to field-desorb the adsorbate species at well-defined instants. The mass spectrum contains mainly H2O+, NO+, NO2+ and peaks that have been assigned to PdNO+. As in the Pd case, NO is used as the imaging gas, and therefore, its presence is dominant in the spectrum. The water peak is supposed to be a contaminant from the microscope since the sample was neither exposed to NO nor to H2 during this experiment. The presence of nitrogen dioxide NO2+ species can be explained by the combination of NO(ads) with O(ads) according to the following steps:

 
NO(g) + S ⇌ NO(ads)(1)
 
NO(ads) + S ⇌ N(ads) + O(ads)(4)
 
NO(ads) + O(ads) ⇌ NO2(ads) + S(8)


image file: c7cy00994a-f4.tif
Fig. 4 1DAP analysis during the NO interaction with a Pd–38 at% Au sample. The mass spectrum differs from the one obtained on pure Pd (Fig. 2a) with the presence of NO2+, a sign of NO dissociation. Conditions of acquisition: T = 300 K, PNO = 10−3 Pa, FS ∼ 14 V nm−1, FP ∼ 12 V nm−1, and 106 cycles at 1000 Hz; Insert: mass spectrum obtained under field conditions proving the dissociation of NO gas. Conditions of acquisition: FS = 0 V nm−1 and FP ∼ 16 V nm−1.

As stated, the absence of N(ads) in the spectrum can be explained by its fast recombination and desorption:

 
2N(ads) ⇌ N2(g) + 2S(9)

Alternatively, the evolution of the dimer (NO)2+ towards NO2+ cannot be excluded but the absence of (NO)2+ in the mass spectrum is not in line with this hypothesis.

 
(NO)2(ads) ⇌ NO2(ads) + N(ads)(10)

Experiments at different amplitudes of FS and FP prove the presence of N+ and O+, a strong indication of NO dissociation (see the inset in Fig. 4). It has to be noted that previous work on Au tips proved the absence of oxygen species in the mass spectra under O2 exposure,52 and showed the presence of (NO)2 and N2O species under NO exposure, without any dissociation,53 as explained by eqn (1)–(3), similar to the Pd case. The presence of Pd in the Au system thus allows the dissociation of NO to occur and the subsequent formation of NO2viaeqn (8). This could be due to either a synergy between Au and Pd atoms, or the presence of a higher Pd/Au ratio at the surface (as will be discussed in the next sections).

C5 NO adsorption on Pd–Au: surface segregation

The previous section focussed on the observation of morphological changes of the surface, as well as the composition of the adsorbate-layer during exposure to NO. However, another process that might occur on alloy systems during gas exposure is the occurrence of segregation/depletion behaviour. This process leads to a surface composition that deviates from the bulk composition. This evolution may affect the catalytic activity of the sample and the selectivity of the process.33,54–57 The extent of the segregation varies with parameters such as the sublimation energy, the lattice constants of the metal elements, and the possible presence of a reactive gaseous environment. It then becomes possible to fine-tune the surface composition by controlling the temperature which increases the diffusion of species, as well as the pressure and nature of different reactive gas species58,59 so as to induce the phenomenon of adsorbate-induced segregation, also referred to as ‘chemical pumping’.38

To study the occurrence of segregation in our system, APT analyses were performed. The APT instrument is equipped with a reaction cell54 allowing exposure of the tip-sample to different physicochemical treatments, at temperatures up to 800–900 K and pressures up to atmospheric pressure. The transfer between the analysis chamber and the reaction cell also avoids the presence of contaminants from exposure to air. The procedure consists of 1) analysing the fresh sample by atom probe tomography, so as to get a blank experiment, 2) transferring the sample into the reaction cell where physicochemical treatments can be applied, 3) transferring the sample in the analysis chamber, and 4) analysing the sample after treatment by atom probe tomography.

Fig. 5a depicts a 3D reconstruction of the Pd–Au sample after atom probe experiments. The sample didn't undergo any treatments and thus represents a blank experiment for the composition of the Pd–Au alloy under UHV conditions. Only palladium and gold ions are detected by mass spectrometry. Since the catalytic activity depends on the surface, the reconstruction has been performed for the surface only and is estimated to have ∼10–13 atomic layers. A rather homogeneous distribution of Pd and Au ions throughout these atomic layers is observed. It has to be noted that the analysis is performed at 100 K to avoid any self-diffusion that may lead to concentration artefacts. One way to rationalise this observation is made via a cumulative diagram plotting the number of ions of one of the metals (Au in this case) as a function of the total number of detected metal ions (Au + Pd). This is represented by the dark curve in Fig. 5b. The slope corresponds to the concentration of the alloy, and any change in the slope reflects changes in the local Au/Pd ratio, which are signs of segregation or depletion behaviour. We can see that the slope remains linear, a sign that both surface and bulk compositions are identical, with a composition of Pd–38 at% Au. If the sample is exposed to NO(g) at a temperature of 573 K for 15 minutes, a rather different slope can be observed (Fig. 5b, grey slope), a sign that the surface composition has evolved to Pd–20 at% Au.38 Segregation of Pd towards the surface thus occurs in the system. In view of 1DAP experiments, this segregation can be correlated with the NO dissociation and the subsequent formation of Pd–O bonds at the surface. Indeed, Pd–O bonds are more easily formed than Au–O bonds,60 so O(ads) is preferentially adsorbed on Pd atoms of the alloy surface and it will be thermodynamically favourable for the Pd to segregate at the surface of the catalyst. We mention that the segregation is observed at 573 K, but exposure to NO at 300 K does not produce any changes in the slope of the cumulative diagram as compared to that of the blank experiment, further confirming that the segregation is thermally activated.


image file: c7cy00994a-f5.tif
Fig. 5 (a) 3D reconstruction of the Pd–38 at% Au tip before any reaction showing the homogeneous distribution of Pd and Au atoms within the probed region corresponding to ∼10 atomic layers; (b) cumulative diagram plotting the total number of Au and Pd ions collected as a function of the number of Au ions before and after exposure to NO gas showing Pd surface segregation. Exposure conditions: T = 573 K, PNO = 6 × 103 Pa, and t = 15 min. Panel (b): Reprinted from ref. 38, Copyright (2009), with permission from Elsevier.

C6 NO + H2 reaction on Pd–Au

Finally, after proving the occurrence of surface reconstruction, surface segregation and NO dissociation on the Pd–Au catalyst, the reaction between NO and H2 can be imaged. Particularly, we intend to prove the formation of an ‘oxide’ layer, i.e. the presence of palladium–oxide, O(ads) and/or O(sub), by titration with hydrogen gas. The reaction between O(ads) and H(ads) is supposed to form water:
 
O(ads) + H(ads) ⇌ OH(ads) + S(11)
 
OH(ads) + H(ads) ⇌ H2O(g) + 2S(12)

The methodology to observe the titration by FIM is as follows: the tip is exposed to NO at 450 K and a pressure of 3 × 10−3 Pa, and the field is increased so as to observe a bright pattern corresponding to the imaging of the sample with NO as the imaging gas, as presented in Fig. 2b for the case of pure Pd. The imaging field is kept constant, the NO gas supply is stopped and the system is evacuated down to a residual pressure in the 10−6 Pa range, insufficient pressure for imaging the tip. An increasing pressure of hydrogen is then introduced into the system up to the moment where changes in the FIM pattern occur. The evolution of the FIM pattern is presented in Fig. 6, where a titration wave can be observed. The electric field used is too low to ionise hydrogen gas. Therefore, the bright patterns correspond to the formation of water at the surface which can be easily ionised. The formed water molecules here play the role of imaging gaseous species. The hydrogen pressure used to observe such titration waves corresponds to ∼10−4 Pa. The lowest temperature necessary to observe this phenomenon is 375 K, which is due to the activation energy for NO dissociation. However, it is known that NO dissociation is facilitated in the presence of an electric field,37 and it is expected that the field could be used to trigger the reaction. It has to be noted that similar experiments on pure Pd generates a somehow more ‘homogeneous’ titration wave, with a more abrupt behaviour, i.e. occurring on a smaller time scale. Indeed, it is supposed that oxygen adsorption occurs preferentially on Pd sites, and that hydrogen dissociation is more favourable on Pd. Subsequently, it is logical that the process would be hindered on a Pd–Au surface as compared to a Pd surface. The presence of Au at the surface can be seen as noxious for the catalytic process, but can also be used to lower the global reactivity and increase the selectivity of the process, as recently shown for Au-based catalysts.4,61,62


image file: c7cy00994a-f6.tif
Fig. 6 FIM imaging during the ongoing reaction between H2 and a NO-precovered Pd–38 at% Au sample, proving the formation of water. NO-exposure conditions: T = 450 K, PNO = 3 × 10−3 Pa, and t = 30 s. Imaging conditions: PH2 ≈ 10−4 Pa and F ∼ 15 V nm−1.

Conclusions

The behaviour during NO adsorption and hydrogenation was studied by field emission techniques on a single catalytic nanoparticle, allowing obtaining a fundamental understanding of the processes. The dissociation of NO does not occur on Au samples, but is effective on Pd samples, leading to a partially oxidized surface. The mixture of NO and H2 leads to a strong hysteresis observed by FIM, where no transition between the two sides of the hysteresis loop occurred when the pressures are kept constant. Abrupt transitions can be observed when the H2 pressure is varied though. The Pd–Au alloy was investigated during NO adsorption and proved that the catalyst undergoes morphological changes and segregation, which can be traced back to the formation of strong Pd–O bonds, leading to chemical pumping of Pd. Indeed, the Pd surface concentration increases from 62 at% to 80 at%, which in turn may affect the reactivity. Chemical probing during the adsorption of NO shows strong arguments towards the dissociation of NO (presence of N+, O+ and NO2+ species in the mass spectrum) and the absence of dimerisation in the system (as it was in the case of pure Au and pure Pd). This difference proved the synergy between Au and Pd in this system, since the process on the alloy is different than that on isolated metals. Finally, the presence of oxygen species is further confirmed by the observation of a titration wave where water molecules correspond to the imaging gas. Similar observations are made on pure Pd, but in a more homogeneous way, probably due to the presence of Au at the surface. The combination of Au and Pd in supported catalysts could therefore provide a means to improve the catalytic performance of NO reduction in SCR exhaust gases systems.

Further work should focus on the systematic study of surface segregation for different Pd/Au compositions, so as to get a better understanding of the segregation process. From a dynamic point of view, it would be interesting to further investigate the reactive behaviour.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

C. B. thanks the Fonds de la Recherche Scientifique (F.R.S.-FNRS) for financial support.

Notes and references

  1. D. D. Beck and J. W. Sommers, Stud. Surf. Sci. Catal., 1995, 96, 721 CrossRef CAS .
  2. C. A. de Wolf and B. E. Nieuwenhuys, Surf. Sci., 2000, 469, 196 CrossRef CAS .
  3. E. Genty, L. Jacobs, T. Visart de Bocarmé and C. Barroo, Catalysts, 2017, 7, 134 CrossRef .
  4. J. Biener, M. M. Biener, R. J. Madix and C. M. Friend, ACS Catal., 2015, 5, 6263 CrossRef CAS .
  5. M. Haruta, Catal. Today, 1997, 36, 153 CrossRef CAS .
  6. F. Besenbacher, I. Chorkendorff, B. S. Clausen, B. Hammer, A. M. Molenbroek, J. K. Nørskov and I. Stensgaard, Science, 1998, 279, 1913 CrossRef CAS PubMed .
  7. J. Creuze, H. Guesmi, C. Mottet, B. Zhu and B. Legrand, Surf. Sci., 2015, 639, 48 CrossRef CAS .
  8. D. Wang, A. Villa, F. Porta, L. Prati and D. Su, J. Phys. Chem. C, 2008, 112, 8617 CAS .
  9. N. Yang, J. Liu, Y. Sun and Y. Zhu, Nanoscale, 2017, 9, 2123 RSC .
  10. C. R. Chang, B. Long, X. F. Yang and J. Li, J. Phys. Chem. C, 2015, 119, 16072 CAS .
  11. T. A. Silva, E. Texeira-Neto, N. López and L. M. Rossi, Sci. Rep., 2014, 4, 5766 CrossRef CAS PubMed .
  12. M. A. Saqlain, A. Hussain, M. Siddig and A. A. Leitão, RSC Adv., 2015, 5, 47066 RSC .
  13. C. Hamill, R. Burch, A. Goguet, D. Rooney, H. Driss, L. Petrov and M. Daous, Appl. Catal., B, 2014, 147, 864 CrossRef CAS .
  14. K. Fujdala, T. Truex, J. Nicholas and J. Woo, SAE Technical Paper, 2008, 01, 0070,  DOI:10.4271/2008-01-0070 .
  15. R. H. Fowler and L. Nordheim, Proc. R. Soc. London, Ser. A, 1928, 119, 173 CrossRef CAS .
  16. R. G. Forbes, Ultramicroscopy, 1999, 79, 11 CrossRef CAS .
  17. R. Bryl, Vacuum, 1997, 48, 333 CrossRef CAS .
  18. J.-S. McEwen, P. Gaspard, Y. De Decker, C. Barroo, T. Visart de Bocarmé and N. Kruse, Langmuir, 2010, 26, 16381 CrossRef CAS PubMed .
  19. C. Barroo, Y. De Decker, T. Visart de Bocarmé and N. Kruse, J. Phys. Chem. C, 2014, 118, 6839 CAS .
  20. C. Barroo, Y. De Decker, T. Visart de Bocarmé and N. Kruse, Phys. Rev. Lett., 2016, 117, 144501 CrossRef PubMed .
  21. C. Barroo, Y. De Decker, L. Jacobs and T. Visart de Bocarmé, Appl. Surf. Sci., 2017, 412, 564 CrossRef CAS .
  22. C. Barroo, N. Gilis, S. V. Lambeets, F. Devred and T. Visart de Bocarmé, Appl. Surf. Sci., 2014, 304, 2 CrossRef CAS .
  23. R. Bryl, T. Olewicz, T. Visart de Bocarmé and N. Kruse, J. Phys. Chem. C, 2011, 115, 2761 CAS .
  24. R. Bryl, T. Olewicz, T. Visart de Bocarmé and N. Kruse, J. Phys. Chem. C, 2010, 114, 2220 CAS .
  25. T. Visart de Bocarmé, T. Bär and N. Kruse, Surf. Sci., 2000, 454, 320 CrossRef .
  26. J. S. McEwen, P. Gaspard, T. Visart de Bocarmé and N. Kruse, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 3006 CrossRef CAS PubMed .
  27. J. S. McEwen, P. Gaspard, T. Visart de Bocarmé and N. Kruse, J. Phys. Chem. C, 2009, 113, 17045 CAS .
  28. Y. Suchorski and G. Rupprechter, Surf. Sci., 2016, 643, 52 CrossRef CAS PubMed .
  29. N. Kruse, G. Abend and J. H. Block, J. Chem. Phys., 1988, 88, 1307 CrossRef CAS .
  30. N. Kruse, Ultramicroscopy, 2001, 89, 51 CrossRef CAS PubMed .
  31. M. Moors, T. Visart de Bocarmé and N. Kruse, Catal. Today, 2007, 124, 61 CrossRef CAS .
  32. P. A. J. Bagot, A. Cerezo, G. D. W. Smith, T. Visart de Bocarmé and T. J. Godfrey, Surf. Interface Anal., 2007, 39, 172 CrossRef CAS .
  33. C. Barroo, P. A. J. Bagot, G. D. W. Smith and T. Visart de Bocarmé, Investigating Nano-structured Catalysts at the Atomic scale by Field Ion Microscopy and Atom Probe Tomography, in Atomically-Precise Methods for Synthesis of Solid Catalysts, ed. S. Hermans and T. Visart de Bocarmé, RSC Catalysis Series, 2015, pp. 248–295 Search PubMed .
  34. E. W. Müller and T. T. Tsong, Field Ion Microscopy Principles and Applications, American Elsevier Publishing Company, Inc., New York, 1969 Search PubMed .
  35. B. Gault, M. P. Moody, J. M. Carney and S. P. Ringer, Atom Probe Microscopy, Springer, New-York, 2012 Search PubMed .
  36. M. K. Miller, A. Cerezo, M. G. Hetherington and G. D. W. Smith, Atom Probe Field Ion Microscopy, Clarendon Press, Oxford, UK, 1996 Search PubMed .
  37. N. Kruse and T. Visart de Bocarmé, Heterogeneous catalysis and high electric fields, in Handbook of heterogeneous catalysis, ed. G. Ertl, H. Knözinger, F. Schüth and J. Weitkamp, Wiley-VCH Verlag GmbH& Co. KGaA, Weinheim, Germany, 2008, 2nd edn, pp. 870–895 Search PubMed .
  38. T. Visart de Bocarmé, M. Moors, N. Kruse, I. S. Atanasov, M. Hou, A. Cerezo and G. D. W. Smith, Ultramicroscopy, 2009, 109, 619 CrossRef PubMed .
  39. T. Visart de Bocarmé and N. Kruse, Ultramicroscopy, 2011, 111, 376 CrossRef PubMed .
  40. P. D. Cobden, B. E. Nieuwenhuys and V. V. Gorodetskii, Appl. Catal., A, 1999, 188, 69 CrossRef CAS .
  41. R. G. Sharpe and M. Bowker, Surf. Sci., 1996, 360, 21 CrossRef CAS .
  42. I. Kojima, E. Miyazaki and I. Yasumori, J. Chem. Soc., Faraday Trans. 1, 1982, 78, 1423 RSC .
  43. G. Ertl, Adv. Catal., 1990, 37, 213 CAS .
  44. H. J. Kreuzer and R. L. C. Wang, Z. Phys. Chem., 1997, 202, 127 CrossRef CAS .
  45. C. Barroo, S. V. Lambeets, F. Devred, T. D. Chau, N. Kruse, Y. De Decker and T. Visart de Bocarmé, New J. Chem., 2014, 38, 2090 RSC .
  46. H. Neergaard-Waltenburg, Q. Gao, R. D. Ramsier and J. T. Yates Jr, J. Phys. Chem., 1994, 98, 12075 CrossRef CAS .
  47. T. Engel and H. Kuipers, Surf. Sci., 1979, 90, 181 CrossRef CAS .
  48. P. D. Cobden, B. E. Nieuwenhuys, V. V. Gorodetskii and V. N. Parmon, Platinum Met. Rev., 1998, 42, 141 CAS .
  49. P. D. Cobden, V. V. Gorodetskii and B. E. Nieuwenhuys, Surf. Sci., 1999, 432, 61 CrossRef CAS .
  50. V. V. Gorodetskii, A. V. Matveev, P. D. Cobden and B. E. Nieuwenhuys, J. Mol. Catal. A: Chem., 2000, 158, 155 CrossRef CAS .
  51. F. Mittendorfer, N. Seriani, O. Dubay and G. Kresse, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 233413 CrossRef .
  52. T. Visart de Bocarmé, T. D. Chau, F. Tielens, J. Andrés, P. Gaspard, R. L. C. Wang, H. J. Kreuzer and N. Kruse, J. Chem. Phys., 2006, 125, 054703 CrossRef PubMed .
  53. T. D. Chau, T. Visart de Bocarmé and N. Kruse, Catal. Lett., 2004, 98, 85 CrossRef CAS .
  54. P. A. J. Bagot, T. Visart de Bocarmé, A. Cerezo and G. D. W. Smith, Surf. Sci., 2006, 600, 3028 CrossRef CAS .
  55. P. A. J. Bagot, A. Cerezo and G. D. W. Smith, Surf. Sci., 2007, 601, 2245 CrossRef CAS .
  56. P. A. J. Bagot, A. Cerezo and G. D. W. Smith, Surf. Sci., 2008, 602, 1381 CrossRef CAS .
  57. C. Barroo, M. M. Montemore, N. Janvelyan, B. Zugic, A. J. Akey, A. P. Magyar, J. Ye, E. Kaxiras, J. Biener and D. C. Bell, J. Phys. Chem. C, 2017, 121, 5115 CAS .
  58. T. Li, P. A. J. Bagot, E. A. Marquis, S. C. E. Tsang and G. D. W. Smith, J. Phys. Chem. C, 2012, 116, 4760 CAS .
  59. T. Li, P. A. J. Bagot, E. A. Marquis, S. C. E. Tsang and G. D. W. Smith, Ultramicroscopy, 2013, 132, 205 CrossRef CAS PubMed .
  60. L. Hilaire, P. Légaré, Y. Holl and G. Maire, Surf. Sci., 1981, 103, 125 CrossRef CAS .
  61. B. Zugic, L. Wang, C. Heine, D. N. Zakharov, B. A. J. Lechner, E. A. Stach, J. Biener, M. Salmeron, R. J. Madix and C. M. Friend, Nat. Mater., 2017, 16, 558 CrossRef CAS PubMed .
  62. M. L. Personick, B. Zugic, M. M. Biener, J. Biener, R. J. Madix and C. M. Friend, ACS Catal., 2015, 5, 4237 CrossRef CAS .

This journal is © The Royal Society of Chemistry 2017