Biosensors and Bioelectronics 20 (2005) 2019–2026

An amperometric glucose biosensor prototype fabricated by thermal

inkjet printing
L. Settia,∗ , A. Fraleoni-Morgerac , B. Ballarinb , A. Filippinia , D. Frascaroa , C. Pianad
a Department of Industrial Chemistry and Materials, University of Bologna, Viale del Risorgimento 4, 40136 Bologna, Italy
b Department of Physical and Inorganic Chemistry, University of Bologna, Bologna, Italy
c Spinner Consortium, Bologna, Italy
d Lesepidado srl, Bologna, Italy

Received 26 July 2004; received in revised form 13 September 2004; accepted 14 September 2004

Abstract

The prototype of an amperometric glucose biosensor was realized by thermal inkjet printing using biological and electronic water-based inks,
containing a glucose oxidase (GOD) from Aspergillus niger and the conducting polymer blend poly(3,4-ethylenedioxythiophene/polystyrene
sulfonic acid) (PEDOT/PSS), respectively. The biosensor was fabricated microdepositing PEDOT/PSS and GOD, in sequence, on ITO-glass,
by a commercial inkjet printer, with the help of a commercial software. High density microdots matrices were so-realized, with a calculated
resolution of about 221 × 221 dpi (dot per inch). By means of a rapid and easy assay it was demonstrated that no activity loss occurred upon
the printing of GOD, despite of the use of a thermal printhead. The device was encapsulated in a semipermeable membrane of cellulose
acetate, applied by dip-coating, in order to prevent dissolution of the enzyme and/or PEDOT/PSS in water. The preliminary response of the
electrode was measured in an aqueous glucose solution in the presence of ferrocenemethanol (FeMeOH) as a mediator, and resulted linear up
to 60 mM in glucose. The best sensitivity value achieved was 6.43 ␮A M−1 cm−2 (447 nA M−1 U−1 cm−2 ). The characteristics of the device,
and the possible performance improvements have been analyzed and discussed. The reported findings indicate that inkjet printing could be a
viable instrument for the easy construction of a working biosensor via direct digital design using biological and conductive polymer based
inks. Such an approach may be seen as an example of “biopolytronics”.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Thermal inkjet printing; Biological ink; Electronic ink; Bioelectronics; Biopolytronics; Biosensors

1. Introduction ditional electronic transducer (Kros et al., 2001). In order to

Biosensors represent a new trend emerging in terms of
technological and theoretical achievements for the develop-
ment and exploitation of analytical devices for detection,
quantification and monitoring of specific chemical species
for clinical, environmental and industrial analysis (Albareda-
Sirvent et al., 2000).
In these devices the detection of a (bio)chemical species
is achieved by means of a biological element, through a tra-
∗ Corresponding author. Tel.: +39 051 2093672; fax: +39 051 2093673.
E-mail address: setti@ms.fci.unibo.it (L. Setti).
measure physiological parameters of complex environmen-
tal matrices in vitro and in vivo the miniaturization of these
sensors is needed.
To this aim, the ability to micropattern surfaces with active
molecules may represent a viable top–down strategy. More-
over, especially for achieving multifunctional devices, the use
of a technology able to dispense onto a solid support quan-
tities of solutions in the order of micro- and nanoliters, and
to ensure at the same time the rapid and reproducible depo-
sition of several sample spots in small areas, is required. To
obtain these results, screen printing and photolithography are
currently the most used approaches, with a preference for the

0956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.bios.2004.09.022
2020 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
first one, which holds advantages over the second one for
an envisioned large scale biosensor production, thanks to the
precision, speed and low cost of the method. This technique
takes advantage of special inks (or pastes), which are forced
through a mesh screen to form an image onto a surface. In this
way different parts of a biosensor has been already produced,
using electronical and biological pastes (Albareda-Sirvent
et al., 2000), with resolutions as high as 390 dots per inch
(dpi), and dot diameters around 40–70 ␮m.
In alternative to this method, the inkjet printing deposition
is attracting a lot of interest as a manufacturing tool. Infact, it
allows to deposit very small droplets (2–12 pL) on virtually
any surface (glass, plastic, metal, etc.), with a precision and
reproducibility potentially higher than that of screen printing,
due to the viscosity of the inks formulated for inkjet printing,
which is lower than that of the screen printing pastes. Indeed,
by this technique, resolutions higher than 1200 dpi (dots per
inch), with a mean dot diameter of about 15–40 ␮m, may
be obtained. Moreover, the use of an array of nozzles con-
nected to a device-driving electronic system (i.e. the print-
head) allows a very good control degree over the layout of
the microdeposited pattern (Mills, 1996). Finally, the absence
of physical contact between printhead and printed support
makes possible to deposit materials on contact-sensitive sub-
strates.
Inkjet printing may be performed with different ink ejec-
tion technologies, of which the most used are the piezoelectric
and the thermal one.
In piezoelectric printheads, crystalline materials undergo
mechanical stress upon application of an electrical field. A
very small contraction or expansion of these crystals, con-
fined into the nozzles, allows for the reduction of the space
available for the ink, thus increasing the pressure and causing
the ejection of a drop (Kenyon, 1996; Le, 1998).
With the thermal technology, the ink emission is obtained
thanks to a vapour bubble that is formed on the surface of a
heating resistor located inside each nozzle. During the heat-
ing, the vapour bubble grows, increasing the pressure inside
the nozzle and forcing an ink droplet through the orifice
(Ping-Hei et al., 1997).
Inkjet printing applications in biotechnology have been
already used for the development of biosensors (Hart and
Turner, 1996; Newmann et al., 1992), biochips (Petridou Xu
et al., 2004) and DNA arrays (Allain et al., 2001; Schena
et al., 1998), as well as for DNA synthesis (Blanchard et al.,
1996), microdeposition of active proteins (Roda et al., 2000;
Setti et al., 2004), free-form fabrication techniques to cre-
ate cellular polymeric scaffolds (Sherwood et al., 2002; Park
et al., 1998) and printing biological materials for cellular scaf-
folds (Roth et al., 2004). It has to be noted that up to now,
the vast majority of the published literature on inkjet printing
for the deposition of organic and biological molecules has
been focused on the piezoelectric technology (Sirringhaus
et al., 2000; Kawase et al., 2001; Chen et al., 2003), due to the
high temperature (about 300 ◦ C) reached inside the nozzles
of thermal printers, which could induce a thermal stress on
the ink, in turn damaging the biological and/or synthetic or-
ganic ink components. The piezoelectric technology is then
usually preferred because it applies no thermal load to the
ink.
Despite of these concerns, our research group has shown
recently that the thermal inkjet technology is compati-
ble with the deposition of the enzyme ␤-galactosidase
(Setti et al., 2004) and of conjugated polymers like
poly(3,4-ethylenedioxythiophene/polystyrene sulfonic acid)
(PEDOT/PSS) (Ballarin et al., 2004), causing no appre-
ciable degradation to the specific function of the organic
molecule.
In this work we show that, through the sequential depo-
sition onto an ITO-coated glass of an electronic ink contain-
ing the conductive polymer blend PEDOT/PSS and of a bi-
ological ink containing the enzyme GOD, both deposited by
thermal inkjet technology, it is possible to obtain a working
printed biosensor for the detection of glucose in an aqueous
environment. In order to preserve the PEDOT/PSS and glu-
cose oxidase from dissolving in water, the printed biosensor
was protected with a cellulose acetate membrane, applied by
dip-coating.
Preliminar chronoamperometric measurements per-
formed on a working prototype, demonstrating its activity,
are presented, and the obtained performances are analyzed
and discussed.
2. Experimental
2.1. Chemicals
Poly(3,4-ethylenedioxythiophene/polystyrene sulfonic
acid) (PEDOT/PSS) 1.3 wt.% dispersion in water, polyoxy-
ethylene-(20)-sorbitan monooleate (Tween 80), ethylene-
diaminetetraacetic acid, tetrasodium salt hydrate (EDTA)
98%, tetrahydrofuran 99%, potassium nitrate 99%, fer-
rocenemethanol 97%, Peroxidase (POD: EC 1.11.1.7.
from horseradish), 3-metil 2-benzothiazolinone hydrazone
(MBTH) were purchased from Sigma–Aldrich. Glucose,
glycerol 98%, and acetone (99.5%) were purchased from
Carlo Erba Reagents. Cellulose acetate (Mn 29.000) and gua-
iacol were purchased from Fluka and Merck, respectively.
Phosphate buffer 0.1 M (pH 6.5) was prepared according to
normal lab procedures.
2.2. Methods
2.2.1. Biological ink preparation
Glucose oxidase (GOD: EC 1.1.3.4. from Aspergillus
Niger, 154 U/mg) was purchased from Sigma–Aldrich (Ger-
many). The biological ink was obtained dissolving 0.6 or
6 mg/mL of GOD in a 0.1 M posphate buffer, pH 6.5, which
contained EDTA 1.5 mM as antimicrobial agent and 10%
(w/v) of glycerol as stabilizer.
 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
2.2.2. Electronic ink preparation
The electronic ink was prepared diluting 20 mL of the
1.3 wt.% PEDOT/PSS dispersion in distilled water, to a final
volume of 50 mL, and filtering the obtained dispersion with a
0.25 ␮m filter (cellulose acetate). In order to obtain the nec-
essary surface tension for the printing device, the dispersion
was added of 0.426 g of Tween 80 (6.50 mM solution).
2.2.3. Description of the printing system
The inkjet printer was developed in a prototype model for
printing on solid surfaces (Setti et al., 2004) by Lesepidado
srl (Bologna, Italy), starting from a commercial model ther-
mal printer supplied by Olivetti Tecnost (Ivrea, Italy). The
printhead, characterized by 208 nozzles with a roof-shooter
design, was supplied by Olivetti I-Jet (Aosta, Italy). The so-
configured system permitted to realize matrices on solid sup-
ports, with a resolution up to 1200 × 1200 dpi, in which each
dot was formed by an ejected ink drop of 10–12 pL. The car-
tridge feeding the printhead was filled with the biological or
the electronic ink using an external flask, connected to the
printhead through a plastic cannula.
2.2.4. Enzyme activity assay
The GOD activity was tested with an assay based on pre-
vious work (Setti et al., 1998), relying on oxidative coupling
between MBTH and guaiacol.
To 3 mL of phosphate buffer 0.02 M, under stirring, 0.1 mL
of GOD solution, 0.1 mL of POD 0.3 mg/mL (279 U/mL) and
0.5 mL of guaiacol 0.1 M were added, followed by 0.5 mL
of MBTH solution (7.4 mM) and 0.5 mL of glucose solu-
tion (4%, w/v). After the addition, the reaction was left for
three minutes, under stirring, at 30 ◦ C, then stopped adding
to the mixture 0.5 mL of 2N H2 SO4 solution and 1 mL of
acetone.
The so-formed red complex was analyzed by UV–vis
spectrophotometry at 505 nm with a Uvikon 923 spectropho-
tometer (Bio-teck Instruments srl, Milano).
2.2.5. Determination of the deposited enzyme
The amount of enzyme deposited by the printhead was
determined mixing the biological ink, containing 0.6 mg/mL
of GOD, with 1.2 mg/mL of Brilliant Blue FCF (E133), pur-
chased from Fiorio (Italy). The printing support was a hy-
drophobic polyester sheet on which the biological ink was
not adsorbed. The polyester sheet was put on the mobile,
flat printer tray, which motion direction was perpendicular
with respect to the one of the printing bar. The distance be-
tween the nozzle array and the support was 2–4 mm, and
the biological ink was put into an external feedstock, con-
nected to the printhead via a plastic cannula. The printer
was then connected to a standard PC and a filled rectangle
(with an area of 16 cm2 ), defined by a word-processing soft-
ware at a resolution of 600 × 600 dpi, was printed on the
substrate.
The so-deposited ink was recovered washing the printed
surface with 10 mL of 0.1 M phosphate buffer. The recovered
2021
E133, hence the amount of ejected ink, was then determined
spectrophotometrically at 628 nm.
2.2.6. Electrochemical measurements
A three electrode cell geometry was used in chronoam-
perometric experiments. The counter electrode was a Pt wire,
the reference electrode was a SCE and the ITO-glass inkjet
printed device was used as working electrode. An Auto-
lab PGSTAT20 (Ecochemie, Utrecht, The Netherlands) po-
tentiostat/galvanostat, interfaced with a personal computer,
was used. The response of the inkjet printed electrode was
measured dipping the electrode in 25 mL of a stirred buffer
solution (0.05 M phosphate buffer + 0.05 M KNO3 , pH 6.5)
in presence or in absence of ferrocenemethanol (FeMeOH)
(14 mg/L), at an applied potential of +0.30 or +0.60 V. Af-
ter that a stable current background was reached (30–60 s),
aliquots of 100 ␮L of a 1.0 M glucose solution were added,
and after 60 s from the addition the system response was mea-
sured.
The cyclic voltammetry was performed cycling the elec-
trode in a range of potential between −0.75 and +0.75 V,
with a scan rate of 50 mV s−1 in the above mentioned buffer
solution.
2.2.7. Biosensor fabrication
ITO-coated glass plates with 12 /m2 (from Technopart-
ner, Italy) were washed with chloroform, water and fi-
nally acetone in order to eliminate dirt traces from the
surface.
Two inkjet printed biosensors were realized. In the first
one, on a rectangle of 7 mm × 50 mm of ITO coated glass, a
film (2.4 cm2 ) of PEDOT/PSS was deposited at 600 × 600 dpi
by thermal inkjet printing, repeating the deposition 10 times
over the same surface.
The film thickness was evaluated by atomic force mi-
croscopy (AFM, Scanning Probe Microscope Vista 100,
Burleigh Instruments Inc.), using a long-range 100 Å tip, op-
erating in contact mode with a 10 nN force constant. The
measurement was made scratching with an edged plastic tool
the surface of the film, obtaining thus a narrow PEDOT/PSS
free surface, whose borderline with the film was included in
the AFM scan.
On the so-obtained polymeric film the biological ink con-
taining 6 mg/mL GOD was then printed, repeating the depo-
sition two times.
The second biosensor was prepared depositing the GOD
directly onto the ITO-coated glass surface.
Both biosensors were finally covered with a cellulose ac-
etate membrane by dip-coating. The dip-coating solution was
prepared adding 2.5% (w/v) of cellulose acetate (Mr 29 000),
under stirring, to a solution of THF/acetone (60/40). The sam-
ple extraction speed was 2.43 mm/s. The membrane thickness
was determined by AFM as mentioned above, taking care to
measure the ITO/membrane step in a region of the electrode
where only the membrane was deposited.
2022 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
3. Results and discussion
3.1. GOD-based biological ink
The biological ink formulation was performed on the ba-
sis of rheological studies, conducted comparing the dynamic
viscosities and surface tensions of typical ink solutions, opti-
mized for the thermal inkjet technology. The desired rheolog-
ical characteristics were obtained mixing specific additives,
such as viscosity regulators and surfactant agents, which are
of capital importance in determining the proper operation of
the inkjet printing device. We have then chosen to use the
thermal technology, which does not require ink with high
viscosity and low surface tension, thus reducing the amount
of needed additives in the system. The thermal inkjet tech-
nology has however a disadvantage with respect to the piezo-
electric one, since the ejection of the microdrop is due to
the formation of a gas bubble generated on the heater lo-
cated inside the nozzle. Such a generation occurs with a ther-
mal flash-time around 2 ␮s, at a temperature around 300 ◦ C,
and this heating phase, although being essentially a “shot”,
given the very short timing of the event, raises concerns
about the possible thermal degradation of the enzyme. In this
view, we have recently demonstrated that the thermal step of
the ink ejection does not represent a limitation for the mi-
crodeposition of a ␤-galactosidase from K. lactis (Setti et al.,
2004).
In this work we prepared another biological ink, starting
from GOD from Aspergillus niger in phosphate buffer, us-
ing EDTA as antimicrobial agent and glycerol as both stabi-
lizer and ink viscosity regulator, to achieve the best possible
performances from the thermal inkjet printer. A food dye
(Brilliant Blue FCF, E133) was added to the mixture as a col-
orimetric probe, in order to visualize the printed area and to
spectrophotometrically determine the volume of the ejected
ink. E133 was chosen for its stability at high temperature and
different pHs, and for having its UV–vis absorption maxi-
mum in the range 500–700 nm (with the peak at 628 nm),
with a high absorption coefficient (112.662 M−1 cm−1) . It is
hence very suitable for spectrophotometric purposes, and it
does not interfere with the product of the peroxidase catal-
ysed oxidative coupling between the MBTH and guaiacol,
in turn obtained by the assay for testing the GOD activity,
which has the maximum absorption at 505 nm (Fig. 1) (Setti
et al., 1998). The effect of the presence of E133 on the GOD
activity was found to be not significant (data not shown), thus
permitting a reliable spectrophotometric determination of the
GOD activity.
The GOD activity was not even affected by the additives
used for the formulation of the biological ink, since experi-
ments comparing a phosphate buffer solution with and with-
out additives showed essentially the same enzyme activity
(145 and 149 U/mL, respectively).
The volume of the recovered biological ink was
0.755 ␮L/cm2 (or 4.92 ␮L/in.2 ), calculated via spectropho-
tometric determination. As the volume of each drop ejected
Fig. 1. Spectrophotometric profiles of the red complex produced by the GOD
activity assay and Brilliant Blue FCF (E133).
(containing 6 pg of enzyme) through a single nozzle was
10 pL, the printhead had ejected 48 700 dots per square inch.
This means that the actual printing resolution achieved was
221 × 221 dpi, which was a quite low value with respect to the
nominal software-selected one of 600 × 600 dpi. The found
poor resolution, much lower than the expected one, was es-
sentially due to the scarce efficiency of the printhead, in which
some nozzles were most likely clogged or not available. No
denaturation effect due to the printhead heater was evidenced,
since the enzyme activity values in the biological ink before
and after the print were respectively 210.3 and 238.0 U/mL.
On this basis the printed activity resulted in 0.18 U/cm2 , cor-
responding to 2.38 × 10−5 U/dot.
The demonstrated enzymatic stability is probably ascrib-
able to a decreasing gradient of temperature from the surface
of the heater to the bulk of the ink solution during the “shot”,
which should allow the enzyme to feel temperatures lower
than the one of the nozzle heater surface. Moreover, also the
glycerol used in the biological ink as a wetting agent, to avoid
the clogging effect on the external nozzle surface, could play
a role in increasing the enzyme stability to the thermal shock,
by means of interactions between the polyol and the protein
(Graber and Combes, 1989; Athés et al., 1999).
3.2. PEDOT/PSS-based electronic ink
In a recent work we showed that it is possible to formu-
late an electronic ink, based on PEDOT/PSS and Tween 80
(a surfactant agent), suitable for being ejected by a thermal
printhead, and that the so-deposited thin film retains electro-
chemical performances perfectly comparable with the ones
measured for a film of the same polymer obtained by tradi-
tional spin-coating deposition (Ballarin et al., 2004). Follow-
ing that line a series of 10 successive depositions was per-
formed on the same surface area of an ITO-coated glass slide,
at a printing resolution of 600 × 600 dpi, in order to realize a
thick, multi-layered film. An AFM analysis performed in dif-
ferent zones of the film confirmed that no passing holes were
produced in the obtained print-coated layer. The film thick-
ness was measured scratching with an edged plastic tool the
surface of the film, obtaining thus a narrow PEDOT/PSS free
surface, and scanning with the AFM tip the polymer-coated
 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
Fig. 2. Three-dimensional and profile images of an inkjet printed film of PEDOT/PSS on ITO obtained by AFM, collected working in contact mode.
and the free surface, obtaining a topographical representation
of the film thickness as of a step. The so-measured thickness
was found to be about 230 nm (Fig. 2). This result, better
than the previously reported one in terms of control over the
film homogeneity (Ballarin et al., 2004), was obtained tuning
properly the printing software parameters.
3.3. Prototype of an amperometric GOD inkjet printed
electrode
A prototype of an amperometric GOD electrode was re-
alized by inkjet printing the active area on ITO-coated glass
with GOD, with the layout depicted in Fig. 3.
As the final sensor construction step, in order to prevent
dissolution of the enzyme and of the PEDOT/PSS film in the
aqueous environment for which the device has been prepared,
the printed device was encapsulated with a water-resistant
semi-permeable membrane of cellulose acetate. The mem-
brane was applied dip-coating the inkjet printed area in a so-
lution of cellulose acetate in acetone:THF 60:40 (Jawaheer
et al., 2003). The composition of the solution was calibrated
for obtaining a semi permeable membrane able, at the same
time, to permit an efficient diffusion of the glucose (Fig. 4a),
and to retain more than 90% of activity of the deposed GOD
inside the membrane after 20 min of immersion of the device
in still water (Fig. 4b; see Section 2).
Fig. 3. Schematic diagram of the prototype of the complete GOD inkjet printed electrode.
2023
Since the organic solvents mixture used for the preparation
of the cellulose acetate solution could have been detrimental
to the enzyme, an evaluation of the activity of the enzyme
after dip-coating was made as follows. A known quantity
of the enzyme was printed on an ITO-coated slide and cov-
ered by dip-coating, using the same solvents mixture above
described, but with a cellulose acetate concentration sensi-
bly lower than that used to protect the working device. The
so-obtained membrane was completely permeable to the en-
zyme (data not shown). After immersion of the so-obtained
assemble in still water for the same time allowed to the
working device, the activity of the recovered enzyme was
seen to be not significantly different, a finding that testifies
for a negligible denaturation effect of the organic solvents.
Possibly, cellulose acetate entraps the enzyme in a stable
state, without the necessity of a covalent modification of the
enzyme.
3.4. Electrochemical characterization
Two different GOD inkjet electrodes were prepared
with and without the PEDOT/PSS inkjet printed layer. All
the biosensors were tested in presence or absence of fer-
rocenemethanol (FeMeOH), which was chosen as a medi-
ator because of its high efficiency in terms of electron trans-
fer with respect to the potassium ferricyanide (Ge et al.,
1998).
2024 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
Fig. 4. Percentage of glucose (a) and GOD (b) permeated through membranes obtained by dip-coating from solutions with different cellulose acetate concen-
trations.
Fig. 5a shows the cyclic voltammogram in the potential
range from +0.75 to −0.75 V, at a scan rate of 50 mV/s, of
a system in which the electrode was constituted by the bare
ITO-coated glass dipped in the buffer solution at pH 6.5 in
presence of FeMeOH (14 mg/L). The anodic peak at +0.43 V
and the corresponding cathodic peak at 0.00 V, due to the
electrochemical reaction of the FeMeOH, are well evident.
After the addition of glucose (40 mM) and GOD into the so-
lution, the cyclic voltammetry curves changed as an effect of
the catalytic process that took place, showing the decrease of
the cathodic peak and the increase of the anodic one (Fig. 5b).
This behaviour is indicative of the regeneration of FeMeOH
from the ferricinium ion by the enzyme glucose oxidase in
its reduced form, as reported in the following scheme of re-
action, in the absence of oxygen:
GODox + glucose → GODred + gluconolactone (1)
GODred + (FeMeOH)ox → GODox + (FeMeOH)red (2)
(FeMeOH)red → (FeMeOH)ox + e− (3)
The voltammetric path obtained with an electrode having
inkjet printed GOD on ITO glass, and then encapsulated
with the cellulose acetate mebrane, showed a shift of the
anodic peak of the FeMeOH at +0.32 V (see Fig. 6a). Af-
ter the addition of glucose into the solution, a decrease of
Fig. 5. Cyclic voltammograms of the ITO-glass electrode in presence of
FcMeOH (14 mg/L) and GOD into the buffer solution, before (a) and after
the addition of glucose (40 mM) (b).
the cathodic peak of the FeMeOH, as well as a relative in-
crease of the anodic one, was noted, in accordance with the
above mentioned observations (see Fig. 6b and c). These vari-
ations were proportional to the concentration of glucose in
solution.
After these preliminar experiments, a complete inkjet
printed GOD electrode (i.e. GOD printed on PEDOT/PSS,
in turn printed on ITO-coated glass, everything encapsu-
lated with the cellulose acetate membrane) with the lay-
out described in Fig. 3, was realized and tested. The
recorded voltammetric path showed an anodic peak at
+0.56 V in the presence of FeMeOH (Fig. 7b), while with-
out the mediator the voltammetric measurement of the
device produced only the PEDOT/PSS signal (Fig. 7a).
Little variations of the voltammetric path were obtained
after the addition of the glucose in solution in both
systems.
The performances of the GOD inkjet printed electrodes
were investigated also by chronoamperometry in absence
of oxygen. According to the voltammograms, +0.30 and
+0.60 V were chosen as the working potential of the enzyme
electrodes, since at these potentials PEDOT/PSS was in its
oxidated, thus conductive, state, and the FeMeOH redox re-
action occurred.
Fig. 6. Cyclic voltammograms of the GOD directly inkjet printed electrode
on ITO-glass electrode in the presence of the FeMeOH (14 mg/L) after the
addition of increasing concentrations of glucose into the buffer solution: (a)
0 mM, (b) 40 mM and (c) 80 mM.
 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
Fig. 7. Cyclic voltammograms of the complete PEDOT/PSS-GOD inkjet
printed on ITO-glass electrode in the absence (a) and in the presence (b) of
the FeMeOH (14 mg/L) before and after the addition of glucose (40 mM)
into the buffer solution.
It should be noted that both the electrodes, with or without
PEDOT/PSS printed layer, showed a weak response in the
absence of mediator.
In the electrode with both the mediator and the PE-
DOT/PSS inkjet printed layer, the response determined
by the slope of the linear interpolation of the curves,
at the working potentials of +0.30 and +0.60 V, resulted
quite linear up to 60 mM in glucose and attained val-
ues of 3.57 ␮A M−1 cm−2 (248 nA M−1 U−1 cm−2 ) and
6.43 ␮A M−1 cm−2 (447 nA M−1 U−1 cm−2 ), respectively
(Fig. 8b and c). The sensitivity of this device was signifi-
Fig. 8. Chronoamperometric response of the two prototype inkjet printed
electrodes in the presence of the FeMeOH (14 mg/L) as a function of the
glucose concentration in the buffer solution: GOD directly inkjet printed
on ITO-glass electrode with an applied potential to the working electrode
of +0.30 V (a), and the complete PEDOT/PSS-GOD inkjet printed electrode
with an applied potential to the working electrode of +0.30 V (b) and +0.60 V
(c).
2025
cantly higher than that observed in the biosensor in which
the GOD was directly printed on ITO glass (Fig. 8a).
The presence of the PEDOT/PSS inkjet printed layer
seems therefore to enhance the sensitivity, a fact that is more
evident testing the preliminary response at the working poten-
tial of +0.60 V, corresponding to the anodic peak. Infact, at
this potential the electrode having the GOD directly inkjet
printed on ITO glass did not present significant response
(data not reported), while the complete electrode having the
PEDOT/PSS inkjet printed layer showed the highest sensi-
tivity value, of 6.43 ␮A M−1 cm−2 (447 nA M−1 U−1 cm−2 )
(Fig. 8c). In general, the direct electron transfer between
the GOD and a conducting polymer is still an open debate.
Nonetheless, the mentioned evidences point to the fact that
FeMeOH could actually mediate the electron transfer from
GOD to the electronically conducting polymer, which is able
to accelerate the heterogeneous electron transfer to the ITO-
glass surface, as already reported (Nagel and Staes, 2001).
The complete inkjet printed sensor here described, though
working, showed electrical responses quite lower than those
commonly reported in the literature for GOD entrapped
in conducting polymers, often comprised between 2 and
5 mA M−1 cm−2 (Piro et al., 2001). This rather poor perfor-
mance was probably due to the film thickness of the conduc-
tive polymer realized in our sensor. Piro et al. (2001) have
infact reported that the sensitivity of a modified GOD incor-
porated within poly(3,4-ethylenedioxythiophene) (PEDOT)
electropolymerised on glassy carbon electrodes decreases
with thicker films, and has a maximum value for thickness of
the film in the range of 10–20 nm, and this finding could be
due to a low diffusion of both the mediator and the substrate
through the polymeric film (Shin and Kim, 1996). The thick-
ness of the conductive film in our experiments was about
230 nm, which likely resulted too high with respect to the
optimum value. Attempts to realize thinner polymer layers
are actually in progress, as well as investigations on other
possible conjugated polymers as potential electrode–enzyme
interfaces.
Another important point for electrodes including immo-
bilized enzymes and mediators is the prevention of their dis-
solution into the environment during in vivo measurements
(Nagata et al., 1995). Although the use of the cellulose mem-
brane demonstrated to fully retain the enzyme on the elec-
trode surface, the mediator stabilization remains an open is-
sue, and work is in progress under this point of view.
4. Conclusions
In this work we have shown that thermal inkjet printing
could be a viable technology to realize biosensors depos-
ing active proteins and conductive polymers on electroni-
cally active substrates, through the formulation of specific
biological and electronic water-based inks. In particular, the
possibility to print the enzyme glucose oxidase (GOD) and
the conducting polymer blend PEDOT/PSS, retaining their
2026 L. Setti et al. / Biosensors and Bioelectronics 20 (2005) 2019–2026
qualifying properties also after the print, has been demon-
strated. Through a proper coupling of the two components,
a completely inkjet printed, working glucose biosensor has
been manufactured, though a decise improvement of the so-
realized device performances, which is the target of the near-
future work, has to be made.
As an outlook, such an approach opens an economical
way for the development of multifunctional, bioelectronic
microdevices. Infact, the realization of printable, active bio-
logical and electronic inks is the first necessary step for the
development of the “biopolytronics”, i.e. the computer aided
design of microsized, multifunctional and complex biosen-
sors, followed by their immediate practical implementation
on virtually any surface via a suitable inkjet multi-ink printer.
Acknowledgements
Ms. Manuela Di Biase is gratefully acknowledged for the
active contribution in part of the presented work. The Italian
Minister for Instruction, University and Research is acknowl-
edged for funding.
References
Albareda-Sirvent, M., Merkoci, A., Alegret, S., 2000. Configurations used
in the design of screen-printed enzymatic biosensors. A review. Sens.
Actuator B 69, 153–163.
Allain, L.R., Askari, M., Stokes, D.L., Vo-Dinh, T., 2001. Microar-
ray sampling-platform fabrication using bubble-jet technology for a
biochip system. Fresenius J. Anal. Chem. 371 (2), 146–150.
Athés, V., Guerra, P., Combes, D., 1999. Effect of soluble additives on
enzyme thermo- and/or baro-deactivation. J. Mol. Catal. B: Enzym.
7, 1–9.
Ballarin, B., Fraleoni-Morgera, A., Frascaro, D., Marazzita, S., Piana, C.,
Setti, L., 2004. Thermal Inkjet Microdeposition of PEDOT/PSS on
ITO-coated glass and characterization of obtained film. Synth. Met.
146, 201–205.
Blanchard, A.P., Kaiser, R.J., Hood, L.E., 1996. High-density oligonu-
cleotide arrays. Biosens. Bioelectron. 11 (6/7), 687–690.
Chen, B., Cui, T., Liu, Y., Varahramyan, K., 2003. All-polymer RC filters
fabricated by all-inkjet printing technique. Solid State Electron. 47,
841.
Ge, F., Zhang, X.E., Zhang, Z., Zhang, X.M., 1998. Simultaneous de-
termination of maltose and glucose using a screen-printed electrode
system. Biosens. Bioelectron. 13 (3/4), 333–339.
Graber, M., Combes, D., 1989. Effect of polyols on fungal ␣-amylase
thermostability. Enzyme Microb. Technol. 11, 532–538.
Hart, A.L., Turner, A.P.F., 1996. On the use of screen- and inkjet print-
ing to produce amperometric enzyme electrodes for lactate. Biosens.
Bioelectron. 11 (3), 263–270.
Kawase, T., Sirringhaus, H., Friend, R.H., Shimoda, T., 2001. Inkjet
printed via-hole interconnections and resistors for all-polymer tran-
sistor circuits. Adv. Mater. 13, 1601.
Kenyon, R.W., 1996. Chemistry and Technology of Printing and Imaging
System. Blackie Academic & Professional, Glasgow.
Kros, A., van Hövel, S.W.F.M., Nolte, R.J.M., Sommerdijk, N.A.J.M.,
2001. A printable glucose sensor based on a poly(pyrrole)-latex hybrid
material. Sens. Actuators B 80, 229–233.
Jawaheer, S., White, S.F., Rughooputh, S.D.D.V., Cullen, D.C., 2003.
Development of a common biosensor format for an enzyme based
biosensor array to monitor fruit quality. Biosens. Bioelectron. 18 (12),
1429–1437.
Le, H.P., 1998. Progress and trends in ink-jet printing technology. J.
Imaging Sci. Technol. 42 (1), 49–62.
Mills, R., 1996. Ink jet printing: past, present and future. In: Rezanka, Es-
chbach (Eds.), IS&T: Recent Progress in Ink Jet Technologies, IS&T,
Springfield, VA, pp. 12–15.
Nagata, R., Yokoyama, K., Clark, S.A., Karube, I., 1995. A glucose sensor
fabricated by the screen printing technique. Biosens. Bioelectron. 10,
261–267.
Nagel, L.J., Staes, E., 2001. Polymer (bio)materials design for ampero-
metric detection in LC and FIA. Trends Anal. Chem. 20 (4), 178–185.
Newmann, J.D., Turner, A.P.F., Marrazza, G., 1992. Inkjet printing for
the fabrication of amperometric glucose biosensors. Anal. Chim. Acta
262, 13–17.
Park, A., Wu, B., Griffith, L.G., 1998. Integration of surface modification
and 3D fabrication techniques to prepare patterned poly(l-lactide)
substartes allowing regionally selsctive cell adhesion. J. Biomater. Sci.
Polym. Ed. 9 (2), 89–110.
Petridou Xu, T., Lee, E.H., Roth, E.A., Vyavahare, N.R., Hickman, J.J.,
Boland, T., 2004. Construction of high-density bacterial colony arrays
and patterns by the inkjet method. Biotechnol. Bioeng. 85 (1), 29–33.
Ping-Hei, C., Wen-Cheng, C., Chang, S.H., 1997. Bubble growth and ink
ejection process of a thermal-bubble-jet printhead. Int. J. Mech. Sci.
39 (6), 683–695.
Piro, B., Dang, L.A., Pham, M.C., Fabiano, S., Tran-Minh, C., 2001. J.
Electroanal. Chem. 512, 101–109.
Roda, A., Guardigli, M., Russo, C., Pasini, P., Baraldini, M., 2000. Protein
microdeposition using a conventional inkjet printer. Biotechnique 28,
492–496.
Roth, E.A., Xu, T., Das, M., Gregory, C., Hickman, J.J., Boland, T.,
2004. Inkjet printing for high-throughput cell patterning. Biomaterials
25, 3707–3715.
Schena, M., Heller, R.A., Theriault, T.P., Konrad, K., Lachenmeier, E.,
Davis, R.W., 1998. Microarrays: biotechnology’s discovery platform
for functional genomics. Trend Biotechnol. 16 (7), 301–306.
Setti, L., Piana, C., Bonazzi, S., Ballarin, B., Frascaro, D., Fraleoni-
Morgera, A., Giuliani, S., 2004. Thermal inkjet technology for the
microdeposition of biological molecules as viable route for the real-
ization of biosensor. Anal. Lett. 37 (8), 1559–1570.
Setti, L., Scali, S., Degli angeli, I., Pifferi, P.G., 1998. Horseradish
peroxidase-catalized oxidative coupling of 3-methyl 2-
benzothiazolinone hydrazone and methoxyphenols. Enzyme
Microb. Technol. 22 (8).
Sherwood, J.K., Riley, S.L., Palazzolo, R., Brown, S.C., Monkhouse,
D.C., Coates, M., Griffith, L.G., Landeen, L.K., Rateliffe, A., 2002.
A three-dimensional osteochondral composite scaffold for articular
cartilage repair. Biomaterials 23 (24), 4739–4751.
Shin, M.C., Kim, H.S., 1996. Electrochemical characterization of polypyr-
role/glucose oxidase biosensor. II. Optimal preparation conditions for
the biosensor. Biosens. Bioelectron. 11, 171–178.
Sirringhaus, H., Kawase, T., Friend, R.H., Shimoda, T., Inbasekaran,
M., Wu, W., Woo, E.P., 2000. High-resolution inkjet printing of all-
polymer transistor circuits. Science 290, 2123.