摘要:

J . Chem. SOC., Furaday Trans. I, 1989, 85(4), 977-990 Simultaneous Alternating Current Impedance/ Electron Spin Resonance Study of Electrochemical Doping in Polypyrrole Andrew M. Waller and Richard G. Compton* Physicd Chemistry Laboratory, Oxford University, South Parks Road, Oxford OX1 3QZ The electrochemical doping of polypyrrole has been studied using simultaneous a.c. impedance and in-situ electrochemical e.s.r. measure- ments. A quantitative estimate of the polypyrrole conductivity as a function of doping level has been obtained and the corresponding change in the polaron concentration measured by in-situ electrochemical e.s.r. It was found that at low doping levels there was a good correlation between these two properties. It is suggested that the polaron may be responsible for the conduction process under such conditions.At higher doping levels conduction would appear to involve the bipolaron. The conducting organic polymer polypyrrole (PP) has been widely studied since its electrical properties were first reported by Diaz and Kanazawa. Of particular interest in relationship to the conductivity of PP films is the mechanism of the transport of electrical charge. This has been postulated as occurring via the movement of charge carriers along the polymer chain. The two possible charge carriers within PP are the polaron and the bzj~olaron.~~~ These are shown schematically in fig. 1 ; the polaron can be considered as a radical cation associated with a local distortion of the PP lattice from a ‘ benzenoid ’ to a ‘ quinoid ’ form. The bipolaron is the corresponding dicationic species.The formation of such localised species is favoured since the energy gained in going from the uncharged to charged state is greater than the energy required for the local distortion of the l a t t i ~ e . ~ Theoretical calculations of the energetics of polaron and bipolaron formation in PP have suggested the formation of a bipolaron to be favourable by 0.45 eV, as compared to the formation of two p01arons.~-~ This has led to suggestions that the fundamental charged species responsible for charge transport in PP is the bip~laron.~-~ Some experimental evidence to support this suggestion has come from optical absorption studies8-l0 and from e.s.r. The latter is discussed next. The use of e.s.r. to study doped PP has been utilised by several workers, through the detection of the polaron, which has spin S = +f.E.s.r. is blind to the bipolaron which has spin S = 0. The presence or absence of an e.s.r. signal has thus been considered to be indicative of whether charge transport in PP occurs via polarons or bipolarons. The earliest such e.s.r. study of the doping of PP involved the chemical doping with oxygen of PP films.” The uptake of oxygen into PP films was measured by the change in weight of the film, and conductivity measurements and e.s.r. spectra were recorded on the doped film. It was observed that the e.s.r. spectra increased in amplitude, with no change in linewidth, until the oxygen uptake was 0.3%. Further oxygen uptake lead to a decrease in the signal amplitude, until at uptakes greater than 0.6% no e.s.r.signal was observed. The conductivity of the same PP films was observed to increase, from to R-l cm-l, with oxygen uptake to 0.3 YO. Further uptake produced no change in the conductivity. The change of both e.s.r. signal intensity and the conductivity with oxygen uptake can be seen in fig. 2. The observed e.s.r. behaviour upon oxygen doping was explained as being not due to the formation of polarons and subsequently bipolarons. 977978 Electrochemical Doping in Polyp yr role r 1 L Fig. 1. The proposed charge carriers in PP: ( a ) the polaro I -------- -O- H (b) the bipolaron. I I 1 Q uptake/% wt. Fig, 2. E.s.r. signal intensity ( S ) and conductivity (a) results for oxygen-doped PP, from ref. (I). Instead the formation within the polymer chain of radical defects, of the type shown in fig.3, was suggested. In the same studyll the authors observed that electrochemically cycled PP films exhibited little or no e.s.r. signal. This result combined with the spins observed in the oxygen uptake experiment being attributed to defects in the PP led the authors to conclude that the principal charge carriers in doped PP were spinless, and therefore bipolarons. The first in-situ electrochemical e.s.r. study of PP doping was described by Ka~fmann,~ who studied the evolution of the e.s.r. signal in PP coats of thickness 200 nm, after the injection of charge into the polymer. Kaufmann interpreted the e.s.r. signal results for the oxidation (of an undoped PP coat) or the reduction (of a doped PP coat) as giving rise to an initial excess concentration of polarons, which then decayed over a period of 103-104 s.Kaufmann interpreted this as due to a recombination of polarons intoA . M. Waller and R. G . Compton 979 N N H H Fig. 3. The defects in PP chains formed upon oxygen doping as proposed in ref. (1). bipolar on^,^ and was presented as strong evidence that conduction in PP was occurring via a mechanism involving the latter. Kaufmann's assertion that the bipolaron is the basic charged species in PP, and hence responsible for charge transport in the doped form, was re-examined by Genoud. 1 2 7 1 3 Genoud and co-workers performed a quantitative e.s.r. study of the formation of polarons and bipolarons upon electrochemical doping. Both steady-state and transient experiments were performed upon PP using in-situ electrochemical e.s.r.experiments. The e.s.r. signal was monitored as a function of time, when the electrode potential was stepped, and as a function of the charge injection in steady-state experiments. Potential steps from undoped to doped PP on coats of identical thickness to those studied by Ka~fmann,~ revealed the initial generation of polarons, which were immediately further oxidised to bipolarons, as evidenced by a fast change in the e.s.r. signal over a timescale of ca. 10 s. Corresponding behaviour was observed in the electrochemical reduction of doped PP. Such results were not in agreement with the observation of Ka~fmann.~ In his experiments the evolution of the e.s.r. signal was over a timescale 100-1000 bigger than this. Under conditions where a PP coated electrode was held at a fixed potential, or open-circuited, Genoud', reported no change in the e.s.r.intensity over a lo4 s timescan. Steady-state experiments performed by Genoud12. l3 examined the quantitative measurement of the e.s.r. signal as a function of the charge injected into the PP coat. It was reported that the spin concentration passed through a maximum for the removal of one electron from a 'unit' of six pyrrole monomers in the PP chain. The experimental data1,. l3 were interpreted in terms of a model involving quasi-equilibrium populations of polarons and bipolarons formed in the following two-step redox process: El PPo --- PP'+ + e E2 PP'+- PP2+ + e where El and E, are the 'quasi-standard' potentials for the formation of polarons (PP'+) and bipolarons (PP").It was shown that at low doping levels a bipolaron could combine with a neutral site in PP to form two polarons, and Genoud concluded that both the polaron and bipolaron need to be considered in the conduction mechanism of PP. The clear discrepancy between the observations of Genoud,l2v l3 and Street'' and Ka~fmann,~ may be further illustrated when the e.s.r. studies of Genoud are combined with recent conductivity data of Murray,14 who studied the change in PP conductivity with the fractional charge per pyrrole monomer sub-unit. Fig. 4 shows Genoud's e.s.r. data12 and Murray's conductivity data14 plotted together. A good agreement can be seen between the increase in PP conductivity and the polaron concentration, with the conductivity reaching a plateau as the polaron concentration begins to fall.This might be attributed to bipolaron formation. The obvious relationship between polaron concentration and conductivity at low doping levels would appear to contradict previous statementsg* l1 asserting the so-called ' spinless conductivity ' of PP. The discrepancies between previous e.s.r. studies of PP, and the relationship between conductivity and polaron concentration (fig. 4), have led us to undertake a further study980 - 0.15 I '€ -0.10 tr, b -0.05 Electrochemical Doping in Polypyrrole - 0.4 2 a d a W 0.2 m I 1 I 1 0.5 1.0 1.5 Q/(electrons per 6 pyrrole units) Fig. 4. Results from Genoud's e.s.r. study,12 showing the spin concentration (R) as a function of charge injection (Q), for comparison with the conductivity (0) results of Murray.14 of PP using in-situ electrochemical e.s.r.experiments at a channel electrode.15 The e.s.r. experiments were undertaken with simultaneous a.c. impedance measurements in order to obtain the polaron concentration and the 'conductivity ' (resistivity-') at the same time. The use of a.c. impedance to study electrochemical processes at the channel electrode has been described previously," and shown to exhibit no significant difference from those studied at more conventional electrodes. The a.c. impedance of PP has been described previously for ~hemically'~, l8 and electrochemicallyls~ 2o produced material, and further reference to this is made below. E.s.r.spectra were also recorded during the pyrrole electropolymerisation so as to detect any paramagnetic species that might occur in this reaction. Experimental PP-coated electrodes were prepared by the electropolymerisation of a solution contain- ing 10 mmol dm-3 pyrrole and 0.1 mol dm-3 tetrabutylammonium tetrafluoroborate (TBAFB, as supporting electrolyte) in ace toni trile. The electropol ymerisation was carried out at platinum electrodes in the following manner. The potential was stepped to + 1.30 V (versus a saturated calomel reference electrode) for chosen lengths of time and the resulting current transient was measured. Integration of the current transient gave the charge passed in the electropolymerisation. This has been related previously to the thickness of PP coats produced by Bard,lg who suggested that a charge of 24 mC cmP2 produced a PP coat 0.1 pm thick.Pyrrole (Aldrich, 98%) was used as supplied, but TBAFB (Fluka, purum) was recrystallised from petroleum ether/ethanol (90 : 10 volume : volume) before use. Acetonitrile (Fisons dried distilled) was stored over 4 A molecular sieves for 24 h prior to use. (Excess pyrrole was removed from the coats, prior to the a.c. and e.s.r. measurements, by washing the PP-coated electrodes in acetonitrile.) Both a.c. impedance and e.s.r. measurements were performed in 0.1 mol dm-3 TBAFB-acetonitrile solution, containing no bulk electroactive species. Experiments were conducted at 25 "C and prior to the recording of data the electrolyte was rigorously deoxygenated, using dried deoxygenated BOC ' white spot ' nitrogen.Rigorous degassing was considered essential to prevent any possible chemical doping of PP by oxygen, which might alter both the conductivity and the polaron concentrations. A.c. impedance measurements were made using a Solartron 1286 Electrochemical Interface coupled with a Solartron 1250 frequency response analyser. This was employedA. M. Wafler and R. G. Compton 98 1 to apply a 10 mV amplitude sine wave to the PP-coated electrode under potentiostatic control, at frequencies from 100 mHz to 65 kHz. Resulting data was displayed on a HP7470 digital plotter. Simultaneous a.c./e.s.r. experiments were performed at a PP- coated channel electrode (vide infra). Independent a.c. measurements were made at a PP- coated platinum disc electrode. For the latter a saturated calomel electrode was employed as a reference electrode, whereas for the e.s.r.and simultaneous a.c./e.s.r. measurements potentials a silver wire was used as a pseudo-reference electrode. In-situ electrochemical e.s.r. experiments were made at a silica channel electrode flow cell which was 50 mm long, with cross-sectional dimensions 0.4 mm x 6.0 mm.15 The square platinum electrode had dimensions of 4 mm x 4 mm. Before use the platinum channel electrode, or the disc electrode, were carefully polished by using successive fine grades of Hyprez Spray (Engis Limited, Kent) diamond lapping compound down to 0.25pm. The channel electrode was used in conjunction with a Bruker ER200D spectrometer, fitted with a TE1,, e.s.r.cavity. E.s.r. data were recorded with the channel electrode placed centrally within the e.s.r. cavity, close to the position of maximum sensitivity . Results and Discussion PP-coated platinum channel electrodes were prepared as described above. The PP films were estimated to have thicknesses ranging from 0.025 to 0.15 pm. Polymerisation was performed with the electrode located in the cavity of the e.s.r. spectrometer, so that any paramagnetic species produced as intermediates in the electrode reaction might be detected. The suggested literature mechanism for the polymerisation of pyrrole is shown below and can beseen to involve the pyrrole radical cation as an intermediate :21 H H H 2 0 .+ c* -(---("+2H+ I \ - + N +N H H No direct evidence has to our knowledge been reported for this mechanism.21g22 In-situ e.s.r.experiments performed upon the electrochemical oxidation of pyrrole, under a wide range of conditions (pyrrole concentrations 0.5 to 50 mmol dmV3, oxidation potentials 0.000 to + 1.500 V us. SCE), gave no spectrum attributable to a solution species. The absence of a spectrum does not necessarily indicate that the proposed mechanism21 is wrong, but more likely that the coupling of two radical cations is rapid, with a second-order rate constant greater than lo6 dm3 mol-1 s-l, which is the probable limit for detection using the channel electr~de.,~ The only spectra observed were those resulting from doped PP coated on the electrode, this is discussed below. The e.s.r. spectra recorded of the PP coats studied here are typical of those that have been observed from polymer-coated electrode~.~~, 25 They consist of a single line, several G wide, with no observable hyperfine interactions.Fig. 5 shows the e.s.r. spectra observed for a doped PP coat of thickness 0.05 pm; the symmetrical lineshape and linewidth of 0.90f0.05 G is in close agreement with the observations of Genoud.12 Under no circumstances was a spectrum observed consistent with a Dysonian lineshape.26 Such lineshapes arise in highly conducting samples where the microwaves982 Electrochemical Doping in Polypyrrole Fig. 5. The e.s.r. spectrum of a doped PP film (charge removed = 1 electron per 6 pyrrole monomer units) of thickness 0.05 pm. employed in the e.s.r. experiment are adsorbed by the sample within a short distance (micrometres) of its surface. Electron diffusion in and out of this ‘skin depth’ on the timescale of the e.s.r.experiment gives rise to the observed asymmetrical spectrum. Dysonian lineshapes are indicative of a ‘metallic ’ conductivity. Although such linescapes have been claimed by A l b e r ~ ~ ~ as occurring in PP-coated electrodes, the absence of such a lineshape in the electrodes studied here can be attributed to the coat thicknesses employed being smaller than the ‘skin depth’ of microwave penetration.26 It is probable that the coats studied by Albery were considerably thicker than the microwave skin depth, resulting in the observed Dysonian lineshape. With the e.s.r. lineshape observed for the PP coats studied here it was possible to study quantitatively the spin concentration (from double integration of the e.s.r.signal) as a function of the charge injected into the PP coat. This is not possible with PP coats exhibiting Dysonian e.s.r. signals, since increasing charge injection leads to enhanced conductivity and thus decreasing skin depths in relation to the thickness of the PP coat. This means that the e.s.r. experiment observes less of the polymer as the doping increases, and so the signal does not directly indicate the spin concentration. The e.s.r. behaviour of the PP coats was studied using both transient and steady-state experiments. Transient experiments were performed in which a potential step was applied to the PP-coated electrode and the e.s.r. signal was monitored with time. Typical results can be seen in fig.6, corresponding to (A) a potential step from the undoped insulating to fully doped conducting PP and (B) a potential step to an intermediate potential. Under all circumstances the e.s.r. response upon application of the potential step reached a steady state within a maximum timescale of 10 to 15 s, as may be seen in fig. 6. This behaviour is identical with the e.s.r. work of Genoud, who also employed potential step experiments12 and also with optical The change in e.s.r. signal amplitude may be used as a measure of the number of spins present since the linewidth was constant throughout the transients (AHpp = 0.90 G). The data shown in fig. 6 suggest that the potential step producing the partially doped material produces solely and directly polarons, whereas in the formation of the fully doped material the time evolution of the number of spins passes through a maximum, suggesting the initial formation of polarons followed by their subsequent combination to form bipolarons.Steady-state experiments were performed at PP coats which had been prepared in situ. After the required film thickness had been obtained, as described above, the potentialA . M. Waller and R. G. Compton 983 5 i0 i5 time/s Fig. 6. The e.s.r. signal response ( S ) to the application of potential steps to a PP-coated electrode : (a) -1.00 to 0.20V U S . Ag, (6) -1.00 to +l.OOV US. Ag. ./x I/ \ ‘X 1 Q/(electrons per 6 pyrrole units) Fig. 7. The e.s.r. signal amplitude ( S ) as a function of the charge injected ((3) into a PP-coated electrode, with a coat thickness 0.5 ,urn.applied to the electrode was stepped to very negative potentials (< - 1.000 V) at which the PP is in its fully reduced and insulating state. E.s.r. spectra were recorded of the insulating PP and showed a complete absence of residual spins. These might, for example, have arisen from charge defects in the undoped PP. The presence of such residual spins from defects has previously been suggested as the source of e.s.r. spectra observed in PP.l’ All our experiments on undoped PP revealed no e.s.r. spectra. Steady- state experiments were then performed on the doped PP. E.s.r. spectra were recorded for differing degrees of charge injection into the PP. The charge passed was calculated from the integration of the current passed as potential steps were applied to the PP-coated electrode.The results of such an experiment can be seen in fig. 7, where the amplitude of the e.s.r. signal is plotted against the charge injected. The signal amplitude may be regarded as proportional to the number of spins (i.e. number of polarons), since for the spectra obtained here the linewidth was again seen to be constant (AH,, = 0.90 G). The charge injected into the PP is expressed as the charge per ‘unit cell’ of six pyrrole monomers, a normalisation akin to that employed by Genoud.12. l3 The behaviour observed in fig. 7 shows the polaron concentration (number of spins) to pass through a maximum corresponding to the removal of one electron per ‘unit cell’984 Electrochemical Doping in Polypyrrole Z'/lO* R Fig.8. The complex plane impedance plot for a PP-coated electrode, with a coat thickness 0.04 pm, measured in 0.1 mol dm-3 TBAP-acetonitrile. The potential was -0.500 V us. SCE. in the PP chain (i.e. the formation of one radical cation). This is in good agreement with the work reported by Genoud.12*13 It was found that such behaviour was reproducible for PP coats over the thickness range 0.025-0.15 ym. The effect of doping on the polaron concentration was reversible. Potential steps to more negative potentials produced identical behaviour, the spin concentration in all cases being dependent solely on the final potential. Cycling of the PP-coated electrode for up to 50 cycles showed no measurable effect on the observed e x . behaviour. This observation is in disagreement with a previous observation on electrochemically cycled PP.l1 Our results obtained from the e.s.r. study of PP were, however, in excellent agreement with those of Genoud,12* l3 and we therefore conclude that the polaron does indeed need to be considered as being involved in the mechanism of charge transport in PP. Further and conclusive evidence for this comes from the simultaneous measurement of the polaron concentration and PP conductivity, which we consider later. The a.c. impedance spectrum of PP has been reported for both ~hemicallyl',~~ and electrochemically19~ 2o synthesised samples. We are unaware of any previous study undertaken of the a.c. impedance behaviour of a PP-coated electrode as a function of the charge injected into the polymer.This we now report. The a.c. impedance was studied for a range of PP coats, thicknesses from 0.025 to 0.15 ym, applied to a platinum disc electrode. A.c. impedance data were recorded as complex impedance plane plots as a function of the charge injected into the PP-coated electrode, this was measured by the integration of the current transient produced from the application of potential steps to the electrode. The complex plane impedance plots typically exhibited two forms, dependent upon the degree of doping of the PP. For PP-coated electrodes, where the potential applied to the electrode was less than 0.OOV (versus SCE), the behaviour observed was seen to be identical with that observed by Jakobs2' under similar conditions. Two semicircles are apparent in the impedance plot and these are inclined below the real impedance axis.A typical impedance plot obtained for a PP-coated electrode under these conditions can be seen in fig. 8. The inclination of the semicircles was attributed by Jakobs2' to the non-A . M. Waller and R. G. Compton 985 uniformity of the electric field, which would be found at the rough electrode surface. Similar observations have been observed at other electrodes,28 and the inclined semicircles observed have been described with reference to the fractal nature of the electrode s ~ r f a c e . ~ ’ - ~ ~ The observed impedance plot is consistent with a series of two parallel combinations of resistance (R) and constant phase-angle admittance (CPA), as seen below. CPAl CPAz where R, is the solution resistance, R,,, is the resistance and CPA,,, is the constant phase-angle admittance.For the case where the constant phase-angle admittance corresponds to a simple capacitance a semicircle situated on the real impedance axis would be observed. Such a situation would be observed with a perfectly smooth electrode surface. However, under real conditions a fractal electrode surface can be shown to generate an interfacial Z-’ = a’(ico). impedance of the where i = J( - l), ~ i ) is the frequency (rad s-’) and d is a frequency-independent real constant. The exponent a is fractional and for ideally polarisable electrodes lies between and 1. Thus perfectly smooth surfaces are characterised by a = 1 and thus show ideal capacitative behaviour. In general a = [D- l]-’, where D is the effective fractal dimension of the surface (2 < D < 3),29-32 from which it may be seen that the case of D = 3 gives a = i, which is a result previously deduced for a porous e l e ~ t r o d e .~ ~ ? ~ ~ The capacitance has then to be replaced by a new element, the constant phase element, which has the property of causing a frequency-independent phase shift between the applied a.c. potential and its current response. Dependent upon the magnitude of this phase shift (0) the semi-circle observed in the impedance plot of an R/C parallel circuit is shifted below the real impedance axis, as seen in fig. 9. From the two inclined semicircles values of R and the ‘pseudo-capacitance’ C*35 can be calculated as described by S l ~ y t e r s ~ ~ from the frequency at maximum Z”, fma,, using; where R is given by the distance between the two intersection points of the semicircle with the real axis.As suggested by Jakobs,20 the relative magnitudes of the respective values of R and C* indicate that the semicircle at low frequency (100 mHz-156 Hz), R = 4350 R and C* = 115 pF cm-2, can be attributed to the double-layer capacitance at the PP-coated electrode/electrolyte interface, and the semicircle at high frequency (156 Hz-65 kHz), R = 85 0 and C* = 97.5 nF cmP2 can be attributed to the bulk PP. At higher doping levels (potentials > 0.00 V us. SCE) the complex plane impedance plots were found to differ from the observations of Jakobs.20 A typical impedance plot can be seen in fig. 10. At low applied a.c. frequency (100 mHz-15.6 Hz) the impedance plot is almost parallel to the imaginary impedance axis.Such behaviour indicates that the double-layer capacitance increasingly determines the impedance. At intermediate frequencies (15.6-134 Hz) the impedance plot is a straight line at an angle of 45 O to the real impedance axis. This behaviour might be interpreted either as that expected for a reversible reaction involving a species (in this case charge-compensating counter ions)986 Electrochemical Doping in Polypyrrole 2” \ Fig. 9. The complex plane impedance plot for (a) simple RC parallel circuit, and (b) RC parallel circuit where C is replaced by a constant phase element. 1 Hz 2 4 Z’lld Icz Fig. 10. The complex plant impedance plot for a PP-coated electrode, with a coat thickness 0.04 pm.The potential was + 0.500 V us. SCE. transferred to a plane electrode by diff~sion,~’ or, as a porous e l e c t r ~ d e . ~ ~ On the basis of previous ob~ervationsl~~ 2o only the porous-electrode interpretation need be considered. At high frequencies (134 Hz45 kHz), R = 71 R and C* = 129.5 nF cm-2 the impedance plot consists of a semicircle inclined below the real impedance axis. The magnitude of R and C* evaluated from this feature would suggest that it may be attributed to the bulk PP. It was possible from these observations in the a.c. impedance of PP to monitor changes in the R and C* of the bulk PP, as a function of the degree of charge removal, and as a function of the PP thickness. Typical results can be seen in fig. 1 1, for the change in resistivity (p) of the PP as a function of charge removal. The charge removed being obtained from the integration of the current transient, after the current had decayed to zero, produced by a potential step at the PP coated electrode.When the conductivity (a) was plotted against the degree of charge removal (fig. 12), good agreement was seen with the results obtained independently by Murray.14 As he observed14 the conductivity of the PP increased as the degree of doping, but then reached a plateau after which further doping lead to no increase in PP conductivity. Fig. 13 shows how the conductivity of theA . M. Waller and R. G. Compton 987 10 !3 C s 5 5 N 0 I I 1 Q/(electrons per 6 pyrrole units) Fig. 11. The resistivity @) of a PP-coated electrode, thickness O.O4pm, as function of charge removed (Q).d'; 0- 0 / I 1 Q/(electrons per 6 pyrrole units) Fig. 12. The conductivity (u) of PP-coated electrode as function of charge removal (Q). PP coats changes with thickness, measured at a constant charge removal of 2 electrons per 6 pyrrole monomers for each sample, the linear behaviour observed being consistent with previous observations.lg Finally, a simultaneous a.c./e.s.r. experiment was performed. The a.c. impedance spectra recorded at the channel electrode in this simultaneous experiment exhibited no differences from those observed at the disc electrode. The data obtained from the simultaneous a.c. impedancelin-situ e.s.r. experiment on the PP-coated electrode can be seen in fig. 14, where the number of spins (S) and conductivity (a) are plotted against the d.c.potential applied to the electrode.988 Electrochemical Doping in Polypyrrole lo-’ Fig. 13. The conductivity I I 0.05 0.10 dlw :a) of PP-coated electrodes as function of coat thickness (d), at a constant charge. removal of 2 electrons per 6 pyrrole monomer units. 3 I 0.0 0.5 EIV us. Ag Fig. 14. The conductivity (a, 0) and polaron concentration (S, x ) as a function of the d.c. potential ( E ) applied to a PP coated electrode, of coat thickness 0.05pm, measured in 0.1 mol dm-3 TBAP-acetonitrile. It can be seen from fig. 14 that a good correlation is observed between the increase in polaron concentration (number of spins) and conductivity (a) up to a potential which corresponds to a charge injection of one hole per unit cell of six monomers.The conductivity behaviour observed here agrees with that found in the independent a.c. experiments, described earlier for a disc electrode. The resistance of the PP coat decreases as the degree of doping increases, consistent with the formation of largerA . M. Waller and R. G. Cornpton 989 numbers of charge carriers within the PP. Such behaviour is seen to occur, until at a given doping level (consistent with one electron removed per PP unit cell of six pyrrole units) the resistance reaches a plateau, after which further charge injection led to no increase in the conductivity (decrease in resistance). It can be seen from fig. 14 that the conductivity reaches its plateau at the same doping level at which the polaron concentration passes through its maximum.That no increase in the conductivity was observed, as the polaron concentration decreased (bipolaron concentration increased), suggests that the bipolaron has the same mobility as the polaron and that the conductivity is dependent on the number of charge carriers (polarons or bipolarons). The bipolaron is not then the sole species responsible for PP conductivity, as has been previously suggested. 5-7 Conclusions The experimental data obtained from this first simultaneous a.c. impedancelzn-situ e.s.r. experiment have indicated that conduction within PP cannot be solely attributed to the bipolaron, as has been previously and that the polaron has to be involved particularly at low doping levels, where it is the dominant species.This conclusion is in agreement with other w o r k e r ~ , ~ ~ . ~ ~ and compares well with the independent results of Genoud12 and M ~ r r a y . ' ~ We conclude that the a.c./e.s.r. results presented here have shown that the so-called 'spinless conductivity' of PP" is not valid, and that the discussion of PP conductivity should consider both the polaron and bipolaron. We thank the S.E.R.C. for financial support and a post-doctoral research assistantship for A.M. W. References 1 A. F. Diaz, K. K. Kanazawa and G. P. Gardini, J . Chem. SOC., Chem. Commun., 1979, 635. 2 J. L. Bredas, R. R. Chance and R. Silbey, Phys. Rev. B, 1982, 26, 5843. 3 J. L. Bredas, B. Themans, J. M. Andre, R. R. Chance and R. Silbey, Synth. Met., 1984, 9, 265. 4 J. L. Bredas, B. Themans, J.G. Fripiat, J. M. Andre and R. R. Chance, Phys. Rev. B, 1984, 29, 5 R. R. Chance, J. L. Bredas and R. Silbey, Phys. Rev. B, 1984, 29, 4491. 6 J. L. Bredas, Mol. Cryst. Liq. Cryst., 1985, 118, 49. 7 J. C. Scott, J. L. Bredas, J. H. Kaufman, P. Pfluger, G. B. Street and K. Yakushi, Mol. Cryst. Liq. 8 J. L. Bredas, J. C. Scott, K. Yakushi and G. B. Street, Phys. Rev. B, 1984, 30, 1023. 9 J. H. Kaufman, N. Colaneri, J. C. Scott and G. B. Street, Phys. Rev. Lett., 1984, 53, 1005. 676 1. Cryst., 1985, 118, 163. 10 E. M. Genies and J. M. Pernaut, Synth. Met., 1984, 10, 117. 11 J. C. Scott, P. Pfluger, M. T. Krounbi and G. B. Street, Phys. Rev. B, 1983, 28, 2140. 12 F. Genoud, M. Guglielmi, M. Nechtstein, E. Genies and M. Salmon, Phys. Rev. Lett., 1985, 55, 13 M. Nechtstein, F.Devreux, F. Genoud, E. Vieil, J. M. Pernaut and E. Genies, Synth. Met., 1986, 15, 14 B. J. Feldman, P. Burgmayer and R. W. Murray, J. Am. Chem. SOC. 1985, 107, 872. 15 R. G. Compton and B. A. Coles, J. Electroanal. Chem., 1981, 127, 37. 16 R. G. Compton, M. E. Laing and P. R. Unwin, J . Electroanal. Chem., 1986, 207, 314. 17 N. Mermilliod, J. Tanguay and F. Petiot, J. Electrochem. SOC., 1986, 133, 1073. 18 J. Tanguay, N. Mermilliod and M. Hoclet, J. Elecirochem. SOC., 1987, 134, 795. 19 R. A. Bull, F. R. F. Fan and A. J. Bard, J. Electrochem. Soc., 1982, 129, 1009. 20 R. C. M. Jakobs, L. J. J. Janssen and E. Bahrendrecht, Red., J. R. Netherlands Chem. SOC., 1984, 104. 21 E. M. Genies, G. Bidan and A. F. Diaz, J. Electroanal. Chem., 1983, 149, 101. 22 A. F. Diaz, A. Martinez, K. K. Kanazawa and M. Salmon, J. Electroanal. Chem., 1980, 130, 181. 23 R. G. Compton, D. J. Page and G. R. Sealy, J . Electroanal. Chem., 1984, 161, 129. 24 W. J. Albery, C. C. Jones and R. G. Compton, J. Am. Chem. SOC., 1984, 106, 469. 25 R. G. Compton, F. J. Davies and H. H. Block, J. Chem. SOC., Chem. Commun., 1984, 890. 26 F. J. Dyson, Phys. Rev., 1955, 98, 349. 118. 59.990 Electrochemical Doping in Polypyrrole 27 W. J. Albery and C. C. Jones, Faraday Discuss. Chem. SOC., 1984, 18, 193. 28 S. Iseki, K. Ohashi and S. Nagaura, Electrochim. Acta, 1972, 17, 2249. 29 L. Nykios and T. Pajkossy, Electrochim. Acta, 1985, 30, 1533. 30 L. Nykios and T. Pajkossy, Electrochim. Acta, 1986, 33, 1347. 31 W. H. Mulder and J. H. Sluyters, Electrochim. Acta, 1988, 33, 303. 32 M. Keddam and H. Takenouti, Electrochim. Acta, 1988, 33, 445. 33 R. de Lmie, Electrochim. Acta, 1964, 9, 1231. 34 R. de Levie, Electrochim. Acta, 1965, 10, 113. 35 D. C. Grahame, J. Electrochem. SOC., 1952, 99, 370. 36 J. H. Sluyters, Red. Trav. Chim. Pays-Bas, 1960, 79, 1092. 37 A. J. Bard and L. R. Faulkner, Electrochemical Methods (Wiley, New York, 1980), pp. 350-352. 38 R. de Levie, Advances in Electrochemistry and Electrochemical Engineering, ed. P. Delahay (Interscience, New York, 1967). Paper 8/02298D; Received 8th June, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500977

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1989, 85(4), 991-998 The Dark and Radiation-induced Microwave Conductivity of Frozen Aqueous Gels Joyce Eden, Dick van Lith, John M. Warman and Andries Hummel* Radiation Chemistry Department, Interfaculty Reactor Institute, Technical University of Delft, Mekelweg 15, 2629 JB Delft, The Netherlands Charge migration has been studied in frozen aqueous gels of the polysaccharide K-carrageenan by measuring the microwave conductivity after pulsed irradiation with high-energy electrons. It is concluded that the conductivity, which is observed above a critical water weight fraction of 27 YO, is due to the migration of highly mobile excess electrons in the ice-like regions of the sample, as previously proposed for hydrated DNA, gelatin and collagen. At low temperatures and in samples with only slighly more than the 'critical' water content the electrons were found to decay non- exponentially over many orders of magnitude in time.Results are also presented on the dark conductivity of the samples in the X-band region and for the water contents and temperatures applicable to the pulse-radiolysis experiments. The dark conductivity shows discontinuities at the same water content as the critical value for the radiation-induced conductivity, further illustrating the transition in the nature of the aqueous medium surrounding the biopolymer. The results are compared with those obtained for other biopolymers. A general conclusion is that charge carriers formed within the biopolymer-bound water regions are immobile or, if mobile, become localised within a timescale much shorter than a nanosecond, thus excluding long-range migration.However, the fully hydrated polymer does not appear to behave as an effective trapping site for the mobile electrons formed in adjacent ice regions. Previous have shown that when hydrated samples of the polynucleotide DNA or the long-chain proteins gelatin or collagen are irradiated below -20 "C, a conductivity signal is observed which can be attributed to the migration of excess electrons through ice-like regions of the medium. The occurrence of this phenomenon has been found to have a sharp onset above a given water content, with the position of the onset being characteristic for the particular biopolymer investigated. In the present work we have extended these studies to the naturally occurring polysaccharide rc-carrageenan.As this is a gelling agent, a larger range of water to polymer ratios could be studied than for DNA or collagen. In addition to the radiation- induced conductivity, the background microwave conductivity of the samples has also been measured and the two are compared in their dependence on water content. The results are also compared with those obtained for the other biopolymers which have been studied. Experimental The gels were prepared by mixing a weighed quantity of rc-carrageenan (Sigma type I11 from Eucheuma cottonii) with a measured volume of water (B.D.H. AnalaR grade), then stirring well and warming to ca. 75 "C for 15 min to dissolve the biopolymer. The sol was then placed in the microwave conductivity cell and left to gel at room temperature for several hours before being stored at -20 "C until required.Samples with water weight fractions, Fw(H20), in the range 0.13-0.99 were prepared. 99 1992 Microwave Conductivity of Frozen Aqueous Gels Those samples consisting of > 80% water could be poured into the cell prior to gelling. Those with a lower water content were too viscous to be poured and had to be scooped into the cell and compressed. Some of the lowest water content samples could not be packed well into the cell and therefore contained air pockets. From the measured density of these samples the fraction of the volume containing the hydrated polymer was determined and the measured conductivity was divided by this fraction to correct for this lack of complete filling.The air pockets in the low water content samples were also thought to be the reason for the observation of short-lived conductivity transients in these media, which is known to result from the formation of electrons in irradiated air. The microwave conductivity cell consisted of a length of metal, X-band waveguide, 2.7 cm long with an internal cross section of 1.0 cm x 2.3 cm, which was closed at one end by a metal plate. During experiments the cell was contained in a cryostat and the temperature was varied between -20 and - 150 "C. The samples were irradiated using 0.3-5 ns pulses of 3 MeV electrons from a Van de Graaff accelerator giving doses per pulse of ca. 0.02-1 .O Gy. Mobile charge carriers formed on pulsed irradiation of the samples result in an absorption of microwave energy and by measuring the change in the microwave power reflected by the sample the mobility of the carriers and their decay kinetics could be investigated.The time resolution was ca. 2 ns. The microwave detection technique has been described fully el~ewhere.~, The dark conductivity of the samples, i.e. the conductivity in the absence of irradiation, and their permittivity were measured by comparing the amount of microwave power reflected by the sample cell with that from an identical but empty cell over the range of frequencies available (8.2-12.4 GHz). These measurements could be carried out for temperatures between -20 and - 135 "C. Below this, the difference in reflected power between the sample cell and the reference cell became too small to measure accurately.Results and Discussion The End-of-pulse Radiation-induced Conductivity As for the other biopolymers which have been studied,lP3 K-carrageenan displays a threshold water concentration below which no radiation induced conductivity can be detected and above which an approximately linear increase in the end-of-pulse con- ductivity with increasing water weight fraction is observed. This effect is shown in fig. 1, which also contains a summary of previously reported data for DNA,l collagen' and gelatin3 for comparison. The 'critical' weight fraction of water, Fw(H20)c, is in the case of carrageenan only 0.27 which is less than the values of 0.30 and 0.32 for the proteins collagen and gelatin and considerably less than the value of 0.45 found for the polynucleotide DNA.Since the threshold has been shown to be due to the onset of normal ice-like regions which support a high electron m~bility,l-~ the conclusion may be drawn that the amount of water which is intimately associated ('bound') with the polysaccharide is comparatively low. The critical weight fractions of water given above for the other polymer systems studied have been found to be close to the estimates of non-freezing water determined using other experimental technique^.^-' As mentioned in the introduction, a particular advantage of carrageenan was its gelling properties, which allowed a very large range of water concentrations to be studied without phase separation occurring on freezing.This was also the case for gelatin. As can be seen from the data in fig. 1 there is a substantial difference in the behaviour of the two gelling agents above the critical water content. For carrageenan the linear increase in conductivity above Fw(H20)c extrapolates almost exactly to the value found for pure ice at F,(H,O) = 1 .O. For gelatin, on the other hand, a linear extrapolation leadsJ . Eden, D. van Lith, J. M. Warman and A . Hummel 2 - I - 993 - (4 /% / / 06 / / 4 / O/ / I n / I 1 I /' / I / I c / / / / 2 d 0 0.2 0.4 0.6 0.8 1.0 Fw (H20) Fig. 1. The conductivity per unit dose after pulsed irradiation of (a) DNA, (b) collagen, (c) gelatin and (d) K-carrageenan as a function of the weight fraction of water at - 135 "C. to a value at Fw(H20) = 1 .O of only approximately one third that for pure ice.This effect in the case of gelatin was explained previously3 in terms of the presence of amorphous regions in which electrons become rapidly trapped even for water contents considerably in excess of the critical concentration. Such amorphous regions would therefore appear to be absent in the case of carrageenan. The Background Conductivity It is interesting at this point to consider the results of the measurements of the background microwave conductivity of the frozen media, since this is also expected to994 Microwave Conductivity of Frozen Aqueous Gels 0 0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 1.0 Fw (H2O) Fw (H20) Fig. 2. Variation in 0, the activation energy, E,, and 0 , the pre-exponential factor, ~(o),,, divided by the polymer volume fraction, F,(P) = 1 - F,(H,O), for the background microwave conductivity of (a) gelatin and (b) K-carrageenan as a function of the water weight fraction F,(H,O).be related to the amount of polymer-bound water in the samples. For all water contents it was found that the background conductivity, a(co), could be described by an Arrhenius- type expression, i.e. a(co, T ) = a(w), exp (- EJk, T ) . (1) The conductivity values used in the temperature plots were the average values required to fit the microwave loss over the available frequency band of 8.2-12.4 GHz. This type of Arrhenius plot has been presented previously for gelatin.3 In fig. 2 the values determined for a(co), and E, are plotted as a function of the water weight fraction. In the case of a(w), the parameter actually plotted is a(co),/F,(P) where F,(P) is the volume fraction of the polymer, i.e.F,(P) = 1 -&,(H,O). The data are plotted in this way since the dielectric loss is expected to result only from those water molecules which are associated with the polymer. The water in ice-like regions should make no contribution to the dielectric loss. Accordingly as the water content increases the value of a(cu),/F,(P) would be expected to reach a constant value when the solvation shell of the polymer is complete. This is in fact seen to occur in fig. 2 for both gelatin and carrageenan, with the onset of a constant level being found at a water weight fraction which is very close to the threshold value found for the radiation-induced conductivity.For water contents below the threshold the value of o(co),/F,(P) is found to decrease close to exponentially with F,(H,O). This type of behaviour has been found in other studies of dielectric relaxation in hydrated biopolymers." pis can be seen from the data in fig. 2, the activation energy for the dielectric loss reaches a plateau for both polymers at a water content which is identical to the threshold found for a(co),/&(P). The plateau activation energy values are 0.19 and 0.21 eV for gelatin and carrageenan, respectively. The activation energy is seen to decrease substantially for water contents below the threshold value. Further discussion of the background microwave conductivity data will be reserved for a separate publication. The data presented here, however, serve to show theJ.Eden, D. van Lith, J. M . Warman and A . Hummel 995 I O - ~ 0 20 40 60 80 I00 tlns Fig. 3. Decay of the radiation-induced conductivity at -40 "C for K-carrageenan samples with weight fractions of water of H, 0.55; W, 0.73; Q, 0.92; X, 0.96; and 0, 0.99; Dose per pulse ca. 0.1 J kg-'. overriding influence of bound water on the behaviour of both the radiation-induced and the dark-conductivity properties of these materials. No evidence could be found for radiation-induced conduction in either the dry polymers or the polymers with a full complement of bound water. We are therefore forced to conclude that the primary charge carriers, whether electrons or electron holes, formed within the polymer and its associated bound water region are either immobile or become rapidly localised on a timescale much shorter than that of our measurement capabilities.Charge migration over large distances would therefore appear to be excluded. The Decay Kinetics of the Radiation-induced Conductivity At the higher temperatures studied (- 20 to - 60 "C) the radiation-induced conductivity signals found above the threshold water content decayed exponentially with time. This is illustrated in fig. 3 by the linear semi-logarithmic plots of the conductivity transients at -40 "C for carrageenan gels with several different water contents. The rate of decay is independent of the water content. An Arrhenius plot of the first-order decay constants is shown in fig. 4, which further illustrates the lack of dependence of the decay kinetics on F,(H,O) in this 'high'- temperature region.An activation energy of 0.56 & 0.05 eV is obtained from the slope of the dashed straight line drawn through the data points in fig. 4. This activation energy is, within experimental error, equal to the activation energy found for the decay of mobile electrons of 0.55 0.05 eV reported for pure ice.", l2 While the activation energy of the decay rate is the same as that found for pure ice, the absolute values of the decay rates are actually almost an order of magnitude lower in the carrageenan samples. This is illustrated by a comparison with the pure-ice data given as the dashed line in fig. 4. In some way the presence of the polymer, with its surrounding bound water layer, appears to be capable of reducing the equilibrium concentration of the defect species responsible for electron localisation in the non-bound996 Microwave Conductivity of Frozen Aqueous Gels I \ \a *\ I \ \ , L! I .\ . \ 6 103 KIT Fig. 4. Arrhenius plot of the first-order trapping constant for hydrated Ic-carrageenan samples with weight fractions of water of A, 0.55; +, 0.73; X, 0.92; 0,0.96; and 0,0.99. The dashed straight line drawn through the points gives an activation energy of 0.56 eV. The dotted line represents the results obtained previously for pure ice. ice regions. How this can be achieved while still retaining the bulk ice activation energy for electron localisation is not completely clear. A further important conclusion which may be reached on the basis of the longer electron lifetimes in the polymer systems containing an excess of water is that the polymer, with its hydration envelope, does not act as an efficient localisation site for those mobile conduction electrons which are formed in the neighbouring regions of ice.In fig. 5 are shown conductivity decays for a sample with a water weight fraction of 0.92 for temperatures down to - 135 "C. The results have been plotted using a logarithmic timescale to accommodate all of the data and also to illustrate the deviation from first-order decay kinetics which occurs at the lowest temperatures. For such a log-log type representation an exponential decay curve has a fixed form which is shown by the dashed line in the figure. Changes in initial yield and in decay constant correspond simply to moving this curve vertically or horizontally, respectively.At the lowest temperatures there is a clear indication of a 'tailing' of the decays, with the disappearance of electrons occurring much more slowly at long times than would be expected on the basis of a single exponential time constant. This disperse kinetic behaviour at low temperatures is most pronounced for the samples containing the smallest amounts of water. This is illustrated in fig. 6, in which it can also clearly be seen that the timescale on which decay occurs has now become veryJ. Eden, D. van Lith, J. M. Warman and A . Hummel 997 I O - ~ I O - ~ 10- tls Fig. 5. log-log plot of the decay of the radiation-induced conductivity at different temperatures in the Ic-carrageenan sample with a weight fraction of water of 0.92: +, -20; Ix], -40; V, -60, X, - 100 and 0, - 135 "C.The dashed lines correspond to the behaviour expected for a pure first-order decay. -5 Fig. 6. weight to give log-log plot of the conductivity decays at - 135 "C for K-carrageenan gels with different fractions of water. Dose per pulse ca. 0.1 J kg-'. All signals shown have been normalised approximately the same initial signal height for ease of comparison of the changes in the kinetics. dependent on the water content, with many orders of magnitude in time separating the highest and lowest Fw(H,O) values. This is in contrast to the independence on Fw(H,O) found at higher temperatures. Similar disperse decay effects have been observed in DNA samples' but a much smaller range of water content was covered.The kinetics resemble those found in certain998 Microwave Conductivity of Frozen Aqueous Gels amorphous semiconductor materials. In the present case, however, it is almost certain that the regions within which the electrons are diffusing are crystalline. This can be concluded from the similarity in the absolute magnitude of the conductivity to that found for pure ice. In bulk ice itself no indication has been found for abnormal decay kinetics even for temperatures down to - 180 "C. We conclude therefore that the disperse kinetics observed in the frozen biopolymer systems at low temperatures are in some way a result of the limited dimensions of the ice-like regions within which the electrons diffuse. An effect which should also be noted is the observed sensitivity of the signals at low temperatures to the accumulated dose given to the sample.A marked decrease in the transient lifetime could be observed for doses as low as 100 Gy. A similar sensitivity was also noted for gelatin samples, whereas the multiple strand polymers, DNA and collagen, seemed to be relatively radiation insensitive. From the effect of accumulated dose on the electron lifetime, a rate constant for reaction with the radiation product responsible for the faster decay could be estimated to be 2 x 1013 dm-3 mol-1 s-l based on a product yield of one molecule per 100 eV absorbed. For all samples, accumulated dose effects could be annealed out by warming temporarily to above -20 "C. All of the traces shown in the figures were obtained for total doses sufficiently small that dose effects could be neglected. References 1 D. van Lith, J. M. Warman, M. P. de Haas and A. Hummel, J. Chem. SOC., Faraday Trans. I , 1986, 2 D. van Lith, J. Eden, J. M. Warman and A. Hummel, J . Chem. Soc., Faraday Trans. I , 1986, 82, 3 J. Eden, D. van Lith, J. M. Warman and A. Hummel, Radiat. Phys. Chem., 1987, 29, 51. 4 P. P. Infelta, M. P. de Haas and A. Hummel, Radiat. Phys. Chem., 1977, 10, 353. 5 J. M. Warman, in The Study of Fast Processes and Transient Species by Electron Pulse Radiolysis, ed. J. H. Baxendale and F. Busi (Reidel, Dordrecht, 1982), p. 129. 6 M. Falk, A. G. Poole and C. G. Goymour, Can. J . Chem., 1970,48, 1536. 7 T. E. Cross and R. Pethig, Znt. J. Quantum Chem., 1983, 10, 143. 8 R. E. Dehl, Science, 1970, 170, 738. 9 I. D. Kunst Jr., T. S. Brassfield, G. D. Law and G. V. Purcell, Science, 1969, 163, 1329. 82, 2933. 2945. 10 R. Pethig, Dielectric and Electronic Properties of Biological Materials (Wiley, Chichester, 1979). 11 M. P. de Haas, M. Kunst, J. M. Warman and J. B. Verberne, J. Phys. Chem., 1983, 87, 4089. 12 J. M. Warman, M. P. de Haas and J. B. Verberne, J. Phys. Chem., 1980, 84, 1240. Paper 8/02600I; Received 29th June, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500991

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J . Chern. Soc., Faraday Trans. I , 1989, 85(4), 999-1007 Supported Palladium Catalyst Prepared from Amorphous Palladium-Zirconium Structural Properties and Catalytic Behaviour in the Hydrogenation of Carbon Dioxide Alfons Baiker* and Daniel Gasser Department of Industrial and Engineering Chemistry, Swiss Federal Institute of Technology, ETH - Zentrum, CH-8092 Zurich, Switzerland Amorphous Pd,Zr, has been used as a precursor for the preparation of CO, hydrogenation catalysts. The active catalysts were obtained by exposing the amorphous alloy to CO, hydrogenation conditions. During this exposure the bulk and surface structure of the alloy changed drastically. The most important changes were : (i) partial crystallization of the amorphous alloy; (ii) the formation of palladium metal particles; and (iii) the oxidation of zirconium to zirconium dioxide. These changes were accompanied by an increase in both the B.E.T.surface area and the palladium surface area of more than two orders of magnitude. Palladium was found to exist in two discernible phases, as pure metal and as solid solution with hydrogen. Two types of palladium particles were distinguishable, small disordered particles and larger crystalline particles of ca. 12 nm mean size. The catalytic behaviour of such catalysts was compared with the corresponding behaviour of a Pd/ZrO, catalyst prepared by wet impregnation of crystalline zirconia with a palladium salt. The activity and selectivity behaviours of the two catalyst types were largely different. Palladium contained in the catalyst derived from amorphous Pd,Zr, was more than an order of magnitude more active for CO, conversion than palladium in the catalyst prepared by wet impregnation. Over Pd/ZrO, derived from Pd,Zr, the products were methane, methanol and CO, whereas over the conventionally prepared Pd/ZrO, no substantial methanol formation was observed.In the past few years several studies have been reported on CO hydrogenation over various amorphous alloys.'-' Yokoyama et a1.l investigated CO hydrogenation over amorphous Pd,Zr, at atmospheric pressure. They prepared a highly active methanation catalyst by in situ activation of the amorphous alloy. This catalyst consisted of an unknown Pd-0-Zr phase and ZrO,. Shibata et al., compared several amorphous zirconium alloys of the composition M,Zr, (M: Rh, Pd, Os, Ir, Pt, Au) for CO hydrogenation at 6 bar.They found that CO hardly reacted over amorphous alloys below 600 K, except over Au,Zr, and Pd,Zr,. The main product formed over Au,Zr, was methane, while over Pd,Zr, methanol formation was prevalent. Au,Zr, was oxidized to metallic gold and ZrO,, while Pd,Zr, was decomposed into a weakly bound Pd-0-Zr type complex oxide and ZrO, under reaction conditions. Comparatively little attention has been paid so far to the hydrogenation of carbon dioxide over supported transition metals. Some aspect of CO, hydrogenation over supported Pd catalysts were discussed by Henderson and Worley.'O They pointed out that CO, hydrogenation is generally thought to proceed via dissociation to some form of adsorbed CO and then along the same reaction pathway as CO hydrogenation.Since Pd-Zr alloys were found to be interesting precursors for the preparation of CO hydrogenation catalysts, it seemed worthwhile to examine also their potential as 9991000 Supported Palladium Catalyst precursors for CO, hydrogenation catalysts. Here we report the structural changes of the amorphous Pd,Zr, precursor during in situ activation and compare the catalytic behaviour of such catalysts with that of a conventionally prepared catalyst of similar composition. Experimental Cat a1 y s t The amorphous Pd,Zr, used as catalyst precursor was prepared from the pre-mixed melt of the pure metals using the technique of melt spinning. For use in the catalytic tests, the 5 mm wide and 2G30 pm thick ribbons fabricated by the melt spinning were ground to flakes of 0.5-1 mm size under liquid nitrogen. The B.E.T.surface area of this material, as measured by krypton adsorption at 77 K, amounted to less than 0.1 m2 g-l. X-Ray diffraction indicated that the ground material was still completely amorphous [fig. 2(a)]. A conventionally prepared zirconia-supported palladium catalyst was used as a reference. The catalyst, containing 5 wt % of palladium, was prepared by impregnation of ZrO, made up of crystalline particles of tetragonal structure." The impregnation was carried out with an aqueous solution of (NH,),PdCl, using the incipient wetness technique. Subsequent to drying for 12 h at 393 K, the precursor was calcined and reduced under a flowing hydrogen-nitrogen mixture (H, : N, = 1 : 10) for 15 h at 473 K, 1 h at 573 K, 1 h at 673 K, 1 h at 773 K and finally 1 h in pure hydrogen.Catalyst Characterization The amorphous Pd,Zr, alloy, as well as the catalysts prepared from it and the reference catalyst, were investigated with regard to their physical and chemical properties using gas adsorption (N,, CO, and krypton), X-ray diffraction (X.r.d.), thermal analysis (t.g., d.s.c.) and scanning electron microscopy (s.e.m.). The adsorption measurements were performed in an apparatus especially designed for the volumetric determination of adsorption isotherms and equipped with a high- precision pressure gauge (Ruska DDR 6000) with an accuracy of 0.2 Pa. B.E.T. surface areas were mcasured by krypton or nitrogen ads9rption at 77 K, using a cross-sectional area of 19.5 A2 for the krypton atom and 16.3 A2 for nitrogen.', CO chemisorption measurements were carried out to evaluate the number of surface palladium atoms. Prior to CO chemisorption measurements, samples were reduced in pure hydrogen at 493 K for 12 h and subsequently evacuated (lo-, Pa) at the same temperature for 24 h.CO chemisorption was carried out at 300K and included subsequent steps: (i) measurement of total CO uptake corresponding to strongly and weakly adsorbed CO; (ii) evacuation of the sample for 1 h at lo-, Pa; (iii) measurement of CO uptake corresponding to weakly adsorbed molecules. The amount of chemisorbed CO was calculated as the difference of the two adsorption uptakes. The palladium surface area was estimated assuming a stoichiometric factor of two for CO chemisorption and a cross-sectional area of 7.87 x lo-,' m2 for the palladium atom.13 Qualitative phase analysis and the determination of the metal particle size were carried out using a powder diffractometer and Cu Ka radiation. Mean palladium particle sizes were determined from the half-width B+ of the Pd(100) reflection using the Scherrer formula.Measured values of B; were corrected to account for the contribution due to instrumental broadening. Scanning electron microscopy was used to investigate textural and morphological properties of the catalysts.A. Baiker and D. Gasser 1001 Catalytic Tests The apparatus used for activity measurements was described in detail in ref. (14). It consisted of a continuous tubular flow fixed-bed microreactor.The product gas mixtures were analysed on line using a gas chromatograph (HP 5890 A) equipped with a column ( 5 m; The experiments were carried out under the following standard conditions : amount of catalyst, 1.2 g for in situ activation, 0.3 g for kinetic measurements; feed rates of reactants : CO,, 2.3 pmol s-l ; H,, 7.6 pmol s-' ; feed rate N,, 1.2 pmol s-' ; total pressure, 15 bar. Since the palladium surface area of the conventionally prepared Pd/ZrO, catalyst was markedly lower, we have used 2.1 g of this catalyst (as compared to 0.3 g of the Pd/ZrO, catalyst prepared from Pd,Zr,) in the comparative kinetic studies. Conversion and selectivity were calculated as follows : id.) packed with Poropak QS (80-100 mesh).moles CO, reacted moles CO, fed to the reactor conversion = x 100% moles of component (i) formed moles CO, reacted selectivity (i) = x 100%. Results Chemical and Physical Changes of Amorphous Precursor During in situ Activation In situ activation of the amorphous Pd,Zr, was performed by exposing it to reaction conditions in the fixed bed reactor. The temperatures used (493-553 K) were significantly lower than the crystallization temperature (ca. 720 K) of the amorphous Pd,Zr, which was determined by d.s.c. in a nitrogen atmosphere at a heating rate of 10 K min-'. Fig. 1 depicts the change of activity and selectivity of the precursor during in situ activation in this temperature range. At all temperatures the activity increased with time on stream and reached a steady state.The steady-state activity exceeded the activities of the original precursor by more than an order of magnitude. Note that the increase in activity with time on stream depended significantly on temperature. After stable activity was reached the major reaction products were methane and water at 523 and 553 K, respectively. At lower temperature (493 K), the formation of CO and methanol also became significant besides that of methane and water. It should be noted that in the initial period of the in situ activation no water was detected in the product stream. The water formed was presumably consumed for the oxidation of the zirconium in the precursor. Thermogravimetric measurements (t .g.) combined with X.r.d. indicated that Zr was also oxidized by CO, according to Zr + 2C0, -+ ZrO, + 2CO.Heating the Pd,Zr, alloy in a carbon dioxide atmosphere resulted in a weight increase corresponding to the complete oxidation of the zirconium to zirconium dioxide. At a heating rate of 10 K min-l the oxidation started at ca. 483 K, i.e. at a temperature considerably lower than the crystallization temperature (ca. 720 K). Fig. 2 shows the bulk structural changes the amorphous Pd,Zr, precursor underwent during activation as seen by X.r.d. The amorphous Pd,Zr, exhibited broad intensity maxima typical for materials without any long-range ordering of the constituents. The diffraction pattern of the active catalyst indicates that the amorphous alloy was transformed into a solid containing palladium and zirconium dioxide. It is interesting to note that some of the palladium formed a solid solution with hydrogen.This is indicated by reflections b (fig. 2) appearing at the lower 28 side of the prominent Pd reflections.1002 Supported Palladium Catalyst time on stream/h Fig. 1. Change of activity of the amorphous Pd,Zr, alloy during in situ activation. CO, conversion and selectivities are plotted versus time on stream. For conditions see Experimental. (0, Conversion; 0, CH,OH; ., CH,; V, CO.) (a) 553 K, (6) 523 K, (c) 493 K. I I I b 1 I 1 I 1 20 30 4 0 50 60 7 0 t 26 Fig. 2. X-Ray diffraction patterns (Cu K,) of (a) the amorphous Pd,Zr, precursor and (h) the active catalyst prepared by in situ activation of the amorphous Pd,Zr, at 553 K (COJH,). a, Pd; b, solid solution of palladium-hydrogen ; c, ZrO,.Table 1.Textural and catalytic properties of catalysts ~ - ~~ A _ _ _ -~ ~- ~ - P Pd/ZrO, catalysts surface palladium prepared area area selectivity to selectivity to selectivity to /m2 g-' /m2 g-' T/K conversion (%) methane (YO) methanol (YO) co (Yo) from Pd,Zr, 16.8 493 K (COJH,) Pd,Zr, 28.4 523 K (COJH,) Pd,Zr, 30.3 553 K (CO,/H,) 5 % Pd/ZrO, 60.3 (impregnated) 5.6 493 483 473 463 443 5.7 523 513 503 493 483 473 463 443 4.4 533 523 51 3 503 493 483 473 463 443 0.5 553 533 523 513 503 493 48 3 15.8 13.4 10.9 8.5 4.3 21.7 19.0 15.1 12.8 10.0 8.3 5.8 2.9 24.9 19.1 16.5 14.6 12.4 9.8 7.6 5.6 2.7 27.3 12.3 8.3 5.3 3.5 2.2 1.5 43.1 31.3 23.1 17.6 11.3 77.2 57.8 43.4 31.2 24.1 18.0 14.1 8.2 77.3 63.5 46.4 34.6 25.6 19.6 15.5 12.4 4.2 100.0 93.9 89.0 83.3 80.3 70.7 63.0 30.2 37.4 41.5 43.3 43.8 6.5 16.1 24.7 32.6 36.2 38.9 39.9 41 .O 5.9 10.8 19.7 26.6 32.6 35.5 36.8 37.2 40.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 26.7 31.3 35.4 39.1 44.9 16.3 26.1 31.9 36.2 39.7 43.I 46.0 50.8 16.8 25.7 33.9 38.8 41.8 44.9 47.7 50.5 55.4 0.0 6. I 11.0 16.7 19.7 29.3 37.0 10.3 8.6 7.2 5.6 2.8 14.2 12.5 9.9 8.4 6.5 5.4 3.8 1.9 19.9 15.3 13.1 11.7 9.9 7.9 6.1 4.5 2.2 25.9 11.6 7.8 4.9 3.3 2. I 1.4 TOF overall /lo" s-' b Q 31004 Supported Palladium Catalyst - 4 F-7 - - 5 v Y fi -0 1.8 1.9 2.0 2.1 2.2 lo3 KIT Fig. 3. Arrhenius plots of CO, hydrogenation rates (turnover frequencies, TOF) of Pd/ZrO, catalysts prepared by in situ activation of amorphous Pd,Zr, and by the wet impregnation technique. 'I, Pd,Zr,, 493 K (COJH,); H, Pd,Zr,, 523 K (CO,/H,); a, Pd,Zr,, 533 K (CO,/H,); 0, 5 YO Pd/ZrO, (impregnated).High-resolution X.r.d. scans indicated that palladium was present in two forms: as disordered small particles and larger crystalline particles of ca. 12 nm mean size. The presence of the disordered small particles is also indicated by the large footings of the diffraction peaks in the wide scan X.r.d. pattern shown in fig. 2. Zirconium dioxide adopted predominantly the structure of baddeleyite15 and of a cubic phase.'' The reflections indicating the solid solution (reflections b in fig. 2), do not coincide with any known reflections of Pd, Zr, PdO, ZrO, or the bimetallic Pd-Zr alloy. However, they are in agreement with the reflections found by Yokoyama et al.,l who attributed them to a weakly bound Pd-0-Zr type complex oxide.Further support for the formation of a solid solution was obtained by heating the sample at higher temperature (873 K) in argon. In fact, the reflections indicating the solid solution disappeared under these conditions and the intensity of the Pd reflections increased. Finally, it should be noted that similar X.r.d. patterns were obtained for all in situ activated Pd,Zr, precursors. The catalyst prepared by impregnation consisted of crystalline palladium particles of about 28 nm mean size supported on well developed zirconia particles of tetragonal structure.'l Thus the morphology and structure of both the metallic palladium as well as of the ZrO, contained in both catalysts differed markedly. Table 1 contains the B.E.T. surface areas and the palladium metal surface areas measured for the differently prepared catalysts.We note that the in situ activation resulted in a marked increase in both the B.E.T. surface area and the metallic palladium surface area. The B.E.T. surface area of the amorphous alloy was less than 0.1 m2 g-l in the as-quenched state and increased to 15-35 m2 g-l depending on the pretreatment. In parallel the palladium surface areas increased from less than 0.03 to &6 m2 g-l. Scanning electron microscopy indicated that the increase in the surface area was a result of the formation of many cracks on the initially flat surface of the amorphous precursor. Plate 1 depicts the typical scaly texture of the surface which was seen with all catalysts prepared from amorphous Pd,Zr,.J.Chem. SOC., Faraday Trans. I , VoE. 85, part 4 Plate 1 Plate 1. Scanning electron micrograph showing the surface morphology of the active Pd/ZrO, catalyst prepared from amorphous Pd,Zr, by in situ activation under CO, hydrogenation conditions at 553 K. The corresponding activity behaviour is plotted in fig. I . A. Baiker and D. Gasser (Facing p . 1004)A. Buiker and D. Gasser 1005 I I I I I 0.01 0.05 0.0 9 0.13 contact time/min g ~ r n - ~ Fig. 4. Influence of the contact time on the product ratio CH,OH/CO for Pd/ZrO, catalyst prepared by in situ activation of amorphous Pd,Zr, at 553 K (COJH,). ., 463 K; V, 443 K; 0 , 483 K. Catalytic Properties in CO, Hydrogenation Preliminary tests with respect to possible influences on the activity and selectivity caused by interparticle and intraparticle mass and heat transfer limitations confirmed that such limitations could be ruled out under the conditions used in the catalytic tests.Table 1 shows the results of CO, hydrogenation performed over Pd/ZrO, catalysts derived from amorphous Pd,Zr, and over the conventionally prepared Pd/ZrO, catalyst. Reported reaction rates represent steady-state values, i.e. after steady state has been reached. Conversion and selectivities at different temperatures are compared. The activities of the differently prepared Pd/ZrO, catalysts are compared on the basis of measured turnover frequencies (TOF). Turnover frequencies were calculated as molecules of product formed per palladium surface atom and per second. The kinetic results of CO, hydrogenation are presented in the form of Arrhenius plots in fig.3. At higher temperatures the Arrhenius lines of the Pd/ZrO, catalysts prepared from the amorphous Pd,Zr, deviate from the straight-line behaviour due to the fact that equilibrium is approached. The most important results emerging from table 1 and fig. 3 are: (i) The conventionally prepared Pd/ZrO, catalyst exhibits only activity for methanation and CO formation (reverse water-gas shift reaction) and not for methanol formation. In contrast, all catalysts prepared from the amorphous Pd,Zr, show methanol formation increasing in importance with decreasing reaction temperature. Methane formation is prevalent with all catalysts at higher temperatures. (ii) CO, conversion rates expressed as TOF are for all catalysts prepared from Pd,Zr, more than an order of magnitude higher than those measured over the conventionally prepared Pd/ZrO, catalyst.The intrinsic activity of the catalyst derived from Pd,Zr, is highest with the catalyst obtained by in situ activation at 493 K, and lowest with the catalyst prepared by in situ activation at 553 K. Finally it is interesting to consider the dependence of the molar ratio of the products formed on the contact time. Fig. 4 shows the ratio CH,OH/CO for the catalysts prepared by in situ activation of the amorphous precursor at 553 K as a function of the contact time. This behaviour was observed with all Pd/ZrO, catalysts prepared from the 34-21006 Supported Palladium Ca fatyst amorphous precursor. We note that favoured by higher temperature and contact time the methanol formed converted to methane (CH,OH + H, -+ CH, + H,O).This is indicated by the decline of the curve (483 K) at higher contact time. The product ratio seems to approach zero as the contact time decreases. The behaviour plotted in fig. 4 indicates that methanol synthesis over the catalysts prepared from the amorphous alloy occurs primarily by hydrogenation of CO, and only to a small extent by hydrogenation of CO,. Carbon monoxide is formed from CO, by the reverse water-gas shift reaction (CO, + H, --* CO -b H,O). Discussion All catalytic tests showed that amorphous Pd,Zr, is only mildly active for CO, hydrogenation. This behaviour is mainly due to the small surface area of the precursor. Furthermore, the surface of palladium-zirconium alloys tends to be covered by zirconium oxide upon longer exposure to air.17 The activity develops as a consequence of drastic changes of the bulk and surface structure of the amorphous alloy.The most obvious changes are the partial crystallization of the amorphous alloy, the formation of palladium metal particles, and the oxidation of zirconium to zirconium dioxide. The surface becomes enriched with palladium as CO-chemisorption and preliminary X.P.S. measurements indicated. Segregation of palladium from the bulk onto the surface results in relatively large palladium metal surface areas. Similar metal segregation has been observed for copper in copper-zirconium al10ys.l~ Although the reason for the observed segregation is not clear, it seems to depend strongly on the disordered structure of the precursor, the gas atmosphere and the temperature.Our results indicate that the preparation method has a marked influence on the structural and chemical properties of the palladium-on-zirconia catalyst. The structural properties of both the metallic palladium as well as of the zirconia are different depending on the precursor and on the preparation. The most striking difference was found in the palladium phase. In the catalyst derived from amorphous Pd,Zr, some of the palladium formed a solid solution with hydrogen, leading to an increase of the Pd d-values of about 2%. A similar phase was not detected in the catalyst prepared by wet impregnation of zirconia. Another important difference is the existence of small disordered metallic palladium particles in Pd/ZrO, prepared from Pd,Zr,. The measured activities and the product distributions (table 1, fig.3) observed with the differently prepared Pd/ZrO, catalysts indicate that these structural differences had a marked effect on the catalytic behaviour. Based on the assumption that palladium is the active species, the different selectivities observed could be attributed to the existence of the palladium-hydrogen phase which was detected in the catalyst prepared from amorphous Pd,Zr, only. Another possible reason for the different selectivity behaviour may originate from the fact that the catalyst prepared from amorphous Pd,Zr, contained small disordered palladium particles besides larger crystalline particles.Fajula et aZ.lS pointed out that on small palladium particles CO is weakly adsorbed leading mainly to methanol formation, whereas on larger palladium particles CO is adsorbed more strongly and forms preferentially methane under CO hydrogenation conditions. Pods et al.19 suggested that cationic sites of Pd are essential for the production of methanol from synthesis gas. Although preliminary XPS studies of our catalysts did not give any indication of the presence of cationic palladium, their presence in the interfacial area cannot be ruled out. From the present investigation, it is not possible to draw final conclusions with regard to the importance of these three factors (presence of solid solution, particle size effect, cationic sites) for the explanation of the different selectivity behaviour of our catalyst preparations.Further work, including characterization of the catalyst surface underA . Baiker and D. Gasser I007 reaction conditions is necessary to assign the different nature of the active sites in these catalyst preparations. Conclusions Palladium-on-zirconia catalysts were prepared by exposing amorphous Pd,Zr, to CO, hydrogenation conditions. In the active catalyst part of the metallic palladium formed a solid solution with hydrogen. Two types of metallic palladium particles could be distinguished, small disordered particles and larger crystalline particles of ca. 12 nm mean size. The intrinsic activity for CO, hydrogenation of the palladium in the catalysts derived from amorphous Pd,Zr, was more than an order of magnitude higher than the activity of palladium in a conventionally prepared (wet impregnation) Pd/ZrO, catalyst containing only crystalline palladium particles of 28 nm mean size.Furthermore, Pd/ZrO, prepared from amorphous Pd,Zr, showed methanol, methane and CO as major products, whereas the conventionally prepared Pd/ZrO, catalyst exhibited only activity for methane and CO formation. The authors thank M. Maciejowski for valuable discussions with regard to the solid solution formation, and P. Wagli for the SEM investigations. Financial support by the ' Swiss National Science Foundation ', the ' Schweizerischer Schulrat ' and Lonza AG is kindly acknowledged. References 1 A. Yokoyama, H. Komiyama, H. Inoue, T. Masumoto and H. Kimura, Chem. Lett., 1983, 195.2 M. Shibata, N. Kawata, T. Masumoto and H. Kimura, Chem. Lett., 1985, 1605. 3 A. Yokoyama, H. Komiyana, H. Inoue, T. Masumoto and H. M. Kimura, Scripta Metallurgica, 1981, 4 A. Yokoyama, H. Komiyana, H. Inoue, T. Masumoto and H. M. Kimura, J . Catul., 1981, 68, 355. 5 M. Peuckert and A. Baiker, J. Chem. Soc., Faraday Trans, 1, 1985, 81, 2797. 6 G. Kisfaludi, K. Lazar, Z. Schay, L. Guczi, Cs. Fetzer, G. Konczos and A. Lovas, Appl. Surf. Sci., 7 H. Yamashita, M. Yoshikawa, T. Funabiki and S. Yoshida, J. Catal., 1986, 99, 375. 8 M. Shibata, Y. Ohbayaski, N. Kawata, T. Masumoto and K. Aoki, J. Catal., 1985, 96, 296. 9 Y. Shimogaki, H. Komiyama, H. Inoue, T. Masumoto and H. Kimura, Chem. Lett., 1985, 661. 15, 365. 1985, 24, 225. 10 M. A. Henderson and S. D. Worley, J . Phys. Chem., 1985, 89, 1417. I 1 ASTM Powder diffraction File 17-923, ed. Joint Committee on Powder Diffraction Standards, 12 S. J. Gregg and K. S. W. Sing, Surf. Colloid Sci. 1976, 9, 254. 13 Introduction to Characterization and Testing of Cutalysts, ed. J. R. Anderson and K. C. Pratt (Academic Press, London, 1985), p. 268. 14 D. Gasser and A. Baiker, submitted for publication. 15 ASTM Powder diffraction File 13-307, ed. Joint Committee on Powder Diffraction Standards, 16 ASTM Powder diffraction File 27-997, ed. Joint Committee on Powder Diffraction Standards, 17 F. Vanini, S. Buchler, Xin-nan Yu, M. Erbudak, L. Schlapbach and A. Baiker, Surf. Sci., 1987, 18 F. Fajula, R. G. Anthony and J. H. Lunsford, J . Catal., 1982, 73, 237. 19 E. K. Poels, R. Koolstra, J. W. Geus and V. Ponec, Stud. Sue- Sci. Catal., 1982, 11, 233. Pennsylvania, 1979, U.S. Bureau of Mines, Albany, Oregon. Pennsylvania, 1979. Lewis, General Electric Co., ANP Dept. Cincinnati 15, Ohio. Pennsylvania, 1979. G. Katz, J . Am. Ceram. Soc., 1971, 54, 531. 189/190, 11 17. Paper 8/02646G; Receiued 4th July, 1988

ISSN:0300-9599    

DOI:10.1039/F19898500999

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. I, 1989, 85(4), 1009-1018 Estimation of Group Dipole Moments from Surface Potential Measurements on Langmuir Monolayers Osvaldo N. Oliveira Jrt, D. Martin Taylor*, T. John Lewis, Susanna Salvagno and Charles J. M. Stirling Institute of Molecular and Biomolecular Electronics, University College of North Wales, Dean Street, Bangor, Gwynedd LL57 I UT Pressure-area isotherms and surface potential data are presented for octadecyl methyl sulphoxide (OMS) and ( + )-octadecyl p-tolyl sulphoxide (OTS). The surface potential measurements indicate very clearly that both compounds are anchored at the water surface by the SO group and that the plateau in the pressure-area isotherm of OTS is the result of significant molecular orientation. The Demchak and Fort model for relating group dipole moments to the surface potential of floating Langmuir monolayers is reviewed and found to be applicable to a number of compounds.However, the values deduced by these authors for the local permittivities are not appropriate for compounds with long aliphatic chains. By drawing on previously published work a new set of values has been deduced which seems to be more applicable to such compounds and to the sulphoxides investigated here. There has been a resurgence of interest in the Langmuir-Blodgett (LB) deposition technique in recent years' because of the possible applications of LB films in sensors,2 in electrical devices3 and as model membranes for mimicking biological membranes. The first step in the formation of an LB film requires that a monolayer of a suitable amphipathic molecule is formed on the surface of an aqueous subphase.Under compression, the initially expanded monolayer will condense into an ordered, two- dimensional structure, in which the long axis of the molecule is approximately normal to the water surface. The transition from the expanded to the condensed phase can be followed by monitoring the so-called pressure-area (n-A) isotherm. Simple, mono- functionalised molecules such as the long-chain alkanoic acids, generally have n-A isotherms which are easily and unambiguously interpreted. For molecules with more complex polar headgroups and certainly for bi- and poly-functionalised molecules, greater detail is observed in the n-A isotherm, e.g. the appearance of an extended p l a t e a ~ , ~ so that interpretation becomes less certain.Alternative measurements are then required if monolayer behaviour is to be correctly interpreted. One important approach is to measure the surface potential, AV, of the film, a quantity which depends on both the packing density and the orientation of the molecules in the film. The technique has been applied to floating monolayers since the early part of the century and although in widespread use, the development of a satisfactory theory relating the magnitude of AV to known molecular and group dipole moments has been slow. In this paper we show that by following an approach suggested by Demchak and Fort6 in 1974 good agreement can be obtained between the dipole moments deduced from surface potential measurements on floating monolayers and those measured by other techniques.t On leave of absence from Instituto de Fisica e Quimica de Sao Carlos, USP, Brasil. 10091010 Group Dipole Moments Theory The floating monolayer is treated usually as a uniform assembly of molecular dipoles which give rise to a polarisation within the layer. The component of polarisation normal to the plane of the film, P,, gives rise to a potential difference AV across the layer. By analogy with a parallel-plate capacitor and remembering that polarisation can also be defined as the dipole moment per unit volume, it is easily shown that Pn is given by P, = E E ~ A V / d = p J A d where E is the relative permittivity of the film, E, is the permittivity of free space, d is the film thickness, p,, is the component of the molecular dipole moment normal to the plane of the monolayer and A is the average area occupied by the molecule.Eqn ( 1 ) then yields the Helmholtz equation7 for an unionised monolayer, p,, = E ~ A A V in which it has been assumed that E = 1 and A V is the difference in potential between a clean water surface and one on which the monolayer is present. Dipole moments calculated by substituting experimental values of A V into the Helmholtz equation are usually much lower' than the true moments measured by other methods, so that ,un is often referred to as the apparent dipole moment. Various reasons have been suggested for the discrepancy: (i) polarisation of the subphase water may oppose the dipole moment of the monolayer ; (ii) the orientation of the molecular dipoles in the condensed film may differ from that assumed; (iii) the relative permittivity of the monolayer is likely to be greater than unity.Adam et al.' suggested 5 < E < 10 should be used. Improvements to the Helmholtz model were made by Davies," who considered the condensed monolayer to be analogous to a three-layer capacitor (fig. 1). The layers correspond broadly with the three main contributions to A V. In the subphase, a moment ,ul represents the polarisation of the water molecules induced by the presence of the monolayer. The hydrophylic headgroup is assigned a moment p,, while the hydrophobic tail (which may be functionalised) has assigned to it a moment p,. In a further refinement, Demchak and Fort' assumed that each of these layers had a different relative permittivity to account for the different polarisabilities of the medium immediately surrounding the dipoles.For example, E, is likely to have a value close to that for bulk water (ca. 80). E, will be much smaller and characteristic of the hydrophobic tail, while E , will lie somewhere between the two values. Eqn (2) for the apparent dipole moment should then be written as Demchak and Fort' tested their assumptions by measuring the surface potential of a series of difunctional p-terphenyl compounds in which the headgroup moment (p2) and the tail end-group moment (p,) were systematically changed. Assuming that all other variables remain constant, they were able to show that if (&/E,) = 40 mD,? E, = 7.6 and E, = 5.3 good agreement with the measured values of AV was obtained by substituting the known values of ,u2 and p 3 into eqn (3).Since E, is determined by the polarisability of the hydrophobic layer, which in their case was formed entirely from vertically stacked terphenyl moieties, a value of 5.3 is not unreasonable. Subsequently, Demchak and Fort proceeded to apply these results to predict the headgroup conformations of a series of condensed aliphatic monolayers. However, we believe that a value of E, closer to that for long-chain paraffins (ca. 2-4) would have been more appropriate for such compounds. This we have confirmed by using the Demchak and Fort approach to calculate E, for a series of w-halogenated alkanoic acids and t 1 D z 3.335 64 x lop3' C m.0.N . Oliveira et al. 101 1 hydrophobic tail _ - - - - - - Fig. 1. The Demchak and Fort three-layer capacitor model for a floating Langmuir monolayer. The contribution of the dipoles in each layer to the surface potential is determined by the local permittivity of each layer [see eqn (3)]. Table 1. Reported values for the normal component of the terminal dipole p,(-CH,X) for various halogenated long-chain compounds together with estimates of E , based on the Demchak and Fort model" compound P,(---CH,X)/D &3 ref. 16-chlorohexadecanoic acid 16-bromohexadecanoic acid 1 6-bromohexadecanoic acid 1 6-bromohexadecanoic acid 16-iodohexadecanoic acid 1 8-bromooctadecanoic acid pH 2.2 1 8-fluorooctadecanoic acid (pH 5.8 I p H 8.6 18,18.18-trifluorooctadecanoic acid 1.87-2.16 1.8 1-2.20 1.81-2.20 1.8 1-2.20 1.62-2.12 1.8 1-2.20 1.85-2.05 1.85-2.05 1.85-2.05 1.85-2.05 1.85-2.05 1.85-2.05 3.44-3 3 8 2.74-3.22 2.74-3.22 2.71-3.19 2.47-3.08 3.12-3.67 2.57-2.80 2.30-2.5 1 2.3 2 -2.53 2.24-2.44 2.68-2.92 2.13-2.32 11 11 10 32 11 11 13 13 13 23 23 23 a The range of values given for each compound reflects the uncertainty in the reported values for group moments.amines using the surface potential data available in the literature. In the calculation, it was assumed that the C-X dipole of the terminal -CH,X moiety (where X is a halogen) was inclined at half the tetrahedral angle (i.e. 54 O 44') with respect to the water surface as suggested by Bernett et d . l l (see scheme 1) and that the group moments have the values given by Smyth.12 In addition it was assumed that the C-H group moment was 0.4 D, the carbon being negatively charged.13 The ranges of values of p3 thus calculated for these halogenated compounds are listed in table 1 together with the corresponding values deduced for E,.As can be seen E, lies within the range 2.13-3.88 with a mean value of 2.8. This is ca. half that determined by Demchak and Fort for their terphenyl compounds and much closer to that expected for saturated aliphatic chains. There has been much discussion in the literature concerning the relative importance of the three contributions to the observed surface p0tentia1.l~ For example, Schulman and Hughes' suggested that the headgroup was dominant. Other authors have attributed an important role to the subphase.15 More recently, Vogel and Mobius'' suggest that the main contribution comes from the dipole located at the air/water interface, i.e.p3. This1012 Group Dipole Moments H2 c - %5? .o 44' 2 Scheme 1. proposal is certainly consistent with experimental result^^'-'^ on monolayers formed from single paraffinic chain compounds where changing the headgroup had little effect on p,. Furthermore, the Demchak and Fort model provides a theoretical justification for their assertion because E , is smaller than either E , or E , so that the contribution to A V from dipoles at the air/water interface will be weighted more heavily than from dipoles in the headgroup or in the subphase. However, the results obtained from experiments on ether and ester phospholipids2' and on octadecylamine m d stearic acid mon01ayers~~-~~ show that the local permittivities are not always sufficiently different in magnitude to ensure that the headgroup contribution to A V is insignificant.Further evidence for this latter view is presented in the present paper, which describes an investigation into the monolayer behaviour of two long-chain sulphoxides. The results obtained confirm also the validity of the Demchak and Fort model for the interpretation of surface potential data, though it is concluded that their values of local permittivities are not generally applicable. Experiment a1 The compounds studied in this work were octadecyl methyl sulphoxide (OMS) and (+)- octadecyl p-tolyl sulphoxide (OTS). These particular compounds were chosen from several that had been screened during a wider investigation into the behaviour of sulphur-containing molecules in Langmuir-Blodgett films and their syntheses have been described fully elsewhere.24 Solutions of the compounds in chloroform (1 mg ~ m - ~ ) were freshly prepared prior to spreading on the Langmuir trough.The trough arrangement was of the continuous-barrier type described by Blight et and was placed on an antivibration table in a clean room. Pure water for the trough was obtained from an Elgastat system which comprised reverse osmosis, activated carbon and ion-exchange cartridges. The temperature of the water was held constant by resting the trough on a thermostatically controlled metal base-plate. The pH of the water was controlled by adding appropriate quantities of HCl or NaOH (AR).The surface pressure of the floating monolayer was measured to an accuracy of 0.1 mN m-' using a Wilhelmy plate connected to an electrobalance. Simultaneously, the surface potential, A V, was recorded using a vibrating plate26 (frequency 300 Hz) located ca. 2 mm above the water surface. The amplitude of vibration of the plate was insufficient to disturb the floating monolayer. Both the vibrating plate and reference electrode located in the water subphase, were made from platinum foil and were cleaned regularly by boiling in ultra-pure water. The potential measurements were reproducible to 10 mV and the system response time was less than 1 s on all ranges.0. N . Oliveira et al. 1013 10.8 area per molecule/A2 Fig. 2. n-A and AV-A plots for octadecyl methyl sulphoxide.Also shown is the apparent dipole moment pn calculated from AV using the Helmholtz equation. T = 18.0 "C, pH 5.6. Results (i) Octadecyl Methyl Sulphoxide The z-A isotherm obtained for OMS during a compression/expansion cycle (fig. 2) is identical to that presented p r e v i o ~ s l y ~ ~ and shows one small feature at ca. 40 mN m-l. Isotherms have now been obtained over wider temperature and pH ranges (14-22 "C and 3-10.7, respectively) than previously explored but no significant changes in monolayer behaviour were observed. Surface potential rises smoothly during film compression. The $orresponding dipole moment, p,, decreases linearly until the area per molecule is ca. 30 A2, the point at which surface pressure begins to rise.I,t then decreases more rapidly on further compression. In the expanded phase ( A > 50 A2), the surface potential varied by as much as 100 mV from film to film, possibly indicating the presence of islands28 of condensed material. In the fully compressed monolayer on tbe other hand, A V was accurately reproducible and at the smallest molecular area (19.5 A2) gave an apparent dipole moment of 0.33 0.01 D. (ii) Octadecyl p-Tolyl Sulphoxide Fig. 3 is the z-A~urve for OTS showing a rise in monolayer pressure at large area per molecule (ca. 90 A2). At ca. 60 A2 a plateau is observed when a marked reduction in area produces very little change in surface pressure. On further compression, the pressure rises rapidly once again. When a fully compressed monolayer is expanded, significant hysteresis is observed. 27 However, if expansion was commenced within the plateau region very little hysteresis occurred.Even when hysteresis was present, the compression curve was completely reproducible, suggesting that the process(es) producing the effect can relax totally when the monolayer is expanded to zero surface pressure. From fig. 3 it can be seen also, that during compression A V increases smoothly until1014 Group Dipole Moments 0 40 80 120 area per molecule/A* Fig. 3. n-A and At/-A plots for octadecyl p-tolyl sulphoxide. The apparent dipole moment calculated from the Helmholtz equation decreases to zero as the monolayer is compressed into the condensed phase. T = 14.9 "C, pH 5.6. I I I area per molecu1e/A2 Fig. 4. Effect of temperature on the n-A isotherms of octadecyl p-tolyl sulphoxide.pH 5.6. the onset of the plateau in the ;n-A curve. On further compression, A V decreased rapidly reaching 0 f 50 mV in the fully compressed film. As far as we are aware, this is the first example to be reported in which A V decreases as the monolayer is compressed. The corresponding curve for p n shows three distinct phases. In the expanded phase, pn decreases almost linearly with decreasing molecular area. When the pressure of the0. N . Oliveira et al. 1015 monolayer begins to rise, the decrease in p, becomes much slower. Finally, when the plateau is reached a rapid downturn occurs with pn falling linearly to zero. Fig. 4 shows that the plateau onset moves to lower molecular areas and higher pressures as the subphase temperature is raised from 15 to 28 "C.Changing the pH within the range >lo did not cause any significant changes in the results. Discussion No effect of pH was observed in these experiments and so it is concluded that the monolayers are not ionised. Consequently, there will be no double-layer (zeta-potential) contribution to A V so that eqn (3) should apply to the surface potential data. The OMS compouqd displays an almost featureless n-A isotherm yielding a molecular area of 19.8 A2, indicating a vertical orientation with the -SOCH3 moiety in the water subphase. The surface behaviour of OTS is unusual in that the plateau in the n-A isotherm is accompanied by a rapid decrease in AV. Several mechanisms could account for the appearance of such a plateau, for example: (i) dissolution of the monolayer; (ii) collapse, stacking or crystallisation of the monolayer; (iii) rearrangement of molecules into a new low-area phase.The first two possibilities can be ruled out in the present cases since the attendant loss of material from the monolayer would have resulted in a shift of the n-A isotherm during successive compression-expansion cycles. Such a shift was not observed, although needle-shaped crystallites were present in films deposited onto substrates at pressures higher than the plateau value.27 The proposal by Lewis et al.27 that molecular rearrangement is the cause of the plateau in this compound is supported by the surface potential measurements presented in fig. 3. A V would be expected to remain constant if dissolution or crystallisation were occurring because under most of the area of the vibrating plate sensor, molecular packing would remain unchanged.The sudden decrease observed in AV at the onset of the plateau in fig. 3 suggests significant rotation of the strongly polar SO group. Energetically, an alignment of the SO dipoles along the water surface is particularly favourable since it leads to strong forces of attraction between the dipoles. In this configuration the p-tolyl moiety would be forced into the water and the aliphatic chains would assume a vertical orientation in the air phase, as suggested previously. Fig. 4 shows that this transition moves to higher pressures and to lower areas per molecule as the subphase temperature is raised.This is to be expected since an increase in the thermal energy of the monolayer would make it more difficult for the SO dipoles to assume an ordered alignment along the water surface. The origin of the small maximum in the curve for T = 28.2 "C is not clear, although a dependence on compression rate suggests that it may be related to the rate at which the molecules are able to reorientate. At the smallest molecular area when all the molecules are vertically oriented, the SO dipoles lie parallel to the water surface and make no contribution to pn. Thus, the positive moment of the terminal -CH, of the aliphatic chain would be compensated by the negative moment of the p-tolyl moiety leading to an almost zero vertical moment for the molecule, as indeed was observed in fig.3. The dipole of the p-tolyl moiety, p2($CH3) is assumed to be inclined at ca. 20" to the vertical and is the only contribution to p2 which, therefore, has a value of -0.35 D.29 Using the experimentally determined values for pn and assuming that E, = 2.8 in eqn (3) we may write the following relationships for the condensed phases of our compounds : Pn(OMS) = ( & / E l ) b2(CH,SO)/%I + b,(CH)/2*8I = 0.33 D (4) &(OTS) = @ i / E i ) 4- [Il2(#cH3)/E21+ L%(CH)/2.81 = 0.00 D* ( 5 )1016 Group Dipole Moments Table 2. The apparent dipole moments of various long-chain compounds of the form R-X where X is the headgroup and = C17H35, Or ‘1gH3ga apparent dipole moment, pn/D calculated headgroup (X) experimental this work ref. (6) cis cis ester 0.250 0.230 0.250 cis methyl ether 0.208 0.197 0.220 cis alcohol 0.215 0.212 0.230 cis cis acid 0.210 0.184 0.210 trans trans acid 0.210 0.2 10 0.230 cis amine 0.200 0.079 0.120 a The nomenclature for the headgroup conformations is that used in ref. (6).The experimental values are also quoted from ref. (6) and were obtained by applying the Helmholtz equation to surface potential data. The calculated moments were derived by combining known group moments with the parameters ~ J E ~ , E, and E, deduced here. Also given for comparison are the values calculated by Demchak and Fort.6 In principle, these equations may be solved to yield p,(CH,SO). However, there remain two other unknowns, namely and E,. In a first approximation, we assume that they are constants independent of the nature of the headgroup so that they may be evaluated from published data on stearic acid.,’ Thus Pu,(SA) = OL1/&1) b2(CO2H)/EJ -t b!,(CH)/2.81 = 0.21 D (6) and assuming the carboxyl moiety to be in the trans(trans) configuration, so that p2(C02H) = 0.99 D6 then eqn (5) and (6) yield (pl/cl) = -65 mD and E , = 6.4. Substituting these values into eqn (4) gives p,(CH,SO) = 1.76 D.The dipole moments of SO and (CH,),SO are 1.55 and 3.96 D, re~pectively,~~ suggesting that the parameters we have deduced are reasonable. The large positive moment of the CH,SO moiety has interesting implications for previously reported measurements on sulph~xides.~~ In this earlier work, both the SO and CO,H dipoles contributed positively to pn in the expanded phase of the midchain sulphoxide acid C,oH21SOCloH,oC02H, resulting in a large apparent moment (1.2 D).In the expanded phase of the near-headgroup sulphoxide acids C,,H,,SO(CH,),CO,H with n = 1 or 2, pn is small and negative, suggesting an antiparallel configuration of the SO and CO,H dipoles and the presence of strong intramolecular interactions. In the condensed phases of these acids when the molecules are vertically oriented, the SO dipole is parallel to the water surface and does not contribute to AV. Therefore, pn should asymptote to the behaviour of stearic acid, as indeed is observed in practice. As a further check on the validity of the parameters deduced above we have repeated the calculations initially carried out by Demchak and Fort on a series of unionised aliphatic monolayers in which the headgroup was systematically changed.6 The new values of pn obtained for these compounds are compared with experimentally obtained values in table 2.Also listed are the values calculated by Demchak and Fort using their parameters @ l / ~ l = 40 mD, E , = 7.6 and E , = 5.3). Within the precision of the calculation (ca. 10mD) both sets of parameters appear to provide equally good0. N . Oliveiru et al. 1017 agreement with experiment, except for the amines. Had we assumed above that the acid headgroup was in the cis(cis) configuration favoured by Demchak and Fort, then poorer agreement would have resulted for most compounds. We believe that the large discrepancy with the amines is evidence that both E , and possibly (,ul/cl) depend on the nature of the headgroup.A similar analysis confined to published data21-23 on octadecylamine, 18-fluorooctadecylamine, stearic and 18- fluorostearic acids yields values of E , in the range 1.84.3. These are much lower than expected for the acids (lower than E , in some cases), indicating that the amine headgroup has a marked effect on both ( , U ~ / E ~ ) and E,. Conclusions The surface behaviour of the two sulphoxides has been studied using a combination of pressure-area and surface potential measurements. The surface potential data have confirmed that the plateau in the n-A isotherm of octadecyl p-tolyl sulphoxide is indeed the result of molecular orientation. In the condensed phase, the molecule adopts a vertical orientation with the sulphoxide dipole lying at and parallel to the water surface, a configuration which leads to an almost zero surface potential.By reviewing published data on a number of compounds we have shown that the surface potential of floating monolayers can be related quantitatively to known group dipole moments using the Demchak and Fort model. However, caution is necessary when assigning values to the local permittivities appearing in the model. For example, their value of 5.3 for E, was obtained from measurements on terphenyls. We have shown that a more appropriate value for compounds with long aliphatic chains is 2.8, a value which fits well with reported data on halogenated fatty acids. Obviously, the higher value may well apply to compounds in which aromatic groups are incorporated into, or are attached to the main chain.Contrary to the conclusions of some authors, it was shown that the headgroup can contribute significantly to the surface potential of Langmuir monolayers, cf. the different effects the two sulphoxide headgroups used in the present work have on surface potential behaviour. The assumption by Demchak and Fort that and E, are constants and not affected by the nature of the headgroup was found to be reasonable for acid, ester, ether and alcohol headgroups, although the values deduced here for these parameters (-65 mD and 6.4, respectively) differ from theirs. Based on our values the dipole moment of the CH,SO group is deduced to be 1.76 D. The behaviour of the amine headgroup shows that while the Demchak and Fort model may be generally applicable, it cannot be assumed that ( ~ J E ~ ) , E, and E , are universal constants.Rather, particular families of compounds will have almost constant values but other families will have a different set of values. The preparation of samples for this work was carried out with financial support from the S.E.R.C. (grant no. GR/C64592) which is gratefully acknowledged. One of us (0. N. 0. Jr) also thanks FAPESP and the ORS for the award of a Research Studentship and financial assistance. We thank the referees for helpful remarks. References 1 Proc. 1st Int. Con$ on Langmuir-Blodgett Films, Thin Solid Films, 1983, 99, nos. 1-3; Proc. 2nd Int. Con$ on Langmuir-Blodgett Films, Thin Solid Films, 1985, 132-134. 2 S. Baker, G. G. Roberts and M. C . Petty, IEE Proc., 1983, 1301, 260.3 P. S . Vincett and G. G. Roberts, Thin Solid Films, 1980, 68, 135. 4 C. N. Kossi and R. M. Leblanc, J. Colloid Sci., 1981, 80, 426; M. C . Wilkinson, B. N. Zaba, D. M. 5 V. Vogel and D. Mobius, Thin Solid Films, 1985, 132, 205. Taylor, D. L. Laidman and T. J. Lewis, Biochim. Biophys. Acta, 1986, 857, 189.1018 Group Dipole Moments 6 R. J. Demchak and T. Fort Jr, J. Colloid Sci., 1974, 46, 191. 7 J. H. Schulman and A. H. Hughes, Proc. R. SOC. London, Ser. A , 1932, 138, 430. 8 G . L. Gaines Jr, Insoluble Monolayers at Liquid-Gas Interfaces (Interscience, New York, 1966). 9 N. K. Adam, J. F. Danielli and J. B. Harding, Proc. R. SOC. London, Ser. A, 1934, 147, 491. 10 J. T. Davies and E. K. Rideal, Can. J. Phys., 1955, 33, 947. I 1 M. K. Bernett, N. L. Jarvis and W. A. Zisman, J. Phys. Chem., 1964, 68, 3520. 12 C. P. Smyth, Dielectric Behaviour and Structure (McGraw-Hill, New York, 1955). 13 M. K. Bernett and W. A. Zisman, J. Phys. Chem., 1963, 67, 1534. 14 Ref. 8, p. 192. 15 J. Goldfarb, S. Gonzalez and B. A. Pethica, Bol. SOC. Chilena Quim, 1962, 12, 24. 16 V. Vogel and D. Mobius, J. Colloid Sci., 1988, in press. 17 J. Marsden and J. H. Schulman, Trans. Faraday Soc., 1938, 34, 748. 18 A. E. Alexander and J. H. Schulman, Proc. R. Soc. London, Ser. A , 1937, 161, 115. 19 T. Fort Jr and A. E. Alexander, J. Colloid Interface Sci., 1959, 14, 190. 20 F. Paltauf, H. Hauser and M. C. Phillips, Biochim. Biophys. Acfa, 1971, 249, 539. 21 J. J. Betts and B. A. Pethica, Trans. Faraday Soc., 1956, 52, 1581. 22 J. Glazer and M. Z. Dogan, Trans. Faraday Soc., 1953, 49, 448. 23 H. W. Fox, J. Phys. Chem., 1957, 61, 1058. 24 C. Georges, T. J. Lewis, J. P. Llewellyn, S. Salvagno, D. M. Taylor, C. J. M. Stirling and V. Vogel, 25 L. Blight, C. W. N. Cumper and V. Kyte, J. Colloid Interface Sci., 1965, 20, 393. 26 A. Noblet, H. Ridelaire and G. Sylin, J. Phys. E, 1984, 17, 235. 27 T. J. Lewis, D. M. Taylor, J. P. Llewellyn, S. Salvagno and C. J. M. Stirling, Thin Solid Films, 1985, 28 W. D. Harkins and E. K. Fischer, J. Chem. Phys., 1933, 1, 852. 29 M. Paluch and P. Dynarowicz, J. Colloid Interface Sci., 1987, 115, 307; J. W. Smith, Electric Dipole Moments (Butterworths, London, I955), p. 242. 30 N. K. Adam and J. B. Harding, Proc. R. Soc. London, Ser. A , 1932, 138, 411. 31 CRC Handbook of Chemistry and Physics, ed. Weast (CRC Press Inc., Boca Raton, Fl., 64th edn, 1983). 32 M. Gerowich and A. Frumkin, J. Chem. Phys., 1936, 4, 624. J. Chem. Soc., Faraday Trans. I , 1988, 84, 1531. 133, 243. Paper 8/02861C: Received 15th July, 1988

ISSN:0300-9599    

DOI:10.1039/F19898501009

出版商:RSC    

年代:1989

数据来源: RSC