Journal of Diabetes Science and Technology
ORIGINAL ARTICLES
Volume 1, Issue 2, March 2007
© Diabetes Technology Society
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Ritesh Tipnis, M.S.,1 SanthiSagar Vaddiraju, M.S.,1 Faquir Jain, Ph.D.,2
Diane J. Burgess, Ph.D.,3 and Fotios Papadimitrakopoulos, Ph.D.1,4
Abstract
Background:
The performance of implantable glucose sensors is closely related to the behavior of the outer membrane. Such
membranes govern the diffusion characteristics of glucose and, correspondingly, the sensitivity of the sensors.
This manuscript discusses the selection of various membrane materials and their effect on the device response.
Methods:
Sensors were fabricated utilizing a 50-µm platinum wire followed by immobilization of the glucose oxidase (GOX)
enzyme. Sequential adsorption of various ionic species via a layer-by-layer process created devices coated with bilayers
of humic acids/ferric cations (HAs/Fe3+), humic acids/poly(diallyldimethylammonium chloride) (HAs/PDDA), and
poly(styrene sulfonate)/poly(diallyldimethylammonium chloride) (PSS/PDDA). The in vitro amperometric response
of the sensors was determined at 0.7 V vs an Ag/AgCl reference electrode in phosphate-buffered saline (37ºC) for
various glucose concentrations. The diffusion coefficients of glucose and hydrogen peroxide (H2O2) through these
membranes were calculated and analyzed.
Results:
Outer membranes based on the sequential deposition of bilayers of HAs/Fe3+, HAs/PDDA, and PSS/PDDA were
grown successfully on immobilized layers of GOX. The amperometric response and reversibility upon changing
the in vitro concentration of glucose were investigated.
Conclusions:
Through alteration of the number of bilayers of the outer membrane, it was possible to modulate the diffusion of
glucose toward the sensor as a result of its flux-limiting characteristics. Semipermeable membranes based on five
HAs/Fe3+ bilayers exhibited a superior behavior with a minimum hysterisis response to glucose cycling and a
lesser current saturation at hyperglycemic glucose concentrations because of a more balanced inward diffusion
of glucose and outward diffusion of H2O2.
J Diabetes Sci Technol 2007; 2:193-200
Author Affiliations: 1 Nanomaterials Optoelectronics Laboratory, Polymer Program, Institute of Materials Science, University of Connecticut, Storrs,
Connecticut;2 Nanomaterials Optoelectronics Laboratory, Electrical and Computer Engineering, University of Connecticut, Storrs, Connecticut; 3 Department
of Pharmaceutical Sciences, University of Connecticut, Storrs, Connecticut; and 4 Department of Chemistry, University of Connecticut, Storrs, Connecticut
Abbreviations: (BSA) bovine serum albumin; (GOx) glucose oxidase; (H2O2) hydrogen peroxide; (HAs) Humic Acid; (LBL) layer by layer;
(PBS) Phosphate Buffer Saline; (PDDA) poly (diallyl-dimethyl ammonium chloride); (PPD) poly (o-phenylene diamine); (PSS) poly (styrene sulfonate)
Keywords: biosensors, outer membrane, layer by layer assembly, humic acids, diffusion through membranes
Corresponding Author: Dr. Fotios Papadimitrakopoulos, Professor of Chemistry, Polymer Program, Institute of Material Science, U-3136,
University of Connecticut, Storrs, CT 06269; email address [email protected]
193
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
Introduction
D
espite considerable research efforts, the fabrication of
long-lasting implantable biosensors for continuous glucose
monitoring remains elusive.1–4 A number of investigators
have employed the design of the Clark-type electrochemical
sensor,5 which is based on a sensing enzyme, and have
developed working prototypes of miniaturized devices.6–9
Although these sensors have exhibited excellent
characteristics, they are prone to long-term gradual
degradation and loss of response in both in vitro and,
more severely, in vivo applications.10 Various mechanisms
are associated with the failure of these sensors, among
which calcification,11 enzyme deactivation,12 and electrode
failure13 are prevalent. In addition, biocompatibility issues
originating from implantation-induced inflammation that
triggers biofouling, tissue fibrosis, and sensor entombment
contribute significantly to sensor failure.14
initial material of choice as an outer coating for glucose
sensors.21 However, the presence of acidic sulfonate groups
(R-SO3H) that coat the hydrophilic channels of Nafion act
as nucleation sites for the dissolved Ca2+ present in serum.
This leads to the formation of calcium phosphate crystals
that clog its pores and eventually crack this membrane,
causing sensor failure.11 It has been reported that spincoated Nafion films, preincubated with ferric chloride,
could prolong membrane lifetime.22 This is because of
the stronger interaction of Fe3+ with sulfonate groups, as
compared to Ca2+, which reduces nucleation of calcium
phosphate crystals and depresses the calcification process.
Enzyme-based glucose sensors work on the principle
of detection of hydrogen peroxide (H2O2) formed by the
oxidation of glucose. The H2O2 is electrochemically oxidized
under an applied potential of 0.7 V, and the current
measured is related to the concentration of glucose in
the system (see diagram in Figure 1). The semipermeable
membrane, depicted in Figure 1, along with assisting in the
prevention of biofouling, should also regulate the diffusion
of glucose. It is well known that for an enzyme-based
implantable sensor to work at its optimum efficiency in the
tissue, the ratio of oxygen, which regenerates the enzyme,
to the permeating glucose has to remain constant.15 Thus,
in the absence of a diffusion-limiting barrier, the kinetics
of the sensor reaction could become oxygen limited as
the result of the large amount of glucose vs oxygen in the
subcutaneous tissue.15 This leads to saturation of the signal,
and any further increase in glucose concentration does not
translate to adequate sensitivity.15 In addition, the surfeit of
glucose leads to the overgeneration of H2O2, which, if not
utilized or diffused out into the bulk, can lead to either
enzyme degradation16 or electrode saturation.17 Thus, the
requirement of a sensor for continuous monitoring of
glucose over the entire physiological concentration with
high sensitivity and short response time demands an outer
membrane with tunable permeability properties.
Figure 1: Various chemical, electrochemical, and diffusion processes
associated with the displayed sensor geometry of the currently studied
glucose sensor.
Layer-by-layer (LBL) deposition of conformal thin films is an
inexpensive and efficient tool for growing multicomponent
molecular architectures that overcomes the problems
associated with other film-growth methods.23 By applying
the simplistic, yet versatile approach of LBL assembly, our
group has grown conformal thin Nafion/Fe3+ films by
alternately dipping substrates into dilute solutions of Nafion
and FeCl3, respectively.18 It was observed that by carefully
controlling the pH and ionic strength of the solutions,
a uniform and timely film growth could be achieved.
Moreover, this led to an order of magnitude reduction
in calcification in these self-assembled membranes, as
compared to spin-coated Nafion films.18
Over the years, we as well as others have worked on
developing biocompatible, semipermeable membranes to
mitigate the plethora of failure modes of glucose sensors
due to an inefficient outer membrane.18-20 Nafion™ (an inert
and oxidatively stable fluoropolymer, which has good
mechanical properties and ionic permittivity) has been an
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
194
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
The hydrophobic micellar interior of Nafion within Nafion/
Fe3+ LBL assemblies inhibits the permeation of hydrophilic
moieties, leading to low sugar diffusion.18 This led us
to investigate more hydrophilic polyelectrolyte humic
acids (HAs) (naturally occurring biopolymers) that, when
assembled with FeCl3, provided the required degree of
glucose permeability.19, 24 A LBL self-assembly was performed
using dilute solutions of HAs and Fe3+ cations. As with
Nafion assembly, a thick and uniform film was obtained
by adjusting the pH of the HAs solution. Moreover, in vivo
results indicated that HAs/Fe3+ coatings were biocompatible,
with reduced tissue fibrosis (as compared to the uncoated
control samples) following implantation.19,20
MW 200,000–350,000] were used as received from Aldrich.
o-Phenylenediamine was obtained from Acros chemicals.
The platinum (Pt) and silver wires were purchased from
Ladd Research. A 1-µm glass fiber membrane was purchased
from Pall Corporation. Deionized water was produced by
a Millipore (Milli-Q) system with resistivity >18 MW and
used to prepare all aqueous solutions.
Sensor Fabrication
Miniaturized sensors were fabricated on a 0.3-mm-diameter
monofilament nylon line, which served as the backbone.
The three electrodes, namely the working, reference, and
counter electrodes, were made by skillfully winding the
respective wires on the nylon backbone in close proximity
to each other.
In the present contribution, application of the aforementioned
LBL technique to grow the semipermeable coating of
the glucose sensors is discussed with an emphasis on
introducing a flux-limiting membrane as a tunable barrier
for the diffusion of glucose toward the immobilized layer
of the redox enzyme (i.e., glucose oxidase). Apart from
the HAs/Fe3+ bilayers, we self-assembled films of HAs/
poly(diallyldimethylammonium chloride) (PDDA) and
poly(styrene sulfonate) (PSS)/PDDA onto the sensory device
with the intent of investigating the relationship between
the permeability of these LBL assembled outer sensor
membranes and the magnitude of the electrochemically
detected current. Using the flexibility afforded by LBL
assembly methodology, we developed these model systems
with markedly different microstructures and permeability
characteristics, which provide a basic understanding of the
system and how to deduce the interdependence between
the various electrochemically participating species and
their effect on sensor behavior. The diffusion coefficients
of glucose and H2O2 through these membrane systems
were investigated in order to explain the individual sensor
response as it pertains to the microstructure of the outer
semipermeable membranes. The hysterisis behavior (defined
as the lag in the current response when the glucose level
is cycled from lower to higher concentration and back) of
these sensors was studied as a function of the permeability
of the outer membrane.
The working electrode was in the form of a 50-µm Pt wire
wound to yield a total surface area of 3 mm2. The sensors
were cleaned via electrochemical oxidation in a 0.5 M H2SO4
solution. This process of fabricating the devices is a variation
of the methodology followed by several researchers.7,8 The
miniaturized Ag/AgCl reference electrode was fabricated
by winding 125-µm-thick silver wire and then converting
it to AgCl via galvanometry in a stirred 0.1 M hydrochloric
acid solution. Auxiliary (counter) electrodes were made
by winding a 50-µm-thick platinum wire on the nylon
line similar to that of the working electrode. Next, a film
of poly(o-phenylenediamine) was electropolymerized on
the Pt working electrode from a 5 mM monomer solution
in acetate buffer.25 While permitting the passage of H2O2
toward the working electrode, this film blocks other
physiological substances, such as ascorbic acid, uric acid,
and acetaminophen, which are likely to interfere with the
operating potential of the sensor.8 The sensing enzyme is
immobilized on the Pt wire by dipping it in a solution of
19.5 mg/ml glucose oxidase and then cross-linking using an
acetate-buffered solution of 25% (w/v) glutaraldehyde and
73.2 mg/ml BSA.
The working electrode was then covered with a
semipermeable membrane, which was grown via LBL
assembly of HAs and Fe3+ ions. This was accomplished by
dip coating the sensor alternately in solutions of HAs and
ferric chloride hexahydrate. A 1-mg/ml solution of HAs was
prepared in deionized water, and the pH was adjusted to 5
using 0.1 M HCl. Care was taken not to assemble the LBL
membrane on the nearby counter and reference electrodes.
Prior studies have indicated that the growth characteristics
of HA-based films are strongly dependent upon the pH
of the solution.19 The acidic functional groups in HAs are
ionized at a pKa of 4–5, and the catecholic functional groups
are ionized at a pKa of 8–11.26 However, at pH values above
5, the HAs attain a planar conformation on the substrate
Materials and Methods
Materials
Glucose oxidase enzyme (GOX) (E.C. 1.1.3.4, 157,500 units/g,
Aspergillus niger), glutaraldehyde [25% (w/v) solution in
water], bovine serum albumin (BSA), and phosphatebuffered saline (PBS) were purchased from Sigma and
used without any further treatment. Humic acids, sodium
salt (Lot No. 11909LR; molecular mass 169 kDa), ferric
chloride hexahydrate (ACS reagent), D-glucose (ACS
reagent), PSS (MW 70,000), and PDDA [20% (w/v) in water;
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
195
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
as opposed to a much more compact aggregation at lower
pH. Consequently, at higher pH, the film growth rates per
dip cycle are considerably slower, thus making a pH of 5
optimum for film assembly. Similarly, a 5-mg/ml solution
of ferric chloride was prepared in deionized water and
used without any further modification of pH. The sensors
were immersed in each solution for 15 minutes and were
separated by a wash step in deionized water to remove any
weakly bonded species. This process, which is termed a dip
cycle, was repeated to assemble bilayers of HAs/Fe3+ into a
membrane and to yield a biosensor (device A). Devices with
2 bilayers of HAs/Fe3+ are henceforth referred to as “device
A2,” whereas those with 5 and 10 bilayers of the membrane
are referred as “device A5” and “device A10,” respectively.
and pathogenic glucose concentrations ranging from
hypoglycemia to hyperglycemia.
Diffusion Studies of Glucose and H 2 O2
Through LBL Assembled Membranes
In order to better understand the influence of the
microstructure of these LBL grown semipermeable
membranes, the diffusion of glucose and H2O2 was
investigated. For this purpose, a dialysis chamber was
built to determine the permeability of these two analytes
through HAs/Fe3+, HAs/PDDA, and PSS/PDDA bilayer
membranes assembled on top of glass fiber substrates with
an average porosity of 1 µm. For this we assume that the
porous glass fiber membrane does not contribute to the
resistance and that the flux of the analyte species is through
the LBL assembled membranes, which block the pores of
the substrate. A diffusion cell equipped with these porous
substrates was filled with 2 ml of 1.0 M analyte solution
(glucose or H2O2) and immersed in an outer chamber
containing 200 ml of 0.1 M PBS solution maintained at 37ºC.
The entire cell was stirred at a fixed rate of 190 rpm. The
outer diffusion of H2O2 was determined amperometrically
on a Pt-working electrode maintained at 0.7 V with respect
to an Ag/AgCl reference electrode while a carbon cloth was
serving as a counter electrode. In the case of glucose as
the diffusing analyte, the GOX enzyme in glutaraldehyde
solution was added to the outer PBS system so as to generate
H2O2. As the analyte diffuses out from the chamber,
the amperometric current was measured at fixed time
intervals. The current measured over time was converted
to a concentration based on a calibration experiment using
the same experimental conditions for known amounts
of analytes. The apparent diffusion coefficient (Dp) of the
analyte species was calculated based on Equation (1):28-30
Similarly, devices were also prepared by assembling
different semipermeable membranes post enzyme
immobilization. One variation was the assembly of
HAs/PDDA bilayers (device B), where the sensors were
dipped alternately in a 1-mg/ml solution of HAs and a
3-mg/ml solution of PDDA in deionized water, respectively.
The pH of HAs was adjusted to 5 as described earlier, whereas
the PDDA solution was used without further modification.
The devices were dipped in each solution for 15 minutes,
separated by a wash step in deionized water. Also, in another
variation, device C had PSS/PDDA bilayers self-assembled
on top of the enzyme-immobilized sensor by dipping the
device in a 2-mg/ml solution of the sodium salt of PSS and a
3-mg/ml solution of PDDA in deionized water, respectively.
The solutions were used without any modification to their
pH, and the devices were immersed for 15 minutes in each
solution, separated by a wash step. As in the earlier case,
there was an attraction between the negatively charged ions
of PSS and the positively charged ions of PDDA, leading to
film growth.27 Five bilayers of either HAs/PDDA or PSS/
PDDA were self-assembled on the sensors to create “device
B5” and “device C5,” respectively. Such devices enable us
to investigate polycations with neutral pH and may possibly
avoid partial enzyme deactivation by the acidic Fe3+ cations.22
where NA = (∆m/A∆t) is the flux over the area (A) of the
membrane, ∆m is the mass of analyte that diffuses out of
the chamber in the time ∆t, L is the thickness of the LBL
assembled membranes (determined ellipsometrically
by assembling the corresponding LBL films onto quartz
substrates19, 31), and [CG]i and [CG] f are the concentrations of
the analyte inside and outside of the diffusion chamber,
respectively. In the calculation of Dp, we doubled the
thickness L to account for membrane growth from both
sides of the glass fiber substrate.
In Vitro Testing
Amperometry experiments were performed using an
electrochemical analyzer made by CH Instruments
(Model CHI660A). The sensors, fabricated as described,
were tested in vitro by immersing in a stirred PBS solution
(pH @ 7.4) maintained at 37ºC and under an applied potential
of 0.7 V vs the miniaturized Ag/AgCl reference electrode.
A measured amount of glucose solution in PBS was added
to the test cell and its concentration was raised every 200
seconds by adding 0.075 mM glucose. The current was plotted
as a function of glucose concentration, and a calibration
curve was generated showing the well-reported
superlinear response of the sensor over physiological
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
Equation (1)
Results and Discussion
The chief purpose of the HAs/Fe3+ bilayer semipermeable
membrane is to tune the glucose permeability by varying
196
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
the thickness (number of dip cycles) of the HAs/Fe3+
bilayers assembly. Figure 2a displays representative
responses of devices A2, A5, and A10, respectively. Each
step in the current corresponds to an increase in glucose
concentration, which in turn leads to an additional
generation of H2O2. It can be seen from the calibration curve
of current vs the glucose concentration (Figures 2b–2d)
that the signal decreases by approximately 50 and 60% for
devices A5 and A10, respectively, as compared to device A2.
As discussed previously, the decrease in the current with
increasing thickness of HAs/Fe3+ bilayers is attributed to the
reduced permeation of glucose. Moreover, all three devices
exhibit superlinear behavior over the requisite pathogenic
and physiological glucose concentrations (2–14 mM). Such
a gradual reduction in signal sensitivity at high glucose
concentrations could be attributed to (1) the passivation
of redox-active sites onto the surface of electrode, which
in the case of the Pt electrode have been determined to be
Pt(OH)2,17 and (2) the lack of O2 in the GOX enzyme layer
when compared to the increasing glucose concentration.15
For a given bulk glucose concentration, the presence of
a diffusion-limiting HAs/Fe3+ membrane increases the
tortuosity of the path of the glucose molecules toward
the sensing enzyme, thus reducing the amount of H2O2
being generated and subsequently sensed. A similar
Fickian behavior toward glucose diffusion through
HAs/Fe3+ bilayers was observed in previous studies.19
A subsequent increase or decrease in bulk glucose sets up a
new concentration gradient across the membrane and thus
brings about a corresponding current change, assuming
that H2O2 also diffuses out through such a semipermeable
membrane (Figure 1). This enables the device to function
over a range of glucose concentrations while allowing a
precise control over its sensitivity by varying the thickness
of HAs/Fe3+ bilayers.
Figure 3: (a–c) Saturation amperometric current vs glucose concentration for
sensors with semipermeable membranes containing five bilayers of HAs/Fe3+,
HAs/PDDA, and PSS/PDDA. (d) Curves a–c are plotted on the same abscissa.
Figure 3 illustrates a typical comparison of response current
from devices B5 and C5 with A5 when tested in an identical
fashion. Although these three devices appear to show a
similar superlinear response (Figures 3a–3c), the current
change at high glucose concentrations (above 12 mM)
Figure 2: (a) Continuous polarization amperometric response during
0.075 mM glucose addition to 50 ml PBS every 200 seconds for devices
coated with a semipermeable membrane of 2, 5, and 10 bilayers of HAs/Fe3+.
(b–d) Saturation amperometric current vs glucose concentration for the
respective three devices.
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
197
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
is greater for HAs/Fe3+ as compared to the other two.
In particular, the ratio of the changes of amperometric
currents between 12 and 18 mM glucose (∆I (12–18 mM))
and between 2 and 18 mM glucose (∆I(2–18 mM)), expressed
as (∆I(12–18 mM)/∆I(2–18 mM)) for the three devices, decreases from
0.133 for A5 to 0.048 and 0.032 for B5 and C5, respectively.
Since the plateau is attributed to either H2O2 catalytic site
deactivation17 or O 2 imbalance with respect to glucose
(at higher glucose concentrations),15 the trends in the
∆I(12–18 mM)/∆I(2–18 mM) ratios should be traced back to these
two effects. Because an excessive amount of H2O2 could be
responsible for such saturation behavior at a high glucose
concentration,17 this ratio provides an initial indication of
an excessive amount of H2O2 being present in devices B5
and C5 as opposed to device A5 in accordance with their
higher amperometric signal. Lower permeability of O2
through these membranes could also lead to decreasing
∆I(12–18 mM)/∆I(2–18 mM) ratios in conjunction with an internal
buildup of H2O2.
Table 1
Diffusion Coefficients of Glucose and Peroxide Through Five Bilayer
Membranes of HAs/Fe3+, HAs/PDDA, and PSS/PDDA
Bilayer system
HAs/Fe
3+
Glucose
Peroxide
4.39 x 10
4.8 x 10-9
-9
HAs/PDDA
8.31 x 10-10
1.17 x 10-9
PSS/PDDA
5.85 x 10-11
1.01 x 10-10
Taking into account that H2O2 levels are directly proportional
to the current values determined in the sensors as shown in
Figure 3, it can be safely concluded that the highest response
of device C5 originates from the poor outer diffusion
of H2O2 from the inner GOX layer, as shown in Figure 1.
Similarly, the highest outer diffusion of H2O2 in device A5
is responsible for the lowest current response. Since the
inner diffusion of glucose is responsible for the production
of H2O2, the ratio of the inner diffusion of glucose to the
outer diffusion of H2O2 (Dp(glucose)/Dp(H2O2)) could provide a
rough qualitative measure of the expected response of these
devices. These ratios are 0.92, 0.71, and 0.58 for HAs/Fe3+,
HAs/PDDA, and PSS/PDDA, respectively. While the
gradual descending values for these diffusion coefficient
ratios are in qualitative agreement with the response of
these three devices, the enzyme kinetics responsible for the
production of H2O2 should also be accounted for. For this,
the relative inner diffusion of O2 with respect to glucose,
as well as the outer diffusion of gluconic acid and H2O2,
must be considered in order to obtain a more refined
analysis of sensor operation. This will be the subject of
future research.
In order to further elucidate the relative interdependence
of the diffusion of electrochemical participating species
through these semipermeable membranes, diffusion studies
were performed for glucose and H2O2 through these three LBL
assembled semipermeable membranes. The close molecular
size of O2 and H2O2 has been reported to result in diffusion
coefficients of similar or slightly higher magnitude than
that of H2O2. As described earlier, a chamber equipped
with a semipermeable membrane (composed of a glass fiber
support on top of which the respective semipermeable LBL
films were assembled) was employed, and the concentration
of analyte (i.e., glucose and H2O2) that diffused through
the membrane was monitored over time. As reported in
Table 1, diffusion coefficients (Dp) of glucose and H2O2
for five bilayers (assembled on both sides of the glass
fiber substrate) were obtained using Equation (1) for all
three LBL assembled systems. The diffusion of glucose is
nearly 5 and 75 times greater through the HAs/Fe3+ bilayer
membrane as compared to HAs/PDDA and PSS/PDDA,
respectively. Similarly, the diffusion of H2O2 was 4 and 48
times greater through the HAs/Fe3+ bilayer membrane as
compared to HAs/PDDA and PSS/PDDA, respectively.
As expected, the Dp of glucose is always smaller than that
of H2O2 for all membranes, consistent with the smaller
molecular size of H2O2 as compared to glucose. The higher
diffusion of these analytes through the HAs/Fe3+ membranes
has been attributed to the globular microstructure of HAs.19
However, PSS/PDDA assemblies have been shown to result
in a much more compact assembly, leaving fewer voids for
the permeation of analytes,32 which is consistent with the
results shown in Table 1.
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
Diffusion coefficient (cm2/s)
These coatings were studied further by subjecting the
various devices to reversible testing and checking for
possible hysterisis during changes in glucose concentration.
In this experiment, the concentration of glucose in the test
cell was cycled from a low to a high glucose concentration
and back to a low concentration. The resulting representative
response is plotted in Figure 4. An equilibration time
of 100 seconds was allowed between successive points.
As can be observed, when the glucose concentration was
decreased, device A5 showed minimal hysterisis as opposed
to devices B5 and C5. The percentage hysterisis at 5.6 mM
glucose concentration in the response current (difference in
current for a particular concentration during the forward
and backward cycle) is 1% for device A5 compared to 6
and 9% for devices B5 and C5. The significant hysterisis for
the two latter devices is attributed to the aforementioned
ratios of (Dp(glucose)/Dp(H2O2)), which are 0.92, 0.71, and 0.58 for
HAs/Fe3+, HAs/PDDA, and PSS/PDDA, respectively. As this
ratio deviates from unity, the amount of outer diffusion of
198
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
H2O2 outweighs the inner diffusion of glucose to create this
hysterisis response. An increase in the equilibration time
from 100 seconds to higher values is expected to readily
eliminate the 1% hysterisis present in device A5, although
the same cannot be said for devices B5 and C5, respectively.
Acknowledgements:
Financial support for this study was obtained through Department
of Defense, U.S. Army Medical Research Grants (DAMD17-02-1-0713,
W81XWH-04-1-0779, and W81XWH-05-1-0539).
References:
1. Reach G, Wilson GS. Can continuous glucose monitoring be used for
the treatment of diabetes. Anal Chem. 1992;64(6):381A-386A.
2. Turner AP, Chen B, Piletsky S. In vitro diagnostics in diabetes: meeting
the challenge. Clin Chem. 1999;45(9):1596-601.
3. Kerner W. Implantable glucose sensors: present status and future
developments. Exp Clin Endocrinol Diabetes. 2001;109 Suppl 2:S341-6.
4. Wang J. Glucose biosensors: 40 years of advances and challenges.
Electroanalysis. 2001;13(12):983-8.
5. Clark LC, Lyons C. Electrode systems for continuous monitoring in
cardiovascular surgery. Ann NY Acad Sci.1962;102:29-45.
6. Shichiri M, Kawamori R,Yamasaki Y, Hakui N, Abe H. Wearable
artificial endocrine pancreas with needle-type glucose sensor. Lancet.
1982 Nov 20;2(8308):1129-31.
7. Bindra DS, Zhang Y, Wilson GS, Sternberg R, Thévenot DR, Moatti D,
Reach G. Design and in vitro studies of a needle-type glucose sensor
for subcutaneous monitoring. Anal Chem. 1991 Sep 1;63(17):1692-6.
8. Moussy F, Harrison DJ, O’Brien DW, Rajotte RV. Performance of
subcutaneously implanted needle-type glucose sensors employing a
novel trilayer coating. Anal Chem. 1993 Aug 1;65(15):2072-7.
9. Ward WK, Jansen LB, Anderson E, Reach G, Klein JC, Wilson GS:
A new amperometric glucose microsensor: in vitro and short-term in
vivo evaluation. Biosens Bioelectron. 2002 Mar;17(3):181-9.
Figure 4: (a–c) Hysterisis analysis of saturation amperometric current
vs glucose concentration for sensors with semipermeable membranes
containing five bilayers of HAs/Fe3+, HAs/PDDA, and PSS/PDDA.
The glucose concentration was increased in PBS (from 2 to 11 mM)
and subsequently decreased (from 11 to 4 mM), with an equilibration
time of 100 seconds at each concentration.
10. Gerritsen M, Jansen JA, Lutterman JA. Performance of subcutaneously
implanted glucose sensors for continuous monitoring. Neth J Med.
1999;54(4):167-79.
11. Mercado RC, Moussy F. In vitro and in vivo mineralization of Nafion
membrane used for implantable glucose sensors. Biosens Bioelectron.
1998 Feb 1;13(2):133-45.
12. Wilkins E, Atanasov P. Glucose monitoring: state of the art and future
possibilities. Med Eng Phys. 1996 Jun;18(4):273-88.
Conclusions
13. Linke B, Kiwit M, Thomas K, Krahwinkel M, Kerner W. Prevention
of the decrease in sensitivity of an amperometric glucose sensor in
undiluted human serum. Clin Chem. 1999 Feb;45(2):283-5.
We have fabricated miniaturized electrochemical sensors
and have successfully demonstrated the growth of a
calcification-free HAs/Fe3+ bilayer membrane as a diffusionlimiting barrier for sampling glucose in the bulk. By varying
the thickness of the coatings, we can effectively tune the
current signal and thus the sensitivity of the device. The
HAs/Fe3+ semipermeable membrane was contrasted with
HAs/PDDA and PSS/PDDA membranes of the same
number of bilayers and was found to exhibit (1) lower
current saturation at hyperglycemic glucose concentrations
and (2) minimal hysterisis in current response in cycling
from low to high to low glucose concentrations. The
superior behavior of HAs/Fe3+ with respect to HAs/PDDA
and PSS/PDDA devices was attributed to a higher and
more balanced inner diffusion of glucose versus the outer
diffusion of H2O2.
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
14. Wisniewski N, Moussy F, Reichert WM. Characterization of
implantable biosensor membrane biofouling. Fresenius J Anal Chem.
2000 Mar-Apr;366(6-7):611-21.
15. Heller A. Implanted electrochemical glucose sensors for the
management of diabetes. Annu Rev Biomed Eng. 1999;1:153-75.
16. Valdes TI, Moussy F. In vitro and in vivo degradation of glucose
oxidase enzyme used for an implantable glucose biosensor. Diabetes
Technol Ther. 2000 Autumn;2(3):367-76.
17. Hall SB, Khudaish EA, Hart AL. Electrochemical oxidation of
hydrogen peroxide at platinum electrodes. Part 1. An adsorptioncontrolled mechanism. Electrochim Acta. 1998;43(5-6):579-88.
18. Galeska I, Chattopadhyay D, Moussy F, Papadimitrakopoulos F.
Calcification-resistant Nafion/Fe3+ assemblies for implantable
biosensors. Biomacromolecules. 2000;1(2):202-7.
19. Galeska I, Hickey T, Moussy F, Kreutzer D, Papadimitrakopoulos F.
Characterization and biocompatibility studies of novel humic acids
based films as membrane material for an implantable glucose sensor.
Biomacromolecules. 2001;2(4):1249-55.
199
www.journalofdst.org
Layer-by-Layer Assembled Semipermeable Membrane
for Amperometric Glucose Sensors
Tipnis
20. Bhardwaj U, Sura R, Papadimitrakopoulos F, Burgess DJ. Controlling
Acute Inflammation with Fast Releasing Dexamethasone-PLGA
Microsphere/PVA Hydrogel Composites for Implantable Devices. J
Diabetes Sci Technol. 2007,1(1):8-17.
21. Harrison DJ, Turner RF, Baltes HP.
Characterization of
perfluorosulfonic acid polymer coated enzyme electrodes and a
miniaturized integrated potentiostat for glucose analysis in whole
blood. Anal Chem. 1988 Oct 1;60(19):2002-7.
22. Valdes TI, Moussy F. A ferric chloride pre-treatment to prevent
calcification of Nafion membrane used for implantable biosensors.
Biosens Bioelectron. 1999 Jun 30;14(6):579-85.
23. Ulman A. In: An introduction to ultrathin organic films from
Langmuir-Blodgett to self-assembly. San Diego: Academic Press;1991.
24. Galeska I, Chattopadhyay D, Papadimitrakopoulos F. Application of
polyanion/Fe3+ multilayered membranes in prevention of biosensor
mineralization. J Macromol Sci Pure Appl Chem. 2002;A39(10):120722.
25. Malitesta C, Palmisano F, Torsi L, Zambonin PG. Glucose fastresponse amperometric sensor based on glucose oxidase immobilized in
an electropolymerized poly(o-phenylenediamine) film. Anal Chem.
1990 Dec 15;62(24):2735-40.
26. March J.: In: Series in advanced chemistry. 2nd ed. New York:
McGraw-Hill; 1977.
27. Hammond PT, Whitesides GM. Formation of polymer microstructures
by selective deposition of polyion multilayers using patterned selfassembled monolayers as template. Macromolecules. 1995;28:7569-71.
28. Heitner-Wirguin C. Infra-red spectra of perfluorinated cationexchanged membranes. Polymer. 1979;20:371-4.
29. Lopez M, Kipling B, Yeager HL. Exchange rates and water content
of a cation exchange membrane in aprotic solvents. Anal Chem.
1976;48:1120-22.
30. Zawodzinski Jr., TA, Springer TE, Davey J, Jestel R. Lopez C,
Valerio J, Gottesfeld S. Comparative study of water uptake by and
transport through ionomeric fuel cell membranes. J Electrochem Soc.
1993;140(7):1981-1985
31. Tompkins HG. In: A user’s guide to ellipsometry. San Diego: Academic
Press; 1993.
32. Mendelsohn JD, Yang SY, Hiller J, Hochbaum AI, Rubner MF. Rational
design of cytophilic and cytophobic polyelectrolyte multilayer thin
films. Biomacromolecules. 2003 Jan-Feb;4(1):96-106.
J Diabetes Sci Technol Vol 1, Issue 2, March 2007
200
www.journalofdst.org