You are on page 1of 10

Anal. Chem.

1994,66, 1520-1529

Prediction of Pocket-Portable and Implantable Glucose


Enzyme Electrode Performance from Combined Species
Permeability and Digital Simulation Analysis
Rathbun K. Rhodes,'ptp$ Mark C. Shults,tp* and Stuart J. Updiket
Department of Medicine, University of Wisconsin Medical Center, 600 Hbhland Avenue,
Madison, Wisconsin 53792, and Markwell Medical Institute, 1202 Ann Street, Madison, Wisconsin 537 13

Development of enzyme electrode sensors can be facilitated by measurement linearity through membrane diffusion control,'
the use of computer simulation modeling techniques. In this the importance of the relative permeabilities of substrate and
work, the model accounts for contacting solution effects, cofactor in defining the usable sensor measurement range,"
changes in species diffusion and partition coefficients across the idea of simulating speciesflux through multiple membrane
multiple membrane layer interfaces,cofactor limitationeffects, layer^,^ transient matching of experiment to theory when
enzyme loading,enzymekineticsand decay, and other variatiolrs enzyme activity is varied and simultaneous step changes occur
in electrode layer geometry. Experimentally determined single for substrate and c0factor,6.~and decay of enzyme concen-
species membrane permeabilities are included in the digital tration over time as a function of contacting solution
simulations to predict the performance characteristics of a composition.8 In most of these simulations, the calculations
multilayered glucose oxidase based enzyme electrode. Infor- were performed in dimensionlessunits and in all cases assumed
mation obtained includes range of sensor linearity under no perturbation of concentrations in external contacting
different substrate and cofactor concentration combinations, solutions.
sensor time response and output as permeability parameters The sensor simulation model used in this paper solves the
and enzymeloading are varied, sensorwashout and interference problem of multiple species diffusion through a diffusion-
effects, and predictionof decay of sensorenzyme concentration controlling resistance layer with reaction in an immobilized
and resultant estimated sensor lifetime as a function of enzyme layer. The mathematics to describe the simulation
contacting solution conditions and sensor permeability. can be reproduced easily by reviewing the diffusional portion
of solutions to prior simulations9and then adding in a reaction
In the development of amperometric enzyme electrodes, term to account for enzyme kinetics. The species concentration
the need arises to quickly identify optimized sensor membrane profiles and sensor outputs are solved by the Newton-Raphson
materials and geometries with a minimumof experimentation. method carried through iteratively to a desired degree of
These sensor materials can serve one or more purposes, convergence.lo
including the following: (1) tailoring of electroactive species This model expands upon the previous enzyme electrode
flux and thus device output and time response, (2) preferential simulations by the inclusion of an unlimited number of
extracting of either substrate or cofactor to linearize signal interfaces, thus allowing for effects due to both membrane/
in the desired measurement regime, (3) providing an immo- solution partition and partition from one membrane material
bilized enzyme reaction layer for conversion of substrate and to another. This permits changes in species diffusion coef-
cofactor to products (any of which may be monitored ficient as an interface is crossed and allows different element
electrochemically), (4) discriminating against either electro- thickness from layer to layer. The model makes provisions
active interferences or other chemically reactive constituents for calculating complete transients where any relevant species
which would degrade underlying materials, and ( 5 ) providing is the limiting reactant and allows the typical calculations for
the requisite hydration and osmotic properties to maintain response variations due to membrane thickness, enzyme
electrode stability. This process can be expedited by the activity, enzyme decay factors and kinetics, and species
development of a sensor simulation model which, with the diffusioncharacteristics. It allows concentration profile carry-
inclusion of a few known or experimentally determined over and parameter modification from one simulation to the
membrane material permeability parameters, can be iterated next, thus making possible the description of significant but
upon in different final geometries to predict transient and previously unreported phenomena, such as sensor washout
steady-state response curves. Thus, the experimenter can
specify the type of material and appropriate dimensions Mell, L. D.; Maloy, J. T. Anal. Chem. 1975, 47, 299-307.
required to fit the desired application. Leypoldt, J. K.; Gough, D.A. Anal. Chem. 19&4,56, 2896-2904.
Leypoldt, J. K.; Gough, D.A. Biotechnol. Bioeng. 1982, 24, 2705-2719.
Previous digital simulation models of enzyme electrodes Leypoldt, J. K.; Gough, D. A. Applied Biochemistry and Bioengineering;
have been used to explain observed electrode transient behavior Academic Press: New York, 1981; Vol. 3, pp 175-206.
( 5 )' Bergel, A.; Comtat, M. Anal. Chem. 1984, Sa, 29W2909.
and to predict steady-state outputs for a variety of membrane (61' Tse, P. H. S.; Gough D.A. Anal. Chem. 1987, 59, 2239-2344.
parameters.1-8These papers explain extension of substrate (7j Lucisano, J.; Gough, D. A. A M I . them. I.WI,~O, 1272-1281.
( 8 ) Tse, P. H. S.; Gough, D.A. Biotechnol. Bioeng. 1987, 29, 705-713.
(9) Brumleve, T. R.; Buck, R. P. J. Electroanal. Chem. 1978,90, 1-31.
University of Wisconsin Medical Center. (10) Camahan B.; Luther H. A.: Wilkcs, J. B. AppliedNumerical Methods; John
* Markwell Medical Institute. Wiley; New York, 1969; pp 274-321, 436540.

1520 Analytical Chemistry, Vol. 66,No. 9, May 1, 1994 0003-27OOf 94/0366-1520%04.50/0


0 1994 Amerlcan Chemlcal Socletv
Tabk 1
paramete9 solutionb resistance enzyme interference hydration
D (GLU) 3.6 X 10-6 2.2 x 10-7 1.0 x 10-8 nmc 1.0 x 10-8
K 1 0.075 0.15 0 0.15
D (02) 2.1 x 10-4 6.0 x 10-7 1.0 x 1od 1.0 x 10-7 1.0 x 10-6
K 1 3.5 1 1 1
D (HzOz) 1.0 x 10-4 5.8 x 10-7 5.0 X 10-8 3.3 x 10-8 5.0 X le
K 1 1.41 1 1 1
D (ASC) (2-4) X 10-6 1.8x 10-7 1.0 x 10-8 nmc 1.0 x 10-8
K 1 0.15 0.25 0 0.25
D (ACETA) (2-4) X 106 5.2 X 10-8 5.0 x 10-7 1.0 x 10-8 5.0 X 10-’
K 1 8.8 1.5 1 1.5
Thickness$ cm 0.089 0.0040 O.OOO64 o.OOO1 O.OOO64

Q Units for D are centimeters squared per seconds and X is dimensionless. K is the ratio of the equilibrium membrane layer concentration
relative to the solution layer concentration. b Note that literature solution layer diffusion coefficientshave been multiplied by 10 to account
for mixed diffusion/convection effects. Not measurable. d Thicknesses are known to be within &20%for the thin interferencelayer and within
+5% for all other layers.

transients and interference effects. We have chosen to carry and partition coefficients. However, for our purposes in
out these simulations without normalizing to dimensionless describing an overall electrode system and in making relative
units since we are working in a relatively small subset of comparisons of parameter variation on electrode performance,
parameter space, and in our experience, the results were easier this technique proved useful.
to explain to nonspecialists who wished to evaluate our work.
Several experimental methods exist for obtaining the EXPERIMENTAL SECTION
necessary single-species diffusion and partition coefficients The permeability parameters of the complete system are
to include in the simulation model. One can assume that the listed in Table 1 under generic headings which describe the
concentration step technique commonly used with membranes contacting solution and membrane layer functions in sequence
contacting gas is valid and that only minimal perturbation of (Le., solution, resistance, enzyme, interference, and hydration).
external concentration occurs.1 However, when membrane Four replicate samples of well-defined thickness for each
permeability approaches that of the contacting solution, this membrane layer type were used to verify the permeability
assumption is badly in error as speciesconcentration gradients parameters obtained. Thicknesses for all layers were directly
extend well into solution. Even rapid stirring cannot com- measurable by micrometer to within f5%. Membranes were
pletely alleviate this problem.12 A better approach to separate given 24 h to hydrate fully prior to parameter measurement.
the membrane from solution permeabilities has been dem- Resistance and interference layer permeability parameters
onstrated by Gough and Leypoldt with a rotating disk were obtained directly from single layer measurements.
technique.13J4Variations of limiting current by adjusting the Enzyme and hydration layer permeability parameters were
electrode rotation rate allow extrapolation to the membrane obtained indirectly from multilayer measurements where all
permeability. Further combination of this technique with
analysis of transients due to voltage steps15 or concentration
steps16 allows an elegant separation of the membrane per-
other layer parameters had already been determined. In the
complete four-layer membrane, the interference layer is 10
times thinner than the samples used for initially obtaining
-
meability into its diffusion and partition components. permeability parameters and thickness is calculated to within
In the absence of rotating disk instrumentation, a further f20% from preparation volume, solution percent solids, and
technique comes to mind and in fact is used in obtaining the resultant membrane square footage.
permeation parameters for our model. A concentration step Enzyme layers of constant thickness, but with concentration
experiment is run and analyzed as for a gas contacting of glucose oxidase (Sigma Type VII) varied from 0.78 pM to
membrane to obtain initial diffusion and partition coefficients. 0.2 mM, were prepared to study effects of enzyme loading.
These values, with the appropriate distance grid describing The0.2 mM concentration resulted in -4% enzyme by weight
membrane and solution regions, are then included in the within the enzyme layer. This by weight percentage scaled
simulation model and iterated through several cycles providing proportionately with lowering of enzyme concentration.
updated values for these parameters until the experimental vs Glutaraldehyde (Sigma Grade I, diluted to 0.1%) served as
calculated transient is matched with respect to both output the cross-linker. All four membrane layers were polyurethane
and time response. This resultant information for single species in nature and prepared by standard coating techniques.
is then included as part of the complete enzyme electrode Membrane samples for study are available from the authors.
model. Since the sensor’s solution boundary layer is not as The complete four-layer membranes define a portion of the
well-defined as in the RDE method, this technique does not system used in the Direct 30/30, a pocket-portable electro-
provide the same degree of accuracy in calculating diffusion chemical glucose sensing unit for diabetic patients.”
(11) Manoy, K. H.; Okun, D. A.; Reilley, C. N. J. Elecrroanal. Chem. 1962, 4, Diffusion and partition coefficients for the species hydrogen
65-92. peroxide, oxygen, glucose, ascorbic acid, and acetaminophen
(12) Malone, D. M.; Anderson, J. L. AIChE J. 1977, 23, 177.
(13) Gough, D. A.; Leypldt, J. K. Anal. Chem. 1979,51,439-444. ~ ~~~~

(14) Gough, D. A.; Leypoldt, J. K. Anal. Chem. 1980,52, 1126-1130. (17) Updike, S.J.; Shults, M. C.; Capelli, C. C.; von Heimburg, D.; Rhodes, R.
(15) Gough, D. A.; Leypoldt, J. K. AIChE J. 1980, 26, 1013-1019. K.; Tipton-Joseph, N.; Anderson, B.; Koch, D. D. Diobetes Care 1988, 11,
(16) Gough, D. A.; Leypldt, J. K. J. Electrochem. Soc. 1980,127, 1278-1286. 801-807.

Ana&ticaI Chemistry, Voi. 66,No. 9, May 1, 1994 IS21


Exposure times for these specific glucose and oxygen con-
centration permutations were varied and then the electrode
assembly was reexposed to storage solution (air saturated with
no glucose). This allowed following of sensor washout curves.
Interference transients were also checked by performing
Figure 1. Cross-sectional view of three-electrode cell for permeability concentration steps using freshly prepared solutions of ascorbic
measurements: (1) bottom case, (2) top case, (3) electrode insert, (4) acid or acetaminophen.
electrodes, (5) snapon membrane holder, (6) membrane, (7) sample
well, and (8) space for electronics. Enzyme deactivation rates were measured using our lower
loading enzyme recipes in a manner equivalent to that described
by Gough.8 This information was then entered into our
in the polymer materials of interest were determined by the simulation model, the simulated steady-state species concen-
combination of permeability measurements with iterative trations and enzyme fractions were calculated as a function
digital simulations. All solutions used were prepared with of contacting solution exposure and membrane permeabilities,
distilled, deionized water (McGaw), reagent grade chemicals and sensor lifetime was simulated.
(J. T. Baker, Aldrich, Amend), and research grade gases To look at the kinetics of the possible decay mechanisms,
(Matheson). Direct 30/30 top cases in standard three- contacting solutions were used which would leave glucose
electrode c~nfigurationl~ were used as the electrode test oxidase in either its resting state or with known fractions of
platform (Figure 1). The working and counter electrodes the reduced forms. The starting enzyme concentration was
were 0.020-in.-diameter Pt wire while the reference electrode */@ofthe standard recipe or lower so that the electrode output
was a 0.020-in.-diameter silver wire. All electrodes were potted to a 11 mM glucose concentration step was approximately
in epoxy, and the final assembly cut to a 0.200 in.-radius linear with enzymeconcentration. Theconcentration of active
dome, which was then polished at the surface. The silver wire enzyme was then calculated from the ratio of sensor output
reference electrode was then chemically chloridized.18 An after a defined decay period to the starting sensor output. The
eight-channel, three-electrode potentiostat (Markwell Medi- averages from four membrane samples were used to calculate
cal) with switching box was used to direct sensor outputs for remaining active enzyme concentration.
computer acquisition. Potentials were set to -0.6 V (vs AgC1) To look at the deactivation of the resting-state enzyme,
for oxygen reduction and +0.6 V for hydrogen peroxide, membranes were suspended in a phosphate-buffered saline
ascorbic acid, and acetaminophen oxidations. Glucose pa- (pH 7.05) contacting solution at a p02 of 150. Two sets of
rameters were determined indirectly by monitoring the samples were tested at temperatures of 22 and 37 OC. Fresh
peroxide formed in a thin layer of laminated polyurethane/ samples were selected periodically over 4 months from these
glucose oxidase on the electrode side of the membranes. storage solutions and tested after mounting on Direct 30/30
The transient responses for solution concentration steps cases.
were acquired with a Macintosh Plus using a Remote
To study peroxide-induced decay pathways of the reduced
Measurement Systems A/D board and Quickbasic software.
enzyme fractions, the phosphate-buffered saline solution also
Since the three electrodes were situated directly beneath the
included 27.5 mM glucose and 5 mM hydrogen peroxide.
membrane, concentration steps were made by blotting dry
Decay measurements were made using membrane samples
the top side of the membrane and then pipeting the solution
exposed to this solution for varying lengths of time. Solution
of interest into a 50-pL well. Data acquisition synchronization
p0z was set to either 0 or 30 mmHg by continuous degassing.
with sample application was assured by applying the sample
This provided a high level of convection and allowed accurate
at the sound of a computer-generated tone. In cases where
calculation of oxygen and glucose concentration within the
oxygen exclusion was required for extended times, a capped
enzyme layer. These oxygen levels produced preferential decay
polycarbonate cylinder was positioned above the electrode to
pathways for the enzyme. At 30 min to 4-h intervals,
increase the solution volume to 0.5 mL. This allowed easy
membrane samples were removed, rinsed in phosphate-
maintenance of a constant pOz. The concentration steps used
buffered saline to quench the decay, and tested as before for
to determine the permeability parameters included 0-5 mM
remaining active enzyme concentration. This process was
for hydrogen peroxide, 0-1 1 mM for glucose, 0-160 mmHg
repeated over four to six time intervals until at least a 60%
for oxygen, 0-6 mM for ascorbic acid, and 0-0.8 mM for
decay in active enzyme concentration occurred.
acetaminophen. A trace of catalase was included in the glucose
solutions to scavenge any enzyme-produced hydrogen peroxide Simulations of electrode output and membrane layer
which diffused back through the membranes to the contacting concentration profiles were run on a Macintosh SE/30
solution. computer using True Basic as the programming language.
Additional concentration steps were performed on the The time required to run any simulation is dependent on the
complete membrane system to provide the data required for number of species, elements, and time steps desired. The
model verification. Linearity was checked by varying the typical transients shown here used eight or less species, 10
glucose level from 0 to 55 mM in air-saturated solution. Further elements per layer, and 120-480 time steps. The species
variation was made by checking the higher glucose level steps included glucose, oxygen, hydrogen peroxide, interferent,
(22-55 mM) as a function of solution oxygen concentration oxidized enzyme, glucose-complexed enzyme, reduced enzyme,
(30-160 mmHg, verified by either Radiometer BMS3 Mk2 peroxide-complexed enzyme, and deactivated enzyme. An
Blood MicroSystemor ColeParmer 551360oxygen monitor). eight-species, 50-element, 120-time step simulation required
-20 min to complete. In addition, the algorithm is designed
(18) Sandifer, J. R.And. Chem. 1981, 53,312-316. in such a way that calculation times will scale directly with

1522 AnalyNcal Chemlsiw, Vol. 88, No. 9, M y 1, 1994


changes in any of these threevariables. To verify the accuracy D, = (zX2)/(?r2t) (4)
of the simulation, the time step size was varied from 0.05 to
5 s to find the times which showed no appreciable distortion where for example z = 1.37 at it/iss= 0.5. The partition
of the simulated current curves due to diffusional/kinetic coefficient then is calculated from
mismatches. In addition, mass balance checks were made on Km = (issu)/ (nFADmCg) (5)
the resultant concentration profiles to ascertain that no
distortion occurred. We typically find the averages for D , and Kmby determining
the times to fractional currents of 0.1, 0.3, 0.5, 0.7, and 0.9
The simulation of enzyme decay as a function of contacting
using the corresponding z factors of 0.66,l .OO,1.37,1.89, and
solution required an important approximation. This was
3.00.
related to the simulation time step that can be used without
These D / K pairs serve as the starting point in the iterative
-
distortion of the concentration profiles. For sensor transient
development, it took 10 min of computer time to simulate
1 min of real time. However, simulating months to years of
process to match experimental data to theoretical transients.
After these values are plugged into the relevant distance grid
enzyme decay with these small time steps presented a which defines our experimental conditions, the concentration
computational bottleneck. This was addressed by realizing step is then simulated with literature values for solution
that enzyme decay occurred slowly. To start the simulation, diffusion coefficients in all solution-layer elements. Curve
initial steady-state concentration profiles for specific contacting matching of simulated to experimental data for a more
solution conditions were first established using a time step of permeable species, such as hydrogen peroxide, is performed
0.25-1 s. However, from this point on we scaled up the decay first to check for mismatches in transient curve shape due to
kinetics from seconds to days as long as two conditions were solution permeability effects. In this experimental situation,
met. First, the product of the enzyme decay kinetics and the where sample application is less well defined than with a RDE,
time step was sufficiently small so that no more than 1% of the classical approximation of D,ff = D a t a sensor boundary
the hydrogen peroxide steady-state concentration was removed layer of defined thickness and D,ff >> D in bulk solution is
during any time step by reaction with enzyme subpopulations. not likely to apply. We have used the approximation of D,ff
Second, after each half-loading of simulated enzyme decay, = 1OD throughout the entire solution layer. This approxima-
the simulation was put into a "zero decay" mode until it was tion is suggested through the high-quality curve shape match
verified that glucose, oxygen, and hydrogen peroxide were of experiment to simulation, but it should be recognized that
indeed at the correct steady-state concentrations for the this factor would be different if applied to a different system
remaining active enzyme. These considerations guaranteed with different permeability parameters and geometry. Once
that enzyme subpopulations were correct and minimized the this modeling of the solution layer is defined, then iterations
underestimation of decay that would occur at highest enzyme in the membrane resistance layer D / K pairs with further
loading when the simulation was being based on hydrogen matching of real to simulated data completes the parameter
peroxide concentration that was a few percent low. Worst- determination. This process is repeated for glucose, hydrogen
caseerrors with this approximation resulted in underestimated peroxide, and oxygen as each successive membrane layer is
added.
decay times of 10-15%. Thus, as long as the concentration
estimates of the enzyme subpopulation fractions are correct The uncertainty introduced by this model of solution
and hydrogen peroxide is close to the correct steady state, permeability has little effect on the reliability of experimentally
propagation of error in enzyme decay during the course of the determined membrane resistance layer permeability param-
simulation is reasonable. eters for the less permeable species, which include glucose,
ascorbic acid, and acetaminophen. However, for more
permeable species such as oxygen and hydrogen peroxide, the
RESULTS
tabulated values of the resistance layer permeability com-
Determination of Model Parameters. The gas contacting ponents are probably not better than 45-896 for D and f 1 0 -
model serves as the starting point and predicts steady-state 12% for K . This does not imply equivalent errors in the total
and transient output currents of the form layer permeability so much as a lack of resolution between its
D and K components.
is, = (nFAD,K,C,)/X (1)
Representative fitted curves are shown for two- and four-
and
layer membranes for hydrogen peroxide and glucose in Figures
it = i,(l + 2~(-l)"e-"") where z = (?r2Dmt)/X2 (2) 2 and 3, respectively. All subsequent simulation predictions
are made for the full sensor system where four of the layers
and C, equals the bulk solution concentration (mol/cm3), X are membrane materials and the fifth layer is the contacting
themembrane thickness (cm), t the time (s), D, the membrane solution. The full permeability parameter sets for these layers
diffusion coefficient (cm2/s), and Km the dimensionless are included in Table 1.
membrane partition coefficient and n, F, A , and ?r have their Oxygen Cofactor ConcentrationLimitationEffects. Figure
usual meaning. 4 shows the simulated oxygen concentration profiles obtained
When expanded, eq 2 leads to through 120 s for the glucose concentration step of Figure 3
(0-1 1 mM glucose, p02 = 160 mmHg = 0.000 25M). Notice
it/iss= 1 - 2e-' + 2e4' -2 +
~ ~2e-16'
' ...
(3) that oxygen reservoirs are present in the membrane layers
Thus, by determining the factor z at different fractional cur- prior to the concentration step. Once the step occurs, new
rents, it becomes possible to solve for D, using the equation profiles are generated with the glucose profiles reaching a

Analytical Chemism, Vol. 66, No. 9, May 1, 1994 1523


40 r

0
0 30 60
0 30 60 TIME (SECONDS)
TIME (SECONDS)
Flgwe 2. Slmulated(solid1ines)vsexperimental sensor outputtransients
for 0.5 mM hydrogen peroxide (filled squares) and 11 mM glucose
(open squares) concentratlon steps through two-layer membrane.
Electrode area is calculated to be 2.0 X cm2.

20 r

0 11 22 33 44 55
mM GLUCOSE

Flgure 5. (a) Simulated output transients for (1) 5.5, (2) 11.O, (3)27.5,
and (4) 55 mM glucose concentration steps with a contactlng blood
pO2 of 30 mmHg. (b) Simulated outputs at (1) 10, (2) 15, and (3) 20
0 60 120 s for the four blood glucose concentration steps at a pOzof 30 mmHg.
Note that best linearity Is obtained at shorter measurement times.
TIME (SECONDS)
Flgure 3. Shnulated (solkl1lnes)vsexperknentalsensor output transients
for 0.5 mM hydrogen peroxlde (filled squares) and 11 mM glucose
(open squares) concentration steps through four-layer membrane. When dissolved oxygen concentration decreases in the
Electrode area is calculated to be 1.8 X lo3 cm2. contacting solution, as is the case for many relevant matrices
Membrane Layers - Function and Thicknesses
such as venous blood samples (0.1-0.025 mM), oxygen flux
into the membrane very badly lags glucose flux. As Figure
hydration interference enzyme resistance solution
0.00064 cm 0.0001 cm 0.00064 cm 0.0040 cm 0.089 cm 5a shows for simulations of 5.5-55 mM glucose with 0.05
mM dissolved oxygen, membrane oxygen quickly becomes
Oxygen Concentration Profiles exhausted and the sensor converts from a glucose-dependent
to an oxygen-dependent sensor. At the higher glucose levels,
the current begins to decay as early as 15 s after sample
application and current is now controlled by oxygen rather
than glucose flux. Thus, steady-state current is no longer
useful for analytical purposes. However, by simulating a
family of output transients as a function of glucose concen-
0
D. trations at the lowest concentration of oxygen to be encoun-
I L
E tered, it is possible to identify a usable time window over
which glucose linearity is maintained. This time window can
then be incorporated into the analysis software, as was done
for the Direct 30/30, to provide a high degree of measurement
linearity. This is demonstrated in Figure 5b.
Discrimination Against Interference Effects. The ability
to discriminate against electroactive interferences is usually
assumed to be due to the complete rejection (i.e., partition
coefficient of 0) of the species at one of the membrane
interfaces. However, as the species partition and diffusion
coefficients of Table 1 show, there must be an additional mode
of interference discrimination since the K for ascorbic acid is
vanishingly small while that for acetaminophen is large. This
second mode can be explained on the basis of the much slower
diffusion coefficient of acetaminophen relative to the other
species. As Figure 6 suggests, we can use the kinetic mode
to measure substrate in the first 30 s before development of
the interference signal at greater times. An analogous type
of time domain discrimination has been demonstrated pre-

1524 Analytical Chemistry, Vol. 66, No. 9, May 1, 1994


viously by Sandifer18 for chloride potentiometric sensors in
the presence of bromide. 'Or 2
Sensor Washout Effects. One phenomenon of significance
which has not been reported in previous enzyme electrode
simulation studies is that of sensor washout. This relaxation
in sensor output occurs when the contacting sample solution
is replaced with one containing no substrate and a normal
level of cofactor. The speed of this relaxation or recovery
process determines the fastest possible sensor cycle time, which 0
is of great practical importance to the analyst trying to run 0 30 60
a series of samples. Three classes of relaxation curves can TIME (SECONDS)
occur during sensor washout. In the simplest case, when Figure6. Simulated output response fora 1 1 mM glucose concentratlon
step in the absence (1) and presence (2) of 0.4 mM acetaminophen
cofactor remains in excess during the entire measurement, interference.
the curve relaxes toward its zero level in a time comparable
to the initial transient development time.
In situations where the enzyme layer cofactor has been fits shown in Figure 8a for transient current as a function of
depleted due to extended exposure at high substrate concen- enzyme loading. The match is quite good at highest loading
tration, the relaxation curve to wash solution is nonmonotonic but appears to lag in time and underestimate current magnitude
(Figure 7a). When the contacting cofactor concentration is as loading is decreased.
low to start with and then is brought back to normal levels, To try to obtain a better match, the different k values are
this transient can be quite striking. This occurs because individually increased. Increases in k2, k3, and k4 by up to
increased cofactor flux reaches and reacts with the ac- a factor of 2 leave essentially the same transient as before.
cumulated glucose in the enzyme layer. The resultant output Increasing kl produces a faster development of the transient
moves to its highest level with cofactor flux maximization and as well as an increase in magnitude. When kl is increased by
holds this level until the residual substrate has been completely a factor of 2, fairly good matches result for ail enzyme recipes,
scavenged from the enzyme layer. At this point, the cofactor as shown in Figure 8b.
begins to replenish itself in the enzyme layer and a region of Enzyme Decay Measurements. Prior work has demon-
relatively fast decay occurs with removal of the last residual strated that glucose oxidase can decay through three primary
substrate and reaction product. pathways.8 An apparent "autoinactivation" can occur when
The last class of relaxation curve occurs for a slowly the enzyme layer is submersed in fluid containing no glucose
diffusing interferent which has not achieved its maximum and no hydrogen peroxide. This resting-state enzyme shows
inward flux. Upon return to wash solution, the output transient a simple first-order decay of the form
appears unaffected and continues to grow in magnitude.
However, at some point the bidirectional flux of species will dxldt = k,,(a - x) where a = total and
begin to diminish the concentration profile near the electrode. x = inactive enzyme (8)
At this point the output transient reverses and begins its slow with the solution
decay back toward zero (Figure 7b). Samples of this type
give an extremely slow decay and can increase the cycle time x = a(1 - with a half-life of r1,2 = 0.693/k,,
(9)
and apparent baseline of the sensor when the interferent In this work deactivation of the resting-state enzyme at 37
concentration is high. OC, as shown in Figure 9a, was fit to project a half-life for
Enzyme Loading Effects. In setting up enzyme loading
permutations, enzyme was lowered by factors of 2 from the
standard recipe to a ' 1 2 5 6 recipe. Assuming homogeneous
this pathway of 105 days. From this information, klt is
-
calculated to be 7.64 X l e 8 s-l. This is 15% slower than
that measured in previous work.8 This rate of decay did not
distribution in the enzyme layer, no enzyme decay during change when externally added peroxide but no glucose was
preparation, and concentrations calculable from normal added to the contacting solution.
geometric factors, enzyme concentrations ranging from 2.00
The other two pathways of enzyme deactivation occur in
X 10-4to 7.81 X M can be predicted. For the standard
the presence of hydrogen peroxide when contacting solution
glucose oxidase mechanism,19
conditions force enzyme to the reduced or peroxide-complexed

E
ki
+ G -,RL -..R + L
ki
(6)
forms. The reduced form is exclusively formed in glucose
solutions with zero oxygen. Both forms are present in
R + 0, -
k,
E*-H202 E -+
k4
H202 (7)
the room-temperature rate constants given in the literature20
significant fractions when glucose solutions are at low PO*.
These decay pathways are "psuedo" first order and of the
form
are kl = 12 200 M d , k2 = 1220 s-l, ks = 2 000 000 Mss-', dx/dt = (kit + k2&)(~ - X) + k&)(c - x)) ( a - X) (10)
and kq = 660 s-l. Combining this information with the
previously determined species permeability data gives the curve where c = peroxide, fR = fraction of reduced enzyme, and fc
= fraction of peroxide-complexed enzyme, thus leading to the
(19)Gibson. Q.H.; Swoboda, B. E. P.; Massey, V. J. Biol. Chem. 1964, 239, solution
3927-3934.
(20) Duke, F. R.; Weibel, M.; Page, D. S.; Bulgrin, V. G.; Luthy, J. J. Am.Chem. = a( 1 - e-(ki,)(t)-(kz,)(/~)(c)(f)-(kjr)(/~)(c)(r))
Soc. 1969, 91, 3904-3909 (1 1)

Analytical Chemistry, Vol. 66, No. 9,May 1, 1994 1525


1.0 s

,0
K
K
3
u ,6
n
W
N
i ,4
a
E
K
0 ,2
z
0 90 180
TINE (SECONDS) .o
0 120 240 360 480 600
TIME (DAYS)
1.0

ii
w -8
K
K
I)
u ,6
n
W
N
.4
E
K
0 90 180 p #2
TINE (SECONDS)
Flgure 7. (a) Experimental (squares) vs simulated (sotid line) output .o
transient for the case where the inittal contacting sdutlon ia 27.5 mM 0 3 6 9 12 15
Inglucose and at a p02of 30 mmHg, while thewash solutionintroduced TINE (HOURS)
at 90 s contains 0 mM glucose and Is at a pO2 of 160 mmHg. (b)
Simulated output transient for the slowly diffusing Interference
acetaminophen Introduced at 0.4 mM concentration at t = 0 s with
a return to an lnterferencafree wash solution at t = 60 s.

10
..................................................................................................
I i\
h

2
v

I-
22OC
E 5
K
K
3
u
0 5 10 15 20 25
0 TINE (HOURS)
0 60 120 Flgure 9. (a) Enzyme decay at 37 O C from restingstate enzyme.
TINE (SECONDS) Exponential fit projects an enzyme half-life of 105 days. Points are the
average value of four sample membranes. (b) Enzyme decay at 22
and 37 O C for a contacting solution of 0 mmHg pop, 27.5 mM glucose,
and 5 mM hydrogen peroxide. The simulation predicts an equilibrium
.I enzyme layer glucose concentration of 4.1 mM. (c) Enzyme decay at

t _I
. / - - 22 and 37 O C for a contacting solution of 30 mmHg p02, 27.5 mM
glucose, and 5 mM hydrogenperoxide. The sknulatkm predictsa steady-
state enzyme layer glucose Concentration of 2.9 mM and an enzyme
layer pop of 18 mmHg.

low oxygen, and externally added hydrogen peroxide are shown


in Figure 9b,c. To determine rate constants from these decay
0 60 120 curves, the enzyme fractions are first calculated. At a
TIME (SECONDS) contacting solution p02 of 0, the enzyme is fully converted to
Flgure 8. (a) Experimental vs simulated transients for 1 1 mM glucose the reduced form by excess glucose. At a contacting solution
step as a function of enzyme concentration. (1)2.0 X lo4, (2)3.13
X lO-O,and (3)7.81 X lo-' M. All kinetic parameters are at literature pOz of 30 mmHg (0.05 mM) with oxygen flux to the enzyme
values. (b) Experimental vs simulated transients for 1 1 mM glucose layer from both sides of the membrane, simulations using the
step as a function of enzyme concentration: (1)2.0 X lo4, (2)1.25 previously defined permeability parameters and modified
X (3)3.13X and (4)7.81 X le7 M. k, istwkethe Herature
value. enzyme kinetics predict enzyme layer glucose and oxygen
concentrations of 2.9 and 0.028 mM. This results in predicted
and enzyme fractions of 0.522 for the reduced form and 0.044 for
the complexed form. Rate constants at 22 and 37 O C are then
fl,2 = 0.693/(k1,+ kZ,CfR)c+ k&&) (12) calculated to be 0.013 and 0.029 (M.s)-' for the reduced
The results for enzyme concentration decay in excess glucose, enzyme and 0.075 (Mas)-' and 0.132 (M.s)-' for thecomplexed

1526 AnalyticalChemlstty, Vol. 66,No. 9, M y 7, 1994


0
5
6
7
a
9
10
11

0 180 360 0 240 480


TIME (SECONDS) TIME (SECONDS)
Flgure 10. Simulated output transients for 22 mM glucose (filled Figure 11. Simulated sensor output vs time for 0-1 1 mM glucose
symbols)vs11 mM glucose (opensymbols)for Direct 30130 membrane concentrationsteps using membranesof implantablepermeabiilty with
sensor covered with an additional 0.0025-cm-thick resistance layer. varied enzyme loading. ID numbers represent the number of half-
The glucose diffusion coefflcient for the extra layer is either (1) 5.50 loadings below the standard recipe of 2.00 X lo4 M: (0) 2.00 X lo4,
X (2) 2.75 X or (3) 1.84 X cm2/s. Oxygen diffusion (5) 6.25 X (6) 3.12 X lo", (7) 1.56 X lo4, (8) 7.83 X lo-', (9)
coefficient did not change from that for the Direct 30130 membrane. 3.92 X lo-', (10) 1.96 X lo-', and (11) 9.80 X M.

-3
enzyme. These experiments compare favorably with the 37
OC decay constants of 0.02 and 0.76 (M.s)-' measured
- -4
W
I

w
previously.8 2 -5
Simulationof Permeability Parameters for an Implantable t;
Glucose Sensor. This model was used in predicting the required Q -6
0
0
membrane permeabilities for an in vivo subcutaneous glucose -I -7
sensor. The challenge here was to continuously maintain linear
response to glucose in the useful clinical range (1 5 - 2 2 mM) -0
with tissue oxygen tension as low as 30 mmHg. Thus, at 0 225 450 675 900
equilibrium the flux of oxygen from contacting environment TIME (DAYS)
to the enzyme layer must always be greater than that for Figure 12. Simulated active enzyme Concentration decay curves vs
time for Implantable membranes continuously exposed to different
glucose. Simply increasing membrane oxygen permeability glucose concentrations of 0.0, 11.0, 22.0, 24.8, and 27.5 mM at a
does minimal good since this will result in a depletion layer contacting solution pOn of 30 mmHg. Starting active enzyme
into the contacting solution and only a small increase in oxygen -
concentrationis 2 X lo4 M; (.I and (- -1 respectively represent the
enzyme concentrations where 90 and 50% of the sensor output
flux to the sensor area. The alternative is to lower glucose remains.
flux through the outermost resistance membrane. This can
be done by lowering the partition or diffusion coefficients.
The former would be preferred since this would not decrease enzyme concentration and its correspondence with sensor
the sensor response time. For us, finding materials with low output is given in Figures 11 and 12. Figure 11 shows the
partition coefficients that maintain reasonable diffusion simulated output transients to a 1 1 mM glucose concentration
coefficients has proved illusive. Thus, theoretical and ex- step for sensors with implantable membrane permeabilities
perimental work has focused on materials with comparable and enzyme concentration lowered in 50% increments from
partition coefficients and decreased diffusion coefficients. 0.2 mM to 0.098 pM. Note that approximately six half-
Membrane and solution permeabilities listed in Table 1 loadings of enzyme must be removed before signal drops more
serve as the simulation starting point. The contacting solution than 10% and that nearly nine half-loadings must be removed
p02 is set to 30 mmHg. Now a fifth membrane layer is to drop the signal more than 50%. These serve as standard
interposed between this solution and the previous membranes. curves for predicted sensor output.
As the glucose diffusion coefficient is lowered by up to a factor In Figure 12, the active enzyme concentration vs time has
of 12 in this fifth layer (with no change in glucose partition been simulated for membranes with implantable sensor
coefficient and oxygen permeability), the family of simulated permeabilities. The contacting solution p02 was 30 mmHg
curves shown in Figure 10 is generated. This desired set of and glucose was varied from 0 to 27.5 mM. Note that these
parameters predicts the development of a linear signal from plotted enzyme concentrations are the lumped sum average
0 to 22 mM. This strategy allowed us to develop membrane of remaining active enzyme throughout all enzyme elements.
resistance layers which, upon continuous aqueous solution These remaining active enzyme concentrations can be com-
exposure, could provide in vitro glucose linearity through pared to Figure 11 to predict the remaining sensor output to
22 mM when pOz is as low as 20 mmHg. In addition, these a standard glucose challenge. Two lines are drawn at the
membranes have now been used in subcutaneous glucose enzyme concentration levels where 90 and 50% of the standard
sensors for periods of up to three months in a dog model.21 loading sensor output remains. At 90% sensor output, only
minimal changes in overall sensor performance will have
Estimation of Implantable Sensor Lifetime Limit Due to
occurred. However, at the 50% sensor output level, the sensor
Enzyme Decay. A useful presentation of remaining active time response will be so slow (2-4 times slower) and the output
(21) Updike, S.J.; Shults, M. C.; Gilligan, B. J.; Rhodes, R . K.;Luebow, J. 0.; so enzyme concentration dependent that the sensor is no longer
vonHeimburg, D. ASAIO Trans., in press. useful.

Analytical Chemistty, Vol. 66,No. 9, May 1, 1994 1527


There are several points to note from this plot. The first with loss of cofactor. The subset of parameter tradeoffs which
is that, even with no glucose exposure, the enzyme concentra- can match this curve type is much smaller than that for the
tion will decrease from resting-state enzyme decay. This simple glucose transient with adequate cofactor, and thus we
represents the best-case limit of electrode lifetime. The second approach convergence on the final parameter set.
is that as long as the glucose flux is lower than the oxygen flux A more stringent test is the prediction of the washout curve
into the enzyme layer, the enzyme decay will be only slightly after cofactor-limited situations. In these situations, substrate
faster than this slowest decay pathway. Finally, once glucose will have accumulated in the enzyme layer after cofactor was
flux is greater than oxygen flux, glucose will accumulate in depleted. Once the contacting solution is replaced with another
the enzyme layer, the fractions of reduced and complexed containing no substrate and normal levels of cofactor,
enzyme concentrations will increase, and enzyme concentration accumulated substrate will begin to either react or diffuse
decay will accelerate. The contacting solution concentrations away. The shape and magnitude of the output washout curve
where this occurs will be very sensitiveto both initial membrane and the time required for removal of residual substrate can
permeabilities and changes in permeability that may occur in be used to verify enzyme kinetic parameters and partition of
the sensor membranes due to aging, foreign body encapsula- substrate from the resistance into the enzyme layer. This is
tion, or membrane pore occlusion. However, if as a starting checked at several widely spaced permutations of substrate
point we are willing to believe this simulated time to 50% and cofactor concentrations to verify that some unknown
output, then we can speculate that glucose concentrations in artifact has not assisted in this match.

-
the typical diabetic patient range can be in continuous contact
with the sensor for 2 years without appreciably accelerating
the speed of enzyme decay. Thus, enzyme lifetime is not
The final test for the model is its ability to predict transient
response curves as a function of enzyme loading. That we
could simply take literature values for the enzyme kinetics as
likely to be the primary limiting factor in implantable sensor determined in solution and have them transfer so well to the
development. immobilized membrane case is quite remarkable. The small
increase in kl to match a 250-fold swing in enzyme loading
DISCUSSI ON implies that only minimal deactivation occurred in the
In our system, there are several experimental situations production of these membranes.
which we need to be able to predict accurately before we can The question arises as to what is considered a good fit of
consider our model complete. These include the following: simulation to experimental data. It is not possible to simply
(1) single species permeability measurements corrected for do a least-squares fit and project from the residuals which of
contacting solution effects, (2) electrode output magnitude the simulation parameters should be changed to get a better
and time responses as a function of membrane thickness (not fit. For this work, the answer has been empirical in that if
cofactor limited), (3) range of substrate linearity as a function all relevant situations are matched without experiencing large
of cofactor concentration, (4)output transients vs time for errors in transient development time response or output, then
cofactor-limited cases, (5) output transients vs time for we will be satisfied with the model. This seems reasonable
different enzyme loadings (at constant membrane thickness since there are unquestionably certain assumptions in the model
and solution conditions), and ( 6 ) washout transients vs time which are oversimplified, such as layer homogeneity, unidi-
for a variety of solution cofactor/substrate concentration mensional diffusion, uniform convection in the contacting
permutations and exposure times. solution resulting in an average “effective” diffusion coefficient,
We start first with the simplest case, that of determining no temperature or membrane stretching effects on DIKpairs,
single species diffusion and partition coefficients. For the no time lags on solution exchanges, etc. Thus, the ability to
directly electroactive species this is easy. However, when we predict all major trends as opposed to an exact fit for one
get to the enzyme converted species measurements, a degree transient is our criterion for “goodness of fit”.
of uncertainty must initially be accepted in making the We have found many applications for this methodology in
measurement. This uncertainty occurs since we are using a our work. This model serves us well in quality control of
single output transient to try to explain several possibly resistance membrane polymer permeability and subsequent
counteracting effects including (1) the bidirectional flux of calculation of required thickness prior to production. It has
peroxide generated in the enzyme layer, (2) enzyme activity been used to follow enzyme decay over the shelf life of
and kinetics, (3) cofactor concentration, and (4) the substrate membranes. It proved valuable in defining the measurement
and cofactor partition coefficients in resistance and enzyme time window in the software for the Direct 30130 and provided
layers. In fact, no single transient of this type is uniquely guidance in establishing the required permeability range for
determined as tradeoffs in species concentrations, species DIK our implantable glucose sensor development. Finally, it was
pair, and enzyme kinetics and loading can provide essentially used in combination with enzyme decay measurements to
indistinguishable curves. estimate the possible lifetime theoretically obtainable for
To assess our initial glucose permeability estimate we sensors in an in vivo environment.
provide an experimental situation that will force a nonmono-
tonic output transient. For our system this is provided simply CONCLUSIONS
by setting up solution conditions where cofactor will quickly A simulation strategy has been developed for modeling the
be depleted. Thus our simulation substrate/cofactor fluxes glucose oxidase enzyme electrode. It provides for the
will have to be matched correctly to reproduce the experimental calculation of sensor performance information, such as range
sensor output magnitude and more importantly the time when of linearity, time response, output magnitudes, cofactor
cofactor becomes the limiting reagent and sensor output decays limitation regimes, and enzyme loading and decay effects. It

1528 AnalyticalChemistry, Vol. 66,No. 9, May 1, 1994


also allows one to establish different sets of contacting solution glucoseoxidase system, this modeling strategy should be easily
conditions (pure diffusion, RDE boundaries, convection, etc.) extendible to other enzyme electrode systems and has the
which can produce strikingly different performance patterns potential to be used with minimal modification to describe the
for the identical set of membrane parameters. Concentration detection of neutral species by spectroscopic as well as
profiles can be simulated under one set of solution boundary electrochemical means.
conditions and the decay then followed for a different set of
conditions. This allows for visualization of sensor washout ACKNOWLEDGMENT
transients, which provides information related to sensor cycle Markwell Medical Institute thanks the SBIR program
time and answers questions about enzyme loading and reaction within NIH for grant support (DK 4065743) pertaining to
mechanisms. Any number of layers can be handled, providing simulated and experimental implantable glucose sensors during
single-layer parameters can be adequately measured. There the course of this work.
is great flexibility in handling of input parameters, and in
fact, if one wishes, these can be changed during the course of
Recehred for revlew October 1, 1993. Accepted February 8,
the simulation (Le., enzyme activity decay, growth of stagnant 1994.'
layer thicknesses, etc.). Finally, while the experimental
verification of this simulation model was performed with the Abstract published in Advance ACS Absrracrs, March 15, 1994.

Anelytical Chemisby, Vol. 66,No. 9, May 1, 1994 1529

You might also like