Sensors and Actuators A 284 (2018) 209–231

Contents lists available at ScienceDirect

Sensors and Actuators A: Physical

journal homepage: [Link]/locate/sna

A review on flexible gas sensors: From materials to devices

Rouba Alrammouz a,b , Jean Podlecki b , Pascale Abboud a , Brice Sorli b , Roland Habchi a,∗
a
EC2M, faculty of science 2, Campus Pierre Gemayel – Fanar, Lebanese University, 90656 Lebanon
b
Institut d’Electronique et des Systèmes (IES), UMR5214, 860 Rue St Priest, Université de Montpellier, 34095 Montpellier Cedex 5, France

a r t i c l e i n f o a b s t r a c t

Article history: This review discusses the recent developments of flexible chemiresistive gas sensors, focusing on the
Received 15 May 2018 main materials and fabrication techniques. This article provides a detailed description of the substrates
Received in revised form 24 October 2018 used in the flexible gas sensing industry. It also highlights the main challenges faced when integrating
Accepted 24 October 2018
a sensing layer on paper, textile or polymeric substrates. A detailed study of the sensing materials in
Available online 26 October 2018
terms of structural properties and sensing mechanisms is presented. This study uses examples of gas
sensors to explain the sensing mechanisms and to emphasize the functionalization effect on the sensing
Keywords:
performance of a material. It also highlights the exceptions in polymers sensing behavior. Since flexible
Flexible
Gas sensor
substrates can’t sustain high temperatures, the main techniques used to fabricate flexible gas sensors are
Chemiresistive explained such as coating, printing, physical vapor deposition, in situ-growth, in situ-polymerization and
Materials chemical vapor deposition. The last technique, which requires high temperatures, is mostly used prior to
Fabrication transferring a sensing layer to a flexible substrate. At the end, the main challenges facing the flexible gas
sensing technology are pointed out.
© 2018 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
2. Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2.1. Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2.1.1. Plastic substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2.1.2. Paper substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
2.1.3. Textiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.2. Sensing material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.2.1. Carbon nanotubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
2.2.2. Conducting polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
2.2.3. Metal oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
2.2.4. Graphene and graphene oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
2.2.5. Metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
3. Device fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
3.1. Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
3.2. Chemical vapor deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
3.3. Physical vapor deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

Abbreviations: WHO, World Health Organization; FET, field effect transistor; PET, polyethylene terephthalate; PEN, polyethylene naphthalate; PI, polyimide; PDMS, poly-
dimethylsiloxane; CNT, carbon nanotubes; SWCNT, single-walled carbon nanotubes; MWCNT, multi-walled carbon nanotubes; FMWCNT, functionalized multi-walled carbon
nanotubes; OLED, organic light emitting diode; OFET, organic field effect transistor; CP, conducting polymer; PANI, polyaniline; PEDOT, poly(3,4-ethylenedioxythiophene);
PSS, polystyrene sulfonated acid; GO, graphene oxide; EDC, 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide; NHS, N-Hydroxysuccinimide; PVP, polyvinylpyrrolidone; PEO,
polyethylene oxide; EC, ethylcellulose; PMS, poly(␣-methylstyrene); IDEs, interdigitated electrodes; DOD, drop-on-demand; R2R, roll to roll; PSSMA, poly(4-styrenesulfonic
acid-co-maleic acid); PAH, poly(allylamine hydrochloride); MPS, 3-mercapto-1-propanesulfonic acid sodium salt; PSS, poly(4-styrenesulfonic acid) sodium salt; Ppy, polypyr-
role; APS, ammonium persulfate; HMTA, hexamethylenetetramine; PEI, poly(ethyleneimine); MOX, metal oxide; PVC, polyvinyl chloride; Cumene-PSMA, cumene terminated
polystyrene-co-maleic anhydride; PSE, poly(styrene-co-maleic acid) partial isobutyl/methyl mixed ester; PVP, polyvinylpyrrolidone.
∗ Corresponding author.
E-mail address: rhabchi@[Link] (R. Habchi).

[Link]
0924-4247/© 2018 Elsevier B.V. All rights reserved.
210 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

3.4. Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223

3.5. Self-assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3.6. In-situ polymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
3.7. In-situ growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
4. Application in wearable devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Biography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

1. Introduction

Ambient air pollution is one of the biggest threats to human

health. Many reports have been published regarding the impact of
air pollution on the burden of diseases such as strokes, lung can-
cers, acute lower respiratory diseases (ALRI), ischemic heart disease
(IHD), chronic obstructive pulmonary disease (COPD) [1,2]. In 2012,
World Health Organization reported that around 3 million deaths
worldwide are caused by atmospheric pollution (Fig. 1) [2]. For all
these reasons, indoor and outdoor air is controlled by a number
of guidelines, laws and air quality standards [3]. Air quality in the
united states is governed by the ‘Clean Air Act’ federal law that
aims to achieve National Ambient Air Quality Standards (NAAQS)
set by EPA (United States Environmental protection agency) [4].
On a European scale, the European commission set a series of
directives that govern the assessment of ambient air quality [5,6]
(Figs. 2 and 3).
The compliance with these laws and regulations requires a con-
stant monitoring of indoor and outdoor air quality. This can only be
achieved using gas sensing systems. These systems can be mounted
indoors, outdoors, or be portable. Many reports have been pub-

Fig. 3. Polymer/SWCNT based gas sensor on fabric substrate for VOC’s detection
[59].

lished regarding the growth of the gas sensors market, which is

expected to reach 3 billion dollars in 2027 [7]. The development
of flexible gas sensing systems is raising a high interest among the
scientific community due to their potential applications in wear-
able and portable electronic products, in RFID tags [8–10], and in
smart packaging for the monitoring of perishable goods [11–13].
Moreover, the techniques used in the flexible gas sensing industry,
such as screen and inkjet printing, enable the large scale fabrica-
tion of low cost effective systems. These techniques also present an
Fig. 1. Wearable electronic nose for armpit odor analysis on plastic substrate [29]. ecofriendly alternative to other techniques, such as etching, that
might produce hazardous chemical waste that needs to be treated
afterwards [14].
Selecting the flexible substrate and the sensing material are the
main challenges faced when developing a flexible gas sensing plat-
form. In fact, some sensing layers might crack when strained or
might separate from the substrate. For all these reasons, research
in this field focuses on optimizing flexible gas sensors in order
to achieve i) sensibilities that are comparable to the sensibilities
achieved with rigid substrates under low operating temperatures,
ii) a high selectivity towards the target analyte, iii) a fast response
and recovery, iv) a good repeatability and stability under bent
and/or stretched and relaxed positions, and v) devices with the abil-
ity to be bent and/or stretched several times without any damage
or performance loss.
Another key factor for developing a flexible gas sensor is its sens-
ing principle, which is directly linked to the target application. The
Fig. 2. Graphene oxide-based gas sensor on paper substrate [57]. main sensing principles used are: chemiresistive [15,16], capaci-
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 211

tive [17], impedancemetric [18], FET [19], mechanical [20–22] and Table 1
Impact of air pollution on the burden of disease in 2012 according to WHO [2].
optical [23,24]. In chemiresistive gas sensing, the resistance of the
sensing layer is proportional to the concentration of the target gas. Disease ALRI COPD Lung cancer Strocke IHD
Capacitive/impedancemetric gas sensing is based on the change of Number of deaths 169 250 242 250 402 350 1 082 750 1 078 800
the dielectric constant/impedance of the sensing layer at high fre- Percentage (%) 5 8 14 36 36
quencies. In FET based sensors, the source-drain current changes
depending on the concentration of the target gas. For mechanical
Table 2
gas sensors, a wave propagates on the surface of the sensing layer. Mechanical properties of PET, PEN and PI.
The exposure of the sensor to a target gas changes the propaga-
tion characteristics of the wave such as its amplitude or frequency. Properties PET PEN PI

These changes are quantified and correlated to a corresponding Friction coefficient 0.2-0.4 0.27 0.42
concentration of the target gas. Another type of mechanical sensors Tensile modulus (GPa) 2-4 5-5.5 2-3
Tensile strength 80 200 70-150
is microcantilever gas sensors. The sensing material is deposited on
a cantilever. The exposure to a target analyte results in a variation
of the mass of the cantilever and thus a shift in its resonance fre- Table 3
quency. An optical gas sensor’s response is reflected in a change Thermal properties of PET, PEN and PI.
in the spectrum of the sensing material. This sensing principle Properties PET PEN PI
is usually based on either absorption spectroscopy, fluorescence
Thermal expansion coefficient (10−6 . K-1 ) 20-80 20-21 30-60
spectroscopy, plasma induced optical emission spectroscopy, Tera Thermal conductivity at 23◦ C (W. m−1 . K−1 ) 0.15-0.4 0.15 0.1-0.35
Hertz spectroscopy or fiber optics. Maximum operating temperature (◦ C) 115-170 155 400
This review focuses on flexible chemiresistive gas sensors. A typ-
ical chemiresitive gas sensor is composed of a substrate, a sensing
Table 4
layer, a pair of electrodes and encapsulation. To ensure the flexibil-
Electrical properties of PET, PEN and PI.
ity of the device, all components must comply with bending to some
degree without any damage or decrease in the sensor’s response. Properties PET PEN PI

This requires a judicious choice of materials and special deposi- Dielectric constant 1MHz 3 3.2 à 10 KHz 3.4
tion techniques. In Section 2 the commonly used substrates and Dissipation factor 1 KHz 0.002 0.005 0.0018
Dielectric strength (KV. mm−1 ) 17 160 à 0.075mm 22
sensing materials in gas sensor fabrication are presented. Plastic,
Surface resistivity (/䊐) 1013 1014 1016
textiles and papers are mainly used as flexible substrates. However, Volume resistivity (.cm) 1014 1015 1018
these substrates have low operating temperatures, which limits the
manufacturing process to low temperature deposition and printing
type techniques. This also limits the performance of metal oxides The substrate’s choice is directly related to the mechanical,
as a sensing layer due to their high operating temperature. Other thermal and electrical constraints of the application. For example,
sensing materials are introduced such as carbon nanotubes, poly- PET and PEN substrates are transparent and low cost. However,
mers, graphene, graphene oxide and metals. In Section 3, we focus they exhibit a low operating temperature (< 150 ◦ C) and thus
on the device fabrication process. Several deposition techniques could not be used as substrates for metal oxide gas sensors, which
used in flexible electronics are discussed in details, such as coat- require high operating temperatures (around 400 ◦ C) [30–32].
ing, chemical vapor deposition, physical vapor deposition, printing, Moreover, plastic substrates exhibit high thermal expansion coef-
self-assembly, in situ-growth and in situ-polymerization. Section 4 ficients (20–80 ppm/◦ C). This makes these substrates unsuitable
concludes this article and highlights the main challenges faced by for low thermal expansion coefficient materials in the case of a
the flexible gas sensing technology. thermal step in the fabrication process. In fact, curing a deposited
layer induces forces and strains between the substrate and the
active layer. The magnitude of these forces and strains depends
2. Materials on the thermal expansion coefficient of both the substrate and the
deposited material. As a result, a high thermal expansion coeffi-
2.1. Substrates cient may lead to a damage in the device. Tables 2–4 summarize
the mechanical, electrical and thermal properties of the three most
2.1.1. Plastic substrates used substrates in the industry [33–35].
The development of flexible electronics goes back to 1960,
where the first array of photovoltaic cells based on crystalline 2.1.2. Paper substrates
silica layer (100 ␮m) was integrated on a plastic substrate. Com- Paper, as a flexible substrate, enables the development of low
mercially, plastic substrates were first used in the fabrication of cost electronic devices using abundant, easily accessible, recyclable,
tape automated bonding interconnects (TAB). In 2001, the first eco-friendly and low cost materials. Moreover, it enables the as fab-
polymer based transistors were introduced by ‘Philips research’. ricated devices to be bent, folded, unfolded and used in the most
During the same year, TITKF/Siemens introduced polymer based unconventional ways that could not be achieved using rigid sub-
organic field effect transistors. These transistors use polymers as strates. Paper based electronic devices are lighter, thinner, roll to
semi-conductors and as insulation. The fast growth in the flexi- roll compatible and might be easily cut using scissors without dam-
ble electronics market drew the attention of many researchers to aging the device. The main advantages of paper substrates over
the possibility and advantages of integrating environmental gas plastic substrates are i) their low cost, ii) recyclability, and iii)
sensors on flexible substrates [25,26]. their low thermal expansion coefficient (2–16 ppm/◦ C) compared
Plastic substrates are the most common flexible substrates used to plastic substrates (20–80 ppm/◦ C).
in flexible electronics, particularly gas sensors. They are highly However, the roughness and absorptiveness of paper limits its
flexible, low cost and roll to roll manufacturable. Polyethylene use in flexible electronics. A solution to this problem is a surface
terephthalate (PET), Polyethylene naphthalate (PEN), parylene, treatment of the paper in order to limit its absorptiveness and
nylon, PDMS, polystyrene and polyimide (PI) are mainly used as smooth its surface. Commercially available paper substrates are
substrates for gas sensing applications [27,28] (Table 1). usually coated with several materials such as kaolin, bentonite, talc
212 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

and polymers [36]. Another solution to this problem is the use of

nanocellusose paper. Other than the small diameter of its fibers
(<10 nm), nanocellulose fibers exhibit a high mechanical strength,
and a large surface area. Nanocellulose has been reportedly used
as a host for several materials such as metal nanoparticles, car-
bon nanotubes and graphene [37]. The literature reports the use of
nanocellulose composites in the fabrication of anti-bacterial filters,
electrochemical sensors, electrochemical gas sensors, biosensors,
fuel cells, solar cells, Li-ion batteries and as a catalyst for the decom-
position of organic pollutants found in the water [38–42]. However,
only cellulose paper was used as a substrate for gas sensors [43–48]:
the efficiency of nanocellulose paper as a substrate for chemiresi-
tive gas sensors has not been yet investigated.
Several gas sensors on paper substrates were reported in the lit-
erature. G. Yang et al. and C. Lee et al. reported the fabrication of a
graphene based nitrogen dioxide gas sensor on commercial white
room paper [49,50]. The graphene sensing layer was deposited
via chemical vapor deposition. J. Safraz et al. reported the fabri-
cation of printed hydrogen sulfide gas sensors on Kaolin coated
paper using polyaniline/copper chloride composites and copper
Fig. 4. Carbon nanotube chiral vector.
acetate as sensing layers [11,51,52]. Another nitrogen dioxide flexi-
ble gas sensor was fabricated by spray coating of graphene oxide on
curtain-coated recyclable paper [53]. This was followed by a chem- Rgas
S [−] = (2)
ical reduction of the deposited graphene oxide. Reduced graphene Rair
oxide was also used for the development of a flexible ammonia
Fractional sensibility is usually only used for metal oxide gas
sensor on filter paper using vacuum filtration [54]. Ammu et al.
sensors due to their high sensibilities (100%–700%). Another impor-
fabricated a nitrogen dioxide and a chlorine gas sensor by inkjet
tant property of a gas sensor is its selectivity. The selectivity is the
printing a carbon nanotube ink on a cellulosic paper [47]. Another
ability of the sensor to detect the target gas without interference
gas sensor on cellulosic paper is reported by M. S. K. Joo-Hyung et al.
from other gaseous species.
The sensor used plasma jet printed MWCNTs as a sensing material
for ammonia detection [55]. Last, Wongchoosuk and al. presented
an ink-jet printed poly(3,4-ethylene dioxythiophene) doped with 2.2.1. Carbon nanotubes
polystyrene sulfonated acid (PEDOT/PSS) ammonia sensor on a [Link]. Structural properties. Ever since their discovery by Ijima
paper substrate [56]. in 1991, carbon nanotubes have drawn the attention of many
researchers due to their unique electrical, mechanical, thermal
2.1.3. Textiles and optical properties. Carbon nanotubes are divided into two
Textiles are of increasing interest as flexible substrates. How- main groups: single walled carbon nanotubes (SWCNT), and
ever, only few articles were published regarding the integration of multi-walled carbon nanotubes (MWCNT). Single walled carbon
gas sensors on textiles. Two main approaches were used. nanotubes are simply a single layer graphene sheet rolled as a tube
The first approach consists of fabricating a gas sensitive textile so with a diameter of several nanometers and a length between 0-
it can act as both substrate and sensing layer. Hong et al. fabricated 100 ␮m. Multi-walled carbon nanotubes consist of several layers
polyaniline-nylon 6 composite textiles. These textiles are highly of graphene sheets [60].
sensitive and selective towards ammonia [10]. The conductivity of a carbon nanotube depends on the orienta-
 [n, m]) (Fig. 4).
tion of its central axis described by its chiral vector (A
The second approach is to integrate a sensing layer or a fabri-
cated flexible sensor on a textile. This method was used by J. Han Metallic conductance occurs whenever the chiral vector’s com-
et al, who deposited single-walled carbon nanotubes on a cotton ponents satisfy the following equation n-m = 3q (q is an integer).
yarn. However, the main drawback of this approach is the bend- We can thus conclude that armchair CNT’s (n = m) are conducting
ing strain, the yarn suffers during the weaving process [9]. The while zigzag (m = 0) and chiral (n = / m and n-m = / 3q) are semi-
same approach was used by Y. Yung et al., who fabricated a car- conducting [61].
bon dioxide gas sensor based on self-assembled reduced graphene Mechanically, carbon nanotubes owe their strength and stiff-
oxide sheets on cotton and polyester yarns [58]. One last example ness to their carbon sp2 bond (C C). Carbon nanotubes present
is the humidity sensor proposed by Mattana et al. The sensor con- the highest Young’s modulus and tensile strength of all fibers and
sists of a spray-coated cellulose acetate butyrate sensing layer on a might be even stronger than diamonds if defect free. The carbon
polyimide substrate. The fabricated sensor was weaved in a textile nanotube’s young modulus is independent of its central axis ori-
[8]. entation, but depends on the tube’s diameter (1 TPa for a diameter
between 1 and 2 nm) [62].
Carbon nanotubes have been extensively investigated in the
2.2. Sensing material
development of electrodes for lithium batteries, organic solar cells
and OLEDS, supercapacitors, organic transistors (OFETs), bistable
This section will cover all materials used in the flexible gas sen-
molecular switches, in the catalysis for fuel cells filter and mem-
sors industry. Before going through the materials in details, we
branes, and sensors [63].
should note that the sensor’s response towards a target gas is cal-
culated using two formulas, the first is the relative sensibility (Eq.
(1)) and the second is the fractional sensibility (Eq. (2)): [Link]. Sensing mechanism. CNTs are well known for their gas
sensing properties. In fact, carbon nanotubes interact with target
Rgas − Rair gases via the Van-der-walls or the donor-acceptor interaction. In
S [%] = × 100 (1)
Rair fact, CNTs act as a p-type semi-conductor whose resistance changes
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 213

the 1980s, conducting polymers were first introduced as an active

layer for gas sensing. In fact, polymers are able to detect several
gaseous species at room temperature. Polymers are easily syn-
thetized using chemical and electrochemical processes, and can
be easily functionalized using metal oxides, metal nanoparticles,
graphene, graphene oxide and other organic compounds such as
amines.
Conducting polymers (CPs), generally used for gas sensing, are
the result of the doping of intrinsic polymers (redox reaction or pro-
tonation for polyaniline), whose conductivity is rather low (<10−5
S cm−1 ). The main chain of pure polymers consists of an alternation
between single and double bonds. The doping process of intrinsic
polymers consists of creating positive charges on the CP’s back-
bone. These charges, called polarons, act as the charge carriers. As
a result, the conductivity of the polymer increases and might reach
conductivities as high as 105 S. cm−1 .

[Link]. Sensing mechanism. A conducting polymer based chemore-

sistive gas sensor converts the concentration of a gaseous analyte
into a variation in the conductivity of the polymer sensing layer. The
conductivity of the sensing layer depends on the doping level of the
Fig. 5. Interaction between nitrogen dioxide and carbon nanotubes. polymer. However, the doping of the polymer is easily changed by
the addition or withdrawal of electrons caused by a redox reaction
between the sensing layer and the target gas. The gas molecules
whenever it interacts with electron-donating (NH3 ) or electron-
are adsorbed at the surface of the sensing layer, altering its conduc-
withdrawing (NO2 , O2 ) molecules (Fig. 5). In addition, CNTs have
tivity. Afterwards, the gas molecules are desorbed and the sensor
a high surface to mass ratio, which increases the contact area
goes back to its initial state. As a result, a p-type polymer placed
between the sensing layer and the target gas and thus increases
under ammonia, a reducing gas, should experience an increase in
the sensibility of the device. All this made CNTs a perfect candidate
its resistance. However, this is not always the case. For example,
for the development of flexible gas sensors.
C. Wongchoosuk and al. developed a flexible gas sensor with a
PEDOT/PSS sensing layer. The sensor exhibits a decrease in its resis-
[Link]. Flexible gas sensors. CNTs were first used for gas sensing tance under ammonia. A possible explanation for this phenomenon
applications in their pristine form [64]. Literature reports dimethyl was the redistribution of charges in the depletion region between
methylphosphonate (DMMP), ammonia (NH3 ) and carbon dioxide the polymer and the metallic contacts [56]. A similar behavior
sensors on plastic substrates using pristine carbon nanotubes as was observed between polyaniline and carbon monoxide [77]. The
a sensing layer [65–68]. These sensors exhibit a good sensibility interaction between polymers and carbon monoxide was studied
towards target gases. Wang et al. report the fabrication of pristine by many and several explanations were proposed, from which the
CNT based gas sensor that can detect DMMP with concentrations as decrease of the charge barrier between the polymer’s grains, and a
low as 1 ppm with a sensibility of 3.6% [69]. Abdellah et al. proposed positive charge transfer from the CO molecule to the amine nitrogen
a flexible CNT based ammonia gas sensor on polyimide foil. The of the polymer [77,78]. Moreover, a PANI sensing layer witnesses a
sensor exhibits a sensibility of 24% towards 10 ppm of ammonia decrease of conductivity when exposed to ammonia. This resistance
[70]. increase, however, does not result from a redox reaction between
However, pristine CNT based sensors’ main drawbacks are their the polymer and the gaseous analyte, but from a deprotonation
irreversibility and lack of selectivity. This problem can be solved process [79]. In fact, ammonia molecules attract a proton from the
by chemical functionalization of the CNTs with metal, metal oxide doped PANI, as shown in Eq. (3) [79]:
nanoparticles and polymers [71]. Researches show that functional-
izing CNTs improves their sensibility and selectivity towards target PANI H + + NH 3  PANI + NH 4 + (3)
gases and reduces their response and recovery time. Literature
reports the development of flexible gas sensors using palladium, [Link]. Flexible gas sensors. Polymers have been widely investi-
iron oxide, reduced graphene oxide, zinc oxide and polymer gated for the development of flexible gas sensors [80,81]. This
functionalized CNTs [72,73]. These sensors can detect hydrogen, is mainly due to their good sensing capabilities at room tem-
nitrogen dioxide, ammonia, methanol, ethanol, and other VOCs. A perature, and their high flexibility which allows the fabrication
SWCNT-Fe2 O3 film exhibits a 15% increase in its sensibility towards of flexible devices without damaging the sensing layer under
nitrogen dioxide compared to a pristine SWCNT film [74]. More- strain. Polyaniline, polypyrrole, polythiophene (Pth) and poly(3,4-
over, a CNT/graphene hybrid layer has a sensibility of 19% towards ethylenedioxythiophene) (PEDOT) are the main polymers used in
10 ppm of NO2 , whereas a pristine CNT layer has a sensibility of only the gas sensing industry.
2% [75]. Palladium decoration also enhances the sensitivity of car- Wongchoosuk et al. reported an ammonia gas sensor based
bon nanotubes. In fact, M. Han et al. demonstrated that a palladium on an inkjet printed poly(3,4-ethylene dioxythiophene) doped
nanoparticle decorated MWCNT senses hydrogen better than pris- with polystyrene sulfonated acid (PEDOT/PSS) sensing layer on a
tine MWCNT and palladium with sensibilities of 17%, 1% and 13% photo paper. The sensor exhibits a 10% response towards 100 ppm
respectively [76]. of ammonia [56]. The sensing response of inkjet-printed PANI
nanoparticles on a PET substrate were also investigated. The fab-
2.2.2. Conducting polymers ricated sensors exhibited a 24% response towards 100 ppm of
[Link]. Structural properties. Ever since the 1970s, conducting ammonia [79].
polymers have drawn the attention of many researchers not only as Polymer based gas sensors have usually slow response and
an insulating substrate for electronics, but as a gas sensing layer. In recovery rate and lack of selectivity. However, these properties are
214 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

enhanced by functionalizing the polymer with other materials [82]. The sensing capabilities of a MOX gas sensor depend not only
For example, the sensing response of polyaniline-FMWCNT com- on the structure of the sensing material (thin film, nanorods,
posites towards 10 ppm of ammonia show an increase of sensibility nanowires. . .) but also on the material itself. In fact, metal oxides,
(30%) compared to pristine PANI (5%) and to pristine FMWCNTs (1%) such as TiO2 , operate at high temperatures (>700◦ C) because they
[83]. The fabricated device is selective. It exhibits a higher sensi- rely on bulk conductance changes. Other metal oxides, such as ZnO
bility towards ammonia compared to other volatile gases (ethanol, and SnO2 , rely on surface conductance changes and thus operate at
methanol, acetone, isopropanol). Another example is the graphene- lower temperatures (<600◦ C) [95]. Since flexible substrates cannot
PEDOT:PSS based gas sensor proposed by Y. Seekaew. The sensor’s sustain high temperatures, metal oxides that follow bulk conduc-
response towards 500 ppm of NH3 is compared to the response of tance changes are rarely used in the development of flexible gas
a PEDOT:PSS and a pristine graphene gas sensor. The sensibilities sensors. Most flexible MOX gas sensors use ZnO and SnO2 as sens-
were calculated to be 9.6%, 4.4%, and 2.4% respectively [84]. The ing materials. Other factors affecting the choice of the material are
proposed sensor is unable to detect acetone, methanol, ethanol and its adsorption ability, catalytic activity, sensitivity and thermody-
toluene. However, it has a slight sensibility towards diethylamine. namic stability. However, there is no metal oxide that satisfies all
As a result, the device isn’t selective. these requirements. This is why metal oxides are usually doped or
The literature reports other flexible gas sensors using function- functionalized by other materials such as metals, carbon nanotubes
alized polymers as a sensing layer. These layers were functionalized and polymers in order to enhance their sensing performance [30]. In
using various compounds such as platinum [85], iron oxide [86], fact, doping a metal oxide with metals such as palladium, platinum,
CNTs [29,72,83,87], reduced graphene oxide [88], SnO2 [89] . . . In copper and silver increases the sensibility by acting as a catalyst for
conclusion, current work in polymer based gas sensors is oriented the reaction between the sensing layer and the target gas and may
towards polymer hybrids sensing layers for their improved per- even lower the sensor’s working temperature. Last, a synergistic
formance compared to pristine polymers in terms of sensibility, effect between the MOX and the functionalizing material (MOX,
selectivity, and reversibility. CNTs, polymers) explains the improvement of the sensing capabil-
ities in terms of sensibility, operating temperature, response and
2.2.3. Metal oxides recovery times.
[Link]. Structural properties. Metal oxides were the first materials
to be used in the gas sensing technology in the 1950s, when Hei-
land demonstrated, for the first time, the change of conductivity of [Link]. Sensing mechanism. The sensing mechanism of a metal
metal oxides, such as ZnO, depending on the changes in the sur- oxide based chemiresistive gas sensor is quite simple. First, we
rounding atmosphere [90]. This discovery was followed by many should point out the fact that the adsorbed oxygen plays a cru-
other discoveries, all leading to the first commercial tin dioxide gas cial part in the gas sensing performance of metal oxides. In fact,
sensor for domestic gas alarms in 1970. oxygen molecules are adsorbed on the surface of the sensing
Nowadays, metal oxides such as ZnO (20.83%), SnO2 (37.39%), layer, extracting electrons from the conduction layer of the metal
TiO2 (12.81%), WO3 (8.65%), Fe2 O3 (4.27%), In2 O3 (6.64%), CuO oxide and trapping them on its surface in the form of anions. This
(3.95%), Co3 O4 (1.13%), Cr2O3(0.94%), NiO (3.29%), Mn3 O4 (0.1%) phenomenon creates a depletion region. Whenever an n-type semi-
are used for gas sensing applications [91]. conductor is exposed to a reducing gas, such as ammonia or carbon
Tin oxide is the most investigated in the field of gas sensing. Tin monoxide, the adsorbed oxygen oxidizes the reducing gas, and the
oxide is an n-type semi-conductor, with a wide band-gap (3.6 ev). resulting free electrons are injected in the semi-conductor core.
Tin oxide is sensible towards many target gases, including ammo- This results in a decrease in the depletion region and an increase of
nia, nitrogen dioxide, carbon monoxide, humidity, carbon dioxide. the conductivity of the sensing layer proportionally to the concen-
It has a high operating temperature between 200 and 400 ◦ C. Tita- tration of the target gas (Fig. 6a).
nium dioxide and zinc oxide are also often used for gas sensing The opposite happens when a p-type metal oxide in exposed
applications. Zinc oxide is a wide band gap (3.4 eV) n-type semi- to a reducing gas, where the depletion layer becomes thicker and
conductor. It has a high exciton binding energy (60 meV). Zinc oxide the concentration of the holes of the semi-conductor decreases
has been widely investigated for the development of gas sensors due to the recombination between the electrons injected into the
due to its thermal/chemical stability and its good sensing prop- semi-conductor following the oxidation of the reducing gas by the
erties. Gas sensors using ZnO as a sensing layer usually operate adsorbed oxygen, and the semi-conductor’s holes (Fig. 6b). As a
at high temperatures (around 500 ◦ C). Titanium oxide is mainly result, the conductivity of the sensing layer decreases. The exam-
used because of its low cross-sensitivity with humidity. It oper- ple of zinc oxide (n-type semi-conductor) interacting with ozone
ates at mid-high temperatures that might reach (400–600 ◦ C and and ammonia is explained by M. Acuaulta and al. [96] (Fig. 7).
700–1000 ◦ C) for oxygen detection [92]. Since the sensing mechanism relies on the interaction between
While metal oxides usually operate at high temperature, the the adsorbed oxygen on the surface of sensing layer, the sensibil-
literature reports some metal oxide based gas sensors working at ity of metal oxide gas sensors depends on the concentration of
room temperature. For example, G. Mani and al. demonstrated that the adsorbed oxygen on its surface. Studies have shown that the
spray deposited zinc oxide thin film may sense target gases at room adsorption of oxygen molecules is enhanced by heating the sensing
temperature [93]. Moreover, I. Kadhim and al. reported a room layer. In other words, heating a metal oxide film will increase the
temperature hydrogen gas sensor based on an SnO2 sensing layer. number of adsorbed oxygen atoms on its surface and thus allow
The fabricated sensor exhibits an increase in its sensibility at high it to interact with a larger number of gas molecules. This results
temperatures [94]. in a higher sensitivity and a faster response and recovery. This
The sensibility of metal oxides towards target gases depends on also justifies the fact that most metal oxide gas sensors on rigid
the thickness of the layer, its porosity, its structure and the grain substrates reported in the literature operate at high temperatures
size and shape. For instance, reducing the grain size increases the (200–700 ◦ C) [30]. Research in the flexible gas sensing field is now
specific area and the number of atoms that are in a direct contact focusing on using metal oxide nanowires, nanorods, nanoribbons
with the target gas and increases the sensibility. A similar effect and nanoflowers as sensing materials in order to increase the con-
happens when increasing the porosity of the metal oxide and when tact surface between the sensing layer and the target analyte. This
using metal oxides nanostructures such as nanorods, nanowires, enables room temperature detection without any performance loss
nanoribbons, nanoflowers. or substrate damage [97].
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 215

Fig. 6. Interaction of a (a) n-type semi-conductor (b) p-type semi-conductor with a reducing gas based on [96].

Fig. 7. Graphene oxide structural models.

[Link]. Flexible gas sensors. Metal oxides were the first to be inves- response and recovery time. It has a response of 1.8 [-] towards
tigated for flexible electronics applications. However, these devices 20 ppm of CO and a response of 35 [-] towards 20 ppm of NO2 at an
main limitation is the low operating temperature of flexible sub- operating temperature of 300 ◦ C [99].
strates. The low temperature constraint limits the sensibility of The literature also reports the use of metal oxides to chemically
metal oxides flexible gas sensors as explained in the previous sec- functionalize other materials such as polymers, carbon nanotubes
tion. An example is the ZnO based gas sensor reported in [98]. The and metal nanoparticles in order to improve their sensibility, selec-
sensor’s response at 180◦ C is 2.5 times its response at room tem- tivity and reversibility [105,106]. A SWCNT-Fe2 O3 composite film
perature. Nevertheless, literature reports many flexible gas sensors was fabricated by chemical vapor deposition of SWCNT followed
based on a metal oxide sensing layer. Most of these sensors are by an annealing process [74]. The sensor exhibited good response
either based on a metal oxide thin layers [96,99], on porous metal (11.3%) towards 10 ppm of ammonia but is irreversible. Another
oxide [100] or on metal oxide nanowires, nanoflowers, nanobelts ammonia gas sensor based on ITO functionalized SWCNT was
and nanoribbons [97,101]. As explained in the previous section, reported by F. Rigoni et al. The sensor detects 10 ppm of ammonia
these structures increase the specific area and thus increase the with a sensibility of 5.6% [107]. Another example of a functional-
chemical interaction rate between the sensing material and the ized metal oxide flexible gas sensor is the ethanol sensor fabricated
target gases. by S. Zhan et al. [108]. The sensing layer is a mixture of SnO2 and
H. Ahn et al. report the fabrication of an ethanol gas sensor by PDDAC. PDDAC that acts as a binder between the sensing layer and
growing ZnO nanorod on polyimide substrate by thermolysis. The the substrate and as a dopant for the SnO2 nanoparticles. The sen-
sensor detects concentrations from 10 to 100 ppm. It has a sensi- sor’s response towards 150 ppm of ethanol is 71.6% and its response
bility of 3.11 [-] towards 100 ppm of ethanol at 300 ◦ C [102]. M. time is 88 s. These results are quite an improvement compared to
Tonezzer et al. fabricated an oxygen and hydrogen gas sensor by those obtained with pristine SnO2 (S = 15.5% and tr = 470 s). Last, the
thermal oxidation of electrodeposited zinc oxide on carbon fibers. fabricated device has a good selectivity towards ethanol compared
The sensor detects 100 ppm of oxygen and hydrogen with a sensi- to NO2 , H2 , SO2 and H2 S.
bility of 1.6 [-] and 2.3 [-] respectively and has a fast response (8 s
and 10 s) and recovery (12 s)[103]. Other ZnO and ZnO nanorods
2.2.4. Graphene and graphene oxide
based sensors were reported on polyimide and nylon fibers for the
[Link]. Structural properties. Graphene consists of sp2 bonded car-
detection of ozone, ammonia and hydrogen [96,98,104,105], Tin
bon atoms in the form of a fullerene. It has good electrical and
oxide is also used for the fabrication of flexible gas sensors: a car-
mechanical properties in terms of electrical conductivity, pliabil-
bon monoxyde and nitrogen dioxide gas sensor was fabricated by
ity, and high specific surface area. All these characteristics made
inkjet printing of SnO2 on a polyimide foil. The sensor exhibits good
graphene and graphene oxide the center of interest for many
216 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

researchers in the field of electronics for the development of sen- acts with the oxygen atoms of the graphene oxide with a binding
sors, electrodes, supercapacitors etc. energy ranging from 0.415 to 8.29 eV [118,119].
Graphene oxide consists of a graphene layer decorated with Basically, graphene oxide acts as a p-type semi-conductor. Upon
functional groups such as carboxyl, hydroxyl, and epoxy. The struc- its exposure to a target gas, the gas molecule binds with the
ture of graphene oxide is still controversial. Many models were graphene oxide functional groups, thus injecting or withdrawing
introduced throughout the years. The first model was proposed electrons from the semi-conductor. This results in a change in
by Hoffman and Holst in 1939 [109]. It consists of planar carbon the conductivity of the sensing layer. The sensing performance
layer with epoxy between carbon 1 and 2 as a functional group. of graphene oxide depends on i) the binding energy between the
The second model consists of a carbon layer with either hydroxyl target analyte and the graphene oxide functional groups, ii) the fab-
or 1,3-epoxy as functional groups. It was introduced by Ruess in rication process of graphene oxide that controls the concentration
1946 [110]. In 1969, another model was proposed by Scholz and of these functional groups, iii) the concentration of the target gas.
Boehm in which epoxy were replaced with carbonyl groups [111].
These models were followed by many others such as Nakajima et al [Link]. Flexible gas sensors. In the last decade, graphene oxide has
[112], Lerf et al. [113], Szabo et al [114], and Drayer and al [115]. The been widely investigated for the fabrication of flexible gas sensor.
structure of graphene oxide depends on the oxidization technique, This is mainly due to its excellent sensing properties, its flexibility,
the oxidant, and the oxidization time. its mechanical strength, and its low operating temperature.
Graphene oxide, in its initial form, is insulating. Its low con- The literature reports many pristine graphene and graphene
ductivity is mainly due to the carbon sp3 (C O) bond. However, oxide based gas sensors on plastic and paper substrates. Dua and al
graphene oxide may be reduced in order to increase its conductivity report the fabrication of an inkjet-printed reduced graphene oxide
and transformed into a semi-conductor by reducing the number of gas sensor. The device has good sensing capabilities. It detects nitro-
sp3 bonds. Graphene oxide can be chemically reduced in its solution gen dioxide with concentrations as low as 500 ppb. Moreover, it has
state using a reducing agent such as hydrazine, sodium borohy- a sensibility of -6%, 21%, 10%, 9%, -10% towards 100 ppm of nitro-
dride, glucose, hydroquinone. Moreover, graphene oxide may be gen dioxide, ammonia, methanol, ethanol and chlorine respectively
thermally reduced by annealing a graphene oxide layer or using [120]. Another spray-coated reduced graphene oxide gas sensor on
electrochemical methods or laser irradiation. curtain-coated recyclable paper was reported by Hassinen et al.
Another interesting property of graphene oxide is the possibility [53]. The sensor exhibited a 70% response towards 10 ppm of NO2 .
to functionalize it. The functionalization’s main goal is to enhance However, the sensor took several hours to recover under N2 . The
the electrical properties of pristine graphene oxide in the case of a process was accelerated using UV radiation.
supercapacitor, its sensing properties in case of a gaz sensor, and As for graphene based flexible gas sensors, one approach is
other properties depending on the target application. Graphene mainly adopted: deposition of graphene on copper foil then trans-
oxide may be decorated with metal oxides, metal nanoparticles ferring it to a flexible substrate. The literature reports two sensors
and even be used to form graphene oxide/polymer composites. fabricated using this principle, one on a PET substrate, and the other
Graphene oxide may also be used to functionalize other materials, on a paper substrate [49,50]. The first is a nitrogen dioxide sensor
such as CNTs, in order to increase their mechanical strength. with a sensibility of 23% towards NO2 under relaxed conditions and
25% under strained conditions (1.39%). The second is an ammonia
[Link]. Sensing mechanism. The sensing mechanism of graphene gas sensor with a sensibility of 32–39% towards ammonia under
oxide based gas sensors in based on the binding between the relaxed and strained conditions (up to 0.5% strain).
graphene oxide’s functional groups and the gas molecules. In fact, Graphene oxide may be covalently functionalized or linked
the epoxy, hydroxyl and carbonyl ( O, OH, COOH) functional to flexible substrates by activating the carbonyl and hydroxyl
groups enhance the adsorption on gas molecules into the graphene groups. For example, graphene oxide is covalently linked to cer-
oxide sheet. The higher the binding energy between graphene oxide tain polymers using the amidation technique. This is the case of the
and target gases, the closer the adsorbed gas molecule is to the sensor reported by Pi-Guey Su et al. Gold electrodes were modified
graphene matrix and the better it is detected. The adsorption of using cysteamine hydrochloride. Afterwards, GO sheets were cova-
a gas molecule depends on its adsorption sites ( OH, COOH, O) lently linked onto the electrodes using a peptide coupling reagent
and on the binding energy between these sites and the gas molecule (EDC/NHS). The sensor detects 5 ppm of NO2 with a sensibility of
itself. 11% and a response and recovery time of 7 s and 28 s respectively.
Many studies were conducted on the binding energy between The fabricated sensor is selective compared to carbon monoxide,
graphene oxide and gas molecules. For example, the adsorption of hydrogen and humidity, but has a sensibility of −10.7% towards
ammonia on graphene oxide is independent of the orientation of 200 ppm of ammonia [121].
the gas molecules, but depends on the hydroxyl and epoxy func- Moreover, the sensibility of graphene and graphene oxide
tional groups. However, the enhancement of the adsorption energy may be improved by chemical modification using polymers, car-
due to the hydroxyl group is higher than the one due to the epoxy bon nanotubes, metals and metal oxides. For example, graphene
group. Thus, the hydroxyl group has a higher contribution in the may be decorated with palladium nanoparticles to enhance its
adsorption of ammonia on the surface of graphene oxide [116,117]. performance. Such sensors detect 1000 ppm of hydrogen with a
Another example is the adsorption of oxygen on graphene oxide sensibility of 32.9% and are flexible (radius of 3 mm) [122]. Another
sheets. In fact, oxygen molecules bind with the carbonyl func- example is the humidity sensor proposed by D. Zhang et al. [123].
tional groups ( COOH) with a binding energy of 0.174 eV whereas The sensor consists of an RGO-PDDA sensing layer on a poly-
the binding energy of oxygen with pristine graphene oxide is of imide substrate. The sensor exhibits a response of 8.69%, 14.99%,
−0.006 eV. The same applies to nitrogen, nitrogen oxide and carbon 20.91%, 28.14%, 37.41% towards 11%, 33%, 52%, 75%, 97% RH. As
monoxide that bind with carbonyl functional groups with binding for carbon nanotubes, vertically aligned carbon nanotubes were
energies of 0.027 eV for nitrogen, 1.187 eV if the binding occurs with grown on a graphene layer to achieve a sensibility of 20% towards
the nitrogen of the nitrogen oxide molecule or 1.157 eV if the bind- 10 ppm of NO2 [75]. The literature also reports other functional-
ing occurs with the oxygen atom of the nitrogen oxide molecule, ized graphene oxide based gas sensors using silver nanoparticles
and 0,045 eV for carbon monoxide [117]. Water molecules bind [124,125], PEDOT:PSS [84] and polyaniline [88].
with the hydroxyl and epoxy functional groups with energies of
0.201 and 0.259 eV respectively whereas nitrogen dioxide inter- 2.2.5. Metals
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 217

Fig. 8. Metallic crystal structures a) FCC b) BCC c) HCP.

Table 5
Thermal properties of noble metals [128].

Thermal
Proterties
Gold Silver Platinum Palladium

Thermal expansion coefficient (10−6 K-1 ) 14.1 19.1 9 11

Thermal conductivity at 23◦ C (W. m−1 . K−1 ) 317 420 107 72
Melting point or degradation (◦ C) 1064 961 1772 1555

Table 6
Electrical properties of noble metals [128].

Electric
Proterties
Gold Silver Platinum Palladium
6
Electrical conductivity (10 Siemens/m) 44.2 62.1 9.3 9.5
Electrical resistivity (10−8 Ohm.m) 2.3 1.6 10.8 10.5

Fig. 9. Hydrogen palladium beta hybrid lattice structure.

- Gray spheres = fcc lattice of Pd.
- Red and green spheres = octahedral and tetrahedral.
[Link]. Structural properties. Metals present a crystalline struc-
ture (Face Centered Cubic FCC, Body Centered Cubic BCC and
Hexagonal Close Packed HCP) (Fig. 8). Noble metals, such as gold,
platinum, palladium and silver adopt a FCC crystallographic struc- Palladium is widely used for hydrogen gas sensing. This is mainly
ture with lattice constants of 407.82 pm, 392.42 pm, 389.07 pm and due to the fact that it absorbs hydrogen up to 900 times its volume
408.53 pm respectively [126,127]. at room temperature. The sensing mechanism can be explained as
Noble metals are characterized by their high conductivity, high follows: The adsorption of a hydrogen molecule onto the surface
melting point and especially their high resistance to oxidation. of a palladium layer results in the dissociation of the H2 molecule
Thermal and electrical properties of gold, silver, platinum and pal- into single hydrogen atoms. These atoms diffuse into the bulk of the
ladium, mainly used in gas sensing devices are summarized in Pd lattice to occupy interstitial sites. This results in an expansion
Tables 5 and 6. of the lattice and thus the formation of PdHx hybrids. Two main
Noble metals have been first investigated for enhancing gas PdHx hybrids form depending on the concentration of hydrogen
sensing properties of metal oxides in 1990s. Metals, such as gold, in the air. Small concentrations of hydrogen (< 200 ppm) lead to a
silver, platinum and palladium, are either integrated in metal slight expansion in the palladium lattice and thus the formation of
oxides as ions on lattices sites or in the form of clusters and alpha-hybrids. On the other hand, higher concentrations of hydro-
nanoparticles [129]. Moreover, metal nanoparticles are used to dec- gen (> 5500 ppm) lead to the formation of beta-hybrids having a
orate carbon nanotubes, graphene and graphene oxide in order larger lattice with more hydrogen atoms filling its interstitial sites
to enhance their sensing properties. In the next section, we will (Fig. 9). The two hybrid types coexist for concentrations around
explain the sensing mechanism of metals in their pristine form, as 1000 ppm. This volume increase of the palladium allows hydrogen
dopants or as nanoparticles decorating other materials (Table 7). to be detected either by a resistance increase in palladium thin film
layers, or by increasing the conductivity of discontinuous palladium
layers by reducing the nanogaps size due to swelling.
[Link]. Sensing mechanism. In gas sensing studies, metals are usu- Metal nanoparticles are widely investigated for their role in
ally used to functionalize a sensing material in order to enhance its the improvement of the sensing capabilities of carbon nano-
sensing properties in terms of sensibility, selectivity, and response tubes, graphene, graphene oxide and metal oxides. The literature
and recovery times. Platinum, palladium, gold and silver are the reports gas sensors functionalized with aluminum, gold, silver,
most used metal for gas sensing. However, the only metal used as platinum and palladium. The sensing mechanisms of graphene, car-
a sensing layer by itself is palladium. Pure metal thin films such bon nanotubes, and graphene oxide decorated with noble metal
as palladium thin films and palladium-silver alloys are used for nanoparticles are similar [122,132]. Let’s explain the sensing mech-
hydrogen sensing [130,131]. In this paragraph we will explain the anism of carbon nanotubes as an example. Upon exposure to
palladium gas sensing principle as well as the role of metal decora- hydrogen, PdHx hybrids form. These hybrids lower the work
tion in the improvement of the sensing response of other materials function of the palladium and facilitate electron mobility from
such as carbon nanotubes, graphene oxide and metal oxides. palladium to CNTs. Since CNTs are usually p-doped, the electrons
218 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

Table 7
Flexible gas sensors.

Target gas Sensing material Substrate Fabrication technique Sensibility Ref.

NH3 SWCNTs PDMS Vaccuum filtration Cg* = 1% [65]

S** = 10%
RT***
CNTs Polyimide Spray deposition Cg = 10 ppm [70]
S = 10%
RT
SWCNT-Fe2O3 Plastic substrate CVD and annealing Cg = 10 ppm [74]
S = 1.5%
RT
MWCNTs/polymer (PVP/PEO/EC/PMS) PI Drop-casting Cg = 1000 ppm [72]
S(PVP) = 10%
S(PEO)7.5%
S(EC) = 5%
S(PMS) = 4%
RT
PEDOT/PSS Photo paper Inkjet printing Cg = 100 ppm [56]
S = 10%
RT
Pani PET Inkjet printing Cg = 100 ppm [79]
S = 24%
RT
Pani/Fe2O3 PET In-situ polymerization Cg = 100 ppm [86]
S = 39%
RT
Pani- Nylon 6 Nylon 6 Polymerization Cg = 1000 ppm [10]
S = 120%
RT
Pani PET In-situ polymerization Cg = 200 ppm [166]
S = 30%
RT
Pani-CNTs PET In-situ polymerization Cg = 10 ppm [87]
S = 6%
RT
Polymer PEN Inkjet printing Cg = 100 ppm [29]
(PVC/Cumene-PSMA/PSE/PVP) - CNTs Smax = 17%
RT
Pani-MWCNTs PET In-situ polymerization Cg = 10 ppm [83]
S = 5%
RT
PPy PET In-situ self-assembly Cg = 50 ppm [163]
S = 77.75%
RT
PEDOT/PSS - graphene Plastic substrate Inkjet printing Cg = 500 ppm [84]
S = 9.6%
RT
Pani-rGO PET In-situ polymerization Cg = 10 ppm [88]
S = 5.5%
RT
Pani-PMMA Polyethylene (PE) electrospinning Cg = 20 ppm [172]
S = 3.5%
RT
SWCNTs-ITO Plastic substrate Drop-casting Cg = 10 ppm [107]
S = 5.6%
RT
SWCNTs Plastic substrate Drop-casting Cg = 10 ppm [107]
S = 2%
RT
GO-Ag Nps**** PET Inkjet printing Cg = 15 ppm [124]
S = 4.25%
RT
rGO PET Inkjet printing Cg = 100 ppm [120]
S = 21%
RT
rGO PET Covalent anchoring Cg = 20 ppm [121]
S = 10.7%
RT
rGO Filter paper Vacuum filtration Cg = 400 ppm [54]
S = 5%
RT
Graphene Plastic substrate Inkjet printing Cg = 500 ppm [84]
S = 2.4%
RT
PEDOT/PSS Plastic substrate Inkjet printing Cg = 500 ppm [84]
S = 4.4%
RT
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 219
Table 7 (Continued)

Target gas Sensing material Substrate Fabrication technique Sensibility Ref.

SWCNT-Au Nps PET Spray deposition of SWCNTs Cg = 6 ppm [136]

E-beam evaporation of gold S = 0.15%
Nps RT
CNT Cellulosic paper Drop-casting Cg = 50 ppm [44]
S = 2%
RT
CNT Cellulosic paper Papermaking filtration Cg = 50 ppm [44]
S = 1%
RT
MWCNTs paper Plasma jet printing Cg = 60 ppm [55]
S = 4%
RT
TiO2-MWCNT Cellulose fibers Covalent anchoring using a Cg = 50 ppm [46]
binding molecule S = 17%
RT
GaN Cellulose paper coating Cg = 50 ppm [48]
S = 5%
RT

NO2 CNT-rGO hybrids Polyimide Growth Cg = 10 ppm [75]

S = 20%
RT
SWCNT-Fe2O3 Plastic substrate CVD and annealing Cg = 10 ppm [74]
S = 2.9%
RT
SnO2-PPy PET Drop-casting of SnO2 Nps Cg = 5 ppm [89]
In situ-polymerization of PPy S = 8.2%
RT
SnO2 PI Inkjet printing Cg = 20 ppm [99]
S = 35
T = 300◦ C
WO3- PEDOT/PSS PI Gravure printing Cg = 5 ppb [106]
S = 1.2
RT
rGO-CNT PI Growth of CNT on GO film Cg = 10 ppm [75]
S = 20%
RT
rGO PET Inkjet printing Cg = 10 ppm [120]
S = 6%
RT
graphene PET CVD Cg = 200 ppm [50]
S = 23%
RT
graphene paper CVD Cg = 200 ppm [49]
S = 32%
RT
rGO Curtain-coated Spray-coating Cg = 10 ppm [53]
recyclable paper S = 70%
RT
Ag-S-rGO PI Gravure printing Cg = 500 ppb [125]
S = 3.5%
RT
rGO PET Covalent anchoring Cg = 5 ppm [121]
S = 11%
RT
Graphene-Palladium Nps PI e-beam and thermal Cg = 1.2 ppm [139]
evaporation S = 32%
RT
Graphene- Aluminum Nps PI e-beam and thermal Cg = 1.2 ppm [139]
evaporation S = 200%
RT
CNT Cellulosic paper Inkjet printing 35% towards [47]
10ppm

H2 MWCNTs/Pd Nps MWCNTs sheet In situ-growth and Cg = 15% [76]

electrodeposition S = 10%
RT
PPy-Pt PET Layer by layer self-assembly Cg = 1000 ppm [85]
S = 11.6%
RT
ZnO nanowires Carbon microfiber Electrodeposition and thermal Cg = 500 ppm [103]
textile oxidation S = 11
T = 280◦ C
ZnO nanorods Nylon substrate Hydrothermal growth 109% towards [98]
500 ppm at RT and
264% at 180◦ C
ZnO-Pd PI Hydrothermal growth of ZnO Cg = 1000 ppm [105]
nanorods S = 91%
RF magnetron sputtering of Pd RT
Nps
220 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

Table 7 (Continued)

Target gas Sensing material Substrate Fabrication technique Sensibility Ref.

Pd-Graphene PET CVD of graphene Cg = 1000 ppm [122]

Thermal evaporation of Pd Nps S = 32.9%
RT
Flower-like palladium nanoclusters PEN Electrodeposition Cg = 10 ppm [138]
S = 5%
RT
Palladium nanotubes PI Chemical reduction around Cg = 100 ppm [134]
hydrothermally grown ZnO S = 247%
nanorods followed by ZnO RT
etching
MWCNT-Pd Nps PET Layer by layer self-assembly Cg = 200 ppm [132]
S = 1.95%
RT
Pd polycarbonate Nanoimprint lithography Cg = 3500 ppm [135]
S = 2.3%
RT
SWCNT-Pd Nps PET CVD and e-beam evaporation Cg = 200 ppm [137]
S = 20%
RT
Pd PI RF sputtering Cg = 20,000 ppm [147]
S = 12%
RT
DIMP SWCNTs PET Dip-coating Cg = 1000 ppm [67]
S = 172%
RT

DMMP CNT Polyimide Self-assembly Cg = 10 ppm [69]

S = 5.5%
RT
SWCNTs PET Dip-coating Cg = 1000 ppm [67]
S = 66%
RT

Ethanol ZnO nanorods PI Thermolysis Cg = 100 ppm [102]

S = 3.11
T = 300◦ C
SnO2 - PDDAC PI Drop-casting Cg = 150 ppm [108]
S = 71.6%
RT
SnO2 PI Drop-casting Cg = 150 ppm [108]
S = 15.5%
RT

O2 ZnO nanowires Carbon microfiber Electrodeposition and thermal Cg = 500 ppm [103]
textile oxidation S = 10
T = 280◦ C

CO SnO2 PI Inkjet-printing Cg = 20 ppm [99]

S = 1.8
T = 300◦ C

O3 ZnO PI Photolithography Cg = 500 ppb [96]

S = 10.07
T = 200◦ C
ZnO PI Laser ablation Cg = 500 ppb [96]
S = 26
T = 200◦ C

H2 O rGO - PDDA PI Layer-by-layer nanoassembly 20.91%, towards [123]

52% of H2O
CNT Cellulosic filter Drop-casting 43% towards 60% [43]
paper RH

Cl2 CNT Cellulosic paper Inkjet printing ∼40% towards [47]

∼20ppm

*Cg: concentration of the target gas.

**S: sensibility.
***RT: Room temperature.
****Nps: Nanoparticles.

transferred from palladium to CNTs combine with holes from the the fabrication of a nanogap-based hydrogen sensor on PDMS sub-
CNTs lowering the conductivity of the sensing layer. A similar strate. The detection principle is based on the increase of the
behavior is observed with carbon nanotubes decorated with gold conductivity on the sensing layer due to the swelling of the pal-
nanoparticles. ladium under hydrogen. In other words, the fabricated sensor acts
as a switch. The sensor detects concentrations as low as 0.4% of
H2 and presents a fast response towards 2% of H2 (0.67 s) [133].
[Link]. Flexible gas sensors. The literature reports several gas sen-
Another way to develop a hydrogen gas sensor is to grow pal-
sors based on palladium for hydrogen sensing. J. Lee et al report
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 221

Fig. 10. Coating techniques a) Drop-casting b) spin-coating [141].

ladium nanotubes on polyimide foil [134]. The sensor exhibits a sensors on polyimide foil [139]. Aluminum increased the sensitiv-
response of 247% towards 100 ppm of H2 . Moreover, the sensor ity of graphene of 200% and the palladium decreased it of 32%
presents an improvement of 18% under strain (10 mm bending under nitrogen dioxide. Under ammonia, the aluminum slightly
radius). However, the sensor’s performance in terms of sensibil- increased the sensibility of graphene (1.94%–2.04%) and the palla-
ity and response time were enhanced by increasing the operating dium slightly decreased it (1.94% to 1.87%). Finally, B. Wu reported
temperature. Other hydrogen gas sensors were developed using graphene-nanosilver based ammonia gas sensors on a PET sub-
nanoimprint lithography by S.H Lim et al. [135]. The fabricated strate [124]. The sensor has a response of 14.7% towards 200 ppm
sensor exhibits a response of 0.15% and 2.3% towards 50 ppm and of ammonia.
3500 ppm of H2 , respectively.
Other researchers have dedicated their work to the integration 3. Device fabrication
of noble metals in a gas sensing layer. For instance, many arti-
cles report the decoration of carbon nanotubes with noble metal 3.1. Coating
nanoparticles for gas sensing. [Link] et al. decorated single-walled
carbon nanotubes with gold nanoparticles for ammonia detection. When dealing with soluble materials, the simplest way to fabri-
The sensor presents a response of 0.15% towards 6 ppm of ammonia cate a flexible sensing device is to coat the substrate with a sensing
while bare SWCNT only have a 0.05% response towards the same gas layer.
concentration. The sensor is also able to detect sub-ppm concentra- The first and the simplest coating technique is drop-casting. Sen-
tions of ammonia (400 ppb) and is reversible under UV radiations sors are fabricated by dropping a solution containing the sensing
[136]. Multi-walled carbon nanotubes may also be modified with material on the substrate and waiting for the solvent evaporation
palladium nanoparticles as reported by P.-G. Su et al. [132]. The (Fig. 10a.). This deposition technique is very simple. However, the
fabricated flexible sensor on PET substrate detects 10,000 ppm of thickness of the sensing layer is hard to control and the deposited
hydrogen with a sensibility of 16.44% whereas the sensibility of layer lacks of uniformity. Literature reports several flexible gas
sensors based on pristine MWCNT does not exceed 3.5%. The sen- sensors fabricated using the drop-casting technique. For exam-
sor exhibits a response time of 300 s and a recovery time of 510 s ple, J.-C. Chiou et al. reported the fabrication of a polymer/carbon
towards 10,000 ppm of hydrogen. However, a deviation of 18% in nanotubes on a polyimide substrate. 2 ␮l of MWCNT (60 ppm)
the sensor’s response is observed when increasing the bending were drop-casted on the substrate with printed interdigitated elec-
angle from 10 to 60◦ at 1000 ppm of hydrogen. Moreover, the sen- trodes. After solvent evaporation, another 2 ␮l of polymer (PVP,
sor’s performance escalates with the increasing temperature and PEO, EC and PMS) were drop-casted on top of the MWCNT layer
humidity. Another SWCNT decorated with palladium nanoparti- [72]. The fabricated sensors were able to detect ammonia, nitrogen
cles was reported in [137]. The sensor exhibits a sensibility of 20% dioxide and toluene with a maximum sensibility of 10% towards
towards 200 ppm of H2 with a response time of 320 s. One drawback 1000 ppm of ammonia (PEO/MWCNTS), 79% towards 1000 ppm of
is that for high concentrations of H2 , the sensor’s response becomes nitrogen dioxide (PEO/MWCNTs) and 2% towards 10 ppm of toluene
similar to the response of pristine SWCNT due to the saturation of (PMS/MWCNTs). Another research group used the drop-casting
the palladium sites. Another drawback is the 4.7% deviation in the method as an intermediate step in the fabrication of a nitrogen
sensor’s response under bent conditions (2 mm radius) and the 13% dioxide flexible gas sensor. In fact, FeCl3 -SnO2 nanoparticles were
deviation after 1000 bending cycles. drop-casted on a PET substrate with Au/Cr interdigitated elec-
Other articles discussed the integration of noble metal nanopar- trodes deposited using e-beam evaporation. The substrate were
ticles in graphene-based gas sensors for hydrogen, ammonia and placed afterwards in a pyrrole saturated vapor for 3 h in order to
nitrogen dioxide sensing. One flexible hydrogen sensor was fab- obtain a polypyrrole layer on the surface of SnO2 nanoparticles [89].
ricated using flower like palladium nanoclusters grown on a F. Rigoni et al. reported the fabrication of an ammonia gas sen-
graphene electrode [138]. Palladium nanoclusters on graphene sor based on single-walled carbon nanotubes functionalized with
electrodes exhibit a response of −5% towards 10 ppm of H2 whereas indium tin oxide (ITO) nanoparticles [107]. ITO nanoparticles were
pristine graphene electrodes do not show any response towards added to a SWCNT dispersion in water. The mixture was then soni-
the same concentration of the target gas. The sensor’s response cated and drop-casted on the plastic substrate. The same deposition
shows a deviation of 2% after 100 bending cycles (1 cm bending technique was also adopted by S. Zhan et al. for the development
radius). Another study aimed to compare to effect of palladium of a flexible ethanol sensor. Following the deposition of Cr/Au IDE
and aluminum nanoparticles on the performance of graphene gas (E-beam evaporation) on polyimide foil, SnO2 powder was added
222 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

Fig. 11. Dip-coating process.

to a PDDAC solution. The mixture was ultrasonicated for 1 h before oxygen plasma, immersed in a PDDA-NaCl solution to increase the
being drop-casted on the substrate [108]. One last example is an ionic strength and then dip-coated in a SWCNT dispersion. This was
ozone gas sensor based on a drop-casting zinc oxide nanoparticles followed by sputtering Ti/Au IDEs.
solution on a kapton polyimide substrate. Titanium and platinum
IDEs were previously deposited on the substrate using photolithog- 3.2. Chemical vapor deposition
raphy and laser ablation. The aim of the study was to study the effect
of the IDE’s fabrication technique on the sensor’s response [96]. Chemical vapor deposition (CVD) is widely used in electronics
Spin-coating is the second coating technique used in flexible and microelectronics. A precursor gas is injected in a chamber with
electronics. It consists of casting a solution on a substrate while it a heated wafer to be coated. Chemical reactions around and on
is rotating (Fig. 10b.). This technique’s main advantages are its sim- the wafer will result in the deposition of a thin film. The chemical
plicity and the uniformity of the deposited layers. However, it is vapor deposition technique presents several variants that change
incompatible to the roll-to-roll process and it is not economic due to depending on the application such as the operating conditions
the high waste of materials: only 10% of the casted solution remains (low pressure LPCVD, high pressure UHVCVD), the physical char-
on the substrate after rotation [140]. In our knowledge, there are acteristics of the vapor (liquid or aerosol) and the type of the
no flexible gas sensors based on purely spin-coated sensing lay- substrate’s heating (cold-wall reactor or hot-wall reactor). Some
ers. Spin-coating is used to transfer sensing layers from a rigid to CVD processes use plasmas (PECVD), photons (PICVD), lasers and
a flexible substrate. This technique is mainly used for graphene. hot filaments (HFCVD) [143]. The main advantage of chemical vapor
For example, B. Cho et al. spin coated PMMA on a graphene layer deposition is that it enables the deposition of a wide variety of
deposited on an Ni substrate in order to transfer it to a polyimide materials with high purity. It also enables the deposition of highly
substrate for the detection of ammonia and nitrogen dioxide [139]. uniform thin films. However, the precursor needs to be highly
Chung et al. used the same technique to transfer a CVD grown volatile at near room temperatures which is not easily achieved for
graphene layer on copper foil to a PET substrate for hydrogen detec- all materials, might be toxic (Ni(CO)4 ), explosive (B2 H6 ), or corro-
tion [122]. sive (SiCl4 ). The chemical reactions inside the chamber might also
The third and last coating method is the dip-coating. Dip-coating release dangerous products such as CO, H2 , or HF.
is the controlled immersion and withdrawal of a substrate in a Chemical vapor deposition is mainly used for the fabrication
precursor solution for the fabrication of thin films using the auto- of gas sensors on rigid substrates due to the sensibility of flexi-
assembly or the sol-gel method [142] (Fig. 11). Auto-assembly ble substrates towards high temperatures. However, the literature
allows the fabrication of monolayers while sol-gel allows the depo- reports several flexible gas sensors based on a CVD grown sens-
sition of layers whose thickness is determined by the viscosity of ing layer. This technique is mainly used for graphene and carbon
the solution and the speed of immersion and withdrawal of the sub- nanotubes based flexible gas sensors. Graphene is usually grown
strate. The final state of the layer depends on the functionalization on a rigid substrate using chemical vapor deposition than trans-
of the substrate, the immersion time, the withdrawal speed, the ferred to a flexible substrate. This is achieved by spin-coating a
number of cycles, the composition and concentration of the pre- PMMA layer on top of the graphene, then etching the rigid sub-
cursor solution, the temperature and the humidity. Several articles strate. The graphene/PMMA layer is then transferred to a flexible
were dedicated to the evaluation of the performance of gas sensors substrate, and PMMA is removed using chloroform. This technique
fabricated by drop-casting the sensing layer on the substrate with was adopted by M.G. Chung et al. in order to fabricate a hydrogen
IDEs. A carbon nanotube based toluene gas sensor was reported in sensor based on a graphene sensing layer decorated with palladium
[66] : SWCNTs (0.16% wt) were dissolved in Polyoxyethylene (10) nanoparticles. Palladium nanoparticles were deposited using ther-
isooctylphenyl ether (0.6% wt.), bath-sonicated (10 min) and ultra- mal evaporation [122]. The same approach was used by Cho et al.
sonicated (35 min), probe-sonicated (24 min), in 0.5 pulses in order to study the effect of aluminum and palladium metal decorations
to form a quasi-stable dispersion. A PET substrate was immersed on the sensing performance of graphene based flexible gas sensors
in the SWCNT’s solution for 15 s and then removed for 20 s. This towards ammonia and nitrogen dioxide [139]. Metal nanoparticles
step was repeated 3 times. The electrodes were deposited using were deposited on top of the graphene layer using thermal evap-
the line patterning technique. Another example is the DMMP gas oration as well. Other researchers used this technique for growing
sensor reported by Cattanach et al. [67]. The sensor was fabricated palladium flower-like nanoclusters on graphene electrodes for
by dip-coating a PET substrate in a SWCNT dispersion using a sim- hydrogen sensing [138]. The palladium nano-flowers were formed
ilar procedure to [66]. Another way to fabricate a DMMP flexible by electrodeposition of palladium in a sulfuric acid solution on
gas sensor is proposed by Wang et al. [69]: a PI was activated by the graphene layer. The nano-flowers growth is controlled by the
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 223

concentration of sulfuric acid and the electrodeposition time. Two

other nitrogen dioxide gas sensors on paper and on PET substrate
were reported in [49,50].
Chemical vapor deposition is also used in the fabrication of car-
bon nanotube based flexible gas sensors. Sun et al. grew SWCNT
on a rigid silicium wafer using chemical vapor deposition. The
SWCNT layer was then transferred to a PET substrate using an
elastomeric PDMS stamp [137]. The carbon nanotubes were after-
wards decorated with palladium nanoparticles using electron beam
evaporation. The fabricated sensor successfully detected hydrogen.
Another nitrogen monoxide, nitrogen dioxide and ammonia sensor
was reported by Hua et al. The sensing layer consists of a CVD grown
carbon nanotube layer annealed in order to obtain a CNT-Fe2 O3
composite [74].

3.3. Physical vapor deposition

Physical vapor deposition is widely used for the fabrication

of electronic devices such as transistors, diodes, solar cells etc.
Physical vapor deposition is divided into two main techniques:
Fig. 12. Piezoelectric inkjet printing [155].
evaporation (thermal, e-beam) and sputtering (DC, DC and RF mag-
netron, RF, ion-beams, DC diode, DC triode, and AC). However, other
specialized techniques have also been used for specific applica- fully detected ammonia. Y. Sun et al. used a similar approach to
tions such as reactive evaporation, activated reactive evaporation, fabricate their hydrogen sensing device. SWCNT were deposited
ion plating, and cluster ion deposition. Physical vapor deposition is using CVD on a silicium wafer, transferred to a PET substrate, then
used to fabricate thin films using a wide variety of material such as decorated with palladium nanoparticles using e-beam evaporation
metals, alloys and some organic materials [144,145]. [137]. One last example is the hydrogen sensor reported by M.
Four main PVD techniques are used in the flexible gas sensor C. Chung et al. The fabrication process is described as follows: i)
industry: DC sputtering, RF magnetron sputtering, thermal evapo- graphene is deposited on copper foil using CVD, ii) the deposited
ration and e-beam evaporation. These techniques are used not only layer is transferred to a PET substrate, iii) graphene is decorated
to grow thin films on flexible substrate, but also to decorate sensing with palladium nanoparticles using thermal evaporation [122].
materials with noble metals nanoparticles.
Sputtering uses high energy argon ions to bombard the target 3.4. Printing
material, the cathode. Atoms of the target material are detached by
the ions and deposit on the substrate, the anode. DC sputtering uses Printing is the most used technique in flexible electronics. In fact,
a constant voltage between the anode (the substrate) and the cath- printing is a low cost roll-to-roll process. Throughout the years,
ode (target material) to ionize argon atoms and thus form a plasma. many printing techniques were developed. These techniques are
The main drawback of this technique is the low deposition rate due divided into two main categories: contact printing and non-contact
to the low ion density of argon ions near the cathode. Magnetron printing. Gravure printing, offset printing, flexography printing,
sputtering uses crossed magnetic and electrical fields to confine micro-contact printing, nano-imprinting and dry transfer print-
the plasma near the target material (cathode) which allows higher ing are the main techniques that require a contact between the
deposition rates than DC sputtering. Moreover, ions cannot reach substrate and the ink’s source. Other techniques such as screen
the substrate and thus the substrate remains intact, whereas in DC printing, inkjet printing, plasma jet printing and slot-die printing
sputtering, the substrate is slightly damaged by the ions [145,146]. use a nozzle to eject the ink on the substrate and do not require any
Literature reports three hydrogen gas sensors using sputtering for contact between the substrate and the ink’s source [25]. In this sec-
device fabrication: the first is a nanogap gas sensor based on a tion, we will only address the printing techniques commonly used
palladium thin film on a PDMS substrate deposited using DC mag- for developing chemoresistive flexible gas sensors: inkjet printing,
netron sputtering [131,133]. The second is a resistive gas sensor screen printing, nanoimprinting and gravure printing.
based on a palladium thin film on polyimide deposited using RF Inkjet printing is a non-contact printing technique that uses a
magnetron sputtering [147]. The third is based on a palladium-gold solute dissolved or dispersed in a solvent as ink. A fixed ink quan-
alloy (Au0.5 Pd0.5 ) sputtered on a polyimide substrate [148]. tity ejected through a nozzle and dries after solvent evaporation.
Unlike sputtering, electron beam evaporation uses six electron Inkjet printers use thermal, piezoelectric or electrohydrodynamic
guns to produce high energy electron beams. A part of the generated inkjet systems for nozzle actuation. These systems differ in the ink
beam is directed towards the target material in order to evaporate ejection mechanism. For instance, thermal and piezoelectric inkjet
it. The other part is directed towards the substrate to preheat it. systems apply a pulse that enables the ejection of small droplets
This enhances the adhesion between the substrate and the coating of ink (Fig. 12): this is called ‘Drop-on-demand’ (DOD) printing.
material. This technique can reach high deposition rates with low The electrohydrodynamic inkjet system enables DOD printing and
contamination [149]. Another evaporation technique is the ther- continuous spraying of the solution, depending on the applied volt-
mal evaporation which is based on heating the target material to age (AC, DC respectively). We should note that the thickness of the
evaporate it. The evaporated material then condenses on the sub- deposited layer is determined by the viscosity of the ink [151–154].
strate. This technique presents contamination issues due to the Most printed gas sensors use this deposition technique. For
evaporation of the container’s material [150]. Many researchers example, graphene oxide-silver nanoparticle ink was inkjet printed
used evaporation for sensing material decoration. For example, on a PET substrate for ammonia detection [124]. Pristine graphene
K. Lee et al. spray deposited SWCNTs on a PET substrate. After- was also inkjet printed on a PET substrate and investigated for
wards, they decorated the sensing layer with gold nanoparticles its sensibility towards VOCs (NO2 , NH3 , methanol, ethanol, Cl2 )
using e-beam evaporation [136]. The fabricated sensor success- [120]. Other ammonia sensors used polymers based inks such
224 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

minated polystyrene-co- maleic anhydride (Cumene-PSMA); (iii)

poly (styrene-co-maleic acid) partial isobutyl/methyl mixed ester
(PSE) and (iv) polyvinylpyrrolidon (PVP) [29]. Inkjet printing was
also used for the fabrication of metal oxide gas sensors. In fact, M.
Rieu et al. printed an SnO2 layer on polyimide foil. The main draw-
back of this device is it’s high operating temperature (200–300 ◦ C)
[99].
Even though inkjet printing is the most commonly used tech-
nique for the fabrication of flexible gas sensors, this technique
presents some limitations. In fact, inkjet-printing cannot be used on
Fig. 13. Plasma jet printing. textiles. Moreover, liquid ink is a requirement, and not all materials
easily achieve a liquid phase. Plasma jet printing allows the depo-
sition of an aerosol on a flexible substrate at atmospheric pressure.
as the poly(3,4-ethylene dioxythiophene) doped with polystyrene Unlike conventional aerosol printing, which requires high temper-
sulfonated acid (PEDOT/PSS) based sensor on commercial pho- atures to achieve a smooth layer, plasma jet printing operates at
topaper proposed by Wongchoosuk et al. [56]. This work was temperatures as low as 40◦ C. A plasma jet printer consists of: i) a
followed by the work of Seekaew et al. [84] who fabricated a quartz nozzle with two electrodes fed by a high current; ii) a neb-
graphene-PEDOT/PSS composite ink for ammonia gas sensing as ulizer to generate the aerosol from a colloid of the target material
well. PET substrates were used in the process. The functionalization iii) Helium as a source of plasma in the quartz nozzle. In fact, the
of PEDOT/PSS with graphene doubled the sensibility of the sensor aerosol is transported by helium gas into the quartz nozzle and is
towards 500 ppm of ammonia (4.4% for PEDOT/PSS and 9.6% for deposited on the substrate, placed at a close distance to the nozzle
graphene-PEDOT/PSS composite). K. Crowley et al. inkjet printed a (Fig. 13) [55,156–158]. R. Gandhiraman and al. reported a flexi-
polyaniline nanoparticles ink on PET substrates for ammonia detec- ble ammonia sensor on paper substrate by plasma jet printing of a
tion [79]. P. Lorwongtragool fabricated an array of eight sensors MWCNTs colloid. The sensor has a sensibility of 4% towards 60 ppm
based on double layer inkjet printing and single (blended) layer of ammonia [55].
inkjet printing of CNTs and polymers. For the double layer inkjet Screen printing is another non-contact printing technique used
printing, MWCNT were printed on the substrate and dried. This in the flexible gas industry. A screen printer’s key components are
step was followed by inkjet printing of a polymer ink. As for sin- the mesh and the squeegee. The squeegee forces the ink to go
gle layer inkjet printing, a MWCNT-polymer ink was directly inkjet through the opened pattern of the mesh to the substrate. There are
printed on the substrate. Moreover, four polymers were used in two types of screen printing: flatbed and rotary screen printing. In
the fabrication process: polyvinyl chloride (PVC); (ii) cumene ter- flat-bed printing, the screen should be lowered on the substrate

Fig. 14. Screen printing a) flatbed b) rotary.

 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 225

ver nanoparticles and is printed on a polyimide substrate using

gravure printing [125]. The last two gas sensors were proposed by
[Link] et al. for ammonia and humidity detection. The ammo-
nia and humidity sensors were based on screen printed polyaniline
hydrochloride colloid and PEDOT/PSS (respectively) [161].
Nanoimprint lithography is the last printing technique used for
the development of flexible gas sensors. This process is explained as
follows: i) a solution is spin-coated on the substrate ii) a patterned
mold is pressed against the spin coated solution at a determined
temperature and pressure in order to imprint the layer iii) After
demolding, patterned layer is thermally treated (Fig. 16) [159]. To
our knowledge, the only flexible gas sensor that uses this technique
is proposed by S.H. Lim et al. The fabrication process started with
a direct nanoimprint lithography of palladium benzyl mercaptide
on a rigid substrate. Afterwards, the palladium layer was annealed
at 250 ◦ C for one hour to form palladium gratings. The resulting
patterned palladium film was transferred into a polycarbonate film
to obtain the flexible sensor [135].

3.5. Self-assembly

Self-assembly is a bottom-up deposition technique that assem-

bles nanostructures in a structured manner by creating covalent
or non-covalent bonds between the material to be deposited and
Fig. 15. Gravure printing [155].
the target substrate. The main advantages of this technique is its
low cost, simplicity, and the mechanical strength and stability of
before the squeegee swipes it and pushes the ink onto the sub- the self-assembled structure [162]. Several techniques are used for
strate. In contrast, rotary screen printers have a cylindrical mesh self-assemble molecules. They are summarized in Fig. 17.
that rotates around a fixed squeegee [155,159] (Fig. 14). Few flexi- Both covalent and non-covalent anchoring were used in the
ble chemoresistive gas sensors are developed using this technology. development of flexible gas sensors. P.G.- Su and al. fabricated
This is mainly due to the advantages of inkjet printing in terms of a hydrogen gas sensor based on self-assembly of MWCNT mod-
R2R, and reliability [160]. ified with palladium nanoparticles [132]. The fabrication process
Gravure printing is the fourth printing technique used for devel- is explained as follows: i) sputtering of Cr/Au electrodes on a PET
oping flexible gas sensors. A gravure printer is composed of a rolling substrate, ii) hydrophilic surface treatment, iii) negatively charg-
cylinder and a printing cylinder. The rolling cylinder coated with ing the electrodes using an MPS solution, iv) deposition of two
copper. Afterwards, a pattern is engraved on the cylinder using layers of PSSMA and PAH by immersing the substrate in the solu-
laser or electrochemical method. The cylinder is then protected tion, v) deposition of a MWCNT/PAH layer in the same manner,
from wear and tear using a chrome coating layer. The ink is intro- vi) deposition of palladium using a PdCl2 solution and sodium
duced using a nozzle on top of the cylinder or an ink bath beneath borohydride as a reducing agent. The same research group used
it. The excess is then removed using a doctor blade. The ink is trans- a similar process to develop an ammonia gas sensor based on self-
ferred to a rolling substrate between the patterned cylinder and a assembled polypyrrole [163]. The substrate and electrodes were
printing cylinder (Fig. 15). The quality of the printed films depends made hydrophilic. The electrodes were modified afterwards using
on the pattern and on the viscosity of the ink. an MPS solution. This was followed by the deposition of a bilayer of
Four flexible gas sensors using gravure printing are reported PAH/PSS using the same technique used in [132]. The Ppy layer was
in the literature. The first is a WO3-PEDOT/PSS nitrogen dioxide in-situ polymerized using a solution containing the monomer and
gas sensor [106]. The WO3 -PEDOT/PSS composite ink was gravure an oxidizing agent (FeCl3 ). The sensor successfully detected ammo-
printed on a polyimide substrate and evaluated for its gas sensing nia. The researcher modified the layer with platinum nanoparticles
capabilities. The second gas sensor detects ammonia and nitrogen by immersing the sensor in a PtCl6 2− acqueous solution and using
dioxide. The ink is fabricated using graphene modified with sil- NaBH4 as a reducing agent. Metal functionalization resulted in a

Fig. 16. Nanoimprint lithography.

226 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

Fig. 17. The classification of self-assembly techniques.

change of the target gas. The device is now sensitive towards hydro- dedope the resulting emeraldine Pani salt. Another Pani/FMWCNT
gen [85]. Another proposed sensor is a covalently anchored reduced (functionalized multi-walled carbon nanotubes) based gas sensor
graphene oxide sensing layer on an PET substrate for nitrogen diox- was fabricated using this same technique [83]. First, PET substrates
ide detection. Gold electrodes were modified using cysteamine were made hydrophilic using oxygen plasma. Second, APS was
hydrochloride. The substrate was then immersed in a graphene gradually added to a mixture of hydrochloride and a previously
oxide suspension with EDC/NHS as a peptide coupling reagent. prepared FMWCNT solution. Third, aniline was dropped in the
The deposited graphene oxide was reduced into rGO using sodium resulting mixture and the hydrophilic substrates were immersed
borohydride [121]. One last example is a humidity gas sensor based in the solution for film deposition. One last example is the sen-
on the layer by layer nanoassembly technique. A plastic substrate sor proposed by Guo and al. [88]. The fabrication process is quite
was successively immersed in PDDA and PSS solutions for charge similar to [83]. A hydrophilic PET substrate was immersed in an
enhancement. The substrate was then immersed successively in rGO suspension containing an aniline monomer. An APS solution
a PDDA solution and a GO dispersion for 5 consecutive times. The was prepared by dissolving APS in HCl. The mixture was gradually
deposited graphene oxide film was reduced using hydrobromic acid added to the rGO-aniline suspension to initiate the polymerization
(HBr) [123]. The literature reports other self-assembled gas sensors process.
[162,164,165].

3.7. In-situ growth

3.6. In-situ polymerization
The last fabrication technique is the in-situ growth of materials
In-situ polymerization is a technique used to assemble a poly- on flexible substrates using either hydrothermal growth or in situ-
mer thin film on a substrate. The substrate is immersed in a solution growth in a CVD chamber using a catalyst.
containing a monomer and an oxidizing agent. The resulting layer Hydrothemal growth is mostly used with carbon nanotubes
might be modified with other components either by introducing and metal oxides, zinc oxide particularly. Hydrothermal growth
nanoparticles in the monomer’s suspension, or by further treat- consists of immersing a substrate, with a metal oxide seed layer,
ment of the grown polymer film. in an autoclave containing a precursor solution at high temper-
In situ-polymerization is the most commonly used technique ature. Many researchers used the hydrothermal method to grow
for the fabrication of polymer based flexible gas sensors. This zinc oxide nanorods on flexible substrates either for their sensing
is mainly due to its simplicity and low cost. L. Kumar and al. capabilities, or to use them as a base to fabricate palladium nano-
used this technique to develop an ammonia gas senor based on tubes. S. Mohammad and al. report the fabrication of a hydrogen
an in situ-polymerized polyaniline sensing layer. The substrates gas sensor based on vertically aligned carbon nanotubes on a nylon
were immersed in a hydrochloride solution containing aniline, the substrate [98]. First, a ZnO layer was sputtered on the nylon sub-
monomer. This was followed by gradually adding an ammonium strate. Second, the substrate was thermally treated at 180 ◦ C to
persulfate (APS) solution to the mixture. The APS solution acts as form ZnO seeds for hydrothermal growth. Finally, the substrates
an oxidant to initiate the polymerization process [166]. The same were placed in a tube containing a precursor solution of zinc nitrate
technique was reported for the fabrication of a Pani-nylon 6 com- hexahydrate [Zn(NO3 )2.6 H2 O] and hexamethylamine (HMT) (1:1)
posites for ammonia detection. Nylon 6 fabrics were immersed in at 85 ◦ C for 4.5 h. Another research group used proposed a palla-
a hydrochloride solution containing the monomer. To initiate the dium nanorods flexible hydrogen gas sensor using hydrothermally
polymerization process, ammonium peroxydisulfate dissolved in grown zinc oxide nanorods as a base. The substrates were coated
an HCl solution were added to the mixture. This was followed by with a ZnO dispersion to form a ZnO nanocrystals seed layer. The
de-doping of the resulting Pani using ammonium hydroxide [10]. hydrothermal growth was carried in a container with a mixture of
The literature also reports functionalized polymers for gas sens- zinc nitrate, HMTA and PEI at 95 ◦ C for 2 h. The grown nanorods
ing fabricated by the same technique. For example, Bandgar et al. were covered with palladium nanoparticles by immersing the sub-
reported an ammonia gas sensor based on a Pani-iron oxide sens- strate in a palladium precursor solution. The ZnO nanowires were
ing layer in-situ grown on a PET substrate [86]. The substrate was dissolved automatically, with no further treatment, by hydrogen
immersed in an aniline monomer solution with APS as an oxidant in ions generated during the reduction of the palladium precursor
an HCl solution. In order to functionalize the resulting sensing layer, [134]. T.R. Rashid and al. developed a flexible hydrogen gas sensor
iron oxide nanoparticles were synthetized using the sol-gel method on polyimide tape using this same technique. A ZnO seed layer was
and introduced into the Pani mixture with ammonium hydroxide to fabricated using the sol-gel method. Afterwards ZnO nanorodes
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 227

were grown on the substrate using the hydrothermal method. on metal oxides, polymers, CNTs, graphene and GO. The main
Finally, platinum was sputtered on the nanorods and acts as a cat- drawback of metals is their low sensibility towards hydrogen if
alyst to enhance the reaction between zinc oxide ant the target gas not functionalized by other materials such as graphene, CNTs
[105]. and polymers. Metal oxides provide sensibilities as high as 400%
In situ- growth is also used for the fabrication of carbon nano- towards several toxic gases, including hydrogen, and have a fast
tubes based gas sensors [76]. For example, carbon nanotubes were response and recovery. However, the good operation of a MOX sen-
grown on a graphene oxide layer for nitrogen dioxide gas sensing. sor requires heating (up to 300 ◦ C for flexible sensors) and thus
Graphene oxide was spin coated on a silicium substrate. This was the integration of heaters in the device. This increases the device’s
followed by the deposition of a patterned iron catalyst. The car- power consumption and might damage flexible substrates, that
bon nanotubes were then grown on the substrate using plasma usually operate at temperatures below 200 ◦ C, except for polyimide
enhanced chemical vapor deposition [75]. The resulting sensing that can withstand temperatures up to 400 ◦ C. MOX nanostructures
layer was transferred to a flexible polyimide substrate. (nanowires, nanoribbons, nanoflowers. . .) have a lower operating
Table 6 summarizes the previous studies of flexible gas sensors temperature than MOX thin films. However, MOX nanostructures
arranged based on the target gas. still exhibit a better response at temperatures around 180◦ C. DMMP
and DIMP gases are mainly detected by carbon nanotubes. CNTs also
exhibit a good response towards others gases such as NH3 , NO2 , Cl2 ,
4. Application in wearable devices
O3 . . . In terms of substrates, CNTs are ideal for paper substrates
since they naturally bond to cellulose. Other interesting carbon-
Wearable devices have attracted the attention of researchers
based materials are graphene and graphene oxide. Graphene and
ever since the 13th century, when the first wearable eyeglasses
graphene oxide are mainly used for NO2 detection. They might also
were introduced. In 1510, the first wearable mechanical clock, the
be used for ammonia and humidity sensing. These two material
‘Nuremberg Egg’, was introduced. These inventions were followed
have the ability to be electrostatically linked to flexible substrates.
by many other: the abacus ring (17th century), the calculator wrist-
The last sensing material discussed in this paper is polymers. Poly-
watch (1975), the Sony Walkman (1979) and digital hearing aids
mers have good ammonia sensing capabilities. This type of material
(1980s). Wearable electronic devices gained their popularity in
is flexible and might be used both as substrate and as a sensing layer
2002 with the invention of the Bluetooth headset. This invention
(ex. PDMS for humidity sensing). The main challenge faced when
was followed by Nike+ (2006), FitBit classic (2008), google glasses
dealing with graphene, graphene oxide, polymers or CNTs is their
(2013), the solar power jacket by tommy Hilfiger (2014), activity
irreversibility after their exposure to a target gas.
trackers (2014) and apple watch (2015).
Many fabrication techniques were tested throughout the years.
Nowadays, wearable electronics for healthcare monitoring are
Inkjet-printing, dip-coating, in situ-growth and polymerization
of increasing interest [167–169]. Gas sensors are a key component
are commonly used for the deposition of a gas sensing layer at
of a wearable gas sensing system for ambient air quality monitor-
low temperature. We should note that some techniques require
ing. The aim of such systems is to detect toxic gases that might
high temperatures or heat treatment (CVD, PVD, nanoimprint
threaten human health. The development of such systems requires
lithography. . .) and are only used prior to the transfer of the layer
the fabrication of flexible gas sensors, operating at room temper-
to a flexible substrate. Moreover, some materials cannot easily
ature and with a low power consumption. Several articles discuss
achieve a liquid phase which limits the use of some printing tech-
the implementation of flexible gas sensor on textiles. Seesaard et al.
niques. All these limitations are leading the development of new
reported the fabrication of a textile gas sensor based on drop-
techniques such as plasma jet printing that operates at tempera-
casted polymer/MWCNTs on a cotton textile embroided with silver
tures as low as 40◦ C and deposits an aerosol on a substrate.
interdigitated electrodes. Four polymers were used: PVC, Cumene-
Last, the performance of a gas sensor is evaluated based on its
PSMA, PSE and PVP. The fabricated sensors were able to detect
sensibility, selectivity, reversibility, fast response and recovery and
ammonia, trimethylamine and ethanol [59]. Another wearable gas
flexibility or ability to be bent or stretched without any perfor-
sensor was proposed by Y. Yun et al, who fabricated a graphene
mance loss. However, there is still no sensing material that satisfies
oxide based ammonia gas sensor on a cotton yarn and then wove it
all these conditions. Most sensors lack of selectivity, reversibil-
into a lab coat [58]. Zheng et al. proposed a ZnO nanoparticle based
ity and/or flexibility. With the increasing interest in flexible gas
ethanol gas sensor on a PET substrate and tested its flexibility for an
sensors, future work will hopefully include new fabrication tech-
extreme bending angle (90◦ ) for its possible integration in a wear-
niques along with new composite materials and substrates aiming
able device [170]. One last example is the stretchable DMMP sensor
to overcome t toward the integration of gas sensors in e-textiles,
based on a graphene and silver nanowires composite sensing layer
smart-packaging and in the most unconventional places.
integrated on the leaves of a plant, in a wristwatch and in a bicycle
light [171].
Acknowledgments

5. Conclusion Rouba Alrammouz, would like to thank the Lebanese Council for
Scientific Research (CNRSL), the University of Montpellier (UM) and
The last twenty years witnessed a significant growth in the flex- the AUF (Agence Universitaire de la Francophonie) for providing the
ible electronics industry, especially in the development of flexible PhD grant.
gas sensors. Many materials have been investigated for their gas
sensing abilities and hundreds of flexible gas sensing devices were References
fabricated for the detection of many gaseous species such as NH3 ,
NO, NO2 , CO, CO2 , H2 S, H2 O and VOCs. [1] A. Prüss-Üstün, C. Corvalán, Preventing Disease Through Healthy
Environment: Towards an Estimate of the Environmental Burden of Disease,
For instance, metals, particularly palladium, are ideal for detect-
World Health Organization, Switzerland, 2006 [Link]
ing hydrogen. Hydrogen detectors are mostly based on a metal quantifying ehimpacts/publications/[Link].
sensing layer or on a sensing material (MOX, CNTs, polymer, [2] WHO, Ambient Air Pollution: A Global Assessment of Exposure and Burden
graphene or GO) doped with a metal. This is mainly due to their of Disease, World Health Organization, Switzerland, 2016 [Link]
int/phe/publications/air-pollution-global-assessment/en/.
ability to absorb hydrogen several times their volume (900 times [3] WHO, Air Quality Guidelines. Global Update 2005. Particulate Matter,
for palladium) at room temperature and to their catalytic effect Ozone, Nitrogen Dioxide and Sulfur Dioxide, World Health Organization,
228 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

Denmark, 2006, pp. 496 [Link] [31] J.F. Chang, H.H. Kuo, I.C. Leu, M.H. Hon, The effects of thickness and
environment-and-health/air-quality/publications/pre2009/air-quality- operation temperature on ZnO:Al thin film CO gas sensor, Sens. Actuators B
guidelines.-global-update-2005.-particulate-matter,-ozone,-nitrogen- Chem. 84 (no. 2-3) (2002) 258–264.
dioxide-and-sulfur-dioxide. [32] Q. Lipeng, X. Jiaqiang, D. Xiaowen, P. Qingyi, C. Zhixuan, X. Qun, L. Feng, The
[4] Clean Air Act, U. S. E. P. Agency 85, 1990. [Link] template-free synthesis of square-shaped SnO 2 nanowires: the temperature
act-overview/clean-air-act-requirements-and-history. effect and acetone gas sensors, Nanotechnology 19 (no. 18) (2008) 185705.
[5] Air Quality Standards, 2008. [Link] [33] (2018). Polyethylene terephthalate (Polyester, PET, PETP)- Film- Material
naaqs-table. information. Available: [Link]
[6] Air Quality - Existing Legislation, E. Commission, 2017. [Link] [34] Polyesters (Thermoplastic) PETP, PBT, PET. 2018 Available: [Link]
environment/air/quality/existing [Link]. [Link]/plastipedia/polymers/[Link].
[7] G. Chanson, D. Pugh, Environmental Gas Sensors 2017-2027 : Technologies, [35] P. Yang, F. Tian, Y. Ohki, Dielectric properties of poly(ethylene terephthalate)
Manufacturers, Forecasts, 2017 [Link] and poly(ethylene 2,6-naphthalate), IEEE Trans. Dielectr. Electr. Insul. 21
reports/[Link]. (no. 5) (2014) 2310–2317.
[8] G. Mattana, C. Ataman, J.J. Ruan, A.V. Quintero, G. Nisato, G. Tröster, D. [36] W.M. Bundy, J.N. Ishley, Kaolin in paper filling and coating, Appl. Clay Sci. 5
Briand, [Link]. Rooij, Woven temperature and humidity sensors on flexible (no. 5-6) (1991) 397–420.
plastic substrates for E-Textile applications, IEEE Sens. J. 13 (no. 10) (2013) [37] J.-H. Kim, B.S. Shim, H.S. Kim, Y.-J. Lee, S.-K. Min, D. Jang, Z. Abas, J. Kim,
3901–3909. Review of nanocellulose for sustainable future materials, Int. J. Precis. Eng.
[9] J.-W. Han, B. Kim, J. Li, M. Meyyappan, A carbon nanotube based ammonia Manuf. Technol. 2 (no. 2) (2015) 197–213, April 01.
sensor on cotton textile, Appl. Phys. Lett. 102 (2013) 193104. [38] G. Zheng, Y. Cui, E. Karabulut, L. Wågberg, H. Zhu, L. Hu, Nanostructured
[10] K.H. Hong, K.W. Oh, T.J. Kang, Polyaniline–Nylon 6 composite fabric for paper for flexible energy and electronic devices, MRS Bull. 38 (no. 4) (2013)
ammonia gas sensor, J. Appl. Polym. Sci. 92 (2004) 37–42. 320–325.
[11] J. Koskela, J. Sarfraz, P. Ihalainen, A. Määttänen, P. Pulkkinen, H. Tenhu, T. [39] J. Zhang, G. Jiang, M. Goledzinowski, F.J.E. Comeau, K. Li, T. Cumberland, J.
Nieminen, A. Kilpelä, J. Peltonen, Monitoring the quality of raw poultry by Lenos, P. Xu, M. Li, A. Yu, Z. Chen, Green solid electrolyte with
detecting hydrogen sulfide with printed sensors, Sens. Actuators B Chem. cofunctionalized Nanocellulose/Graphene oxide interpenetrating network
218 (2015) 89–96. for electrochemical gas sensors, Small Methods 1 (2017) 1700237.
[12] S.F. Liu, A.R. Petty, G.T. Sazama, T.M. Swager, Single-walled carbon [40] L. Hu, G. Zheng, J. Yao, N. Liu, B. Weil, M. Eskilsson, E. Karabulut, Z. Ruan, S.
Nanotube/Metalloporphyrin composites for the chemiresistive detection of Fan, J.T. Bloking, M.D. McGehee, L. Wågberg, Y. Cui, Transparent and
amines and meat spoilage, Angew. Chemie Int. Ed. 54 (no. 22) (2015) conductive paper from nanocellulose fibers, Energy Environ. Sci. 6 (no. 2)
6554–6557. (2013) 513–518.
[13] E. Abad, S. Zampolli, S. Marco, A. Scorzoni, B. Mazzolai, A. Juarros, D. Gómez, [41] T. Abitbol, A. Rivkin, Y. Cao, Y. Nevo, E. Abraham, T. Ben-Shalom, S. Lapidot,
I. Elmi, G.-C. Cardinali, J.M. Gómez, F. Palacio, M. Cicioni, A. Mondini, T. O. Shoseyov, Nanocellulose, a tiny fiber with huge applications, Curr. Opin.
Becker, I. Sayhan, Flexible tag microlab development: gas sensors Biotechnol. 39 (2016) 76–88.
integration in RFID flexible tags for food logistic, Sens. Actuators B Chem. [42] Y.J. Kang, S.J. Un, S.-S. Lee, B.-Y. Kim, J.H. Kim, H. Chung, S.-Y. Lee, W. Kim,
127 (no. 1) (2007) 2–7. All-solid-State flexible supercapacitors fabricated with bacterial
[14] V. Lakafosis, A. Rida, R. Vyas, L. Yang, S. Nikolaou, M.M. Tentzeris, Progress nanocellulose papers, carbon nanotubes, and triblock-copolymer ion gels,
towards the first wireless sensor networks consisting of inkjet-printed, ACS Nano 6 (no. 7) (2012) 6400–6406.
paper-based rfid-enabled sensor tags, Proc. IEEE 98 (no. 9) (2010) [43] J.-W. Han, B. Kim, J. Li, M. Meyyappan, Carbon nanotube based humidity
1601–1609. sensor on cellulose paper, J. Phys. Chem. C 116 (no. 41) (2012) 22094–22097.
[15] I. Fratoddi, I. Venditti, C. Cametti, M.V. Russo, Chemiresistive [44] J.-W. Han, B. Kim, J. Li, M. Meyyappan, A carbon nanotube based ammonia
polyaniline-based gas sensors: a mini review, Sens. Actuators B Chem. 220 sensor on cellulose paper, RSC Adv. 4 (no. 2) (2014) 549–553.
(2015) 534–548. [45] K. Joo-Hyung, M. Seongcheol, U.K. Hyun, Y. Gyu-Young, K. Jaehwan,
[16] S.-W. Chiu, K.-T. Tang, Towards a chemiresistive sensor-integrated Disposable chemical sensors and biosensors made on cellulose paper,
electronic nose: a review, Sensors (Basel, Switzerland) 13 (no. 10) (2013) Nanotechnology 25 (no. 9) (2014) 092001.
14214–14247, 10/22 2013. [46] S. Mun, Y. Chen, J. Kim, Cellulose–titanium dioxide–multiwalled carbon
[17] J. COURBAT, Gas Sensors on Plastic Foil With Reduced Power Consumption nanotube hybrid nanocomposite and its ammonia gas sensing properties at
for Wireless Applications, PhD thesis manuscript, Ecole polytechnique room temperature, Sens. Actuators B Chem. 171-172 (2012) 1186–1191.
fédérale de Lausanne, Suisse, 2010. [47] S. Ammu, V. Dua, S.-R. Agnihotra, S.P. Surwade, A. Phulgirkar, S. Patel, S.K.
[18] M.C. Santos, A.G.C. Bianchi, D.M. Ushizima, F.J. Pavinatto, R.F. Bianchi, Manohar, Flexible, all-organic chemiresistor for detecting chemically
Ammonia gas sensor based on the frequency dependant impedance aggressive vapors, J. Am. Chem. Soc. 134 (no. 10) (2012) 4553–4556.
characteristics of ultrathin polyaniline films, Sens. Actuators A Phys. 253 [48] Y. Chen, S.-D. Jang, J. Kim, Gas sensing properties of gallium nitride-coated
(2017) 156–164. cellulose nanocomposite, Sens. Lett. 10 (no. 3-4) (2012) 748–753.
[19] A.L. Spetz, M. Skoglundh, L. Ojamäe, FET gas-sensing mechanism, [49] G. Yang, C. Lee, J. Kim, F. Ren, S.J. Pearton, Flexible graphene-based chemical
experimental and theoretical studies, in: E. Comini, G. Faglia, G. Sberveglieri sensors on paper substrates, Phys. Chem. Chem. Phys. 15 (no. 6) (2013)
(Eds.), Solid State Gas Sensing, Springer US, Boston, MA, 2009, pp. 1–27. 1798–1801.
[20] C. Caliendo, P. Verardi, E. Verona, A.D.’ Amico, C. Di Natale, G. Saggio, M. [50] C. Lee, J. Ahn, K.B. Lee, D. Kim, J. Kim, Graphene-based flexible NO2 chemical
Serafini, R. Paolesse, S.E. Huq, Advances in SAW-based gas sensors, Smart sensors, Thin Solid Films 520 (no. 16) (2012) 5459–5462.
Mater. Struct. 6 (1997) 689–699. [51] J. Sarfraz, A. Määttänen, P. Ihalainen, M. Keppeler, M. Lindén, J. Peltonen,
[21] J. Devkota, P.R. Ohodnicki, D.W. Greve, SAW sensors for chemical vapors and Printed copper acetate based H2S sensor on paper substrate, Sens. Actuators
gases, Sensors (Basel, Switzerland) 17 (no. 4) (2017) 801. B Chem. 173 (2012) 868–873.
[22] H.P. Lang, Cantilever-based gas sensing, in: E. Comini, G. Faglia, G. [52] J. Sarfraz, P. Ihalainen, A. Määttänen, J. Peltonen, M. Lindén, Printed
Sberveglieri (Eds.), Solid State Gas Sensing, Springer US, Boston, MA, 2009, hydrogen sulfide gas sensor on paper substrate based on polyaniline
pp. 1–24. composite, Thin Solid Films 534 (2013) 621–628.
[23] K. Eguchi, Optical gas sensors, in: G. Sberveglieri (Ed.), Gas Sensors: [53] J. Hassinen, J. Kauppila, J. Leiro, A. Määttänen, P. Ihalainen, J. Peltonen, J.
Principles, Operation and Developments, Springer Netherlands, Dordrecht, Lukkari, Low-cost reduced graphene oxide-based conductometric nitrogen
1992, pp. 307–328. dioxide-sensitive sensor on paper, Anal. Bioanal. Chem. 405 (no. 11) (2013)
[24] J. Hodgkinson, R.P. Tatam, Optical gas sensing: a review, Meas. Sci. Technol. 3611–3617.
24 (no. 1) (2013) 012004. [54] R. Ghosh, A. Singh, S. Santra, S.K. Ray, A. Chandra, P.K. Guha, Highly sensitive
[25] S. Alberto, W.S. Wong, Flexible Electronics: Materials and Applications large-area multi-layered graphene-based flexible ammonia sensor, Sens.
(Electronic Materials: Science & Technology), Springer, New York, 2009. Actuators B Chem. 205 (2014) 67–73.
[26] D. Briand, F. Molina-Lopez, A.V. Quintero, C. Ataman, J. Courbat, [Link]. Rooij, [55] R.P. Gandhiraman, E. Singh, D.C. Diaz-Cartagena, D. Nordlund, J. Koehne, M.
Why going towards plastic and flexible sensors? Procedia Eng. 25 (2011) Meyyappan, Plasma jet printing for flexible substrates, Appl. Phys. Lett. 108
8–15. (no. 12) (2016) 123103.
[27] D. Briand, A. Oprea, J. Courbat, N. Bârsan, Making environmental sensors on [56] C. Wongchoosuk, P. Jangtawee, S. Lokavee, S. Udomrat, P. Sudkeaw, T.
plastic foil, Mater. Today 14 (no. 9) (2011) 416–423. Kerdcharoen, Novel Flexible NH3 Gas Sensor Prepared By Ink-Jet Printing
[28] V. Zardetto, T.M. Brown, A. Reale, A.D. Carlo, Substrates for flexible Technique, Adv. Mat. Res. 506 (2012) 39–42.
electronics: a practical investigation on the electrical, film flexibility, optical, [57] R. Alrammouz, B. Sorli, J. Podlecki, R. Garcia, A. Vena, P. Abboud, R. Habchi,
temperature, and solvent resistance properties, J. Polym. Sci. Part B: Polym. Graphene oxide based ammonia gas sensor on paper substrate, Presented at
Phys. 49 (no. 9) (2011) 638–648. the IdtechEx Show (2017).
[29] P. Lorwongtragool, E. Sowade, N. Watthanawisuth, R.R. Baumann, T. [58] Y.J. Yun, W.G. Hong, N.-J. Choi, B.H. Kim, H.-L.Y. Jun, Ultrasensitive and
Kerdcharoen, A novel wearable electronic nose for healthcare based on highly selective graphene-based single yarn for use in wearable gas sensor,
flexible printed chemical sensor array, Sensors 14 (no. 10) (2014) Sci. Rep. 5 (2015) 10904.
19700–19712. [59] T. Seesaard, P. Lorwongtragool, T. Kerdcharoen, Development of
[30] C. Wang, L. Yin, L. Zhang, D. Xiang, R. Gao, Metal oxide gas sensors: fabric-based chemical gas sensors for use as wearable electronic noses,
sensitivity and influencing factors, Sensors 10 (no. 3) (2010) 2088–2106. Sensors 15 (no. 1) (2015) 1885–1902.
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 229

[60] Q. Cao, J.A. Rogers, Ultrathin films of single-walled carbon nanotubes for [89] X. Chen, D. Li, S. Liang, X. Li, S. Zhan, M. Liu, Novel flexible room temperature
electronics and sensors: a review of fundamental and applied aspects, Adv. NO2 gas sensor based on polypyrrole coated SnO2 nanoparticles, Presented
Mater. 21 (no. 1) (2009) 29–53. at the The 8th Annual IEEE International Conference on Nano/Micro
[61] L.-C. Qin, Determination of the chiral indices (n,m) of carbon nanotubes by Engineered and Molecular Systems (2013).
electron diffraction, Phys. Chem. Chem. Phys. 9 (no. 1) (2007) 31–48. [90] G. Heiland, Zum einfluss von wasserstoff auf die elektrische leitfähigkeit von
[62] M. Meyyappan, Carbon Nanotubes, Science and Applications, 1st edition, ZnO-kristallen, Zeitschrift für Physik 148 (no. 1) (1957) 15–27.
CRC Press, 2004, ISBN 9780849321115. [91] H.-J. Kim, J.-H. Lee, Highly sensitive and selective gas sensors using p-type
[63] S. Park, M. Vosguerichian, Z. Bao, A review of fabrication and applications of oxide semiconductors: overview, Sens. Actuators B Chem. 192 (2014)
carbon nanotube film-based flexible electronics, Nanoscale 5 (no. 5) (2013) 607–627.
1727–1752. [92] R. Ramamoorthy, P.K. Dutta, S.A. Akbar, Oxygen sensors: materials, methods,
[64] Y. Wang, J.T.W. Yeow, A review of carbon nanotubes-based gas sensors, J. designs and applications, J. Mater. Sci. 38 (no. 21) (2003) 4271–4282.
Sens. 2009 (2009) 493904. [93] G.K. Mani, J.B.B. Rayappan, A highly selective room temperature ammonia
[65] C.-S. Woo, C.-H. Lim, C.-W. Cho, B. Park, H. Ju, D.-H. Min, C.-J. Lee, S.-B. Lee, sensor using spray deposited zinc oxide thin film, Sens. Actuators B Chem.
Fabrication of flexible and transparent single-walled carbon nanotube gas 183 (2013) 459–466.
sensors by vacuum filtration and poly(dimethyl siloxane) mold transfer, [94] I.H. Kadhim, H.A. Hassan, Room temperature hydrogen gas sensor based on
Microelectron. Eng. 84 (no. 5-8) (2007) 1610–1613. nanocrystalline SnO2 thin film using sol–gel spin coating technique, J Mater
[66] K. Parikh, K. Cattanach, R. Rao, D.-S. Suh, A. Wu, S.K. Manohar, Flexible Sci: Mater Electron 27 (no. 5) (2016) 4356–4362.
vapour sensors using single walled carbon nanotubes, Sens. Actuators B [95] A. Dey, Semiconductor metal oxide gas sensors: a review, Mater. Sci. Eng. B
Chem. 113 (no. 1) (2006) 55–63. 229 (2018) 206–217.
[67] K. Cattanach, R.D. Kulkarni, M. Kozlov, S.K. Manohar, Flexible carbon [96] M. Acuautla, S. Bernardini, L. Gallais, T. Fiorido, L. Patout, M. Bendahan,
nanotube sensors for nerve agent simulants, Nanotechnology 17 (no. 16) Ozone flexible sensors fabricated by photolithography and laser ablation
(2006) 4123–4128. processes based on ZnO nanoparticles, Sens. Actuators B Chem. 203 (2014)
[68] S.J. Young, Z.D. Lin, Sensing performance of carbon dioxide gas sensors with 602–611.
carbon nanotubes on plastic substrate, Adv. Solid state electrochem. Sci. [97] E. Comini, Integration of metal oxide nanowires in flexible gas sensing
Technol. 6 (no. 5) (2017) M72–M74. devices, Sensors 13 (no. 8) (2013) 10659–10673.
[69] Y. Wang, Z. Yang, Z. Hou, D. Xu, L. Wei, E.S.-W. Kong, Y. Zhang, Flexible gas [98] S.M. Mohammad, Z. Hassan, R.A. Talib, N.M. Ahmed, M.A. Al-Azawi, N.M.
sensors with assembled carbon nanotube thin films for DMMP vapor Abd-Alghafour, C.W. Chin, N.H. Al-Hardan, Fabrication of a highly flexible
detection, Sens. Actuators B Chem. 150 (no. 2) (2010) 708–714. low-cost H2 gas sensor using ZnO nanorods grown on an ultra-thin nylon
[70] A. Abdellah, A. Abdelhalim, F. Loghin, P. Köhler, Z. Ahmad, G. Scarpa, P. Lugli, substrate, J. Mater. Sci.: Mater Electron 27 (no. 9) (2016) 9461–9469.
Flexible carbon nanotube based gas sensors fabricated by large-scale spray [99] M. Rieu, M. Camara, G. Tournier, J.-P. Viricelle, C. Pijolat, N.F. de Rooij, D.
deposition, IEEE Sens. J. 13 (no. 10) (2013) 4014–4021. Briand, Fully inkjet printed SnO2 gas sensor on plastic substrate, Sens.
[71] V. Mittal, Carbon nanotubes surface modifications: an overview, in: V. Mittal Actuators B Chem. 236 (2016) 1091–1097.
(Ed.), Surface Modification of Nanotube Fillers, Wiley-VCH Verlag GmbH & [100] M. Tiemann, Porous metal oxide as gas sensors, Chem. Eur. J. 13 (no. 30)
Co. KGaA, 2011. (2007) 8376–8388.
[72] J.-C. Chiou, C.-C. Wu, A wearable and wireless gas-sensing system using [101] E. Comini, Metal oxide nanowire chemical sensors: innovation and quality
flexible polymer/multi-walled carbon nanotube composite films, Polymers of life, Mater. Today 19 (no. 10) (2016) 559–567.
9 (no. 9) (2017) 457. [102] H. Ahn, J.-H. Park, S.-B. Kim, S.H. Jee, Y.S. Yoon, D.-J. Kim, Vertically aligned
[73] Z. Gao, Z. Lou, S. Chen, L. Li, K. Jiang, Z. Fu, W. Han, G. Shen, Fiber gas ZnO nanorod sensor on flexible substrate for ethanol gas monitoring,
sensor-integrated smart face mask for room-temperature distinguishing of Electrochem. Solid-state Lett. 13 (no. 11) (2010) J125–J128, November 1,
target gases, Nano Res. 11 (no. 1) (2017) 511–519. 2010.
[74] C. Hua, Y. Shang, Y. Wang, J. Xu, Y. Zhang, X. Li, A. Cao, A flexible gas sensor [103] M. Tonezzer, R.G. Lacerda, Zinc oxide nanowires on carbon microfiber as
based on single-walled carbon nanotube-Fe2O3 composite film, Appl. Surf. flexible gas sensor, Physica E 44 (no. 6) (2012) 1098–1102.
Sci. 405 (2017) 405–411. [104] M. Acuautla, S. Bernardini, E. Pietri, M. Bendahan, Nozzle-less ultrasonic
[75] H.Y. Jeong, D.-S. Lee, H.K. Choi, D.H. Lee, J.-E. Kim, J.Y. Lee, W.J. Lee, S.O. Kim, spray deposition for flexible ammonia and ozone gas sensors, Sens.
S.-Y. Choi, Flexible room-temperature NO 2 gas sensors based on carbon Transducers J. 201 (no. 6) (2016) 59–64.
nanotubes/reduced graphene hybrid films, Appl. Phys. Lett. 96 (no. 21) [105] T.-R. Rashid, D.-T. Phan, G.-S. Chung, A flexible hydrogen sensor based on Pd
(2010) 213105. nanoparticles decorated ZnO nanorods grown on polyimide tape, Sens.
[76] M. Han, D. Jung, G.S. Lee, Alladium-nanoparticle-coated carbon nanotube Actuators B Chem. 185 (2013) 777–784.
gas sensor, Chem. Phys. Lett. 610-611 (2014) 261–266. [106] L.H.Y. Lin, L. Chen, J. Zhang, L. Shen, Q. Chen, W. Shi, Fully gravure-printed
[77] S. Watcharaphalakorn, L. Ruangchuay, D. Chotpattananont, A. Sirivat, J. NO2gas sensor on a polyimide foil using WO3-PEDOT:PSS nanocomposites
Schwank, Polyaniline/polyimide blends as gas sensors and electrical and Ag electrodes, Sens. Actuators B Chem. 216 (2015) 176–183.
conductivity response to CO–N2 mixtures, Polym. Int. 54 (no. 8) (2005) [107] F. Rigoni, G. Drera, S. Pagliara, A. Goldoni, L. Sangaletti, High sensitivity,
1126–1133. moisture selective, ammonia gas sensors based on single-walled carbon
[78] V. Dixit, S.C.K. Misra, B.S. Sharma, Carbon monoxide sensitivity of vacuum nanotubes functionalized with indium tin oxide nanoparticles, Carbon 80
deposited polyaniline semiconducting thin films, Sens. Actuators B Chem. (2014) 356–363.
104 (no. 1) (2005) 90–93. [108] S. Zhan, D. Li, S. Liang, X. Chen, X. Li, A novel flexible room temperature
[79] K. Crowley, A. Morrin, A. Hernandez, E. O’Malley, P.G. Whitten, G.G. Wallace, ethanol gas sensor based on SnO2 doped poly-diallyldimethylammonium
M.R. Smyth, A.J. Killard, Fabrication of an ammonia gas sensor using chloride, Sensors 13 (no. 4) (2013) 4378–4389.
inkjet-printed polyaniline nanoparticles, Talanta 77 (no. 2) (2008) 710–717. [109] U. Hofmann, R. Holst, Über die Säurenatur und die Methylierung von
[80] J.J. Miasik, A. Hooper, B.C. Tofield, Conducting polymer gas sensors,J̈ournal of Graphitoxyd, Berichte der deutschen chemischen Gesellschaft (A and B
the Chemical Society, Faraday Trans. 1: Phys. Chem. Condensed Phases 82 Series) 72 (no. 4) (1939) 754–771.
(no. 4) (1986) 1117–1126. [110] G. Ruess, Über das Graphitoxyhydroxyd (Graphitoxyd), Monatshefte für
[81] J. Janata, M. Josowicz, Conducting polymers in electronic chemical sensors, Chemie und verwandte Teile anderer Wissenschaften 76 (no. 3-5) (1947)
Nat. Mater. 2 (2003) 19–24. 381–417.
[82] S.J. Park, C.S. Park, H. Yoon, Chemo-electrical gas sensors based on [111] W. Scholz, H.P. Boehm, Untersuchungen am Graphitoxid. VI. Betrachtungen
conducting polymer hybrids, Polymers 9 (no. 5) (2017) 155. zur Struktur des Graphitoxids, Zeitschrift für anorganische und allgemeine
[83] P. Wan, X. Wen, C. Sun, B.K. Chandran, H. Zhang, X. Sun, X. Chen, Flexible Chemie 369 (no. 3-6) (1969) 327–340.
transparent films based on nanocomposite networks of polyaniline and [112] T. Nakajima, A. Mabuchi, R. Hagiwara, A new structure model of graphite
carbon nanotubes for high-performance gas sensing, Small 11 (no. 40) oxide, Carbon 26 (no. 3) (1988) 357–361.
(2015) 5409–5415. [113] A. Lerf, H. He, M. Forster, J. Klinowski, Structure of graphite oxide revisited, J.
[84] Y. Seekaew, S. Lokavee, D. Phokharatkul, A. Wisitsoraat, T. Kerdcharoen, C. Phys. Chem. B 102 (no. 23) (1998) 4477–4482.
Wongchoosuk, Low-cost and flexible printed graphene–PEDOT:PSS gas [114] T. Szabó, O. Berkesi, P. Forgó, K. Josepovits, Y. Sanakis, D. Petridis, I. Dékány,
sensor for ammonia detection, Org. Electron. 15 (no. 11) (2014) 2971–2981. Evolution of surface functional groups in a series of progressively oxidized
[85] P.-G. Su, C.-C. Shiu, Flexible H2 sensor fabricated by layer-by-layer graphite oxides, Chem. Mater. 18 (no. 11) (2006) 2740–2749.
self-assembly of thin films of polypyrrole and modified in situ with Pt [115] D. Dreyer, S. Park, C. Bielawski, R. Ruoff, The chemistry of graphene oxide,
nanoparticles, Sens. Actuators B Chem. 157 (no. 1) (2011) 275–281. Chem. Soc. Rev. 39 (no. 1) (2010) 228–240.
[86] D.K. Bandgar, S.T. Navale, M. Naushad, R.S. Mane, F.J. Stadler, V.B. Patil, [116] Y. Peng, J. Li, Ammonia adsorption on graphene and graphene oxide: a
Ultra-sensitive polyaniline-iron oxide nanocomposite room temperature first-principles study, Front. Environ. Sci. Eng. 7 (no. 3) (2013) 403–411.
flexible ammonia sensor, RSC Adv. 5 (no. 84) (2015) 68964–68971. [117] A. Omidvar, A. Mohajeri, Edge-functionalized graphene nanoflakes as
[87] L. Xue, W. Wang, Y. Guo, G. Liu, P. Wan, Flexible polyaniline/carbon selective gas sensors, Sens. Actuators B Chem. 202 (2014) 622–630.
nanotube nanocomposite film-based electronic gas sensors, Sens. Actuators [118] L. Guo, H.B. Jiang, R.Q. Shao, Y.L. Zhang, S.Y. Xie, J.N. Wang, X.B. Li, F. Jiang,
B Chem. 244 (2017) 47–53. Q.D. Chen, T. Zhang, H.B. Sun, Two-beam-laser interference mediated
[88] Y. Guo, T. Wang, F. Chen, X. Sun, X. Li, Z. Yu, P. Wan, X. Chen, Hierarchical reduction, patterning and nanostructuring of graphene oxide for the
graphene–polyaniline nanocomposite films for high-performance flexible production of a flexible humidity sensing device, Carbon 50 (no. 4) (2012)
electronic gas sensors, Nanoscale 8 (no. 23) (2016) 12073–12080. 1667–1673.
230 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231

[119] W.L. Yim, X.G. Gong, Z.F. Liu, Chemisorption of NO2 on carbon nanotubes, J. [148] E. Menumerov, B.A. Marks, D.A. Dikin, F.X. Lee, R.D. Winslow, S. Guru, D. Sil,
Phys. Chem. B 107 (no. 35) (2003) 9363–9369. E. Borguet, P. Hutapea, R.A. Hughes, S. Neretina, Sensing hydrogen gas from
[120] V. Dua, S.P. Surwade, S. Ammu, S.R. Agnihotra, S. Jain, K.E. Roberts, S. Park, atmospheric pressure to a hundred parts per million with nanogaps
R.S. Ruoff, S.K. Manohar, All-Organic vapor sensor using inkjet-printed fabricated using a single-step bending deformation, ACS Sens. 1 (no. 1)
reduced graphene oxide, Angew. Chemie Int. Ed. 49 (no. 12) (2010) (2016) 73–80.
2154–2157. [149] J.F. Groves, H.N.G. Wadley, A.P. Ritenour, D.D. Hass, P.L. Ratnaparkhi,
[121] P.-G. Su, H.-C. Shieh, Flexible NO2 sensors fabricated by layer-by-layer Electron beam directed vapor deposition, in: Proceedings of the Electron
covalent anchoring and in situ reduction of graphene oxide, Sens. Actuators Beam Melting and Refining Conference, 1997, pp. 46.
B Chem. 190 (2014) 865–872. [150] W. Prusseit, Methods of HTS deposition: thermal evaporation, in: A. Goyal
[122] M.G. Chung, D.-H. Kim, D.K. Seo, T. Kim, H.U. Im, H.M. Lee, J.-B. Yoo, S.-H. (Ed.), Second-Generation HTS Conductors, Springer, Boston, MA, 2005, pp.
Hong, T.J. Kang, Y.H. Kim, Flexible hydrogen sensors using graphene with 81–96.
palladium nanoparticle decoration, Sens. Actuators B Chem. 169 (2012) [151] B.J. de Gans, P.C. Duineveld, U.S. Schubert, Inkjet printing of polymers: state
387–392. of the art and future developments, Adv. Mater. 16 (no. 3) (2004) 203–213.
[123] D. Zhang, J. Tong, B. Xia, Humidity-sensing properties of chemically reduced [152] P. Calvert, Inkjet printing for materials and devices, Chem. Mater. 13 (no. 10)
grapheneoxide/polymer nanocomposite film sensor based on (2001) 3299–3305.
layer-by-layernano self-assembly, Sens. Actuators B Chem. 197 (2014) [153] B. Derby, Inkjet printing of functional and structural materials: fluid
66–72. property requirements, feature stability, and resolution, Annu. Rev. Mater.
[124] B. Wu, X. Zhang, B. Huang, Y. Zhao, C. Cheng, H. Chen, High-Performance Res. 40 (no. 1) (2010) 395–414.
wireless amonia gas sensors based on reduced graphene oxide and [154] G. Cummins, M. Desmulliez, Inkjet printing of conductive materials: a
nano-silver ink hybrid material loaded on a patch antenna, Sensors 17 (no. review, Circuit World 38 (no. 4) (2012) 193–213.
9) (2017) 2070. [155] R.R. Søndergaard, M. Hösel, F.C. Krebs, Roll-to-Roll fabrication of large area
[125] L. Huang, Z. Wang, J. Zhang, J. Pu, S.X.Y. Lin, L. Shen, Q. Chen, W. Shi, Fully functional organic materials, J. Polym. Sci. Part B: Polym. Phys. 51 (no. 1)
printed, rapid-response sensors based on chemically modified graphene for (2013) 16–34.
detecting NO2 at room temperature, ACS Appl. Mater. Interfaces 6 (no. 10) [156] R.P. Gandhiraman, V. Jayan, J.-W. Han, B. Chen, J.E. Koehne, M. Meyyappan,
(2014) 7426–7433. Plasma jet printing of electronic materials on flexible and nonconformal
[126] C.S. Barret, Structure of Metals: Crystallographic Methods, Principle and objects, ACS Appl. Mater. Interfaces 6 (no. 23) (2014) 20860–20867.
Data, first ed., McGraw-Hill Book company. INC, USA, 1943 (Mettalurgy and [157] R.P. Gandhiraman, D. Nordlund, V. Jayan, M. Meyyappan, J.E. Koehne,
metallurgical engineering). Scalable low-cost fabrication of disposable paper sensors for DNA detection,
[127] C.J. Smithells, Metals Reference Book, fifth ed., Butterworth & co, United ACS Appl. Mater. Interfaces 6 (no. 24) (2014) 22751–22760.
Kingdom, 1976, ISBN: 9781483192529. [158] B.P. Ajayi, S. Kumari, D. Jaramillo-Cabanzo, J. Spurgeon, J. Jasinski, M.
[128] Properties Table of Stainless Steel, Metals and Other Conductive Materials, Sunkara, A rapid and scalable method for making mixed metal oxide alloys
2011, Available: [Link] for enabling accelerated materials discovery, J. Mater. Res. 31 (no. 11) (2016)
[129] D. Kohl, The role of noble metals in the chemistry of solid-state gas sensors, 1596–1607.
Sens. Actuators B Chem. 1 (no. 1-6) (1990) 158–165. [159] S. Khan, L. Lorenzelli, R.S. Dahiya, Technologies for printing sensors and
[130] M. Wang, Y. Feng, Palladium–silver thin film for hydrogen sensing, Sens. electronics over large flexible substrates: a review, IEEE Sens. J. 15 (no. 6)
Actuators B Chem. 123 (no. 1) (2007) 101–106. (2015) 3164–3185.
[131] W. Kim, B. Jang, H.-S. Lee, W. Lee, Reliability and selectivity of H2 sensors [160] G. Dubourg, A. Segkos, J. Katona, M. Radović, S. Savić, G. Niarchos, C. Tsamis,
composed of Pd film nanogaps on an elastomeric substrate, Sens. Actuators V. Crnojević-Bengin, Fabrication and characterization of flexible and
B: Chem. 224 (2016) 547–551. miniaturized humidity sensors using screen-printed TiO2 nanoparticles as
[132] P.-G. Su, Y.-S. Chuang, Flexible H2 sensors fabricated by layer-by-layer sensitive layer, Sensors 17 (no. 8) (2017) 1854.
self-assembly thin film of multi-walled carbon nanotubes and modified [161] P. Kuberský, T. Syrový, A. Hamáček, R. Nešpůrek, J. Stejskal, Printed flexible
in situ with Pd nanoparticles, Sens. Actuators B Chem. 145 (no. 1) (2010) gas sensors based on organic materials, Procedia Eng. 120 (2015) 614–617.
521–526. [162] C.E. Cava, R.V. Salvatierra, D.C.B. Alves, A.S. Ferlauto, A.J.G. Zarbin, L.S.
[133] J. Lee, W. Shim, E. Lee, J.-S. Noh, W. Lee, Highly mobile palladium thin films Roman, Self-assembled films of multi-wall carbon nanotubes used in gas
on an elastomeric substrate : nanogap-based hydrogen gas sensors, Angew. sensors to increase the sensitivity limit for oxygen detection, Carbon 50 (no.
Chemie Int. Ed. 50 (no. 23) (2011) 5301–5305. 5) (2012) 1953–1958.
[134] M.A. Lim, D.H. Kim, C.-O. Park, Y.W. Lee, S.W. Han, Z. Li, R.S. Williams, I. Park, [163] P.-G. Su, C.-T. Lee, C.-Y. Chou, Flexible NH3 sensors fabricated by in situ
A new route toward ultrasensitive, flexible chemical sensors: metal self-assembly of polypyrrole, Talanta 80 (no. 2) (2009) 763–769.
nanotubes by wet-chemical synthesis along sacrificial nanowire templates, [164] P.-G. Su, Y.-S. Chuang, Flexible H2 sensors fabricated by layer-by-layer
ACS Nano 6 (no. 1) (2012) 598–608. self-assembly thin film of multi-walled carbon nanotubes and modified
[135] S.H. Lim, B. Radha, J.Y. Chan, M.S.M. Saifullah, G.U. Kulkarni, G.W. Ho, in situ with Pd nanoparticles, Sens. Actuators B Chem. 145 (no. 1) (2010)
Flexible palladium-based H2 sensor with fast response and low leakage 521–526.
detection by nanoimprint lithography, ACS Appl. Mater. Interfaces 5 (no. 15) [165] D. Zhang, J. Tong, B. Xia, Q. Xue, Ultrahigh performance humidity sensor
(2013) 7274–7281. based on layer-by-layer self-assembly of graphene oxide/polyelectrolyte
[136] K. Lee, V. Scardaci, H.-Y. Kim, T. Hallam, H. Nolan, B.E. Bolf, G.S. Maltbie, J.E. nanocomposite film, Sens. Actuators B Chem. 203 (2014) 263–270.
Abbott, G.S. Duesberg, Highly sensitive, transparent, and flexible gas sensors [166] L. Kumar, I. Rawal, A. Kaur, S. Annapoorni, Flexible room temperature
based on gold nanoparticle decorated carbon nanotubes, Sens. Actuators B ammonia sensor based on polyaniline, Sens. Actuators B Chem. 240 (2017)
Chem. 188 (2013) 571–575. 408–416.
[137] Y. Sun, H.H. Wang, High-performance, flexible hydrogen sensors that use [167] S. Patel, H. Park, P. Bonato, L. Chan, M. Rodgers, A review of wearable sensors
carbon nanotubes decorated with palladium nanoparticles, Adv. Mater. 19 and systems with application in rehabilitation, J. Neuroeng. Rehabil. 9 (no.
(no. 19) (2007) 2818–2823. 1) (2012) 21.
[138] D.H. Shin, J.S. Lee, J. Jun, J.H. An, S.G. Kim, K.H. Cho, J. Jang, Flower-like [168] S.L.P. Tang, Recent developments in flexible wearable electronics for
palladium nanoclusters decorated graphene electrodes for ultrasensitive monitoring applications, Trans. Inst. Meas. Control. 29 (no. 3-4) (2007)
and flexible hydrogen gas sensing, Sci. Rep. 5 (2015) 12294. 283–300.
[139] B. Cho, J. Yoon, M.G. Hahm, D.-H. Kim, A.R. Kim, Y.H. Kahng, S.-W. Park, Y.-J. [169] M. Stoppa, A. Chiolerio, Wearable electronics and smart textiles: a critical
Lee, S.-G. Park, J.-D. Kwon, C.S. Kim, M. Song, Y. Jeong, K.-S. Nam, H.C. Ko, review, Sensors 14 (no. 7) (2014) 11957–11992.
Graphene-based gas sensor: metal decoration effect and application to a [170] Z.Q. Zheng, J.D. Yao, B. Wang, G.W. Yang, Light-controlling, flexible and
flexible device, J. Mater. Chem. C 2 (no. 27) (2014) 5280–5285. transparent ethanol gas sensor based on ZnO nanoparticles for wearable
[140] M.D. Tyona, A theoritical study on spin coating technique, Adv. Mater. Res. 2 devices, Sci. Rep. 5 (2015) 11070.
(no. 4) (2013) 195–208. [171] J. Park, J. Kim, K. Kim, S.-Y. Kim, W.H. Cheong, K. Park, J.H. Song, G.
[141] Spin Coating: A Guide to Theory and Techniques, 2018, Available: https:// Namgoong, J.J. Kim, J. Heo, F. Bien, J.-U. Park, Wearable, wireless gas sensors
[Link]/pages/spin-coating. using highly stretchable and transparent structures of nanowires and
[142] C.J. Brinker, G.C. Frye, A.J. Hurd, C.S. Ashley, Fundamentals of sol-gel graphene, Nanoscale 8 (no. 20) (2016) 10591–10597.
dip-coating, Thin Solid Films 201 (no. 1) (1991) 97–108. [172] X.-S. Jia, C.-C. Tang, X. Yan, G.-F. Yu, J.-T. Li, H.-D. Zhang, J.-J. Li, C.-Z. Gu, Y.-Z.
[143] J.-H. Park, T.S. Sudarshan, Chemical Vapor Deposition (Surface Engineering Long, Flexible polyaniline/poly(methyl methacrylate) composite fibers via
Series Volume 2), The materials Information Society, USA, 2001. electrospinning and in situ polymerization for ammonia gas sensing and
[144] D.M. Mattox, Handbook of Physical Vapor Deposition (PVD) Processing, strain sensing, Journal of Nanomaterials 2016 (2016) 9102828.
second edition, William Andrew Publishing, Boston, 2010.
[145] W.D. Westwood, Physical vapor deposition, in: R.A. Levy (Ed.),
Microelectronic Materials and Processes, vol. 164, Springer, Dordrecht, Biographies
1989, pp. 133–201 (NATO ASI Series (Series E: Applied Sciences).
[146] P.J. Kelly, R.D. Arnell, Magnetron sputtering: a review of recent
developments and applications, Vacuum 56 (no. 3) (2000) 159–172. Rouba Alrammouz is a PhD candidate in a joint program between the University of
[147] S. Oztürk, N. Kılınç, Pd thin films on flexible substrate for hydrogen sensor, J. Montpellier France and the Lebanese University in Lebanon. She is currently working
Alloys. Compd. 674 (2016) 179–184. on flexible gas sensors, with a focus on detecting ammonia, for eventually apply the
results to wearable electronics.
 R. Alrammouz et al. / Sensors and Actuators A 284 (2018) 209–231 231

Jean Podlecki is a researcher in the “Institut d’Electronique et de Systèmes” at the Brice Sorli is a researcher in the “Institut d’Electronique et de Systèmes” at the
University of Montpellier France. He is currently working on optoelectronic devices University of Montpellier France. He is working on gas sensors, thin films technology,
through CVD deposition. He is also working on micro-systems and micro-sensors, nanomaterials, RFID and materials characterization.
energy autonomy, sensors network, and flexible electronics.
Roland Habchi is a professor of Physics at the Lebanese University and the director
Pascale Abboud is a researcher at EC2M laboratory in the Lebanese University of EC2M laboratory. He is currently working on nanomaterials and smart structures,
and an associate professor at Saint Joseph University. She worked on microcrys- magnetic materials, energy storage, and wearable electronics.
talline silicon thin layers for solar cells and electronic transport modeling. She is
currently working on wearable electronic and materials characterization for energy
harvesting applications.