ANALYST, AUGUST 1992, VOL. 117 1251 Mobile Species Uptake by Polymer-modified Electrodes* A. Robert HillmanJ David C. Loveday and Marcus J. Swann School o f Chemistry, University o f Bristol, Bristol BS8 ITS, UK Stanley Bruckenstein Department of Chemistry, State University o f New York at Buffalo, Buffalo, NY 14214, USA C. Paul Wilde Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada KIN 6N5 The use of mobile species uptake by a polymer-modified electrode as a probe of solution composition is discussed. The over-all mobile species exchange process between the polymer and solution phases is monitored gravimetrically, using the electrochemical quartz-crystal microbalance (EQCM). The influence of electroneutrality and activity constraints is considered under both equilibrium and transient conditions.The generality of detection by mass and the high sensitivity of the EQCM have analytical advantages. Selectivity requires separation of the total mass change into single species' components. The thermodynamic and kinetic approaches to the problem of selectivity are explored. Keywords: Polymer-modified electrodes; mobile species; permselectivity; quartz crystal microbalance; redox polymer Introduction The Problem The electrochemical quartz-crystal microbalance (EQCM) is presently finding considerable application as a probe of interfacial processes.1-4 In this paper its use for monitoring the exchange of mobile species between a polymer-modified electrode and its bathing solution is discussed. The key problem of deconvoluting the observed signal into its com- ponents is addressed, each being associated with a single mobile species.This problem exists as a consequence of the generality of detection by mass and the application of the EQCM to analytical problems requires its solution. Initially, the nature of the problem is illustrated by showing how a variety of phenomena alter the simple pattern of exchange predicted by electroneutrality arguments. Secondly, two methods are proposed for the separation of mobile species transfer processes. One approach involves the application of thermodynamic arguments to (pseudo-) equilibrium" data; the other exploits kinetic measurements. Then the principles of these methods are demonstrated by applying them to mobile species transfers for model systems. Background The modification of electrode surfaces with polymer films allows us to tailor chemically the properties of the electrode/ electrolyte interface.5.6 This facility has been exploited in a number of ways to give electrodes with analytical utility.Potentiometric responses have been reported for H+ using poly( 1,2-diaminobenzene) ,7 K+ using poly( acrylic acid) ,8 and SCN - using pol y [ trialk yl (vin ylbenz yl) ammonium] .8 Polymer film immobilization of suitable mediators allows ampero- metric detection of solution species. Choice of mediator is based upon the selectivity and kinetics of its reaction with the target species. The numerous cases (see refs. 5 and 6 for reviews) of amperometric detection are exemplified by the use of immobilized iridium mediators for nitrite9 and ascorbatelo detection.Selectivity and sensitivity of amperometric sensors are strongly related to partition of the target species into the polymer film. Strictly, the solution species need only interact with the film at the polymer/solution interface; however, the advantages of increasing film thickness576 are then not realized. In this paper, in situ nanogravimetric measurements are used to examine the solution/polymer phase exchange process. Insight into this process should be helpful both in the rational design of modified electrodes, and the development of alternative analytical strategies to existing potentiometric and amperometric approaches. Electroanalytical Strategy Generally, the solution composition determines the uptake by the polymer of mobile species, the target species and possible interferents.The basis of a sensor is inherent in this uptake if (i) there exists a means for monitoring its extent, and (ii) there is a known relationship between it and the analyte composi- tion. In the first instance, the response of the EQCM is related directly to the uptake process. In the second instance, two general interpretational issues arise: that of separating the total transfer process into its components and that of exploiting target species preconcentration. In the latter respect, polymer-modified electrodes appear promising, e. g . , partition coefficients in excess of 106 have been reported for ion-exchange into Nafion.11 Simple strategies for mobile species uptake are typically based on a single type of interaction.One approach, ion exchange into charged polymer films,12 has been utilized for the determination of cationic11 and anionic13 metal species, and neurotransmitters.14 Selectivity is primarily (see below) governed by ion charge-type.13 Another approach builds on the fact that coordination complexes have been electropoly- merized onto electrodes. 15 Electrodeposition of a metal-free polymer yields an electrode capable of complexing metal ions from the bathing solution.16.17 Here, selectivity is (primarily) governed by complexation chemistry. The chemical process * Presented at the meeting on Analytical Applications of Chemi- T Present address: Department of Chemistry, University of cally Modified Electrodes, Bristol, UK, January 7-8, 1992.Leicester, University Road, Leicester LE1 7RH, UK. a In some situations, a variety of metastable states exist. By (pseudo-)equilibrium we mean that on the timescale of the measure- ment one of these metastable states predominates, so that the system appears to be in equilibrium. On a different timescale another metastable state might predominate.1252 ANALYST, AUGUST 1992, VOL. 117 may be as simple as protonation7 for a pH sensor, or as sophisticated and specific as antibody interactions.18 Usually more than one type of interaction will contribute to selectivity patterns.11.14 In this paper, the ion-exchange and coordination uptake strategies are examined and it is shown how a single polymer may exhibit both characteristics. Partitioning of ion and neutral species is not independent, as there is a general thermodynamic requirement19 that neutral species (notably solvent) transfer accompanies ion transfer into a solvent-containing polymer film.Electrochemical Quartz Crystal Microbalance (EQCM) The quartz crystal microbalance (QCM) technique measures the variation (Af) in resonant frequency of a quartz crystal oscillator from its base value cfo) that accompanies a change (AM) in the mass attached to the crystal. When the additional mass is small and rigidly coupled20 (Af/Hz) = -(2/pv)f02(AM/g cm-2) where p is the density of the quartz and v is the wave velocity in the quartz. Characterization21722 of crystal oscillation in a liquid prompted in situ electrochemical application of the QCM. The EQCM technique has since been applied to a variety of electrochemical problems.1-4 Here we focus on its ability to follow quantitatively mobile species exchange between a polymer film and its bathing electrolyte.23-37 We draw attention to four attractive characteristics of the EQCM.The first is high sensitivity. For the 10 MHz AT-cut quartz crystals used, eqn. (1) shows that a frequency change of 1 Hz (routinely measurable) corresponds to a mass change of 4.4 ng cm-2. This areal density corresponds to 10 pmol for exchange of a solution species of molar mass 100 into the electrodes used in this work (area 0.23 cm2). The second is known sensitivity. Conversion of measured signal (An to moles of partitioning species only requires eqn. (1) and molar mass; the latter conversion is free of matrix effects unlike, for example , calculations based upon molar absorption coeffi- cients or cross-sections in spectroscopic methods.The third is in situ applicability, which allows analytical measurements to be made under optimum electrochemical control conditions. Certain powerful structural techniques, such as Auger,38 are inherently ex situ methods. This restricts their application to problems where removal from solution is acceptable. The fourth is the time resolution of the EQCM, which enables dynamic measurements to be made. This relatively unex- plored capability is crucial to one of the strategies described to extract a single species contribution from the over-all mass response. All the EQCM data analyses used are based on eqn. (l), which is appropriate for rigid films.If the polymer is extensively solvent-swollen, this rigidity requirement may not be satisfied.39 The data presented here are for rigid films. Although the QCM has been used for sensing biologically relevant species such as glucose40 and proteins,*s the full capabilities of the EQCM for biosensors have not yet been realized. A primary reason for this is the need to identify that component of the total response that is caused by the target species. This problem is addressed here. Theory The electrochemical potential, pj, of a species j in a single phase is given by: p, = p0, + RT In(.,) + z,F@ = p, + z,F@ pi is the chemical potential component, which is all that need be considered for neutral species (z, = 0). At equilibrium, the (2) - - p, values for each species in polymer and solution phases will be equal.p, imposes two constraints on each mobile species. They are associated with the two terms on the right-hand side of eqn. (2), and are the activity and electroneutrality constraints, respectively. Generally, these two constraints are fulfilled on different timescales. The EQCM responses on long timescales (typically >lo0 s) reflect satisfaction of both constraints by populations of all mobile species. The EQCM responses on shorter timescales (typically (1 s) are determined not by equilibrium paramet- ers, but rather by the rate(s) of mobile species transfer(s). These considerations suggest two procedures, one thermody- namic the other kinetic, for separating the total mass response into the components associated with individual transfer processes.Thermodynamic Aspects of the Exchange Process The problem The objective is to determine the population change of an individual species (the target species) within the polymer film. AM necessarily pertains to a sum of all mobile species population changes. We now describe a procedure for extraction of individual species contributions from the over-all gravimetric response for a modified electrode immersed in a solution of a single electrolyte (C+A-). This case involves three species [counter ion, co-ion and solvent (solv)], and is a first step towards solving the general problem. The analysis is presented for reduction of a polymer containing unipositively charged redox sites (the ‘redox’ couple): [(ox+A-).a(C+A-).P(s~lv)]~ + e- e [red.(cx - G)(C+A-).(B - E)(so~v)]~ + 6C,+ + (6 + l)As- + Esolv, (3) Subscripts p and s denote polymer and solution phases, respectively. The stoichiometric coefficients (Y and P b 0, and 6 and E may be of either sign or zero; none of these quantities need be integral.19 It is emphasized that the analysis is thermodynamic and pertains to over-all changes between (pseudo-)equilibrium states.The total mass change is the sum of three components, associated with counter ion, solvent19 and salt6 transfer: (4) Extraction of the three unknowns (population changes of counter ion, solvent and salt) requires three pieces of information. These are provided by AM (the experimental data), the electroneutrality constraint (commonly recog- nized5.6) and the concept of ‘constant solvent transfer’.Recently the background to this latter concept was pre- sented,41 and its application is demonstrated here for the first time. Counter ion contribution Electroneutrality dictates that the difference in counter- and co-ion population changes6 must balance the electronic charge injected. The counter ion contribution, AMcounter, to the mass change is given by where Q (/C cm-2) is the charge passed per unit area, and m and z represent the molar mass and charge number of the designated species (counter ion in this instance). b Electroneutrality demands that transferred co-ion be ac- companied by an equivalent amount of counter ion; these can be considered as ‘salt’.ANALYST, AUGUST 1992, VOL. 117 1253 Solvent contribution Partition coefficients (K) for solvent between the polymer (in its fully reduced and oxidized states) and the solution are defined in terms of activities (a) by - SOIV,o~n T, S0IVred Ksolv, red - asolv, red/asolv, soh (6) SOIVs,~n 2 soIvox (7) - Ksolv, ox - asolv, ox/aso~v, soln We have shown41 that - - - (Ksolv, ox/Ysolv. ox)} (8) where V is the volume of the polymer phase and y is the activity coefficient for the designated species and phase.For dilute solutions, when the activity coefficients are unity, the limiting expression for the change in solvent concentration in the polymer film upon reduction is AMsolv = msolv csolv, soh (&oh, red - Ksolv, ox) (9) Regardless of its oxidation state, the polymer is equilibrated with the same solution phase, so we can combine eqns. (6) and (7): soIvox * S0IVred Ksolv, red/ox = asolv, red/asolv, ox (10) This contains the key idea that the ratio of solvent activities in the oxidized and reduced states of the polymer is constant and does not depend on the solvent activity in the bathing solution.The solvent activity coefficient in the polymer phase cannot change in the permselective regime [a = 6 = 0 in eqn. (3)], so the redox-induced change in the polymer phase solvent population is independent of electrolyte concentration in the bathing solution .c As the electrolyte concentration is increased, salt will be partitioned into the film (permselectiv- ity failure: a, 6 # 0) and the activity coefficient of solvent in all phases will change. However, as the polymer phase (whatever its redox state) is equilibrated with the same solution phase [see eqn.(lo)], the ratio of the activity coefficients will not change significantly. This leads to the conclusion that the amount of water transferred is the same in the permselective and non-permselective regimes. We refer to this concept as ‘constant solvent transfer’. The constant solvent transfer concept will fail predictably at very high c,, typically in excess of 5 mol dm-3 (ref. 41) where the activity coefficient of water in the bathing solutions starts to decrease significantly. Salt contribution At sufficiently high electrolyte concentration, permselectivity fails: co-ion (accompanied by an equivalent amount of counter ion, and designated ‘salt’) partitions into the polymer film.We describe AMsalt, the extent of permselectivity failure, as a function of bathing solution and polymer compositions for the case where salt partitions only into one redox form (here, ox). The partition of anions and cations into a polymer contain- ing fixed (cationic) sites M+ (concentration c ~ + , ~ ) must satisfy two conditions. The first is electroneutrality cA-,p = cM+,p + cC++p (11) The second is the activity constraint,*g described by the salt partition coefficient (at a specified polymer oxidation state) Under these conditions the activity of water in the bathing Modulated by the thermodynamic function T [eqn. (15)]. solution hardly differs from unity. Combination of eqns. (11) and (12) yields a quadratic in CC+ +. The solution is expressed in terms of three dimensionless parameters R = CC+,dCM++p (13) s = cs/cM+,p (14) T = Ksalt4[Y k ,sly k ,pl (15) The extent of salt partitioning is described by R , which we term the ‘permselectivity index’.We have predicted41 that the departure of R from zero (when ST + 0) is initially quadratic (for ST <0.1), and eventually linear with electrolyte concen- tration (for ST >lo). Consequently AMsalt will vary in a quadratic mannerd as c, is increased from zero. Separation of contributions to AM The general expression for the mass change at any electrolyte concentration is given by eqn. (4). It is convenient to rewrite this as where AMperm describes the mass change at low electrolyte concentrations. For a single electrolyte, the ‘constant solvent transfer’ concept allows separation of experimental values of AM into three components, as follows.AM is measured over the widest range of c, available. At low c,, AM is independent of cs and becomes AMperm, defined by eqn. (16). We calculate AMcounter from eqn. (5). Subtrac- tion of AMcOunter from AMperm yields AMsolv. AMcounter and AMsolv are independent of c,. Consequently, their sum, the experimental quantity AMperm, can be subtracted from AM in the non-permselective regime [eqn. (4)] to yield AMsalt. This strategy is illustrated schematically in Fig. 1. Note that AMsalt and AMs0lV may each be of either sign. The sign of AMs01, is determined from the magnitude of AM in dilute solution (ST +- 0), and that of AMsalt from the behaviour of AM at high concentration (ST >0.1).When AMsalt and AMsolv are of opposite sign, interesting compensa- tory mass change effects can result (see below). Kinetics of Mobile Species Transfer We also discuss EQCM measurements on shorter timescales, following the application of a potential step. Starting with eqn. (2) the motion of charged and neutral species can be 8 6 w m C 5 4 rn s 2 0 -2.0 - 1 .o Fig. 1 Schematic diagram showing the separation of the total mass change, A M , into its components [see eqn. (4) . Solid line shows AM as a function of ST [see eqns. (14) and ( 1 5 i . Diagonally shaded, cross-hatched and open regions show (in cumulative manner) the single species contributions to A M . For illustrative purposes, contri- butions are normalized with the counter ion molar mass, and the (normalized) solvent contribution is taken to be 1.51254 ANALYST, AUGUST 1992, VOL.117 Table 1 Modified electrode deposition conditions (electrode area 0.23 cm-2) Deposition potential/ Concentration/ Species Electrode V versus SCE mmol dm-3 Electrolyte Solvent Thionine Au Poly(viny1ferrocene) Au 2,2‘-Bithiophene Pt 1.10 ca. 0.04 0.05 rnol dm-3 H20 (monomer) HC104 CH2C12 0.70 1 .o 0.1 mol dm-3 (pol ymer-reduced form) (C4H9)4 NC104 1.225 2.0 0.1 rnol dm-3 CH&N (monomer) (C2H.514 NBF4 described. Charged species are influenced by both activity and potential gradients while neutral species are influenced by activity gradients, but not potential gradients. This leads to the general expectation that ion transfers will ordinarily be more rapid than net neutral species transfers.Transient EQCM responses are therefore likely to be dominated by the transfer of charged species at short times; neutral species transfers will dominate the longer time responses. One manifestation of this field-induced transport rate enhancement is an apparently permselective response (no salt, or indeed any other neutral species, transfer) on short timescales. This will be true regardless of whether the thermodynamic behaviour of the film is permselective. In a preliminary paper,42 this effect was observed and termed ‘kinetic permselectivity’. Within the broad ‘ion’ and ‘net neutral’ classes of species, size effects and other more specific chemical interactions can lead to an observable dispersion of transfer rates. These observations suggest that individual transfers might be re- solved in the time domain.Now the analytical utility of this approach is discussed by considering the EQCM response of a polymer-modified electrode under transient conditions. A significant extension of the ‘constant solvent transfer’ concept is possible provided that salt transfer does not influence the transfer rates of other mobile species.e When this is the case we may differentiate eqn. (16) with respect to time to obtain dMldt = dM,,,,~,,/dt + dM,,,,/dt + dM,,ltldt = dMperm/dt + dM,,l,/dt (17) Rearranging eqn. (17) yields dM,,,t/dt = dM/dt - dM,,,,/dt (18) This result reveals that differences between kinetically con- trolled mass changes obtained in the non-permselective and permselective regions will provide quantitative information about instantaneous salt populations in the polymer.Experimental The instrumentation has been described previously.21~2+31 All films were deposited potentiostatically . Poly(thionine) (PTh) and poly(bithiophene) (PBT) films were deposited by electro- polymerization of the respective monomers. Electrodeposi- tion of poly(viny1ferrocene) (PVF) films exploited the lower solubility of the polymer upon oxidation.43 Details of solutions and deposition conditions are summarized in Table 1. For the experiments described here, all the modified electrodes were transferred to monomer/polymer-free solutions (see figure legends for composition). Polymer coverages were deter- mined by integration of slow scan rate (4 mV s-1) voltammetric current responses and are reported in terms of moles of electroactive sites, I‘Irnol cm-2.Coverages were in the ranges 3-10 nmol cm-2 (PTh), 5-12 nmol cm-2 (PVF), e This includes, but is not restricted to, the situation where all mobile species transfers are independent of each other. 20 - 10 im8 0 . . . . . . . . . . . 10 5 (v -- $ 1 8 m 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I .. ..... c 0 -” 0 2 4 PH 6 Fig. 2 Experimental mass change data for complete redox switching of poly(thionine) films immersed in aqueous acid solutions: U, HC104, pH varied via c,; and 0,O.l mol dm-3 total acetate, pH varied by addition of NaOH. Data taken from cyclic voltammetric scans at 5 mV s-1 and 16-38 nmol cm-2 (PBT). Comparison of gravimetric data for films with different coverages is made by dividing experimental mass changes, AMlng cm-2, by r (=Q/nF, where n is the number of electrons passed per redox site converted).Values of nAMFIQ correspond to the mass change per mole of redox sites converted. Results and Discussion Thermodynamics of Mobile Species Uptake Contributory phenomena Operationally, one seeks a single valued (ideally linear) relationship between the film mass change (AM) and the target species concentration in solution. In this section we illustrate some of the phenomena which contribute to the value of AM, emphasizing circumstances under which the desired behaviour will not prevail. Fig. 2 shows data for the redox switching of PTh films in aqueous solutions of a strong and a weak acid. The thionine redox couple in solution is a 2e--3H+ system: TH+ + 2e- + 3H+ e TH42+ where T represents the free base (see Scheme 1). Electroneut- rality requires that each redox site in a PTh film in the oxidized (reduced) state be associated with one (two) anions.The consequence of this is that 2AM,,,,t,,F/Q will be numerically equal to the anion mass. Including the protons, these electroneutrality considerations alone predict 2AMFIQ values of 102.5 and 62 g mol-1 for HC104 and CH3C02H electro- lytes, independent of solution composition (cs). These predic- tions fail, even at a qualitative level. The value of 2AMFIQ is dependent on solution composition, as required of a sensor, under some but not all circumstances (see acetic acid data at low pH). The variations of AM with electrolyte composition (19)ANALYST, AUGUST 1992, VOL.117 100 H - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lPerchlorateJ I I I 1255 Scheme 1 are qualitatively different for the two acids; even the sign of AM is dependent upon HC104 concentration. These results pertain to complete redox switching of the modified electrode, under slow-scan voltammetric conditions and, therefore, relate to changes in (pseudo-)equilibrium populations of mobile species. We cannot invoke kinetic effects to rationalize these observations. The effects are of thermodynamic origin, but are different in the two examples. In HC104 solutions, the activity constraint requires transfer of two net neutral species: water and hydronium perchlorate.Equilibrium hydronium perchlorate populations in the film, for each polymer redox state, are dependent upon the acid concentration in solution. The extent of transfer (the differ- ence between these two equilibrium populations) is conse- quently dependent upon solution concentration. It is not possible a priori to calculate the (variation of) activity coefficients in the film. We are, therefore, required to determine the relationship between film and solution compo- sitions empirically. As different partitioning species would be characterized by different activity coefficients, it would be necessary to produce a calibration plot (of the form of Fig. 2) for each species of interest. The CH3C02H data represent a special case, of particular relevance to sensors employing specific reagents.Neutral acetic acid interacts very strongly with the polymer film,30 to the extent that a coordination model applies. Provided a minimum concentration of CH3C02H is present in the solution, all redox sites in the film are saturated with CH3C02H. Upon redox switching, one of these coordinated acetic acid molecules provides the counter ion required by electroneutrality and one of the protons [see eqn. (19)]. The result is that the only ionic species required to move during switching are two protons. The over-all 2AMFlQ value of -16 g mol-1 indicates expulsion of a species of molar mass 18, namely water. On this basis, the suggested30 half-reaction in acetic acid solutions of low pH is [(ThH+A-).(H20)-X.(HA)1, + 2e- + 2H3O+, = where X is permanently present in the film and may be either acetic acid or water.Attention is drawn to three points. Firstly, the counter ion-conjugate acid (acetate-acetic acid) is always resident in the film, i.e., does not transfer. Secondly, different species dominate the charge and mass responses, proton and solvent, respectively. Thirdly, the invariance of EQCM response at low pH (high CH3C02H concentration) is a consequence of the strong interaction between the acetic acid and the polymer. It is common for a sensor to exploit a strong, specific interaction between the target species and the polymer. If this interaction is too strong, saturation of all the available sites may result, even at fairly moderate target species concentra- tions. In the PTh example here, a composition-dependent response is seen only for pH > pK, when the solution concentration of neutral acetic acid is low and the polymer sites are no longer saturated with acetic acid.{ [ThH42+(A-)2]X}, + 3H20, (20) Separation of single species contributions We now apply the ‘constant solvent transfer’ concept to separate over-all (thermodynamic) mass change data into 500 I m rn Fig. 3 AMFIQ data for PVF oxidation as a function of NaC104 concentration. Symbols represent experimental data. Counter ion and solvent contributions (determined as in text) are shown cumulatively their components. We consider PVF films exposed to aqueous NaC104 solutions, for which the electroneutrality and activity constraints are sufficient to describe the mass changes. Fig.3 shows AMFIQ data for PVF oxidation as a function of NaC104 concentration, c,. At low concentration, AMFIQ is independent of c,, and given by AMperm [see eqn. (16)J. Eqn. (5) shows that AMcounterF/Q corresponds to +99.5 g mol-1. By difference, AMSoIvF/Q is 110 g mol-1. The magnitudes of the counter ion and solvent contributions to AM (experimen- tal points) are represented additively by the two marked regions in Fig. 3. In the light of the constant solvent transfer concept we are now able to extend our deconvolution of AM to the non-permselective region, here c, >1 mol dm-3. The low c, values of AM,,u,,erF/Q and AM,,IvF/Q in Fig. 3 have been extrapolated to high c,, using the strategy of Fig. 1. By difference [see eqn. (16)], we obtain AMSaltF/Q. Our theory predicts that the salt contribution should rise quadratically with NaC104 activity in solution.Figs. 1 and 3 are not combined, as the experimental data have a ‘concentration’ abscissa, whereas the theoretical prediction is in terms of ‘ST. Future work will involve reconciliation of theory and experi- ment: according to eqn. (15) this will require fitting activity coefficients and a partition coefficient. The central point is that we now have the theoretical machinery to effect complete separation of each single species contribution to the total mass change. Compensatory mass changes Previous PBT studies31 illustrate a more subtle problem. As for PVF, counter ion, salt and solvent are transferred upon PBT redox switching. The difference is that, for PBT, the transfers of the two net neutral species (salt and solvent) are in opposite directions. In terms of eqn.(3), 6 = -0.36 and 0.11 < E < 0.32 for a range of CH3CN-tetraalkylammonium salt solutions. Reduction of the PBT+ is accompanied by solvent exit and salt entry, in addition to counter-ion exit. The values of molar masses, 6 and E, are such that AMsolv = -AM,,It. This has the totally fortuitous result that AM = AMcounter. We term this set of circumstances ‘apparent permselectivity’. Detailed experiments31 for different electrolytes over a range1256 ANALYST, AUGUST 1992, VOL. 117 Y a1 2000 1 I x "3 .. I 1600 - N 1200 - m C 5 800 - 400 - 0 400 800 #pC cm-2 1200 Fig. 4 AM versus Q plots for the oxidation and subsequent reduction of a PVF film in 0.1 mol dm-3 (0) and 3 mol dm-3 (x) NaC104 solutions.The potential was stepped from 0 to +0.7 V, held at +0.7 V for 15 s, then stepped back to 0 V. The origin corresponds to reduced PVF at 0 V. The small intercept on the abscissa is attributed to double layer charging. The lines correspond to the predicted AM/Q relation- ship for anion and anion + solvent transfer, as marked of concentrations revealed deviations from AM = AMcounter, due to the concentration dependence of AMsalt (see above). We suspect that compensatory mass changes are fairly common. Analytical measurements made without due atten- tion to this process could lead to serious errors. Kinetic Aspects of the Exchange Process Kinetic permselectivity : the strategy In the previous section, we considered the over-all mass changes associated with complete oxidation-reduction of a film, which was then allowed to come to equilibrium with the solution phase.In this section we consider the time course of the redox transformation, i.e., the transfer of species during the oxidation-reduction process. When permselectivity does not exist in a thermodynamic sense (for example at high electrolyte concentration), we propose that it may be possible to achieve it transiently. We aim to exploit the differing rates of mobile ion and neutral species transfers. In a transient experiment the response on a short timescale will be dominated by the fastest moving species. The converse will be true on a longer timescale. We suggest that this approach might be exploited at two levels. Firstly, field assistance of ion transfer (migration) will lead to their being more rapid than neutral species transfers. Secondly, size effects (for ions or neutral species) will lead to a diversity of transport rates.The latter effect is likely to be more pronounced in the confined geometry of polymer films than for the same species in solution. Also, the extent to which transfer of a given species dominates the net transfer process (on a given timescale) will depend on its availability, i.e., the difference between its film and solution concentrations. Kinetic permselectivity: an example We first observed this effect during rapid scan voltammetry of PVF films in aqueous NaC104 solutions,42 for which the over-all mass changes accompanying redox switching have been described above (see Fig.3). The EQCM data acquired during chronoamperometric experiments on a PVF film exposed to 0.1 and 3 mol dm-3 NaC104 are shown in Fig. 4. Two straight lines are drawn corresponding to the mass- charge relationships that would result if (i) only counter ions and (ii) counter ion and solvent (but not salt) were trans- ferred. These correspond to AMcounter and AMperm, respec- tively. 2500 2000 ," 1500 c I 1000 m . 500 P m . -1000 I I I I I I -0.2 0 0.2 0.4 0.6 0.8 1.0 tls Fig. 5 Salt mass flux (dM,,,,ldt) as a function of time for the oxidation stage of the data in Fig. 4. Points were generated by taking the time differential of the difference in the two sets of data in Fig. 4 [see eqn. C18)l During the initial stage of the oxidation process, the experimental data in both bathing solutions superimpose on the line for AMcounter.This result implies that only C104- transfers on that timescale. Neutral species transfers are rather slower and progressively become more predominant at longer timescales. Fig. 5 gives the result of applying eqn. (18) to the first (oxidative) stage of the two data sets of Fig. 4. It shows the variation of AMsalt with time in 3 mol dm-3 sodium perchlor- ate (non-permselective conditions). As generally expected, salt (a net neutral species) transfer lags counter-ion transfer. For the particular case of our PVF films in this medium, salt transfer is the faster of the two neutral species transfers. Conclusions We have considered the transfers of mobile species accom- panying redox switching of a redox polymer film immersed in a solution of a single electrolyte.The problem has been cast in terms of the associated mass changes, which can be deter- mined by using the EQCM. It is possible to separate the total mass change observed during film redox switching into the components associated with counter ion, solvent and salt transfer under both equilibrium and transient conditions. This separation exploits the concept of constant solvent transfer, which asserts that the extent of solvent transfer is independent of electrolyte concentration in the bathing solution. Transient measurements allow separation of charged and the various net neutral species transfers in time. We suggest that this will have analytical utility when determining mixtures of mobile species via their uptake by a polymer film.The SERC (GFUE132946 and GR/E/78104), NATO (86/0830) and the National Science Foundation (CHE-9115462) are thanked for financial support. M. J. S. thanks the SERC for a research studen tship. References 1 Deakin, M. R., and Buttry, D. A . , Anal. Chem., 1989, 61, 1147A. 2 Schumacher, R., Angew. Chem., Znt. Ed. Engl., 1990,29,329. 3 Ward, M. D . , and Buttry, D. A . , Science, 1990,249, 1OOO. 4 Buttry, D. A., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1991, vol. 17, p. 1. 5 Murray, R. W., in Electroanalytical Chemistry, ed. Bard, A. J., Marcel Dekker, New York, 1984, vol. 13, p. 192. 6 Hillman, A. R., in Electrochemical Technology of Polymers, ed. Linford, R . , Elsevier, London, 1987, p.241. 7 Heineman, W. R., Wieck, H. J., and Yacynych, A . M., Anal. Chem., 1980, 52,345.ANALYST, AUGUST 1992, VOL. 117 1257 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Lawton, R. S., and Yacynych, A. M., Anal. Chim. Acta, 1984, 160, 149. Cox, J. A., and Kulesza, P. J., J. Electroanal. Chem., 1984,175, 105. Facci, J. S., and Murray, R. W., Anal. Chem., 1982,54, 772. Szentirmay, M. R., and Martin, C. R., Anal. Chem., 1984, 56, 1898. Guadalupe, A. R., and Abruna, H. D., Anal. Chem., 1985,57, 142. Cox, J. A., and Kulesza, P. J., Anal. Chim. Acta, 1983,154,71. Nagy, G., Gerhardt, G. A., Oke, A. F., Rice, M. E., Adams, R. N., Moore, R. B., Szentirmay, M. N., and Martin, C. R., J. Electroanal. Chem., 1985, 188, 85. Denisevich, P., Abruna, H.D., Leidner, C. R., Meyer, T. J., and Murray, R. W., Inorg. Chem., 1982,21,2153. Wier, L. M., Guadalupe, A. R., and Abruna, H. D., Anal. Chem., 1985,57,2009. Itaya, K., Akahoshi, H., and Toshima, S., J. Electrochem. Soc., 1982, 129,762. Ebersole, R. C., and Ward, M. D., J. Am. Chem. Soc., 1988, 110, 8623. Bruckenstein, S., and Hillman, A. R., J. Phys. Chem., 1988,92, 4837. Sauerbrey, G. Z., 2. Phys., 1959,155,206. Bruckenstein, S., and Shay, M., Electrochim. Acta, 1985, 30, 1295. Kanazawa, K. K., and Gordon, J. G., Anal. Chim. Acta, 1985, 175, 99. Orata, D., and Buttry, D. A., J. Am. Chem. SOC., 1987, 109, 3574. Baker, C. K., and Reynolds, J. R., J. Electroanal. Chem., 1988, 251, 307. Reynolds, J. R., Sundaresan, N. S., Pomerantz, M., Basak, S., and Baker, C. K., J. Electroanal. Chem., 1988, 250, 355. Baker, C. K., and Reynolds, J. R., Synrh. Met., 1989, 28, C21. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., J. Electroanal. Chem., 1989, 274, 157. 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Hillman, A. R., Loveday, D. C., Swann, M. J., Eales, R. M., Hamnett, A., Higgins, S. J., Bruckenstein, S., and Wilde, C. P., Faraday Dkc. Chem. SOC., 1989, 88, 151. Bruckenstein, S., Wilde, C. P., Shay, M., and Hillman, A. R., J. Phys. Chem., 1990,94, 787. Bruckenstein, S., Wilde, C. P., and Hillman, A. R., J. Phys. Chem., 1990,94,6458. Hillman, A. R., Swann, M. J., and Bruckenstein, S., J. Electroanal. Chem., 1990, 291, 147. Kelly, A. J., Ohsaka, T., Oyama, N., Forster, R. J., and Vos, J. G., J. Electroanal. Chem., 1990,287, 185. Inzelt, G., J. Electroanal. Chem., 1990, 287, 171. Naoi, K., Lien, M., and Smyrl, W. H., J. Electrochem. SOC., 1991, 138,440. Borjas, R., and Buttry, D. A., Chem. Muter., 1991,3,872. Hillman, A. R., Swann, M. J., and Bruckenstein, S., J. Phys. Chem., 1991,95,3271. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., J. Electroanal. Chem., 1991,300, 67. Bruce, J. A., and Wrighton, M. S., J. Am. Chem. SOC., 1982, 104,74. Borjas, R., and Buttry, D. A., J. Electroanal. Chem., 1990,280, 73. Lasky, S. J., and Buttry, D. A., ACSSymp. Ser., 1989,403,237. Bruckenstein, S., and Hillman, A. R., J. Phys. Chem., 1991,95, 10748. Hillman, A. R., Loveday, D. C., Bruckenstein, S., and Wilde, C. P., J. Chem. Soc., Faraday Trans., 1990,86, 437. Hillman, A. R., Loveday, D. C., and Bruckenstein, S., Langmuir, 1991, 7, 191. Paper 2102236 B Received April 30, 1992 Accepted May 14, 1992