Dip pen nanolithography-deposited zinc oxide nanorods on a CMOS MEMS platform for ethanol sensing

Article abstract of DOI:10.1039/c5ra04584c

This paper reports on the novel deposition of zinc oxide (ZnO) nanorods using a dip pen nanolithographic (DPN) technique on SOI (silicon on insulator) CMOS MEMS (micro electro mechanical system) micro-hotplates (MHP) and their characterisation as a low-cost, low-power ethanol sensor. The ZnO nanorods were synthesized hydrothermally and deposited on the MHP that comprise a tungsten micro-heater embedded in a dielectric membrane with gold interdigitated electrodes (IDEs) on top of an oxide passivation layer. The micro-heater and IDEs were used to heat up the sensing layer and measure its resistance, respectively. The sensor device is extremely power efficient because of the thin SOI membrane. The electro-thermal efficiency of the MHP was found to be 8.2°C mW^-1, which results in only 42.7 mW power at an operating temperature of 350°C. The CMOS MHP devices with ZnO nanorods were exposed to PPM levels of ethanol in humid air. The sensitivity achieved from the sensor was found to be 5.8% ppm^-1 to 0.39% ppm^-1 for the ethanol concentration range 25-1000 ppm. The ZnO nanorods showed an optimum response at 350°C. The CMOS sensor was found to have a humidity dependence that needs consideration in real-world application. The sensors were also found to be selective towards ethanol when tested in the presence of toluene and acetone. We believe that the integration of ZnO nanorods using DPN lithography with a CMOS MEMS substrate offers a low cost, low power, smart ethanol sensor that could be exploited in consumer electronics.

Full text of DOI:10.1039/c5ra04584c

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Dip pen nanolithography-deposited zinc oxide
nanorods on a CMOS MEMS platform for ethanol
sensing
Cite this: RSC Adv., 2015, 5, 47609
S. Santra, a A. De Luca,b S. Bhaumik,a S. Z. Ali,c F. Udrea,bc J. W. Gardner,cd S. K. Raya
*
and P. K. Guhae
This paper reports on the novel deposition of zinc oxide (ZnO) nanorods using a dip pen nanolithographic
(DPN) technique on SOI (silicon on insulator) CMOS MEMS (micro electro mechanical system) micro-
hotplates (MHP) and their characterisation as a low-cost, low-power ethanol sensor. The ZnO nanorods
were synthesized hydrothermally and deposited on the MHP that comprise a tungsten micro-heater
embedded in a dielectric membrane with gold interdigitated electrodes (IDEs) on top of an oxide
passivation layer. The micro-heater and IDEs were used to heat up the sensing layer and measure its
resistance, respectively. The sensor device is extremely power efficient because of the thin SOI
membrane. The electro-thermal efficiency of the MHP was found to be 8.2 ꢀC mWꢁ1, which results in
only 42.7 mW power at an operating temperature of 350 ꢀC. The CMOS MHP devices with ZnO
nanorods were exposed to PPM levels of ethanol in humid air. The sensitivity achieved from the sensor
was found to be 5.8% ppmꢁ1 to 0.39% ppmꢁ1 for the ethanol concentration range 25–1000 ppm. The
ZnO nanorods showed an optimum response at 350 ꢀC. The CMOS sensor was found to have a
humidity dependence that needs consideration in real-world application. The sensors were also found to
be selective towards ethanol when tested in the presence of toluene and acetone. We believe that the
integration of ZnO nanorods using DPN lithography with a CMOS MEMS substrate offers a low cost, low
power, smart ethanol sensor that could be exploited in consumer electronics.
Received 16th March 2015
Accepted 18th May 2015
DOI: 10.1039/c5ra04584c
www.rsc.org/advances
better substrates (e.g. sensors can be developed on ceramic,
silicon, exible polymer or a CMOS-MEMS platform).
1. Introduction
Carbon nano material (carbon nanotube, graphene) can
sense gases close to room temperature, but their response is
generally disappointing.3–6 Metal oxides on the other hand are
another class of material which shows large response but
usually at elevated temperatures (e.g. 350 ꢀC), so needs high
power.
Gas sensors have enjoyed a wide range of applications since the
1970s that is steadily growing. Nowadays, they are extensively
used to detect hazardous (i.e. toxic and explosive) gases present
in industrial areas, indoors, and coal mines. More recently they
are being explored to monitor trace levels of volatile organic
compounds (VOCs) in human breath, which contains
biomarkers for specic diseases.1,2 However, commercially
available gas sensors are still bulky, expensive and, with the
exception of electrochemical ones, consume a large amount of
power (ꢂ0.5 W). Hence, recent research on gas sensors has
shied towards the development of miniaturised, low power,
inexpensive and easy to integrate devices. Such sensors can be
achieved by advances on two fronts, (A) better sensing layers
(e.g. polymer, carbon nano material or metal oxide) and (B)
Ceramic and silicon substrate have been extensively used for
sensor development. Flexible substrates3,4,6 have recently been
introduced to develop large area exible sensors. However all
these platforms require separate interface electronics that
incurs additional costs making them commercially limited. In
this respect, CMOS-MEMS7–9 technology is very promising for a
new generation of smart low-cost gas sensors; because the
fabrication process is well established, devices are miniaturised
(<4 mm2) and their performances are reliable and reproducible
(because of batch fabrication). Such sensors can also have on-
chip interface electronics, which results in a further system
miniaturization and cost reduction. MEMS technology inte-
gration with CMOS is critical for achieving low power
consumption of the sensor device through micro-heater
thermal isolation schemes.9,10 Resistive sensors have an
advantage when compared to other classes of gas sensors (e.g.
aDepartment of Physics, Indian Institute of Technology, Kharagpur, India, 721302.
E-mail: susmita.santra@phy.iitkgp.ernet.in
bEngineering Department, University of Cambridge, Cambridge, CB3 0FA, UK
cCambridge CMOS Sensors Ltd., Cambridge, CB4 0DL, UK
dSchool of Engineering, University of Warwick, Coventry, CV4 7AL, UK
eE & ECE Department, Indian Institute of Technology, Kharagpur, India, 721302
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 47609–47616 | 47609

View Article Online
RSC Advances
Paper
electrochemical and optical), because resistive sensors are solution. This solution was drop casted on a beaker preheated
easier to integrate with a CMOS-MEMS platform.
to 200 ꢀC. White ZnO nanoparticles were formed from the rapid
Nano-materials have a very high surface to volume ratio, thus evaporation of the solution. This seed layer was well dispersed
providing a signicant gas response even with a small amount in 400 ml of water by continuous ultrasonication for two hours.
of material. Therefore nano-materials are well suited for mini- In this solution 20 mM zinc nitrate and hexametheline tetra-
aturised CMOS gas sensor devices. Commercially available solid mine were added and vigorously stirred for 10 minutes. This
state gas sensors are generally metal oxide based (e.g. tin oxide solution was kept on a hotplate at 90 ꢀC for ve hours to allow
or tungsten oxide). Gas sensors using different metal oxide ZnO nanorods extrusion. The nal solution was cleaned several
nano-materials are already reported in the literature.1,11–15 times with deionised water and air dried, in order to remove any
Among them, zinc oxide (ZnO) is one of the most reported chemical residue present in NRs.
sensing materials.11,14–22 Because its preparation method is
simple, inexpensive, it has good thermal and chemical stability,
high electron mobility and it responds to several gases or
volatile organic compounds. Different nano-structure of ZnO
including nanoparticles,17,21 nanorod,17–20,23 nanowire,14,22
nanotube,16 thin lm,24 thick lm25 are reported as gas sensing
layers.
2.2 CMOS MEMS device fabrication
The gas sensors devices are based on a SOI (silicon on insulator)
CMOS MEMS micro-hotplate (MHP) technology. The cross-
sectional view and the optical micrograph of the device are
shown in Fig. 1(a) and (b). The silicon die is only 1 mm ꢃ 1 mm
in size. The devices were designed in Tanner Tools LEdit 15.2,
using a 1.0 mm SOI CMOS process and fabricated at a
commercial foundry. Micro-hotplates comprised a micro-heater
and interdigitated electrodes (IDEs), suspended in a dielectric
membrane. Tungsten was used as metal for the micro-heater, in
place of Al or polysilicon, because of its superior electro-thermal
properties that resulted in extended device lifetime at high
operating temperatures. The membrane structure was created
using a deep reactive ion etching (DRIE) technique. The buried
oxide was used as the etch stop with the total membrane
thickness being ca. 5 mm. The micro-heater and membrane
diameters were 250 and 650 mm, respectively, with a ratio
chosen to optimise the chip size with respect to power
consumption. The micro-heater design was also optimised in
order to achieve highly temperature uniformity of the sensing
material. The details of this study are reported in ref. 10 and 29.
A silicon nitride layer passivated the whole chip area with
exception of the pad areas to contact heater and IDEs. The gold
electrodes and the gold interconnects to the pads were made by
a post-CMOS process within the same foundry. Gold electrodes
were found to have a good ohmic contact with the sensing
materials, a long time stability and a smaller feature size in
comparison to Al IDEs. The width of each electrode nger is 5
mm. The separation between the electrodes is 5 mm (aspect ratio
800). As a consequence of the high thermal isolation offered by
the thin dielectric membrane and the heat sink effect provided
by the sensor package,10,30 the temperature of the substrate area
of the chip is close to room temperature. CMOS circuitry, for
heater temperature control or sensing layer resistance readout,
can thus be integrated on-chip.
Nano-materials are prepared in several ways like chemical
and physical methods (chemical vapour deposition methods,
physical vapour deposition methods, sputtering and evapora-
tion), mechanical exfoliation, chemical exfoliation etc. Deposi-
tion of nano-materials on CMOS-MEMS devices is a challenging
task. Chemical routes are not suitable for CMOS sensors as
some chemicals can have a detrimental effect on the integrity of
the chip. Physical methods need either shadow or lithographic
masking. Several methods have been reported to integrate
nano-materials with CMOS devices, e.g. local growth tech-
nique9,26 to grow carbon nanotubes (CNTs), hydrothermal
method to grow zinc oxide nanowires.14 These reported mate-
rials grown directly on the sensor device, however seed layers
were required that need to be sputter coated. Inkjet printing27,28
is a technique suitable for the deposition of nano-materials.
However, making a stable ink which can be deposited to
required precision and stability is under development.
Dip pen nano-lithography (DPN) is an extremely exible
deposition method, with possible wafer level scalability, which
has proved to be suitable for nano-materials integration on
membrane based CMOS devices. Additionally, DPN does not
require time-consuming development of the material to be
deposited (as for ink-jet printing); a highly viscous slurry is
acceptable for most applications.
Here in this work zinc oxide nanorods (NRs) were prepared
using a hydrothermal method. The ZnO nanorods were depos-
ited on a predened electrode area of CMOS sensor devices
using DPN. Finally the devices were tested in presence of
ethanol in humid air. The humidity dependence on ethanol
sensing was experimentally investigated. A cross-sensitivity
study to other VOCs was also performed.
The power versus temperature plot of the device is presented
in Fig. 2. The heater can reliably reach temperatures of 600 C
ꢀ
with 73 mW of power in 15 ms, with an electro-thermal effi-
ciency of 8.2 ꢀC mWꢁ1, and cool down to ambient temperature
in about 30 ms.
2. Experimental
2.1 Zinc oxide nanorods preparation
Initially a ZnO nano-particle seed layer was prepared for zinc
oxide nanorods growth. The seed layer was needed to support
2.3 Dip pen nanolithography deposition
nucleation and subsequent growth. The seed layer was prepared ZnO nanorods were integrated onto the CMOS microhotplates
by dissolving 0.5 g zinc-acetate in 100 ml of 2-propanol. Then with a dip pen nanolithographic system (NLP2000 by NanoInk).
the mixture was ultrasonicated for one hour to make a uniform Dip pen nanolithography31 is
a unique scanning-probe
47610 | RSC Adv., 2015, 5, 47609–47616
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Fig. 1 (a) Cross sectional view of the CMOS MEMS micro-hotplate technology (b) optical micrograph of fabricated device (1 mm by 1 mm).
slurry was ꢂ10 pL per deposition. The chemical optimization of
the slurry for obtaining a highly uniform layer is under study.
2.4 Characterisation
Optical microscope pictures were taken with Nikon L-UEPI tted
with Nikon DX1200F digital camera. The morphological and
structural studies of ZnO NRs were performed using eld
emission scanning electron microscope (FESEM, Zeiss Gemini-
Sigma), transmission electron microscopy (TEM, FEI – Tecnai
G2 20S – Twin operated at 200 kV), X-ray diffraction (Philips X-
Pert MRD with Cu Ka radiation). Electrical measurements were
performed using a commercial precision instrument (Keithley
2401).
Fig. 2 Power versus temperature plot of tungsten micro-heater.
2.5 Automated gas testing station
lithographic technique enabling the deposition of patterns of
The gas sensing measurements were performed at the Micro-
sensors and Bioelectronics Laboratory, Warwick University
using fully-automated custom rig. The CMOS die were mounted
and ultrasonically wire bonded onto gold TO5 packages, in
order to be connected to a custom made printed circuit board
and housed inside the sensor chamber. Micro-heater and
sensing material's resistance were respectively driven and
measured via a LabView soware. Control of the different valves
and digital mass ow controllers of the gas system was also
performed automatically through a NI Labview VI.
various materials with sub-micron resolution. To date, DPN has
been investigated as a possible nanofabrication technology in
four main areas: (i) direct-writing of bio-molecular micro arrays,
(ii) creating tailored chemical surfaces for the investigation of
biological recognition processes down to cellular level, (iii)
generating chemical templates for orthogonal materials'
assembly, (iv) low-cost and rapid sub-micron structures proto-
typing with both top-down and bottom-up approaches. All the
previously mentioned application areas exploit the excellent
DPN in-plane resolution for the delivery of materials in pico-
litre quantities. In this work, we used DPN to functionalize
fragile CMOS membrane based microhotplates, exploiting the
DPN system off-plane resolution and the high mechanical
compliance offered by the cantilever-type pens, in order to
3. Results and discussion
3.1 Morphology and structure
“gently” deposit ZnO nanorods (mixed with terpineol) slurry on An optical microscope image of the DPN deposited ZnO nano-
a relatively wide area (>250 mm diameter) at once. The slurry rods on MHP is shown in Fig. 3(a). A typical SEM image of the
contains 95% of ZnO and 5% of terpineol by weight. In order to MHP with ZnO nanorods is shown in Fig. 3(b). From these
achieve such a wide deposition area a double cantilever pen was micrographs, it can be clearly seen that the nanorods are
used. Each cantilever is approximately 50 mm wide and 150 mm deposited mainly over the micro-heater region, without creating
long and they are 50 mm apart from each other. By fully dipping a thermal bridge from the hot-area to the substrate – thus
the cantilevers and part of the pen in the material reservoir (not avoiding undesired extra power dissipation. In our previous
only the cantilever tip underneath the cantilevers' end as reports9,14 vertically aligned ZnO nanowires12 and carbon
usually done in a DPN process for deposition of tiny amount of nanotubes,7 were grown directly on the MHP. Here the nano-
material) and by appropriately controlling the deposition time rods (single or bundles) are horizontally bridging electrodes,
(the contact time between cantilever and substrate is in the resulting in a better lateral heat transfer, making it more
seconds range) it is possible to load and then deposit enough important to avoid growing the nanorods in the membrane area
slurry to coat the whole IDEs area. The volume of the deposited between the heater and substrate. A magnied view of the
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 47609–47616 | 47611

View Article Online
RSC Advances
Paper
Fig. 3 (a) Optical microscope picture of micro-hotplate with ZnO nanorods, (b) typical SEM image of ZnO NRs on MHP, a magnified view of
nanorods is shown in the inset (c) TEM image of ZnO nanorods, (d) XRD spectrum of ZnO nanorods.
nanorods is presented in the inset of Fig. 3(b). The hexagonal affinity then the contact is Ohmic. In our case the work function
structure of ZnO nanorods can be clearly noticed. Nanorods are of Au (5.1 eV) is larger than the electron affinity (4.2 eV) of zinc
100–300 nm in diameter and more than 2 mm long. Trans- oxide. Consequently, at ambient temperatures the contact is
mission electron microscopy (TEM) image of ZnO as seen in non-linear. But as we increase the temperature then the I–V slope
ꢀ
Fig. 3(c) clearly signies the rod like nano-structure. The X-ray increases up to 350 C and becomes more linear (i.e. Ohmic).
diffraction (XRD) measurement, shown in Fig. 3(d), has been Thus the contact resistance is lower and can be ignored at higher
performed to have a detailed insight of the structure of ZnO temperature while performing the gas sensing measurements.
nanorod, which is shown to be polycrystalline and correspond
to wurtzite structure.
3.3 Gas response
In this work, the sensor response (S) is dened by the ratio Ra/Rg
where Ra and Rg are the sensor electrical resistance in air and
3.2 Electrical characterisation
The current voltage characteristics (I–V) were measured at ethanol, respectively. Initially the ZnO sensor was tested in the
different operating temperatures and are shown in Fig. 4(a). A presence of 1000 ppm ethanol in order to obtain the optimum
contact between a metal and semiconductor can be either Ohmic working temperature of the sensor, as shown in Fig. 4(b). The
or Schottky in nature, depending upon the difference in the work sensor response increases with temperature and maximum
function of the metal and electron affinity of semiconductor. response is found at 350 ꢀC. The response decreases above this
When the work function is equal or lower than the electron temperature. Initially the reaction of gases with chemisorbed
Fig. 4 (a) I – V characteristic ZnO NRs on gold IDEs at different temperatures, (b) Sensor response as a function of temperature for 1000 ppm of
ethanol in 10% RH air.
47612 | RSC Adv., 2015, 5, 47609–47616
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Fig. 5 (a) Dynamic response of ZnO NRs in presence of ethanol at 350 ꢀC, (b) response and recovery time against ethanol concentration plot at
350 ꢀC.
oxygen sites increases with increase in temperature. Above result the response and recovery time will be better in this case.
350 ꢀC, the desorption rate is higher than the adsorption, so the The obtained results (response, recovery and response time
response decreases with the further increase in temperature. towards ethanol) are better than or comparable to the reported
Therefore, all the measurements have been performed at a results available in literature.14,23,32–34 Some reports are also
temperature of 350 ꢀC. The DC power consumption required to available which are better than our results.16,18,20,21 It may be due
operate at this temperature is 42.7 mW.
to the different morphology and size of the nanostructures. The
The time-dependent response to ethanol pulses at 10% measurements were performed in dry air for some reports which
relative humidity (RH) is shown in Fig. 5(a). The peak gas provides better response than humid air measurements.
concentration increases from 25 to 1000 ppm. The response is However, most of the reported results are on non-CMOS platform
found to increase with increasing concentration as expected. which are not ideal for batch fabrication and commercialisation.
The responses (Ra/Rg) at 25 and 1000 ppm are calculated to be
The response of semiconductor metal oxide gas sensor is
1.45 and 3.9, respectively. Sensitivity in percentage (dened as empirically represented by the following power law:
response/concentration) varies from 5.8–0.39% ppmꢁ1 for the
concentration range from 25–1000 ppm.
b
b
S ¼ 1 + AgPg ¼ 1 + AgCg
(1)
The response time is dened as the time taken by the sensor
to reach 90% of the full response, whereas recovery time is the
time required to reach the 10% of the sensor baseline resistance.
Response and recovery times at different concentrations of
ethanol are presented in Fig. 5(b). It is observed that with increase
in gas concentrations, the response time decreases whereas the
recovery time increases. Response time ranged from 150–8 s and
recovery time 100–560 s when the concentration is varied from
25–1000 ppm. The performance of DPN deposited ZnO nanorods
is much better than our earlier report,14 e.g. the sensitivity is
much higher (145% higher – in present work the sensitivity at
750 ppm is 0.49% ppmꢁ1 as compared to 0.2% ppmꢁ1 at
809 ppm of ethanol in earlier work). The response and recovery
time is also much faster (in present work tresponse is only 9 s and
trecovery is 480 s at 750 ppm whereas tresponse is 200 s and trecovery is
more than 10 min at 809 ppm in earlier work). In the present
work we have detected ethanol concentration as low as 25 ppm
(the lowest concentration we could measure in our previous work
was 809 ppm). These improved results can be attributed to the
fact that ethanol molecules can react more quickly with the
chemisorbed oxygen species at the surface of ZnO nanorod. This
is because nanorods are horizontally aligned on the electrodes
instead of vertically standing, thus offering a wider surface for
chemical interaction. Also, as the nanorods are horizontally
aligned on the IDEs, so we believe heat transfer will be better than
the rooted growth vertical nanorods. So the temperature of the
where Ag is the prefactor which depends on sensing materials,
operating temperature and the type of gas interacting with the
sensor. Pg is the partial pressure of gas that is directly propor-
tional to concentration (Cg) of gas and b is the exponent factor.
The ideal value of b is either 0.5 or 1 depending on the charged
state of surface oxygen species and the stoichiometry of the
elementary reactions on the surface. The value of b was calcu-
lated from the interaction between chemisorbed surface oxygen
species and the reducing gases like ethanol. It was reported35
that b is 1 or 0.5 when the surface oxygen species are Oꢁ and
O2ꢁ state, respectively. In Fig. 6 the experimentally measured
Fig. 6 Response of ZnO nanorods sensor as a function of concen-
tration plot. Solid line represents the power law fitted through the
lm will be much more uniform than the rooted growth. As a experimental points.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 47609–47616 | 47613

View Article Online
RSC Advances
Paper
a basic oxide, ethanol will be converted to acetaldehyde by the
following equation.
CH3CH2OH / CH3COH + H2
(6)
Acetaldehyde will be further oxidised by interacting with the
chemisorbed oxygen species. As our working temperature is
350 ꢀC, O2ꢁ species will mainly interact with ethanol by the
following equation
CH3CHO + 5O2ꢁ(ads) / 2CO2(ads) + 2H2O(vap) + 10eꢁ (7)
Fig.
7 Selectivity measurements of ZnO NRs for 1000 ppm
concentration.
It can be seen from the above equation that the electrons will
be freed due to the reaction with ethanol and returned back to
the conduction band of nanorods. Hence, the resistance of the
ZnO nanorods will decrease in presence of ethanol as presented
in Fig. 8(b).
response versus concentration is plotted. A power law has been
tted through the experimental points according to eqn (1). The
value of b is 0.438 ꢄ 0.022, which is slightly lower than the ideal
value (0.5). This could be due to the agglomeration of nano-
materials or some area is less sensitive to ethanol.36 The expo-
nent b is nearer to 0.5 suggesting that the chemisorbed oxygen
species are in O2ꢁ state.
Zinc oxide sensors are known to be sensitive towards various
organic vapours. Hence some selectivity measurements have
also been performed and presented in Fig. 7. These sensors
were tested in presence of 1000 ppm of ethanol, acetone and
toluene. It is found that responses of acetone and toluene are
1.29 and 2.2, respectively. The response is 3 and 1.77 times
lower than ethanol. Hence the ZnO sensors are selective
towards ethanol. This observation is also in agreement with
other reported result on ethanol sensor.21
3.5 Humidity effect on ethanol sensing
The ZnO sensor was also tested in presence of ethanol at dry air,
10% RH, 40% RH to see the humidity effect on sensing as
shown in Fig. 9. In this case concentrations were varied from
100–1000 ppm. At dry air, response varied from 2.9–6.1 as the
concentrations were changing from 100–1000 ppm. The corre-
sponding values for 10% and 40% RH are 2–3.7 and 1.7–2.8,
respectively. At 1000 ppm the response decreased by 1.67 and
2.17 times for 10% and 40% RH, respectively. Hence humidity
plays a crucial role in sensing even at this high temperature.
When 10% RH was introduced inside the chamber, the
conductivity of the ZnO nanorods increased very sharply as
shown in inset of Fig. 9 (red circular area). This fast change in
resistance is due to the non-dissociative adsorption of water
molecules as shown by the following equation39
3.4 Sensing mechanism
The scheme for sensing mechanism of ZnO NR is presented in
Fig. 8. When the semiconductor metal oxides are exposed to air,
the oxygen molecules are adsorbed on the surface of the
materials. At room temperature the oxygen molecules are
H2O(gas) / H2O(ads) / H2O+(ads) + eꢁ
(8)
According to the above equation water molecules donate the
electron to the conduction band of ZnO, which is responsible
for the fast change in resistance with the introduction of
humidity. Illyaskutty et al.39 also reported the similar conduc-
(2) tivity change results in presence of humidity of molybdenum
oxide. It was also observed that aer the sharp decrease, the
physisorbed. As the temperature is increased, oxygen molecules
ꢁ
are dissociated to atomic oxygen. Different oxygen species (O2
,
Oꢁ, O2ꢁ) are created depending on the working temperature.23
O2(g) / O2(ads)
O2(ads) + eꢁ / O2ꢁ(ads) T < 100 C
(3)
ꢀ
resistance increases slowly. This can be due to another second
slow reaction started along with the process presented in eqn
(8). In the slow reaction, adsorbed water molecules broke into
hydrogen and hydroxyl group39 due to the high working
O2ꢁ(ads) + eꢁ / 2Oꢁ(ads) 100 C # T # 300 C
(4)
ꢀ
ꢀ
Oꢁ(ads) + eꢁ / O2ꢁ(ads) T > 300 C
(5)
ꢀ
temperature.
H2O(ads) / OH(ads) + H(ads)
(9)
These oxygen species will take electrons from the conduction
band of the ZnO and chemisorbed on the surface. This creates
an electron depletion layer on the surface of the material which
increases the resistance of the sensing layer as shown in
Fig. 8(a).
This hydroxyl group interacts with Zn and hydrogen reacts
with lattice oxygen by the following equation.14,39
H2O(ads) + Znlat + Olat / (HO ꢁ Znlat
)
+ (OlatH)+ + Vo + eꢁ
2+
(10)
Ethanol interaction with metal oxide is complicated.
Depending on the acid, base property of metal oxide ethanol
2+
will be decomposed to an intermediate state either to ethylene where Vo is the oxygen vacancy. Aer initial increase of
(dehydration) or acetaldehyde (dehydrogenation).37,38 As ZnO is conductivity, the reactions presented in (8) and (10) progress
47614 | RSC Adv., 2015, 5, 47609–47616
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Fig. 8 (a) ZnO NR in presence of air, (b) ethanol and their corresponding band diagrams.
nanorods were prepared hydrothermally. CMOS MEMS gas
sensors are found to be power efficient and on-chip electronic
circuit integration is possible. The ethanol sensing performance
ꢀ
was investigated in the temperature range 250–450 C and the
optimum operating temperature was found to be at 350 ꢀC.
Ethanol sensitivity was found to be 5.8–0.39% ppmꢁ1 for the
concentration range 25–1000 ppm. The response and recovery
times are found to be 9 s and 480 s at 750 ppm of ethanol,
respectively. The CMOS sensors were tested in presence of
ethanol, toluene and acetone in humid air. The sensor is found
to be selective towards ethanol. The response was found to
decrease in humid ambience as compared to dry air environ-
ment. The study indicates that DPN deposited ZnO nanorods on
CMOS substrate may open up a scalable, batch produced
approach to develop power efficient, low cost smart ethanol
sensor.
Fig. 9 Ethanol sensing measurements in presence of dry, 10% RH and
40% RH air.
simultaneously, thus slowly decreasing the conductivity of the
nanorods. In this condition when ethanol is introduced the
response was found to decrease, in comparison to dry air
condition. The reasons for such behaviour are two fold.40,41
Firstly, some active sites are blocked by hydroxyl groups.
Secondly, there is a competition between water molecules and
ethanol to react with adsorbed species which eventually decrease
response towards ethanol. Similar observations have also been
demonstrated for other metal oxides towards ethanol sensing.39,41
Acknowledgements
S. Santra acknowledges Department of Science and Technology
(DST) for Ramanujan Fellowship (project no. SR/S2/RJN-104/
2011). This work was (partly) supported through the EU FP7
projects SOI-HITS (288481), GRAPHOL (2852754) and MSP
(611887).
References
4. Conclusions
1 J. Shin, S.-J. Choi, I. Lee, D.-Y. Youn, C. O. Park, J.-H. Lee,
H. L. Tuller and I.-D. Kim, Adv. Funct. Mater., 2013, 23,
2357–2367.
Zinc oxide nanorods were successfully integrated with SOI CMOS
MEMS platform using dip pen nanolithography. The zinc oxide
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 47609–47616 | 47615

View Article Online
RSC Advances
Paper
2 G. Peng, M. Hakim, Y. Y. Broza, S. Billan, R. Abdah-Bortnyak, 21 S. Wei, S. Wang, Y. Zhang and M. Zhou, Sens. Actuators, B,
A. Kuten, U. Tisch and H. Haick, Br. J. Cancer, 2010, 103, 542–
551.
3 S. Ammu, V. Dua, S. R. Agnihotra, S. P. Surwade,
2014, 192, 480–487.
22 T.-J. Hsueh, C.-L. Hsu, S.-J. Chang and I. C. Chen, Sens.
Actuators, B, 2007, 126, 473–477.
A. Phulgirkar, S. Patel and S. K. Manohar, J. Am. Chem. 23 M. Z. Ahmad, A. Z. Sadek, K. Latham, J. Kita, R. Moos and
Soc., 2012, 134, 4553–4556. W. Wlodarski, Sens. Actuators, B, 2013, 187, 295–300.
4 K. A. Mirica, J. G. Weis, J. M. Schnorr, B. Esser and 24 G. K. Mani and J. B. B. Rayappan, Sens. Actuators, B, 2013,
T. M. Swager, Angew. Chem., Int. Ed., 2012, 51, 10740–10745. 183, 459–466.
5 F. Schedin, A. K. Geim, S. V. Morozov, E. W. Hill, P. Blake, 25 M. S. Wagh, G. H. Jain, D. R. Patil, S. A. Patil and L. A. Patil,
M. I. Katsnelson and K. S. Novoselov, Nat. Mater., 2007, 6,
652–655.
6 J. W. Han, B. Kim, J. Li and M. Meyyappan, J. Phys. Chem. C,
2012, 116, 22094–22097.
Sens. Actuators, B, 2006, 115, 128–133.
26 M. S. Haque, K. B. K. Teo, N. L. Rupensinghe, S. Z. Ali,
I. Haneef, S. Maeng, J. Park, F. Udrea and W. I. Milne,
Nanotechnology, 2008, 19, 025607.
7 C. Hagleitner, A. Hierlemann, D. Lange, A. Kummer, 27 S. M. C. Vieira, P. Beecher, I. Haneef, F. Udrea, W. I. Milne,
N. Kerness, O. Brand and H. Baltes, Nature, 2001, 414, 293–
M. A. G. Namboothiry, D. L. Carroll, J. Park and S. Maeng,
296.
Appl. Phys. Lett., 2007, 91, 203111.
8 J. S. Suehle, R. E. Cavicchi, M. Gaitan and S. Semancik, IEEE 28 F. Torrisi, T. Hasan, W. Wu, Z. Sun, A. Lombardo,
Electron Device Lett., 1993, 14, 118–120.
T. S. Kulmala, G.-W. Hsieh, S. Jung, F. Bonaccorso,
P. J. Paul, D. Chu and A. C. Ferrari, ACS Nano, 2012, 6,
2992–3006.
9 S. Santra, S. Z. Ali, P. K. Guha, G. F. Zhong, J. Robertson,
J. A. Covington, W. I. Milne, J. W. Gardner and F. Udrea,
Nanotechnology, 2010, 21, 7.
10 S. Z. Ali, F. Udrea, W. I. Milne and J. W. Gardner, J.
Microelectromech. Syst., 2008, 17, 1408–1417.
29 P. K. Guha, S. Z. Ali, C. C. C. Lee, F. Udrea, W. I. Milne,
T. Iwaki, J. A. Covington and J. W. Gardner, Sens. Actuators,
B, 2007, 127, 260–266.
11 Y. Qiu and S. Yang, Adv. Funct. Mater., 2007, 17, 1345–1352. 30 S. Santra, P. K. Guha, S. Z. Ali, I. Haneef and F. Udrea, IEEE
12 C. Wang, D. Cai, B. Liu, H. Li, D. Wang, Y. Liu, L. Wang, Sens. J., 2008, 10, 997–1003.
Y. Wang, Q. Li and T. Wang, J. Mater. Chem. A, 2014, 2, 31 D. S. Ginger, H. Zhang and C. A. Mirkin, Angew. Chem., Int.
10623–10628. Ed., 2004, 43, 30–45.
13 S. Vallejos, P. Umek, T. Stoycheva, F. Annanouch, E. Llobet, 32 H. J. Pandya, S. Chandra and A. L. Vyas, Sens. Actuators, B,
X. Correig, P. De Marco, C. Bittencourt and C. Blackman,
2012, 161, 923–928.
Adv. Funct. Mater., 2013, 23, 1313–1322.
33 N. Van Hicu and N. D. Chien, Phys. B, 2008, 403, 50–56.
14 S. Santra, P. K. Guha, S. Z. Ali, P. Hiralal, H. E. Unalan, 34 W.-Y. Wu, J.-M. Ting and P.-J. Huang, Nanoscale Res. Lett.,
J. A. Covington, G. A. J. Amaratunga, W. I. Milne, 2009, 4, 513–517.
J. W. Gardner and F. Udrea, Sens. Actuators, B, 2010, 146, 35 S. Choopun, N. Hongsith, P. Mangkorntong and
559–565. N. Mangkorntong, Phys. E, 2007, 39, 53–56.
15 N. Gogurla, A. K. Sinha, S. Santra, S. Manna and S. K. Ray, 36 R. W. J. Scott, S. M. Yang, G. Chabanis, N. Coombs,
Sci. Rep., 2014, 4, 6483.
D. E. Williams and G. A. Ozin, Adv. Mater., 2001, 13, 1468–
16 Q. Yu, C. Yu, J. Wang, F. Guo, S. Gao, S. Jiao, H. Li, X. Zhang,
1472.
X. Wang, H. Gao, H. Yang and L. Zhao, RSC Adv., 2013, 3, 37 T. Rakshit, S. Santra, I. Manna and S. K. Ray, RSC Adv., 2014,
16619–16625. 4, 36749–36756.
17 R. C. Singh, O. Singh, M. P. Singh and P. S. Chandi, Sens. 38 S. Shi, Y. Liu, Y. Chen, J. Zhang, Y. Wang and T. Wang, Sens.
Actuators, B, 2008, 135, 352–357. Actuators, B, 2009, 140, 426–431.
18 S. Tian, F. Yang, D. Zeng and C. Xie, J. Phys. Chem. C, 2012, 39 N. Illyaskutty, H. Kohler, T. Trautmann, M. Schwotzer and
116, 10586–10591. V. P. M. Pillai, Sens. Actuators, B, 2013, 187, 611–621.
19 L. Wang, Y. Kang, X. Liu, S. Zhang, W. Huang and S. Wang, 40 P. R. S. Medeiros, J. G. Eon and L. G. Appel, Catal. Lett., 2000,
Sens. Actuators, B, 2012, 162, 237–243. 69, 79–82.
20 Z. Wen, L. Zhu, Z. Zhang and Z. Ye, Sens. Actuators, B, 2015, 41 F. Pourfayaz, A. Khodadadi, Y. Mortazavi and
208, 112–121.
S. S. Mohajerzadeh, Sens. Actuators, B, 2005, 108, 172–176.
47616 | RSC Adv., 2015, 5, 47609–47616
This journal is © The Royal Society of Chemistry 2015