J. MATER. CHEM., 1991, l(6) 989-996 Correlation of Surface Acidity with Electrical Conductivity of Silica-supported Heteropoly Compounds studied by the Complex- impedance Method Naoto Azuma,* Reiji Ohtsuka, Yoshio Morioka, Hiroko Kosugi and Jun-ichi Kobayashi Department of Applied Chemistry and Materials Technology, Faculty of Engineering, Skizuoka University, Hamamatsu 432, Japan A.c. electrical measurements and impedance analysis have been carried out to characterize silica-supported heteropoly compounds (HPC), viz. 12-molybdophosphoric acid (H,PMo,,O,), 12-tungstophosphoric acid (H,PW,,O,,) and their sodium, potassium, and caesium salts under various conditions of humidity. The absorption of water was found to enhance the electrical conductivity of silica-supported HPC.In H,PMo,,O,, and its salts supported on silica, complex-impedance plots showed one semicircular arc due to adsorbed water. For H,PW,,O, and its salts supported on silica, the plots showed two arcs; these are ascribed to fast and slow relaxation processes of orientation polarization for conductive species, supposedly protons in occluded and adsorbed water, respectively. As the relative humidity was increased, the conductivity increased logarithmically, showing humidity-sensing characteristics. The conduction behaviour was affected by the loaded amount of HPC and the substitution of cations in HPC. The activation energy for electrical conduction of low-loaded heteropoly acid (HPA) on silica was lower than that for bulk HPA. It changed both with the amount of loaded HPA and with the substitution for hydrogen ions in the loaded HPA.The acidity function, H,, of the silica-supported HPC has been examined also in connection with their electrical behaviour. The electrical conductivity increased with increasing acidity. A linear relationship was found between the electrical conductivity and the acidity function. Keywords: Supported heteropoly compound; Acidity; Ionic conductivity; Complex impedance spectroscopy Inorganic ion-exchangers are used as solid electrolytes in many electrical devices, such as chemical sensors and fuel cells. The electrical conductivity of the ion-exchanger is depen- dent both on the concentrations of ionic and electronic conductors involved and on their mobilities. Dissociative ions, such as protons, play an important role in the electrical behaviour, and these are affected by the surface properties and structure. In order to clarify the electricai properties and develop applications to electrical devices, it is very important to elucidate the interactions between the conductive species and the material.Previous studies of chemical sensors based on ion-exchangers, however, have been concerned with the charac- terization of the conduction mechanism, the sensitivity, and the level of electrical resistance. They have been performed mainly with unmodified bulk materials. However, insufficient information is available on surface properties and structures to understand the electrical behaviour.The relationship between electrical conduction and surface states has been studied in only a few cases, and so the dependence of electrical conduction on surface acidity or structure in silica-supported heteropoly acids (HPAs), e.g. 12-molybdophosphoric acid (H3PMo12040)and 12-tungstophosphoric acid (H3PW12040), and their salts (heteropoly compounds; HPC) will be reported here. In recent years, HPCs, have received considerable attention in the hydration, dehydration, and dehydrogenation of organic compounds' -6 since SOH10 catalysts, which were developed for the synthesis of acrylonitrile by Standard Oil of Ohio, had been applied to the oxidation of a$-unsaturated aldehydes.' Some HPA catalysts not only behave as strong acids, but also play roles as oxidizing agents.* Recently, it has been reported that H3PMoI2O4,, and H3PW12040 show high protonic conducti~ities.~ HPCs may be used as chemical sensors in many electrical devices.We have recently studied HPAs, such as H3PW12040, H3PM012040 and their salts, supported on silica in an effort to gain a better understanding of their reactivities and structures."-'* If the HPAs are loaded onto inorganic supports such as silica or alumina, their mechanical and thermal toughness do not increase, but the solubility of supported HPAs in water decreases. Furthermore, it is easy to modify the surface characteristics of supported HPAs by changing the supported amount and the extent of cation exchange for hydrogen ions in the HPA.The purpose of this work is to characterize the electrical properties of silica-supported H3PMo12040, H3PW 12040 and their sodium, potassium, and caesium salts. The efiect of surface modification on the electrical behaviour is also investigated. A.c. electrical measurements and complex impedance analyses are used to construct an equivalent-circuit representation for the silica-supported heteropoly compounds. The relation between electrical behaviour and surface acidity is also investi- gated. This paper also describes several experiments designed to help determine the conduction mechanism in HPCs and to assess their usefulness as humidity sensors. Experimental Preparation of Samples The raw materials used in preparing the specimens were the HPAs 12-molybdophosphoric acid (H3PMo12040) and 12- tungstophosphoric acid (H3PW 12040), obtained from Japan New Metals and the cation-exchanged HPAs.The silica used was AEROSIL-200 (Aerosil Nippon; denoted throughout as SiO2-20O) which had a surface area of ca. 200m2 g-I. Solutions or emulsions of the HPCs were prepared as follows. Potassium salt emulsions of the HPAs were prepared by slow addition of an aqueous potassium carbonate solution to an HPA solution. Sodium salt solutions and caesium salt emul- sions were obtained by a similar procedure. The samples supported on SiO2-2O0 were obtained as described previously. lo For completeness, it is described briefly here. The support was calcined at 773 K for 3 h and then impregnated with the required amount of HPC solution.The resulting solution was dried at 383 K for 24 h. The solid thus obtained was ground and then calcined at 573 K for 3 h. The amount of supported HPC, rn [in mmol HPC g- '(support)], ranged from 0.05 to 0.45. The HPC samples supported on Si0,-200 will be denoted hereafter as H3PMo12040/Si02-200, M,P3 ~,Mo120,0/Si02-200, H3PW 12040/Si02-200 or M,P3 -,,W 12040/Si02-200, respectively, where M stands for sodium, potassium or caesium, and n denotes the number of substituted hydrogen ions. Fig. 1 shows schematically the surface structure of the HPC/Si02-200 sample. The SiO2-2O0 particle is considered to be a spherical body with no micro- pores. The particle size is ca. 14 nm, calculated from the surface area (200m2 g-') and the density (2.2g~m-~) of SiO2-2O0.For the formation of the first monolayer of HPC on the surface of the silica support, a critical value of m= 0.30 can be deduced, using the effective cross-section per HPC molecule." The structures of loaded HPC on SiO2-2O0 support are also shown in Fig. 1. The surface acidity of the sample is affected by the amount and species of the HPC supports. The acidity function of the sample was measured by means of a Hitachi model 340 spectrophotometer equipped with a head-on detector at room temperature. The quartz cell had a light pathlength of 0.5 mm. The sample was dispersed in dry decalin, and dicinnamylidene- acetone (DCA) was used as an acid-base indicator.DCA adsorbed on H2S04/Si02 and Si0,-200 was adopted as the standard for the acid and base colour, respectively. The acidity function was estimated from the changes in intensities of absorption bands at 400 and 560nm. Details have been described previously." Form of Element and Electrode The samples were pressed into cylindrical pellets (1 3 mm in diameter, 2 mm in thickness) at 1000 kgcm-2, and then calcined at 573 K for 1 h. For electrical measurements, vapour- deposited gold electrodes were applied on both surfaces of the samples. The shapes and positions of the evaporated electrodes were controlled by a thin metal mask. The forms of electrodes and elements are shown in Fig. 2. Two electrodes (inner and outer) were made up on a surface of the element, and one electrode on another surface.The outer electrode served as an earthed guard ring in electrical measurements. This form of electrode can eliminate edge and fringe effects caused by leakage electric fields in electrical measurements. Conductive paint (DOTITE D-500, Fujikura Kasei) was used to make the electrical connection between the evaporated Au electrodes and the Cu lead wires, which were connected to a measuring device. The electrical conduction was measured loaded HPC low-loaded HPC high-loaded HPC 0ca. 14 nm Si02Si02 support 0 : HPC molecule J. MATER. CHEM., 1991, VOL. 1 G H / samp e-I I I+ electrical lead L Fig. 2 Construction of electrode and sample between electodes H and L (Fig. 2).The outer electrode G was joined to the ground terminal of an electrical device. Electrical Measurements It is well known that for poorly conducting materials, perma- nent changes in physical and chemical properties, caused by e.g. dielectric breakdown and electrolysis, can be induced by applying a d.c. voltage. Silica-supported HPC samples, used in this study, showed high resistivity in dry atmospheres. Therefore, we used a complex-impedance method' in order to estimate the electrical conductivity of the silica-supported HPC at various humidities. The frequency dependencies of the impedance, 2, and loss tangent, tan 6, were measured at at least nine frequencies between 100 Hz and 100 kHz with a commercial LCR meter (AG-43 1 1, Ando Electric) under an applied a.c.voltage of 1 V. The variation of impedance with relative humidity (r.h.) were measured in the humidity range 30-90% r.h. at 323 K. The change of electrical conductivity with temperature was deter- mined at 60% r.h. and five temperatures between 303 and 343 K. In order to obtain the equivalent values in each measurement, we waited for at least 6 h after having set a sample in a test chamber. The results were obtained automati- cally with a personal computer system (PC 9801 VX2, Nippon Electric), which was connected to the LCR meter by an IEEE- 488 standard digital interface. l4 In the above measurements, we used a commercial humidity generator (ETAC JLH-400- 20, Kusumoto Kasei). Results and Discussion Complex Impedance Analysis of Silica-Supported HPC in the Presence of Adsorbed Water Complex Impedance Diagram of Supported HPC The resistance of silica-supported H3PMo12040 and H3PW12040(rn =0.05) was beyond the limit of our measure- ments (>lo0 MQ).The measurements were made at room temperature after the samples had been dried at 383 K. The bulk conductivity of supported HPA/Si02-200 is inherently low, and the remaining intrinsic surface protons do not Fig. 1 Schematic illustration for supported HPC on silica-200 and contribute to any detectable surface conductivity under the stacking structure of supported HPC dry atmosphere. When the samples had been left to stand J. MATER. CHEM., 1991, VOL. I (ca. 80% r.h.) for some time before measurements, the resist- ance of the samples became low enough to make our measure- ments possible.The electrical conductivity of supported HPA/Si02-200 is enhanced by the water molecules in the ambient atmosphere. Complex-impedance analysis was applied to elucidate the electrical behaviour of several of the samples. The complex impedance, Z(o),of a material at an applied angular frequency can be represented as follows: Z(o)=Z’(o)+jZ”(w) (1) where Z’(w)and Z”(o)stand for the real part (resistance) and the imaginary part (reactance) of the impedance, respectively. Plots of Z’(o)us. Z(o)suggest one or more possible equival- ent circuits leading to the circuit parameters, i.e. resistance Iand cond~ctance,’~ for a system. Fig.3 shows the typical complex-impedance plots for the 0.05 mmol g- ‘(support) H3PMo ,,0,,/Si0,-200 element as functions of the angular frequencyf=o/2n in 30, 40, 50, and 60% r.h. at 323 K. A computer program was used to fit the best arc to the data, which is shown by the solid lines in Fig. 3. These profiles are semicircular arcs. The impedance of H3PMo12040/Si02-200 decreased with increasing relative humidity. The electrical behavour of H3PMo12040/Si02-200 was found to be closely related to the humidity. Fig. 4 shows the complex-impedance plots for 0.05 mmol g -‘(support) H3PW120,0/Si02-200 in 50% r.h. at 323 K. The diagram comprises two neighbouring arcs, which are attributed to fast and slow relaxation processes of orientation polarization for conductive species, respectively.In the low-humidity region, the high-frequency spur in Fig. 4 diminished or disappeared 1.5 c:2 1.0> ru 0.5 0 0.5 1.0 1.5 2.0 2.5 Z ’/MQ Fig. 3 Complex-impedance plots for supported H3PMo,,0, on silica-200 of 0.05 mmol g -‘(support) measured at various humidities at 323 K. 0,30; 0,40; A,50; 0,60% r.h. 10 1 5 10 15 20 25 Z’jhnQ Fig. 4 Complex-impedance plot for supported H3PW,,0, on silica- 200 of 0.05 mmol g-’(support); relative humidity was 50% at 323 K 99 1 from the complex-impedance diagram. The low-frequency spur, however, prevailed at all humidities. The loci of the complex-impedance plots exhibited significant humidity dependence. For the higher-loaded HPA on silica, there appeared to be another low-frequency tilted spike at the right- hand sides of the arcs in Fig.3 and 4. The spike may be interpreted as representing the specimenlelectrode interface process. As shown in Fig. 3 and 4, the loci of complex-impedance plots are not complete semicircles. The loci in Fig. 4 cross each other. If the electrical behaviour of specimen HPA/Si02- 200 involves a single relaxation process of orientation polariz- ation of the conductive species and the electrodes are complete blocking electrodes, the complex-impedance diagram should be a complete ~emicircle’~ and the spike should be a perpen- dicular line to the Z’(o) axis. Therefore, in this case, there may exist either another relaxaiton process of orientation polarization or some other conductive species responsible for the electrical conduction in HPA/Si02-200.The causes of the distributed relaxation processes and the appearance of the tilted spike in the complex-impedance plots have not yet been identified. One plausible cause is that the supported HPA crystalline sizes are distributed around the most probable size, since HPC molecules tend to aggregate in an earlier stage of loading, even under low The coarse surface of the sample and the specimen/electrode interface cause the current inhomogeneity in the a.c. electrical measurement. l6 Such a possibility is supported by the following experimental result. When measurements were made on uniform films of highly dispersed HPA particles prepared by the sol-gel method, the loci of the complex-impedance plots were complete semicircles.These results will be reported in detail elsewhere. Similar results were also obtained for the complex-impedance plots of other samples of H3PMo12040/Si02-200 and H3PW12040/ SiO2-2O0 with different degrees of loading and for cation- exchange ones (M,H3 -,PMo12040/Si02-200 and M,H3 -,PW 12040/Si02-200). The nature of absorbed water in HPC/Si02-200 is classified into four categories: occlusion as pseudo-liquid into the HPC m~lecules;’~adsorption as a multilayer on the surface of HPC; adsorption as a multilayer on the SiO2-2O0 support; and condensation as a liquid at the contact zone of SiO2-2O0 particles (cf. Fig. 1). Under these experimental conditions, the last form of water is negligible, since condensation as a liquid at the contact zone of SiO2-2O0 particles does not occur below 90% r.h., which can be calculated by the Kelvin equa- tion and the meniscus radius at the contact zone.Furthermore, it was found that the resistance of silica support (SiO2-2O0) without HPC, in which the absorbed water exists only in the form of adsorption as a multilayer on the SiO2-2O0 support, was much higher than that of silica-supported HPC by ca. 3 orders of magnitude over these experimental humidity ranges. For electrical conduction, water adsorption on the SiO2-2O0 support seems to be less pronounced in HPC/Si02-200 samples. Thus, it is concluded that both the water occlusion as pseudo-liquid into HPC molecules and the water adsorp- tion as a multilayer on the surface of HPC are pronounced for ionic conduction of HPC-supported SiO2-2O0 sample under these experimental conditions.The arcs in Fig. 3 and 4 are correlated with the electrical behaviour of the above two absorption methods of water. The change of water content in variously loaded HPA/ SiO2-2O0 samples was investigated vs. humidity at 323 K. It has been reported for HPA that a maximum of 30 water molecules can be contained per Keggin unit,’* made up of [MO,,PO~~]~-or [W12P040]3- anions.lg At low humidity (30% r.h.), the water content was ca. 1 water molecule per Keggin unit for any loaded HPA/Si0,-200 specimen. With an increase in relative humidity, the water content was found to increase gradually.It reached a maximum of ca. 18-30 water molecules per Keggin unit at 90% r.h. For the same amount of loaded HPA on SiO2-2O0 support, the effect of occluded water in H3PW 12040/Si02-200 is more significant than that in H3PMo12040/Si02-200 since the growth of HPA crystal in H3PW12040/Si02-200 exceeds that in H3PMo12040/Si02-200,as shown by the X-ray diffraction patterns of these samples. Furthermore, in a highly loaded H3PMo,2040/Si02-200 sample, in which the growth of the supported H3PMo12040 crystal and the amount of occluded water is high compared to the sample in Fig. 3, there appeared another high-frequency spur at the left-hand sides of the arcs in Fig. 3. Thus, the high-frequency spur in Fig.4 may be linked to the behaviour of the occluded water in the H3PW12040 crystal in H3PW12040/Si02-200. The low-frequency spur in Fig.4 and the arcs in Fig. 3 indicate the behaviour of the adsorbed water on the surface of HPA. Equivalent Circuit for Supported HPC We suggest a total equivalent circuit representation of H3PMo12040 and H3PW12040/Si02-200 as depicted in Fig. 5 and 6, respectively. The complex-impedance diagrams are also shown. The equivalent circuit associated with loop A in these figures consists of resistance R,, and capacitance Cwl, rep-resenting the supported HPC which contains some adsorbed water in HPC/Si02-200. Loop A in Fig. 6 consists of resist- ance Rw2 and capacitance Cw2,representing the HPC which contains some occluded water in M,H3 -,PW 12040/Si02-200.Spike B is associated with the electrical process of the elec- trode/specimen interface and is represented by constant phase- angle impedance 2, and capacitance for electrode polarization Ci.?The electrode/specimen interface process in the a.c. electri- cal measurement seems to be less pronounced in this frequency range and for this applied voltage since the impedance of the samples did not change with time and spike B only appeared for the element of higher-loaded HPA on silica. Thus, we will not consider the specimen/electrode interface process in the following paragraphs. The impedance 2 of the equivalent circuit for M,H3 ~,PMo12040/Si02-200 (Fig. 3) is given by where o is the angular frequency, R,, and CW1or cb are the resistance and capacitance of the adsorbed water on the surface of supported M,H3 -,PMo120q0 crystal or silica sup- port, respectively.In this equation, we do not consider the bulk resistance of the silica support. The contribution of resistance Rb, the bulk resistance, to the a.c. electrical process can be ignored in these systems because the silica support is an insulator. Rewriting eqn. (2) leads to eqn. (3) =(Rw1/2)2 (3)(2'-17,,/2)2 +zr2 This equation shows that the impedance diagram of H3PMo12040/Si02-200becomes a semicircle of radius RW1/2. The loci of Fig. 3 are approximately described by the above equation. From the above relationships, we obtained t Impedance Z, is expressed by: 2, =K,W-~[cos(pz/2)-j sin(pn/2)] where w is the angular frequency, K, and p are independent of o and lpl<1.18 J.MATER. CHEM., 1991, VOL. 1 t-IoopA -+ spikeB -Fig. 5 Equivalent-circuit representation and impedance plot of M,H, -,PM0,,0,~/Si0~-200. R, ,=resistance for adsorbed water; CW1=capacity for adsorbed water; Rh =bulk resistance; Cb =capaci- tance of silica support; Z, =constant-phase angle impedance; Ci = specimen/electrode interfacial capacity Rw2 Rw1 cb I I Fig. 6 Equivalent-circuit representation and impedance plot of M"H3 -,P~Wl,0,,/Si0,-200. -R,, =resistance for -adsorbed- water; CW1=capacity for adsorbed water; R,, =resistance for condensed water; C,, =capacity for condensed water; Rb=bulk resistance; Cb= the capacitance af silica support; Z, =constant phase-angle impedance; Ci specimen/electrode interfacial capacity the resistance, capacitance values and conductivity of the system.$ The impedance 2 of the equivalent circuit for M,H3 ~,PW,2040/Si02-200 (Fig. 6) can also be obtained.It $ Eqn. (2) indicates that the resistance values of the system are derivable from the circular-arc intercepts on the Z axis; the capaci- tance values (the sum of CW1 and C,) can be derived from the frequency (f) at the peaks of the circular arcs (w,) and the resistance (R)by using the following relation, which can be derived also from eqn. (1): = /c(cwl +ch) Rwll The d.c. conductivity of the sample, 0, can be obtained from the resistance by correcting for specimen geometry.J. MATER. CHEM., 1991, VOL. 1 is expressed as follows: z=CRwl+Rw2+j~RW1Rw2Kwl+ Cw2)l/ c1 -W2RwlRw2(CwlCw2 + CbCwl’ cbcw2) +jo(CwlRwl +Cw2Rw2 +CbRwl + CbRw2)l (4) where o is the angular frequency, Rwl or Rw2 and Cwl or Cw2 stand for the resistance and capacitance of surface- adsorbed water or occluded water in crystalline M,H3 -,PW12O40 in the M,H, -nPW12040/Si02-200 speci- men. Cb is the capacitance of the silica support. In this case, although the impedance shows a more complex form com- pared with eqn. (2), the complex-impedance diagram for M,H3 -,PW 12040/Si02-200 can be explained. The resistance and capacitance for H3PW12040/Si02-200 can also be obtained in the same way as for the H3PMo12040/Si02-200 system above.Typical parameters were Rwl= 17.2 MR, Cwl=23.1 pF, Rw2=5.4 MQ Cw2=4.9 pF for H3PW12040/Si02-200 (rn= 0.05, 323 K, 80% r.h.). In the humidity range 30-90% r.h., the conductance of HPC/Si02-200 samples was almost constant. The resistances, however, changed with relative humidity. In order to compare the M,H3 -,PW 12040/Si02-200 specimen with the M,H3 -,,PM0~~0~~/Si0~-200 one, we will consider only the low-frequency spur A in Fig. 5 and 6. Variation of Electrical Conductivity of Supported HPC with Humidity Fig. 7 shows plots of resistivity us. relative humidity for the H,PM0~,0~~/Si0~-200samples as a function of amount supported. As the relative humidity increases from 30 to 90% r.h., the magnitude of the resistivity decreases logarithmi- O\ -lo6 lo5 -0 E q4 -.$I0.-CI v).-v)L lo3 --\lo2 \ 0 10’ -30 40 50 60 70 80 93 relative humidity (YO) Fig.7 Plots of resistivity against relative humidity for the as-prepared H,PM0,,0~~/Si0~-200with various loadings of H,PMol,04, meas-ured at 323 K. Values of mlmmol g-’(support): 0,0.05; A, 0.1; 0, 0.25; V, 0.45 993 cally. The behaviour of the electrical conduction depends upon the amount of H3PMo12040 supported. The electrical conduction increases with an increase in the amount of loaded HPA. The electrical behaviour of highly loaded (rn=0.45) H3PMo12040/Si02-300 is different from that of other low- loaded H,PMo12040/Si02-200 samples. In a previous study,12 we clarified by EPR the difference in the structures of loaded H,PMO,~O~~ silica supports between high- on loaded and low-loaded ones (cJ: Fig. 1).The H~PMo~~O~~ molecules tend to aggregate in an early stage of loading, owing to the network formation of H3PMo12040 and/or the stacking of on themselves. The difference in the electrical behaviour of highly loaded H,PMo 12040/Si02-200, compared with other loaded samples, is due to the structural difference in HPC/Si02-200. Similar results were also obtained for the other HPC/Si02-200 samples. Observation of conductivity changes with temperature is essential for an understanding of conductivity phenomena. When electrical conductivity, 6, is supported by ionic charge carriers, it varies approximately exponentially with the absol- ute temperature, T,” i.e.CT =go exp(-E/RT) (5) where R is the gas constant, E stands for the activation energy, and go is a constant. Fig. 8 shows the Arrhenius plots for electrical conduction in H,PMo12040/Si02-200 at various loading levels. Other silica-supported HPC plots gave negative temperature coefficients similar to those in this figure. From these results, it was concluded that ionic conduction occurs in the HPC/Si0,-200 system. Dissociative protons produced in adsorbed water may be taken as charge carriers in the adsorbed water-HPC/Si02-200 system, since bulk HPA is a good protonic condu~tor.~ Table 1 shows the activation energies for the electrical 0-0-0-0-0 I I I 1 1 2.9 3.0 3.1 3.2 3.3 103 K/T Fig.3 Arrhenius plots for the electrical conductivities of H3PMo,,040/Si0,-200. Values of m/mmol g-’(support): 0,0.05; A, 0.10; 0,0.25; V,0.45 Table 1 Activation energies for the electrical conduction of supported H3PMo12040 on silica-200 sample m"/mmol g- '(support) Eb/kJ mol-' 3PMo 1Z040 0.05 7.1 0.10 9.2 0.25 16.3 0.45 16.5 H3PM012040 bulk 15.5' J. MATER. CHEM., 1991, VOL. 1 1o8 1 lo7 -5 56210 -.-c. v).-v)E -lo5 0 lo4 t \ 30 40 50 60 70 relative humidity (%) The measuring temperature range is 303-343 K in 60% r.h. " m= amount of HPC supported. E =activation energy; error =ca. +1 kJ mol-'. 'From ref. 9. conduction obtained from Fig. 8. The activation energy for bulk H3PMo12040 is also shown.The activation energy for rn=0.05 and 0.10 is lower than that for bulk H3PMo12040. The activation energy for the electrical conduction, however, increases with the amount of loaded H3PMo12040, becoming closer to the activation energy for bulk H3PMo12040. The reason for the lower activation energy for electrical conduction in lower-supported HPA compared to that in bulk HPA is not yet clear. Since the humidity sensing properties of various loaded H3PMo12040/Si02-200 specimens show the same SiO2-2O0 (cf: Fig. 7), a charge in the conductive species responsible for the electrical conduction process is unlikely in this system. Thus, the difference of the activation energy is ascribed to the difference of the surface-stacking structure of supported H3PMo12040 on silica.There may arise a potential barrier for the electrical conduction with increased loadings of HPA on Si02-200. Table 2 lists activation energies for the electrical conduction of supported H3PMo12040, Na3PMo12040. K3PMo12040, and on SiO2-2O0. Since the activation energies for the H3PMo12040/Si02-200 series vary with the loaded amount, we include the data for rn=0.05 in this table. The activation energy is changed when cations are substituted for hydrogen ions in HPA. Replacement of the cation in HPC by H, K, and Cs ions increases the activation energy for electrical conduction in HPC/Si02-200 in this order. The lowest activation energy for Na3PMo 12040/Si02-200 com- pared to other cation-exchanged HPC/Si02-200, may be due to the participation of some other conductive species, such as Na' ions, in the electrical conduction, since Na3PMo12040 is a hydrated, water-soluble ~alt.~.~~ Correlation between the Electrical Conductivity and the Acidity Function H,, of supported HPC Fig.9 shows plots of resistivity us. relative humidity for silica- supported K,H3 -,PM012040 (n= 1-3) samples. The electrical conductivity decreases with an increase of the number, n, of cations substituted for hydrogen ions in H,PMO,~O~~. A similar trend was observed for the Cs,H3 -nPMo12040/Si02- tendency except for high-loaded (rn=0.45) H3PMo12040/Fig. 9 Plots of resistivity us. relative humidity for m =0.05 mmol g-'(support) HPC supported on SiO2-2O0 measured at 323 K.0, KH,PMol,0,0/Si02-200; 0, K,HPMol,040/Si0,-200; A, K3PMo,z0,,/Si02-200 sample, however, there was not a clear correlation between the conductivity and the number of cations substituted for hydrogen ions. Fig. 10 shows plots of resistivity us. relative humidity for silica-supported Na3PMo12040, K3PMo12040 and samples. The behaviour for the H3PMo12040/Si02-200 sample is also shown in this figure. When the hydrogen ions in HPA were completely substituted with sodium, potassium, or caesium cations, the electrical conductivity of the specimens decreased in this order. Further- I o9 I O8 5 lo7 5> c.-.-> -6.5 10 200 sample examined. For the Na,H3 -,PM0,~0~~/Si0~-200 Table 2 Activation energies for the electrical conduction of supported HPC on silica-200 ~~~ sample E"/kJ mol-' H3PMo,2040/Si02-200 7.1 Na3PMo,,0,,/Si0,-200 3.6 K3PMo 12040/Si02-200 9.2 Cs3PMo 12040/Si0,-200 10.4 The amount of HPC supported is 0.05 mmol g- '(support).The measuring temperature range is 303-343 K in 60% r.h. " E =acti-vation energy; error =ca. f1 kJ mol- '. I o5 \\ 6 I o4 30 40 50 60 70 relative humidity (%) Fig. 10 Plots of resistivity vs. relative humidity for m=0.05mmol g-'(support) HPC supported on Si0,-200 measured at 323 K. 0, H3PMo12040/Si0,-200; A, Na3PMo120,0/Si0,-200;0,K,PMo1,0,,/Si0,-200; V, Cs3PMo1,0,,/Si0,-200 J. MATER. CHEM., 1991, VOL. 1 more, the electrical behaviour of Na3PMo12040/Si02-200 differs from that of H3PMo12040/Si02-200, K3PMoI2O40/ SiO2-2O0 and Cs3PMo12040/Si02-200.The surface acidity, which is a measure of the tendency to donate the proton, of HPC/Si0,-200 depends both on the number of cations substituted for hydrogen ions and the species of substituted cations. Previously, we have presented a quantitative evaluation of the acidity of HPC supported on silica (Si0,-200) using the acidity function, Ho.ll The value of Ho for the H3PMo12040/Si02-200 samples was found to increase with the amount of support. It was also affected by the substitution of hdyrogen ions in HPA. Thus, it seems that the change in electrical conduction in HPC/Si02-200 with the number and species of cations substituted for hydrogen ions may correspond to the change in the surface acidity (Br~nsted acidity) of HPC/Si02-200.We have assumed that Ho is closely related to the electrical conduction of ion- exchangers such as HPC. Hereafter, we discuss the surface electrical conductivity for HPA/Si02-200 samples with various loadings in relation to the Ho value. Fig. I1 shows the electrical conductivity of supported H3PMo12040/Si02-200 against Ho. The solid line was drawn to fit the plots. The figure shows a good linear relationship between the electrical conductivity and Ho. The electrical conductivity increases with increasing Ho. Thus, it is con- cluded that the surface acidity affects the electrical conduction of supported HPA. It is seen from Fig. 11 that the relation between the surface acidity and the electrical conductivity can be empirically formulated as follows: logo=-A&+B (6) where A and B are constants.The slope A of the solid line in Fig. 11 may reflect the mobility of electrical conductors (protons). The constant B may reflect the concentration of electrical conductors. Fig. 12 shows plots of electrical conductivity against the acidity function Ho for various cation-exchanged MnH3-nPMo12040/Si02series. In this case, it is also clear that the surface acidity of HPC/Si02-200 has an influence on the electrical conductivity; plots for H3PMo12040/Si02-200, K,H3 -,PM0~~0~~/Si0~-200, and Cs,H3 -nPMo12040/Si02- 200 are superimposed on each other with a slope of -1.9, / P0 A/ 0/ I 6" 1ii4 r 5 r I G .-:lo 5 .-c0 7J 0 1Ci6 I ci7 -2 -3 -4 acidity function H, Fig.12 Relation between the electrical conductivity for HPC/Si02- 200 in 60% r.h. at 323 K and the acidity function H,. 0, H,PMo O4,/Si0,-2O0; A, Na,H, -,PMo ,O4,/SiO,-20O; 0, K,H, ~,PMo,,04,/Si0,-20~ V,Cs,H, -,PMo,,04,/Si0,-200 whereas a plot for Na,H3 ~,PMo12040/Si02-200 gives a sep- arate line with a slope of -0.64. Fig. 13 shows a plot of the electrical conductivity us. Ho for the M,H3 -,PW 12040/Si02- 200 series. The behaviour shows the same tendency as that for the M,H3 -,PM0~~0~~/Si0~-200 series. Plots for K,H -,PW 040 and Cs,H -,P W 2040/Si02- H PW 040, 200 fall approximately on a straight line with a slope of -2.0. The behaviour of the Na,H3 -,PW 12040/Si02-200 sample, however, differs from other samples with a slope of -0.76./ / / 10" -Ln -3n -4-2 -3 -4 acidity function H, acidity function H, Fig. 13 Relation between electrical conductivity for HPC/Si0,-200 Fig. 11 Relation between electrical conductivity for H,PMO,,O~~/ in 60% r.h. at 323 K and the acidity function Ho. 0, H3PW12040/Si0,-200 in 60% r.h. at 323 K and the acidity function H,. Values SiO2-2O0; A, Na,H3~,PW,,04,/Si0,-200;0,K,H, -nPW12040/ of m/mmol g-'(support): 0,0.05; A, 0.1; 0,0.25; V, 0.45 SiO2-2O0; V, Cs,H, -,PW,2040/Si02-200 The electrical conductivity of sodium cation-substituted HPC/Si0,-200 is larger than that of other cation-substituted HPC/SiO,-200, when compared at the same H, value.The slopes of plots for sodium cation-substituted HPC/Si02-200 in Fig. 12 and 13 are small, compared with the slopes for other HPC/Si0,-200. In sodium cation-substituted HPC/ Si0,-200, it is assumed that the substituted cations can dissolve into the adsorbed water and participate in the electri- cal conduction. Such a possibility is supported by the fact J. MATER. CHEM., 1991, VOL. 1 7 Standard Oil (Sohio), Jpn. Pat., 10 308, 1960. 8 I. V. Kozhevnikov and K. I. Matveev, Appl. Catal., 1983,5, 135. 9 0.Nakamura, T. Kodama, 1. Ogino and Y. Miyake, Chem. Lett., 1979, 1, 17. 10 R. Ohtsuka, Y. Morioka and J. Kobayashi, Bull. Chem. SOC. Jpn., 1989, 62, 3195. 11 R. Ohtsuka, Y. Morioka and J. Kobayashi, Bull. Chem. SOC. Jpn., 1990, 63, 2071.12 R. Ohtsuka and J. Kobayashi, Bull. Chem. SOC.Jpn., 1990, 63, 2076. 13that the electrical conductivity of Na,H, -,,PMo~~O~,/S~O~- 200 is affected by the number of hydrogen ions which have been replaced by sodium. It decreased with decreasing number of n (cJ Fig. 12 and 13). Furthermore, the electrical conduc- tivity for HPC/Si02-200 substituted by low concentrations of sodium cations approached that for HPC/Si02-200 exchanged by other cations. References K. Urabe, K. Fujita and Y. Izumi, Shokubai (Catalyst), 1980, 22, 223; Y. Onoue, Y. Mizutani, S. Akiyama, Y. Izumi and H. Ihara, Shokubai (Catalyst), 1976, 18, 180. T. Okuhara, A. Kasai, N. Hayakawa, Y. Yoneda and M. Misono, J. Catal., 1983, 83, 121. M. Ai, J. Catal., 1981, 71, 88.M. Akimoto, Y. Tsuchida, K. Sat0 and E. Echigoya, J. Catal., 1981, 72, 83. H. Hayashi and J. B. Moffat, J. Catal., 1982, 77, 473. Y. Izumi, R. Hasebe and K. Urabe, J. Catal., 1983, 84, 40%. S. H. Chu and M. A. Seitz, J. Solid State Chem., 1978, 23, 297; J. G. Thevenin and R. H. Muller, J. Electrochem. SOC.,1987, 1-34, 273. 14 ANSI/IEEE Std, 488 Standard Digital Interface for Programm- able Instrumentation, 1975, revised 1978, 1980. 15 M. Watanabe. S. Oohashi, K. Sanui, N. Ogata, T. Kobayashi and Z. Ohtaki, Macromolecules, 1984. 17, 2908. 16 P. K. Botteleberghs, Solid Electrolytes, General Principles, Characterisation, Materials, Applications, ed. P. Hagenmuller and W. Van Gool, Academic Press, New York, 1978, p. 145; M. Watanabe, Dodensei Kobunshi, ed. N. Ogata, Kodansya, Tokyo, 1990, p. 40. 17 G. M. Brown, M. R. Noe-Spirlet, W. R. Busing and H. A. Levy, Acta Crystallogr., Sect. B, 1977, 33, 1038; J. B. Moffat, Poly-hedron, 1986, 5, 261. 18 G. A. Tsigdinos, Ind. Eng. Chem. Prod. Res. Dev., 1974, 13, 267. 19 J. F. Keggin, Nature (London), 1983, 131,908; H. A. Lavy, F. A. Agron and M. D. Danford, J. Chem. Phys., 1959, 30, 1486. 20 L. Glasser, Chem. Reu., 1975, 75(1), 21. 21 M. Misono, N. Mizuno, K. Katamura, A. Kasai, Y. Konishi, K. Sakata, T. Okuhara and Y. Yoneda, Bull. Chem. SOC. Jpn., 1982, 55, 400. Paper 1/02400K; Received 22nd May, 1991