You are on page 1of 5

Nicole Teresa Furtak,

Ewa Skrzetuska,
Development of Screen-Printed Breathing
Izabella Krucińska Rate Sensors
Department of Material and Commodity Sciences Abstract
and Textile Metrology, This paper presents a method of screen-printing a conductive carbon nanotube printing
Lodz University of Technology, paste onto textile substrates to prepare textronic strain sensors for measuring breathing
ul. Żeromskiego 116, 90-924 Łódź, Poland. rate. Screen-printed sensors can be integrated with garments after construction and do not
Email: require significant modification of the construction process. Personalisation of the garment
is used to optimize sensor placement for improved measurement accuracy. Changes in the
electrical resistance of the sensor as a result of changes in strain are detected and used to
determine the breathing rate.

Key words: textronics, textile sensor, breathing rate, carbon nanotubes, screen-printing.

n Introduction garment dimensions and a further change

in resistance that is used to determine the
Textronics can be used to create wireless, breathing rate.
wearable systems to monitor physiologi-
cal parameters in daily activities [1]. Sensors of this type, however, require the
The breathing rate is an important physi- introduction of electrically conductive
ological parameter that can be monitored yarns into the garment’s structure using
[1 - 3] by measuring mechanical signal insertion techniques (using weft or warp
[4]. These low frequency mechanical sig- thread or sewing). The use of printing is
nals are generated by movements of the
advantageous in sensor construction be-
human body while breathing and can be
cause it does not require intervention in
measured using strain gauges [1]. Com-
the knitting process, but instead can be
mercial piezoelectric sensors can be used
applied to a finished shirt. Sensors can
to ascertain the breathing rate [5, 6] and
be printed using traditional printing tech-
garments made with textronics can be
niques, such as screen-printing, as well
used to distribute the sensors and elec-
as modern techniques, such as jet-print-
trodes optimally on the body [4]. Gar-
ing. A previous work described a printed
ments with sensors are designed to fit the
sensor for measuring the respiratory rate
body so that the garment is subjected to a
made in the form of a belt surrounding
moderate amount of pressure [7].
the chest [12].
Textronic sensors can monitor vital
Current printing with conductive inks
signs, such as the breathing rate, with-
out hindering comfort or performance. has been used to create electric circuits
Precise sensor placement is important on a textile surface, and techniques from
for monitoring physiological processes this approach can be utilised to print an
and personalised garment construction elongation strain sensor. Direct printing
can ensure ideal sensor placement [4,8]. is one method of printing conductive ink
Textronic sensors that do not require spe- whereby a mask is used to cover areas
cial methods of garment construction are that are not to be printed [8]. Screen-
beneficial for sensor integration in gar- printing is one type of direct printing
ments. that has been utilised on materials such
as nonwovens [9]. Another method to ap-
This paper presents a sensor that is based ply a conductive circuit is to use ink jet
on the detection of changes in resist- printing to apply the conductive ink [10].
ance, due to changes in the dimensions Most conductive inks contain metallic
of textiles (strain). Other sensors made nanoparticles, but carbon nanotubes in
with electrically conductive yarn incor- an aqueous dispersion can also be used
porated directly into the fabric are cur- [11, 12].
rently implemented and detect a change
in resistance, constructed [3]. One meth- By incorporating electronic functions
od of using this type of sensor is to in- into traditional textiles, increased func-
sert a band that wraps around the chest tionality is possible. Textronics allow
incorporated into a textronic shirt [3]. physiological signal monitoring during
Deformations of the human body while everyday activity, using wireless and
breathing subsequently cause a change in wearable systems [1]. Functional materi-

84 Furtak NT, Skrzetuska E, Krucińska I. Development of Screen-Printed Breathing Rate Sensors.

FIBRES & TEXTILES in Eastern Europe 2013; 21, 6(102): 84-88.
als allow the creation of garments with Table 1. Textile substrate properties; 1. PES (Polyester), PU (Polyurethane), PA (Polyam-
distributed sensors and electrodes [4]. ide).
Textronic systems must be able to sense, Mass per unit Thickness, Apparent density,
Name Fibre content1 Structure
as well as pre-process, transmit, process area, g/m2 mm kg/m3
and manage data [1]. Smart fabrics can Shirt A
81 % PES
warp knit 205 0.49 418.37
19 % PU
act as flexible circuit boards, with sensors 64% PA warp knit with
Shirt B 245 0.41 597.56
that record electronic signals and send 36% elastane weft insertion
them to a base station (via conductive
pathways) for processing [1]. The nanotube dispersion was 90% pure The main goal of the research performed
with a carbon nanotube molecular weight was determining the sensors sensitivity
To build a correct system to measure the
of 0.5 - 1.5%. The average nanotube to mechanical deformations. The kinetics
frequency of breathing, it is necessary to
length was 1.5 µm with a diameter of of the change in surface resistance under
prepare a measuring sensor, equipment
9.5 nm. The solution was thoroughly the influence of the action in stress used
that processes the signal, and a monitor-
mixed, and excess water was removed by was recorded with the use of the Keithley
ing system that converts the signal col-
gravity drainage through filter paper for multimeter.
lected into readable information. The
possibility to send the information via an approximately 10 minutes. A squeegee
and printing screen were used to apply Constant conditions of conditioning and
antenna to part of the monitoring system
printing paste to the fabric material of researching the sample were maintained:
exists for each of the systems studied.
the personalised shirts. After printing, the a temperature of 23 °C, RH = 65% [12].
A truly intelligent material would allow
sensing and processing while providing materials were put in a 30 °C heat cham-
To fabricate personalised shirts from
instantaneous feedback to the user [16]. ber and then placed under UV light with
each substrate material, a test subject
In addition to electronic functionality, the a dosage of 3.5 J/cm2 for cross-linking. was scanned using a Model NX-16 [TC]2
textile must be wearable with a high de- body scanner to determine their measure-
gree of comfort. To maintain comfort, it Specimens were prepared for preliminary ments. Due to technological limitations,
is important that the measuring devices tensile testing by cutting 6 cm × 55 cm the body scan was not used to generate a
are of a small size [15]. strips from the textile substrate materi- custom pattern; however, the bust meas-
als, with each containing a 20 cm length urement from the body scan was used to
The authors of this paper focused only on screen-printed band. Electrodes were determine the size required in a standard
the first stage of such a system: a sensor hand sewn using silver conductive yarn shirt pattern. A shirt was then sewn from
devoted to determining changes in re- (Polyamide yarn X-Static (34 filaments each substrate material. The test subject
sistance related to breathing, which can modelled the shirts for custom tailor-
coated with a layer of 15% Ag), with
be assesed per unit of time. ing to personalise the garment fit. While
each filament mass doubled and sewn in
modelling each shirt, the sensor location
two straight lines 4.5 cm apart. An In-
n Methodology desired was marked for preparation of
stron Model 5944 tensile tester was used
screen-printing. Two 5 cm × 5 cm sen-
A specially prepared printing paste doped to cyclically load the sensor for five cy- sors were printed on each shirt. One sen-
with carbon nanotubes, investigated in a cles at each of the following strains: 1.25, sor was applied along the bust line, while
previous work, was used to screen-print 2.5, 5, 7.5, and 10%. The electrodes of the other was placed under the bust line,
ready-made shirts [12]. As breathing oc- the sample were connected to a Keithley as illustrated in Figure 1. Silver conduc-
curs, the garment changes in dimension, Model 2000 digital multimeter to meas- tive thread (doubled) was sewn in two
which causes variations in the resist- ure the resistance during strain testing. different configurations in order to create
ance measured for further calculating the
breathing rate.

Table 1 shows two substrate materials

used to manufacture the shirts, that were
screen-printed. The textile substrates
were selected based on preliminary stud-
ies that assessed their effectiveness for
athletic shirts.

The screen-printing procedure developed

in a previous work was followed [12].
A carbon nanotube paste containing
55.8 ml of aqueous carbon nanotube
dispersion (AquaCyl AQ0101, Nanocyl-
company), 6.25 ml aliphatic urethane
acrylate (Ebecryl2002), and 0.45 ml pho-
toinitiator (Esacure DP250) was mixed. Figure 2. Conductive wire configuration.

FIBRES & TEXTILES in Eastern Europe 2013, Vol. 21, No. 6(102) 85
the connections. The “I” configuration Hysteresis properties of the textile sub- G1 – circumference of the body part (or
consists of two parallel lines of stitches, strate materials were determined with the cylinder), in cm; and
while the “L” configuration has two par- tensile tester by subjecting a 5 cm × 5 cm W – width of the knitted fabric strip, in
allel lines of stitches and a zigzag stitch sample to ten cycles of 20% strain along cm.
extending out from the sensor, as illus- the wale direction, based on previous
trated in Figure 1. The effect of laundering on the ink prop-
work. The pressure of the garment on
the subject tested, which related to the erty is an important consideration and
Measurements were conducted with the
garment fit and elastic properties of the can be determined by comparing initial
subject in a seated position, which is not
textiles used, was calculated based on measurements with those after several
typical for most situations where breath-
the body and garment dimensions and cycles of washing [13].
ing rate measurements are required, but is
practical for the purposes of preliminary measurements. The force versus strain
was plotted for the relaxation portion of To assess the washability of the screen-
research. Electrical connections to the
the final strain cycle, and the relationship printed strain sensor, the garments were
Keithley multimeter and corresponding
lap top computer necessitated the posi- between the force and strain determined subjected to one washing cycle of 40 °C
tioning of the subject in a sedentary state. by curve fitting. The Laplace formula, for 30 minutes according to Standard PN
The resistances recorded were analysed (Equation 1), was used to determine the EN ISO 105-C06:2010. The garments
using the msProcess package in the sta- garment pressure during wearing. were hung to dry in ambient conditions
tistical analysis program. The breathing (20 °C and 65% relative humidity). Pre-
2pF vious research indicated that for up to
rate was determined based on the number P (1)
of local resistance minima detected. For 20 cycles of washing, the conductivity
comparison, the breathing rate was man- where: decrease observed was not significant
ually determined by counting the number F – force in the knitted fabric strip of enough to affect the maximum or mini-
of inhalations over five minutes. width W, in cN; mum peak measurements recorded [12].

Resistance, W
Resistance, W
Strain, %
Strain, %
Resistance, W
Resistance, W

Strain, %
Strain, %

Time, s Time, s

Figure 2. Sensor resistance and strain versus time – Shirt A. Figure 3. Sensor resistance and strain versus time – Shirt B.
Force per distance, cN/cm

Force per distance, cN/cm

Strain, % Strain, %

Figure 4. Force versus strain, Shirt A. Figure 5. Force versus strain, Shirt B.

86 FIBRES & TEXTILES in Eastern Europe 2013, Vol. 21, No. 6(102)
Table 2. Personalised shirt pressures. Table 3. Breathing rate determined from screen-printed sensors.

Pressure, hPa Breathing rate, breaths per minute

Shirt A Shirt B Shirt A Shirt B
Bust 8.75 7.03 Bust Underbust Bust Underbust
Underbust 4.11 5.60 I L I L I L I L
Mean 9.5 12.5 13.6 13.1 11.9 13.1 10.7 14.6
Standard Deviation 0.71 3.16 2.18 1.60 1.29 1.05 1.37 0.80
n Results and discussion Coefficient of Variance 0.07 0.25 0.16 0.12 0.11 0.08 0.13 0.05

There was no noticeable difference in the

application of printing paste and finished bust placements were calculated, shown equations determined by stretching per-
products for shirts A and B. A different in Table 2. The higher pressure of Shirt A pendicularly were used. The key garment
application method, such as direct print- at the bust may account for the more ac- measurements for the breathing rate sen-
ing, may reduce the amount of excess curate breathing rate measurements, but sor are those of the bust circumference
dyestuff present on the final product. further experimentation is also needed. and underbust circumference; the pres-
sure force was calculated in these loca-
Plots of the sensor resistance and strain The graphs were completed after the fifth tions. Results of the calculations can be
over time examined using the tensile measurement cycle. Based on the graphi- seen in Table 2.
tester for Shirts A & B can be seen in cal analysis performed, we can see that
Figures 2 and 3. An Instron tensile tester a quadratic equation is the most accurate Representative plots of the resistance
was used, with the distance between representation of the data experimentally measured while breathing (for the “I”
clamps of 45 cm, and the constant rate obtained for Shirts A and B. The coef- configuration of connections and bust
of movement of 2 mm/min. Even at low ficient of determination, R2, shown for sensor) can be seen in Figures 6 and 7.
strains, the change in sensor resistance each of the fit curves indicates the good- Table 3 shows the mean breathing rate and
with varying strain was sufficiently large ness of fit, with the highest degree of fit statistical analysis for the different sensor
to count the minima using a detection al- for R2 close to one. The force, in cN/cm, configurations. A seated person breath-
gorithm, while higher strains resulted in corresponding to a given strain in % ing normally at a rate of 10.6 breaths
larger, easier-to-detect changes in resist- can be calculated with known validity per minute was determined by observ-
ance. The magnitude of resistance meas- for strains below 20% using a quadratic ing and counting breaths. Comparing
ured is dependent on the textile material. equation, and the coefficients determined results presented in Figures 6 and 7
were based on test work. Strains above especially with baseline values, we can
The pressure of the garment on the wear- 20% can be used, but there is uncertainty see that the “I” sensor configuration pro-
er influences sensor accuracy, hence the about the force obtained above this level. vides more accurate results than the “L”
pressure of the garments used was de- sensor configuration. Shirt A indicates
termined. Figures 4 and 5 show the re- Using the final measurements of the gar- no significant difference in measurement
lationship between force and strain, as ments, the body measurements obtained efficacy between the bust and underbust
well as curves of best fit for the different with the body scan, and the force equa- sensors, while Shirt B shows more accu-
textile substrates. The garment and body tion determined via hysteresis testing, rate results for the bust sensor.
measurements, as well as the quadratic the pressure exerted by the garment on
relationship between force and strain, the body was calculated. As the garments The slight drift in measurements ob-
were used to determine the force for the were constructed so that breathing strains served in Figure 7 may be attributed to
Laplace equation. Using the Laplace the fabric in a direction perpendicular to the movement of the shirt relative to the
equation, pressures at the bust and under- the columns of knit stitches, the force subject’s body. This movement and re-


Resistance, W

Resistance, W




8,800 12,000

0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
Time, s Time, s

Figure 6. Resistance when breathing, Shirt A, Bust configuration Figure 7. Resistance when breathing, Shirt B, Bust configuration
“L”. “L”.

FIBRES & TEXTILES in Eastern Europe 2013, Vol. 21, No. 6(102) 87
lated variation in resistance magnitude Pressure calculated on the basis of the care monitoring applications. Biomedi-
do not have an impact on the breathing Laplace equation from data obtained cal Engineering, Trends in Material Sci-
rate determined as only the number of using sensors on shirt A is higher than ence 2011: 541-546.
peaks is used in the calculation and not that calculated using data from sensors 7. Dunne LE, Brady S, Smyth B, Diamond
the magnitude. on shirt B. It was also observed that in D. Initial development and testing of a
the case of both shirts, pressure obtained novel foam-based pressure sensor for
Preliminary work carried out at the De- from the Laplace equation on the basis of wearable sensing. Journal of NeuroEn-
partment of Material and Commodity results registered by sensors on the bust gineering and Rehabilitation 2005; 2, 4.
Sciences and Textile Metrology indicat- are higher than the same pressure calcu- 8. Lorussi F, Rocchia W, Scilingo EP, Tog-
ed, according to our assumptions that the lated from sensors located at the under- netti A, De Rossi D. Wearable, redun-
sensor resistance varies with temperature bust. dant fabric-based sensor arrays for re-
and humidity changes. However, this construction of body segment posture.
should not have a significant impact on Screen-printing with conductive paste is IEEE Sensors Journal 2004; 4, 6: 807-
the final measuring effect, which depends an effective method of applying a strain 818.
sensor to a garment manufactured us- 9. Merritt CR, Nagle T, Grant E. Fabric-
on the counting of resistance pulses and
ing standard techniques. However, it is based active electrode design and fab-
not the level of the resistance value.
necessary to carry out further work to rication for health monitoring clothing.
Nevertheless in future investigations the
refine the composition of printing paste IEEE Transactions on Information Tech-
boundary limits should be determined
and the method of application for print- nology in Biomedicine 2009; 13, 2: 274-
depending on the threshold of the future
ing. Changes in conductivity and sensi- 280.
counting system.
tivity observed do not cause significant 10. Kim D, Jeong S, Park BK, Moon J. Di-
changes in breathing rate measurement, rect writing of silver conductive patterns:
n Future work an important factor for garments. Screen- Improvement of film morphology and
conductance by controlling solvent com-
printed strain sensors for measuring the
In sporting applications, considerable positions. Applied Physics Letters 2006;
breathing rate show promise for monitor-
laundering is required as the garments 89: 264101-1 - 264101-3.
ing physiological signals.
are exposed to large amounts of sweat. 11. Chiolerio A, et al. Inkjet printing and low-
Further testing to determine the suitabil- er power laser annealing of silver nano-
ity of the printing paste with additional particle traces for the realization of low
laundering cycles, as well as resistance to Acknowledgements resistivity lines for flexible electronics.
perspiration, is required. Other printing The authors would like to thank M. Klonow-
Microelectronic Engineering 2011; 88:
techniques could also be explored to im- ska, PhD from the Department of Knitting 2481-2483.
prove the tactile feel of the screen-print- Technology, Lodz University of Technology, 12. Krucińska I, Skrzetuska E, Urbaniak-
ed sensor, increase aesthetics, and opti- for preparing the shirts. Domagala W. Prototypes of carbon
mise print paste delivery. A quite differ- nanotube-based textile sensors manu-
ent problem is the designing of the whole factured by the screen printing method.
textronic system, which means counting, Fibres and Textiles in Eastern Europe
monitoring and eventual transmitting of
References 2012; 20, 2, 91: 85-89.
the breathing rate signal. 1. Paradiso R. Wearable health care sys- 13. Schwarz A, et al. Electro-conductive
tem for vital signs monitoring. In: 4th and elastic hybrid yarns – The effects of
International IEEE EMBS Special Topic stretching, cyclic straining and washing
n Conclusions Conference on Information Technology on their electro-conductive properties.
Applications in Biomedicine 2003, Bir-
Materials and Design 2011; 32, 4247-
To obtain a personalised garment that mingham, 2003: 283-286.
2. Lymberis A, Dittmar A. Advanced wear- 4256.
fits snugly without being uncomfortably
able health systems and applications. 14. Maklewska E, Nawrocki A, Ledwoń J,
tight, pressure calculations based on the
IEEE Engineering in Medicine and Biol- Kowalski K. Modelling and designing of
Laplace equation are useful. We can see
ogy Magazine 2007: 29-33. knitted products used in compressive
that the pressure varies based on the seg- 3. Frydrysiak M, Zięba J. Textronic Sensor therapy. Fibres & Textiles in Eastern Eu-
ment of the body that is examined, due for Monitoring Respiratory Rhythm. Fi- rope 2006; 14, 5, 59: 111-113.
to the changing circumference. As the bres & Textiles in Eastern Europe 2012; 15. Paradiso R. Wearable health care sys-
pressures calculated for all of the sam- 20, 2, 91: 74-78.
tem for vital signs monitoring. In 4th
ple garments are less than 10 hPa, we 4. Scilingo EP, et al. Performance evalu-
ation of sensing fabrics for monitoring International IEEE EMBS Special Topic
can see that those obtained are comfort- Conference on Information Technology
physiological and biomechanical vari-
able. Clothing elongations greater than Applications in Biomedicine 2003, Bir-
ables. IEEE Transactions on Information
5% observed indicate that the garments Technology in Biomedicine 2005; 9, 3: mingham, 2003, 283-286.
fit closely to the body, as required for a 345-352. 16. Ferri A, Corbellini S, Parvis M, Splen-
breathing rate sensor. 5. De Rossi D, et al. Electroactive fabrics dore R. Compact RF sensors for moni-
and wearable biomonitoring devices. toring thermo-physiological comfort of
Sample A poses better efficiency of de- AUTEX Research Journal 2003; 3, 4:
sportswear. In: 2010 IEEE Instrumenta-
tection of the breathing rate, which is 180-185.
6. Seoane F, Marquez JC, Ferreira J, tion & Measurement Technology Confer-
comparable to real breathing, than sam- Buendia R, Lindecrantz K. The chal- ence, Austin, 2010, 1262-1267.
ple B. Moreover a better relation between lenge of the skin-electrode contact in
the measurement and real breathing was textile-enabled electrical bioimpedance,
obtained for sensors located at the bust. measurements for personalized health- Received 06.08.2012 Reviewed 12.07.2013

88 FIBRES & TEXTILES in Eastern Europe 2013, Vol. 21, No. 6(102)