|Home | About | Journals | Submit | Contact Us | Français|
We have developed a theoretical description of the amperometric response of ultramicroelectrode (UME) biosensors formed via enzyme entrapment. Our model allows for multiple enzymes and co-substrates, and results in a closed-form analytical expression for the steady-state current response of the disk ultramicroelectrode. It captures the effects of enzyme-entrapment domain size, species transport properties (which can be different in the polymer and surrounding electrolyte), enzyme kinetics, and axisymmetric diffusion. Assumptions inherent in the derivation are carefully explained, as are the resulting limits on the applicability of the results. The ability to theoretically predict the response of enzyme entrapped UMEs should enable improved design, operation, and data interpretation for this important class of biosensors.
Ultramicroelectrodes have revolutionized the field of analytical electrochemistry, leading to novel methodologies and making new time and space scales accessible to experimentation [1, 2]. For instance, the goal of (electro)chemical imaging with submicron spatial resolution and sub-millisecond temporal resolution has resulted in considerable interest in the use of scanning electrochemical microscopy (SECM), in which UMEs are employed to measure local fluxes of molecules and ions [3–7]. In amperometric operation, the UME, a small electrode (diameter 50 nm–25 μm), induces an electrochemical reaction, and the resulting Faradaic current is proportional to the net flux of a redox species to the electrode surface [8, 9]. An oft-stated advantage of this technology is the ability to predict or interpret the measured electrode current based on a first principles theoretical description of the relevant phenomena [1, 9].
For biological applications the versatility of the UME as a probe is dramatically improved when combined with biosensor technology . A biosensor is characterized by a biological recognition element in contact with the transducer surface. Enzymes are frequently chosen as the recognition element due to their remarkable specificity and inherent biocatalytic signal amplification [10–12]. Some biologically relevant analytes require a complex biosensing interface, with co-immobilization of multiple enzymes [5, 13, 14]. Current state-of-the-art immobilization strategies include (i) crosslinking (e.g., attachment of the biological recognition element based on biotin/avidin chemistry  or antigen-antibody binding ), (ii) covalent attachment to self assembled monolayers [17, 18], and (iii) polymer entrapment. Enzyme entrapment in polymer matrices has been demonstrated in conducting polymers [19–21], sol gels , and hydrogels , including electrodeposited electrophoretic paints (EDPs) , which are polymer suspensions deposited via a pH shift induced desolubilization. Electrophoretic paint entrapment exhibits favorable characteristics, including ease of implementation, retention of enzyme activity, and extended sensor stability [13, 25]. The model presented in this work was developed to describe the response assuming a probable geometry and likely dominant physics for micro-disk electrodes and polymeric matrices formed via pH shift induced polymer deposition (see section 2.1) [13, 24, 25].
Theoretical prediction of the amperometric response for miniaturized biosensors utilizing polymer entrapment requires knowledge of the permeable polymer domain shape and transport properties, reaction pathway(s), and enzyme kinetics. Even with known properties and reaction parameters, there is currently no comprehensive theoretical description for predicting the current signal at the microelectrode for a given target molecule concentration. This is because most available models such as those of Savéant and co-authors [14, 26–29] and others [30–37] are based on 1-D planar diffusion, usually with the enzyme reaction confined to a monolayer. But 3-D steady-state diffusion dominates in amperometric UME operation , and to date analytical models to describe these conditions are limited to extremely large  or infinite  entrapment domains, which electrodeposition is intended to avoid. Hence, a first principles theoretical model describing the electrodeposited polymer-entrapped enzyme microelectrode experiment is missing, and both interpretation of results and design of the microprobes are at present entirely empirical.
In this contribution an analytical description of entrapped enzyme ultramicroelectrode (UME) steady state response is presented. The formulation is particularly suitable for micro-disk electrodes and polymeric matrices formed via pH shift induced polymer deposition ; however, the model may be extended to biosensors based on different enzyme entrapment methods. Experimental validation of the theory and demonstration of its application are provided in a companion paper .
The amperometric response of the electrode is related to the net rate of diffusional mass transport of redox species R to the electrode, JE,R, by
where the determination of the algebraic sign of the current signal is based on the direction of electron transfer, n is the number of moles of electrons transferred per mole of electroactive species oxidized or reduced, and F is the Faraday constant. For unmodified disk UMEs with sufficiently thick insulating sheaths JE,R depends upon the electrode size, (radius a), the diffusivity of the redox species in the electrolyte, DR, and its bulk concentration, C∞,R. At steady-state, assuming a diffusion-limited electrochemical reaction, the resulting current is theoretically predicted to be I=±4nFaDRC∞,R .
Amperometric microbiosensor response can also be described by Eq. (1), but determination of the net rate of mass transfer of the electroactive species R to the electrode surface, JE,R, is more complicated (Fig. 1). For a polymer entrapped enzyme UME, the production of R occurs via enzyme-catalyzed reactions confined to a polymer domain adjacent to the electrode surface. The rate of species R production depends upon the bulk concentration of the target analyte, C∞,A, on enzyme reaction kinetics, and on mass transport in two linked domains (polymer and electrolyte solution). Example enzyme catalyzed strategies linking redox species production to analyte species concentrations are depicted in Table I. For three of the example strategies (Uni Uni Reaction, Bi Bi Reaction, and Consecutive Bi Bi Reactions) an increase in C∞,A results in an increased rate of generation of electroactive species R and consequently an increase in magnitude of current I. Competitive reaction electrodes, such as the last example in Table I, utilize a strategy whereby the electrode current has a nonzero baseline value in the absence of the analyte because the electroactive species, R, is produced by an enzymatic reaction that does not involve the analyte; therefore, the effect of the analyte, A, is to reduce the current, due to participation in a competitive enzyme catalyzed reaction. An advantage of our theoretical description of polymer entrapped enzyme UMEs is its ability to accurately predict the current for electrodes based on many strategies, including all of those listed in Table I.
Our formulation is based on a number of standard assumptions frequently used in modeling of electrochemical processes. These assumptions include Fickian diffusion with uniform transport properties (within a phase) and negligible double layer and migration effects . The electroactive species R has a specified uniform bulk concentration, C∞,R (which may be zero). Sufficient overpotential is applied on the electrode so that the concentration of the electroactive species, CR, is brought to zero on that surface. Local equilibrium of species’ concentration at the polymer/electrolyte interface is assumed, so that at the polymer/electrolyte interface, the concentration inside the polymer is related to the concentration in the solution by the partitioning coefficient θR. For the electroactive species, with diffusivity in the electrolyte DR, the diffusivity in the polymer is described using a proportionality constant αR, which is the ratio of the diffusivity in the polymer compared to its diffusivity in the electrolyte. Inside the polymer, volumetric generation of the electroactive species due to the enzyme reaction occurs at a rate given by SR, which is generally a function of position due to spatial variation in concentrations of the species which participate in the reaction. At steady state the concentration of the electroactive species is therefore described by the coupled reaction-diffusion equations:
The boundary conditions prescribe the bulk concentration
the concentration on the electrode
and no flux on the insulating surfaces
where denotes a derivative in the direction normal to the surface. Boundary conditions also link the two domains. They are the continuity of flux condition
and a statement of local thermodynamic equilibrium
The net flux of the electroactive species to the electrode, JE,R, is given by the area integral
Equations (1)–(9) comprise a general model for predicting the current, I, measured at an electrode due to a redox reaction with generation of reactant R limited to a finite domain adjacent to the electrode surface. The current depends on the bulk concentration of redox species R, its transport properties in the polymer and electrolyte, the geometry of the system, and the source term for species R in the polymer, SR. It is through the source term, SR, that the analyte species A affects the electrode response. Hence, the prediction of the electrode current, I, as a function of bulk analyte concentration, C∞,A, requires at minimum knowledge of the parameters listed in Table II. Some of the parameters, if not known initially, can be determined via a calibration procedure presented in a companion article . The electrode current prediction also requires evaluation of the source term, SR, which will depend upon the concentrations of other species. In general, the steady-state concentrations of these species will be governed by diffusion-reaction equations analogous to those for species R. Thus, with the prerequisite parameters in hand, the steady-state current can be obtained by solution of a set of reaction-diffusion equations (inhomogeneous partial differential equations and associated boundary conditions) in linked domains (see Supporting Information).
In general, the solution of this problem can only be accomplished numerically, using, for instance, finite volume, finite element, or finite difference techniques, implemented through either commercially available or user-written computer codes (e.g., see Supporting Information). When using numerical simulation, for any given combination of input parameters and species concentrations, a single electrode current prediction is obtained. The resulting parametric space, even after nondimensionalization, is large, filling it is time consuming and tedious, and extracting useful insight from the results is difficult. To be useful for the broad community of experimentalists, a theoretical tool must be accurate yet simple enough not to require special skills for implementation, and expressed in terms of meaningful and accessible physical parameters. Analytical solutions are therefore profoundly more useful than numerical simulations, and it is often worth introducing carefully validated simplifications to obtain them.
To this end, our main original result is a closed-form analytical description of the relationship between the steady-state Faradaic current, produced by the oxidation or reduction of the electroactive species on the UME surface, and the bulk concentration of the analyte species, C∞,A. The functional relationship is expressed in terms of an equivalent redox species bulk concentration, K:
The equivalent redox species bulk concentration, K, which captures effects of geometry and enzyme kinetics, depends upon the bulk concentrations of the analyte species and all co-substrates, and the parameters listed in Table II. The derivation of Eq. (10) is provided as Supporting Information. Much of the present work is devoted to explaining the algorithm(s) by which K can be determined. First we explain the assumptions made in the derivation and the conditions under which they should be reasonable.
A subset of conditions exists for which the general model can be solved exactly, and for which numerical simulation is not necessary to predict the electrode current. These conditions are satisfied when the assumptions listed in Table III are appropriate. Some of these assumptions are standard and their limits are well understood. Others require explanation. The first such condition is that the polymer domain shape is accurately approximated as an oblate hemispheroid. The assumption of this oblate hemispheroid geometry is partially justified by consideration of the fact that in electrophoretic paint entrapment the polymer is deposited due to diffusion of protons from the electrode into the solution and isoconcentration surfaces in steady state diffusion from a disk into a semi-infinite medium are oblate hemispheroids . In our analysis, after nondimensionalization of all lengths with the disk electrode radius a, the assumed oblate hemispheroid shape and size are specified by a single dimensionless parameter, Z (see Supporting Information). It is worth noting that with a small Z the entrapment domain is a thin film just covering the electrode.
Another necessary condition for the presented analyses is the requirement of disparate length scales. Specifically, for all species other than the electroactive species R, the ratios of the length scales for changes in their concentration due to enzymatic action in the polymer to the length scale for molecular transport in the polymer must be either much greater than or much less than unity. When the ratio is small, the limiting reactant for the relevant reaction will be entirely consumed close to the polymer/electrolyte interface and an approximation of total consumption of the limiting reactant on the polymer surface is applied. We call this approximation the heterogeneous reaction approximation. In contrast, if the length scale ratio is large there will be little variation in concentrations of the non-electroactive species, and therefore little variation in the magnitude of source terms within the polymer. Thus the source terms can be approximated as uniform throughout the polymer domain. We call this approximation the uniform homogeneous reaction approximation. An illustration of the effect of these approximations on predicted concentrations is provided in Fig. 2.
Figure 3 provides a guide for application of the polymer entrapped enzyme microbiosensor model. In addition to obtaining all parameters listed in Table II and checking the assumptions of Table III, one of the steps is the determination of the relevant species for the analysis. This is accomplished by examination of the enzyme reaction mechanism(s), which should reveal those species that affect the reaction rate: they are at least the substrate and co-substrates, and include products if the reaction is reversible. Next, equations for source terms are obtained for steady-state conditions. These source terms can be derived by writing species conservation equations for differential control volumes, or, equivalently, by applying the schematic method of King and Altman . Source terms for some irreversible reactions are included in Table I. Examples with enzyme reaction mechanisms that yield more complicated source terms are included in the companion paper . Once these initial steps are completed, Fig. 3 with the following explanation provides sufficient guidance to predict electrode current. Considerable detail on the derivation of the methodology is available in the Supporting Information.
The source term expressions mentioned above are needed to determine whether the conditions that allow for analytical solution are met. This determination requires calculation of the nondimensional parameters,
which are the ratios of the length scale for variation of concentration due to diffusive transport of species i in the polymer, to the length scale due to the variation of concentration of species i as a result of its production or consumption in reaction j, described by the source term Sj. In Eq. (11) Z is a dimensionless parameter specifying the size of the polymer domain, a is the radius of the disk electrode, Di is the diffusivity of species i in the electrolyte, αi is the polymer to electrolyte diffusivity ratio for species i, θi is the coefficient for equilibrium partitioning of species i at the polymer/electrolyte interface, and C∞,i is the bulk concentration of species i. Each source term scale, Sj,s, is a characteristic volumetric rate of production of species by enzyme-catalyzed reaction j, and can be obtained by replacing the concentrations in the source terms using Ci ~ θiC∞,i.
A single species may be produced or consumed by more than one reaction, and each reaction will involve more than one species. Our analytical method requires using Eq. (11) to determine for each reaction j which, if either, of two approximations can be applied. For a given reaction j described by source term Sj, the maximum and minimum Λji are found by considering all species i with nonzero bulk concentrations that participate in the reaction. The uniform homogeneous reaction approximation should be used if
and the heterogeneous reaction approximation is valid in the case of
If the source term describing the production of the redox species, i.e. SR, satisfies Eq. (12) then the uniform homogeneous reaction approximation is applied to the production of species R, and
with the evaluation of SR described in iii (below). If instead SR satisfies Eq. (13) then the heterogeneous reaction approximation is applied, and
where ωAR is a stoichiometric coefficient specifying the number of moles of species R produced per mole of the analyte A consumed in the enzyme catalyzed reaction. In Eqs. (14) and (15), two parameters that capture the polymer shape and transport property effects are used:
In the case that the heterogeneous reaction approximation applies to the redox species producing reaction, then Eqs. (15) and (10) are sufficient to predict the electrode current. If, however, the homogeneous reaction approximation is used to describe the generation of the redox species, then Eqs. (14) and (10) cannot be used without first evaluating the uniform source term SR.
The source term used in Eq. (14) requires an assumption of uniform concentration in the polymer for one or more species. The approximation of the uniform concentrations is obtained by considering a non-uniform concentration in the solution so as to satisfy global mass conservation for species i. Global mass conservation is satisfied by balancing the net rate of transport of species i from the polymer with the net rate of generation of species i in the polymer. The result is a system of algebraic equations for approximate uniform concentrations in the polymer, Cp,i, that must be solved simultaneously. To obtain this system of equations, first, for all relevant species:
Once steps (a)–(c) have been performed for all relevant species, there will be a system of equations of the type given by Eq. (18), written for all species that participate in reactions treated under the uniform homogeneous reaction approximation. The system is generally non-linear, and must be solved numerically, for example, using the method of successive approximations with Cp,i= θiC∞,i as initial guesses.
With Eq. (10), the steady-state electrode current is calculated using the equivalent bulk concentration (K) of redox active species R, obtained from either Eq. (14) (uniform homogeneous reaction approximation for SR) or from Eq. (15) (heterogeneous reaction approximation for SR).
We demonstrate aspects of our method using two of the example strategies depicted in Table I: a Uni Uni reaction (one reactant, one product) with A as reactant and R as product; and two Bi Bi reactions competing for a common (shared) co-substrate, with A as one of the reactants for one reaction and R as one of the products of the other reaction. In both cases, we assume that the all necessary parameters are known (Table II) and that the conditions listed in Table III are met. We limit consideration to zero bulk concentration of the redox active species, C∞,R = 0, and irreversible enzyme catalyzed reactions. We do not consider the complication of participation in the electrode reaction by species that also participate in the enzyme catalyzed reaction (other than species R). Note that the methodology can be applied without the latter restrictions as demonstrated in the companion paper .
First an irreversible Uni Uni enzyme reaction is considered: . In accordance with Fig. 3, we begin by identifying the relevant species (A and R) and writing expressions for the source terms. The steady state source terms are given in terms of the kinetic constants Vmax, which is the maximum reaction velocity, and KA, which is the Michaelis constant:
With Eq. (21) one can determine whether the uniform homogeneous approximation applies (ΛA ≤ 0.6) or the heterogeneous reaction approximation applies (ΛA ≥ 3).
The uniform source terms are obtained by replacing CA in Eq. (20) with Cp,A. The stoichiometric coefficients νAR and ωA are zero, so that Eq. (18), written for species A, is Cp,A = θA(C∞,A + Ga2SA/(3DA)). Combined with the expression for the source term, i.e., Eq. (20) with CA = Cp,A, this allows one to solved for the uniform concentration of analyte in the polymer:
The results can also be described in terms of a measure of enzyme electrode effectiveness, K′ |I|/(4nFDAaC∞,A):
In the opposite limit of consumption of the analyte by the redox-species-producing reaction very close to the polymer surface, Eq. (15) is used. For a Uni Uni reaction the stoichiometric coefficient is unity, i.e. ωAR = 1, so that from Eq. (15) one obtains K = TCDAC∞,A/DR, which, with Eq. (10), gives I=±4nFaTCDAC∞,A. Rewriting this expression in terms of the previously defined measure of enzyme electrode effectiveness, K′, yields
The motivation behind consideration of the uniform homogeneous reaction approximation was the recognition that as entrapped enzyme electrodes become smaller, the ratio of the length scale for concentration variation as a result of the enzyme reaction to the length scale for molecular transport in the polymer becomes larger. Hence, there is less variation of concentrations within the polymer domain, and therefore less variation of the rates of homogeneous reactions. Conversely, if reaction rates are very high then concentration variation due to reaction can occur over very short length scales, and appreciable reaction rates become confined to a region close to the polymer/electrolyte interface. The consequences of these conclusions are illustrated in Fig. 4. In Fig. 4 predictions of enzyme electrode effectiveness are plotted against polymer domain size, given by the dimensionless parameter Z, for constant αRθR, ΨA, and ΠA, using the two approximations given by Eqs. (24) and (25). Also plotted are results of numerical simulations generated using a commercial finite volume method solver  (see Supporting Information). The electrode effectiveness predicted from numerical simulation shows a smooth transition with increasing Z from values correctly predicted using the uniform homogeneous reaction approximation, Eq. (24), to values correctly predicted using the heterogeneous reaction approximation, Eq. (25). In Fig. 4, plots of analyte centerline concentration for large and small polymer deposits are presented and clearly illustrate the qualitative difference in concentration variation in the polymer for electrodes well approximated by the different formulations.
The plots in Fig. 4 also confirm experimental observations that an optimal size of the entrapment domain exists [24, 45, 46]. The existence of a Z corresponding to a maximum electrode effectiveness K′ is evident in Fig. 4, and the occurrence of the optimum can be explained in terms of the two approximate models. The uniform homogeneous reaction approximation model exhibits a monotonic increase with Z of predicted electrode effectiveness, while the heterogeneous reaction approximation model displays the opposite trend. The maximum K′ occurs at a polymer domain size, with corresponding Z, within the transition regime between the two limiting behaviors. Note that other factors could be more important than maximum effectiveness for actual sensor design. For instance, a smaller Z would provide better temporal resolution. Finally, the plots in Fig. 4 illustrate another application of the results of our analysis to experimental design. For a given sensor, the heterogeneous reaction approximation predicts the maximum possible electrode effectiveness. Thus, Fig. 4 provides an indication of the improvement in effectiveness that is possible by increasing, for instance, enzyme loading while decreasing deposit size. By quantifying the potential for improvement, the results allow an experimentalist to make an informed decision when contemplating attempts to increase sensitivity.
We next model sensors with two enzymes co-immobilized to catalyze the following reactions:
The analyte A does not participate in the redox species (R) producing reaction, Reaction 1. Instead, A reduces the production of the electroactive species R due to involvement in a competitive reaction, Reaction 2, which also consumes shared co-substrate, B. Therefore, application of the heterogeneous reaction approximation to the redox species producing reaction, i.e., use of Eq. (15), is not appropriate for competitive reaction electrodes.
Because in this example the reactions are irreversible and none of the products, with the exception of R, participate in the electrode reaction, the relevant species, in addition to R, are the analyte A, F, and the shared reactant, B. Equations for the source terms for irreversible Bi Bi reactions are derived assuming steady-state conditions. In general, for irreversible Bi Bi reactions, three mechanisms are possible: random; compulsory ordered; and double displacement (ping pong) [47, 48]. Regardless of the actual mechanism, the source terms due to Reaction 1 are
where the kinetic constants V1, KiF, KF, and KB1 depend on the rate coefficients and the concentration of total active enzyme sites (for a ping-pong mechanism KiF = 0). The source terms due to Reaction 2 are
The source term for the shared reactant, B, has components from both reactions:
The values of four length scale ratios must be calculated to determine which, if any, approximation applies to each reaction: Λ1B; Λ1F; Λ2B; and Λ2A. These parameters are calculated by replacing the corresponding concentrations in the source term, i.e., Eq. (28) or (29), with θiC∞,i, leading to Eq. (11), e.g.,
If the maximum length scale ratio for the reaction that produces the redox active species, Reaction 1, satisfies the criteria for application of the uniform homogeneous reaction approximation, i.e., max(Λ1B, Λ1F) ≤ 0.6, then the presented methodology can be applied. If this criterion is not met, however, including, as noted previously, the possibility that the heterogeneous reaction approximation may apply, then the analytical solution results cannot be used.
Provided the uniform homogeneous reaction approximation applies to Reaction 1, then we first note that νiR = 0:i=B, F, A. If max(Λ2B, Λ2A) ≤ 0.6 then the uniform homogeneous reaction approximation is applied to both reactions. The system of equations for the uniform concentrations in the polymer is obtained by writing Eq. (18) for A, F, and B:
where we have used SB =SA + SF. In the expressions for the source terms, the local concentrations are all replaced with uniform concentrations, e.g., CB = Cp,B. The system of equations, Eq. (31)–(33) is solved for Cp,F, Cp,A, and Cp,B, which then allows the uniform source term for the redox species, SR, to be evaluated using Eq. (28). Finally, SR is substituted into Eq. (14) to obtain the equivalent bulk redox species concentration, K, which is used to obtain the electrode current prediction from Eq. (10).
Alternatively, if min (Λ2B, Λ2A) ≥ 3 then the heterogeneous reaction approximation is applied to Reaction 2. First we must verify that the analyte, A, is the limiting reactant for Reaction 2, a requirement that is met if DAC∞,A<DBC∞,B The stoichiometric coefficients ω2B = 1, and the mass conservation equation, Eq. (18), is then written for species F and B,
The resulting system of equations is solved, yielding the uniform concentrations of species B and F in the polymer, which allows evaluation of SR and therefore, evaluation of the electrode current.
In the companion paper  we present application of our theory to a real competitive reaction electrode, similar in principle to that explained above, though including some additional complications.
We have developed a theoretical model of the steady state amperometric response of polymer entrapped enzyme ultramicroelectrodes. The model is analytical, although not explicit except in the simplest of cases. For realistic situations, solution of a small system of nonlinear algebraic equations is required. Two limiting cases are considered, based on the assumption that (i) homogenous reaction rates can be treated as spatially uniform, or (ii) reactions are confined to a region very close to the entrapment domain boundary
A notable advantage of the presented theory is that UME calibration becomes a measurement of transport properties and enzyme kinetics parameters. Because the results of calibration are quantities with specific meanings in a physics-based electrode model, the resulting description of electrode response can be used to predict the changes in electrode behavior as experimental conditions or electrode parameters are altered, leading to targeted optimization efforts and improved data interpretation.
This research was supported by the NIH grant RO1 EB000508-01A1.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.