IEEE SENSORS JOURNAL, VOL. 18, NO.

10, MAY 15, 2018 4097

Design and Characterization of Screen-Printed

Textile Electrodes for ECG Monitoring
Andrea Achilli, Annalisa Bonfiglio, Member, IEEE, and Danilo Pani, Member, IEEE

Abstract— This study was conducted to develop a screen- low cost and can provide high-quality and stable signals. Such
printed approach for the fabrication of textile electrodes and electrodes are typically composed of a metallic snap fastener
analyze their characteristic features for use as electrocardio- that is crimped to an Ag/AgCl button, which is eventually
graphy (ECG) signal sensors. The electrodes were obtained
by printing poly(3,4-ethylenedioxythiophene) doped with surrounded by a conductive layer to enlarge the sensing area.
poly(4-styrenesulfonate) organic bio-compatible polymer-based The coupling with the skin is supported by an ionic element,
ink on a finished fabric using a polyester mesh with 43 lines/cm. such as solid or liquid hydrogel, to improve the quality of the
The physical–chemical properties of the polymer solution contact, thus reducing the associated impedance. This type of
were varied to obtain a simple and reproducible fabrication electrodes cannot be adopted for long-term monitoring as they
process and a good electrode performance simultaneously.
The electrode properties were first assessed by a benchtop can cause skin reactions [2] to either the glue adopted for the
measurement instrument and then on human subjects. Several adhesive support or the hydrogel and thus cannot comply with
tests under different conditions, for instance, by adding liquid the comfort requirements of wearable applications.
and solid electrolytes, revealed that the electrodes possess The non-invasive nature of the ECG recording and the low
the majority of characteristics required by the ANSI/AAMI cost of the devices employed to acquire it make it possible for
EC12:2000 standard for gelled ECG electrodes. Furthermore,
the benchtop measurements demonstrated that these electrodes its exploitation in any scenario, beyond the clinical use. For
preserve their electrical properties and functionality even after instance, health self-management devices, wearable activity
several washing cycles (performed under mild conditions), while trackers and fitness performance recorders have now become
they suffered physical stretching. While using some electrolytes, mainstream consumer products.
the skin contact impedance and the features extracted from the This study focuses on the technology behind the develop-
physiological signals were found to be highly similar to those
obtained with disposable gelled Ag/AgCl electrodes (ρ > 0.99). ment of the so-called ‘smart textiles’ [3], [4]. The principal
Through the proposed fabrication process, it is possible to feature of these wearable devices is their ability to be used
change the electrical properties of a specific region of a finished as chemical or physical sensors exploiting the properties of
fabric or a garment with an excellent geometrical resolution. the material, or of the conductive elements woven into the
This opportunity, along with the obtained results in terms of fabric, to analyse and/or respond to signals acquired from
the electrical characteristics as ECG electrodes, indicates that
the proposed approach could possibly be adopted in the future the surrounding environment [5]–[12]. Smart textiles can be
for the development of smart garments that could comfortably exploited for creating smart garments that can be used to
detect ECG signals. monitor physiological parameters. In the scientific literature,
Index Terms— Screen-printed electrodes, ECG, textile elec- it is possible to find several examples of smart textiles for ECG
trodes, PEDOT:PSS. monitoring [13]–[20]. In recent years, research in this field has
been making great progresses, developing new technologies
I. I NTRODUCTION that can provide the textiles the ability to communicate,
transform, conduct energy and self-repair [7].
D IAGNOSTIC electrocardiography (ECG) is a consoli-
dated clinical practice performed in adults and children,
which enables the identification of pathological conditions
The majority of such smart textiles are composed of textile
substrates that have been treated to change their physical prop-
even before the diagnosis of structural changes in the heart by erties, in particular, switching from electrically non-conductive
other methods [1]. Clinical ECG is generally acquired using to conductive fabrics. To achieve this aim, textile substrates
disposable gelled Ag/AgCl electrodes, which are available at are typically modified by adding conductive elements (such as
metallic fibres inside the yarns) or treated by coating fibres,
Manuscript received December 5, 2017; revised January 24, 2018 and yarns or textile surfaces with Au or Ag nanoparticles [21].
March 9, 2018; accepted March 9, 2018. Date of publication March 26, 2018; The conductivity of the textile substrate can also be
date of current version April 23, 2018. The associate editor coordinating the
review of this paper and approving it for publication was Prof. Alper Bozkurt. modified by treating the textile with a polymer, thereby
(Corresponding author: Andrea Achilli.) avoiding the use of metallic components in contact with the
A. Achilli is with the Department of Electrical and Electronic skin. Typically, poly(3,4-ethylenedioxythiophene) doped with
Engineering, University of Cagliari, 09123 Cagliari CA, Italy, and also
with DIBRIS, University of Genoa, 16145 Genova GE, Italy (e-mail: poly(4-styrenesulfonate), termed as PEDOT:PSS, an organic
andrea.achilli@diee.unica.it). bio-compatible polymer, is used for this purpose [22]–[27].
A. Bonfiglio and D. Pani are with the Department of Electrical and PEDOT:PSS has been also used as an anti-static coating
Electronic Engineering, University of Cagliari, 09124 Cagliari CA, Italy
(e-mail: annalisa@diee.unica.it; pani@diee.unica.it). material, to produce electrodes for capacitors or photodiodes
Digital Object Identifier 10.1109/JSEN.2018.2819202 and as a hole transport layer in organic LEDs [28]–[31].
1558-1748 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 4098 IEEE SENSORS JOURNAL, VOL. 18, NO. 10, MAY 15, 2018

An important advantage of this polymer is the possibility

to easily cover the surfaces using simple coating methods
such as spin-coating [32]. The conductivity of PEDOT:PSS
can be enhanced by adding elements such as surfactants,
reticular agents and organic solvents [33]. Some of these
dopants can change the viscosity of the resulting PEDOT:PSS
solution, thus enabling different production processes, includ-
ing dip coating [34], direct fabric patterning [13] and ink-
jet printing [17]. Differently from dip coating, the other
approaches can be used to selectively functionalise limited
areas of a finished garment, thereby avoiding seaming the
electrodes on it. For the same reason, screen-printing was also Fig. 1. Screen-printed textile electrode fabrication. On the left, different
views of the prototypical electrodes; on the right, their expanded parts view.
proposed in the past to develop circuits [35], antennas [36],
transmission lines [37] and supercapacitors [38] on textile
substrates, by using inks based on silver or carbon. The
same technique and inks were also used for the fabrication According to several studies, the conductivity of a polymer
of electrodes for electrophysiological signals. At the state of increases dramatically upon the addition of organic solvents
the art, these sensors, named screen-printed electrodes (SPE), such as dimethyl sulfoxide (DMSO) and ethylene glycol [36].
exploit flexible plastic substates as polyethylene naphthalate The ink used for screen printing is composed of two compo-
(PEN) or polyethylene terephthalate (PET) [39], [40]. Only nents of PEDOT:PSS, one component being ethylene glycol
in [41], a silver ink was deposited on the fabric by screen- and the other being 3-glycidoxypropyltrimethoxysilane [29].
printing, but after a pre-conditioning of the substrate wih a To improve the geometrical resolution of the screen printing
polyurethane layer. process, the viscosity of the ink was increased by reducing
This study describes and validates a technique for the devel- the water:polymer ratio [43]–[45]. This was accomplished by
opment of ECG textile electrodes based on screen-printing of drying the solution in an oven at 90°C until its density became
a PEDOT:PSS solution. This approach allows realising any similar to that of acrylic inks.
electrode shape with high reproducibility at a very low cost, To deposit the ink on a specific area, it was transferred
resulting in completely seamless smart garments. This further onto the textile substrate through a polyester screen with 43T
reduces the occurrence of skin irritation by preserving the mesh delimitated by a wax stencil using a rubber squeegee.
smooth texture of the fabric. To assess the primary features of Immediately after the deposition, the electrodes were dried
the proposed electrodes, both benchtop and in-vivo measure- in the oven at 70°C. This process needs to be executed in a
ments were performed, and the obtained results were compared timely manner since the ethylene glycol would start spreading
with those of disposable Ag/AgCl electrodes. Based on the through the fabric, thereby affecting the printing precision.
literature in this field, tests were carried out under dry and wet This is the primary obstacle for drying the printed electrodes
conditions and with different electrolytes to analyse different at room temperature.
application scenarios. Both stretchable and non-stretchable To measure the ECG signal, the electrodes were designed
substrates were tested, simulating daily use, including stretch- with a 4 cm2 -sensitive area, so that they could be comparable
ing and washing cycles. Results were also evaluated in terms of in size with commercial disposable surface electrodes. This
the ANSI/AAMI EC12:2000/(R)2010 standard for disposable area has a square shape, with an additional conductive path
ECG electrodes. (1.5 × 0.5 cm) employed to apply a connector and pick up the
signal (Fig. 1).
1) Prototypical Electrode Assembly: To obtain a usable
II. M ATERIALS AND M ETHODS
textile ECG electrode from the functionalised fabric, we added
This section first describes the methodology followed for some features such as a medical-grade Ag/AgCl fastener to
the electrode production, along with the protocols adopted easily connect the electrode to snap-on patient cables of the
for electrode stress tests. Then, the benchtop measurements recording devices.
on such electrodes are described, followed by a detailed Furthermore, the textile electrodes included a 2-mm-thick
description of the measurements performed on human subjects. layer of foam placed on the fabric on the opposite side of the
active area. This foam was added to maintain both a homoge-
neous and continuous pressure [41], and a wet skin condition
A. Screen-Printed Electrode Fabrication Process during the acquisition [46] when a liquid electrolyte is added,
The proposed textile ECG electrodes consist of a sheet as performed in the experimental sessions. In fact, the foam
of cotton fabric with conductive properties. As mentioned absorbs the liquid, preserving a homogeneous hydration level
earlier, this fabric has been treated by screen-printing throughout the measurement. The foam was coupled to the
with an ink based on PEDOT:PSS. The chosen HERAUS fabric using the same medical tape that was used to apply the
CLEVIOUSTH PH1000 PEDOT:PSS solution is composed electrodes on the skin. This tape was shaped and placed on
of 1.0 wt%–1.3 wt% of polymer in a water suspension, with the fabric, with the precaution of isolating the channel and the
a nominal viscosity ranging from 15–50 cps. snap button from the skin. A spot was also provided to apply

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 ACHILLI et al.: DESIGN AND CHARACTERIZATION OF SCREEN-PRINTED TEXTILE ELECTRODES FOR ECG MONITORING 4099

a few drops of saline solution between the electrode and the TABLE I
skin surface. ANSI/AAMI S TANDARD EC 12:2000 L IMITS
2) Electrolytes and Substrates: In principle, the mixed
electronic/ionic conduction of PEDOT:PSS [34] allows using
such textile electrodes under dry conditions for ECG record-
ing. The possibility of creating dry electrodes integrated
into a smart garment represents a fascinating goal. However,
it is known from the literature that the interface with the
skin in this condition has a negative effect on the quality
of the acquired signal [46]. The signal can be corrupted
by a significant amount of baseline drift, motion artefacts the excess water by pressing the wet fabrics with paper towels,
and power line interference [41]. This is due to both the the fabrics are dried in an oven at 70°C for 15 min.
high skin contact impedance and the non-uniform and pos- Stretching was simulated by recreating the conditions that
sibly insufficient pressure exerted by the electrode on the could be experienced by the electrodes when they are embed-
skin [41]. ded in a stretchable smart garment. To achieve this aim,
The former problem can be mitigated by the adoption of the electrodes, which were screen-printed on the stretchable
an electrolyte, either solid [14] or liquid [46]. In the present substrate, were stretched to 100% of their length three times.
study, the electrodes were tested under dry conditions and
also with the addition of different electrolytes, namely, 0.9% B. Benchtop Measurements
NaCl physiological solution (saline) and solid adhesive KCl This section describes the methods used for electrode char-
hydrogel. The saline solution can mimic sweat, for example, acterisation by a benchtop measurement equipment, with spe-
produced during physical activity, whereas the solid hydrogel cific reference to the electrical properties and the requirements
enables a fair comparison with commercial disposable gelled for electrophysiological sensing.
Ag/AgCl electrodes. 1) Sheet Resistance: The sheet resistance was evaluated
The need of improving the pressure of the electrode on using an LCR meter (4284A by Agilent/HP Keysight) for
the skin can be fulfilled by inserting a foam layer behind the a four-probe resistance measure. This measurement allows
sensitive area, as described earlier, and also by printing the monitoring the planar conductivity of the electrodes by remov-
electrodes on an elastic substrate. The latter approach could ing the probe contact impedance. The resistance measure-
also be used for the development of smart sportswear that can ment was performed by applying an AC electrical current at
monitor the ECG signal during an athletic workout. However, 20 Hz between the external probes and measuring the voltage
the negative effect of stretching the substrate needs to be inves- between the internal ones, the four tips being equally spaced
tigated, as will be described in the next section. To achieve by 2.54 mm (1 mil). The sheet resistance value (Rs ) was
this aim, we developed textile electrodes by functionalising computed by solving (1).
both non-stretchable and stretchable fabrics. The first one was    
π V V
composed of 100% cotton fibres with a weight of 116 g/m2 and Rs = × ≈ 4.53 × (1)
threads weft of 25 n° /in. It was chosen because of its natural ln 2 I I
fibers composition and its breathability. The stretchable fabric 2) Xtratek Electrode Benchtop Tester Measurements:
(84% polyester and 16% spandex), with a weight of 230 g/m2 Although there is a lack of standards for textile electrodes,
and threads weft 20 n° /in, was chosen to represent the typical it is possible to assess their performance by observing the
material used to develop smart sportswear. Remarkably, any limits reported in the ANSI/AAMI EC12:2000/(R)2010 for
other similar fabrics could have been used. disposable ECG electrodes [48] in terms of DC offset, AC
3) Washing and Stretching Procedure: As screen-printed impedance and simulated defibrillation recovery (TABLE 1).
textile electrodes are designed for the development of smart The standard refers to adhesive gelled electrodes that are
clothes, some technical aspects related to real-world applica- tested in couples by putting their active surfaces in contact.
tion scenarios need to be considered. A lacking aspect in the Since such electrodes present either solid or liquid hydrogel,
majority of scientific articles concerning the development of and an adhesive support, the measurements are characterized
textile electrodes is the resistance to physical stresses such by the perfect contact between the two electrodes under
as stretching and washing cycles. The following procedures test. To achieve comparable results with non-gelled and non-
were applied to investigate how the ink conductivity and the adhesive electrodes, such as those proposed in this study,
performance of ECG electrodes change. the physical coupling between the two electrodes was ensured
Washing was simulated following the ISO 105-C10:2006 by pressing them one against the other with a reciprocal force
standard that was designed to establish the effect of washing of 9 N. This force was removed while testing the textile
cycles on the ink fastness on textile substrates [47]. To recre- electrodes with the adhesive solid hydrogel.
ate mild washing conditions, as per standard requirements, To perform such measurements, the Xtratek ET-65A ECG
the printed textiles are placed in an aqueous solution con- electrode tester (Xtratek, Lenexa, Kans., USA) was used [49].
taining 5% of a low-duty laundry detergent for clothes and This instrument was developed to benchmark coupled gelled
gently stirred at 40°C for 20 min. At the end of the process, electrodes respecting the conditions required by the stan-
the soap is washed away with clean water and, after removing dard [48], and it is used in research and development (either

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 4100 IEEE SENSORS JOURNAL, VOL. 18, NO. 10, MAY 15, 2018

in academia or industry) to assess the conformity of ECG measurement, a disposable gelled Ag/AgCl off-the-shelf elec-
electrodes to the standard. Compared to other custom solu- trode (CDES000024 by Spes Medica Srl) was used.
tions, the adoption of this instrument allows evaluating all To compare the performance of the screen-printed textile
the required parameters simply connecting a paired couple of electrodes and that of traditional, disposable, gelled Ag/AgCl
electrodes to the device and pressing a single button to start electrodes, the same placement was adopted for both. For this
every measurement. purpose, the measurement with the proposed textile electrodes
The sample size chosen for these tests consisted of five pairs was first performed, either under dry conditions or with differ-
of electrodes. ent electrolytes. Then, the electrodes were replaced with the
3) Benchtop Impedance Measurement by LCR Meter: Using disposable Ag/AgCl electrodes for a comparative recording.
Xtratek ET-65A, the impedance can be measured only at the As an alternative to such an asynchronous recording, two
specific frequency of 10 Hz (TABLE 1), since this value couples of electrodes (textile and Ag/AgCl) could also be
is imposed by the standard [48]. Because of this limitation, applied for a synchronous recording [54]. However, a different
to investigate how the PEDOT:PSS electrodes work at different signal amplitude can be experienced on the two parallel leads
frequencies, the same LCR meter used for the evaluation of only because of the slight difference in the heart vector pro-
the sheet resistance was employed. The measurements were jection [34]. As the ECG signal will be analysed in this study,
performed coupling the active surfaces of two electrodes at including its amplitude features, the asynchronous approach
a time. This device allows acquiring the impedance in a was preferred.
frequency range from 20 Hz to 1 MHz. 2) Electrophysiological Recording System: The ECG signal
The characteristics of the ECG signal must be considered was recorded using a 32-channel Porti7 system by TMSI.
while defining the upper limit of the frequency range for A single bipolar channel was used to perform the ECG lead I
this analysis. The QRS complex is the sharpest waveform measurement, and a passive ground lead was connected to the
in the signal; although low-amplitude signal components as reference electrode placed on the subject. The device features a
high as 500 Hz have been detected and studied [50], the level high-resolution analogue-to-digital converter (22 bits) without
of the noise could overcome the signal of interest at such any analogue high-pass filtering stage. This method compen-
frequencies. At the same time, the QRS complex of infants sates for the large converter range for the baseline drift of the
can contain significant components up to 250 Hz [51]. This is signal. The resolution on the bipolar channels is 71.5 nV, and
why the analysis was carried out applying an AC sinusoidal the noise level is <1 μV at the lowest sampling rate. The
current, with a 0.1 mA amplitude, at different frequencies from sampling frequency for the ECG signals was set to 2048 Hz.
20 Hz to 300 Hz (10 frequency points: 20, 25, 30, 40, 60, 80, The instrument features an active shielding noise-protection
100, 150, 250 and 300). Remarkably, this benchtop impedance strategy, with micro-coaxial cables terminated by a snap-
measurement does not depend on the skin condition; hence, the on connector. This strategy significantly reduces the com-
variability can only be ascribed to the electrode characteristics. mon mode signal coupled with the cable leads. The input
impedance of the recording systems is in the order of 1 T,
which is high enough to consider the instrument to be suffi-
C. Human Tests ciently insensitive to the skin–electrode impedance level. The
This study was performed according to the principles out- signal was recorded and visualised in real time on TMSi
lined in the Declaration of Helsinki of 1975, as revised PolyBench, and the digital data were exported into Matlab
in 2000. Four healthy voluntary subjects belonging to the for all the subsequent signal analyses.
research group (aged 33 ± 6 years, BMI 24 ± 2 kg/m2 ) 3) Skin Contact Impedance: The same electrode configura-
were enrolled, and their informed consent for the acquisition tion adopted for the ECG was used to acquire the skin contact
protocol was obtained. impedance as well. This measurement was performed using the
During the measurements, the subjects were comfortably same LCR meter adopted for the impedance benchtop mea-
seated in a room with a constant temperature of about surements, at the same frequencies of the previous analysis, but
23°C. A 1-min delay between electrode positioning and applying an AC sinusoidal signal with a 500 mV. It is worth
impedance measurement was adopted to allow the stabilisation mentioning that this type of measurement on the human skin
of the skin–electrode interface. For evaluating the relationship is heavily affected by the skin conditions; hence, the results
between the skin contact impedance and the signal quality, the should be carefully considered and cannot be generalised.
different electrolytes introduced in Section II.A.2) were added To estimate the skin–electrode impedance related to a single
to the textile electrodes. electrode, we divided by two the modulus of the measured
1) ECG Recording and Feature Extraction: Lead I of impedance, which is actually the series of three impedances,
the ECG was recorded according to the Mason-Likar mod- i.e. the skin contact impedance of the first electrode, the
ification [52] of the standard 12-leads system [53]. Such negligible impedance contribution of the body, and the skin
an approach, typically adopted for ECG stress tests, allows contact impedance of the second electrode [34]. Obviously,
reducing the electromyographic interference compared to the such an approach only leads to a rough average estimate of
standard positioning. The signal ground electrode was also the skin contact impedance, which would be justified by the
placed according to the Mason-Likar scheme, thus in a assumption that the skin condition under the two electrodes is
neutral position with respect to the recording site. Since similar. Even though this hypothesis is questionable when the
this electrode is not directly involved in the bipolar lead skin contact impedance is high, such as for dry electrodes, it is

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 ACHILLI et al.: DESIGN AND CHARACTERIZATION OF SCREEN-PRINTED TEXTILE ELECTRODES FOR ECG MONITORING 4101

good enough when some electrolytes are added, because of the TABLE II
better coupling with the skin. In any case, due to the variability ACZ AND N OISE M EASURED BY X TRATEK ET-65A
C OUPLING T EXTILE E LECTRODES
of the skin condition between subjects, and from time to time
for the same subject [55], such an estimate provides acceptable
information for the purposes of this work.
4) ECG Signal Analysis: The spectrum of the ECG signal
was extracted using the Welch’s power spectral density (PSD)
estimation method [56] over a 4-s window with a 1-s overlap.
A comparative analysis between the PSD values estimated
on the signals acquired through the textile and Ag/AgCl
electrodes was performed.
Beyond the PSD, some morphological ECG signal aspects
in the time domain can be analysed, as they are used for represents the mean of the parameter over six measures for
diagnostic purposes. A normal ECG presents three primary the impedance and ten measures for the noise, while the error
epochs, which are associated to three events in the cardiac represents the standard deviation.
cycle: In addition to such important parameters, the DC offset was
• the P wave represents the atrial depolarisation also measured to assess the electrode performance. A stable
• the QRS complex (composed of the Q, R and S waves) value of 0 mV both in the dry condition and with saline
represents the ventricular depolarisation was observed, whereas the addition of a solid hydrogel film
• the T wave represents the ventricular repolarisation resulted in a DC offset value of 2 ± 1 mV.
The QRS is the highest and sharpest waveform in the The measurements performed to simulate a defibrillation
ECG and, as such, is commonly used at the first stage of recovery using the screen-printed organic ink exceeded the
ECG segmentation algorithms and as a reference for the limits imposed by the ANSI/AAMI standard and those of the
computation of the heart rate, under the assumption of a sinus Xtratek ET-65A (400 mV and 9.9 mV/s) for both offset and
rhythm. In this study, the QRS complex amplitude, measured recovery slopes.
peak-to-peak, its duration and the RR interval, defined as For the stretchable substrate, tested only under dry condi-
the time interval between consecutive R wave peaks, were tions, the same results can be achieved in terms of DC offset
extracted and analysed. and SDR, while the results in terms of impedance and noise
To extract the QRS complex features, the ECG delineation are 243 ± 1  and 47.9 ± 1.5 μV, respectively.
algorithm presented in [57] was chosen because of its precision 3) Daily Use Simulation: Daily use was simulated by
and reliability. The algorithm is based on a multi-scale wavelet repeated washing cycles and stretching tests. To avoid the
analysis and can be also used to delineate P and T waves. introduction of confounding factors in the analysis, the former
It was implemented in Matlab, and its performance was test was performed only on non-stretchable samples (cotton),
assessed on the same data sets used in [57] for functional whereas the latter test was carried out only on the stretchable
validation. samples.
In this study, the aforementioned QRS features were com- Fig. 2a shows the variation of the sheet resistance (average
puted as the average of those extracted from the individual values over six samples, standard deviation and linear regres-
beats that were detected and delineated by the algorithm over sion) after each washing cycle, measured with the LCR meter.
a 60-s segment of the ECG signal. In Fig. 2b, the variation of the noise and impedance
parameters, measured with the Xtratek ET-65A according to
the ANSI/AAMI EC12:2000/(R)2010 standard, by coupling
III. R ESULTS
the dried pairs of washed electrodes, is presented.. For each
A. Benchtop Measurements washing cycle condition, a single couple of electrodes was
1) Sheet Resistance: The production of screen-printed elec- tested.
trodes involves treating a sheet of fabric with the PEDOT:PSS Stretching tests under dry conditions revealed an increment
solution to obtain a conductive textile. The electrical sheet in the impedance and noise values, respectively increasing
resistance values of the screen-printed fabrics were (mean ± from 243 ± 1  to 670 ± 30  and from 48 ± 2 μ V to
standard deviation) 29 ± 3 /sq for the cotton fabric and 139 ± 2 μV after one stretching cycle of 100% in length.
35 ± 8 /sq for the stretchable substrate. Despite the deterioration, the impedance was still within
2) Xtratek Electrode Tester: In addition to sheet the ANSI/AAMI limits, whereas the noise slightly exceeded
resistance, the parameters required by the ANSI/AAMI the limits. Even after repeating this stretching several times,
EC12:2000/(R)2010 standard for disposable ECG no significant variations were obtained with respect to just one
electrodes [48], whose limits are reported in TABLE 1, cycle.
were analysed. Impedance (ACZ), noise and SDR were 4) Electrode Benchtop Impedance Measurements: The
analysed under the controlled conditions described in results of electrode impedance are shown in Fig. 3. As
Section II.B.2) under dry conditions and with the addition indicated by the impedance results obtained with the Xtratek
of electrolytes. The two former parameters are shown in ET-65A, the response to the addition of different electrolytes
TABLE 2 for the cotton samples. The value indicated showed a similar trend over the different frequencies.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 4102 IEEE SENSORS JOURNAL, VOL. 18, NO. 10, MAY 15, 2018

Fig. 2. (a) Sheet resistance. (b) Impedance and noise for washed electrodes Fig. 4. (a) Skin contact impedance measured before every ECG measurement.
under dry condition. (b) Impedance measured with the addition of electrolytes and compared with
that of Ag/AgCl electrodes.

TABLE III
P EARSON ’ S C ORRELATION C OEFFICIENTS O BTAINED BY C OMPARING THE
T EXTILE E LECTRODES W ITH THE A DDITION OF E LECTROLYTES AND
THE C OMMERCIAL G ELLED E LECTRODES

allow the assessment of the electrode performance in real

working conditions.
1) Skin Contact Impedance: The skin contact impedance
values are shown in Fig. 4a, in terms of their mean and
Fig. 3. Impedance benchtop measurements using LCR meter.
standard deviation over the study population. To better identify
the contribution of the electrolytes to this parameter, the same
plot is shown in Fig. 4b without the curve related to the dry
B. Human Tests condition. These curves exhibit a similar trend, as revealed by
The previous results revealed that the proposed electrodes the Pearson’s correlation coefficients presented in TABLE 3.
could be used as ECG sensors in applications that do not Remarkably, such coefficients present an error lower
require defibrillator proof electrodes. The following results than 10−13.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 ACHILLI et al.: DESIGN AND CHARACTERIZATION OF SCREEN-PRINTED TEXTILE ELECTRODES FOR ECG MONITORING 4103

Fig. 7. QRS features obtained with textile electrodes with added saline (S),
Fig. 5. ECG signal recorded with different electrolytes with saline solution hydrogel (H), compared with disposable Ag/AgCl electrodes (Ag).
(top), hydrogel (center) and Ag/AgCl electrodes (bottom).

Fig. 8. ECG signal recorded with textile electrodes washed 20 times under
dry condition (top) and saline solution (bottom).
Fig. 6. ECG power spectral density Welch’s estimations with saline solution
(top), hydrogel (center) and Ag/AgCl electrodes (bottom).
Based on the lower skin contact impedance exhibited by the
dry textile electrodes that underwent 20 washing cycles, it is
possible to record an ECG signal even under dry condition.
The textile electrodes that were washed 20 times showed
A comparison between the signal recorded with the washed
a higher skin contact impedance compared to that of the
electrodes under dry condition and that with the addition of
brand new electrodes when the saline solution was used as
saline solution is presented in Fig. 8.
the electrolyte (134 ± 13 k), whereas under dry conditions,
the impedance was lower (313 ± 175 k).
2) ECG Analysis: The performance of the screen-printed IV. D ISCUSSION
textile electrodes for ECG acquisition was assessed by Sheet resistance has an important role in the evaluation of
comparing them with disposable gelled Ag/AgCl electrodes the electrical properties of the functionalised fabric and is
(CDES000024 by SpesMedicaSrl). The results presented in strongly dependent on the conductivity of the PEDOT:PSS
this section are related to a single subject because of the ink. However, the quality of the physiological signals that
physiological variability of the signals and, consequently, can be acquired using the electrodes is marginally affected by
of the extracted features. Because of the excessively high value this parameter [14], [17], [34]. In fact, the signal quality also
of the skin contact impedance under dry conditions, only the depends on other parameters that will be investigated hereafter.
results with an electrolyte are reported hereafter. Three ECG At first, the results achieved in controlled conditions, i.e. by
excerpts obtained with brand new textile electrodes (saline and directly coupling the conductive surfaces of the two electrodes,
hydrogel) and Ag/AgCl electrodes are shown in Fig. 5. getting rid of the variability introduced by the coupling with
The signals appear very similar, as also demonstrated by the skin, are discussed.
the PSD analysis, reported in Fig. 6. The results presented in TABLE 2 clearly show how the
The QRS features (RR interval, QRS amplitude and QRS electrolyte increases both the impedance and the noise, dou-
duration), extracted from the signals of the same subject over bling the values obtained under dry condition. It should be
60 s, are presented in Fig. 7. noted that these results are within the limits imposed by

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 4104 IEEE SENSORS JOURNAL, VOL. 18, NO. 10, MAY 15, 2018

the standard, except for the noise level measured with the of the skin and its better coupling with the functionalised
hydrogel. The effect of the electrolytes on the impedance is fabric [18]. Remarkably, the skin contact impedance without
confirmed by the analysis over different frequencies in the any electrolyte is strongly affected by the skin characteristics
typical ECG band, presented in Fig. 3. In this figure, the fluc- of the different subjects [34], also considering that no skin
tuation below 100 Hz is partially due to the lower electrical preparation was performed to simulate a real application
mobility of organic semiconductors at low frequencies [58]. scenario. This aspect, along with the limited validity of the
Moreover, the difference in the impedance can be explained assumptions behind the skin contact impedance estimation
by the increase in the capacitive component introduced by model, hamper a detailed analysis of the curve shape that,
the addition of an electrolyte layer in between the two-faced conversely, could be easily performed in presence of elec-
electrodes. In both cases (TABLE 2 and Fig. 3), the impedance trolytes (Fig. 4b). A decrease of up to two orders of magnitude
is higher when the hydrogel layer is applied, mainly because of between the values recorded with electrolytes and the values
the physical separation between the electrodes surfaces. This without electrolytes under the same measurement conditions
condition also influenced other parameters, such as the DC was recorded, leading to a performance close to that achievable
offset, since the hydrogel layer leads to an ionic charge distri- with commercial electrodes. Remarkably, this behaviour is
bution at the interface with the polymer [59], were interactions opposite to what was observed in benchtop measurements,
between the solid hydrogel polymer and the PEDOT:PSS can where the effect of the signal conduction through the skin
occur. was not present. As shown in TABLE 3, a strong statistical
Even though textile electrodes are aimed at the development correlation was observed between the skin contact impedance
of smart garments, and not for clinical use, we also performed curves obtained with Ag/AgCl electrodes and textile electrodes
the analyses related to the simulated defibrillation recovery. with addition of hydrogel and saline.
The obtained results (Section III.A.2)) are in agreement with In addition to the aforementioned conditions, there are
the limits of the organic polymers, which present a threshold several other aspects related to the production process that
voltage of about 1.6 V for a non-reversible polarisation [24]. can affect the skin contact impedance. This makes it unfair
In this case, the application of the defibrillation discharge any detailed comparison with other studies in the scientific
irreparability polarises the electrodes, making them unusable, literature. First, the polymer-based ink could be composed
so that the other limits imposed by the standard are no more of different PEDOT:PSS formulations commercially (Heraeus
respected. CleviosTM PH500, CleviosTM PH1000, etc.), characterized
The simulation of the daily use conditions (washing and by a different conductivity. Furthermore, to produce the ink,
stretching) showed interesting results. The variation in sheet different secondary dopants and other surfactants are typically
resistance due to the repeated washing procedure (Fig. 2a) added to improve the conductivity and change the ink viscos-
could be approximated by a linear trend with a 99% con- ity. This, in turn, also influences the amount of polymer that is
fidence. Again, this parameter (whose slope was evaluated effectively transferred to the fabric. Moreover, the deposition
as 4.0 ± 0.2 /(sq· cycle)) cannot completely describe the processes have a strong influence on this aspect. Nevertheless,
electrode performance. In fact, both impedance and noise the skin-contact impedance results are comparable with those
(Fig. 2b) after 10 washing cycles were stabilised, reaching the reported in [17], where inkjet printing on a commercially
respective values of about 0.9 k and 230 μV. Remarkably, available pantyhose (100 wt% polyamide) was used, only
the dry washed electrodes showed a skin contact impedance when eight layers of PEDOT:PSS and five layers of liquid
compatible with the acquisition of an ECG signal, as clearly ionic gel were deposited on the fabric. Also direct patterning
depicted in Fig. 8. It is conceivable that this effect could be of PEDOT:PSS on fabric, with the same ionic gel, presents
due to the changes in the morphology of the polymer chains similar values for this parameter [13]. Compared to [34],
caused by the washing procedure. A rigorous analysis of this a worse impedance is achieved, as clearly demonstrated by
phenomenon is hampered by the irregular surface of the textile the acceptable performance of the dry electrodes in that work.
substrate. However, in [34] a mild skin treatment was performed, and
Stretching tests revealed a high increment in the impedance the electrodes were produced by dip coating, leading to a
and noise values compared to that with the brand new elec- significantly higher amount of polymer deposited on the fabric,
trodes, reaching 250% of their starting values. This is due to compared to screen-printing. The importance of this aspect is
polymer film cracking on the electrode surface caused by the evident when considering the number of layers that must be
physical separation of the fibres, which permanently breaks printed on the fabric to achieve a good conductivity, when
the polymer chains. Nevertheless, the screen-printed textile using inkjet printing [17].
electrodes can be used for producing smart technical sports- The results reported in Section 0 show the comparison
wear, provided that special care is taken to avoid excessive between disposable gelled Ag/AgCl electrodes and the screen-
stretch forces that could irreparably damage the conductive printed electrodes with addition of electrolytes. These results
surface. confirm the possibility of using the proposed textile electrodes
After benchtop measurement characterization, we evaluated as ECG sensors, since the signal quality and the spectral
the features of these screen-printed electrodes as ECG sensors analyses reveal high similarity with the commercial Ag/AgCl
on human subjects. electrodes.
The skin contact impedance results decreased with addition From a spectral perspective, the PSD values obtained using
of different electrolytes (Fig. 4a) because of the hydration the different electrodes overlap. This is reasonable in the

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 ACHILLI et al.: DESIGN AND CHARACTERIZATION OF SCREEN-PRINTED TEXTILE ELECTRODES FOR ECG MONITORING 4105

light of the impedance behaviour of the different electrodes approach, such inks present a higher conductivity. However,
(Fig. 4b). Furthermore, the estimated PSD values are anal- for wearable applications, SPEs cannot be directly realized on
ogous to those obtained in other studies with textile elec- the garment, requiring an assembly procedure. Furthermore,
trodes realised by stainless steel threads twisted around a the plastic substrate is less conformable and breathable than a
standard continuous viscose textile yarn, with addition of solid fabric, leading to reduced comfort and potential skin irritation
hydrogel [14]. in long-term monitoring applications.
Obviously, the quality of the signals recorded with the Inkjet printing was also proposed for the realization of
different electrodes is similar also when looking at the ECG polymer-based textile ECG electrodes. In [17], a Dimatix
morphology in the time domain (Fig. 5). Baseline drift is DMP-2800 inkjet printer was used to deposit a PEDOT:PSS
comparable, even though this is a critical aspect for textile ink on synthetic fabric. The same commercial polymeric
electrodes [41], and the primary waves and the complex of solution adopted in this work was used, but with different
the ECG signal are clearly recognisable without significant additives to adapt the final ink viscosity to the printing process.
distortion. In fact, viscosity influences the ink drops size, which in turn
The QRS morphological features (QRS amplitude and QRS defines the geometrical resolution and the shape reproducibil-
duration) and the RR interval reveal a substantial coherence, ity. The best resolution achievable with that printer, by using
with the variability depending on their physiological changes 1 pL cartridges, is around 20 μm, whereas screen-printing,
over time (60 s), across the different electrodes confirmed, depending on the screen used, can reach only 200 μm. Despite
respectively, by a 96% ± 4% confidence level for saline the better resolution, inkjet printing introduces huge limitations
solution and a 70% ± 8% confidence level for hydrogel in terms of throughput since, even with cartridges by 10 pL,
compared with the results computed using disposable Ag/AgCl the drops deposition is a slow process. In fact, the reduced
electrodes. Such results reveal how the best condition for ECG amount of ink deposited by each drop requires several layers
acquisition with screen-printed textile electrodes is represented (up to 10 in [17]) to provide enough conductivity.
by the addition of saline to the electrodes. In fact, even Conversely, screen-printing allows producing at the same
though the hydrogel addition could help in contrasting motion time, with a single ink layer, several electrodes, even though
artefacts because of the better adhesion to the skin, the characterized by a lower resolution. Notably, such a high
previously discussed higher noise of this solution hampers a resolution is not required for textile ECG electrodes whereas
correct delineation. the production costs, significantly higher for inkjet printing
The novelty of the proposed approach lies in the function- because of the limited throughput, are of paramount impor-
alization procedure used to develop textile ECG electrodes tance.
directly on the fabric, by screen-printing with a biocompatible Finally, screen-printing is a well-known and widely used
conductive polymer. In fact, PEDOT:PSS-based textile elec- printing technique in fashion industry, and could be easily
trodes can be fabricated in several ways, i.e. by coating single adapted to enable smart garments development by replacing
yarns [60], by immersion of the whole fabric [34] or by a the ink, compared to more custom printing procedures [13].
controlled deposition of a polymer solution on the surface
of a finished fabric [13], [17]. Excluding the first approach,
V. C ONCLUSION
which cannot be applied to finished fabrics, a major difference
exists between the other two. In fact, in the second process, This study presented a novel approach for the development
the complete encapsulation of all the fibres is achieved without of PEDOT:PSS-based textile electrodes for ECG by screen
any control on the polymer amount, whereas in the third printing. A complete characterization, including benchtop
approach, the amount of polymer deposited on the fabric measurements and human tests, was carried out, comparing the
surface is finely controlled. electrode performance with that of disposable gelled Ag/AgCl
However, by treating only the fabric surface, the volume commercial electrodes and taking into account the recom-
of PEDOT:PSS transferred to the fabric by screen printing is mendations of the ANSI/AAMI EC12:2000/(R)2010 standard
several times lower than the amount transferred by immersion, for disposable gelled ECG electrodes. The results revealed
where a bulk of conductive material is obtained. This is why the possibility of using them as ECG sensors, except when
the skin contact impedance under dry condition for screen- defibrillation proof properties are required. Due to the limited
printed electrodes is higher, thereby hampering the ECG amount of conductive polymer deposited onto the fabric sur-
acquisition [34]. face, the exploitation under dry conditions is also limited by
Nevertheless, by the controlled deposition of the ink, screen the high skin contact impedance. The conductive properties
printing allows a good reproducibility, a high geometrical are preserved after several washing cycles, even though the
resolution and the possibility to functionalise components of ANSI/AAMI limits are exceeded. On stretchable fabrics, spe-
finished garments, enabling the realisation of seamless smart cial care should be taken to avoid damaging the electrodes by
textiles. Using the polyester mesh applied in this study, with applying a significant level of stretch.
43 lines/cm, it is possible to achieve a geometrical resolution The simplicity and high reproducibility, in terms of the
of 0.3 mm. geometry and electrical properties, of the proposed fabrication
Previous works on SPEs adopted the same technique but process pave the way to the possibility of exploiting this
with a different ink (such as carbon-based [40] or silver- approach for the development of seamless smart garments
based [39]) on a plastic substrate. Compared to the proposed without the limitations imposed by other approaches. The soft

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 4106 IEEE SENSORS JOURNAL, VOL. 18, NO. 10, MAY 15, 2018

touch of the material on the skin also enables a comfortable [21] M. G. Honarvar and M. Latifi, “Overview of wearable electronics and
use, thereby avoiding the stress on the skin due to metallic smart textiles,” J. Textile Inst., vol. 108, no. 4, pp. 631–652, 2017.
[22] T. Yamashita, S. Takamatsu, K. Miyake, and T. Itoh, “Antistiction
elements or seams. Clinical trials on healthy and pathological technique using elastomer contact structure in woven electronic textiles,”
populations have been scheduled to investigate the applicabil- Jpn. J. Appl. Phys., vol. 53, no. 4S, p. 04EK03, 2014.
ity for ECG monitoring with clinical accuracy. [23] T. Yamashita, S. Takamatsu, K. Miyake, and T. Itoh, “Fabrication and
evaluation of a conductive polymer coated elastomer contact struc-
ture for woven electronic textile,” Sens. Actuators A, Phys., vol. 195,
ACKNOWLEDGMENT pp. 213–218, Jun. 2013.
[24] Y. Wen and J. Xu, “Scientific importance of water-processable
The authors wish to thank Dr. Matteo Mistretta and PEDOT–PSS and preparation, challenge and new application in sensors
Dr. Elena Cotella by Spes Medica Srl for their help on of its film electrode: A review,” J. Polym. Sci. A, Polym. Chem., vol. 55,
electrodes testing with Xtratek ET-65A. no. 7, pp. 1121–1150, 2017.
[25] M. Papaiordanidou, S. Takamatsu, S. Rezaei-Mazinani, T. Lonjaret,
A. Martin, and E. Ismailova, “Cutaneous recording and stimulation of
R EFERENCES muscles using organic electronic textiles,” Adv. Healthcare Mater., vol. 5,
no. 16, pp. 2001–2006, 2016.
[1] S. Stern, “Electrocardiogram: Still the cardiologist’s best friend,” Circu-
[26] A. S. Karimullah, D. R. S. Cumming, M. Riehle, and N. Gadegaard,
lation, vol. 113, no. 19, pp. e753–e756, 2006.
“Development of a conducting polymer cell impedance sensor,” Sens.
[2] J. Crawford and L. Doherty, Practical Aspects of ECG Recording.
Actuators B, Chem., vol. 176, pp. 667–674, Jan. 2013.
Keswick, U.K.: MK Update Ltd, 2012.
[27] M. P. Aleksandrova, G. H. Dobrikov, and G. D. Kolev, “Electrical
[3] T. Dias, Electronic Textiles: Smart Fabrics and Wearable Technology.
characterization of PEDOT:PSS based flexible organic optoelectronic
Amsterdam, The Netherlands: Elsevier, 2015.
devices,” Int. J. Tech. Phys. Problems Eng., vol. 8, no. 26, pp. 71–76,
[4] T. Page, “A forecast of the adoption of wearable technology,” Int.
Mar. 2016.
J. Technol. Diffusion, vol. 6, no. 2, pp. 12–29, 2015.
[5] I. Gualandi, M. Marzocchi, A. Achilli, D. Cavedale, A. Bonfiglio, and [28] J. E. Mark, Science and Technology of Polymers and Advanced Materi-
B. Fraboni, “Textile organic electrochemical transistors as a platform for als, vol. 15. New York, NY, USA: Springer, 1999.
wearable biosensors,” Sci. Rep., vol. 6, Sep. 2016, Art. no. 33637. [29] A. Elschner, S. Kirchmeyer, W. Lovenich, U. Merker, and K. Reuter,
[6] M. Amjadi, K.-U. Kyung, I. Park, and M. Sitti, “Stretchable, skin- PEDOT: Principles and Applications of an Intrinsically Conductive
mountable, and wearable strain sensors and their potential applications: Polymer. Boca Raton, FL, USA: CRC Press, 2010.
A review,” Adv. Funct. Mater., vol. 26, no. 11, pp. 1678–1698, 2016. [30] S. H. Eom et al., “Polymer solar cells based on inkjet-printed
[7] P. Kumari, L. Mathew, and P. Syal, “Increasing trend of wearables PEDOT:PSS layer,” Organic Electron., vol. 10, no. 3, pp. 536–542, 2009.
and multimodal interface for human activity monitoring: A review,” [31] S. Garreau, G. Louam, S. Lefrant, J. P. Buisson, and G. Froyer, “Optical
Biosensors Bioelectron., vol. 90, pp. 298–307, Apr. 2017. study and vibrational analysis of the poly (3,4-ethylenedioxythiophene)
[8] T. Q. Trung and N.-E. Lee, “Flexible and stretchable physical sensor (PEDT),” Synth. Met., vol. 101, nos. 1–3, pp. 312–313, 1999.
integrated platforms for wearable human-activity monitoringand per- [32] A. N. Aleshin, R. Kiebooms, and A. J. Heeger, “Metallic conductivity of
sonal healthcare,” Adv. Mater., vol. 28, no. 22, pp. 4338–4372, 2016. highly doped poly(3,4-ethylenedioxythiophene),” Synth. Met., vol. 101,
[9] S. Takamatsu, T. Lonjaret, E. Ismailova, A. Masuda, T. Itoh, and nos. 1–3, pp. 369–370, 1999.
G. G. Malliaras, “Wearable keyboard using conducting polymer elec- [33] Z. A. King, C. M. Shaw, S. A. Spanninga, and D. C. Martin,
trodes on textiles,” Adv. Mater., vol. 28, no. 22, pp. 4485–4488, 2016. “Structural, chemical and electrochemical characterization of poly(3,4-
[10] V. Lebedev et al., “Investigation of sensing capabilities of organic bi- Ethylenedioxythiophene) (PEDOT) prepared with various counter-ions
layer thermistor in wearable e-textile and wireless sensing devices,” and heat treatments,” Polymer, vol. 52, no. 5, pp. 1302–1308, 2011.
Organic Electron., vol. 42, pp. 146–152, Mar. 2017. [34] D. Pani, A. Dessì, J. F. Saenz-Cogollo, G. Barabino, B. Fraboni, and
[11] S. Choi, H. Lee, R. Ghaffari, T. Hyeon, and D.-H. Kim, “Recent A. Bonfiglio, “Fully textile, PEDOT:PSS based electrodes for wearable
advances in flexible and stretchable bio-electronic devices integrated ECG monitoring systems,” IEEE Trans. Biomed. Eng., vol. 63, no. 3,
with nanomaterials,” Adv. Mater., vol. 28, no. 22, pp. 4203–4218, 2016. pp. 540–549, Mar. 2016.
[12] I. Gualandi, D. Tonelli, F. Mariani, E. Scavetta, M. Marzocchi, and [35] Y. Kim, H. Kim, and H. J. Yoo, “Electrical characterization of screen-
B. Fraboni, “Selective detection of dopamine with an all PEDOT:PSS printed circuits on the fabric,” IEEE Trans. Adv. Packag., vol. 33, no. 1,
organic electrochemical transistor,” Sci. Rep., vol. 6, Oct. 2016, pp. 196–205, Feb. 2010.
Art. no. 35419. [36] T. Kellomäki, J. Virkki, S. Merilampi, and L. Ukkonen, “Towards
[13] S. Takamatsu, T. Lonjaret, D. Crisp, J.-M. Badier, G. G. Malliaras, washable wearable antennas: A comparison of coating materials for
and E. Ismailova, “Direct patterning of organic conductors on knitted screen-printed textile-based UHF RFID tags,” Int. J. Antennas Propag.,
textiles for long-term electrocardiography,” Sci. Rep., vol. 5, Oct. 2015, vol. 2012, Sep. 2012, Art. no. 476570.
Art. no. 15003.
[37] I. Locher and G. Tröster, “Screen-printed textile transmission lines,”
[14] E. P. Scilingo, A. Gemignani, R. Paradiso, N. Taccini, B. Ghelarducci,
Textile Res. J., vol. 77, no. 11, pp. 837–842, 2007.
and D. De Rossi, “Performance evaluation of sensing fabrics for
monitoring physiological and biomechanical variables,” IEEE Trans. Inf. [38] K. Jost et al., “Knitted and screen printed carbon-fiber supercapacitors
for applications in wearable electronics,” Energy Environ. Sci., vol. 6,
Technol. Biomed., vol. 9, no. 3, pp. 345–352, Sep. 2005.
no. 9, pp. 2698–2705, 2013.
[15] M. K. Yapici, T. Alkhidir, Y. A. Samad, and K. Liao, “Graphene-clad
textile electrodes for electrocardiogram monitoring,” Sens. Actuators B, [39] L. Rattfält, F. Björefors, D. Nilsson, X. Wang, P. Norberg, and P. Ask,
Chem., vol. 221, pp. 1469–1474, Dec. 2015. “Properties of screen printed electrocardiography smartware electrodes
[16] Y.-H. Chen et al., “Soft, comfortable polymer dry electrodes for investigated in an electro-chemical cell,” Biomed. Eng. OnLine, vol. 12,
high quality ECG and EEG recording,” Sensors, vol. 14, no. 12, p. 64, Jul. 2013.
pp. 23758–23780, 2014. [40] A. A. Chlaihawi, B. B. Narakathu, A. Eshkeiti, S. Emamian,
[17] E. Bihar et al., “Fully printed electrodes on stretchable textiles for long- S. G. R. Avuthu, and M. Z. Atashbar, “Screen printed MWCNT/PDMS
term electrophysiology,” Adv. Mater. Technol., vol. 2, no. 4, p. 1600251, based dry electrode sensor for electrocardiogram (ECG) measurements,”
2017. in Proc. IEEE Int. Conf. Electro Inf. Technol., May 2015, pp. 526–529.
[18] H. Zhou et al., “Stimulating the comfort of textile electrodes in [41] G. Paul, R. Torah, S. Beeby, and J. Tudor, “The development of screen
wearable neuromuscular electrical stimulation,” Sensors, vol. 15, no. 7, printed conductive networks on textiles for biopotential monitoring
pp. 17241–17257, 2015. applications,” Sens. Actuators A, Phys., vol. 206, pp. 35–41, Feb. 2014.
[19] P. Shyamkumar, P. Rai, S. Oh, M. Ramasamy, R. Harbaugh, and [42] O. P. Dimitriev, D. A. Grinko, Y. V. Noskov, N. A. Ogurtsov, and
V. Varadan, “Wearable wireless cardiovascular monitoring using textile- A. A. Pud, “PEDOT:PSS films—Effect of organic solvent additives and
based nanosensor and nanomaterial systems,” Electronics, vol. 3, no. 3, annealing on the film conductivity,” Synth. Met., vol. 159, nos. 21–22,
pp. 504–520, 2014. pp. 2237–2239, 2009.
[20] I. G. Trindade et al., “Design and evaluation of novel textile wearable [43] Z. Taleat, A. Khoshroo, and M. Mazloum-Ardakani, “Screen-printed
systems for the surveillance of vital signals,” Sensors, vol. 16, no. 10, electrodes for biosensing: A review (2008–2013),” Microchim. Acta,
p. 1573, 2016. vol. 181, nos. 9–10, pp. 865–891, 2014.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.
 ACHILLI et al.: DESIGN AND CHARACTERIZATION OF SCREEN-PRINTED TEXTILE ELECTRODES FOR ECG MONITORING 4107

[44] J. Kim, R. Kumar, A. J. Bandodkar, and J. Wang, “Advanced materials [59] G. Baura, Medical Device Technologies. Cambridge, MA, USA:
for printed wearable electrochemical devices: A review,” Adv. Electron. Elsevier, 2012.
Mater., vol. 3, no. 1, p. 1600260, 2017. [60] S. Tsukada, H. Nakashima, and K. Torimitsu, “Conductive polymer
[45] M. Li, Y.-T. Li, D.-W. Li, and Y.-T. Long, “Recent developments combined silk fiber bundle for bioelectrical signal recording,” PLoS
and applications of screen-printed electrodes in environmental assays— ONE, vol. 7, no. 4, p. e33689, 2012.
A review,” Anal. Chim. Acta, vol. 734, pp. 31–44, Jul. 2012.
[46] J. Löfhede, F. Seoane, and M. Thordstein, “Textile electrodes for eeg
recording—A pilot study,” Sensors, vol. 12, no. 12, pp. 16907–16919,
2012. Andrea Achilli received the M.Sc. degree in physics
[47] M. Kisan, S. Sangathan, J. Nehru, and S. G. Pitroda, Textiles—Tests for from the University of Bologna, Italy, in 2015. He is
Colour Fastness—Part C10: Colour Fastness to Washing With Soap or currently pursuing the Ph.D. degree in bioengineer-
Soap and Soda, ISO Standard 105.C10.2006, 2006. ing and robotics with the University of Genova, and
[48] Disposable ECG Electrodes, document ANSI/AAMI EC12:2000, he is part of the Advanced Electronic Devices Group,
ANSI/AAMI, Arlington, VA, USA, 2000. Department of Electrical and Electronic Engineer-
[49] H.-J. Yoo and C. van Hoof, Bio-Medical CMOS ICs, vol. 53. New York, ing, University of Cagliari, Italy. His research inter-
NY, USA: Springer, 2013. ests are in the field of biosensors on flexible and
[50] J. W. Mason, E. W. Hancock, and L. S. Gettes, “Recommendations for textile substrates, including electrodes for biopoten-
the standardization and interpretation of the electrocardiogram: Part II: tial recordings and biomechanical sensors based on
Electrocardiography diagnostic statement list. A scientific statement OECT structures.
from the American Heart Association Electrocardiography and Arrhyth-
mias Committee, Council on Clinical Cardiology; the American College
of Cardiology Foundation; and the Heart Rhythm Society. Endorsed by
the International Society for Computerized Electrocardiology,” J. Amer.
College Cardiol., vol. 49, no. 10, pp. 1128–1135, 2007. Annalisa Bonfiglio received the degree in
[51] P. R. Rijnbeek, J. A. Kors, and M. Witsenburg, “Minimum bandwidth physics from the University of Genova, Italy,
requirements for recording of pediatric electrocardiograms,” Circulation, in 1991, and the Ph.D. degree in bioengineering
vol. 104, no. 25, pp. 3087–3090, 2001. from the Politechnical School, Milan, in 1996.
[52] R. E. Mason and I. Likar, “A new system of multiple-lead exercise She is currently a Full Professor of Electronic
electrocardiography,” Amer. Heart J., vol. 71, no. 2, pp. 196–205, 1966. Bioengineering with the University of Cagliari.
[53] J. Malmivuo and R. Plonsey, Bioelectromagnetism: Principles and Appli- She authored more than 130 papers in international
cations of Bioelectric and Biomagnetic Fields. Oxford, U.K.: Oxford journals, conference proceedings, and book chapters.
Univ. Press, 2012. She also holds 9 patents. Her research activity
[54] V. Marozas, A. Petrenas, S. Daukantas, and A. Lukosevicius, “A compar- is focused on innovative materials (in particular
ison of conductive textile-based and silver/silver chloride gel electrodes organic semiconductors) and devices for wearable
in exercise electrocardiogram recordings,” J. Electrocardiol., vol. 44, electronics and biomonitoring.
no. 2, pp. 189–194, 2011.
[55] L. Beckmann et al., “Characterization of textile electrodes and conduc-
tors using standardized measurement setups,” Physiol. Meas., vol. 31,
no. 2, p. 233, 2010. Danilo Pani (S’05–M’07) received the Ph.D. degree
[56] P. D. Welch, “The use of fast Fourier transform for the estimation of from the University of Cagliari, Italy. He is tenure-
power spectra: A method based on time averaging over short, modified track Assistant Professor of Bioengineering with
periodograms,” IEEE Trans. Audio Electroacoust., vol. AU-15, no. 2, the Department of Electrical and Electronic Engi-
pp. 70–73, Jun. 1967. neering, University of Cagliari, Italy. His research
[57] J. P. Martinez, R. Almeida, S. Olmos, A. P. Rocha, and P. Laguna, interests are in the field of real-time digital process-
“A wavelet-based ECG delineator evaluation on standard databases,” ing of bio-signals and medical devices. His main
IEEE Trans. Biomed. Eng., vol. 51, no. 4, pp. 570–581, Apr. 2004. achievements regard fetal ECG, textile electrodes,
[58] J. Rivnay et al., “Structural control of mixed ionic and electronic real-time embedded processing of signals from the
transport in conducting polymers,” Nat. Commun., vol. 7, Apr. 2016, peripheral nervous systems for motor neuroprosthe-
Art.no. 11287. ses, and telehealth systems.

Authorized licensed use limited to: ULAKBIM UASL ISTANBUL TEKNIK UNIV. Downloaded on January 05,2024 at 07:43:16 UTC from IEEE Xplore. Restrictions apply.