ArticlePDF Available

Chemiresistive and Gravimetric Dual-Mode Gas Sensor toward Target Recognition and Differentiation


Abstract and Figures

We demonstrate a dual-mode gas sensor for simultaneous and independent acquisition of electrical and mechanical signals from the same gas adsorption event. The device integrates a graphene field-effect transistor (FET) with a piezoelectric resonator in a seamless manner by leveraging multiple structural and functional synergies. Dual signals resulting from independent physical processes, i.e., mass attachment and charge transfer can reflect intrinsic properties of gas molecules and potentially enable target recognition and quantification at the same time. Fabrication of the device is based on standard Integrated Circuit (IC) foundry processes and fully compatible with system-on-a-chip (SoC) integration to achieve extremely small form factors. In addition, the ability of simultaneous measurements of mass adsorption and charge transfer guides us to a more precise understanding of the interactions between graphene and various gas molecules. Besides its practical functions, the device serves as an effective tool to quantitatively investigate the physical processes and sensing mechanisms for a large library of sensing materials and target analytes.
Content may be subject to copyright.
Chemiresistive and Gravimetric Dual-Mode Gas Sensor toward
Target Recognition and Dierentiation
Yan Chen,
Hao Zhang,
Zhihong Feng,
Hongxiang Zhang,
Rui Zhang,
Yuanyuan Yu,
Jin Tao,
Hongyuan Zhao,
Wenlan Guo,*
Wei Pang,
Xuexin Duan,
Jing Liu,
and Daihua Zhang*
State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin University, Tianjin 300072, China
College of Precision Instrument and Optoelectronics Engineering, Tianjin University, Tianjin 300072, China
ABSTRACT: We demonstrate a dual-mode gas sensor for simultaneous and independent acquisition of electrical and
mechanical signals from the same gas adsorption event. The device integrates a graphene eld-eect transistor (FET) with a
piezoelectric resonator in a seamless manner by leveraging multiple structural and functional synergies. Dual signals resulting
from independent physical processes, i.e., mass attachment and charge transfer can reect intrinsic properties of gas molecules
and potentially enable target recognition and quantication at the same time. Fabrication of the device is based on standard
Integrated Circuit (IC) foundry processes and fully compatible with system-on-a-chip (SoC) integration to achieve extremely
small form factors. In addition, the ability of simultaneous measurements of mass adsorption and charge transfer guides us to a
more precise understanding of the interactions between graphene and various gas molecules. Besides its practical functions, the
device serves as an eective tool to quantitatively investigate the physical processes and sensing mechanisms for a large library of
sensing materials and target analytes.
KEYWORDS: gas sensor, graphene, solidly mounted resonator (SMR), resonators, electrical sensing
The development of modern gas sensors is currently driven by
two major trends. On one hand, research and development
eorts are targeting ever smaller form factors to promote sensor
integration with wearables and smart phones for enhanced
human interfacing and wireless connectivity, aiming to create
new application scenarios such as real time health monitoring,
remote diagnostics,
environmental sensor networks and others
yet to be imagined. Industry leaders including Bosch and
Cambridge CMOS Sensors have made successful penetration
into cell phone markets with their newly released environ-
mental sensor chips.
On the other hand, diverse real-life
applications such as breath analysis
and air quality
are calling for extended functionalities including
quantitative identication of gas mixers with unknown
concentrations. The target recognition and dierentiation
capabilities are becoming particularly and increasingly im-
portant to modern gas sensors.
Achieving satisfactory size reduction and functionality
enhancement at the same time, however, has proven extremely
challenging. Gas chromatography (GC)
and optical
oer excellent recognition capabilities, whereas
the space consuming components including GC columns and
optical pathways make system-on-chip (SoC) integrations
almost impossible. Power eciency is another concern as the
systems need to host active components (e.g., gas pumps,
photo diodes and coolers) with continuous power consumption
during operation. However, solid-state sensors based on
and piezoelectric resonators are ideal in
size and power eciency,
but the poor selectivity is severely
limiting their adoption in a large number of real life
Approaches to implementing target recognition solid-state
sensors are based upon primarily two mechanisms. The rst
one integrates the sensing membrane with a microheater, and
employs temperature transient response to dierentiate
Received: March 3, 2016
Accepted: July 26, 2016
Published: July 26, 2016
Research Article
© 2016 American Chemical Society 21742 DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
The approach has been fairly mature but yields
limited improvements. The second approach works by
functionalizing the sensing interface with molecular groups
that can either discriminatively interact with dierent analytes
or modulate the property of the sensing materials toward
specic targets.
In this case, multiple sensors with dierent
surface functionalizations are needed, and because of imperfect
specicity of each sensor, signal processing always involves
sophisticated pattern recognition algorithms.
In addition,
surface functionalization introduces complexity into device
fabrication, generally incompatible with standard IC foundry
processes for SoC integrations.
In this report, we propose a new approach to enabling target
recognition in solid-state gas sensors through dual-mode
sensing. Specically, the device is congured to simultaneously
capture the electrical and mechanical signals from the same
absorption event. As each type of gas molecule possesses a
unique set of physical properties (i.e., electron anity, polarity,
and mass) and undergoes qualitatively and/or quantitatively
dierent interactions with the sensing material, the combined
electrical and mechanical signals are expected to reect specic
and intrinsic information from individual targets. In fact, dual-
and multimode sensing have been utilized in biosensing
systems to investigate protein adsorption kinetics
and to
expand dynamic sensing range.
However, the concept has not
been demonstrated in gas sensors presumably due to challenges
in multimodal and real-time detection of gas molecules, which
are signicantly smaller and lighter compared to biomolecules.
2.1. Fabrication of SMR. The devices were fabricated on 100 mm
undoped Si wafers, starting from deposition of the Bragg reector. The
alternating AlN and SiO2layers were deposited through PVD and
CVD, respectively. The thicknesses of each AlN/SiO2pair was set to
be 1200/700, 1000/1300, and 1000/650 nm from bottom to top. BE
of the SMR was made of 600 nm thick Mo and deposited via PVD on
top of the Bragg reector. The lm was then patterned through
photolithography and plasma etching into isolated islands. PVD was
used again to deposit the piezoelectric layer, a 1000 nm thick AlN lm
on top of the BE. Orientation of the AlN crystal was along c-axis. X-ray
diraction spectrum indicated a sharp peak at (002) with a Full-Width-
at-Half-Maximum (FWHM) of 1.5°. In the nal step, the SMR was
capped with a pair of Au top electrodes deposited using Ebeam
evaporation followed by wet etch. Thicknesses of the Au electrodes
and the underlying Cr adhesive layer were 300 and 50 nm,
respectively. It is worth mentioning that the inner electrode (TIE)
was intentionally shaped like a pentagon in order to suppress spurious
resonance in the device. The electrode area was congured to be 3.0 ×
104μm2so that the SMR has a characteristic impedance of 50 Ωin
order to match the impedance of external circuits. The outer electrode
(TOE) was separated 15 μm away from the TIE and grounded during
operation. The area of TOE was several times larger than that of TIE.
This geometrical arrangement ensures good E-eld connement
within the active area under the TIE electrode.
2.2. Graphene Transfer and Patterning. The CVD-grown
monolayer graphene lm was purchased from VIGON Technologies.
The lm rst went through a thorough RCA cleaning process
remove surface contaminants. Prior to transfer, the device wafer was
pretreated with O2plasma (100 W, 1 min) to clean the surface and
condition it to a hydrophilic interface. The graphene lm was then
pressed against the wafer with the support of a poly methyl
methacrylate (PMMA) stamp and held in position for 2 h at room
temperature. The xture was then heated to 150 °C with a hot plate
and kept at the temperature for 15 min. The heating promotes van der
Waals binding at the graphene-SMR interface, leaving an even and
strongly adhered graphene lm after removal of the PMMA
stamp.Trimming of the graphene lm was done with Ebeam
lithography followed by O2plama etching. We used a negative
Ebeam resist from Allresist, which introduced much less surface
contaminants compared to most photo resists. Recipe of the O2
plasma etching was 100 W/70 sccm/50 mTorr. The treatment
duration was 60 s.
Our device consists of a thin-lm piezoelectric resonator and a
graphene eld eect transistor (FET) seamlessly integrated
with each other. The resonator is made of ultralight
microstructures and operates at gigahertz frequency, which is
signicantly higher than those of the multimode biosensors.
This enables the device to develop detectable frequency shift
upon attachment of very tiny masses like gas molecules. At the
same time, the graphene FET senses charge transfers at the
sensing interface and provides independent signals from the
same adsorption event. Due to its atomically thin nature, the
graphene induces negligible perturbation to the underlying
Figure 1. (a) Cross-sectional illustration of the multilayered device structure. From bottom to top are high-resistivity silicon substrate, Bragg
reector (containing 6 AlN and SiO2layers to attenuate acoustic waves leaking to the Si substrate (inset)), bottom electrode (BE), piezoelectric
layer, top electrodes (TIE and TOE), and monolayer graphene lm. (b) The entire process ow includes Bragg reector deposition (1), BE
deposition and patterning (2), piezoelectric layer deposition (3), TIE/TOE deposition and wet etch (4), and graphene transfer and patterning (5).
(c) SEM image of a completed device. Scale bar is 100 μm.
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
piezoelectric resonator. Instead, it considerably enhances the
sensitivity of the resonator by providing a highly active surface
for gas adsorption.
We also note that the device fabrication is
entirely compatible with IC processes and system-level
integration. The simple structure also supports integration
with microheaters and surface functionalizations to further
expand the dimensionality of parameter space. In our studies,
we used two common air pollutants, nitrogen dioxide (NO2)
and ammonia (NH3), and two volatile organic compounds
(VOCs), ethanol (C2H5OH) and n-hexane (C6H14) as target
gases to demonstrate the operation and performance of the
dual-mode sensor.
Figure 1a shows the schematic of the dual-mode device,
consisting of a piezoelectric solidly mounted resonator (SMR)
and a monolayer graphene lm attached on top of it. The inset
of Figure 1a depicts the prole of the strain waves across
dierent layers. Thickness of each AlN/SiO2layer is set to be
one forth of the wavelength. Sharp acoustic impedance
mismatch between AlN and SiO2results in quick attenuation
of the strain wave across the Bragg reector, and eectively
minimizes energy dissipation to the underlying substrate. We
chose Bragg reector over air cavity as the impedance mismatch
layer for the consideration of process stability. Compared to a
suspended membrane, a rigid structure can eectively mitigate
stress issues commonly present in AlN lms, and form a sturdy
base for subsequent processes and functional structures built on
the top.
The active layers of the SMR, including the bottom electrode
(BE), piezoeletric AlN layer and two top electrodes (denoted as
TIE and TOE for inner and outer electrodes) were sequentially
deposited on the Bragg reector. Thickness and quality of the
lms determine key parameters of the SMR including its
impedance, resonant frequcies, and eective coupling coef-
cient. Deposition of the AlN layer was under strict process
control with well-optimized parameters to ensure good
crystallinity. The lm exhibited excellent piezoelectric proper-
ties as demonstrated in our previous studies on Bulk Acoustic
Wave (BAW)
and Contour Mode Resonators (CMRs).
After completion of the SMR, a monolayer graphene lm was
then transferred onto the surface to form a semiconducting
channel between the two top electrodes. The graphene lm was
then trimmed to within the active area of the SMR to avoid
undesired coupling with neighboring devices. The entire
process ow is summarized in Figure 1b. The scanning
electron microscopy (SEM) image of Figure 1c shows a
completed device. Contours of the TIE and TOE electrodes are
clearly visible through the semitransparent graphene lm. The
bottom electrode is buried under the AlN layer therefore
unseen from this image.
It can be seen that the device architecture leverages several
synergies between the SMR and the graphene FET. The TIE
and TOE of the SMR serve as the source and drain electrodes
of the FET at the same time; the BE is used to create an
equipotential surface to shape the RF eld inside the
piezoelectric layer. The same electrode can also be biased to
gate the graphene lm to modulate its carrier concentration.
The sandwiched AlN lm generates and hosts electro-
mechanical resonance within the SMR, and simultaneously
works as the gate dielectric of the graphene FET. As mentioned
above, the graphene lm functions not only as a FET channel,
but also as an eective surface coating to promote surface
adsorption of gas molecules and boost mass detection
sensitivity of the SMR.
Figure 2a shows the measurement setup to characterize the
dual-mode device. The chip was mounted onto an evaluation
board with the TIE and TOE connected to the center pin and
outer shell of a coaxial small-A-type (SMA) connector (Figure
2a inset). The board was then connected to a Bias-Tee to split
radio frequency (RF) and direct-current (DC) signals from
each other. The conguration allows the dual modes to operate
independently and simultaneously at distinct frequencies
(gigahertz vs DC) with negligible interference. The two output
terminals of the Bias-Tee were separately hooked up to a
network analyzer (Agilent E5071B) and a source meter
(Agilent B1500) to characterize the RF and DC responses.
We note that signals from both channels can be electrically read
out, which gives the device great advantage over other
multimode sensors that rely on optical means.
capability of electrical readout eliminates the need of complex
and expensive optical components, making the sensor chip fully
compatible with system-level integrations in wearable and
mobile platforms.
The measurement of Figure 2b was taken from the RF port
of the Bias-Tee. The red and blue curves plot the magnitude
and phase of the device impedance as a function of frequency
from 1.55 to 1.80 GHz. The dual-mode device behaves
essentially the same as a bare SMR despite a moderate decrease
in Q factor from 480 to 260. The additional energy loss is
primarily due to nite resistance of the graphene lm, which
adds an extra energy dissipation channel in parallel with the
SMR. The valley and peak of the red curve at 1.660 and 1.693
GHz correspond to series and parallel resonance, respectively.
Both frequencies (denoted as fsand fp) are sensitively
dependent on mass load on the device, and can therefore be
Figure 2. (a) Test conguration and picture of the evaluation board (inset). The Bias-Tee separates RF and DC signals from the same port (TIE)
and feeds them to network analyzer and semiconductor parameter analyzer, respectively. (b) RF output of the device. Red and blue curves are the
magnitude and phase of electrical impedance at various frequencies, exhibiting sharp transitions at series (1.660 GHz) and parallel (1.693 GHz)
resonance frequencies. (c) DC output of the device showing the transfer and output (inset) characteristics of the graphene FET. Current through the
drain and source electrodes (Ids) is conducted by holes under zero gate-source bias (Vgs) according to the IdsVgs curve. The inset shows the IdsVds
curves under dierent Vgs (10, 20, 30, and 40 V).
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
used to precisely quantify molecular adsorptions upon gas
exposures. The phase change (blue curve) of the SMR indicates
transitions between capacitive (negative phase) and inductive
(positive phase) reactance occurring at the two resonance
frequencies, a characteristic feature of BAW resonators
operating in thickness extension (TE) mode.
Output from the DC port of the Bias-Tee reects DC
characteristics of the dual mode device. Due to nearly innite
DC impedance of the SMR, the device behaves the same as a
standard graphene FET. Figure 2c and the inset present the
transfer and output characteristics of the bipolar FET. Surface
doping of oxygen and moisture from ambient air leads to a p-
type channel and shifts the charge neutrality point (CNP)
toward positive gate bias (30 V). Hole mobility of the device
is derived to be 87.4 cm2/(V s),
which is comparatively low
among CVD-grown graphene lms (1002000 cm2/(V s)).
This implies a relatively large number of intrinsic defects/
impurities as scattering centers. Despite the resultant reduction
in carrier mobility, these local defects provide active bonding
sites for gas molecules and are favorable to sensing
The measurements in Figure 3a,b characterize the mechan-
ical (RF) response of the sensor to mass adsorption in diluted
NH3and C2H5OH. High purity (99.999%) N2was used as the
carrier gas with a constant ow rate of 1000 sccm. The two
gures plot the shift in series resonance frequency (Δfs) of the
SMR during exposure to three dierent concentrations. The
resonance frequency (fs) drops as the gas molecules attach onto
the graphene lm and plateaus when surface adsorption and
desorption reach equilibrium. The following equation allows us
to correlate the frequency shift with the amount of mass
adsorption on the graphene surface:
where msis the surface density of the adsorbed mass, in kg/m2.
The minus sign indicates that fsdrops with increasing ms. The
ρ0(in kg/m3) and d0(in m) are the density and thickness of
the SMR. In this particular case, the product of ρ0and d0is the
sum of three components contributed by the BE (Mo), TIE
(Au), and the piezoelectric layer (AlN), respectively:
ρρρ=++dd d d
00Mo Mo Au Au AIN AI
plugging the numbers (ρMo = 10.2 g/cm3,dMo = 0.6 μm, ρAu =
19.3 g/cm3,dAu = 0.3 μm, ρAlN = 3.3 g/cm3,dAlN =1μm) in to
eq 2 results in ρ0d0= 1.52 ×102kg/m3.
We carried out measurements with all four gases and plotted
the frequency shifts at equilibrium as a function of gas
concentration in Figure 3c. The lowest detectable concen-
trations were 10, 100, 500, and 1000 ppm for NO2,NH
C2H5OH, and C6H14, respectively. For the two inorganic gases
NH3and NO2,Δfsshows a clear trend of saturation at high
concentrations (red and blue curves), whereas the trend is
absent with the two VOC gases, C2H5OH, and C6H14 (green
and black curves), which result in linear response of Δfsacross
the entire concentration range. For NH3and NO2, the data can
be modeled with Langmuir Isotherm:
Le (3)
where csis dened as the surface adsorption density of gas
molecules (in mol/m3), which relates to msas cs=ms/M, with
M being the molar mass in kg/mol. cmax is a constant denoting
the maximum value of cswhen adsorption saturates the surface.
Ce(in ppm) is the volume concentration of target gas, and aL
(in 1/ppm) the ratio between adsorption and desorption
constants. Since the change in resonance frequency, as
discussed above, is proportional to surface adsorption density,
we can equate cs/cmax with Δfs/Δfmax, with Δfmax being the
maximum frequency shift when all adsorption sites on graphene
are fully occupied. This way we relate Δfsdirectly to Ce:
max Le
Le (4)
The two coecients, Δfmax and aLcan then be derived from
the intercept and slope of the linear tting between 1/Δfsand
1/Ce. The Δfmax of NH3and NO2are calculated to be 21.4 and
112 kHz, respectively, which correspond to cmax = 11.6 ×106
Figure 3. Shift in series resonance frequency (Δfs) during exposure to dierent concentrations of NH3(a) and C2H5OH (b). (c) Magnitude of
frequency shift (|Δfs|) as a function of gas concentration (Ce)ofNH
3(red), NO2(blue), C2H5OH (green), and C6H14 (black). The red and blue
lines are ttings based on Langmuir Isotherm. Source-drain bias was 5 mV. Parts (d) and (e) are real-time change in FET current recorded
simultaneously with the Δfsdata in (a) and (b). (f) Magnitude of current change (|ΔI|) at various gas concentrations when system reaches
equilibrium. Data points of NH3(red) and NO2(blue) use the scales on the left axis, while C2H5OH and C6H14 share the scales on the right.
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
mol/m2(NH3) and 22.2 ×106mol/m2(NO2) according to
eq 1. We can then use the length of NH and NO bonds to
estimate the area each adsorbed molecule occupies on the
graphene surface, and obtain a rough estimation of the
maximum coverage ratio at saturation adsorption. The numbers
are derived to be 8% for NH3and 20% for NO2. The results
suggest that only a small portion of the graphene surface is
attached with NH3or NO2molecules at saturation. This is
presumably due to the fact that the molecular attachment is
through chemisorption toward preferential sites, which is more
selective compared to physisorption that primarily relies on van
der Waals force. In fact, we believe that physisorption
dominates the case of C2H5OH, and C6H14 despite weak
charge transfers (as will be discussed in the next section). Using
eq 1 and assuming single layer adsorption, the coverage ratios
of C2H5OH, and C6H14 are estimated to be 0.2 and 0.73 at Ce=
3000 ppm. The adsorption kinetics is unable to be tted with
the Langmuir Isotherm in this case, and multilayer adsorption
will likely occur after the rst layer saturates. We note that
direct measurement of surface adsorption density is a unique
and important advantage of gravimetric sensors over chemir-
esistors. The latter is more prevalent but unable to correlate
sensing signals directly to the amount of surface adsorptions.
Figure 3df discusses the FET sensing signals acquired
through the DC channel. Figure 3d,e shows real time
monitoring of the change in source-drain current (ΔI) (at
source-drain bias of 5 mV) during NH3and C2H5OH exposure.
The data were simultaneously recorded with the Δfsmeasure-
ments in Figure 3a,b. The DC current responded to NH3and
C2H5OH in opposite directions. NH3is a reducing gas and
withdraws holes from graphene,
thus lowering the carrier
concentration within the p-type channel. On the contrary,
C2H5OH is a weak oxidizing gas and able to raise the hole
concentration by withdrawing electrons.
Similar to Figure 3c,
we summarize the magnitude of current change (|ΔI|)at
various concentrations (Ce) for all four gases in Figure 3f. We
note that the sign of ΔIfor NH3and C6H14 was inverted in the
plot for better presentation and comparison. The detection
limits to NO2and NH3were about 0.5 and 2.5 ppm with the
graphene sensor, signicantly lower than those of the SMR.
NO2and C6H14 are another pair of oxidizing and reducing
gases that result in opposite changes in source-drain current. In
addition, charge transfers induced by the two VOC gases,
C2H5OH and C6H14 are signicantly less than those of NH3
and NO2, as evidenced by the contrast between the scales on
left and right axes. It is interesting to note that the proles of
|ΔI|Cecurves considerably dier from the |Δf|Cecurves in
Figure 3c. The dierence is attributed to changes in carrier
mobility in addition to charge transfers during gas adsorption.
We will continue the discussion in later sections.
The simultaneous dual-mode measurements enable a very
important analysis, that is, to directly correlate the current
change |ΔI|with surface adsorption densities cs.Figure 4 shows
the correlation for all four gases. We have also included the Δfs
data for easy comparison, which, by the denition of eq 1, are
always proportional to csin all cases. Since the scales are quite
dierent for various gases, we split the curves into separate
parts in Figure 4ad. Previous studies on FET gas sensors
generally assumed that changes in source-drain current were
proportional to densities of surface adsorption. The assumption
holds well for small molecules with strong charge transfer
tendency (e.g., NH3and NO2) at very low concentrations,
while Figure 4 gives a more precise idea about how the sensing
behavior deviates from this simplest approximation in more
generic cases. In general, |ΔI|cscurves are not straight lines.
They are either concave (NH3and C6H14) or convex (NO2and
C2H5OH) depending on direction of charge transfer at the
sensing interface. When gas molecules are adsorbed on the
graphene surface, they reduce hole mobility by adding extra
scattering centers in the conduction channel.
Since the
current density is proportional to the product of carrier
mobility and concentration, the net eect of mobility reduction
and charge transfer leads to pronounced deviation from a linear
|ΔI|csrelationship. Specically, for reducing gases that
withdraw holes from the channel, the |Δ|cscurves are
concave (Figure 4ad) as the changes in mobility and carrier
concentration act in the same direction, leading to accelerat-
ingdecrease in current density; whereas for oxidizing gases,
the decrease in mobility counteracts the increase in hole
concentration and weakens the sensing response at high
concentrations, resulting in convex |ΔI|cscurves in Figure
We redraw the |ΔI|cscurves of NH3and NO2with more
details in Figure 5 for quantitative analysis. The source-drain
current (I) can be written as (assuming metallic contacts as
suggested by the output characteristics management in Figure
2c inset)
where Vis the source-drain bias and set to be 5 mV throughout
the measurements; Wand Lare width and length of the FET
conduction channel, which are 400 and 15 μm, respectively; e=
1.6 ×1019 C is the elementary charge; nand μare hole
concentration (in 1/m2) and mobility (in m2/V·s) of the
graphene FET. We assume that mobility change at very low gas
concentrations is largely negligible, as suggested by previous
observations in graphene
and black phosphorus sensors.
can therefore use the initial carrier mobility μ0=87cm
make a linear tting of the |ΔI|vs cscurve at small
concentrations, as shown in Figure 5a,b. Slope of the linear
tting can be used to derive the amount of charge each gas
molecule exchanges with the graphene lm. The number is
generally a small portion of the elementary charge (e) and
Figure 4. Correlation between |ΔI|(red), |Δfs|(blue), and surface
adsorption density (cs)ofdierent gas molecules. Reducing gases like
NH3(a) and C6H14 (d) follow concave curves, while oxidizing gases
like NO2(b) and C2H5OH (c) have a convex shape.
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
reects very intrinsic property of the target gas associated with
its electron anity. We dene nias the charge transfer amount
by single gas molecule (in C) divided by e. The numbers are
calculatedtobe0.008 (NH3), 0.049 (NO2), 0.002
(C2H5OH), and 0.003 (C6H14) for the four gases. The plus
and minus signs denote hole donating and withdrawing,
respectively. The results of NO2and NH3agree relatively well
with previous studies.
The data of C2H5OH and C6H14 are,
however, hard to nd references to compare with due to the
lack of studies on this particular topic.
At higher gas concentrations, the eect of mobility
modulation starts to take a more dominating role. As a
reasonable approximation, we assume the change of carrier
mobility is governed by Matthiessens rule:
μμ μ=+ +
−− −
where μlattice is the carrier mobility that the material would have
if only lattice phonon scattering exits; μimpurity and μadsorption are
the mobilities when assuming only impurity scattering or
surface adsorption scattering, respectively. To the rst-order
approximation, μadsorption is inversely proportional to the density
of surface adsorption (cs).
Therefore, the plot of 1/μcs
should display a linear trend when csis suciently large. This
has been veried in Figure 5c with both NH3and NO2. Slope
of the linear tting is an important parameter that characterizes
the eciency of mobility modulation by unit density of surface
adsorption. The two tting lines have similar intercepts around
105 V·s/m2. This is not surprising as the intercept corresponds
to μlattice
1, which is an intrinsic value of the
graphene FET and independent of the type of target gases.
The correlation between ΔIand cs(or equivalently ΔIvs
Δfs) not only provides valuable insights into the adsorption
kinetics and gas-graphene interactions, but also enables a very
unique and important feature of the dual mode sensor. In real-
life applications such as human breath and auto exhaust
analysis, the application environment is much more complex
compared to most lab experiments, which commonly assume
the onset of target gas to follow perfect step functions with
well-dened start/stop time and a steady ow in between.
realistic situations, however, concentrations of target gases
follow continuous and relatively random waveforms. This
introduces additional variables and makes the measurement and
recognition extremely challenging with single-mode sensors. In
contrast, the dual mode sensor is theoretically able to
discriminate gas species by generating a ΔIΔfscurve in
real time as the concentration varies. Each specic gas type (or
mixture) would correspond to a unique trace on the ΔIΔfs
plane independent of instantaneous concentration or shape of
the waveform. At the same time, the start and end points of the
trace mark the minimum and maximum concentrations during
the exposure. In the case of dual mode sensing, the variation
(or instability) of concentration is in fact benecial for target
recognition. With small portable sensors, users can even scan
the concentration initiatively by varying the sensor-to-source
distance to achieve higher accuracy.
In the last part, we demonstrate the usage of the dual mode
sensor for quantitative analysis of multicomponent gas
mixtures. For simplicity, we consider a trinary gas mixture
containing trace amount of gas A and B diluted in carrier gas C.
The types of all components are known but the concentrations
of A and B are wanted. We note that this simplied problem
setting is actually very representative of a large number of
application scenarios, in which the absolute or relative
concentrations of the two most important labelsare
substantially decisive. Examples include ethanethiol and
methanethiol in human breath for diagnosis of chronic liver
acetone and ammonia for diabetes/hyperglycemia,
and pentane and dimethyl selenide for early detection of cystic
To give specic numbers, we set A = NO2,B=
C2H5OH, and C = air, and use the tting curves of the sensing
data from single component to construct the plot in Figure 6a.
Figure 5. |ΔI|cscorrelations of NH3(a) and NO2(b). Slope of the linear tting at low concentrations reects the amount of charge transfer per
gas molecule. (c) Reciprocal of carrier mobility (1/μ) as a function of cswith linear tting lines for data points at high concentrations.
Figure 6. (a) Calculated ΔIand Δfsvalues under all concentration combinations of gas A (NO2) and gas B (C2H5OH) within the range of CA= (0,
100 ppm) and CB= (0, 7500 ppm). (b) Experimental data of the gas mixture. (The red circles indicate actual concentrations and the green dots
indicate the values derived from the ΔI/Δfsmeasurements). The dashed lines shows the ΔIand Δfscontour lines projected on CACBplane.
Coordinates of each intersection point correspond to denitive (CA,CB) values.
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
When all possible concentrations (CAand CB) are sampled, the
measured ΔIand Δfsvalues form two curved surfaces in the
parameter space. Each xed ΔI(or Δfs) value corresponds to a
series of CA/CBcombinations that form a specic contour line
in Figure 6b. The intersection point of each pair of the ΔI/Δfs
contour lines determines the corresponding CAand CBvalues
We have selected 16 pairs of CA/CBcombinations to
experimentally verify the hypothesis. ΔIand Δfswere recorded
under these concentrations and used to derive the CA/CB
values. These numbers are also plotted in Figure 6 (green dots
with error bars) to be compared with the actual concentrations
(red circles in Figure 6b). It can be seen that most data points
match reasonably well with the actual values (within ±520%).
The demonstration gives a good example of gas mixture
analysis by leveraging independent readings from both the
current and frequency signals. The conclusion should also hold
for more practical carrier gases (e.g., breath air) with multiple
background components, which would induce a shift in
background signal but preserve the systematic trends with
varying CAand CB. Finally, we note that the model assumes
independent interactions of dierent components. Cross-talks
in terms of screening eect or other interactions were not
considered. We believe it is reasonable when the concentrations
of the target species are substantially low.
In this work, we developed a dual-mode gas sensor based on
SMR and graphene FET to simultaneously measure the mass
adsorption and charge transfer occurring at the sensing
interface. Four target gases including NO2,NH
and C6H14 were used to evaluate the function and performance
of the device. The SMR provides precise measurement of the
amount of gas molecules adsorbed on the graphene surface.
The surface adsorption density is crucial information for us to
understand the adsorption kinetics of various gas molecules,
and more importantly, when combined with FET signals, it
provides intrinsic information on the target gas such as redox
properties and electron anity. This capability can therefore be
tailored to recognize and dierentiate targets, or even quantify
the concentration of individual component in mixtures without
the help of complicated data processing algorithms. Our
scheme can also work together with other approaches such as
surface functionalization and microheaters to further expand
the parameter space and eventually achieve accurate target
recognition and quantication in complex gaseous environment
for real-life applications.
Corresponding Authors
*E-mail: (W.G.).
*E-mail: (D.Z.).
The authors declare no competing nancial interest.
The authors acknowledge nancial support by Tianjin Applied
Basic Research and Advanced Technology (13JKYBJC37100)
and 111 Project (B07014).
(1) Weisskopf, M. G.; Grodstein, F.; Ascherio, A. Prospective Study
of Smoking and Decline in Cognitive Function in Parkinsons Disease.
Mov. Disord. 2006,21, S103S103.
(3) Santra, S.; Guha, P. K.; Ali, S. Z.; Hiralal, P.; Unalan, H. E.;
Covington, J. A.; Amaratunga, G. A. J.; Milne, W. I.; Gardner, J. W.;
Udrea, F. ZnO Nanowires Grown on SOI CMOS Substrate for
Ethanol Sensing. Sens. Actuators, B 2010,146, 559565.
(4) Jerger, A.; Kohler, H.; Becker, F.; Keller, H. B.; Seifert, R. New
Applications of Tin Oxide Gas Sensors II. Intelligent Sensor System
for Reliable Monitoring of Ammonia Leakages. Sens. Actuators, B 2002,
81, 301307.
(5) Seifert, R.; Keller, H. B.; Frank, K.; Kohler, H. ProSens-An
Efficient Mathematical Procedure for Calibration and Evaluation of
Tin Oxide Gas Sensor Data. Sens. Lett. 2011,9,710.
(6) Shehada, N.; Bronstrup, G.; Funka, K.; Christiansen, S.; Leja, M.;
Haick, H. Ultrasensitive Silicon Nanowire for Real-World Gas Sensing:
Noninvasive Diagnosis of Cancer from Breath Volatolome. Nano Lett.
2015,15, 12881295.
(7) Hakim, M.; Broza, Y. Y.; Barash, O.; Peled, N.; Phillips, M.;
Amann, A.; Haick, H. Volatile Organic Compounds of Lung Cancer
and Possible Biochemical Pathways. Chem. Rev. 2012,112, 5949
(8) Broza, Y. Y.; Haick, H. Nanomaterial-Based Sensors for Detection
of Disease by Volatile Organic Compounds. Nanomedicine 2013,8,
(9) Mai, L. Q.; Xu, L.; Gao, Q. A.; Han, C. H.; Hu, B.; Pi, Y. Q. Single
beta-AgVO3 Nanowire H2S Sensor. Nano Lett. 2010,10, 26042608.
(10) Moon, S. E.; Lee, H. K.; Choi, N. J.; Lee, J.; Yang, W. S.; Kim, J.;
Jong, J. J.; Yoo, D. J. Low-Power-Consumption Metal Oxide NO2Gas
Sensor Based on Micro-Heater and Screen Printing Technology. J.
Nanosci. Nanotechnol. 2012,12, 55435546.
(11) Buszewski, B.; Ulanowska, A.; Ligor, T.; Denderz, N.; Amann,
A. Analysis of Exhaled Breath from Smokers, Passive Smokers and
Non-Smokers by Solid-Phase Microextraction Gas Chromatography/
Mass Spectrometry. Biomed. Chromatogr. 2009,23, 551556.
(12) Junger, M.; Bodeker, B.; Baumbach, J. I. Peak Assignment in
Multi-Capillary Column-Ion Mobility Spectrometry using Compara-
tive Studies with Gas Chromatography-Mass Spectrometry for VOC
Analysis. Anal. Bioanal. Chem. 2010,396, 471482.
(13) Maharana, P. K.; Jha, R.; Padhy, P. On the Electric Field
Enhancement and Performance of SPR Gas Sensor Based on
Graphene for Visible and Near Infrared. Sens. Actuators, B 2015,
207, 117122.
(14) Wang, Y. J.; Wang, Y. T.; Liu, J. A Novel CO Gas Optical Fiber
Sensor Based on Spectrum Absorption. Proceedings of the First
International Symposium on Test Automation &Instrumentation 2006,
13, 602605.
(15) Zou, X. M.; Wang, J. L.; Liu, X. Q.; Wang, C. L.; Jiang, Y.;
Wang, Y.; Xiao, X. H.; Ho, J. C.; Li, J. C.; Jiang, C. Z.; Fang, Y.; Liu,
W.; Liao, L. Rational Design of Sub-Parts per Million Specific Gas
Sensors Array Based on Metal Nanoparticles Decorated Nanowire
Enhancement-Mode Transistors. Nano Lett. 2013,13, 32873292.
(16) Feng, P.; Shao, F.; Shi, Y.; Wan, Q. Gas Sensors Based on
Semiconducting Nanowire Field-Effect Transistors. Sensors 2014,14,
(17) Tukkiniemi, K.; Rantala, A.; Nirschl, M.; Pitzer, D.; Huber, T.;
Schreiter, M. Fully Integrated FBAR Sensor Matrix for Mass
Detection. Procedia Chem. 2009,1, 10511054.
(18) Koshets, I. A.; Kazantseva, Z. I.; Shirshov, Y. M.; Cherenok, S.
A.; Kalchenko, V. I. Calixarene Films as Sensitive Coatings for QCM-
Based Gas Sensors. Sens. Actuators, B 2005,106, 177181.
(19) Llobet, E.; Vilanova, X.; Brezmes, J.; Sueiras, J. E.; Correig, X.
Transient Response of Thick-Film Tin Oxide Gas-Sensors to
Multicomponent Gas Mixtures. Sens. Actuators, B 1998,47, 104112.
(20) Hayashi, K.; Yokota, Y.; Tachibana, T.; Kobashi, K.; Fukunaga,
T.; Takada, T. Characteristics of Diamond Film Gas Sensors upon
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
Exposure to Semiconductor Doping Gases. J. Electrochem. Soc. 2001,
148, H17H20.
(21) Broza, Y. Y.; Mochalski, P.; Ruzsanyi, V.; Amann, A.; Haick, H.
Hybrid Volatolomics and Disease Detection. Angew. Chem., Int. Ed.
2015,54, 1103611048.
(22) Wang, B.; Cancilla, J. C.; Torrecilla, J. S.; Haick, H. Artificial
Sensing Intelligence with Silicon Nanowires for Ultraselective
Detection in the Gas Phase. Nano Lett. 2014,14, 933938.
(23) Goda, T.; Maeda, Y.; Miyahara, Y. Simultaneous Monitoring of
Protein Adsorption Kinetics Using a Quartz Crystal Microbalance and
Field-Effect Transistor Integrated Device. Anal. Chem. 2012,84,
(24) Goda, T.; Miyahara, Y. Interpretation of Protein Adsorption
through Its Intrinsic Electric Charges: A Comparative Study Using a
Field-Effect Transistor, Surface Plasmon Resonance, and Quartz
Crystal Microbalance. Langmuir 2012,28, 1473014738.
(25) Zhu, A. Y.; Yi, F.; Reed, J. C.; Zhu, H.; Cubukcu, E.
Optoelectromechanical Multimodal Biosensor with Graphene Active
Region. Nano Lett. 2014,14, 56415649.
(26) Sakata, T.; Fukuda, R. Simultaneous Biosensing with Quartz
Crystal Microbalance with a Dissipation Coupled-Gate Semiconductor
Device. Anal. Chem. 2013,85, 57965800.
(27) Quang, V. V.; Hung, V. N.; Tuan, L. A.; Phan, V. N.; Huy, T.
Q.; Quy, N. V. Graphene-Coated Quartz Crystal Microbalance for
Detection of Volatile Organic Compounds at Room Temperature.
Thin Solid Films 2014,568,612.
(28) Zhang, Z. X.; Liang, J.; Zhang, D. H.; Pang, W.; Zhang, H. A
Novel Bulk Acoustic Wave Resonator for Filters and Sensors
Applications. Micromachines 2015,6, 13061316.
(29) Liang, J.; Zhang, H. X.; Zhang, D. H.; Duan, X. X.; Zhang, H.;
Pang, W. Design and Fabrication of Aluminum Nitride Lamb Wave
Resonators Towards High Figure of Merit for Intermediate Frequency
Filter Applications. J. Micromech. Microeng. 2015,25, 035016.
(30) Kumar, S.; Peltekis, N.; Lee, K.; Kim, H. Y.; Duesberg, G. S.
Reliable Processing of Graphene using Metal Etchmasks. Nanoscale
Res. Lett. 2011,6, 390.
(31) Hashimoto, K. RF Bulk Acoustic Wave Filters for
Communications[M]. Artech House 2009,1, 9781596933217.
(32) Xu, H. L.; Zhang, Z. Y.; Xu, H. T.; Wang, Z. X.; Wang, S.; Peng,
L. M. Top-Gated Graphene Field-Effect Transistors with High
Normalized Transconductance and Designable Dirac Point Voltage.
ACS Nano 2011,5, 50315037.
(33) Reina, A.; Jia, X. T.; Ho, J.; Nezich, D.; Son, H. B.; Bulovic, V.;
Dresselhaus, M. S.; Kong, J. Large Area, Few-Layer Graphene Films on
Arbitrary Substrates by Chemical Vapor Deposition. Nano Lett. 2009,
(34) Kumar, B.; Min, K.; Bashirzadeh, M.; Farimani, A. B.; Bae, M.
H.; Estrada, D.; Kim, Y. D.; Yasaei, P.; Park, Y. D.; Pop, E.; Aluru, N.
R.; Salehi-Khojin, A. The Role of External Defects in Chemical Sensing
of Graphene Field-Effect Transistors. Nano Lett. 2013,13, 19621968.
(35) Dan, Y. P.; Lu, Y.; Kybert, N. J.; Luo, Z. T.; Johnson, A. T. C.
Intrinsic Response of Graphene Vapor Sensors. Nano Lett. 2009,9,
(36) Schedin, F.; Geim, A. K.; Morozov, S. V.; Hill, E. W.; Blake, P.;
Katsnelson, M. I.; Novoselov, K. S. Detection of Individual Gas
Molecules Adsorbed on Graphene. Nat. Mater. 2007,6, 652655.
(37) Li, J.; Lu, Y. J.; Ye, Q.; Cinke, M.; Han, J.; Meyyappan, M.
Carbon Nanotube Sensors for Gas and Organic Vapor Detection.
Nano Lett. 2003,3, 929933.
(38) Zhang, D. H.; Liu, Z. Q.; Li, C.; Tang, T.; Liu, X. L.; Han, S.;
Lei, B.; Zhou, C. W. Detection of NO2Down to ppb Levels using
Individual and Multiple In2O3Nanowire Devices. Nano Lett. 2004,4,
(39) Qi, P.; Vermesh, O.; Grecu, M.; Javey, A.; Wang, O.; Dai, H. J.;
Peng, S.; Cho, K. J. Toward Large Arrays of Multiplex Functionalized
Carbon Nanotube Sensors for Highly Sensitive and Selective
Molecular Detection. Nano Lett. 2003,3, 347351.
(40) Cui, S. M.; Pu, H. H.; Wells, S. A.; Wen, Z. H.; Mao, S.; Chang,
J. B.; Hersam, M. C.; Chen, J. H. Ultrahigh Sensitivity and Layer-
Dependent Sensing Performance of Phosphorene-Based Gas Sensors.
Nat. Commun. 2015,6, 8632.
(41) Peng, S.; Cho, K. J. Chemical Control of Nanotube Electronics.
Nanotechnology 2000,11,5760.
(42) Sze, S M, Ng, K K. Physics of Semiconductor Devices; John Wiley
& Sons; New York, 2006, 3, 9780471143239.
(43) Shuster, G.; Baltianski, S.; Tsur, Y.; Haick, H. Utility of
Resistance and Capacitance Response in Sensors Based on Monolayer-
Capped Metal Nanoparticles. J. Phys. Chem. Lett. 2011,2, 19121916.
(44) OHara, M. E.; Clutton-Brock, T. H.; Green, S.; Mayhew, C. A.
Endogenous Volatile Organic Compounds in Breath and Blood of
Healthy Volunteers: Examining Breath Analysis as a Surrogate for
Blood Measurements. J. Breath Res. 2009,3, 027005.
(45) Konvalina, G.; Haick, H. Effect of Humidity on Nanoparticle-
Based Chemiresistors: A Comparison between Synthetic and Real-
World Samples. ACS Appl. Mater. Interfaces 2012,4, 317325.
(46) Liang, X. L.; Sperling, B. A.; Calizo, I.; Cheng, G. J.; Hacker, C.
A.; Zhang, Q.; Obeng, Y.; Yan, K.; Peng, H. L.; Li, Q. L.; Zhu, X. X.;
Yuan, H.; Walker, A. R. H.; Liu, Z. F.; Peng, L. M.; Richter, C. A.
Toward Clean and Crackless Transfer of Graphene. ACS Nano 2011,
5, 91449153.
ACS Applied Materials & Interfaces Research Article
DOI: 10.1021/acsami.6b02682
ACS Appl. Mater. Interfaces 2016, 8, 2174221749
... The devices were fabricated on 100 mm undoped silicon wafers. The alternating aluminum nitride (AlN) and silicon dioxide (SiO2) layers were deposited onto the silicon substrate to form the Bragg reflector [32], as shown in Figure 2a. The Bragg reflector transforms the impedance of the substrate to a near-zero or infinite value to avoid wave energy dissipation into the substrate [33]. ...
... The devices were fabricated on 100 mm undoped silicon wafers. The alternating aluminum nitride (AlN) and silicon dioxide (SiO 2 ) layers were deposited onto the silicon substrate to form the Bragg reflector [32], as shown in Figure 2a. The Bragg reflector transforms the impedance of the substrate to a near-zero or infinite value to avoid wave energy dissipation into the substrate [33]. ...
... This geometrical arrangement ensures good electric field confinement within the active area under the TIE electrode. More details have been reported in the literature by us [32]. ...
Full-text available
Low-abundance biomolecule detection is very crucial in many biological and medical applications. In this paper, we present a novel electrolyte-gated graphene field-effect transistor (EGFET) biosensor consisting of acoustic tweezers to increase the sensitivity. The acoustic tweezers are based on a high-frequency bulk acoustic resonator with thousands of MHz, which has excellent ability to concentrate nanoparticles. The operating principle of the acoustic tweezers to concentrate biomolecules is analyzed and verified by experiments. After the actuation of acoustic tweezers for 10 min, the IgG molecules are accumulated onto the graphene. The sensitivities of the EGFET biosensor with accumulation and without accumulation are compared. As a result, the sensitivity of the graphene-based biosensor is remarkably increased using SMR as the biomolecule concentrator. Since the device has advantages such as miniaturized size, low reagent consumption, high sensitivity, and rapid detection, we expect it to be readily applied to many biological and medical applications.
... Multivariable sensor, which outputs multiple response signals, has attracted tremendous attention as a cost-effective and compact alternative to sensor arrays [18][19][20] . The different responses, being partially or fully independent, arise typically from the same sensing material, which favors not only drift alleviation but also size and cost reduction. ...
Full-text available
Highly efficient gas sensors able to detect and identify hazardous gases are crucial for numerous applications. Array of conventional single-output sensors is currently limited by problems including drift, large size, and high cost. Here, we report a sensor with multiple chemiresistive and potentiometric outputs for discriminative gas detection. Such sensor is applicable to a wide range of semiconducting electrodes and solid electrolytes, which allows to tailor and optimize the sensing pattern by tuning the material combination and conditions. The sensor performance is boosted by equipping a mixed-conducting perovskite electrode with reverse potentiometric polarity. A conceptual sensor with dual sensitive electrodes achieves superior three-dimensional (sub)ppm sensing and discrimination of humidity and seven hazardous gases (2-Ethylhexanol, ethanol, acetone, toluene, ammonia, carbon monoxide, and nitrogen dioxide), and enables accurate and early warning of fire hazards. Our findings offer possibilities to design simple, compact, inexpensive, and highly efficient multivariate gas sensors.
Toxic and flammable gases pose a major safety risk in industrial settings; thus, their portable sensing is desired, which requires sensors with fast response, low‐power consumption, and accurate detection. Herein, a low‐power, multi‐transduction array is presented for the accurate sensing of flammable and toxic gases. Specifically, four different sensors are integrated on a micro‐electro‐mechanical‐systems platform consisting of bridge‐type microheaters. To produce distinct fingerprints for enhanced selectivity, the four sensors operate based on two different transduction mechanisms: chemiresistive and calorimetric sensing. Local, in situ synthesis routes are used to integrate nanostructured materials (ZnO, CuO, and Pt Black) for the sensors on the microheaters. The transient responses of the four sensors are fed to a convolutional neural network for real‐time classification and regression of five different gases (H2, NO2, C2H6O, CO, and NH3). An overall classification accuracy of 97.95%, an average regression error of 14%, and a power consumption of 7 mW per device are obtained. The combination of a versatile low‐power platform, local integration of nanomaterials, different transduction mechanisms, and a real‐time machine learning strategy presented herein helps advance the constant need to simultaneously achieve fast, low‐power, and selective gas sensing of flammable and toxic gases. A low‐power, multi‐transduction nanosensor array is demonstrated by integrating nanostructured materials on bridge‐type microheaters for accurate sensing of flammable and toxic gases. The nanosensor array operates based on chemiresistive and calorimetric mechanisms for enhanced selectivity. By applying transient responses of the nanosensor array to a convolutional neural network, it is possible to accurately identify flammable and toxic gases in real time.
Li dendrite and electrolyte leakage are common causes of Li-ion batteries failure. H 2 , generated by Li dendrite, and electrolyte vapors have been regarded as gas markers of the early safety warning of Li-ion batteries. SnO 2 -based gas sensors, widely used for a variety of applications, are promising for the early safety detection of Li-ion batteries, which are necessary and urgent for developments of Li-ion battery system. However, the traditional SnO 2 sensor with a single signal cannot show intelligent identification of multi-gas. The single dual-mode (direct and alternating current modes) SnO 2 sensor was proved the discrimination of electrolyte vapors and H 2 , released in different status of Li-ion batteries, togethering with principal component analysis (PCA) analysis. This work provides insight for the intelligent technology of single gas sensors.
This paper provides an overview of recent developments in the field of volatile organic compound (VOC) sensors, which are finding uses in healthcare, safety, environmental monitoring, food and agriculture, oil industry, and other fields. It starts by briefly explaining the basics of VOC sensing and reviewing the currently available and quickly progressing VOC sensing approaches. It then discusses the main trends in materials' design with special attention to nanostructuring and nanohybridization. Emerging sensing materials and strategies are highlighted and their involvement in the different types of sensing technologies is discussed, including optical, electrical, and gravimetric sensors. The review also provides detailed discussions about the main limitations of the field and offers potential solutions. The status of the field and suggestions of promising directions for future development are summarized.
Highly selective, sensitive and fast gas sensing has attracted increasing attention in the fields of environmental protection, industrial production, personal safety as well as medical diagnostics. Field effect transistor (FET) sensors have been extensively investigated in gas sensing fields due to their small size, high sensitivity, high reliability and low energy consumption. This comprehensive review aims to discuss the recent advances in FET gas sensors based on materials such as carbon nanotubes, silicon carbide, silicon, metal oxides-, graphene-, transition metal dichalcogenides- and 2-dimensional black phosphorus. We first introduce different types of sensor structures and elaborate the gas-sensing mechanisms. Then, we describe the optimizing strategies for sensing performances, response parameters, FET based dual-mode sensors and FET based logic circuit sensors. Moreover, we present the key advances of the above materials in gas sensing performances. Meanwhile, shortcomings of such materials are also discussed and the future development of this field is proposed in this review.
Volatile organic compounds (VOCs) are pervasive in the environment. Since the early 1980s, substantial work has examined the detection of these materials as they can indicate environmental changes that can affect human health. VOCs and similar compounds present a very specific sensing problem in that they are not reactive and often nonpolar, so it is difficult to find materials that selectively bind or adsorb them. A number of techniques are applied to vapor sensing. High resolution molecular separation approaches such as gas chromatography and mass spectrometry are well-characterized and offer high sensitivity, but are difficult to implement in portable, real-time monitors, whereas approaches such as chemiresistors are promising, but still in development. Gravimetric approaches, in which the mass of an adsorbed vapor is directly measured, have several potential advantages over other techniques but have so far lagged behind other approaches in performance and market penetration. This review aims to offer a comprehensive background on gravimetric sensing including underlying resonators and sensitizers, as well as a picture of applications and commercialization in the field.
A paper based chemiresistor has been fabricated to selectively sense ethanol in human breath. The chemiresistor was composed of a sensing mixture of multiwall-carbon-nano-tubes (MWCNTs), poly (diallyl-dimethyl-ammonium chloride) (PDDA), alcohol dehydrogenase (ADH), and coenzyme (NADH). The aluminum electrode was deposited on the paper surface, followed by drop-casting of the aforementioned sensing mixture. The resistance of the sensors was measured by exposing the same in gas-vapor mixture as well as the sample solution. The surface-modified MWCNTs specifically broke down ethanol present in the gas-vapor mixture or in a solution to generate a quantitative electronic response proportional to the ethanol concentration. Subsequently, the interference of other volatile organic materials was also tested to prove the selectivity and sensitivity of the sensor towards ethanol in the presence of different volatile organic compounds (VOCs). The variation of the resistance during the interaction between sensor and ethanol was also characterized by measuring the surface potential of the channel material under ethanol exposure using Kelvin probe force microscopy (KPFM). The sensor was integrated with a voltage divider circuit, a display, and a microcontroller unit to make a proof-of-concept prototype for the point-of-care (POC) detection of ethanol in human breath.
We studied sol-gel-prepared crystalline (anatase) TiO2 nanopowder coatings, exhibiting simultaneously three different oxygen-sensitive signals under ultraviolet illumination: photoconductivity, photoluminescence of intrinsic defects, and photoluminescence of Sm³⁺ impurity ions. The material was subjected to a flow of O2/N2 gas mixture at normal pressure, where the volume fraction of oxygen was precisely controlled. All three signals responded to changes in oxygen content but also exhibited long-term drifts. The signals were fused by a multivariable model, which related the logarithm of oxygen concentration to a polynomial composed of signal logarithms. Extensive testing by the ordinary least squares optimization with randomly generated O2 concentrations was used for training the model. It was demonstrated that the intrinsic luminescence in combination with one of the remaining two signals suppressed the drifts and significantly improved the precision in predicting the actual oxygen levels. The rms errors were improved by five times in the experiments where the oxygen concentration was randomly varied between 0.21% and 21% during two days. The combination of photoconductivity and Sm³⁺ luminescence resulted in somewhat smaller improvement (2–3 times), because of the mutual dependence of these two signals caused by similarities in the underlying physical mechanisms.
Full-text available
Two-dimensional (2D) layered materials have attracted significant attention for device applications because of their unique structures and outstanding properties. Here, a field-effect transistor (FET) sensor device is fabricated based on 2D phosphorene nanosheets (PNSs). The PNS sensor exhibits an ultrahigh sensitivity to NO 2 in dry air and the sensitivity is dependent on its thickness. A maximum response is observed for 4.8-nm-thick PNS, with a sensitivity up to 190% at 20 parts per billion (p.p.b.) at room temperature. First-principles calculations combined with the statistical thermodynamics modelling predict that the adsorption density is 1/410 15 cm '2 for the 4.8-nm-thick PNS when exposed to 20p.p.b. NO 2 at 300K. Our sensitivity modelling further suggests that the dependence of sensitivity on the PNS thickness is dictated by the band gap for thinner sheets (<10nm) and by the effective thickness on gas adsorption for thicker sheets (>10nm).
Full-text available
Bulk acoustic wave (BAW) resonators are widely applied in filters and gravimetric sensors for physical or biochemical sensing. In this work, a new architecture of BAW resonator is demonstrated, which introduces a pair of reflection layers onto the top of a thin film bulk acoustic resonator (FBAR) device. The new device can be transformed between type I and type II dispersions by varying the thicknesses of the reflection layers. A computational modeling is developed to fully investigate the acoustic waves and the dispersion types of the device theoretically. The novel structure makes it feasible to fabricate both type resonators in one filter, which offers an effective alternative to improve the pass band flatness in the filter. Additionally, this new device exhibits a high quality factor (Q) in the liquid, which opens a possibility for real time measurement in solutions with a superior limitation of detection (LOD) in sensor applications.
Full-text available
A design guideline for one-port aluminum nitride (AlN) Lamb wave resonators (LWRs) working at S0 mode with high performance is reported. A fabricated 252 MHz LWR, with an aperture of 200 μm, 12 fingers, and 1.5 μm thick AlN, is found to have a remarkably high figure of merit (FOM), which exhibits a high ratio of the resistance at parallel resonance (Rp) to the resistance at series resonance (Rs) of 1317 and a corresponding product of the effective coupling coefficient (k2eff) and quality factor (Q) exceeding 52. Consisting of such resonators, a 6-stage ladder filter with a low pass-band insertion loss (IL) of 4.5 dB and steep filter skirts is achieved, offering significant advantage of size savings.
Full-text available
We report on ultrasensitive molecularly-modified silicon nanowire field effect transistor that marries between the lock-and-key and cross-reactive sensing worlds for the diagnosis of (gastric) cancer from exhaled volatolome. The sensor is able to selectively detect VOCs that are linked with gastric cancer conditions in exhaled breath and to discriminate them from environmental VOCs that would exist in exhaled breath samples but do not relate to the gastric cancer per se. Using breath samples collected from real patients with gastric cancer and from volunteers that have no cancer, blind analysis validated the ability of the reported sensor to discriminate between gastric cancer and control conditions, irrespective of important confounding factors such as tobacco consumption and gender with >85% accuracy. The reported sensing approach paves the way for utilizing the power of silicon nanowires in simple, inexpensive, portable, and non-invasive diagnosis of both cancer and other diseases conditions.
Full-text available
One-dimensional semiconductor nanostructures are unique sensing materials for the fabrication of gas sensors. In this article, gas sensors based on semiconducting nanowire field-effect transistors (FETs) are comprehensively reviewed. Individual nanowires or nanowire network films are usually used as the active detecting channels. In these sensors, a third electrode, which serves as the gate, is used to tune the carrier concentration of the nanowires to realize better sensing performance, including sensitivity, selectivity and response time, etc. The FET parameters can be modulated by the presence of the target gases and their change relate closely to the type and concentration of the gas molecules. In addition, extra controls such as metal decoration, local heating and light irradiation can be combined with the gate electrode to tune the nanowire channel and realize more effective gas sensing. With the help of micro-fabrication techniques, these sensors can be integrated into smart systems. Finally, some challenges for the future investigation and application of nanowire field-effect gas sensors are discussed.
Full-text available
The detection of volatile organic compounds (VOCs) at room temperature with rapid response and recovery is important for early explosive alarm. Herein, we demonstrate significantly enhanced VOC-sensing properties of quartz crystal microbalance (QCM) coated with monolayer graphene film. The monolayer graphene was firstly synthesized by the chemical vapour deposition (CVD) method on Cu foil and then transferred to the gold electrode of the QCM for VOC-sensing application. The gas-sensing properties of the graphene-coated QCM sensor were examined at room temperature for various concentrations of numerous VOCs, including butanol, isopropanol, acetone, and ethanol. The results revealed that the graphene-coated QCM sensor exhibits the best performance with ethanol gas. The gas-sensing mechanism of the graphene-coated QCM sensor was attributed to the adsorption and desorption of VOC molecules on the defect sites of graphene sheet.
This Review presents a concise, but not exhaustive, didactic overview of some of the main concepts and approaches related to "volatolomics"-an emerging frontier for fast, risk-free, and potentially inexpensive diagnostics. It attempts to review the source and characteristics of volatolomics through the so-called volatile organic compounds (VOCs) emanating from cells and their microenvironment. It also reviews the existence of VOCs in several bodily fluids, including the cellular environment, blood, breath, skin, feces, urine, and saliva. Finally, the usefulness of volatolomics for diagnosis from a single bodily fluid, as well as ways to improve these diagnostic aspects by "hybrid" approaches that combine VOC profiles collected from two or more bodily fluids, will be discussed. The perspectives of this approach in developing the field of diagnostics to a new level are highlighted. © 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.
A general, overarching theme in nanotechnology is the integration of multiple disparate fields to realize novel or expanded functionalities. Here, we present a graphene enabled, integrated opto-electro-mechanical device and demonstrate its utility for biomolecular sensing. We experimentally achieve an ultra-wide linear dynamic sensing range of 5 orders of magnitude of protein concentration, an improvement over state-of-the-art single mode nanosensors by approximately 2-3 orders of magnitude, while retaining a sub-picomolar lowest detection limit. Moreover, the ability to monitor and characterize adsorption events in the full opto-electro-mechanical space allows for the extraction of key intrinsic parameters of adsorbates, and has the potential to extend the capabilities of nanosensors beyond the traditional binary-valued test for a single type of molecule. This could have significant implications for molecular detection applications at variable concentrations, such as early disease detection in biomedical diagnostics.
We investigate the utility of resistance and capacitance responses, as derived by impedance spectroscopy, in well-controlled and real-world applications of monolayer-capped metal nanoparticle (MCNP) sensors. Exposure of the MCNP films to well-controlled analytes showed stable sensing responses and low baseline drift of the pertinent capacitance signals, when compared with equivalent resistance signals. In contrast, exposure of the MCNP films to breath of chronic kidney disease patients under dialysis, as a representative example to real-world multicomponent mixtures, showed low baseline drift but relatively scattered signals when compared with the equivalent resistance response. We ascribe these discrepancies to the level and fluctuating concentration of water molecules in the real-world samples.