Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Room-temperature carbon monoxide oxidation by oxygen over Pt/Al2O3 mediated by reactive platinum carbonates

Room-temperature carbon monoxide oxidation by oxygen over Pt/Al2O3 mediated by reactive platinum... ARTICLE Received 3 May 2015 | Accepted 18 Sep 2015 | Published 22 Oct 2015 DOI: 10.1038/ncomms9675 OPEN Room-temperature carbon monoxide oxidation by oxygen over Pt/Al O mediated by reactive 2 3 platinum carbonates 1 2 2 2 2 Mark A. Newton , Davide Ferri , Grigory Smolentsev , Valentina Marchionni & Maarten Nachtegaal Room-temperature carbon monoxide oxidation, important for maintaining clean air among other applications, is challenging even after a century of research into carbon monoxide oxidation. Here we report using time-resolved diffuse reflectance infrared spectroscopy, X-ray absorption fine structure spectroscopy and mass spectrometry a platinum carbonate- mediated mechanism for the room-temperature oxidation of carbon monoxide. By applying a periodic reduction–oxidation mode of operation we further show that this behaviour is reversible and can be formed into a catalytic cycle that requires molecular communication between metallic platinum nanoparticles and highly dispersed oxidic platinum centres. A new possibility for the attainment of low-temperature oxidation of carbon monoxide is therefore demonstrated. 1 2 Department of Physics, University of Warwick, Gibbet Hill Road, Coventry CV4 7AL, UK. Paul Scherrer Institut, CH-5232 Villigen, Switzerland. Correspondence and requests for materials should be addressed to M.A.N. (email: M. Newton.2@warwick.ac.uk). NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 he catalytic oxidation of CO by O by platinum is one of 5-wt% Pt/Al O (Type-94, Johnson Matthey) to 5 vol% CO/Ar 2 2 3 the longest established and industrially important hetero- (50 ml min ). Supplementary Fig. 3 shows equivalent data for a Tgeneous catalytic conversions. The generally accepted second commercial catalyst (2 wt% Pt/g-Al O , Umicore), 2 3 Langmuir–Hinshelwood (LH) mechanism requires the adsorp- whereas Supplementary Fig. 4 compares both catalysts from the tion and reaction of molecular CO with atomic oxygen over perspective of X-ray diffraction, see also Supplementary 1–5 metallic platinum surfaces . At temperatures where adsorbed Discussion. The feed is then switched to 21 vol% O /Ar at the molecular CO becomes stable, the catalytic cycle cannot be same flow rate. This sample had been previously reduced in situ completed as the Pt surfaces become poisoned by adsorbed CO, (5 vol% H /Ar to 573 K) before being returned to an Ar flow which prevents dissociation of O . Catalysis therefore becomes before the subsequent CO/O exposure cycle at 298 K. Figure 1b 2 2 efficient only when the steady-state coverage of CO is diminished shows the CO production resulting from extending the experi- to the degree where dissociation of O can efficiently occur. mental protocol of Fig. 1a to a periodic operation over repeated Practically speaking, this restricts effective CO oxidation using CO/O switches. Alongside the CO production from the 5 wt% 2 2 Pt-Al O catalysts to temperatures 4400 K (Supplementary Fig. 1 Pt/Al O the red curve shows that obtained from an unloaded 2 3 2 3 shows this for a conventional test of catalytic light off using the Al O sample (Condea Puralox (g)). Finally, Fig. 1c shows the 2 3 5 wt% Pt/Al O used in this study). However, Pt remains the cumulative CO production achieved over the duration of the 2 3 2 metal of choice in many applications for effective CO conversion periodic operation of Fig. 1b. to CO where sufficient temperatures are intrinsic or can be easily CO is produced instantaneously on the admission of CO 2 2 applied. before rapidly returning towards baseline levels within the cycle. There are strong drivers, be they economic or regulatory in At the same time (not shown), an exotherm (ca. 3 K) is also nature, to achieve CO oxidation at lower, ideally ambient, transiently observed. On O /Ar admission, CO is again observed 2 2 temperatures. This is especially the case in the purification of fuel to be formed but with some delay from the switching out of CO (H ) feeds, emission control and the maintenance of clean air. in favour of O . The integrals show that in each branch of the 2 2 Given the fundamental limitations of the LH mechanism for redox switch practically the same number of CO molecules are simple Pt catalysts, research to achieve low-temperature CO produced, indicating a quantitative reversibility. This reversibility oxidation has proceeded in other directions: Pt may be favourably is confirmed (Fig. 1b,c) by repeated gas switching and is modified through contact with more exotic oxides, for instance compared with a similar experiment carried out over a Pt-free CeO (refs 6,7) and, most recently, iron–nickel hydroxides ,or Al O (red lines). The total number of CO molecules produced 2 2 3 2 other metals, most notably Au, may be employed to achieve low- over the first cycle shown in Fig. 1a is equivalent to almost 0.2 9–11 temperature CO conversion . CO per Pt atom present in the bed. This implies that in each half Beyond the classic LH mechanism, some attention has recently of the first cycle the active phase comprises only ca. 10% of the Pt. been drawn to reactive Pt surface oxides and carbonates in CO Figure 1b shows that the CO production in the CO cycle actually 12–17 12 conversion . Ackermann et al. determined the presence of improves significantly after the first cycle (Fig. 1a) and is two reactive, single-layer surface oxides formed on Pt(110). The maintained at a higher level thereafter; as such, the 10% estimate incommensurate oxide (essentially a sheet of hexagonal PtO ) has of the active fraction of the Pt represents a lower limit, as been latterly observed during CO oxidation over Al O -supported subsequent cycles might indicate that a range of 10–20% may be 2 3 Pt nanoparticles . The second reactive and commensurate appropriate. Over the extended period of the experiment (Fig. 1c), (1  2) surface oxide observed by Ackermann et al. has yet to the total amount of CO produced is some five times the number be directly observed on high surface area nanosized catalysts. of Pt atoms present in the catalyst bed. Intriguingly, however, calculations suggested that this oxide requires the stabilization of an additional carbonate species. Most recently, Moses-DeBusk et al. have theoretically DRIFTS during periodic CO oxidation operation at 298 K. constructed an entire catalytic cycle that converts CO to CO Figure 2a shows the evolution of infrared bands during the single solely through isolated Pt centres, adsorbed on Al O (010) CO/O cycle of Fig. 1a. Figure 2c shows a colour map of the 2 3 2 surfaces, which requires the involvement of reactive carbonates. 1,200–2,000 cm region of the DRIFTS spectra to emphasize The existence of such Pt carbonates on low loaded (0.18 and the temporal behaviour of infrared-active species in this region. 1 wt%) Pt/Al O catalysts was verified using diffuse reflectance Figure 3a,b shows the temporal variations in a number of 2 3 infrared Fourier transform (DRIFTS), although their intrinsic different infrared-visible bands observed in Fig. 2 and how they reactivity was not addressed. relate to the observed production of CO . As might be expected, Herein, we reveal a route to CO that is mediated through such under CO the DRIFTS spectra are dominated by bands due to 1  1 Pt carbonates and active at ambient temperature. This is achieved linear (2,094 cm ) and bridged CO (1,845 cm ) adsorbed on using a commercially available catalyst system comprising 5 wt% metallic Pt . These appear instantaneously on introduction of Pt (average particle diameter 3 nm, see Supplementary Fig. 5) CO and persist beyond the introduction of O before attenuating supported on a majority y-Al O co-existing with some g-Al O to any significant degree. Alongside these bands a number of 2 3 2 3 phase (see Supplementary Fig. 4). We show, using periodic redox other transient species are also seen, to rapidly evolve and then operation, that this route to CO can be made catalytic, and that attenuate. The largest of these under CO is a sharp feature at phys achieving such reversibility requires a synergy between metallic Pt 2,345 cm due to physisorbed CO (CO ). Concomitantly, a nanoparticles and non-metallic Pt centres that are the precursors strong band at 1,695 cm transiently appears along with less to the carbonates. A new mechanism for low-temperature intense bands at 1,547, 1,402 and 1,330 cm . CO conversion using Pt/Al O -based catalysts is therefore As can be seen in Fig. 2b the bands due to CO adsorbed on 2 3 demonstrated. metallic Pt have little relation to the evolution of gas phase CO (CO ): their temporal character indicates little more than 2(g) adsorption/desorption processes related to the presence in this Results sample of metallic Pt nanoparticles. However, it may be noted phys Periodic CO oxidation of the 5 wt% Pt/Al O catalyst at 298 K. that the lower wavenumber bands, and that due to CO , only 2 3 Figure 1a shows variations in the CO production measured by appear in the oxygen cycle when the removal of linear and mass spectrometry (MS) during room-temperature exposure of a bridged, adsorbed CO species starts to become appreciable. This 2 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. Absorbance (a.u.) CO per Pt atom Time (s) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE a b c 16 19 7 × 10 2.5 × 10 4.0 × 10 0.20 CO O 3.5 × 10 6 ×10 2 × 10 0.15 3.0 × 10 5 × 10 5 wt% Pt/Al O 16 2 3 2.5 × 10 19 16 1.5 × 10 4 × 10 0.10 2.0 × 10 3 × 10 16 1 × 10 1.5 × 10 2 × 10 0.050 1.0 × 10 5 × 10 1 × 10 5.0 × 10 Al O 2 3 0.0 0.0 0 0 0 50 100 150 200 250 0 500 1,000 1,500 2,000 0 500 1,000 1,500 2,000 Time (s) Time (s) Time (s) Figure 1 | Periodic redox operation of the 5 wt% Pt/Al O catalyst at 298 K. (a) Transient evolution of CO observed during exposure of a pre-reduced 2 3 2 Type-94 (Johnson Matthey) 5 wt% Pt/Al O catalyst to 5 vol% CO/Ar (shaded area), followed by a switch to 21 vol% O /He at 298 K. The left-hand axis 2 3 2 reports the evolution of CO in terms of molecules per second (black), the right-hand axis shows the cumulative CO production as a fraction of the total 2 2 number of Pt atoms in the catalyst bed (blue). (b) Repeated cycles of a similar (shorter oxidizing cycle) experiment shown in a: black ¼ 5 wt% Pt/Al O ; 2 3 red ¼ Al O .(c) Cumulative CO (molecules) production during the experiment shown in b. 2 3 2 Pt(CO ) Al(CO ) Pt CO Pt CO a b 0.05 Pt CO phys Pt(CO ) CO 3 2 0.00 c 250 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 CO 0.00 0 50 2,200 2,400 1,600 1,800 2,000 1,400 CO –1 Wavenumber (cm ) 1,200 1,400 1,600 1,800 2,000 –1 Wavenumber (cm ) Figure 2 | DRIFTS during periodic redox operation at 298 K. (a) DRIFTS spectra derived from a first CO/O cycle shown in Fig. 1. (b) The 1  1 0 1,200–2,000 cm region of the DRIFTS shown to emphasize the reactive behaviour of species below ca. 1,750 cm as compared with the bridging Pt CO (ca. 1,850 cm ). (b) Individual absorbance spectra corresponding to the different arrow/lines shown in c. The band positions expected for different 21,22 aluminum carbonates (red), Pt(CO ) (blue) and Pt CO (black) are also given .(c) Colour map representation of the same set of DRIFTS spectra of a. The red arrow shows the changeover point in time between the CO/Ar flow and the O /Ar flow; the solid black line highlights the delay between the switch 0 0  1 to O , the removal of bridging Pt CO species and the transient re-appearance of the bridging Pt CO band at ca. 1,700 cm . 2 2 phys shows that the reserve of ‘inactive’ CO adsorbed on metallic DRIFTS via the CO band at 2,345 cm ) is found for the Pt nanoparticles is the source through which the low-temperature band at ca. 1,695 cm . This principal band is very similar to that active phase of Pt in this system is replenished and can convert calculated by Moses-Debusk et al. for bidentate carbonates CO to CO in the absence of gas phase CO. This can occur as a formed at oxidized and atomic Pt centres, lying midway between 2(g) result of CO desorption being mediated by physisorbed precursor that calculated for a Pt(CO ) species (1,730 cm ) and that 19,20  1 states . At ambient temperature, these have an appreciable observed by experiment (1,659 cm –0.18 wt% Pt and 1 17 lifetime and can therefore search out the active Pt centres required 1,637 cm –1 wt% Pt) . for reaction. The low-temperature catalytic production of CO Although the summation of the evidence derived from MS and observed therefore requires a communication between reduced DRIFTS strongly suggests that a single Pt carbonate species is nanoparticles and other highly dispersed, oxidic Pt centres. responsible for the majority of the room-temperature production The bands at lower wavenumber all show similar, although not of CO , it would be remiss of us not to note that the number of identical, profiles that correlate well with CO production. The bands, however weak, observed to correlate with the CO 2 2 strongest correlation in both halves of the cycle (both in MS and production would indicate the presence of more than a single NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. –1 CO (s ) –1 CO (s ) Abs. (a.u.) Time (s) CO cumulative ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ab 0.14 0.06 CO O CO O 2 2 0.12 0.05 0.1 0.04 Linear CO 0.08 × 0.3 –1 1,695 cm 0.03 0.06 Bridged 0.02 –1 CO 1,402 cm 0.04 phys 0.01 CO 0.02 –1 1,330 cm 0 0 0 50 100 150 200 250 0 50 100 150 200 250 Time (s) Time (s) phys 1  1 Figure 3 | Temporal evolution of adsorbed species at 298 K. (a) Linear CO (blue: 2,094 cm ), bridged CO (red: 1,845 cm ) and CO 1  1 (black: 2,345 cm ). (b) Other bands at 1,695, 1,402 and 1,330 cm as indicated. The shaded area indicates the period of exposure to 5% CO. carbonate species. However, and in addition to the negligible and bridging CO bands in DRIFTS, the Pt in this sample activity of an unloaded alumina support (Fig. 1c, red line), we principally comprises metallic nanoparticles. This is verified for a may exclude the possibility that carbonates formed at the Al O fresh sample by transmission electron microscopy (TEM, 2 3 surface are participating in the chemistry shown in Fig. 2. Supplementary Fig. 5). However, alongside this, and to a first Although such species show infrared absorption in the region of approximation (Fig. 4c), a significant proportion (20–25%) of the 21,22 interest (Fig. 2b), none of them, by way of band position, Pt is present in a Pt (IV) oxidation state that would correspond to combinations of bands and their expected relative intensities , the relatively large, low Z (oxygen) coordination detected by the explains our observations. For instance, Al bicarbonates EXAFS (Fig. 4a,b and and Table 1). All or a portion of this 1 21,22 (1,655 cm ) and bidentate carbonates (1,660– oxidized Pt may correspond to the smaller Pt entities unambigu- 1 21  1 1,730 cm ) always show bands in the 1,200–1,300 cm ously detected by high-angle annular dark-field scanning TEM region that should be detectable if they were contributing analysis (Supplementary Fig. 6). Moreover, the spatial proximity 21,22 significantly to the chemistry . Similarly, Al monodentate of these species to the reduced Pt nanoparticles makes the carbonates show spectral features only between 1,400 and required molecular communication between these two types of Pt 1,650 cm (refs 21,22). In this region, however, only the weak plausible. band at 1,402 cm shows evidence for a diminution after its Figure 4d shows the behaviour of the Pt during room- formation (although, by and large, it persists) and one that is very temperature exposure to CO and then 21 vol% O /Ar, from the much slower than the majority CO production. As such, it perspective of the height of the Pt L edge white line. This reveals 2 3 cannot be deemed responsible for the majority CO turnover we that the reduction of the Pt—corresponding with the onset of observe that is far more highly correlated to the much stronger CO production in the CO cycle—is extremely rapid and band at 1,695 cm . essentially complete in o10 s. On returning the sample to the Therefore, the above demonstrates that Pt carbonates are oxidizing flow we clearly see that re-oxidation of the active Pt is intrinsically capable of forming rapidly under CO and converting much slower and subject to an induction time that matches well that CO to CO at 298 K. It also shows that re-oxidation of the Pt with the observations made using DRIFTS. From this we may sites responsible for this chemistry is much slower and acts to conclude that re-oxidation of the carbonate precursor requires limit the overall efficiency of this process. dissociation of O at the metallic sites, and that this can only occur as and when the linear and bridging CO species start to desorb. XAFS during periodic CO oxidation on 5 wt% Pt/Al O at 298 K. 2 3 DRIFTS and MS analyses tell us little about the Pt itself Discussion during these events. Importantly, these measurements cannot The combined (DRIFTS, MS and time-resolved XAFS) evidence discriminate between a direct re-oxidation of isolated Pt centres obtained allows us to derive a room-temperature cyclic mechan- that is subject to a relatively high activation energy—as ism, involving both Pt carbonates and Pt nanoparticles for the CO considered by Moses-DeBusk et al. —or whether, and as oxidation under the periodic conditions of operation employed. might be implied by the DRIFTS, desorption of molecular CO Under CO: from metallic Pt nanoparticles has to occur first. To resolve this 0 0 0 Pt þ CO ! PtðÞ CO þ PtðÞ CO ð1Þ issue we therefore conducted static and time-resolved X-ray L 2 B absorption fine structure (XAFS) spectroscopy at the Pt L edge IV under identical experimental conditions. Some results of this are Pt ðÞ O þ CO ! PtðÞ CO ð2Þ shown in Fig. 4. phys Figure 4a shows the k-weighted extended XAFS (EXAFS) PtðÞ CO ! PtðÞ O þ CO ! PtðÞ O þ CO ð3Þ 3 2gðÞ obtained from the 5 wt% Pt/Al O catalyst, whereas Fig. 4b shows 2 3 the corresponding Fourier transform. The results of best fitting Under O : the EXAFS spectrum are given in Table 1. These measurements 0 0 phys 0 PtðÞ CO þ PtðÞ CO ! 2CO þ 2Pt ð4Þ show that, as might be deduced from the dominance of linear L 2 B 4 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. Absorbance NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE a b 2468 10 12 14 16 01234567 –1 k (Å ) R (Å) c d Blue = PtO 2 CO O Red = Pt foil Oxidation 75%Pt + 25%PtO reduction 5 wt% Pt/Al O 2 3 80%Pt + 20%PtO 11.54 11.57 11.6 11.64 060 120 180 240 300 Energy (keV) Time (s) Figure 4 | XAFS during periodic redox operation at 298 K. (a) k -weighted EXAFS derived in situ from the 5 wt% Pt/Al O catalyst along with theoretical 2 3 fit (red); (b) the corresponding Fourier transform (again with the theoretical fit in red). (c) Comparison of Pt L edge X-ray absorption near edge structure 0 0 from this sample with that from Pt foil (red) and PtO (blue). The sample spectrum is also compared with linear combinations (80% Pt and 75% Pt as indicated) of these two reference spectra. (d) The temporal variation observed in the Pt L white line intensity during exposure of the catalyst to 5 vol% CO/Ar and then 21 vol% O /Ar. 0 0 where Pt is the metallic platinum, Pt (CO) is the CO adsorbed Table 1 | Best-fit parameters obtained for the 5 wt% 0 in linear geometry on metallic Pt, PtðÞ CO is the CO adsorbed in 2 B Pt/Al O sample measured in situ at room temperature and 2 3 twofold bridge geometry on metallic Pt, Pt(CO ) is the platinum after reduction at 573 K. IV carbonate intermediate, Pt (O) is the isolated oxidic platinum species, O is the adsorbed oxygen, Pt(O) is the product of * w 2  2 z phys phys Element CN R (Å) DW (2r /(Å ) carbonate decomposition, CO and CO are physisorbed Pt 8 2.75 0.014 CO and CO, respectively, and O ,CO and CO are gas 2 2(g) 2(g) (g) O 1.9 2.00 0.007 phase species. Within this mechanism, the oxidative regeneration of the EXAFS, extended X-ray absorption fine structure. E ¼ 10.31¼ the edge position relative to the vacuum zero (Fermi energy). precursor to carbonate formation, equation (6), limits the T E 3 E 3 T E R%¼ 33.37 ¼ ( [w  w ]k dk/[w ]k dk) 100%, where w and w are the theoretical and reformation of the carbonates in oxidizing conditions. However, experimental EXAFS and k is the photoelectron wave vector. The Debye–Waller factor¼ 2s , where s is the root mean square internuclear separation. the combined evidence suggests that this is mediated via the Pt Other parameters: Attentuation factor (AFAC), related to the proportion of electrons performing nanoparticles and therefore can only happen once free sites are an EXAFS-type scatter on absorption, is 0.875. Structural data were obtained by fitting the EXAFS in k space in the range: k ¼ 2.5–16.5 Å . created on the Pt nanoparticles rather than through a direct *Coordination number ( ca. 10% stated value). dissociation of O at the precursor sites to the carbonates. As such, wDistance of scatterer atom from central atom ( ca. 1.5% stated value). zDebye–Waller factor. we suggest that it is CO desorption (equation (4)) that effectively limits the efficacy of this reactive pathway. In this respect, our mechanism differs from that of Moses-DeBusk et al. who considered only a direct dissociation of O by atomically dispersed Pt precursors to the carbonate species. As a result, and in respect of steady-state operation, our room- 0 0 Pt þ O ! Pt þ 2O ð5Þ 2gðÞ a temperature mechanism still suffers from the poisoning effect of CO that hobbles the classic LH mechanism. Indeed, steady-state IV PtðÞ O þ O ! Pt ðÞ O ð6Þ operation (Supplementary Fig. 2) under a 4O :CO flow leads to a very small, but non-zero, production of CO at room tempera- phys 2 CO ! CO ð7Þ ðÞ g ture; at present, it is only through adopting a periodic operation that this reactive impasse may be, to some degree, circumvented. phys IV CO þ Pt ðÞ O ! PtðÞ CO ð8Þ A priori we cannot definitively say where these carbonates actually reside. It is clear from our experimental methodology phys PtðÞ CO ! PtðÞ O þ CO ! PtðÞ O þ CO ð9Þ 3 2gðÞ 2 that the oxidized Pt entities that lead to their formation can survive reduction in H /Ar to at least 573 K. This IV PtðÞ O þ O ! Pt ðÞ O ð10Þ 2 characteristic mitigates against their existence within a NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. Normalized XAFS f(k)χ(k ) XANES intensity FT intensity ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 initiated (TB423 K). At each temperature, the DRIFTS and MS data were commensurate and reactive surface oxide in the manner proposed 12 collected, with the latter being used to calibrate the CO response in terms of the by Ackermann et al. Instead, in this respect we would tend to overall CO conversions. concur with the propositions of Moses-DeBusk et al. who only Having conducted this calibration experiment, the sample was re-cooled under considered atomically dispersed oxidic Pt centres adsorbed on the reaction mixture, purged again before a second CO/O switching experiment was conducted and again DRIFTS and MS data simultaneously acquired. It is the alumina surfaces. Further, high-angle annular dark-field scanning results of this second switching experiment that are reported in Figs 1–3. We note TEM measurements (Supplementary Fig. 6) also support this that both first and second switching experiments, as well as the 16 cycle view and are consistent with recent works that have shown that experiment, return the same global results save for some minor differences in similar atomic or quasi atomically adsorbed Pt and Pd species can overall CO production. In addition, although not shown here, essentially the 23,24 same results can be achieved without any pre-reduction of the catalyst and the be highly reactive for oxidation reactions . Although we species responsible for the chemistry we have reported are present in the as- cannot rule out a role for interfacial Pt-O-Al sites, these would received catalyst. not be sufficient (based on the particle size distribution from TEM, see Supplementary Fig. 5) to yield the ca. 10–20% of active X-ray absorption spectroscopy. Pt L edge XAFS was collected in transmission Pt that our results indicate are mediating carbonate formation mode at the SuperXAS beamline at the Swiss Light source (Villigen, Switzerland) and CO turnover in the current case. using a newly installed fast, Si (111) channel cut monochromator system coupled to To summarize, we have demonstrated that room-temperature gridded N -filled ionization chambers for detection. This bidirectional scanning CO oxidation over Pt/Al O is feasible and mediated by oxidized, 2 3 system was operated at 2 Hz, yielding four spectra per second. The static and isolated and reduction-resistant Pt centres that may form reactive reference spectra are obtained as averages over 3 min of acquisition. The in-situ time-resolved X-ray absorption near edge structure data were carbonates from CO and be re-oxidized under O . Using periodic extracted from individual (250 ms time resolution) spectra during CO switching operation we have shown that this process is quantitatively from 5 vol% CO/Ar and 21 vol% O /Ar, and using the same sample environment as reversible and can form a catalytic cycle. At present, the reactive for the DRIFT/MS. Online MS was recorded as for the DRIFTS-based Pt centres necessary for this low-temperature conversion exist as measurements. Data reduction was made using PAXAS and analysis of the EXAFS using EXCURV . a minority species within the conventional catalyst used for these studies: most of the Pt (estimated to be ca. Z80%) is not directly active but does have a role to play in completing the catalytic References cycle. We have shown that at room temperature metallic 1. Langmuir, I. The mechanism of the catalytic action of platinum in the reactions nanoparticulate Pt may act as a reservoir for CO, which may be of 2CO þ O ¼ CO and 2H þ O ¼ H O. Trans. Faraday Soc. 17, 0621–0654 2 2 2 2 2 (1921). transferred to the minority quasi atomic oxidic Pt phase, and 2. Campbell, C. T., Ertl, G., Kuipers, H. & Segner, J. A molecular beam study of converted to CO . The rate-limiting re-oxidation of the carbonate the catalytic-oxidation of CO on a Pt(111) surface. J. Chem. Phys. 73, precursor sites (Pt(O)) also requires the presence of Pt 5862–5873 (1980). nanoparticles that can facilitate O dissociation and oxygen 3. Wintterlin, J., Volkening, J., Janssens, T. V. W., Zambelli, T. & Ertl, G. Atomic transfer to these sites. and macroscopic reaction rates of a surface-catalyzed reaction. Science 278, 1931–1934 (1997). As such, this new pathway for achieving low-temperature CO 4. Alavi, A., Hu, P., Deutsch, T., Silvestri, P. & Hutter, P. L. CO oxidation oxidation immediately poses a number of further challenges to on Pt(111): an ab initio density functional theory study. Phys. Rev. Lett. 80, material synthesis and fundamental understanding. Foremost 3650–3653 (1998). among these is the fundamental question of whether a 5. Allian, A. D. et al. Chemisorption of CO and mechanism of CO oxidation on direct oxidation of the precursor to the Pt carbonates can be supported platinum nanoclusters. J. Am. Chem. Soc. 133, 4498–4517 (2011). 6. Trovarelli, A. Catalysis by Ceria and Related Materials (Imperial College Press, attained and how we may promote CO desorption from the 2002). Pt nanoparticles to permit more efficient O dissociation. Second 7. Vayssilov, G. N. et al. Support nanostructure boosts oxygen transfer to would be to understand, from a material synthesis perspective, catalytically active platinum nanoparticles. Nat. Mater. 10, 310–315 (2011). how to optimize the levels of such carbonate forming Pt species 8. Chen, G. et al. Interfacial effects in iron-nickel hydroxide–platinum on Al O from ca. 10–20% minority to a working majority. nanoparticles enhance catalytic oxidation. Science 344, 495–499 (2014). 2 3 9. Haruta, M. Size- and support-dependency in the catalysis of gold. Catal. Today 36, 153–166 (1997). Methods 10. Bond, G. C. & Thompson, D. T. Catalysis by gold. Catal. Rev. Sci. Eng. 41, Infrared spectroscopy. Combined DRIFTS/MS experiments were carried at the 319–388 (1999). Swiss Light Source using a Bruker Vertex 80V infrared spectrometer fitted with a 11. Hashmi, A. S. K. & Hutchings, G. J. Gold catalysis. Angew. Chem. Int. Ed. 45, narrow-band high-sensitivity mercury cadmium telluride (MCT) detector and with 7896–7936 (2006). a Pfeiffer mass spectrometer. DRIFTS spectra were collected with a time resolution 12. Ackermann, M. D. et al. Structure and reactivity of surface oxides on Pt(110) of 0.433 s and 2 cm resolution. during catalytic CO oxidation. Phys. Rev. Lett. 95, 255505 (2005). The in-situ cell used for these experiments was that recently described by 13. Newton, M. A., Chapman, K. W., Thompsett, D. & Chupas, P. J. Chasing Chiarello et al. for transient experimentation and was windowed with CaF . This changing nanoparticles: a time resolved total X-ray scattering study of the cell was connected to a gas handling system equipped with fast solenoid switching behavior of Supported Pt nanoparticles during CO oxidation catalysis. J. Am. valves (Parker) controlled using the infrared spectrometer. The reactor exit was Chem. Soc. 134, 5036–5039 (2012). connected to a Pfeiffer mass spectrometer that recorded both the sample 14. Butcher, D. R. et al. In situ oxidation study of Pt(110) and its interaction with temperature and the gas switching events alongside a range of masses compatible CO. J. Am. Chem. Soc. 133, 20319–20325 (2011). with the experiments being made. 15. Miller, D. et al. Different reactivity of the various platinum oxides and Gas flows (50 ml min ) were controlled by Bronkhorst mass flow controllers chemisorbed oxygen in CO oxidation on Pt(111). J. Am. Chem. Soc. 136, with a 2-bar pressure of gases behind the controllers themselves. This set-up 6340–6347 (2014). allowed the purging of the system with Ar, then reduction using 5 vol% H /Ar and, 16. Szlachetko, J. et al. Sub-second and in situ chemical speciation of Pt/Al O 2 3 finally, gas switching experiments between 5 vol%CO/Ar and 21 vol%O /Ar. during oxidation reduction cycles monitored by high-energy resolution off- In the experiments described, 25–30 mg of the 5 wt% Pt/Al O catalyst (Type- 2 3 resonant X-ray spectroscopy. J. Am. Chem. Soc. 135, 19071–19074 (2013). 94, Johnson Matthey) was loaded into the DRIFTS cell. After purging in Ar, this 17. Moses-DeBusk, M. et al. CO oxidation on supported single Pt atoms: sample was then reduced under 5 vol% H /Ar to 573 K under a linear heating ramp experimental and ab initio density functional studies of CO interaction with Pt (10 K min ) and then held at 573 K for 30 min. The sample was then cooled atom on theta-Al O (010) surface. J. Am. Chem. Soc. 135, 12634–12645 (2013). under 5 vol% H /Ar to 323 K whereon the reducing flow was substituted for 2 3 18. Hollins, P. The influence of surface-defects on the infra-red spectra of adsorbed flowing Ar once more. At 298 K, the sample was then exposed to 5 vol% CO/Ar for species. Surf. Sci. Rep. 16, 51–94 (1992). 52 s before the flow is switched to 21 vol% O /He for 208 s, whereas DRIFTS and 19. Kisliuk, P. The sticking probabilities of gases chemisorbed on the surfaces of MS data were collected. This experiment was then repeated for a total of 16 CO/O solids. J. Phys. Chem. Solids 3, 95–101 (1957). cycles. Subsequently, the sample was then exposed to a 4O :1CO reaction mixture 20. Bowker, M., Bowker, L. J., Bennett, R. A., Stone, P. & Ramirez-Cuesta, A. In 1 17 (again 50 ml min total flow, 4.2  10 molecules CO per second) and the consideration of precursor states, spillover and Boudart’s ‘collection zone’ and temperature incremented until light-off of conventional CO oxidation catalysis was of their role in catalytic processes. J. Mol. Cat. A 163, 221–232 (2000). 6 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE Financial support from the Swiss National Science Foundation (SNF, project number 21. Busca, G. & Lorenzelli, V. Infrared spectroscopic identification of species 200021–138068) is greatly acknowledged. arising from reactive adsorption of carbon oxides on metal oxide surfaces. Mater. Chem. 7, 89–126 (1982). 22. Liu, X., Korotkikh, O. & Farrauto, R. Selective catalytic oxidation of CO in H : Author contributions structural study of Fe oxide-promoted Pt/alumina catalyst. Appl. Catal. A 226, Experiments were conceived and carried out by M.A. Newton with experimental 293–303 (2002). assistance from V. Marchionni and D. Ferri (DRIFTS/MS) and M. Nachtegaal and 23. Narula, C. K., Allard, L. F., Stocks, G. M. & Moses-DeBusk, M. Remarkable NO G. Smolentsev (XAFS). M.A. Newton wrote the manuscript with D. Ferri and oxidation on single supported platinum atoms. Sci. Rep. 4, 7238 (2014). M. Nachtegaal. All authors have proofread the manuscript. 24. Peterson, E. J. et al. Low temperature carbon monoxide oxidation catalyzed by regenerable atomically dispersed Palladium on Alumina. Nat. Commun. 5, 4885 (2014). Additional information 25. Chiarello, G. L. et al. Adding diffuse reflectance infrared Fourier transform Supplementary Information accompanies this paper at http://www.nature.com/ spectroscopy capability to extended X-ray absorption fine structure in a new naturecommunications cell to study solid catalysts in combination with a modulation approach. Rev. Sci. Instr. 85, 074102 (2014). Competing financial interests: The authors declare no competing financial interests. 26. Binsted, N. PAXAS: Programme for the Analysis of X-ray Adsorption Spectra Reprints and permission information is available online at http://npg.nature.com/ (University of Southampton, 1988). reprintsandpermissions/ 27. Binsted, N. EXCURV98 (CCLRC Daresbury Laboratory Computer Programme, 1998). How to cite this article: Newton, M. A. et al. Room-temperature carbon monoxide oxidation by oxygen over Pt/Al O mediated by reactive platinum carbonates. 2 3 Nat. Commun. 6:8675 doi: 10.1038/ncomms9675 (2015). Acknowledgements M.A.N. thanks the Swiss Light Source/Paul Scherrer Institut and the Department of This work is licensed under a Creative Commons Attribution 4.0 Physics at the University of Warwick, for the visiting scientist positions that facilitated International License. The images or other third party material in this this work. Johnson Matthey (Dr D. Thompsett) and Umicore are also thanked for article are included in the article’s Creative Commons license, unless indicated otherwise provision of the samples used in these studies. We thank Frank Krumeich and Maksym in the credit line; if the material is not included under the Creative Commons license, Kovalenko (ETH Zu¨rich) for enabling high-angle annular dark-field scanning TEM. Electron microscopy was carried out at the Scientific Center for Optical and Electron users will need to obtain permission from the license holder to reproduce the material. Microscopy (ETH Zurich). We are thankful to Lorenz Ba¨ni for technical support. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nature Communications Springer Journals

Room-temperature carbon monoxide oxidation by oxygen over Pt/Al2O3 mediated by reactive platinum carbonates

Loading next page...
 
/lp/springer-journals/room-temperature-carbon-monoxide-oxidation-by-oxygen-over-pt-al2o3-ISFIF3c3kv

References (59)

Publisher
Springer Journals
Copyright
Copyright © 2015 by The Author(s)
Subject
Science, Humanities and Social Sciences, multidisciplinary; Science, Humanities and Social Sciences, multidisciplinary; Science, multidisciplinary
eISSN
2041-1723
DOI
10.1038/ncomms9675
Publisher site
See Article on Publisher Site

Abstract

ARTICLE Received 3 May 2015 | Accepted 18 Sep 2015 | Published 22 Oct 2015 DOI: 10.1038/ncomms9675 OPEN Room-temperature carbon monoxide oxidation by oxygen over Pt/Al O mediated by reactive 2 3 platinum carbonates 1 2 2 2 2 Mark A. Newton , Davide Ferri , Grigory Smolentsev , Valentina Marchionni & Maarten Nachtegaal Room-temperature carbon monoxide oxidation, important for maintaining clean air among other applications, is challenging even after a century of research into carbon monoxide oxidation. Here we report using time-resolved diffuse reflectance infrared spectroscopy, X-ray absorption fine structure spectroscopy and mass spectrometry a platinum carbonate- mediated mechanism for the room-temperature oxidation of carbon monoxide. By applying a periodic reduction–oxidation mode of operation we further show that this behaviour is reversible and can be formed into a catalytic cycle that requires molecular communication between metallic platinum nanoparticles and highly dispersed oxidic platinum centres. A new possibility for the attainment of low-temperature oxidation of carbon monoxide is therefore demonstrated. 1 2 Department of Physics, University of Warwick, Gibbet Hill Road, Coventry CV4 7AL, UK. Paul Scherrer Institut, CH-5232 Villigen, Switzerland. Correspondence and requests for materials should be addressed to M.A.N. (email: M. Newton.2@warwick.ac.uk). NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 he catalytic oxidation of CO by O by platinum is one of 5-wt% Pt/Al O (Type-94, Johnson Matthey) to 5 vol% CO/Ar 2 2 3 the longest established and industrially important hetero- (50 ml min ). Supplementary Fig. 3 shows equivalent data for a Tgeneous catalytic conversions. The generally accepted second commercial catalyst (2 wt% Pt/g-Al O , Umicore), 2 3 Langmuir–Hinshelwood (LH) mechanism requires the adsorp- whereas Supplementary Fig. 4 compares both catalysts from the tion and reaction of molecular CO with atomic oxygen over perspective of X-ray diffraction, see also Supplementary 1–5 metallic platinum surfaces . At temperatures where adsorbed Discussion. The feed is then switched to 21 vol% O /Ar at the molecular CO becomes stable, the catalytic cycle cannot be same flow rate. This sample had been previously reduced in situ completed as the Pt surfaces become poisoned by adsorbed CO, (5 vol% H /Ar to 573 K) before being returned to an Ar flow which prevents dissociation of O . Catalysis therefore becomes before the subsequent CO/O exposure cycle at 298 K. Figure 1b 2 2 efficient only when the steady-state coverage of CO is diminished shows the CO production resulting from extending the experi- to the degree where dissociation of O can efficiently occur. mental protocol of Fig. 1a to a periodic operation over repeated Practically speaking, this restricts effective CO oxidation using CO/O switches. Alongside the CO production from the 5 wt% 2 2 Pt-Al O catalysts to temperatures 4400 K (Supplementary Fig. 1 Pt/Al O the red curve shows that obtained from an unloaded 2 3 2 3 shows this for a conventional test of catalytic light off using the Al O sample (Condea Puralox (g)). Finally, Fig. 1c shows the 2 3 5 wt% Pt/Al O used in this study). However, Pt remains the cumulative CO production achieved over the duration of the 2 3 2 metal of choice in many applications for effective CO conversion periodic operation of Fig. 1b. to CO where sufficient temperatures are intrinsic or can be easily CO is produced instantaneously on the admission of CO 2 2 applied. before rapidly returning towards baseline levels within the cycle. There are strong drivers, be they economic or regulatory in At the same time (not shown), an exotherm (ca. 3 K) is also nature, to achieve CO oxidation at lower, ideally ambient, transiently observed. On O /Ar admission, CO is again observed 2 2 temperatures. This is especially the case in the purification of fuel to be formed but with some delay from the switching out of CO (H ) feeds, emission control and the maintenance of clean air. in favour of O . The integrals show that in each branch of the 2 2 Given the fundamental limitations of the LH mechanism for redox switch practically the same number of CO molecules are simple Pt catalysts, research to achieve low-temperature CO produced, indicating a quantitative reversibility. This reversibility oxidation has proceeded in other directions: Pt may be favourably is confirmed (Fig. 1b,c) by repeated gas switching and is modified through contact with more exotic oxides, for instance compared with a similar experiment carried out over a Pt-free CeO (refs 6,7) and, most recently, iron–nickel hydroxides ,or Al O (red lines). The total number of CO molecules produced 2 2 3 2 other metals, most notably Au, may be employed to achieve low- over the first cycle shown in Fig. 1a is equivalent to almost 0.2 9–11 temperature CO conversion . CO per Pt atom present in the bed. This implies that in each half Beyond the classic LH mechanism, some attention has recently of the first cycle the active phase comprises only ca. 10% of the Pt. been drawn to reactive Pt surface oxides and carbonates in CO Figure 1b shows that the CO production in the CO cycle actually 12–17 12 conversion . Ackermann et al. determined the presence of improves significantly after the first cycle (Fig. 1a) and is two reactive, single-layer surface oxides formed on Pt(110). The maintained at a higher level thereafter; as such, the 10% estimate incommensurate oxide (essentially a sheet of hexagonal PtO ) has of the active fraction of the Pt represents a lower limit, as been latterly observed during CO oxidation over Al O -supported subsequent cycles might indicate that a range of 10–20% may be 2 3 Pt nanoparticles . The second reactive and commensurate appropriate. Over the extended period of the experiment (Fig. 1c), (1  2) surface oxide observed by Ackermann et al. has yet to the total amount of CO produced is some five times the number be directly observed on high surface area nanosized catalysts. of Pt atoms present in the catalyst bed. Intriguingly, however, calculations suggested that this oxide requires the stabilization of an additional carbonate species. Most recently, Moses-DeBusk et al. have theoretically DRIFTS during periodic CO oxidation operation at 298 K. constructed an entire catalytic cycle that converts CO to CO Figure 2a shows the evolution of infrared bands during the single solely through isolated Pt centres, adsorbed on Al O (010) CO/O cycle of Fig. 1a. Figure 2c shows a colour map of the 2 3 2 surfaces, which requires the involvement of reactive carbonates. 1,200–2,000 cm region of the DRIFTS spectra to emphasize The existence of such Pt carbonates on low loaded (0.18 and the temporal behaviour of infrared-active species in this region. 1 wt%) Pt/Al O catalysts was verified using diffuse reflectance Figure 3a,b shows the temporal variations in a number of 2 3 infrared Fourier transform (DRIFTS), although their intrinsic different infrared-visible bands observed in Fig. 2 and how they reactivity was not addressed. relate to the observed production of CO . As might be expected, Herein, we reveal a route to CO that is mediated through such under CO the DRIFTS spectra are dominated by bands due to 1  1 Pt carbonates and active at ambient temperature. This is achieved linear (2,094 cm ) and bridged CO (1,845 cm ) adsorbed on using a commercially available catalyst system comprising 5 wt% metallic Pt . These appear instantaneously on introduction of Pt (average particle diameter 3 nm, see Supplementary Fig. 5) CO and persist beyond the introduction of O before attenuating supported on a majority y-Al O co-existing with some g-Al O to any significant degree. Alongside these bands a number of 2 3 2 3 phase (see Supplementary Fig. 4). We show, using periodic redox other transient species are also seen, to rapidly evolve and then operation, that this route to CO can be made catalytic, and that attenuate. The largest of these under CO is a sharp feature at phys achieving such reversibility requires a synergy between metallic Pt 2,345 cm due to physisorbed CO (CO ). Concomitantly, a nanoparticles and non-metallic Pt centres that are the precursors strong band at 1,695 cm transiently appears along with less to the carbonates. A new mechanism for low-temperature intense bands at 1,547, 1,402 and 1,330 cm . CO conversion using Pt/Al O -based catalysts is therefore As can be seen in Fig. 2b the bands due to CO adsorbed on 2 3 demonstrated. metallic Pt have little relation to the evolution of gas phase CO (CO ): their temporal character indicates little more than 2(g) adsorption/desorption processes related to the presence in this Results sample of metallic Pt nanoparticles. However, it may be noted phys Periodic CO oxidation of the 5 wt% Pt/Al O catalyst at 298 K. that the lower wavenumber bands, and that due to CO , only 2 3 Figure 1a shows variations in the CO production measured by appear in the oxygen cycle when the removal of linear and mass spectrometry (MS) during room-temperature exposure of a bridged, adsorbed CO species starts to become appreciable. This 2 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. Absorbance (a.u.) CO per Pt atom Time (s) NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE a b c 16 19 7 × 10 2.5 × 10 4.0 × 10 0.20 CO O 3.5 × 10 6 ×10 2 × 10 0.15 3.0 × 10 5 × 10 5 wt% Pt/Al O 16 2 3 2.5 × 10 19 16 1.5 × 10 4 × 10 0.10 2.0 × 10 3 × 10 16 1 × 10 1.5 × 10 2 × 10 0.050 1.0 × 10 5 × 10 1 × 10 5.0 × 10 Al O 2 3 0.0 0.0 0 0 0 50 100 150 200 250 0 500 1,000 1,500 2,000 0 500 1,000 1,500 2,000 Time (s) Time (s) Time (s) Figure 1 | Periodic redox operation of the 5 wt% Pt/Al O catalyst at 298 K. (a) Transient evolution of CO observed during exposure of a pre-reduced 2 3 2 Type-94 (Johnson Matthey) 5 wt% Pt/Al O catalyst to 5 vol% CO/Ar (shaded area), followed by a switch to 21 vol% O /He at 298 K. The left-hand axis 2 3 2 reports the evolution of CO in terms of molecules per second (black), the right-hand axis shows the cumulative CO production as a fraction of the total 2 2 number of Pt atoms in the catalyst bed (blue). (b) Repeated cycles of a similar (shorter oxidizing cycle) experiment shown in a: black ¼ 5 wt% Pt/Al O ; 2 3 red ¼ Al O .(c) Cumulative CO (molecules) production during the experiment shown in b. 2 3 2 Pt(CO ) Al(CO ) Pt CO Pt CO a b 0.05 Pt CO phys Pt(CO ) CO 3 2 0.00 c 250 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 CO 0.00 0 50 2,200 2,400 1,600 1,800 2,000 1,400 CO –1 Wavenumber (cm ) 1,200 1,400 1,600 1,800 2,000 –1 Wavenumber (cm ) Figure 2 | DRIFTS during periodic redox operation at 298 K. (a) DRIFTS spectra derived from a first CO/O cycle shown in Fig. 1. (b) The 1  1 0 1,200–2,000 cm region of the DRIFTS shown to emphasize the reactive behaviour of species below ca. 1,750 cm as compared with the bridging Pt CO (ca. 1,850 cm ). (b) Individual absorbance spectra corresponding to the different arrow/lines shown in c. The band positions expected for different 21,22 aluminum carbonates (red), Pt(CO ) (blue) and Pt CO (black) are also given .(c) Colour map representation of the same set of DRIFTS spectra of a. The red arrow shows the changeover point in time between the CO/Ar flow and the O /Ar flow; the solid black line highlights the delay between the switch 0 0  1 to O , the removal of bridging Pt CO species and the transient re-appearance of the bridging Pt CO band at ca. 1,700 cm . 2 2 phys shows that the reserve of ‘inactive’ CO adsorbed on metallic DRIFTS via the CO band at 2,345 cm ) is found for the Pt nanoparticles is the source through which the low-temperature band at ca. 1,695 cm . This principal band is very similar to that active phase of Pt in this system is replenished and can convert calculated by Moses-Debusk et al. for bidentate carbonates CO to CO in the absence of gas phase CO. This can occur as a formed at oxidized and atomic Pt centres, lying midway between 2(g) result of CO desorption being mediated by physisorbed precursor that calculated for a Pt(CO ) species (1,730 cm ) and that 19,20  1 states . At ambient temperature, these have an appreciable observed by experiment (1,659 cm –0.18 wt% Pt and 1 17 lifetime and can therefore search out the active Pt centres required 1,637 cm –1 wt% Pt) . for reaction. The low-temperature catalytic production of CO Although the summation of the evidence derived from MS and observed therefore requires a communication between reduced DRIFTS strongly suggests that a single Pt carbonate species is nanoparticles and other highly dispersed, oxidic Pt centres. responsible for the majority of the room-temperature production The bands at lower wavenumber all show similar, although not of CO , it would be remiss of us not to note that the number of identical, profiles that correlate well with CO production. The bands, however weak, observed to correlate with the CO 2 2 strongest correlation in both halves of the cycle (both in MS and production would indicate the presence of more than a single NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. –1 CO (s ) –1 CO (s ) Abs. (a.u.) Time (s) CO cumulative ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ab 0.14 0.06 CO O CO O 2 2 0.12 0.05 0.1 0.04 Linear CO 0.08 × 0.3 –1 1,695 cm 0.03 0.06 Bridged 0.02 –1 CO 1,402 cm 0.04 phys 0.01 CO 0.02 –1 1,330 cm 0 0 0 50 100 150 200 250 0 50 100 150 200 250 Time (s) Time (s) phys 1  1 Figure 3 | Temporal evolution of adsorbed species at 298 K. (a) Linear CO (blue: 2,094 cm ), bridged CO (red: 1,845 cm ) and CO 1  1 (black: 2,345 cm ). (b) Other bands at 1,695, 1,402 and 1,330 cm as indicated. The shaded area indicates the period of exposure to 5% CO. carbonate species. However, and in addition to the negligible and bridging CO bands in DRIFTS, the Pt in this sample activity of an unloaded alumina support (Fig. 1c, red line), we principally comprises metallic nanoparticles. This is verified for a may exclude the possibility that carbonates formed at the Al O fresh sample by transmission electron microscopy (TEM, 2 3 surface are participating in the chemistry shown in Fig. 2. Supplementary Fig. 5). However, alongside this, and to a first Although such species show infrared absorption in the region of approximation (Fig. 4c), a significant proportion (20–25%) of the 21,22 interest (Fig. 2b), none of them, by way of band position, Pt is present in a Pt (IV) oxidation state that would correspond to combinations of bands and their expected relative intensities , the relatively large, low Z (oxygen) coordination detected by the explains our observations. For instance, Al bicarbonates EXAFS (Fig. 4a,b and and Table 1). All or a portion of this 1 21,22 (1,655 cm ) and bidentate carbonates (1,660– oxidized Pt may correspond to the smaller Pt entities unambigu- 1 21  1 1,730 cm ) always show bands in the 1,200–1,300 cm ously detected by high-angle annular dark-field scanning TEM region that should be detectable if they were contributing analysis (Supplementary Fig. 6). Moreover, the spatial proximity 21,22 significantly to the chemistry . Similarly, Al monodentate of these species to the reduced Pt nanoparticles makes the carbonates show spectral features only between 1,400 and required molecular communication between these two types of Pt 1,650 cm (refs 21,22). In this region, however, only the weak plausible. band at 1,402 cm shows evidence for a diminution after its Figure 4d shows the behaviour of the Pt during room- formation (although, by and large, it persists) and one that is very temperature exposure to CO and then 21 vol% O /Ar, from the much slower than the majority CO production. As such, it perspective of the height of the Pt L edge white line. This reveals 2 3 cannot be deemed responsible for the majority CO turnover we that the reduction of the Pt—corresponding with the onset of observe that is far more highly correlated to the much stronger CO production in the CO cycle—is extremely rapid and band at 1,695 cm . essentially complete in o10 s. On returning the sample to the Therefore, the above demonstrates that Pt carbonates are oxidizing flow we clearly see that re-oxidation of the active Pt is intrinsically capable of forming rapidly under CO and converting much slower and subject to an induction time that matches well that CO to CO at 298 K. It also shows that re-oxidation of the Pt with the observations made using DRIFTS. From this we may sites responsible for this chemistry is much slower and acts to conclude that re-oxidation of the carbonate precursor requires limit the overall efficiency of this process. dissociation of O at the metallic sites, and that this can only occur as and when the linear and bridging CO species start to desorb. XAFS during periodic CO oxidation on 5 wt% Pt/Al O at 298 K. 2 3 DRIFTS and MS analyses tell us little about the Pt itself Discussion during these events. Importantly, these measurements cannot The combined (DRIFTS, MS and time-resolved XAFS) evidence discriminate between a direct re-oxidation of isolated Pt centres obtained allows us to derive a room-temperature cyclic mechan- that is subject to a relatively high activation energy—as ism, involving both Pt carbonates and Pt nanoparticles for the CO considered by Moses-DeBusk et al. —or whether, and as oxidation under the periodic conditions of operation employed. might be implied by the DRIFTS, desorption of molecular CO Under CO: from metallic Pt nanoparticles has to occur first. To resolve this 0 0 0 Pt þ CO ! PtðÞ CO þ PtðÞ CO ð1Þ issue we therefore conducted static and time-resolved X-ray L 2 B absorption fine structure (XAFS) spectroscopy at the Pt L edge IV under identical experimental conditions. Some results of this are Pt ðÞ O þ CO ! PtðÞ CO ð2Þ shown in Fig. 4. phys Figure 4a shows the k-weighted extended XAFS (EXAFS) PtðÞ CO ! PtðÞ O þ CO ! PtðÞ O þ CO ð3Þ 3 2gðÞ obtained from the 5 wt% Pt/Al O catalyst, whereas Fig. 4b shows 2 3 the corresponding Fourier transform. The results of best fitting Under O : the EXAFS spectrum are given in Table 1. These measurements 0 0 phys 0 PtðÞ CO þ PtðÞ CO ! 2CO þ 2Pt ð4Þ show that, as might be deduced from the dominance of linear L 2 B 4 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. Absorbance NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE a b 2468 10 12 14 16 01234567 –1 k (Å ) R (Å) c d Blue = PtO 2 CO O Red = Pt foil Oxidation 75%Pt + 25%PtO reduction 5 wt% Pt/Al O 2 3 80%Pt + 20%PtO 11.54 11.57 11.6 11.64 060 120 180 240 300 Energy (keV) Time (s) Figure 4 | XAFS during periodic redox operation at 298 K. (a) k -weighted EXAFS derived in situ from the 5 wt% Pt/Al O catalyst along with theoretical 2 3 fit (red); (b) the corresponding Fourier transform (again with the theoretical fit in red). (c) Comparison of Pt L edge X-ray absorption near edge structure 0 0 from this sample with that from Pt foil (red) and PtO (blue). The sample spectrum is also compared with linear combinations (80% Pt and 75% Pt as indicated) of these two reference spectra. (d) The temporal variation observed in the Pt L white line intensity during exposure of the catalyst to 5 vol% CO/Ar and then 21 vol% O /Ar. 0 0 where Pt is the metallic platinum, Pt (CO) is the CO adsorbed Table 1 | Best-fit parameters obtained for the 5 wt% 0 in linear geometry on metallic Pt, PtðÞ CO is the CO adsorbed in 2 B Pt/Al O sample measured in situ at room temperature and 2 3 twofold bridge geometry on metallic Pt, Pt(CO ) is the platinum after reduction at 573 K. IV carbonate intermediate, Pt (O) is the isolated oxidic platinum species, O is the adsorbed oxygen, Pt(O) is the product of * w 2  2 z phys phys Element CN R (Å) DW (2r /(Å ) carbonate decomposition, CO and CO are physisorbed Pt 8 2.75 0.014 CO and CO, respectively, and O ,CO and CO are gas 2 2(g) 2(g) (g) O 1.9 2.00 0.007 phase species. Within this mechanism, the oxidative regeneration of the EXAFS, extended X-ray absorption fine structure. E ¼ 10.31¼ the edge position relative to the vacuum zero (Fermi energy). precursor to carbonate formation, equation (6), limits the T E 3 E 3 T E R%¼ 33.37 ¼ ( [w  w ]k dk/[w ]k dk) 100%, where w and w are the theoretical and reformation of the carbonates in oxidizing conditions. However, experimental EXAFS and k is the photoelectron wave vector. The Debye–Waller factor¼ 2s , where s is the root mean square internuclear separation. the combined evidence suggests that this is mediated via the Pt Other parameters: Attentuation factor (AFAC), related to the proportion of electrons performing nanoparticles and therefore can only happen once free sites are an EXAFS-type scatter on absorption, is 0.875. Structural data were obtained by fitting the EXAFS in k space in the range: k ¼ 2.5–16.5 Å . created on the Pt nanoparticles rather than through a direct *Coordination number ( ca. 10% stated value). dissociation of O at the precursor sites to the carbonates. As such, wDistance of scatterer atom from central atom ( ca. 1.5% stated value). zDebye–Waller factor. we suggest that it is CO desorption (equation (4)) that effectively limits the efficacy of this reactive pathway. In this respect, our mechanism differs from that of Moses-DeBusk et al. who considered only a direct dissociation of O by atomically dispersed Pt precursors to the carbonate species. As a result, and in respect of steady-state operation, our room- 0 0 Pt þ O ! Pt þ 2O ð5Þ 2gðÞ a temperature mechanism still suffers from the poisoning effect of CO that hobbles the classic LH mechanism. Indeed, steady-state IV PtðÞ O þ O ! Pt ðÞ O ð6Þ operation (Supplementary Fig. 2) under a 4O :CO flow leads to a very small, but non-zero, production of CO at room tempera- phys 2 CO ! CO ð7Þ ðÞ g ture; at present, it is only through adopting a periodic operation that this reactive impasse may be, to some degree, circumvented. phys IV CO þ Pt ðÞ O ! PtðÞ CO ð8Þ A priori we cannot definitively say where these carbonates actually reside. It is clear from our experimental methodology phys PtðÞ CO ! PtðÞ O þ CO ! PtðÞ O þ CO ð9Þ 3 2gðÞ 2 that the oxidized Pt entities that lead to their formation can survive reduction in H /Ar to at least 573 K. This IV PtðÞ O þ O ! Pt ðÞ O ð10Þ 2 characteristic mitigates against their existence within a NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. Normalized XAFS f(k)χ(k ) XANES intensity FT intensity ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 initiated (TB423 K). At each temperature, the DRIFTS and MS data were commensurate and reactive surface oxide in the manner proposed 12 collected, with the latter being used to calibrate the CO response in terms of the by Ackermann et al. Instead, in this respect we would tend to overall CO conversions. concur with the propositions of Moses-DeBusk et al. who only Having conducted this calibration experiment, the sample was re-cooled under considered atomically dispersed oxidic Pt centres adsorbed on the reaction mixture, purged again before a second CO/O switching experiment was conducted and again DRIFTS and MS data simultaneously acquired. It is the alumina surfaces. Further, high-angle annular dark-field scanning results of this second switching experiment that are reported in Figs 1–3. We note TEM measurements (Supplementary Fig. 6) also support this that both first and second switching experiments, as well as the 16 cycle view and are consistent with recent works that have shown that experiment, return the same global results save for some minor differences in similar atomic or quasi atomically adsorbed Pt and Pd species can overall CO production. In addition, although not shown here, essentially the 23,24 same results can be achieved without any pre-reduction of the catalyst and the be highly reactive for oxidation reactions . Although we species responsible for the chemistry we have reported are present in the as- cannot rule out a role for interfacial Pt-O-Al sites, these would received catalyst. not be sufficient (based on the particle size distribution from TEM, see Supplementary Fig. 5) to yield the ca. 10–20% of active X-ray absorption spectroscopy. Pt L edge XAFS was collected in transmission Pt that our results indicate are mediating carbonate formation mode at the SuperXAS beamline at the Swiss Light source (Villigen, Switzerland) and CO turnover in the current case. using a newly installed fast, Si (111) channel cut monochromator system coupled to To summarize, we have demonstrated that room-temperature gridded N -filled ionization chambers for detection. This bidirectional scanning CO oxidation over Pt/Al O is feasible and mediated by oxidized, 2 3 system was operated at 2 Hz, yielding four spectra per second. The static and isolated and reduction-resistant Pt centres that may form reactive reference spectra are obtained as averages over 3 min of acquisition. The in-situ time-resolved X-ray absorption near edge structure data were carbonates from CO and be re-oxidized under O . Using periodic extracted from individual (250 ms time resolution) spectra during CO switching operation we have shown that this process is quantitatively from 5 vol% CO/Ar and 21 vol% O /Ar, and using the same sample environment as reversible and can form a catalytic cycle. At present, the reactive for the DRIFT/MS. Online MS was recorded as for the DRIFTS-based Pt centres necessary for this low-temperature conversion exist as measurements. Data reduction was made using PAXAS and analysis of the EXAFS using EXCURV . a minority species within the conventional catalyst used for these studies: most of the Pt (estimated to be ca. Z80%) is not directly active but does have a role to play in completing the catalytic References cycle. We have shown that at room temperature metallic 1. Langmuir, I. The mechanism of the catalytic action of platinum in the reactions nanoparticulate Pt may act as a reservoir for CO, which may be of 2CO þ O ¼ CO and 2H þ O ¼ H O. Trans. Faraday Soc. 17, 0621–0654 2 2 2 2 2 (1921). transferred to the minority quasi atomic oxidic Pt phase, and 2. Campbell, C. T., Ertl, G., Kuipers, H. & Segner, J. A molecular beam study of converted to CO . The rate-limiting re-oxidation of the carbonate the catalytic-oxidation of CO on a Pt(111) surface. J. Chem. Phys. 73, precursor sites (Pt(O)) also requires the presence of Pt 5862–5873 (1980). nanoparticles that can facilitate O dissociation and oxygen 3. Wintterlin, J., Volkening, J., Janssens, T. V. W., Zambelli, T. & Ertl, G. Atomic transfer to these sites. and macroscopic reaction rates of a surface-catalyzed reaction. Science 278, 1931–1934 (1997). As such, this new pathway for achieving low-temperature CO 4. Alavi, A., Hu, P., Deutsch, T., Silvestri, P. & Hutter, P. L. CO oxidation oxidation immediately poses a number of further challenges to on Pt(111): an ab initio density functional theory study. Phys. Rev. Lett. 80, material synthesis and fundamental understanding. Foremost 3650–3653 (1998). among these is the fundamental question of whether a 5. Allian, A. D. et al. Chemisorption of CO and mechanism of CO oxidation on direct oxidation of the precursor to the Pt carbonates can be supported platinum nanoclusters. J. Am. Chem. Soc. 133, 4498–4517 (2011). 6. Trovarelli, A. Catalysis by Ceria and Related Materials (Imperial College Press, attained and how we may promote CO desorption from the 2002). Pt nanoparticles to permit more efficient O dissociation. Second 7. Vayssilov, G. N. et al. Support nanostructure boosts oxygen transfer to would be to understand, from a material synthesis perspective, catalytically active platinum nanoparticles. Nat. Mater. 10, 310–315 (2011). how to optimize the levels of such carbonate forming Pt species 8. Chen, G. et al. Interfacial effects in iron-nickel hydroxide–platinum on Al O from ca. 10–20% minority to a working majority. nanoparticles enhance catalytic oxidation. Science 344, 495–499 (2014). 2 3 9. Haruta, M. Size- and support-dependency in the catalysis of gold. Catal. Today 36, 153–166 (1997). Methods 10. Bond, G. C. & Thompson, D. T. Catalysis by gold. Catal. Rev. Sci. Eng. 41, Infrared spectroscopy. Combined DRIFTS/MS experiments were carried at the 319–388 (1999). Swiss Light Source using a Bruker Vertex 80V infrared spectrometer fitted with a 11. Hashmi, A. S. K. & Hutchings, G. J. Gold catalysis. Angew. Chem. Int. Ed. 45, narrow-band high-sensitivity mercury cadmium telluride (MCT) detector and with 7896–7936 (2006). a Pfeiffer mass spectrometer. DRIFTS spectra were collected with a time resolution 12. Ackermann, M. D. et al. Structure and reactivity of surface oxides on Pt(110) of 0.433 s and 2 cm resolution. during catalytic CO oxidation. Phys. Rev. Lett. 95, 255505 (2005). The in-situ cell used for these experiments was that recently described by 13. Newton, M. A., Chapman, K. W., Thompsett, D. & Chupas, P. J. Chasing Chiarello et al. for transient experimentation and was windowed with CaF . This changing nanoparticles: a time resolved total X-ray scattering study of the cell was connected to a gas handling system equipped with fast solenoid switching behavior of Supported Pt nanoparticles during CO oxidation catalysis. J. Am. valves (Parker) controlled using the infrared spectrometer. The reactor exit was Chem. Soc. 134, 5036–5039 (2012). connected to a Pfeiffer mass spectrometer that recorded both the sample 14. Butcher, D. R. et al. In situ oxidation study of Pt(110) and its interaction with temperature and the gas switching events alongside a range of masses compatible CO. J. Am. Chem. Soc. 133, 20319–20325 (2011). with the experiments being made. 15. Miller, D. et al. Different reactivity of the various platinum oxides and Gas flows (50 ml min ) were controlled by Bronkhorst mass flow controllers chemisorbed oxygen in CO oxidation on Pt(111). J. Am. Chem. Soc. 136, with a 2-bar pressure of gases behind the controllers themselves. This set-up 6340–6347 (2014). allowed the purging of the system with Ar, then reduction using 5 vol% H /Ar and, 16. Szlachetko, J. et al. Sub-second and in situ chemical speciation of Pt/Al O 2 3 finally, gas switching experiments between 5 vol%CO/Ar and 21 vol%O /Ar. during oxidation reduction cycles monitored by high-energy resolution off- In the experiments described, 25–30 mg of the 5 wt% Pt/Al O catalyst (Type- 2 3 resonant X-ray spectroscopy. J. Am. Chem. Soc. 135, 19071–19074 (2013). 94, Johnson Matthey) was loaded into the DRIFTS cell. After purging in Ar, this 17. Moses-DeBusk, M. et al. CO oxidation on supported single Pt atoms: sample was then reduced under 5 vol% H /Ar to 573 K under a linear heating ramp experimental and ab initio density functional studies of CO interaction with Pt (10 K min ) and then held at 573 K for 30 min. The sample was then cooled atom on theta-Al O (010) surface. J. Am. Chem. Soc. 135, 12634–12645 (2013). under 5 vol% H /Ar to 323 K whereon the reducing flow was substituted for 2 3 18. Hollins, P. The influence of surface-defects on the infra-red spectra of adsorbed flowing Ar once more. At 298 K, the sample was then exposed to 5 vol% CO/Ar for species. Surf. Sci. Rep. 16, 51–94 (1992). 52 s before the flow is switched to 21 vol% O /He for 208 s, whereas DRIFTS and 19. Kisliuk, P. The sticking probabilities of gases chemisorbed on the surfaces of MS data were collected. This experiment was then repeated for a total of 16 CO/O solids. J. Phys. Chem. Solids 3, 95–101 (1957). cycles. Subsequently, the sample was then exposed to a 4O :1CO reaction mixture 20. Bowker, M., Bowker, L. J., Bennett, R. A., Stone, P. & Ramirez-Cuesta, A. In 1 17 (again 50 ml min total flow, 4.2  10 molecules CO per second) and the consideration of precursor states, spillover and Boudart’s ‘collection zone’ and temperature incremented until light-off of conventional CO oxidation catalysis was of their role in catalytic processes. J. Mol. Cat. A 163, 221–232 (2000). 6 NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms9675 ARTICLE Financial support from the Swiss National Science Foundation (SNF, project number 21. Busca, G. & Lorenzelli, V. Infrared spectroscopic identification of species 200021–138068) is greatly acknowledged. arising from reactive adsorption of carbon oxides on metal oxide surfaces. Mater. Chem. 7, 89–126 (1982). 22. Liu, X., Korotkikh, O. & Farrauto, R. Selective catalytic oxidation of CO in H : Author contributions structural study of Fe oxide-promoted Pt/alumina catalyst. Appl. Catal. A 226, Experiments were conceived and carried out by M.A. Newton with experimental 293–303 (2002). assistance from V. Marchionni and D. Ferri (DRIFTS/MS) and M. Nachtegaal and 23. Narula, C. K., Allard, L. F., Stocks, G. M. & Moses-DeBusk, M. Remarkable NO G. Smolentsev (XAFS). M.A. Newton wrote the manuscript with D. Ferri and oxidation on single supported platinum atoms. Sci. Rep. 4, 7238 (2014). M. Nachtegaal. All authors have proofread the manuscript. 24. Peterson, E. J. et al. Low temperature carbon monoxide oxidation catalyzed by regenerable atomically dispersed Palladium on Alumina. Nat. Commun. 5, 4885 (2014). Additional information 25. Chiarello, G. L. et al. Adding diffuse reflectance infrared Fourier transform Supplementary Information accompanies this paper at http://www.nature.com/ spectroscopy capability to extended X-ray absorption fine structure in a new naturecommunications cell to study solid catalysts in combination with a modulation approach. Rev. Sci. Instr. 85, 074102 (2014). Competing financial interests: The authors declare no competing financial interests. 26. Binsted, N. PAXAS: Programme for the Analysis of X-ray Adsorption Spectra Reprints and permission information is available online at http://npg.nature.com/ (University of Southampton, 1988). reprintsandpermissions/ 27. Binsted, N. EXCURV98 (CCLRC Daresbury Laboratory Computer Programme, 1998). How to cite this article: Newton, M. A. et al. Room-temperature carbon monoxide oxidation by oxygen over Pt/Al O mediated by reactive platinum carbonates. 2 3 Nat. Commun. 6:8675 doi: 10.1038/ncomms9675 (2015). Acknowledgements M.A.N. thanks the Swiss Light Source/Paul Scherrer Institut and the Department of This work is licensed under a Creative Commons Attribution 4.0 Physics at the University of Warwick, for the visiting scientist positions that facilitated International License. The images or other third party material in this this work. Johnson Matthey (Dr D. Thompsett) and Umicore are also thanked for article are included in the article’s Creative Commons license, unless indicated otherwise provision of the samples used in these studies. We thank Frank Krumeich and Maksym in the credit line; if the material is not included under the Creative Commons license, Kovalenko (ETH Zu¨rich) for enabling high-angle annular dark-field scanning TEM. Electron microscopy was carried out at the Scientific Center for Optical and Electron users will need to obtain permission from the license holder to reproduce the material. Microscopy (ETH Zurich). We are thankful to Lorenz Ba¨ni for technical support. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ NATURE COMMUNICATIONS | 6:8675 | DOI: 10.1038/ncomms9675 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved.

Journal

Nature CommunicationsSpringer Journals

Published: Oct 22, 2015

There are no references for this article.