HongYoonki1
YunJonghyun1
HanDong Jin1
LeeSung-Tae2
YunWooSung3
-
(School of Electrical and Electronics Engineering, Pusan National University 46241,
Busan, Republic of Korea)
-
(School of Electronic and Electrical Engineering, Hongik University 04066, Seoul, Republic
of Korea)
-
(School of Electronic and Electrical Engineering, Kyungpook National University 41566,
Daegu, Republic of Korea)
Copyright © The Institute of Electronics and Information Engineers(IEIE)
Index Terms
Humidity sensing, Si field-effect transistor-type sensor, tungsten trioxide (WO3), pulse measurement, transduction mechanism
I. Introduction
Humans are constantly developing technologies to pursue smart living. Currently, the
Internet of Things (IoT) is among the most promising technologies for making our lives
smarter. IoT refers to the networked interconnection of embedded devices that allows
them to interact with each other, services, and people on a global scale [1]. Such level of connectivity can increase reliability, sustainability, and efficiency
through the integration of every device for interaction via embedded systems to improve
access to information [1,2]. Conceptual examples of IoT-based smart technologies are as follows. In houses, gas
leakages generated in the kitchen are detected by smart gas sensors, and alerts are
sent in the form of mobile messages via a wireless network. Simultaneously, self-inspection
is performed to remove the leakage using a smart automation system based on IoT technology
[3,4]. Depending on occupancy, the lights in a smart home are automatically turned on and
off, and heating or air-conditioning systems operate to adjust the indoor temperature
[5]. Moreover, through the monitoring of the home environment, power efficiency can be
improved, and the well-being of residents can be enhanced [6]. In factories, various types of smart sensors, such as temperature, pressure, and
vibration sensors, can contribute to automated process control to satisfy the quality
and yield requirements of products [7,8]. Many concepts and research results on IoT-based smart technologies have been reported
[3,4,5,6,7,8,9,10,11]. Furthermore, a considerable amount of cash flow is expected through the global market
in the coming years for IoT [12]. For the development and realization of IoT technology, the development of sensor
technology, which is one of the components of IoT system architecture and where the
workflow of IoT systems begins, is critical [13]. These sensors detect external changes and transduce them into electrical signals.
Subsequently, the electrical signals are conditioned by interface circuits, such as
a noise filter and an amplifier, and digitally processed and transmitted to the data
collection center or other devices through a wireless network [14]. Simultaneously, humans can monitor the transmitted data in real time using a user
interface-based mobile application [14].
Sensor technology requires high sensitivity and high resolution. To improve sensitivity
and resolution, extensive research on the morphological transformation or composition
tuning of sensing materials has been conducted for a long time [15,16,17]. Recently, a novel electrical operation scheme was proposed to achieve high resolution
without any transformation or tuning of the sensing materials which is complex and
difficult to apply in device fabrication [18]. However, in the recent development of IoT sensor technology, reliability and high
operating speed have gained increased importance over the sensing abilities described
above [19]. Moreover, IoT sensors necessitate miniaturization and low power consumption for
application to portable devices. CMOS technology can satisfy these requirements for
IoT sensors [20]. In addition, CMOS technology is compatible with interface circuits and can reduce
fabrication costs [20].
Humidity control technologies play an important role in human life. Accurate humidity
monitoring and control is critical for maintaining a pleasant home environment. In
factories and farms, the maintenance of a constant humidity level significantly affects
the product yield. Accordingly, several studies on humidity sensors have been performed,
but most of them have focused on resistive- [21,22] and capacitive-type [23,24] humidity sensors. Although resistive and capacitive sensors are inexpensive and easily
fabricated, they have the disadvantages of a large size [21], output signal drift [22], and hysteresis [24]. Si field-effect transistor (FET)-type sensors are expected to be the most suitable
candidates for overcoming these disadvantages owing to their scalability via the use
of CMOS technology and compatibility with integrated circuits (ICs) and electrical
control schemes which improve the sensing performance [18].
This study presents the sensing characteristics of a Si FET-type humidity sensor with
a tungsten trioxide (WO${}_{3}$) sensing layer prepared using Si CMOS fabrication
technology. The key process steps of the humidity sensor are explained, and the experimental
setups for humidity generation and electrical measurements with pulse bias are described.
The humidity-sensing characteristics are measured and analyzed based on the chemical
reaction mechanism between WO${}_{3}$ and H${}_{2}$O. Subsequently, the chemical-to-electrical
transduction that occurs in the FET part of the sensor during humidity sensing is
explained comprehensively using schematic energy band diagrams. Finally, the transient
drain current ($I_{\rm D}$) as a function of relative humidity (RH) is discussed.
II. Device Fabrication
The humidity sensor is prepared using the conventional Si FET technology. An ${n}$-type
Si (100) wafer is used as the substrate for the sensor because ${p}$-channel MOSFET
is superior to ${n}$-channel MOSFET in terms of low-frequency noise [25]. Fig. 1(a) displays the top-view SEM image of the humidity sensor. The humidity sensor comprises
two parts: sensing (top of Fig. 1(a)) and FET (bottom of Fig. 1(a)) parts. The sensing part comprises the control-gate (CG), the floating-gate (FG),
and the WO${}_{3}$ film functioning as a humidity-sensing layer. The CG and FG are
formed in an interdigitated structure and face each other with a gap of 500 nm. The
WO${}_{3}$ sensing layer partially covers them by filling the gap. The FET part comprises
the FG, the source, the drain, and the Si active layer. Fig. 1(b) depicts a three-dimensional structure of the humidity sensor, viewed from above at
an oblique angle. Figs. 1(c) and 1(d) present schematic cross-sectional views of the
sensor along the FET channel width and length directions, respectively. The key fabrication
process for the humidity sensor is described below. To define the channel region,
a stack of pad oxide and silicon nitride is patterned on a bare Si wafer and a local
field oxide is thermally grown. Subsequently, the pad oxide and silicon nitride are
removed and a 10 nm thick gate oxide is grown back on the Si active. Thereafter, a
350 nm thick $n ^{+}$ poly-Si is deposited and patterned to form the FG. The source
and drain (S/D) regions are defined by the self-aligned arsenic ion implantation.
The wafer is covered with an oxide-nitride-oxide (O/N/O, 10 nm/20 nm/10 nm) stack
to passivate the FET part of the sensor. The nitride layer of the O/N/O stack is known
to prevent impurities from penetrating into the channel region [26]. The metal contact areas in the S/D regions are defined, followed by the formation
of a Cr/Au metal stack for the source, drain, and CG electrodes. At the end, a 30
nm thick semiconducting WO${}_{3}$ sensing layer is formed via radio frequency (RF)
magnetron sputtering. The chamber pressure is 5 mTorr with Ar flow and the RF power
is 230 W during sputtering.
Fig. 1. (a) Top-view SEM image of the fabricated Si FET-type humidity sensor having
a WO${_{3}}$ sensing layer. The blue-shaded region indicates the WO${}_{3}$ sensing
layer. (b) Oblique view of the humidity sensor. Cross-sectional views of the sensor
cut along the red dotted lines in (a): (c) A-A' and (d) B-B'.
III. Experimental Setup
Fig. 2 shows a schematic diagram of the humidity-sensing measurement system, which is composed
of two parts: a humidity generator and a test chamber. Three dry N${}_{2}$ gas lines
controlled via a mass flow controller (MFC) are used. The first gas line is connected
to the mixing container via a bubbler for N${}_{2}$ gas to carry water vapor. The
sidewall of the bubbler is wrapped with an isolated flow line from the heating circulator
through which the heated water flows. This heated water is used to warm the water
inside the bubbler to facilitate the generation of water vapor. The second gas line
is directly connected to the mixing container, where N${}_{2}$ gas from the second
gas line is intermixed with that from the first gas line to adjust the RH. If the
intermixed N${}_{2}$ gas with a certain RH flows toward the test chamber for humidity
sensing, it passes through a commercialized RH reader to recognize the RH before reaching
the test chamber. The third gas line connected directly to the RH reader is used to
recover the sensor.
Fig. 2. Schematic diagram of the measurement system for humidity sensing.
IV. Results and Discussion
1. Pulse Operation Schemes for Humidity Sensing
The humidity-sensing properties of the sensor are demonstrated using pulse operation
schemes. Figs. 3(a) and 3(b) show the pulse signal waveforms for operating the sensor
to obtain the pulsed $I_{\rm D}$-$V_{\rm CG}$ (PIV) characteristics and the transient
$I_{\rm D}$ behaviors, respectively. The CG bias ($V_{\rm CG}$) is swept ranging from
2 to $-2$ V for PIV measurement, while a fixed value of $-0.2$ V is applied to the
CG read bias ($V_{\rm rCG}$) for transient $I_{\rm D}$ measurement. For both PIV and
transient $I_{\rm D}$ measurements, the base voltages ($V_{\rm base}$s) of the CG
and drain pulses are fixed at 0 V, and the drain read bias is $-0.1$ V. A single pulse
signal is applied to the CG terminal with the high and low levels corresponding to
$V_{\rm CG}$ (or $V_{\rm rCG}$) and $V_{\rm base}$, respectively. The durations of
the high and low levels are defined as $t_{\rm on}$ and $t_{\rm off}$, respectively,
which are set to 30 $\mu$s and 100 ms in this study. The drain pulse signal is synchronized
with the CG pulse signal, following the same timing scheme.
The pulse operation scheme in Fig. 3(b) has been previously proposed and its effect has been verified [27]. According to [27], through the application of a pulse operation scheme, the $I_{\rm D}$ drift of a
Si FET-type sensor caused by the DC bias is significantly reduced, which facilitates
the obtaining of stable sensing characteristics.
Fig. 4 displays the transfer ($I_{\rm D}$-$V_{\rm CG}$) characteristics of the fabricated
FET-type humidity sensor. The channel width and length of the FET are both 2 $\mu$m.
The $I_{\rm D}$-$V_{\rm CG}$ curves are measured at an RH of 3.4% and room temperature
using the DC and PIV methods. The two logarithmic $I_{\rm D}$-$V_{\rm CG}$ curves
obtained by the DC and PIV methods coincide well with each other at $|I_{\rm D}|>
1$ nA. This implies that the $t_{\rm on}$ of 30 $\mu$s is appropriate to operate normally
the FET of the humidity sensor.
Fig. 3. Pulse signal waveforms for operating the humidity sensor to measure (a) PIV
and (b) transient $I_{\rm D}$ behavior.
Fig. 4. DC and PIV characteristics of the humidity sensor.
2. Chemical Reaction and Chemical-to-Electrical Transduction Mechanisms
Fig. 5(a) shows the shift in the $I_{\rm D}$-$V_{\rm CG}$ curves of the sensor obtained using
the PIV method with varying RH inside the test chamber. Each $I_{\rm D}$-$V_{\rm CG}$
curve is measured 300 s after injecting N${}_{2}$ gas with a certain RH into the test
chamber to saturate the ambient RH inside the test chamber. Fig. 5(b) illustrates the variation of threshold voltage ($V_{\rm th}$) with respect to RH.
$V_{\rm th}$ is extracted from each $I_{\rm D}$-$V_{\rm CG}$ curve in Fig. 5(a) using the linear extrapolation method. The increase in RH from 3.4% to 80.3% leads
to the change in $V_{\rm th}$ ($\Delta V_{\rm th}$) by $-134$ mV.
Fig. 5. (a) $I_{\rm D}$-$V_{\rm CG}$ characteristics of the humidity sensor as a function
of relative humidity ranging from 3.4% to 80.3% measured by using the PIV method.
(b) Variation of $V_{\rm th}$ with RH.
Fig. 6. Schematic diagram of adsorption of water molecules on the WO${_{3}}$ sensing
layer.
Here, we explain the reason for the $\Delta$$V_{\rm th}$ being negative with the increase
of RH. As shown in Fig. 6, adsorption of water molecules onto the surface of the WO${}_{3}$ humidity-sensing
layer is divided into two steps. At a lower RH, water molecules dissociate into hydroxide
(OH${}^{-}$) and hydrogen (H${}^{+}$) ions when exposed to the WO${}_{3}$ layer [28].
The OH${}^{-}$ ions interact with the tungsten cations on the WO${}_{3}$ surface.
Whereas, the H${}^{+}$ ions interact with the lattice oxygens or the oxygen ions existing
on the surface of the WO${}_{3}$ layer, forming the hydroxyl groups that are chemically
bonded to the tungsten cations [28]. Finally, both OH${}^{-}$ and H${}^{+}$ ions are chemically adsorbed (chemisorbed)
onto the surface of the WO${}_{3}$ layer in the form of a hydroxyl group ($-$OH).
Note that the O-H chemical bond in this hydroxyl group is a polar covalent bond with
the negatively polarized oxygen atom and the positively polarized hydrogen atom.
Meanwhile, we must recall a fact that has already been demonstrated experimentally
in our previous study [18]. In [18], it was revealed that the sensor response changes depending on the pre-bias ($V_{\rm
pre}$) applied to the CG. When $V_{\rm pre}$ is negative, the electrons of the ZnO
sensing layer are accumulated near the interface between the ZnO layer and the O/N/O
layer, and the chemical reaction between NO${}_{2}$ (target gas) and ZnO mainly occurs
near the interface, which is relatively close to the FET channel. Therefore, the sensor
response increases compared to that in the case of $V_{\rm pre} = 0$ V. However, when
$V_{\rm pre}$ is positive, the ZnO layer near the interface is depleted. Consequently,
the chemical reaction mostly happens in the bulk region of the ZnO layer. In this
case, the sensor response eventually decreases compared to that in the case of $V_{\rm
pre} = 0$ V because the amount of chemical reaction at the interface between the ZnO
layer and the O/N/O layer is reduced due to the depletion of the ZnO layer. In summary,
the chemical reaction occurring in the sensing layer near the interface between the
sensing layer and the O/N/O layer dominantly affects the FET of the sensor. Therefore,
in this study, it is assumed that the chemical reaction between H${}_{2}$O and WO${}_{3}$
occurs predominantly near the interface between the WO${}_{3}$ layer and the O/N/O
layer.
At the interface between the WO${}_{3}$ layer and the O/N/O layer near the FG, the
positively polarized hydrogen atoms of the hydroxyl groups generated by H${}_{2}$O
chemisorption on the WO${}_{3}$ layer induce a negative sheet charge at the interface
of the FG in contact with the O/N/O layer. This negative sheet charge consists of
the electrons, which are the majority carriers of the $n$-type poly-Si FG. Simultaneously,
a positive sheet charge, which consists of the depletion charge of the $n$-type poly-Si
FG, is induced at the interface of the FG in contact with the gate oxide. This positive
sheet charge reduces the hole concentration of the $p$-type FET channel of the sensor.
In summary, chemisorption of water molecules onto the surface of the WO${}_{3}$ layer
shifts $V_{\rm th}$ of the $p$-type FET sensor in the negative direction, causing
$|I_{\rm D}|$ to decrease. At a higher RH, additional water molecules are physically
adsorbed (physisorbed) onto the hydroxyl groups formed by chemisorption of water molecules
[28]. The first physisorbed layer of water molecules are formed by the hydrogen bond between
the hydrogen atoms of the hydroxyl groups already bonded to the WO${}_{3}$ layer via
chemisorption of water molecules at a lower RH, and the oxygen atoms of the additional
water molecules [28]. From the second physisorbed layer, hydrogen bonds occur between the hydrogen atoms
of the previously physisorbed water molecules and the oxygen atoms of added water
molecules [28]. Similar to the chemisorbed layer, the molecular arrangement of these multiple physisorbed
layers can be regarded as a dipole with the direction of the dipole moment being toward
the FET channel of the sensor, thereby reducing the hole concentration of the channel.
Therefore, $V_{\rm th}$ and $|I_{\rm D}|$ of the sensor decrease as RH increases,
regardless of whether water molecules are chemisorbed or physisorbed onto the surface
of the WO${}_{3}$ layer.
3. Change of Energy Band Structure during Humidity Sensing
Next, we examine the change in energy band structure of the sensor when water molecules
are adsorbed onto the WO${}_{3}$ sensing layer. Fig. 7(a) shows the schematic energy band diagram of the sensor under flat-band condition.
Here, $\Phi_{\rm CG}$, $\Phi_{\rm WO3}$, $\Phi_{\rm FG}$, and $\Phi_{\rm sub}$ represent
the work functions of CG, WO${}_{3}$ sensing layer, FG, and Si substrate, respectively.
$\chi_{\rm WO3}$, $\chi_{\rm FG}$, and $\chi_{\rm sub}$ stand for the electron affinities
of WO${}_{3}$ sensing layer, FG, and Si substrate, respectively. Because the Au of
the CG occupies most of the direct contact area with the WO${}_{3}$ layer, the $\Phi_{\rm
CG}$ is regarded as the work function of Au (5.1 eV). The electron concentrations
of the FG and the substrate are ${\sim} 1\times 10^{21}$ and ${\sim} 1\times10 ^{16}$
cm${}^{-3}$, respectively, thus the $\Phi_{\rm FG}$ and the $\Phi_{\rm sub}$ are ${\sim}
4.05$ and ${\sim} 4.26$ eV, respectively. WO${}_{3}$ is a well-known $n$-type semiconductor
[29] with an $\chi_{\rm WO3}$ value of ${\sim} 3.3$ eV [30]. The $\chi_{\rm FG}$ and the $\chi_{\rm sub}$ are both 4.05 eV. Note that the FG
is in a fresh state, as shown in Fig. 7(a). This implies that electrons or holes are not stored in the FG. Fig. 7(b) shows the schematic energy band diagram under equilibrium condition where all the
electrodes in the sensor are grounded (0 V). As the $|I_{\rm D}|$ at $V_{\rm CG} =
0$ V is ${\sim} 0.2$ $\mu$A in the $I_{\rm D}$-$V_{\rm CG}$ curve in Fig. 4, it can be inferred that the FET of the sensor is in weak inversion. In addition,
considering the band structure shown in Fig. 7(a), it is evident that electrons are initially stored in the FG, as shown in Fig. 7(b). This phenomenon is similar to the programmed state of a flash memory device with
a poly-Si floating-gate; we have confirmed the program and erase characteristics of
the sensor in the previous study [18]. Fig. 7(c) illustrates the change of the energy band diagram before and after humidity sensing
at $V_{\rm CG}$ = $V_{\rm rCG} = -0.2$ V. As described in Fig. 6, the chemical reaction between water molecules and the WO${}_{3}$ sensing layer results
in the formation of hydroxyl groups by chemisorption or dipoles by physisorption at
the interface between the WO${}_{3}$ layer and the O/N/O layer. These hydroxyl groups
and dipoles both exhibit the dipole moments in the same direction from the interface
to the FET channel. This induces a negative sheet charge at one interface of the FG
in contact with the O/N/O layer and a positive sheet charge at the opposite interface
of the FG in contact with the gate oxide. The localization of these charges results
in a change in the electric field, which consequently decreases $|I_{\rm D}|$ of the
FET.
Fig. 7. Schematic band diagrams at (a) flat-band condition, (b) equilibrium condition
($V_{\rm CG} = 0$ V) and (c) the read bias ($V_{\rm CG} = V_{\rm rCG} = -0.2$ V).
In (c), the gray and the black lines stand for the states before and after humidity
sensing, respectively.
4. Sensor Response as a Function of Relative Humidity
Fig. 8(a) shows the transient $I_{\rm D}$ behaviors as a parameter of RH using the pulse measurement
method presented in Fig. 3(b). Similar to the experimental method used when obtaining the results shown in Fig. 5, N${}_{2}$ gas with a certain RH is injected into the test chamber for 300 s to saturate
the humidity sensing before the pulse biases are applied to the sensor for 10 s to
measure every transient $I_{\rm D}$ curve. At this time, the $V_{\rm rCG}$ and the
$V_{\rm rDS}$ are $-0.2$ V and $-0.1$ V, respectively. $I_{\rm D}$ decreases with
the increase in RH and there is no drift in any of the transient $I_{\rm D}$ curves
owing to the pulse measurement method. In the previous studies [31,32], drifts in output signals have been observed due to the application of DC bias to
the sensors, which are undesirable because they degrade the accuracy of sensing and
therefore induce errors in sensor operations. However, this study confirms that the
pulse bias to the sensor can suppress $I_{\rm D}$ drift. This method also offers the
advantage of reduced power consumption during operation. Fig. 8(b) shows the sensor response (${R}$) versus RH calculated from the $I_{\rm D}$s in Fig. 8(a), where ${R}$ is defined as the rate of change in $I_{\rm D}$ divided by $I_{\rm D}$
at an RH of 3.4%. The $R$s are 22.3% and 45.5% at RHs of 54.9% and 80.3%, respectively.
As RH increases from 3.4% to 54.9%, $R$ increases and then appears to saturate. However,
it linearly increases above an RH of 54.9%. It is speculated that a change in the
adsorption process of water molecules on the WO${}_{3}$ layer from chemisorption to
physisorption occurs at this inflection point (RH $= {\sim}$54.9%), which needs to
be proven through further research.
Fig. 8. (a) Transient $I_{\rm D}$ behaviors of the humidity sensor as a parameter
of relative humidity ranging from 3.4% to 80.3%. Each curve is obtained by adopting
the pulse scheme explained in Fig. 3(b) for 10 s. (b) $R$ versus relative humidity.
Fig. 9. Dynamic response of the humidity sensor monitored by changing relative humidity
of the injected N${_{2}}$ carrier gas. The relative humidity of the dry N${}_{2}$
gas for recovery is 3.4%.
The dynamic response of the sensor is also demonstrated in Fig. 9 via the injection of N${}_{2}$ gas with a certain RH and the reference N${}_{2}$
gas with an RH of 3.4% alternately into the test chamber for 180 and 300 s, respectively.
To measure the transient $I_{\rm D}$ curve for the dynamic response, N${}_{2}$ gases
with RHs of 11.5%, 28.2%, 54.9%, and 68.5% are sequentially injected, and the reference
N${}_{2}$ gas is injected for sensor recovery. The $V_{\rm rCG}$ and the $V_{\rm rDS}$
are $-0.2$ V and $-0.1$ V, respectively. Similar to the results shown in Fig. 8(a), $|I_{\rm D}|$ decreases with the increase in RH of the injected N${}_{2}$ gas. The
response and recovery speeds of the sensor are also obtained by defining response
time as a time duration for which $|I_{\rm D}|$ decreases by 90% of the change in
$I_{\rm D}$ during response, and recovery time as a time duration for which $|I_{\rm
D}|$ increases by 90% of the change in $I_{\rm D}$ during recovery. The response and
recovery times of the sensor are 97 and 190 s, respectively, at an RH of 68.5%.
V. Conclusions
In this study, a Si FET-type humidity sensor was fabricated using a WO${}_{3}$ thin
film as the sensing material. The sensor comprised a sensing part that functioned
as a gate as well as an RH sensor and an FET part, wherein an electrical change was
induced by humidity sensing. The humidity-sensing characteristics were obtained by
measuring the $I_{\rm D}$-$V_{\rm CG}$ curves and transient $I_{\rm D}$s. In contrast
to the general methods, stable performance was achieved without drift of $I_{\rm D}$
by using the pulse measurement method. In addition, the signal transduction of the
sensor generated by the reaction between the WO${}_{3}$ sensing material and the water
molecules was analyzed using energy band diagrams. The FET-type humidity sensor with
Si CMOS technology and pulse-driving method proposed in this study is expected to
be the most promising candidate for IoT-based smart humidity sensor.
ACKNOWLEDGMENTS
This work was supported by a New Faculty Research Grant of Pusan National University,
2024 and in part by NRF funded by the Korean government (RS-2024-00405200).
References
S. C. Mukhopadhyay and N. K. Suryadevara, ``Internet of Things: Challenges and opportunities,''
Internet of Things: Challenges and Opportunities, 1st ed., S. C. Mukhopadhyay, Ed.,
Springer International Publishing, Cham, Switzerland, pp. 1-17, 2014.
F. Xia, L. T. Yang, L. Wang, and A. Vinel, ``Internet of Things,'' International Journal
of Communication Systems, vol. 25, no. 9, pp. 1101-1102, Septermber 2012.
J. M. Chatterjee, R. Kumar, M. Khari, D. T. Hung, and D.-N. Le, ``Internet of Things
based system for smart kitchen,'' International Journal of Engineering and Manufacturing,
vol. 8, no. 4, pp. 29-39, July 2018.
R. Blasco, A. Marco, R. Casas, D. Cirujano, and R. Picking, ``A smart kitchen for
ambient assisted living,'' Sensors, vol. 14, no. 1, pp. 1629-1653, January 2014.
R. Kadam, P. Mahamuni, and Y. Parikh, ``Smart home system,'' International Journal
of Innovative Research in Advanced Engineering, vol. 2, no. 1, pp. 81-86, January
2015.
G. Bedi, G. K. Venayagamoorthy, and R. Singh, ``Internet of Things (IoT) sensors for
smart home electric energy usage management,'' Proc. of 2016 IEEE International Conference
on Information and Automation for Sustainability (ICIAfS), Galle, Sri Lanka, pp. 1-6,
2016.
T. Kalsoom, N. Ramzan, S. Ahmed, and M. Ur-Rehman, ``Advances in sensor technologies
in the era of smart factory and industry 4.0,'' Sensors, vol. 20, no. 23, pp. 6783-6804,
November 2020.
M. Pech, J. Vrchota, and J. Bedná˘r, ``Predictive maintenance and intelligent sensors
in smart factory: Review,'' Sensors, vol. 21, no. 4, pp. 1470-1509, February 2021.
J. Gao, J. Wang, L. Zhang, Q. Yu, Y. Huang, and Y. Shen, ``Magnetic signature analysis
for smart security system based on TMR magnetic sensor array,'' IEEE Sensors Journal,
vol. 19, no. 8, pp. 3149-3155, April 2019.
C. Sisavath, and L. Yu, ``Design and implementation of security system for smart home
based on IOT technology,'' Procedia Computer Science, vol. 183, pp. 4-13, April 2021.
J. Zhang, Y. Wang, S. Li, and S. Shi, ``An architecture for IoT-enabled smart transportation
security system: A geospatial approach,'' IEEE Internet of Things Journal, vol. 8,
no. 8, pp. 6205-6213, November 2020.
S. Al-Sarawi, M. Anbar, R. Abdullah, and A. B. Al Hawari, ``Internet of Things market
analysis forecasts, 2020–2030,'' Proc. of 2020 Fourth World Conference on Smart Trends
in Systems, Security and Sustainability (WorldS4), London, UK, pp. 449-453, 2020.
N. Paudel and R. C. Neupane, ``A general architecture for a real-time monitoring system
based on the internet of things,'' Internet Things, vol. 14, 100367, June 2021.
W. Gao, S. Emaminejad, H. Y. Y. Nyein, S. Challa, K. Chen, A. Peck, H. M. Fahad, H.
Ota, H. Shiraki, D. Kiriya, D.-H. Lien, G. A. Brooks, R. W. Davis, and A. Javey, ``Fully
integrated wearable sensor arrays for multiplexed in situ perspiration analysis,''
Nature, vol. 529, no. 7587, pp. 509-514, January 2016.
X. Wan, J. Wang, L. Zhu, and J. Tang, ``Gas sensing properties of Cu2O and its particle
size and morphology-dependent gas-detection sensitivity,'' Journal of Materials Chemistry
A, vol. 2, no. 33, pp. 13641-13647, June 2014.
S. Kang, J. Lee, S. Lee, S.G. Kim, J.-K. Kim, H. Algadi, S. Al-Sayari, D.-E. Kim,
D.E. Kim, and T. Lee, ``Highly sensitive pressure sensor based on bioinspired porous
structure for real-time tactile sensing,'' Advanced Electronic Materials, vol. 2,
no. 12, 1600356, November 2016.
C.-H. Hsia, A. Wuttig, and H. Yang, ``An accessible approach to preparing water-soluble
Mn2-doped (CdSSe)ZnS (core)shell nanocrystals for ratiometric temperature sensing,''
ACS Nano, vol. 5, no. 12, pp. 9511-9522, October 2011.
J. Shin, Y. Hong, M. Wu, Y. Jang, J. S. Kim, B.-G. Park, C. S. Hwang, and J.-H. Lee,
``Highly improved response and recovery characteristics of Si FET-type gas sensor
using pre-bias,'' Proc. of 2016 IEEE International Electron Device Meeting (IEDM),
San Francisco, CA, USA, pp. 18.1.1-18.1.4, 2016.
R. Krishnamurthi, A. Kumar, D. Gopinathan, A. Nayyar, and B. Qureshi, ``An overview
of IoT sensor data processing, fusion, and analysis techniques,'' Sensors, vol. 20,
no. 21, pp. 6076-6098, October 2020.
J. M. Nassar, G. A. T. Sevilla, S. J. Velling, M. D. Cordero, and M. M. Hussain, ``A
CMOS-compatible large-scale monolithic integration of heterogeneous multi-sensors
on flexible silicon for IoT applications,'' Proc. of 2016 IEEE International Electron
Device Meeting (IEDM), San Francisco, CA, USA, pp. 18.6.1-18.6.4, 2016.
A. Sharma, Y. Kumar, K. Mazumder, A. K. Rana, and P. M. Shirage, ``Controlled Zn$_{\rm
1-x}$Ni$_{\rm x}$O nanostructures for an excellent humidity sensor and a plausible
sensing mechanism,'' New Journal of Chemistry, vol. 42, no. 11, pp. 8445-8457, April
2018.
S. Capone, A. Forleo, L. Francioso, R. Rella, P. Siciliano, J. Spadavecchia, D. S.
Presicce, and A. M. Taurino, ``Solid state gas sensors: State of the art and future
activities,'' Journal of Optoelectronics and Advanced Materials, vol. 5, no. 5, pp.
1335-1348, August 2003.
Y. Zhao, B. Yang, and J. Liu, ``Effect of interdigital electrode gap on the performance
of SnO$_2$-modified MoS$_2$ capacitive humidity sensor,'' Sensors and Actuators B:
Chemical, vol. 271, pp. 256-263, October 2018.
T. Fei, H. Zhao, K. Jiang, X. Zhou, and T. Zhang, ``Polymeric humidity sensors with
nonlinear response: Properties and mechanism investigation,'' Journal of Applied Polymer
Science, vol. 130, no. 3, pp. 2056-2061, May 2013.
M. Aoki, Y. Sakai, and T. Masuhara, ``Low 1/f noise design of Hi-CMOS devices,'' IEEE
Transactions on Electron Devices, vol. 29, no. 2, pp. 296-299, February 1982.
T. Ito, T. Nakamura, and H. Ishikawa, ``Advantages of thermal nitride and nitroxide
gate films in VLSI process,'' IEEE Jorunal of Solid-State Circuits, vol. 17, no. 2,
pp. 128-132, April 1982.
J. Shin, Y. Hong, M. Wu, J.-H. Bae, H.-I. Kwon, B.-G. Park, and J.-H. Lee, ``An accurate
and stable humidity sensing characteristic of Si FET-type humidity sensor with MoS$_2$
as a sensing layer by pulse measurement,'' Sensors and Actuators B: Chemical, vol.
258, pp. 574-579, April 2018.
P. Shankar, and J. B. B. Rayappan, ``Gas sensing mechanism of metal oxides: The role
of ambient atmosphere, type of semiconductor and gases - A review,'' Science Letters
Journal, vol. 4, no. 4, pp. 126, 2015.
F. E. Annanouch, S. Roso, Z. Haddi, S. Vallejos, P. Umek, C. Bittencourt, C. Blackman,
T. Vilic, and E. Llobet, ``$p$-Type PdO nanoparticles supported on $n$-type WO$_3$
nanoneedles for hydrogen sensing,'' Thin Solid Films, vol. 618, pp. 238-245, November
2016.
C. W. Walter, C. F. Hertzler, P. Devynck, G. P. Smith, and J. R. Peterson, ``Photodetachment
of WO$_3^-$: The electron affinity of WO3,'' The Journal of Chemical Physics, vol.
95, no. 2, pp. 824-827, July 1991.
C. Wang, Y. Zhang, X. Sun, Y. Sun, F. Liu, X. Yan, C. Wang, P. Sun, and G. Lu, ``Preparation
of Pd/PdO loaded WO$_3$ microspheres for H2S detection,'' Sensors and Actuators B:
Chemical, vol. 321, 128629, October 2020.
T. Samerjai, N. Tamaekong, C. Liewhiran, A. Wisitsoraat, A. Tuantranont, and S. Phanichphant,
``Selectivity towards H$_2$ gas by flame-made Pt-loaded WO$_3$ sensing films,'' Sensors
and Actuators B: Chemical, vol. 157, pp. 290-297, April 2011.
Yoonki Hong received his B.S. and Ph.D. degrees in electrical engineering and computer
science from Seoul National University (SNU), in 2013 and 2019, respectively. In 2019,
he joined Samsung Electronics, Hwaseong, Gyeonggi-do, Korea, where he has been involved
in the development of conventional and next-generation DRAM devices and fabrication
processes. Since March 2024, he has been an Assistant Professor in the School of Electrical
and Electronics Engineering at Pusan National University, Busan, Korea. His current
research interests include the development of neuromorphic multimodal sensor platforms
and the design of next-generation memory devices.
Jonghyun Yun received his B.S. degree in electrical and electronics engineering
from Pusan National University, in 2025. He is currently pursuing a master's degree
in electrical and electronics engineering at Pusan National University. His current
research interests include the design and fabrication of metal oxide thin film transistor
and the development of advanced sensing devices.
Dong Jin Han received his B.S. degree in electrical and electronics engineering
from Pusan National University, in 2025. He is currently pursuing a master's degree
in electrical and electronics engineering at Pusan National University. His current
research interests include gas sensors for early detection of thermal runaway in lithium-ion
batteries, and semiconductor devices based on AlGaN/GaN heterostructures.
Sung-Tae Lee received his B.S. and Ph.D. degrees in electrical and computer engineering
from Seoul National University (SNU), Seoul, Korea, in 2016 and 2021, respectively.
He was a Post-Doctoral Fellow with the Georgia Institute of Technology, from 2021
to 2022. He has been a Professor at the School of Electronic and Electrical Engineering,
Hongik University, since 2023. His current research interests include neuromorphic
devices and their application in computing, NAND flash memory, noise analysis of semiconductor
devices.
Sung Yun Woo received his B.S. degree in electronics engineering from Kyungpook
National University, in 2014. He then received an integrated M.S./Ph.D Program in
electrical engineering from Seoul National University (SNU), in 2021. In 2021, he
joined Samsung Electronics, Hwaseong, Gyeonggi-do, Korea, where he has worked on developing
conventional DRAM, high bandwidth memory (HBM), and fabrication processes. Since March
2023, he has been an Assistant Professor in the School of Electronics Engineering
at Kyungpook National University, Daegu, Korea.