Selective dehydrogenation of propane to propene with O2–H2 on bifunctional Pt-H3PMo12O40 catalysts

Applied Catalysis A: General 477 (2014) 1–7 Contents lists available at ScienceDirect Applied Catalysis A: General jou rn al hom ep age: www.elsev ier .com/ locate /apcata Selecti ene bifunct N.I. Kuzn ovs Yu.A. Che kho a Boreskov Inst b Novosibirsk S c Institute of Hy d Rohm and Ha a r t i c l Article history: Received 26 A Received in re Accepted 1 Ma Available onlin Keywords: Dehydrogenation Propane Propene Oxygen Hydrogen Platinum ed by H2 m ith E has b 0 ◦C, 94–96% at propane conversion of 2–4%). The process probably proceeded through oxidation of propane with peroxo species to C3-alcohols on Pt sites, followed by dehydration of the alcohols on the acid sites of HPA. The catalytic properties were strongly dependent both on the feed composition (the C3H8/H2/O2/N2 ratio) and on the composition of catalysts, optimum HPA loading being 20–30 wt.% at 0.2–1 wt.% of Pt. Under optimal operating conditions, a space time yield of propene was at a level of 420 g kgcat−1 h−1, without a tendency to decline with time on stream. 1. Introdu The low their produ in chemical drogenation cost–effecti forward co proceeds at temperatur rapid carbo tion tempe the feed, as equilibrium achieved at of oxidants tion, thus p with a high process bec ∗ Correspon E-mail add http://dx.doi.o 0926-860X/© © 2014 Elsevier B.V. All rights reserved. ction -molecular-weight alkenes and new technologies for ction from corresponding alkanes are in great demand industry. The conventional steam-cracking and dehy- processes suffer from many disadvantages and are ve only on a very large scale. In particular, the straight- nversion of propane to propene is endothermic and reasonable rate and to acceptable degree only at a high e (>600 ◦C), which facilitates side reactions and causes nization and deactivation of catalysts [1–3]. The reac- rature can be decreased upon addition of oxidants to the conversion of H2 to H2O shifts the thermodynamic , and rather high conversions of propane have been 550 ◦C or even lower temperatures [4–6]. The addition simultaneously eliminates or reduces the coke forma- roviding stable catalytic activity and allowing the feed er concentration of alkane to be used. At last, the overall omes exothermic, thereby saving energy and enhancing ding author. Tel.: +7 383 3269720; fax: +7 383 3269529. ress: [email protected] (N.I. Kuznetsova). capital and operational efficiencies. The oxidative dehydrogena- tion (ODH) is therefore considered a promising alternative to the conventional dehydrogenations and is currently under intensive study. Various catalysts, promoters and oxidants have been offered and can be found in the extensive literature on ODH. Despite the strong efforts made, there remain problems to be solved, how- ever, the easy combustion of alkenes under the oxidizing conditions being the most serious one [7]. With molecular oxygen as a sole oxidant, the process still occurs at too high temperatures where the alkene burning accelerates. Some but limited improvement has been achieved with the use of milder oxidants, such as CO2 [8]. Strong oxidants, e.g., N2O [9], can also help in gaining a high selectivity to alkenes, in a given case through lowering the tem- perature required for the oxidation of starting alkanes; however, such oxidants are expensive. Meanwhile, the problem could be cir- cumvented even with O2 or air – the cheapest and most attractive oxidant – provided that the oxygen was activated by an indepen- dent reductant. For generating in situ the more reactive oxygen species, a com- bination of O2 with H2 – another simple and cheap reagent – was used most often. It is well-known that the partial reduction of O2 by H2 on Au, Pd, or Pt catalysts produces H2O2 or peroxide compounds rg/10.1016/j.apcata.2014.03.001 2014 Elsevier B.V. All rights reserved. ve dehydrogenation of propane to prop ional Pt-H3PMo12O40 catalysts etsovaa,∗, G.Ya. Popovaa, L.I. Kuznetsovaa, V.I. Zaik salova, T.V. Andrushkevicha, A.S. Lisitsyna, V.A. Li itute of Catalysis, Novosibirsk 630090, Russian Federation tate University, Novosibirsk 630090, Russian Federation drocarbons Processing, Omsk 644040, Russian Federation as Company, Spring House, PA 19477-0904, USA e i n f o ugust 2013 vised form 27 February 2014 rch 2014 e 12 March 2014 a b s t r a c t Properties of Pt/SiO2 catalysts modifi genation of propane with an O2 and IRS, XPS, and HAADF-STEM coupled w and well-mixed on the silica surface obtained at temperatures as low as 15 with O2–H2 on kiia,b, S.V. Koscheeva, lobovc, S. Hand a heteropoly acid (HPA) were studied in oxidative dehydro- ixture. The catalysts were characterized with N2 adsorption, DX, and the possibility for Pt and HPA to be highly dispersed een shown. In the presence of O2 and H2, propene could be but heating to 200 ◦C provided much better selectivity (up to 2 N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 (e.g., review [10]), which are capable of oxidizing hydrocarbons at a low temperature. For example, propylene was oxidized with O2 and H2 on Au/TiO2 at 100 ◦C [11] and on Pd/TiO2 at ambient temperature [12]. Hydroxylation of benzene by the O2–H2 oxidant has been ca composite fied with va acetylaceto ketones wi catalyst at alkanes wit Lin and Sen used to con of Pd0.08Cs2 ence of Pd/ Temperatur and ethane Several and H2. black/VO(ac acetone, ac in CF3COOH produced p CO2 [25]. In catalyst at 95%, probab Ti OOH [26 oxidations b of the inter and isoprop n-propanol oxygenates While t are oxygen recently re Au/TiO2 cat [29]. In the bilities of a O2–H2 mix properties, and a heter previously benzene in shown activ were: (1) w Pt-HPA cata that govern estimate th conditions 2. Experim 2.1. Catalys Pt-H3PM 10–60 wt.% through sim a silica sup solution co proportion. (2 h at 100 H2 at 300 ◦ prepared t ple was ob reduced sa of H3PMo12O40, followed by final drying in air at 100 ◦C. Pt-free samples, 20H3PMo12O40/SiO2 and 20H3PMo12O40-0.9H3PO4/SiO2, were obtained analogously, with the use of bare support instead of Pt/SiO2. In the text, the catalysts are denoted as xPt-yH3PMo12O40- 4(S), ) ind the c sour 3 g− ed (0 % Cl, s use tion. talys pos en ad 2400 met (X-r n ES- on ( red s d to coun f reg imp EM -STE n El scop (JEO epos ere spac rogr (Infr ) we stan 0 mg −1, t ts. talyt alytic r (5 m lled tal v g of feed as d o ca 0 ml gas w 2, C3H 2; Po 2; ac olecu trati irecte sitio ered rried out in one step on Pd/Ti-silicalite [13], Pd-Sil-1 membrane [14], Pd–Cu [15], Pd, Rh, Ir and Pt modi- nadium oxide [16,17], as well as with V, Fe, La, and Y nates [18]. Cycloalkanes were oxidized to alcohols and th O2 and H2 on a heterogeneous Pt/Eu2O3/TiO2/SiO2 40 ◦C [19]. Low-temperature (85 ◦C) oxidation of light h O2 and H2 on Pd catalyst has been reported first by [20]. The O2 and H2 containing mixtures were later vert CH4 to HCOOH and methyl esters in the presence .5H0.34PVMo11O40 at 150–320 ◦C [21], and in the pres- C and Cu(CH3COO)2 or V(V) compounds at 80 ◦C [22]. es >300 ◦C were required for the oxidation of methane on FePO4 in the absence of precious metals [23]. papers describe oxidation of propane with O2 Electrochemical oxidation of propane over Pd- ac)2/Carbon-fiber cathode at 25 ◦C resulted in etaldehyde and CO2 [24]. In the presence of EuCl3 /CH3COOH solutions and Zn, instead of H2, propane ropyl alcohols, acetone, propanal and large amount of gaseous phase, propane was oxidized on an Au/TS-1 170 ◦C to acetone and isopropanol with selectivity of ly through intermediate formation of H2O2 and then ]. The identity of the products which are formed in the y O2 and H2 and by H2O2 confirms the peroxide nature mediates. For instance, aqueous H2O2 led to acetone anol on titanosilicate [27] and acetone, isopropanol, , propanal in Fe2+ solution [28], the selectivity to C3 being close to 100% in both the cases. ypical products of the oxidations with O2 and H2 -containing compounds, Oyama and co-workers have ported a predominant formation of propene on an alyst (selectivity of 64% at propane conversion of 2.8%) present study we examine more in detail the possi- selective alkane-to-alkene conversion with the aid of ture. To have more “freedom” in influencing catalytic bi-functional catalysts composed of Pt nanoparticles opolyacid (HPA) have been used. Such catalysts have demonstrated good performance in the oxidation of solution [30] and in vapor phase [31] and have also ity in the oxidation of light alkanes [32]. Our objectives ith propane as the substrate, to assess the feasibility of lysts to produce alkenes, (2) to reveal the main factors the process in the presence of O2 and H2, and (3) to e extent to which the catalyst composition and reaction can help raise the selectivity to propene. ental ts preparation o12O40-H3PO4/SiO2 catalysts with 0.2-2.0 wt.% Pt, H3PMo12O40 and 0–0.9 wt.% H3PO4 were prepared ultaneous introduction of the starting compounds onto port. Silica was impregnated with a freshly prepared ntaining H2PtCl6, H3PMo12O40 and H3PO4 in required The resulted samples were dried and calcined in air ◦C and 2 h at 450 ◦C) and finally treated in flowing C for 1 h. For comparison, several samples were also hrough successive impregnation. First, Pt/SiO2 sam- tained under the conditions specified above, and the mple was further impregnated with a fresh solution zH3PO and (S tion. As native 0.84 cm was us K, 0.04 Pt) wa extrac 2.2. Ca Com Nitrog (ASAP- by BET XPS with a radiati powde fastene under in Pt4 minum HRT HAADF missio spectro 2200FS were d maps w Lattice DigMic IR 4 cm−1 ter by per 50 470 cm catalys 2.3. Ca Cat reacto contro 4 ml to heatin bed. A ance w Prior t air (10 outlet for CO C2H4O C3H6O and m concen was d Compo consid where x, y, z stand for wt.% of corresponding component icates samples prepared through successive impregna- atalyst support, an ordinary commercial silica (KSK, ce) with BET surface area 263 m2 g−1, pore volume 1, pore diameter 126 A˚, and particle size 0.2–0.5 mm .81% Na, 0.54% Ca, 0.35% Al, 0.10% Mg, 0.06% Fe, 0.06% 0.01% S as main impurities). H2PtCl6·6H2O (37.56 wt.% d as supplied, and H3PMo12O40 was purified by ether t characterization ition of all catalysts was checked by XRF analysis. sorption–desorption isotherms were obtained at 77 K , Micromeritics), specific surface areas were calculated hod and pore-size distributions by BJH method. ay Photoelectron Spectroscopy) spectra were recorded 300 (Kratos Analytical) spectrometer using Mg anode h� = 1253.6 eV) under fixed transmission mode. The amples were dusted onto an adhesive tape, which was a sample holder. Narrow XPS regions were recorded t storage mode. Spectra of Pt-20H3PMo12O40 catalysts ion were processed to exclude Al2p signal from alu- urity in the support [33]. (High Resolution Transmission Electron Microscopy), M (High Angle Annular Dark Field Scanning Trans- ectron Microscopy), EDX (Energy Dispersive X-ray y) and elemental mapping were made with a JEM- L) microscope. Suspensions of the catalysts in hexane ited on carbon-film-coated copper grids. The elemental obtained on Mo L (E = 2.29 keV) and Pt L (E = 9.44 keV). ing was calculated by FFT (Fast Fourier Transform) using aph (GATAN) soft. ared) spectra (4000–250 cm−1, 30 scans, resolution re recorded on a BOMEM MB-102 FTIR spectrome- dard technique of pressing the sample with KBr (2 mg KBr). Normalized to the intensity of Si–O–Si band at he spectrum of SiO2 was subtracted from the spectra of ic testing testing was performed in a fixed-bed quartz flow l in volume, i.d. 8 mm) mounted inside a temperature- furnace. Catalysts were mixed with crushed quartz to olume to ensure uniform gas flow and prevent over- catalyst. Thermocouple was placed inside the catalyst consisting of C3H8 (99.94% purity), O2, H2 and N2 bal- osed by mass flow controllers at ambient pressure. talytic reaction, the catalyst was treated in a flow of min−1) at 200 ◦C for 1 h. During catalytic testing, the as analyzed by GC using 4 columns: Porapak T (TCD) 8, H2O, C3H4O, C3H6O, 1-C3H7OH, 2-C3H7OH, HCOOH, rapak Q (FID) C3H8, C2H4O, C2H4O2, C3H4O, C3H4O2, tivated Al2O3 (FID) for CH4, C2H6, C2H4, C3H8, C3H6; lar sieves (TCD) for H2, O2, CH4, CO. To measure the ons of C3H8, O2, H2 and products in the outlet gas, it d to chromatograph through heated lines (≥110 ◦C). n of the outlet gas 1 h after starting the experiment was corresponding to a steady-state. N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 3 Fig. 1. IR spec (b), 0.5Pt-20 20H3PMo12O4 (SiO2). 3. Results 3.1. The sta A chem XPS. In th tra of 0.5P 20H3PMo12 with BE Pt4 tion of Pt [3 or simultan Changes of of Pt4f sign only signals HPA, but th impregnati Mo5+ (BE M to the litera The pos supporting troscopy (F which refle adsorbed s 1060–780 c ting H3PMo Fig. 1a). IR samples (Fi ples. Hence structure of HPA, at least seriously. It is known that H3PO4 can cause restructuring of the PMo12O403− anion in aqueous solutions but such restructuring needs time [37]. Strong broadening and some the IR bands were only observed after the treatment of sup- HPA s in HPA sente ith n re v phas an p unifo silic taneo acid mina -typ t-20 nm i ima show FFT o refl dista 9]. T d du trosc yed. artic t and t th shift of ported change ions of Pre tion w and po ported the me rather out the (simul phoric Exa “island of 0.5P 20–30 HRTEM Fig. 2c tures. narrow to the unit [3 reduce IR spec destro Pt p suppor ticles. A tra of solid H3PMo12O40 (a) and supported samples: 20H3PMo12O40 H3PMo12O40(S) (c), 20H3PMo12O40-0.9H3PO4 (d), and 0.5Pt- 0 (e); indicated by stars are the bands related to the support and discussion te of supported components ical state of the main elements was probed with e region of Pt4f binding energies (BE), the spec- t-20H3PMo12O40(S), 0.5Pt-20H3PMo12O40 and 0.5Pt- O40-0.9H3PO4 catalysts had characteristic doublets f7/2 71.4–71.6 eV, thus indicating a complete reduc- 4] irrespective of the preparation method (successive eous impregnation) or presence of phosphoric acid. Pt loading (0.2–2 wt.%) were reflected in the intensity als but did not influence BE. In the region of Mo5d BE, from Mo6+ were observed for samples after supporting e reductive treatments in H2 (the case of simultaneous on, Section 2.1) gave rise to some contribution from o3d5/2 232.1–232.5 eV and 230.9 eV, which are close ture values for Mo6+ and Mo5+, respectively [35]). sible changes in the structure of H3PMo12O40 after and further treatments were monitored with IR spec- ig. 1). Except for broadening of the band at 865 cm−1, cts some deformation of the heteropoly anion in the tate, there were no visible changes in the range of m−1 (characteristic of H3PMo12O40 [36]) after suppor- 12O40 onto SiO2 or Pt/SiO2 (Fig. 1b and c; cf. with spectra of freshly prepared 20H3PMo12O40-H3PO4 g. 1d) were very similar to those of H3PO4-free sam- , the addition of phosphoric acid does not affect the Pt crystallit with EDX g (Fig. 2h). In neous prese signals corr sample. For atomic rati strengthene concentrate Mo and ove highly disp 3.2. Catalyt In the p able to cat temperatur ple 1). How exclusively consumed the additio among the containing acetone, pro in minor am ticles (XPS but the mo selectivity. The eve 3.1) gives o the product burning (CO ature (Tabl catalytic pr in H2 (a sample of 0.5Pt-20H3PMo12O40 in Fig. 1e). Such IR spectra are usually attributed to reduction of Mo6+ [38], and this is consistent with the XPS data. d in Table 1 are the results of catalysts characteriza- itrogen adsorption/desorption. Both BET surface areas olumes per gram of total sample (support plus sup- e) decreased with increasing H3PMo12O40 loading, but ore diameter remained nearly constant. This indicates a rm distribution of the heteropoly compound through- a surface, irrespective of the impregnation method used us or successive impregnation) or addition of phos- . tion of the samples with HAADF-STEM revealed an e” distribution of HPA over the support (a sample H3PMo12O40-0.9H3PO4 in Fig. 2). Plain HPA particles n diameter and ∼1 nm thick predominated (Fig. 2a). ge of HPA particles in Fig. 2b and a scaled-up fragment in periodical fringes with weakly ordered atomic struc- f a small area inside the fragment (Fig. 2d) detected ections with ∼0.25 nm periodicity, which corresponds nce between Mo layers inside the Keggin PMo12O403− hus, although some of the Mo6+ ions have been partly ring the catalyst treatment in H2 (as shown by XPS and opy), the structure of the HPA has not been completely les appear undistinguishable on the background of the HPA. This indicates a rather small size of the metal par- orough examination of many images, one large (visible) e has only been observed (Fig. 2e, label 1). Probing it ave a strong signal from Pt but negligible one from Mo all other cases the probes with EDX showed simulta- nce of Pt and Mo, and the intensity ratio between their esponded well to the total contents of Pt and Mo in the example, EDX from region 2 in Fig. 2e resulted in Pt/Mo o of 0.04 (Fig. 2i). There could be places with a locally d Mo signal from bulky HPA particles but Pt was always d in the places abandoned with HPA (Figs. 2f and g with rlay Mo–Pt maps). It follows that both HPA and Pt are ersed and well-mixed on the silica surface. ic properties resence of O2 and H2, all Pt-containing samples were alyze oxidation of propane at 150 ◦C and even lower es, while HPA alone left propane intact (Table 2, exam- ever, HPA-free Pt/SiO2 catalysts directed the process to CO2, and all hydrogen and oxygen was rapidly in the formation of water (Table 2, example 2). Only n of HPA to Pt resulted in the appearance of propene products, along with a number of C2 and C3 oxygen- compounds (propyl and isopropyl alcohols, propanal, panoic acid, acetaldehyde, acetic acid, and some others ounts). It follows that the presence of Pt(0) nanopar- data, Section 3.1) is prerequisite for catalytic activity, dification of Pt surface with HPA is needed to provide n distribution of Pt and HPA on the support (Section pportunity for Pt and HPA to interact. Nevertheless, s of C C bond cleavage (C2-oxygenates) and complete 2) still dominated, despite the low reaction temper- e 2, example 3), and no substantial improvements in operties were achieved under the given conditions upon 4 N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 Table 1 Texture characteristics of catalyst samples. Catalyst BET surface area (m2 g−1) Pore volume (cm3 g−1) Average pore diameter (nm) 1Pt-10H3PMo12O40 230 0.74 13 1Pt-30H3PM 0.56 0.5Pt-20H3P 0.62 0.5Pt-20H3P 0.62 0.5Pt-40 H3P 0.44 0.5Pt-60 H3P 0.26 0.5Pt-20 H3P 0.62 Table 2 Catalytic prop No. C 1 2 2 1 3 0 4 1 a Feed: (1–3 time 0 b As detecte c Carbonyl a d Predomina Table 3 Reaction prod No. sb 1 2 3 4 a 2Pt-20H3P alanc b (Mol)(mol c Carbonyl a varying the Some, but catalysts w propane con (from the da under simil argued agai Pt and HPA by means o preparation Surprisin resulted in CO2 could b (Table 4). U practically No. 2, 5, 8) a and were w be expected large HPA l latter is pro Pt at the ov ical blockag particles (F We have not be don the possible effect from acid to HPA the propane The exa tion. The ra o12O40 173 Mo12O40 (S) 204 Mo12O40 204 Mo12O40 143 Mo12O40 88 Mo12O40-0.9H3PO4 195 erties of mono- and bi-component samples at 150 ◦C. atalyst Substratea Resting substrate and productsb C3H8 C3H6 C3H7OH 0 H3PMo12O40 C3H8 ≥99.9 0 0 Pt C3H8 99.5 0 0 .5Pt-20H3PMo12O40 C3H8 99.6 0.14 0.02 Pt-20H3PMo12O40 Propene 38.2 55.8 0.03 ) C3H8/O2/H2 18/5/10 and (4) C3H6/O2/H2 3.4/5/10 (vol.%, N2 balanced); retention d in outlet gas, in terms of (mol)(mol of starting substrate)−1 × 100. nd carboxy compounds. ntly acetic acid. ucts at 150 ◦C and different concentrations of H2 in the inlet gasa. [H2]0 vol.% Conversion, % Product H2 O2 C3H8 C3H6 0 – ∼0 0 0 5 100 66 0.84 0 8.5 95 82 0.73 0.41 10 85 85 0.69 0.43 Mo12O40 as catalyst; 18 vol.% of C3H8 and 5 vol.% of O2 in the inlet gas (H2 and N2 b of starting propane)−1 × 100. nd carboxy compounds. composition of Pt-HPA catalysts and/or pretreatments. limited, increase in selectivity was only observed for ith higher contents of Pt; for instance, the fraction of verted to propene over a 2%Pt-HPA catalyst rises to 60% ta in Table 3, case 4), as compared to 45% for 0.5%Pt-HPA ar conditions (Table 2, case 3). All the preliminary data nst the possibility to gain an intimate contact between and thereby provide satisfactory catalytic properties f the simple impregnation method used for the catalyst . gly, the increase in reaction temperature to 200 ◦C much better catalytic performance – the burning to e depressed and propene obtained as the main product nder the given conditions, the catalytic properties were independent of Pt content within 0.2–1 wt.% (Table 4, nd of HPA content within 20–30 wt.% (Table 4, No. 2, 3) ell reproducible in independent preparations. As could , they responded negatively to both very small and very oadings (Table 4, No. 1 and No. 6, 7, respectively). The bably caused by a drop in the hydrogenation activity of ersaturation of its surface with HPA species, but a phys- e of the small Pt particles by the relatively large HPA ig. 2) is also not excluded. tried to increase acidity of the catalysts. As it could e through increasing the HPA content, we examined effect of other acids and observed a strong promoting H3PO4. The addition of small amounts of phosphoric did not affect the selectivity to propene but enabled conversion to be doubled (Table 4, No. 9, 10). mples in Figs. 3 and 4 show effects of feed composi- te of propane conversion increased at increasing the concentrati being follow (Fig. 3). Mo straightly f time yield selectivity. (Table 2, ca and low con The pro O2, but the tion rate an H2 (Fig. 4). With none absence of H propene am be a sole pr tion zone ( initial conce selectivity t ered indica if proper re Figs. 3 and appreciable measureme The data H3PO4 do n however, t of HPA spe at the high some (posi Neverthele 13 12 12.5 12 12 13 H2/O2 conversion, % Other C3-oxc C2-oxc CO2 H2 O2 0 0 0 0 0 0 0 1.2 >98 >98 0.02 0.11 0.18 75 73 0.5 0.8d 3.6 97 88 .4 s. Alcohols C3–oxc C2–oxc CO2 0 0 0 0 0 0 0.01 2.50 0.02 0.01 0.09 0.54 0.01 0.01 0.22 0.30 ed); flow rate 150 cm3/min, catalyst mass 0.45 g. on of propane in the feed, the initially linear dependence ed by a weaker one in the region of high concentrations st important, the rates of propene and CO2 formation ollowed the rate of propane conversion, and a space- of propene could be enhanced without a decrease in The extensive burning and hydrogenation of propene se 4) appear strongly diminished at high concentrations versions of the starting propane. pane conversion increased at higher concentrations of increase was followed by a decline, both in the reac- d selectivity, if the concentration of O2 exceeded that of The concentration of hydrogen had most strong effect. of the catalysts the catalytic process occurred in the 2, and rather large amounts of H2 were required to have ong the products. As exemplified in Table 3, CO2 could oduct if all hydrogen had been consumed in the reac- Table 3, case 2), while seemingly small increase in the ntration of H2 could be sufficient to provide rather high o propene (Table 3, cases 3 and 4). This might be consid- tive of instability in the catalyst performance. However, action conditions are found (e.g., those in Table 4 and 4), the catalytic process proceeds smoothly, without changes both in the rate and selectivity for the time of nts (5–10 h). in Figs. 1 and 2 indicated that the small amounts of ot violate the structure of HPA. It does not exclude, hat phosphoric acid assists in a proper distribution cies over Pt surface. The better catalyst performance er temperature (200 ◦C) could also be attributed to tive) changes in the state of supported components. ss, the reaction mechanism seems to contribute to N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 5 Fig. 2. HAADF (e); mapping p the improv tions. Scheme the organic -STEM image of a 0.5Pt-20H3PMo12O40-0.9H3PO4 sample (a); HRTEM image of HPA part icture of MoL (f), overlay mapping picture of MoL and PtL (g) and EDX spectra from place ement of catalytic performance under such condi- 1 shows the possible reactions on Pt-HPA catalysts. As substrate starts to react only in the presence of O2 and H2, we hav has enough in Scheme role of per icles (b), scaled-up part C (c) and FFT from region D (d); STEM image s 1 (h) and 2 (i). e to assume that O2 and H2 produce some species that reactivity to attack the saturated hydrocarbon. [H2O2] 1 denotes both H2O2 and transient species. The key oxide intermediates in the oxidation of alkanes with 6 N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 Table 4 Catalytic properties of different Pt-HPA samples at 200 ◦Ca. No. Catalyst Propane conversion (%) Selectivitiesb (%) C3H6 1 68.0 2 86.0 3 83.7 4 87.2 5 86.0 6 70.6 7 25.7 8 9 10 a Other cond b Carbon effi c Propanal, Fig. 3. Rates o in the inlet 20H3PMo12O4 conversion con Fig. 4. Conver tion in the inle 0.2Pt-20H3PM the O2–H2 (e.g., ref. [21 oxygen-con that C3-oxy and acetone concentrati tions used, 1Pt-10H3PMo12O40 0.6 1Pt-20H3PMo12O40 0.7 1Pt-30H3PMo12O40 1.0 0.5Pt-20H3PMo12O40(S) 1.0 0.5Pt-20H3PMo12O40 0.8 0.5Pt-40H3PMo12O40 0.9 0.5Pt-60H3PMo12O40 0.8 0.2Pt-20H3PMo12O40 0.9 87.0 0.5Pt-20H3PMo12O40-0.9H3PO4 1.6 87.1 0.2Pt-20H3PMo12O40-0.9H3PO4 1.8 85.5 itions: C3H8/O2/H2/N2 = 18/6/10/66 (vol.%); retention time 0.4–0.5 s; in all experiments, ciencies (amount of carbon in product/total carbon in converted propane) × 100. acetone, propanoic acid, acetaldehyde, and acetic acid. f C3H8 conversion and C3H6/CO2 formation vs. C3H8 concentration feed (initial concentrations of O2 and H2 12 vol.% each; 0.2Pt- 0-0.9H3PO4 as catalyst; 200 ◦C; retention time was varied to keep C3H8 stant (3.5%). sion of C3H8 (C) and selectivities to C3H6 and CO2 (S) vs. O2 concentra- t feed (initial concentrations of C3H8 and H2 30 and 12%, respectively; o12O40-0.9H3PO4 as catalyst; 200 ◦C; retention time 0.45 s). mixture has been confirmed in a number of studies ,26,40]). Typical products in the previous studies were taining compounds, and it seems reasonable to suggest genates form in our case as well. Propanol, isopropanol were always present among the products, yet in a low on. The latter is easily explained by the reaction condi- which facilitate further conversion of oxygenates. The Scheme 1. Sc dehydratio stage but it in the catal acid acts on ture can ex verified the experiment at 150–200 with 100% s It is evid the catalyti the oxidatio an importa tone and pr to propene oxidation, o alcohols an abnormally formance ( on inappro tributes to (Tables 2 an The stab implies tha and there a C3H7OH Other oxygenatesc CO2 3.5 14.0 14.5 1.8 5.9 6.3 0.8 8.7 6.8 0.9 5.5 6.4 1.1 6.6 6.3 0 21.1 8.3 0 7.6 66.7 0.4 6.6 6.0 0 4.2 8.7 0 5.6 8.9 conversion of H2 ≥90%, O2 ≥80%. hematic of possible reactions leading to propene and side products. n of C3-alcohols to propene is a key process at the final requires acid sites to occur. Obviously, the role of HPA ytic process is not limited to be a modifier for Pt but the its own, and the increased acidity of (HPA + H3PO4) mix- plain the promoting effect of phosphoric acid. We have rapid conversion of C3-alcohols to propene in special s with addition of isopropyl alcohol into C3H8/N2 feed; ◦C and in the absence of O2 and H2 the process proceeds electivity to propene. ent that Pt nanoparticles play a crucial role in initiating c process, because HPA alone is not capable of catalyzing n of H2 (neither to H2O2 nor H2O). Besides, Pt must have nt role in the hydrogenation of the intermediate ace- opanal to C3 alcohols, as only the latters can dehydrate . The C3 carbonyl intermediates seem liable to further n Pt and/or HPA sites, and must be rapidly converted to d then to propene to avoid burning. This can explain the strong effect of H2 concentration on the catalytic per- Table 3). In its turn, strong adsorption of the carbonyls priate surface sites at a low temperature probably con- the worse catalytic properties under such conditions d 4). le catalyst performance during the time of experiments t detrimental reactions do not accelerate with time re no undesirable changes in the state of supported N.I. Kuznetsova et al. / Applied Catalysis A: General 477 (2014) 1–7 7 components. Nevertheless, in view of the catalyst preparation through the simple impregnation, there could hardly be a strong interaction between Pt and HPA, and the search for additional, more specific promoters for Pt seems to be fruitful. This especially concerns the excessive hydrogen consumption in the process. The permanent and strong decline in the concentration of H2 along the catalyst bed explains the deviation from linearity for the plots in Figs. 3 and 4. The unproductive consumption of H2 still remains a challenging problem for the oxidations with O2–H2. It was out of the scope of the present study because the task needs specific approaches and equipment. At present it can be stated that about five molecules of H2 were consumed at best per a molecule of the targeted product. It is comparable with or better than the level attained in other processes with O2–H2 as the oxidant, in partic- ular, in the propane oxidation on Au/TS-1 to acetone [26] and on Au/TiO2 to propene [29]. 4. Conclusions The hydrogen-assisted oxidative dehydrogenation of propane on Pt-HPA catalysts manifests itself as a multi-stage process that needs a delicate balance between the red–ox and acidic functions of the catalyst. It probably starts from the formation of reactive peroxo species and alcohols, fo spite of the c side reactio edly high se 2–4% and sp taneously w of propertie As emer nanoparticl system; the therefore o synergistic and prepar raise the alk finds confir here with th and use of s help decrea process. References [1] L.R. Ment [2] O.A. Barias, A. Holmen, E.A. Blekkan, J. Catal. 158 (1996) 1–12. [3] J.C. Bricker, Top. Catal. 55 (2012) 1309–1314. [4] S. Sugiyama, T. Osaka, Y. Hirata, K.-I. Sotowa, Appl. Catal. A: Gen. 312 (2006) 52–58. [5] Y.-F. Sun, G.-C. Li, X.-D. Pan, C.-J. Huang, W.-Z. Weng, H.-L. Wan, Acta Phys. Chim. Sin. 28 (2012) 2135–2140. [6] F. Cavani, F. Trifiro, Catal. Today 24 (1995) 307–313. [7] W. Ruettinger, A. Benderly, S. Han, X. Shen, Y. Ding, S.L. Suib, Catal. Lett. 141 (2011) 15–21. [8] R. Wu, P. Xie, Y. Cheng, Y. Yue, S. Gu, W. Yang, C. Miao, W. Hua, Z. Gao, Catal. Commun. 39 (2013) 20–23. [9] A. Ates, C. Hardacre, A. Goguet, Appl. Catal. A: Gen. 441–442 (2012) 30–41. [10] C. Samanta, Appl. Catal. A: Gen. 350 (2008) 133–149. [11] B. Chowdhury, J.J. Bravo-Suarez, M. Date, S. Tsubota, M. Haruta, Angew. Chem. Int. Ed. 45 (2006) 412–415. [12] S. Hikazudani, T. Tatsuya, N. Matsuo, K. Nagaoka, T. Ishihara, H. Kobayashi, Y. Takita, J. Mol. Catal. A: Chem. 358 (2012) 89–98. [13] T. Tatsumi, K. Yuasa, H. Tominaga, Chem. Commun. 19 (1992) 1446–1447. [14] Y. Guo, X. Wang, X. Zhang, Y. Wang, H. Liu, J. Wang, J. Qiu, K.L. Yeung, Catal. Today 156 (2010) 282–287. [15] K. Sasaki, S. Ito, A. Kunai, Stud. Surf. Sci. Catal. 55 (1990) 125–131. [16] T. Miyake, M. Hamada, Y. Sasaki, M. Oguri, Appl. Catal. A: Gen. 131 (1995) 33–42. [17] H. Li, Z. Fu, X. Peng, D. Yin, Stud. Surf. Sci. Catal. 170B (2007) 1399–1404. [18] N. Miyake, M. Hamada, H. Niwa, M. Nishizuka, M. Oguri, J. Mol. Catal A: Chem. 178 (2002) 199–204. [19] I. Yamanaka, Y. Suzuki, M. Toida, Catal. Today 157 (2010) 286–290. [20] M. Lin, A. Sen, J. Am. Chem. Soc. 114 (1992) 7307–7308. [21] J.-S. Min, H. Ishige, M. Misono, N. Mizuno, J. Catal. 198 (2001) 116–121. [22] E.D. Park, Y.-S. Hwang, J.S. Lee, Catal. Commun. 2 (2001) 187–190. ang, amana tsuka Bravo- mun . Cler spro, 06) 25 Bravo –126 Kuzn azard Kuzn 05) 19 . Pop holobo ian, C .A. Sh er an Mould ctron imino occhi ctrosc etters Kuzne gorod 06) 70 . Pope 24. einric proceeds through the oxidation of propane to propyl llowed by the dehydration of the alcohols to propene. In omplexity of the process and a great number of possible ns, it has proved possible to provide an unprecedent- lectivity to propene of 94–96% at propane conversion of ace-time propene yield up to 420 g kgcat−1 h−1, simul- ith the stable catalytic activity and good reproducibility s. ged from the results of physical methods, HPA and Pt(0) es are two principal components of the given catalytic y are evenly distributed over the support and have pportunity for contacting each other and providing a effect. The simplicity of the catalyst, catalyst support, ation method used gives one fair chances to further ene yield through catalyst design and promoters. This mation in the improvements that have been achieved e aid of phosphoric acid. More detailed kinetic studies pecial reactors seem also to be of interest, as they can se the excessive hydrogen consumption in the catalytic asty, O.F. Gorriz, L.E. Cadus, Ind. Eng. Chem. Res. 38 (1999) 396–404. [23] Y. W [24] I. Y [25] K. O [26] J.J. Com [27] M.G [28] C. E (20 [29] J.J. 114 [30] N.I. J. H [31] N.I. (20 [32] G.Y Lik Dal [33] P.M Aug [34] J.F. ele [35] A. C [36] C. R Spe [37] L. P [38] L.I. Nov (20 [39] M.T pp. [40] S. H K. Otsuka, J. Catal. 171 (1997) 106–114. ka, S. Hasegawa, K. Otsuka, Appl. Catal. A: Gen. 226 (2002) 305–315. , I. Yamanaka, Y. Wang, Stud. Surf. Sci. Catal. 119 (1998) 15–24. Suarez, K.K. Bando, T. Akito, T. Fujitani, T.J. Futher, S.T. Oyama, Chem. . (2008) 3272–3274. ici, Appl. Catal. 68 (1991) 249–261. F. Arena, F. Tasselli, A. Regina, E. Drioli, A. Parmaliana, Catal. Today 118 3–258. -Suarez, K.K. Bando, J. Lu, T. Fujitani, S.T. Oyama, J. Catal. 255 (2008) . etsova, N.V. Kirillova, L.I. Kuznetsova, M.Y. Smirnova, V.A. Likholobov, . Mater. 146 (2007) 569–576. etsova, L.I. Kuznetsova, V.A. Likholobov, G.P. Pez, Catal. Today 99 3–198. ova, L.I. Kuznetsova, N.I. Kuznetsova, T.V. Andrushkevich, V.A. v, S. Han, Proc. 3-rd Asia-Pacific Congress on Catalysis, Oct. 12–15, hina, 2003, pp. 874–875. erwood, in: D. Briggs, M.P. Seach (Eds.), Practical Surface Analysis by d X-ray Photoelectron Spectroscopy, Wiley, Chichester, 1983, p. 522. er, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-Ray Photo- Spectroscopy, Perkin–Elmer, Eden Prairie, 1992. , B.A. de Angelis, J. Catal. 36 (1975) 11–22. ccioli-Deltcheff, R. Thouvenot, R. Franck, Spectrochim. Acta A: Mol. . 32A (1976) 587–597. son, I. Andersson, L.-O. Öhman, Inorg. Chem. 25 (1986) 4726–4733. tsova, N.I. Kuznetsova, S.V. Koshcheev, V.A. Rogov, V.I. Zaikovskii, B.N. ov, L.G. Detusheva, V.A. Likholobov, D.I. Kochubei, Kinet. Catal. 47 4–714. , Heteropoly and Isopoly Oxometalates, Springer-Verlag, Berlin, 1983, h, M. Plettig, E. Klemm, Catal. Lett. 141 (2011) 251–258. Selective dehydrogenation of propane to propene with O2–H2 on bifunctional Pt-H3PMo12O40 catalysts 1 Introduction 2 Experimental 2.1 Catalysts preparation 2.2 Catalyst characterization 2.3 Catalytic testing 3 Results and discussion 3.1 The state of supported components 3.2 Catalytic properties 4 Conclusions References