J. CHEM. SOC. FARADAY TRANS., 1994, 90(14), 2099-2106 Physicochemical and Catalytic Properties of Polyaniline Protonated with 12-Molybdophosphoric Acid M. Hasik and A. Pront Department of Materials Science and Ceramics, Academy of Mining and Metallurgy, 30459 Krakow, Mickie wicza 30, Poland J. Pozniczek and A. Bielanski Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences, 30-239 Krako w, Niezapomina jek , Poland Z. Piwowarska, K. Kruczala and R. Dziembaj Faculty of Chemistry, Jagiellonian University, 30-060 Krako w, lngardena 3, Poland Using acid-base-type doping of polyaniline with H,PMo, 2040,two series of catalyst samples were prepared. In series SI the doping and polymerization were carried out simultaneously, leading to a uniform distribution of the dopant over the whole volume of the polymer. In series SII the doping was achieved by protonation of already formed polyemeraldine base with H,PMo, 2040.Since this process must involve diffusion in the solid matrix, the surface becomes enriched in heteropolyanions (Keggin units) due to their limited diffusivity.The examination of physicochemical properties of the samples by XP, EPR and FTIR spectroscopies indicated that the Keggin units incorporated into the polymer matrix retain their identity and represent catalytically active centres. The insertion of the dopant species increases the electrical conductivity of the polymer by several orders of magnitude. The catalytic activity was tested in ethyl alcohol conversion and both the products of acid-base [C,H, and (C,H,),O] and redox (CH,CHO) type reactions were observed.Owing to the surface enrichment of the SII series of cata-lysts their catalytic activity was much higher than that of samples of the SI series nominally doped to the same level. In both types of catalysts the selectivity to the redox reaction product (CH,CHO) was greatly enhanced in comparison with unsupported H,PMo, 2040. Heteropolyacids (HPA) and their salts have been extensively studied because of their interesting catalytic properties.' Sig- nificant research effort has been directed towards the entrap- ment of heteropolyanions in suitable polymeric matrices with the main goal of preparing a new type of polymer-supported catalyst for various applications in heterogeneous and elec- trocatalysis.This entrapment can be conveniently achieved if electroactive conjugated polymers are used as host matrices. Two approaches to this problem can be used: (i) Hetero- polyanions can be incorporated into the growing polymeric matrix during electrochemical or chemical polymerization of a suitable monomer. In this case the growing polymer chains are of a cationic nature and the inserted heteropolyanions serve as charge-compensating species. This approach has .~~~been used by Keita et ~21 and Bidan et in the electro- ~22.~9~ chemical preparation of heteropolyanions containing poly- aniline and polypyrrole. (ii) Alternatively, heteropolyanions can be incorporated into the polymer matrix oia the so-called doping reaction.In this case the process of polymer matrix formation and the process of incorporation of hetero-polyanions into this matrix are separated in time. In other words heteropolyanions are introduced into the already formed, neutral solid polymer. The doping process consists therefore of the transformation of neutral polymer chains into polycations and the simultaneous incorporation of het-eropolyanions. This doping reaction can be oxidative in its t Also Department of Chemistry, Technical University of Warsaw, 00-664 Warszawa, Noakowskiego 3, Poland. nature, as in the case of p~lyacetylene,~or acid-base, as in the case of polyaniline.* Taking into account the large size of heteropolyanions and their limited diffusivity into the bulk of the solid polymer, it is expected that doping will be limited to the surface or near-to-surface layer of the polymer; this has been proven e~perimentally.~ Previously7 we demonstrated that polyacetylene doped with 1Zmolybdophosphoric acid exhibited a high catalytic activity in ethyl alcohol conversion compared to the unsupported crystalline 12-molybdo-phosphoric acid studied under the same experimental conditions.The overall increase in catalytic activity is accom- panied by a significant change in the redox activity over the acid-base one. Polyaniline is another example of a polymer host which can accommodate anions originating from 12-molybdo-phosphoric acid. This can be achieved either by poly- merization of aniline in the presence of H,PMo,,O,, in one step" or by doping, i.e.the protonation of the poly-emeraldine base with H,PMo,,O,, in a two-step pro-cedure.* Our preliminary experiments with H,PMo~~O~,- doped polyaniline have shown that distinct differences exist in the catalytic behaviour of the samples obtained by these two methods." The catalyst prepared by the latter method exhibited much higher activity in the ethyl alcohol conver- sion than the catalyst prepared according to the former method, despite the fact that both had similar surface areas. In order to explain these differences the present study of the physicochemical properties of our catalysts was undertaken. FTIR, XP and EPR spectroscopies as well as electrical conductivity measurements were applied. Catalytic experi- ments were used in this case as an additional method of char- acterizing surface properties of the samples.Experimental Preparation of Polyaniline Protonated with 12-Molybdophosphoric Acid Polyaniline can be rendered conductive either by the oxida- tive or acid-base doping according to the scheme: 1 ti -2e 2 Polyaniline in the form of polyleucoemeraldine can be oxi- datively doped by withdrawal of electrons from its n-bonding system. This oxidation process results in electrical charge introduction into the polymer matrix and its transformation into an organic conductor. If one electron per two polymeric units of polyleucoemeraldine is withdrawn, the polymer reaches the oxidation state of polyemeraldine, which is stable in air both in its conducting polyemeraldine salt form and in its insulating polyemeraldine base form.The base can easily be converted into the salt form by protonation with a suffi- ciently strong acid.12 The anions of the protonating acid are simultaneously incorporated into the polymer, thus neutral- izing the positive charge of polymer chains. As already stated, this doping of the polyemeraldine base with dodecamolybdo- phosphoric acid can be achieved either in a one-step or a two-step process. Both were used in the present study. One-step Procedure Polyaniline protonated with H,PMo, 2040was prepared in a one-step procedure using a modification of the method described in ref.10. In all preparations the amounts of aniline and H,PMo,,O,, were fixed (8.25 and 2.60 mmol, respectively) whereas varying amounts of the oxidant were added (increasing amounts from 1.29 to 6.88 mmol). Typically aniline and H,PMo120,, were dissolved in 50 ml of acetoni- trile. Then ammonium persulfate dissolved in 1.5 ml of H20 was added dropwise. The reaction mixture was kept at room temperature for 24 h with constant stirring. Reaction was ter- minated by pouring the reaction mixture to 250 ml of acetone which caused immediate precipitation of a black powder. The powder was then separated by centrifugation and repeatedly washed with acetone until the filtrate became colourless. Polyaniline-H,PMo, 2040prepared in such a manner con- tains significant amounts of sulfur as determined by elemental analysis.Evidently the HSO, ions created upon the reduction of the persulfate ion in the acidic medium are co- inserted with heteropolyanions into the polymer matrix. These hydrogenosulfate anions can conveniently be removed from polyaniline by prolonged washing with water. Washing results in effective removal of HSO, ions via deprotonation, whereas heteropolyanions remain essentially intact. Because of the HSO, co-insertion mentioned above, the content of heteropolyanions in the polymer matrix must be dependent on the concentration of the oxidant in the reaction medium. More persulfate will create more HSO, and thus fewer sites will be available for the protonation with heteropolyanions.This is indeed the case. The heteropolyanion content in the J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 polymer matrix decreases markedly with increasing oxidant/ aniline ratio (see Table 1). In the subsequent text all samples prepared in the one-step procedure are denoted as SI, (n = 1, 2, 3, . . .). Two-step Procedure In the two-step procedure the polyemeraldine base was first synthesized and then protonated with 12-molybdophosphoric acid. The polyemeraldine base was prepared by condensation-polymerization of aniline in aqueous HCl using (NH,),S208 as the oxidant. The polymerization under these conditions leads to polyemeraldine hydrochloride, which must be then converted into polyemeraldine base by deprotonation with a suitable base (NH,).In a typical prep- aration 10.2 g (0.1093 mol) of aniline was dissolved in 125 ml of 1.5 mol 1-' HCl. Then 12.5 g (0.0548 mol) of ammonium perchlorate was dissolved in 125 ml of 1.5 mol I-' HCl and added dropwise to the aniline-HC1 solution. The reaction was carried out at 0-1 "C for 4 h as recommended in ref. 13. The black precipitate obtained was filtered off, washed with water and subsequently with methanol and diethyl ether until in each washing the filtrate was colourless. The polymer was then vacuum dried to constant mass and then deprotonated with 3% NH, aqueous solution for 2 h. Deprotonated poly- emeraldine was dried to constant mass and washed again with water, methanol and diethyl ether in a Soxhlet appar- atus.This final washing is essential for total removal of all oligomeric species still extant in the system. The polyemeraldine base was then protonated in acetonitrile-HPA solutions. Note that for a given HPA : polyemeraldine ratio in the protonating medium a partition equilibrium of HPA between two phases (polyemeraldine and acetonitrile) is easily estabilished. Thus the protonation level can conveniently be varied by changing the HPA : polyemeraldine ratio (Table 2). The samples pre- pared according to the two-step procedure are abbreviated as SII,(n = 1,2,3...). Characterization of Unprotonated and Protonated Pol yerneraldine The samples of polyemeraldine base and polyemeraldine protonated with 1Zmolybdophosphoric acid were subjected to elemental analysis.Nitrogen, carbon, hydrogen, molyb- denum and sulfur were determined using classical methods. The specific surface area of the obtained powders was mea- sured using the BET method on a Carlo Erba Sorpty 1750 apparatus. Thermogravimetric (TG and DTG) and differential thermal analyses of the samples were carried out with a Mettler Thermoanalyser TA-2 using Al,O, as the standard substance. Table 1 Elemental analysis and conductivity of C6H4.5N(H3PMo12040)yprepared in the 'one-step procedure' involving oxidation of aniline in the presence of 12-molybdophosphoric acid aniline : persulfate empirical formula conductivity sample molar ratio based on N : Mo /S cm-' SI, SI2 SI3 SI4 SI5 1.2: 1 4.8 : 1 6.4 : 1 2.0 : 1 3.2 : 1 c,H,,~N(HPA),,,,,~ C,H4~5N(HPA),,,,3 C6H4,5N(HPA),,l,, C,H4~5N(HPA)o.056 C6H4,5N(HPA)o,075 7.0 x 10-4 1.8 x 1.9 x 1.3 X 1.7 x lop3 a HPA = 12-molybdophosphoric acid.J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 2 Protonation of polyemeraldine base with 12-molybdophosphoric acid ~~ ~ HPA :polyaniline molar ratio in the protonating empirical formula conductivity sample medium'" based on N : Mo /S cm-* SII 0.0025 C6H,,sN(HPA)o,oo24b 8.2 x SII, 0.0050 C,H,,,N(HPA),,,,,, 6.1 x loa6 SII 0.010 C6H4,sN(HPA)O~O077 6.4 x lo-, SII, 0.015 C6H,,5N(HPA)o,o13 3.6 X SII5 0.020 C6H,~SN(HPA)o~o,5 2.3 X SII, 0.025 C6H,~SN(HPA)o,o16 2.8 X lo-, SII , 2.000 C,H,,SN(HPA)o,o,7 7.0 x lo-, a Concentration of the doping solution 0.001 mol 1-'.HPA = H,PMo,,O,, * The electrical conductivity of the protonated polyaniline was measured on pressed pellets (4 t cm-') using a classical four-probe method. Since the conductivity of polyaniline depends on the water vapour pressure in air, the samples were dried in vacuum and the conductivity was also mea- sured in vacuum. FTIR spectra were recorded on a Digilab FTS-60 spec- trometer in transmission mode using KBr pellets. The X-ray photoelectron (XP) spectra were taken on powders of polyaniline base or of polyaniline protonated with H,PMo,,O,, pressed onto a copper mesh. XP spectra recorded on an ESCA 100 (VSW Manchester) using Mg K, radiation of 1253.6 eV under pressure lower than 5 x lo-' mbar. This low pressure was achieved in the case of proto- nated polyemeraldine after 24 h of evacuation at 27 "C. As an internal standard C 1s = 284.5 eV was applied.In the course of recording the spectra the samples were cooled to -123"C to improve the quality of the spectra. EPR spectra were recorded by means of a computer-controlled spectrometer operating in the X-band with 100 kHz modulation. DPPH was used as the g-factor standard (g = 2.0036). EPR spectra of the polyemeraldine base and of polyemeraldine protonated with H,PMo,,O,, at room and at liquid-nitrogen temperatures were recorded. Additionally the EPR measurements were carried out while conducting the catalytic reaction on sample SII, in the specially designed reactor that was placed in the spectrometer cavity at elevated temperatures (175-240 "C) and after annealing the reaction mixture to -170°C.Catalytic Experiments As in the previous st~dies,~,'~ conversion of ethyl alcohol was used as the catalytic test reaction. The experiments were carried out in the same pulse micro- reactor in which 0.3 g of the catalyst mixed with 0.3 g of powdered quartz was placed. Before the catalytic experiments the samples were standarized by heating them for 30 min at 50, 100 and 150°C and for 1 h at 200°C. The sample pre- pared in the one-step procedure was only weakly catalytically active and in order to obtain measurable amounts of pro- ducts it was necessary to carry out catalytic testing at 300 "C. On the other hand the sample obtained by the two-step pro- cedure was much more active and the catalytic test could be carried out at 240"C, ensuring better thermal stability of the polymer matrix.Results The experiments were carried out using three samples: the undoped polyaniline base (PA), polyaniline doped with 21G1 heteropolyacid using the one-step procedure (SI,) the com- position of which corresponded to the formula C6H,~,N(HPA),,,,, (36.9 wt.% of H3PMo,,0,,) and poly- aniline doped with HPA (SII,) obtained in the two-step pro- cedure of the composition C6H,~,N(HPA),,,,, (35 wt.% H,PMo,,O,,). The specific surface areas of samples PA, SI, and SII, were 20.1, 20.2 and 24.7 m2 g-',respectively. The results of TG, DTG and DTA analyses are shown in Fig.1 Both SI, and SII, exhibited a mass loss between CQ. 50 and 150°C characterized by a DTA peak at 100°C. This was interpreted as the desorption of molecules of adsorbed (or possibly occluded) solvent molecules or adsorbed water ii IL I I I I I I I 50 100 200 300 400 500 600 Tl" C iii-4-cad I1 1 I I 50 100 200 300 400 500 600 T/OC Fig. 1 Thermogravimetric (i), differential thermal (ii) and differential gravimetric (iii) analyses of: (a) unsupported H,PMO~,O~~f26.3 H,O, (b)undoped polyemeraldine base, (c) HPA-doped polyaniline in a one-step procedure, SI,, (6)HPA-doped polyaniline in a two- step procedure SII,. Heating rate 10°C min-', sample weight 50 mg. vapour. In this period SI, lost 7.5% of its mass and sample SII, only 4.3%.Both samples exhibited practically constant mass between 200 and 270-280°C, above which temperature slow decomposition of the polymer was recorded. FTIR spectra of the samples prepared in both one- and two-step procedures for a given protonation level are essen- tially indistinguishable. In Fig. 2 IR spectra of samples of the SII series are presented and compared with those of the poly- emeraldine base PA and the crystalline heteropolyacid. With an increasing protonation level the modes originating from heteropolyanions grow in intensity and become dominant for higher protonation levels. "-'Note that the two bands which are diagnostic of Keggin structural units are clearly observed in heteropolyanions incorporated into the polymer : v(M-0,) at 954 cm-' and v(x-0,) at 1067 cm-' [Fig. 2(d)].The v(M-0,) band is not obscured by peaks originat- ing from polyaniline chains, whereas the v(X-0,) band strongly overlaps with polyaniline bands but it can be clearly identified. The most intense HPA band, i.e. v(M-OO,-M) is also present. Coulombic interactions between hetero-polyanions and the protonated polymeric support induce a shift in M-Oc-M vibrations by ca. 10 cm-' [compare Fig. 2(d)and 2(e)]. The changes in the modes associated with the poly-emeraldine chain are essentially the same as observed for polyaniline protonated with inorganic acids such as HCl or H2S04 and correspond to the benzenoid and quinoid sequences of the polyaniline chain.' The doping-induced 1140 cm- 'band characteristic of the charged (protonated) chain increases markedly with the level of protonation.Additional information about the protonation of poly- emeraldine with HPA can be extracted from N 1s XP spectra. Fig. 3 and 4 show deconvoluted N 1s XP spectra of the pris- tine undoped polyemeraldine base and the pristine sample SII,. The deconvolution was carried out assuming in each case the same weighting of Gaussian (75%) and Lorentzian (25%) functions. Under these conditions the binding energy 1494.0i\1300.8 I15rO 829.4 154.2 I\ 1613.1dln I L I I I 400 600 800 -1001 0 1200 1400 1600 wavenumberfcm-' Fig. 2 FTIR spectra of the SII series: (a) undoped polyemeraldine base PA, (b) sample SI12, (c) sample SI14, (d) sample SI16, (e) H,PMo,,O,, J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I,I ,*,,I,, IIII,,,I,.III.II,,,,, ,,.I,,,III~II .,.IIIII,II.II,I,I~III.I,,I,,I,,I,,, 408 403 398 393 EbW Fig. 3 N 1s XP spectrum of polyemeraldine base deconvoluted into the imine (398.4 eV), amine (399.5 eV) and protonated nitrogen atom (401.9 eV) contributions (Eb)of N 1s in the imine state (==N-) has been obtained within a very narrow range of 398.4 If: 0.1 eV. Similarly the Eb values of the amine state (-NH-) have been determined within the range of 399.4 &0.1 eV. The values are higher than those obtained in ref. 19 but very similar to those found by Kang et ~1.~'The third peak at 401.9 eV must be ascribed to a protonated nitrogen species (symbol N').The data con- cerning N 1s spectra of undoped polyemeraldine and the heteropolyacid-doped samples are collected in Table 3. Note that the analysis of the N 1s spectra of poly-emeraldine doped with H3PM~12040 requires first the sub- traction of Mo 3p3/2 from the N 1s envelope. This molybdenum peak is expected at ca. 396 eV as the standard value in the pure metallic state is 392.3 eV.21 The second line of the Mo 3p doublet lies 17.4 eV to higher energy, i.e. way above the N 1s envelope. The parameters of Mo 3p3,2 were practically constant in all the samples investigated (Table 3), independent of their pretreatment. Fig. 5 and 6 show the deconvolution of the 0 1s XP spectra of undoped polyemeraldine and sample SII, as pre- pared.Analysis of all spectra listed in Table 4 allowed us to distinguish between three different states of oxygen atoms. The fractional peak at 530.5 k0.1 eV, which was present only in the case of heteropolyacid-doped polyemeraldine, has been ascribed to oxide species as the one showing the lowest binding energy. This value is very close to frequently quoted data (530.4 eV) for MOO,. 22,23 The second 0 Is peak was characterized within the range 532.05 0.15 eV in all samples investigated. It is ascribed to surface OH- anions and its binding energy is similar to the value of 531.80 eV found in ln(OH), .24*2s The fractional 0 1s peak showing the highest E, (533.55 &0.15 eV) and also appearing in the spectra of all the samples investigated is ascribed to oxygen atoms in IIIIII,II I, (I,,,...IIII.III.,ILIII IIIIIIIIII,II 11,,1,,,,,,,,, 406 402 398 394 EbIeV Fig.4 N 1s XP spectrum of SII, deconvoluted as in Fig. 3 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 3 N 1s XPS data for the polyemeraldine base and polyaniline doped with 12-molybdophosphoric acid proportion (%) sample E, "+)lev -N= PA as prepared PA activated" 401.9 401.2 47 50 SI, as prepared SII, as prepared SII, activated" 400.6 400.5 400.6 401.6 401.6 401.8 33 41 40 SII, after catalytic reaction 400.6 401.7 48 "Activation under the conditions of catalytic reaction. water molecules. The literature data for E, in H20 give 533.20 eV.24.26 Table 4 presents the share of fractional peaks in the total intensity of 0 1s XPS signals for all the samples investigated.Fig. 7 shows the Mo 3d5,,doublet obtained for SII,(two- step preparation) comprising peaks at 235.6 and 232.5 eV. lss~***m*,l IIIII*I!.(II I I I I IIIII1II.III,I.IIIIIII,II,IIII'IIII,,,,,,,,(,,,,,,,,, I 542 537 532 527 Eb/eV Fig. 5 0 1s XP spectrum of polyemeraldine base deconvoluted into peaks corresponding to the oxygen atoms in OH-groups (532.05 eV) and in H,O molecules (533.55 eV) ~I.~III1II~IIIII.I~I~IIIIIIIII~IIIIIIIII~II.IIIII.~IIIIIII.I)III..I.I,I'I., "' 540 535 530 525 Eb/eV Fig. 6 0 1s XP spectrum of SII, deconvoluted into peaks corre- sponding to oxide oxygen atoms (530.55 eV), oxygen atoms in OH-groups (532.05 eV) and in water molecules (533.55 eV) Table 4 0 1s XPS data for the polyemeraldine base and polyanil- ine doped with 1Zmolybdophosphoric acid proportion (%) sample O2-OH-H2O ratio -NH- N+ -NH- :-N- J% (Mo 3P5,2)/ev 42 11 0.9 44 6 0.9 47 20 1.4 397.0 41 18 1.o 396.9 40 20 1.o 396.9 35 17 0.7 396.9 In the case of the one-step preparation (SI,) the analogous values were 235.3 and 232.2 eV.The analytical peak of Mo 3d5,2 is located at 232.6 eV for MOO,, 22923i27 at 232.5 eV for A12(MOO& 28 and Na2Mo0, 2H20,29therefore no Mo is present in oxidation states lower than +6 in the above experiments. Fig. 8 and 9 show the results of catalytic measurements using SI, and SII,. Undoped polyemeraldine (PA) did not exhibit any catalytic activity.The main product of catalytic I I ,,,.,,),, ,.,,,,.I'*".''<,~IIII~.IIIIIIII.II.I(..IIIIIIIIIIIIIII1I(IIII1,IIIIIII 244 239 234 229 EbW Fig. 7 Mo 3d5,, XP spectrum of SII, deconvoluted into a doublet (235.6 and 232.5 eV) 2.5 \ \ \ \ \2.0 \0) dI 0 F\-g1.5 m 5 'c E 1.0 3 0.5 PA as prepared -37 63 PA activated" -72 28 0.1 SI, as prepared 45 32 24 1 5 10 15 20 ene; ., no. ofpulsesSII, as prepared 52 36 12 SII, activated' 66 24 10 Fig. 8 Results of catalytic experiments at 300 "Cfor SI, :x ,ethyl-unreacted alcohol; (---)deficit of alcohol SII, after catalytic reaction 53 33 14 acetaldehyde; 0, as a function of the pulse number. Amounts of products are "Activated under the conditions of catalytic reaction.expressed as the equivalent amount of alcohol. 2104 :::;I I 15 10 15 20 25 no. of pulses Fig. 9 Catalytic experiments at 240 "C for sample SII,: x , ethyl-acetaldehyde; 0,ene; 0,diethyl ether; ., unreacted alcohol; (---) deficit of alcohol as a function of the pulse number. Amounts of pro-ducts are expressed as in Fig. 8. ethanol conversion was acetaldehyde which is produced in the redox-type reaction. The yield of ethylene in the case of sample SI, and ethylene and diethyl ether in the case of sample SII,, products of acid-base type reaction, is low. Note that in our previous studies in which we studied poly- aniline supported H,SiW 120,0the acid-base activity was predominant, acetaldehyde being a minor product.8 No hydrogen was detected in our catalytic experiments, although separate experiments have shown that an amount of hydro- gen corresponding to the amount of acetaldehyde produced (dehydrogenation of ethyl alcohol) could be detected chro- matographically in our apparatus.No hydrogen was detected in the case of unsupported H3PMo120,, and most probably 3.0.-I \ \a-2.5 \P, \tI z \ \5 2.0.-\ t 5 .4-2 1.5-3 G 1.o ,r 0.5 0.1 5 10 15 20 no. of pulses Fig. 10 CatH,PMo,,O,, alytic experiments . Symbols as in Fig. 9. at 240 "C for unsupported J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 hydrogen evolved in the course of acetaldehyde formation is immediately used for the hydrogenation of the carbonaceous deposit, the product of side-reactions (oide infra).Water present in the reaction products gave a diffuse peak in the gas chromatograph that was unsuitable for quantitative determi- nations. The sum of the amounts of alcohol necessary to produce acetaldehyde, ethylene and (in case of SII,) diethyl ether as well as the unreacted alcohol detected in the pulse leaving the reactor was always smaller than the amount of alcohol intro- duced on the catalyst in one pulse (3.6 x lo-, g). A similar effect was observed in the case of polyacetylene' and p~lypyrrole'~ supported H,PMo, 2040as well as polyaniline-supported H,SiW120,, .8 This might be due to the irreversible penetration of alcohol into the bulk of cata- lyst and/or to the oligomerization of ethylene and the forma- tion of coke.This deficit of alcohol was highest in the initial pulses, reaching 80-70% of the introduced alcohol, but rapidly decreased to ca. 50% after 20-25 pulses. The course of the catalytic reaction at 240°C in the case of unsupported H,PMo,,O,, catalyst is shown in Fig. 10. Discussion Tables 1 and 2 show that by changing the preparation condi- tions the amount of heteropolyanions incorporated into poly- emeraldine matrix can be varied to a significant extent. Generally the one-step procedure results in a higher HPA content, but a small overlap of the doping ranges exists for the one-step and two-step methods. It is therefore possible to compare samples with very close nominal HPA contents pre- pared by the two different methods: C,H,.,N(HPA),.,,, (one-step) and C,H,~,N(HPA),~,,, (two-step).In the one-step procedure doping is effected simultaneously with the poly- merization of aniline, and the dopant may be evenly distrib- uted throughout the whole volume of the polymer. On the other hand, in the two-step procedure the doping is necessar- ily connected with the chemisorption of heteropolyacid at the surface and subsequent diffusion into the bulk, which cannot be fast and may stop completely without reaching an even spatial distribution. Hence, the formation of samples with lower contents of heteropolyacid but concentrated mainly on the surface and in the near-to-surface layer is expected. Additional information indicating an uneven distribution of HPA in the two-step series samples can be extracted from the dependence of the conductivity on the heteropolyanion content in the polymer matrix.If the conducting (i.e. protonated) phase is limited to the surface and its penetration into the bulk of the polymer is almost negligible, percolation- type conductivity behaviour is expected with a low perco- lation threshold. This is indeed the case. For the two-step series samples a clear percolation threshold exists for a com- position of ca. C,H,,,N(HPA),~,,, (Fig. 11). Some information concerning the structure of the doped and undoped polymers or at least the chemical structure of the surface layer could be obtained from XPS measurements. First, deconvoluted XP spectra of the undoped poly-emeraldine base contained N 1s peaks not only at 398.4 and 399.5 eV corresponding to the imine (=N-) and amine (-NH-) nitrogen atoms but also a peak at 401.9 eV ascribed to protonated nitrogen atoms (symbol N') (Table 3).The existence of a small number of protonated nitrogen atoms can be ascribed either to incomplete deprotonation in the transformation from polyemeraldine hydrochloride to polyemeraldine base or possibly to the dissociation of chemi-sorbed water in accordance with the 0 1s XP spectrum (Table 4). This spectrum can be deconvoluted into two peaks: one, at 33.55 eV, is characteristic of adsorbed water molecules J. CHEM. SOC. FARADAY TRANS., 1994, VOL.90 -* I -3 YOt 00 0 0 Table 5 product catalyst H,PMo 12040unsupported H,PMo 2040aunsupported doped polyaniline, SII, doped polyaniline, SI 0 0.01 0.05 0.10 Y Fig. 11 Conductivity us. HPA content (y) in the samples C6H4.5N(HPA),: 0,one-step preparation; x ,two-step preparation and the other, at 532.05 eV, corresponds to oxygen atoms in OH -ions. In the strictly stoichiometric polyemeraldine base, the structure of which is shown schematically in the previous section, the molar ratio of -NH- : =N-is equal to unity. Considering the fact that a certain number of imine nitrogen atoms in the polyaniline sample had been protonated we would expect an increase of this ratio beyond unity. However, the opposite is observed (Table 3): the concentra- tion of imine nitrogen atoms exceeds that of amine atoms.A plausible explanation for this is the assumption that in the polyemeraldine base as prepared by the above method (or at least in its near-to-surface layer accessible for XPS measurements) strict stoichiometry is not preserved and there is an excess of quinoid groups with respect to benzenoid groups if one compares it with the stoichiometry of perfect polyemeraldine. Assuming that in the course of partial proto- nation protons were localized on imine atoms we can calcu- late the hypothetical amine : imine nitrogen atom ratio in the sample which is not protonated at all: z = [-NH-]/([=N-] + "'1) = 42/(47 + 11) = 0.72 The values in parentheses refer to the proportions of the par- ticular species in the sample from XPS measurements (Table 3).Doping of the polyemeraldine base with H3PMo,,04, results in an increase in the proportion of protonated nitro- gen atoms with increasing HPA content. Thus the HPA Brernsted acidity is the main source of N+.The protonation occurs preferentially on imine sites since the N+ intensity grows at the expense of the -N= peak. Again, as in the case of the undoped sample, we can determine the oxidation state of the polymer being protonated. For example, in the case of SII,, the observed -NH- :-N= ratio (assuming proto- nation of only imine sites) is: z = 41/(41 + 18) = 0.70 i.e. practically the same as in the polyemeraldine base dis- Yield per pulse calculated per g H3PMo,,04, and expressed as the amount of C,H,OH (g x cussed above.The analogous value calculated for sample SI, is 0.85, indicating that the one-step preparation of doped polyemeraldine results in the formation of a product, the structure of which is closer to that shown in the scheme (vide supra). The structural difference between SI, and SII, does not give rise to any essential difference in their FTIR spectra. The spectra also show that heteropolyanions (structural Keggin units) preserve their identity when they are dispersed in the polymer matrix. Despite the fact that the specific surface areas of both SI, and SII, are similar, their catalytic behaviour is very differ- ent. At 240"C, a temperature that ensures satisfactory thermal stability of the polymer (Fig.l), in contrast to sample SII, sample SI, was catalytically inactive and it was neces- sary to heat it to 300°C in order to obtain a measurable reaction yield. Table 5 shows that the total catalytic activity calculated per g of heteropolyacid for SII, is 5.6 times greater than that of unsupported H3PMo,,04, in comparable conditions. On the other hand, dispersion of heteropolyacid in the one-step preparation (SI,) resulted in a decrease of specific activity to ca. 58% of that of unsupported H3PMo,,04,. This shows that in the SII, catalyst the Keggin units are more accessible and/or more active than in the unsupported acid, probably owing to their higher concentration on the polymer surface.On the other hand, the one-step procedure results in the dis- persion of HPA over the whole volume of the polymer, which in turn leads to a distinctly lower concentration of HPA at the catalyst surface. From the catalytic point of view the most striking property of both catalysts, independent of the differences in their activ- ity, is the very high selectivity to acetaldehyde, the product of the redox-type reaction, while the products of acid-base-type reaction, i.e. ethylene and diethyl ether, are formed with very low selectivity. Table 6 shows that the redox selectivity of polyaniline-supported H,PMo,,O,, is more than three times higher than that found in the case of the unsupported hetero- polyacid and is also much higher than in the case of polyacetylene-and polypyrrole-supported H3PMo 2040.For polyaniline-supported H,SiW ,2040only very weak redox activity was observed.8 This is not unexpected since the molybdenum in the HPA is much more easily reduced than the tungsten. We expected to detect a distinct reduction in the amount of molybdenum in the samples of the catalyst after the catalytic reaction. However, the XP spectra of such samples gave only the XP signals for Mo 3d,,! at 396.9 eV characteristic of MoV* atoms. This could be ascribed either to the real absence of reduced molybdenum at the surface of the working catalyst or, more probably, to the oxidation of reduced Mo atoms when the sample was exposed to air when it was removed from the reactor.Note that any catalytic model describing redox reactions must include oscillations between the higher and the lower oxidation states of the active centre. One can conclude that in the case of ethanol oxidation to acetaldehyde the reoxidation step of surface centres is very fast and spontaneous. used for formation of the given sum of T/"C CH4 C2H6 C2H4 (C2H5)20 CH3CH0 products 240 0.39 0.76 0.32 1.45 320 0.32 0.28 4.67 1.35 2.10 8.12 240 0.93 0.94 6.08 7.95 300 1.18 traces 3.55 4.73 14.Data from ref. ,, 2106 J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 Table 6 Selectivity to particular products after 20 pulses of ethanol (%) catalyst T/"C CH4 C2H6 C2H4 (C2HS)20 CH,CHO ~~ H3PMoI2O4, unsupported 240 -H,PMo,,O,," unsupported 320 3.7 doped polyaniline, SII, 240 -doped polyaniline, SI 300 -doped polyacetyleneb 230 -doped polypyrrole" preheated in He 320 - -25.7 51.7 22.6 3.3 53.5 15.5 24.1 -11.4 11.6 77.0 --25.0 75.0 -16.4 27.8 55.8 -45.5 3.5 51.0 doped polypyrrole" preheated in air 320 - -58.4 13.5 28.1 " Data from ref.14. Data from ref. 7 This problem was studied in a more detailed way by EPR spectroscopy. The EPR signal of HPA-doped polyaniline depends on the content of the heteropolyacid in the matrix. All samples of the SI series exhibited only a narrow signal with g = 2.003, characteristic of the presence of free radicals in the polyaniline matrix. In the SII series the signal de- 2 3 4 5 B. Keita, L. Nadjo and J. P.Haeussler, J. Electroanal. Chem., 1988,243,481. B. Keita, K. Essadi and L. Nadjo, J. Electroanal. Chem., 1989, 259, 127. B. Keita, D. Bouazis and L. Nadjo, J. Electroanal. Chem., 1988, 255, 303. G. Bidan, E. M. Genies and M. Lapkowski, J. Electroanal. pended on the content of HPA, and for low contents of HPA a weak but distinct signal which could be ascribed to MoS+ appeared. However, for higher contents the Mo5+ signal was no longer present. Since SII, was highly protonated, its EPR spectra both in the pristine state and after its use as a catalyst revealed only one signal at g = 2.003 which was ascribed to 6 7 8 Chem., 1988,251,297. G. Bidan, E. M. Genies and M. Lapkowski, J. Chem. SOC.,Chem. Commun., 1988,533. J. Poiniczek, I. Kulszewicz-Bajer, M. Zagorska, K.Kruczala, K. Dyrek, A. Bielanski and A. Pron, J. Catal., 1991,132, 311. M. Hasik, J. Poiniczek, Z. Piwowarska, R. Dziembaj, A. Biel-anski and A. Pron, J. Mol. Catal., submitted. the presence of unpaired spins in the n system of the polymer backbone. A very similar signal, although much weaker, was observed in the pristine polyemeraldine base. Catalytic reaction strongly influences the EPR signal of the catalysts. Joint EPR-catalytic experiments were carried out with SII, in a small reactor situated in the EPR spectrometer cavity. The catalytic reaction took place at 175-240 "C and the EPR measurements were taken at the reaction tem- 9 10 11 12 13 I. Kulszewicz-Bajer, M. Zagorska, A. Pron, D. Billaud and J. J. Ehrhardt, Mater. Res. Bull., 1991,26, 163.A. Pron, Synth. Met., 1992,43,277. M. Hasik, A. Pron, I. Kulszewicz-Bajer, J. Pozniczek, A. Biel-anski, Z. Piwowarska and R. Dziembaj, Synth. Met., 1993, A. G. MacDiarmid and A. Epstein, Faraday Discuss. Chem. SOC., 1989,88,3 17. Y. Cao, A. Andreatta, A. J. Heeger and P. Smith, Polymer, 1989, 55-57,972. perature as well as at -170"C immediately after the reaction was interrupted (without exposing the sample to air). EPR spectra recorded at 240 "C during the reaction exhibited only a signal at g x 2. However, at -170°C a poorly resolved signal of MoS+ (g < 2) became visible. Obviously its shape 14 15 30,2305. J. Poiniczek, A. Bielanski, I. Kulszewicz-Bajer, M. Zagorska, K. Kruczala, K. Dyrek and A. Pron, J. Mol. Catal., 1991,69,223. C. Rocchiccioli-Deltcheff, R.Thouvenot and R. Franck, Spectro-chim. Acta, Part A, 1976, 32, 587; C. Rocchiccioli-Deltcheff, M. Fournier, R.Franck and M. Thouvenot, Znorg. Chem., 1983, 22, was influenced by molecules of the reacting species adsorbed on the catalyst. When air was allowed into the reactor the EPR parameters of the sample changed, giving a better defined signal with g1 = 1.95 and gll = 1.88. Note that the observed g values are different from those of Keggin and close to those observed for MOO, or to those 16 17 18 207. C. Rocchiccioli-Deltcheff, M. Amirouche, G. Herve, M. Four-nier, M. Che and J. M. Tatibouet, J. Catal., 1990, 126, 591. C. Rocchiccioli-Deltcheff, M. Amirouche and M. Fournier, J. Catal., 1992, 138, 445. J. Tang, X.Jing, B. Wang and F. Wang, Synth.Met., 1988, 24, 213. of MoS+-grafted samples.,, This last observation may suggest that the catalyst may undergo some decomposition under catalytic conditions. A detailed understanding of this phenomenon requires more detailed studies involving other spectroscopic techniques such as 31PMAS NMR studies.34 19 20 21 22 23 J. Yue and A. J. Epstein, Macromolecules, 1991,24,4441. E. T. Kang, K. G. Neoh and K. L. Tan, Polym. J., 1989,21,873. K. T. Ng and D. M. Hercules, J. Phys. Chem., 1976! 80,2095. D. D. Sarma and C. N. R. Rao, J. Electron. Spectrosc. Relat. Phenom., 1980,25.R.J. Colton, A. M. Guzman and J. W. Rabelais, J. Appl. Phys., 1978,49,409. Conclusions 24 C. D. Wagner, D. A. Zatko and R. H. Raymond, Anal. Chem., 1980,52, 1445. Protonation of polyaniline with 12-molybdophosphoric acid is a convenient route for preparing a new type of catalyst that is highly selective towards the formation of acetaldehyde in ethyl alcohol conversion. The catalytic activity of this catalyst is much higher in those samples in which the polymerization of aniline and the protonation reaction are separate in time.In this case catalytically active centres are located only on the polymer surface and are more accessible for alcohol mol- ecules. 25 26 27 28 29 30 31 K. Kishi and S. Ikeda, Bull. Chem. SOC. Jpn., 1973,46,341. V. I. Nefedov, D. Gati, B. F. Dzurinskii, N. P. Sergushin and Y. W. Salyn, Zh. Neorg. Khim., 1975,20, 2307. P. Gajardo, D. Pirotte, C. Defosse, P. Grange and B. Delmon, J. Electron. Spectrosc. Relat. Phenom., 1979,17, 121. T. A. Patterson, J. C. Carver, D. E. Leyden and D. M. Hercules, J. Phys. Chem., 1976,80, 1702. S. 0.Grim and L. J. Matienzo, Znorg. Chem., 1975,14,1014. C. Sanchez, J. Livage, J. P.Launay, M. Fournier and J. Jeannin, J. Am. Chem. SOC., 1982,104,3194.R.Fricke and G. Ohlman, J. Chem. SOC., Faraday Trans., 1986, This work was financially supported by KBN (Committee for Scientific Research of Poland), grant no. 20898 91/p01 and p02. 32 33 34 82, 263. G. Centi, J. Lopez Nieto, C. Iapalucci, K. Bruckman and E. M. Serwicka, Appl. Catal., 1989,46, 197. C. Louis and M. Che, J. Phys. Chem., 1987,91,2875. R. Thouvenot, C. Rocchhicciolo-Deltcheff and M. Fournier, J. References Chem. SOC., Chem. Commun., 1991, 1352. 1 M. Misono, Proc. 10th Int. Congr. Catal. Akademiai Kiado, Budapest, 1993, vol. A, p. 69. Paper 4/01411A; Received 9th March, 1994