Biosensors: Blockbuster or Bomb?
Electrochemical Biosensors for Diabetes Monitoring
by Lance S. Kuhn
onvenience is an expectation of patient. In 1987, MediSense marketed
life these days, from bank the Exactech glucose sensor and,
machines to cellular phones to although it never gained a majority
fast food. The medical world is share of the market, it did generate a
not immune to this drive. There movement toward electrochemical sen-
is a continual effort to push sors within the medical diagnostics
testing to the lowest level of profes- community. In 1989, Eli Lilly began to
sional skill possible, in order to allow market the Direct 30/30, a reusable
patients easy access to the information biosensor that promised to revolu-
they need to live their lives more fully. tionize the home glucose monitoring
The ultimate in this battle is to make market. Unfortunately, the user inter-
systems that patients can use by them- face was not robust enough for the
selves, at home. market, and this system was unsuc-
Electrochemical sensors have been a cessful. Others have now followed,
significant part of the move toward learning from the first systems, and
convenience and ease of use. For many three of the four largest self-blood glu-
years, sensors have helped reduce labor- cose monitor (SBGM) makers have sig-
intensive tests to simple one-step nificant electrochemical sensor-based
analysis, or even to hand-held home- systems.
use devices. More recently, I-Stat introduced a
The earliest tests to become sensor- hand-held system that has cartridges of
based were potentiometric ion-selective up to six clinical tests at a time, with a
electrodes (ISEs), such as for pH, K+, total of 18 tests, all with electrochemical
Na+, and Cl- and gas-sensing electrodes, sensors. Wampole introduced a hand-
such as for O2 and CO2. The use of held instrument that measures hemat-
these sensors is now standard proce- ocrit by conductance. Next-generation
dure. Later electrochemical sensors sensor-based systems are beginning to
included conductance tests for hemat- emerge, including systems from Via
ocrit (red blood cell volume), and then Medical and TheraSense.
enzyme-based methods. Based on this history, one might
FIG. 1. The Accu-Chek™ Comfort Curve™
In 1956, Leland Clark stimulated the think that electrochemical sensors
biosensor test strip in the Accu-Chek™
electrochemical biosensor endeavor have taken the world by storm. Complete™ test device.
when he described a method for Although there have been isolated suc-
making oxygen sensors that could be cesses, the market has not grown to
The complications of battling this
combined with enzymatic systems (the the levels predicted ten years ago. Even
life-changing disease are numerous:
“bio-” part) to measure a whole new in the glucose test market, easily the
adults with diabetes have heart disease
array of analytes.1 The test he started largest for electrochemical sensors, it
death rates two to four times those
with, and the one that is still the most took many years to accept them as a
without diabetes; 60-65% of diabetics
important enzymatic sensor in the standard. Why is this true?
have high blood pressure; end-stage
market, is for glucose. His method was
renal (kidney) disease is common
patented in 1965, and applied to the The Problem among the diabetic population;
Yellow Springs Instrument analyzer first
retinopathy causes loss of sight; and the
sold in the mid-1970s. In this test, glu- Diabetes is a world-wide public
list goes on.
cose oxidase oxidizes glucose in the health problem. When the body no
Much of the burden of this disease
presence of oxygen, turning the oxygen longer produces insulin, or has devel-
can be reduced or eliminated by early
over as peroxide, which is then oxidized oped a tolerance to it, and does not
detection and improved self-care. The
at the working electrode of an electro- properly convert glucose into energy,
Diabetes Control and Complications
chemical cell. Following Clark’s lead, diabetes is the result. Approximately 16
trial3 (DCCT), a 10-year nationwide
others have taken this technology and million Americans have diabetes; one
study of 1,441 diabetics, conclusively
applied it to lactate, creatinine, choles- third or more have yet to be diagnosed.
demonstrated that good control of
terol, and other analytes of medical Worldwide, these numbers are even
blood sugar delayed or prevented many
importance. more staggering. A 1994 World Health
of these complications at rates of at
It took more than a decade longer Organization report estimated that
least 50% better than poorly-controlled
before such a system was reduced to a there are at least 110 million diabetics,
subjects. Important to this good control
hand-held instrument, reaching the and this number is expected to more
is frequent, consistent, and accurate
ultimate level of convenience for the than double in the next 30 years.2
self-testing of blood glucose.
26 The Electrochemical Society Interface • Winter 1998
Finally, the market has matured to a
point where it is difficult for small
players to compete, and necessary for
the very large to be highly cost-con-
We will describe our efforts to
develop a marketable biosensor,
shown in Fig. 1, that is state-of-the-
art, novel, inexpensive, and high-per-
forming. A large part of our success is
due to our ability to use two identical
electrodes, and to allow the chemistry
of the strip to control our glucose
measurement. Using this biampero-
metric approach simplifies the cell
requirements for the sensor. The
FIG. 2. An exploded view of the Accu-Chek™ Comfort Curve™ biosensor test strip. product’s life cycle will be followed to
show how biosensors can make a dif-
ference in medical diagnostics. We will
also peer briefly into the future of
biosensors in this and other medical
markets. Many excellent reviews delve
deeper into biosensor technology; one
recent paper broadly reviews the
There are several keys to making a
competitive biosensor for the medical
devices marketplace. Since this is a
near-commodity market, cost (to the
FIG. 3. The reaction sequence for glucose measurement on the Accu-Chek Advantage sensor.
manufacturer) and price (to the con-
sumer) of the individual sensor are
Hurdles convoluted with specificity, in that
major issues. As these devices are med-
many other compounds that might
ical devices, which are used to diagnose
The first hurdle any test in medical produce a competing signal are at
potentially life-threatening incidents
diagnostics faces is the need for speci- similar, or even higher, concentra-
every day, they must be of very high
ficity; it is no different for biosensors. tions. Fortunately, this is not the case
quality, and the information displayed
A test is only as good as its ability to for glucose, which is at millimolar
to the user must be accurate. The sen-
separate the signal due to the analyte concentrations in blood.
sors must be easily manipulated by
of interest from another signal. In
sight-impaired users, and the system
medical tests this is typically done Market View
must be very user-friendly to encourage
with a biological specifier, such as an
more frequent testing for better con-
enzyme or an antibody. For glucose The slow acceptance of electro-
trol. In today’s glucose testing market,
this specifier was defined long ago to chemical sensors in the SBGM market
features differentiate products in the
be the enzyme glucose oxidase; more was probably due to several effects
marketplace; the ability of the system
recently, glucose dehydrogenase has unrelated to detection of the analyte
to interface with the physician’s work,
been used successfully. molecule. First, the market for glucose
and software that allows users to track
The use of a biological molecule as biosensors, i.e., the diabetic popula-
their results and regimen changes, are
a specifier brings an inherent second tion and their physicians, has changed
of growing importance.
hurdle with it: instability. Large bio- drastically in the past 15 years. The
In our sensor, shown in an
logical molecules are typically not hand-held optical instruments first
exploded view in Fig. 2, two identical
stable outside the environment for introduced 15 years ago have evolved
electrodes are used for amperometric
which they were designed. The use of into very sophisticated instruments,
detection, with no need for either a ref-
them in a test, especially in a test capable of accurate and precise read-
erence or a counter electrode of dif-
where they must be dried and stored ings with very little effort from the
ferent size or material. The sensor
for months or years, then used in patient. Any new tests, biosensor or
reagent includes three active ingredi-
extreme environments of temperature otherwise, must meet or exceed these
ents and several support ingredients.
and humidity, requires considerable performance standards, which are
The active ingredients in most sensors
work, and often is not successful at all. based upon a calorimetric chemistry
of this type are a glucose-specific agent,
A third common hurdle for med- that has been under development for
an electron shuttle or mediator, and
ical tests is sensitivity, as many mole- decades. Secondly, the manufacture of
stabilizing agents for the glucose speci-
cules of interest in the body are at the electrochemical strips with the
fier to ensure long shelf life. The glu-
concentration of 10-6 M and below. required electrode tolerances has
cose specifier has typically been glucose
This is not only a difficulty in proven to be both more difficult and
oxidase or glucose dehydrogenase,
absolute sensitivity, however, but is more expensive than expected.
The Electrochemical Society Interface • Winter 1998 27
although the hexokinase reaction has
also been used. The electron shuttle is
often a ferrocene or ferricyanide deriva-
tive. In our case, glucose oxidase was
used in the original Accu-Chek™
Advantage® product in 1994. Then,
following a continuous improvement
path, a glucose dehydrogenase-based
sensor was introduced two years later.
Potassium ferricyanide has been the
mediator in all cases.
The mechanism of action of the
sensor is shown in Fig. 3. The glucose
in a sample reacts with glucose oxidase
to make gluconic acid and the reduced
form of glucose oxidase. The reduced
glucose oxidase then reacts with ferri-
cyanide to make ferrocyanide. The
working electrode, poised at a potential
positive of the rest potential of the
mediator, oxidizes all ferrocyanide as it
diffuses to the electrode. This generates
a current directly proportional to the
concentration of glucose in the solu- FIG. 4. A comparison of the effect of oxygen on the glucose test result, for a glucose-oxidase-based system (red
tion. It also allows for a catalytic cycle regression lines) versus a glucose dehydrogenase-based system (blue lines). Three levels of oxygen are shown: a
standard venous level of 21.7 mm Hg (red circles), a standard capillary level of 94.3 mm Hg (green squares), and
of regeneration of the two key reagent
a hyper-oxygenated sample at 182.8 mm Hg (yellow triangles). The five glucose levels were tested at n=12, and
constituents. error bars represent two standard deviations from the mean.
Sensor reagent improvements high-
light a key to the marketability of this in samples must be described in the chemical cell does not necessarily have
type of product. In a glucose oxidase- product literature, and claims must be to be complex.
based test, the first interferent in the approved by the Food and Drug Admin- In a typical electrochemistry system,
system is oxygen, which is the natural istration (FDA) for use in such settings. a reference electrode is used to stan-
substrate for the enzyme. We are This has led to few systems being dardize the potential of the working
attempting to replace oxygen with approved for use for all samples. electrode to a known value. This allows
potassium ferricyanide as an electron The route to an improved sensor one to know, very accurately, at what
acceptor. The ferricyanide is not as effi- that can be used in all medical situa- potential the working electrode is
cient at shuttling electrons with the tions and, more importantly, gives poised, and to assure that that potential
enzyme as oxygen; therefore, any unvarying results in these wide-ranging is both sufficient for linear current gen-
oxygen in the solution can compete circumstances was to replace the glu- eration and low enough to avoid as
effectively for the enzyme site, pro- cose oxidase in the strip with glucose many biological oxidations as possible.
ducing a signal that is related to glu- dehydrogenase. Glucose dehydroge- But it complicates the manufacture of
cose, but not in the same way as the nase does not interact with oxygen, the disposable strip, as it takes extra
ferricyanide-mediated signal. Oxygen and therefore is unaffected by variable steps to create a reference electrode. In
can give a positive bias in such a oxygen concentrations in the sample. the case of the Advantage® strip, we
system, meaning measurements are Fig. 4 shows the reduction in oxygen have eliminated many of these difficul-
skewed higher, limiting the accuracy at dependence afforded by such a change. ties by using two noble metal elec-
low glucose values. Note that there is a > 10% difference in trodes that are exactly the same.
Additionally, biological specimens calibration slope for venous versus cap- A depiction of the Accu-Chek™ Com-
contain widely varying oxygen levels. illary data in the oxidase system. In the fort Curve™ strip is shown in Fig. 2. Note
The oxygen partial pressure, pO2, (a dehydrogenase system, the calibration the simplified architecture with the
measure of dissolved oxygen content) of slopes are essentially superimposable. similar electrode configuration. Essen-
an average venous blood sample is Similar results have been shown in a tially, two identical palladium elec-
about 40 mm Hg. This corresponds to a clinical setting.5 trodes are sealed between two thin
dissolved oxygen level of approximately Other concerns for a successful busi- sheets of plastic, with cut-outs for elec-
0.06 mmoles/L. For an arterial sample, ness venture in biosensors are cost and trode contact and chemistry/sample
one would expect much larger oxygen complexity of manufacturing. Biosen- interface. Palladium was used in this
levels, with pO2 reaching as high as 110 sors, by definition, include biological sensor because of its relatively low cost
mm Hg, or 0.15 mmoles/L dissolved components; and as they must make for sputtered films (at that time, much
oxygen. Capillary samples typically an accurate measurement in some of lower than for gold) and its resistance
have oxygen levels a little lower than the worst samples, there is a tendency to surface oxidation. A small volume (6
arterial samples. Therefore, if one were to create a rather complex system that µL in the configuration shown) of the
to make measurements of glucose in the becomes too difficult to manufacture. sensor reagent is dispensed and dried in
three different sample matrices, a glu- To date, the sensor reagent mixture the open window of the strip. Data is
cose oxidase-based sensor could give remains complex, for both biological collected using chronoamperometry;
three different results. Due to the seri- (enzyme degradation) and manufac- after a short reaction period, a potential
ousness of this problem, any differences turability reasons. However, the electro- difference is applied between the elec-
28 The Electrochemical Society Interface • Winter 1998
In the case of our sensor, the two
electrodes of the electrochemical cell
are identical. One way we can create a
difference is to apply a potential differ-
ence between the two, changing the
energy levels such that current must
flow. If that potential difference is large
enough (150 mV), diffusion-limited
(Cottrellian) current can occur at each
The limiting current will be
obtained at the electrode at which the
limiting reaction is occurring; for
instance, if [ox]>>[red], the oxidation
of [red] to [ox] will be the limiting reac-
tion (and in this case, the one we wish
to measure at the working electrode).
The two electrodes must quickly come
to agreement, with a separation of 150
mV and with currents equal in magni-
tude. Because the working electrode
FIG. 5. A standard Clark Error Grid analysis of the calibration curve assignment for an Advantage® biosensor reaction generates a small amount of
product. The lettered zones represent clinically relevant regions of performance: A: clinically accurate readings;
current, the potential of the other elec-
B: results that would lead to benign action or inaction by the user; C: results that would lead to unnecessary
corrections; D: results that would lead to inaction when action is necessary; E: results that would lead to trode (the counter electrode) will posi-
treatment opposite of what clinical accuracy would call for. tion itself just far enough negative to
give an equal cathodic current to the
anodic current measured at the
working electrode, as shown in Fig. 6.
The working electrode will then poise
itself as far positive as necessary to
maintain the 150 mV potential differ-
ence. This should always be in a region
of purely diffusion-limited current gen-
eration. According to the Nernst equa-
tion, the actual potential of the
working electrode will be more nega-
tive as the concentration of glucose
(and, therefore, ferrocyanide) is
We need not concern ourselves
with the true (referenced) potential.
Different mediators will have different
rest potentials and, therefore, actual
FIG. 6. The general principle of the electrochemical mechanism of Advantage® sensors is illustrated. The electrode potentials will vary
yellow bar shows the 150 millivolt potential window with equal oxidized and reduced mediator concentra- depending on the chemical system. A
tions (a situation we must never reach in the sensor); the purple bar shows the potential window with slightly mediator should be chosen with as low
more oxidized than reduced mediator (high glucose concentrations), and the
E° as possible, so that the oxidation of
red bar shows the situation at oxidized >> reduced (low glucose.)
extraneous compounds, e.g., uric acid
or ascorbic acid is limited. However,
trodes and current collected for a speci- applies to equilibrium, i.e., zero net cur-
regardless of the mediator used, very
fied time. Fig. 5 demonstrates a typical rent. This is the case, for instance, if
high redox indicator concentrations
calibration curve for a set of about 300 one allows a reaction to proceed to
must be used, such that [ox] is always
measurements using capillary blood equilibrium, then measures the energy
greater than [red] and the limiting cur-
samples. Precision, accuracy, and lin- level of the system, i.e., the open circuit
rent is always generated at the working
earity are excellent. potential. Such a case is considered
electrode. The use of a high concentra-
Of course, there is no free lunch, here, where we always begin with the
tion of a redox mediator allows control
and we pay for this by a slightly more same ratio of [ox]/[red] (approx. 0.05%
of the actual potentials of the identical
complicated or less well-defined elec- ferrocyanide in 99.95% ferricyanide).
working and counter electrodes by the
trochemical mechanism. Our sensor The glucose-specific reaction described
glucose reaction. As long as the initial
works through a combination of very earlier converts a fixed amount of the
mediator concentration is greater than
basic electrochemical principles, ferricyanide to ferrocyanide under a
two times the stoichiometric concen-
though in an unusual way. condition of open circuit. Thus, if we
tration of sample analyte, one main-
The Nernst equation, allow the reaction to proceed, different
tains analyte-limited currents. In the
glucose concentrations will give us dif-
case of potassium ferricyanide, the
0.059 ferent [ferricyanide]/[ferrocyanide]
E = E° +
n ( (log [ox]))
[red] ratios and, therefore, different open cir-
high concentrations of mediator do
not create a problem, since this com-
cuit solution potentials.
The Electrochemical Society Interface • Winter 1998 29
pound is highly soluble, inexpensive, tion of the meter by the sample. The we try to provide the customer with
and easy to obtain. If too low a concen- strip is large enough to be easily manip- useful information.
tration of the oxidized mediator in the ulated, has a colored sample touch pad, A look into the future shows a sig-
strip was used, the signal due to glucose and the end nearest the patient can be nificant shift in the way glucose biosen-
would peak, then diminish, even as handled without fear of touching the sors are used.6-8 The results from the
glucose increased, because the limiting sample or adulterating the test. The DCCT, and several studies since then,
reaction would become the reduction resulting device is an electrochemical clearly demonstrate that diabetics who
of the oxidized form of the mediator at sensor system with a high degree of lin- closely monitor their glucose, and act
the counter electrode. This would result earity, excellent accuracy and precision according to the test results, have better
in a very dangerous test system for the (Coefficient of Variation < 3%) across long-term outcomes. The obvious way
patient. Great care has been taken to the clinically-relevant range of glucose to make this possible for our customers
insure that this will not be the case, concentration. Of course, there are still would be to provide them with a
loading the sensor reagent with several improvements that can be made. sensing system that continually, or
times the amount of mediator neces- nearly continually, displays accurate
sary to meet even very high glucose Future Directions for glucose readings, which are generated
levels. Electrochemical Biosensors from as minimally invasive a system as
Now the steps to address the keys to possible. The “holy grail” of this work is
making a competitive biosensor for the The newest sensor in the the noninvasive systems that allow
medical device market have been taken. Advantage® line, the Accu-Chek™ testing with no finger sticking and
The cost of the sensor has been reduced Comfort Curve™ in FIGS. 1 and 2, takes nothing in the body. However, these
by eliminating one of the complicating advantage of capillary fill and an systems appear still to be far away.
features of an electrochemical sensor, improved reagent to test with even On the other hand, many groups
the reference electrode. By developing smaller sample volumes and more accu- have reported methods for minimally
the reagent according to the chemical rate results. The former is important to invasive glucose measurement on a rel-
and electrochemical principles this patients who prick their finger several atively continuous basis. Among these
simple test has been made very accurate times daily, because smaller sample vol- are many electrochemical methods.
and reproducible. The electrochemical umes necessarily mean finger punctures One, the VIA Medical probe, is a
sensor eliminates some of the draw- that are less severe. More accurate, and venous catheter system that constantly
backs of an optical test system, such as especially more robust, systems are measures glucose in the blood, and is
stray light interference and contamina- always at the top of our priority list as already on the market for hospital use.
30 The Electrochemical Society Interface • Winter 1998
A few others, further from the where the research groups of Buck and 2. D. McCarty and P. Zimmet in an Interna-
market, are worthy of note. Wilson and others have developed arrays of tiny tional Diabetes Institute report: Diabetes
1994 to 2010: Global Estimates and Projec-
coworkers at the University of Kansas electrodes that monitor heart electrical
tions, ISBN 0646202448 (1994).
and in Europe, have been pursuing the activity and important clinical parame- 3. The Diabetes Control and Complications
implantable glucose sensor for many ters.16 Another use of microfabrication Trial Research Group, The New England
years.9-11 Their sensor uses glucose oxi- showed 400 individually-addressable Journal of Medicine, 329, 977 (1993).
dase immobilized at the surface of a microelectrodes on a single 1 cm2 chip, 4. J. L. Anderson et al., Anal. Chem., 70, 519R
working/reference electrode combina- allowing spatial resolution of analyte (1998).
5. G. J. Kost, H.-T. Vu, J. H. Lee, P. Bourgeois,
tion. Here, no mediator is employed, distribution in a small area.17
C. Martin, S. S. Miller, A. O. Okorodudu, J. J.
and the sensor relies on the generation These are only a few of the vast Podczasy, R. Webster, and K. J. Whitlow,
of peroxide from the native oxygen array of research efforts currently Crit. Care Med., 26, 581 (1998).
reaction; the peroxide is then oxidized exploring biomedical, especially glu- 6. G. Spera, Medical Device and Diagnostic
at the working electrode. Polymer layers cose, sensors. All these efforts will Industry, November 1997, 44.
surround the chemistry to protect it eventually face the same issues as those 7. C. G. Masi, R&D Magazine, July 1996, 23.
8. M. Vreeke, IVD Technology, July 8, 1997, 39.
from the body, and the body from the we have encountered with first genera-
9. D. Moatti-Sirat, V. Poitout, V. Thome, M. N.
chemistry. This sensor is used subcuta- tion sensors. In addition, they will face Gangnerau, Y. Zhang, Y. Hu, G. S. Wilson, F.
neously to measure glucose on a con- the difficult tasks of making a measure- Lemonnier, J. C. Klein, and G. Reach, Dia-
tinuous basis. Similar sensors have been ment in the very harsh environment of betologia, 35, 224 (1992).
reported by other research groups, some the body, and making these tests very 10. V. Poitout, D. Moatti-Sirat, G. Reach, Y.
relying on the direct measurement of stable and reproducible. Their success Zhang, G. S. Wilson, F. Lemonnier, and J. C.
Klein, Diabetologia, 36, 658 (1993).
oxygen consumption, and others on in passing these tests will bring signifi-
11. V. Thome-Duret, G. Reach, M. N. Gang-
the oxidation of the peroxide produced cant changes in near-patient testing of nerau, F. Lemonnier, J. C. Klein, Y. Zhang,
in the enzymatic reaction. many medical and biological com- Y. Hu, and G. S. Wilson, Anal. Chem., 68,
MiniMed has recently described a pounds. s 3822 (1996).
similar system to the FDA in an 12. MiniMed Press Release, December 17, 1997.
attempt to gain approval for use in Acknowledgments 13. A. Heller, J. Phys. Chem., 96, 3579 (1988).
14. B. A. Gregg and A. Heller, J. Phys. Chem., 95,
hospitals and some patients.12 The
results of FDA’s review of their data will A special thank you to the Test 15. E. Csoregi, D. W. Schmidtke, and A. Heller,
have a dramatic effect on research in Development, Product Performance, Anal. Chem., 67, 1240 (1995).
many companies and research labora- Marketing Communications, and Math 16. V. V. Cosofret; M. Erdosy, T. A. johnson, R.
tories around the world. and Statistical Analysis groups at Roche P. Buck, R. B. Ash, and M. R. Newman,
Heller and associates13-15 have Diagnostics-Boehringer Mannheim Anal. Chem., 67, 1647 (1995).
17. H. Meyer, H. Drewer, B. Grundig; K. Cam-
taken a different approach. In their Corp. for their help in preparing this
mann, R. Kakerow, Y. Manoli, W. Mokwa,
sensor, a wired enzyme/mediator com- manuscript, and their hard work in and M. Rospert, Anal. Chem., 67, 1164
bination is stated to reduce oxygen developing the biosensor products in (1995).
dependency of the sensor, and to pro- the Accu-Chek™ Advantage® line.
vide a reliable result continuously. Some of this work was originally
They attach glucose oxidase or glucose presented at the 1994 Third World
dehydrogenase to a poly(vinylimida- Congress on Biosensors, in New About the Author
zole) polymer backbone, and attach a Orleans, LA, USA. Lance Kuhn is a Principal Scientist in
redox mediator (eg., Os(bpy)2Cl) to Methods Development for biosensors at
another part of the polymer. These two References Roche Diagnostics-Boehringer Corporation
molecules are then relatively free to 1. L. C. Clark, Jr., Trans. Am. Soc. Artif. Intern.
interact, exchanging electrons which Organs, 2, 41 (1956).
then travel “through” the polymer
“wire” and lead to a final signal from
the mediator at the electrode.
Even more fascinating, but also
more difficult, are electrochemical
methods to measure glucose without
the aid of biological specifiers. Such
methods would allow reduced biolog-
ical reaction to the testing device cycle,
and less-complicated sensors. Several
groups have used pulsed AC and cyclic
voltametric methods to measure the
direct oxidation of glucose at a cat-
alytic electrode surface such as plat-
inum or metal oxides.
Microfabrication techniques are
leading to the proliferation of
microsensing devices, which will lead
biomedical sensors into entirely new
fields and allow for arrays of tests on
single small devices. One example of
such research is at Duke University,
The Electrochemical Society Interface • Winter 1998 31