sensors

Article
Development of Miniaturized Water Quality
Monitoring System Using Wireless Communication
Hsing-Cheng Yu 1, * , Ming-Yang Tsai 1 , Yuan-Chih Tsai 1 , Jhih-Jyun You 2 , Chun-Lin Cheng 2 ,
Jung-How Wang 2 and Szu-Ju Li 2
1 Department of Systems Engineering and Naval Architecture, National Taiwan Ocean University,
2 Pei-Ning Road, Keelung 20224, Taiwan
2 Material and Chemical Research Laboratories, Industrial Technology Research Institute, 195, Sec. 4,
Chung Hsing Road, Chutung, Hsinchu 31040, Taiwan
* Correspondence: hcyu@ntou.edu.tw; Tel.: +886-2-24622192 (ext. 6059)




Received: 26 July 2019; Accepted: 28 August 2019; Published: 30 August 2019


Abstract: Recently, environmental pollution resulting from industrial waste has been emerging in
an endless stream. The industrial waste contains chemical materials, heavy metal ions, and other
toxic materials. Once the industrial waste is discharged without standards, it might lead to water or
environmental pollution. Hence, it has become more important to provide evidence-based water
quality monitoring. The use of a multifunctional miniaturized water quality monitoring system
(WQMS), that contains continuous monitoring, water quality monitoring, and wireless communication
applications, simultaneously, is infrequent. Thus, electrodes integrated with polydimethylsiloxane
flow channels were presented in this study to be a compound sensor, and the sensor can be adopted
concurrently to measure temperature, pH, electrical conductivity, and copper ion concentration,
whose sensitivities are determined as 0.0193 ◦ C/mV, −0.0642 pH/mV, 1.1008 mS/V·cm (from 0 mS/cm
to 2 mS/cm) and 1.1975 mS/V·cm (from 2 mS/cm to 5.07 mS/cm), and 0.0111 ppm/mV, respectively.
A LoRa shield connected into the system could provide support as a node of long range wide area
network (LoRaWAN) for wireless communication application. As mentioned above, the sensors,
LoRa, and circuit have been integrated in this study to a continuous monitoring system, WQMS.
The advantages of the multifunctional miniaturized WQMS are low cost, small size, easy maintenance,
continuous sampling and long-term monitoring for many days. Every tested period is 180 min,
and the measured rate is 5 times per 20 min. The feedback signals of the miniaturized WQMS
and measured values of the instrument were obtained to compare the difference. In the measured
results at three different place-to-place locations the errors of electrical conductivity are 0.051 mS/cm,
0.106 mS/cm, and 0.092 mS/cm, respectively. The errors of pH are 0.68, 0.87, and 0.56, respectively.
The errors of temperature are 0.311 ◦ C, 0.252 ◦ C, and 0.304 ◦ C, respectively. The errors of copper ion
concentration are 0.051 ppm, 0.058 ppm, 0.050 ppm, respectively.

Keywords: embedded system; water quality monitoring; wireless communication

1. Introduction
Industrial waste contains toxic substances such as chemical substances, which cause water and
environmental pollution. Once the industrial waste is discharged without standards, it might lead to
water or environmental pollution. Thus, it has become more important to inquire into water quality
monitoring. With the progress of semiconductor, large data processing, and wireless communication
technologies, a miniaturized water quality monitoring system (WQMS) could be developed in this study.
Generally, in order to identify the level of water pollution, water quality indicators are used to analysis
the water properties: physical (e.g., temperature, odor, turbidity, and electrical conductivity), chemical

Sensors 2019, 19, 3758; doi:10.3390/s19173758 www.mdpi.com/journal/sensors

Sensors 2019, 19, 3758 2 of 14

(e.g., pH value, copper ion concentration), and biological (e.g., content of organisms) characteristics.
Among them, the water temperature indirectly affects some measured standards of the indicators,
such as electrical conductivity, dissolved oxygen, and so on. Acidity is also called the pH value, which
is the hydrogen ion concentration in water. Aquatic organisms are sensitive to acid and alkali, so the
pH value is crucial to the survival of aquatic organisms. Electrical conductivity, related to the ionic
strength of water, can be used to infer the quality of dissolved materials in water. It can be used to
determine whether water is polluted or to what extent saltwater infusion occurs in a coastal area.
The copper ion concentration in general water is between 20 and 75 ppb [1], but there may be a large
amount of copper ions of water in old pipelines or contaminated water sources. If such water becomes
household drinking water, it may affect human health. Hence, the US Environmental Protection
Agency has made a standard policy to limit the copper ion concentration in lakes and rivers to less
than 1 ppm, and drinking water must be less than 1.3 ppm [2]. Zhou et al. published a proposed low
cost and miniature sensor chip with multi-parameters [3]. The sensor chip can detect the conductivity,
temperature, and pH of water simultaneously. Iridium oxide film was electrodeposited as pH-sensing
material to fabricate a chip. The linearity of the pH sensor was R2 = 0.9997 and sensitivity was
−67.60 mV/pH. Furthermore, LoRa technology features a long range, low power consumption (i.e., long
battery life), multi-node network, and low cost [4]. Otherwise, ZigBee technology is a low-power LAN
protocol based on the IEEE802.15.4 standard. ZigBee is a two-way wireless communication technology
with a closed range, low complexity, low power consumption, low speed, and low cost. It is mainly
used for data transmission between various electronic devices over a short distance with low power
consumption and a low transmission rate, and has typical applications with periodic data, intermittent
data and low response time data transmission. It can be seen that the transmission distance can be
lengthened by using LoRa technology, so LoRa has been chosen to adopt in this study. According to the
National Environmental Water Quality Monitoring Information Network of Environmental Protection
Administration, Executive Yuan, R.O.C. (Taiwan), it is necessary to directly measure the pollutant
content in the water environment, including physical, chemical and biological data, to provide accurate
and reliable data to describe the water quality, and to compare with the water quality standards of the
water body for reference to the assessment of the conformity of water quality standards. According to
the regulation of the National Environmental Water Quality Monitoring Information Network of
Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan), most industrial areas have
to meet the specified values. For example, the temperature must be below 38 ◦ C, the pH measurement
range should be from 6.0 to 9.0, and the copper ion content should be 3.0 ppm or less [5–7]. Based on
the above reasons, this research proposes a miniaturized WQMS with multiple functions for wireless
communication application of water quality monitoring.

2. Measurement Principle
The multifunctional WQMS using wireless communication technology is presented in this
study, and it can be utilized to measure temperature, pH, electrical conductivity, and copper ion
concentration, simultaneously.

2.1. Measurement of Temperature

The initial capacitance of the capacitive sensor is very small. The stray capacitance of the measuring
circuit and the capacitance of the sensor plate and its surrounding conductors are large, which reduces
the sensitivity of the sensor; on the other hand, these capacitors often change randomly, which can
destabilize the sensor and affect measurement accuracy [8].
A resistance temperature detector (RTD) has many advantages of natural properties of the metal,
including high stability, reliability and accuracy. When the temperature rises due to heat, the resistance
of the metal will rise. Most RTDs are made of materials such as white gold, copper or nickel. The thin
film type RTD is coated with a metal film on a substrate. The characteristics of RTD sensing resistance
Sensors 2019, 19, 3758 3 of 14

with temperature variation can be applied to measurement of temperature. The functional relationship
between resistance value and temperature is in form of:

RT = R0 (1 + αT ) (1)

where RT is the resistance value of RTD at T ◦ C, R0 is the resistance value of the RTD at 0 ◦ C, α is the
temperature coefficient of the resistance, and T is the temperature in Celsius.

2.2. Measurement of pH Value in Water

The architecture of ion sensitive field effect transistor (ISFET) was set on removing the gate of
the metal-oxide-semiconductor field-effect transistor (MOSFET) [9]. In the transistor, the currents
are varied by the concentration of hydrogen ions in the solution when elements are soaked into the
solution. The most serious shortcoming of the ISFET architecture is that the ion selection effect is poor,
and the small size is difficult to process. When the measurement is performed, the immersion of the
transistor in the solution is easily eroded by the object to be tested, resulting in damage or electrical
changes [10]. Aiming to deal with the shortcoming of ISFETs, in 1983, Jan van der Spiegel et al. brought
forth the architecture of the extend-gate field-effect transistor (EGFET) [11]. The main change was
to keep the metal gate of the MOSFET separate from the gate and extend out with the wire, which
required a low-impedance and a conductive film with high conductivity. The most importance part
of the adjustment was to separate the sensing block from the component block, which improved the
problem of vulnerability and damage. Moreover, the sensing area became adjustable in dimension and
shape, which greatly improved the adaptability of the measurement [12].

2.3. Measurement of Electrical Conductivity

Electrical conductivity is a physical quantity that refers to the ability of an object to pass current.
The rate of movement of ions is affected by temperature. The temperature is higher, and the movement
of ions is faster. The solution carries a current through the movement of anion and cation. Thus,
the electrical conductivity decreases with a temperature increase. Generally, differences in composition
of water cause different temperature effects, and each increase of 1 ◦ C causes the electrical conductivity
to increase by 1.9%.
The two-pole electrode contains two parallel plate electrodes, and it applies alternating current
at both ends. The voltage is measured between the two electrodes, so that the resistance of the
solution can be determined. The reciprocal of the resistance is the electrical conductivity. Additionally,
the conductance influence can be separated by using AC method, and it can measure the exact solution
resistance [13].

2.4. Measurement of Copper Ion Concentration

Copper is an important trace element needed to sustain human life, but the amount should not be
too great. When the copper ion content in water reaches 0.01 ppm, the self-purification ability of the
water itself is suppressed. When the copper ion content exceeds 3 ppm, the water has a significant
odor. Besides, the water is unsuitable to drink if the copper ion content exceeds 15 ppm.
Ion selective electrodes, also known as thin film electrodes, are an electrochemical type of sensor.
The ion selective electrode can measure the potential of a specific ion. When a specific ion in the
solution to be tested is in contact with the sensing film, a film potential is generated on the sensing film.
This film potential can generate a potential difference from the potential of the reference electrode,
which can be adopted to convert the concentration of the ion to be measured one.

2.5. Measurement of LoRa Wireless Transmission

LoRa technology has been developed as a long-distance wireless communication technology in
recent years. Its transmission range is about 15 km, and the battery power supply of the wireless
Sensors 2019, 19, 3758 4 of 14

communication device
Sensors 2019, 19, x FOR can be used for several years. Additionally, low power wide area network
PEER REVIEW 4 of 15
(LPWAN), a power-saving and long-distance transmission method, was created in response to the
demand for for the
the wireless
wirelesscommunication.
communication.Furthermore,
Furthermore,long longrange
rangewide
widearea
areanetwork
network (LoRaWAN)
(LoRaWAN) is
is one
one ofof a varietyofofLPWAN
a variety LPWANspecifications
specificationsthat
thatfocuses
focuseson onlow-power
low-power and
and long-range
long-range transmission,
transmission,
defines the
which defines the communication
communication architecture
architecture ofof the
the system
systemarchitecture
architectureand
andnetwork.
network.

3. Experimental Setup
The miniaturized WQMS consists of four modules, which are the water quality monitoring
module, control circuit module, wireless transmitter module, and the sample and cleaning module,
as shown in Figure 1. The
The front
front and back views
views of the
the experimental
experimental device diagram of the real WQMS
are shown in Figure 2.

Figure 1. A block diagram of the miniaturized water quality monitoring system.

Sensors 2019, 19, 3758 5 of 14
Sensors 2019, 19, x FOR PEER REVIEW 5 of 15

Figure
Figure 2. The experimental
2. The experimental device
device diagram
diagram of
of the
the real
real water
water quality
quality monitoring
monitoring system:
system: (a)
(a) front
front view;
view;
(b)
(b) back
back view.
view.

3.1. Water Quality Monitoring Module

3.1. Water Quality Monitoring Module
The designed sensing electrodes of the water quality monitoring module integrate three measured
The designed sensing electrodes of the water quality monitoring module integrate three
parameters including temperature, electrical conductivity, and pH values. The temperature indication
measured parameters including temperature, electrical conductivity, and pH values. The
used a gold-plated electrode to make an RTD-type temperature sensor, and its characteristics have good
temperature indication used a gold-plated electrode to make an RTD-type temperature sensor, and
linearity. The conductivity indication was made of a two-pole electrode, and the platinum electrode
its characteristics have good linearity. The conductivity indication was made of a two-pole electrode,
was used in the measurement area, which can reduce the possibility of reacting with the object tested.
and the platinum electrode was used in the measurement area, which can reduce the possibility of
The material quality of the platinum electrodes have a degree of protection for the sensing electrodes of
reacting with the object tested. The material quality of the platinum electrodes have a degree of
the electrical conductivity. The pH sensing electrode was made of indium tin oxide (ITO) as the sensing
protection for the sensing electrodes of the electrical conductivity. The pH sensing electrode was
film material, and the reference electrode was made of the non-reactive platinum electrode. Many
made of indium tin oxide (ITO) as the sensing film material, and the reference electrode was made of
studies of sensing terminals are based on commercially available stable pen-type sensors. This study
the non-reactive platinum electrode. Many studies of sensing terminals are based on commercially
mainly achieved the same effect as commercially available sensors at a low cost. In general, the reference
available stable pen-type sensors. This study mainly achieved the same effect as commercially
electrode is configured with a glass film. Herein, the reference electrode was configured with an ITO
available sensors at a low cost. In general, the reference electrode is configured with a glass film.
film [14]. The gold-plated electrodes at the contact point could produce good conductivity and reduce
Herein, the reference electrode was configured with an ITO film [14]. The gold-plated electrodes at
noise. A part of the flow channel adopted a positive engraved acrylic mold, and it only needed to be
the contact point could produce good conductivity and reduce noise. A part of the flow channel
filled with a proper amount of polydimethylsiloxane (PDMS). After drying, it could be combined with
adopted a positive engraved acrylic mold, and it only needed to be filled with a proper amount of
the sensing electrode. The copper ion electrode adopted a commercially available copper ion electrode.
polydimethylsiloxane (PDMS). After drying, it could be combined with the sensing electrode. The
The electrode was an ion selective electrode in the form of a solid membrane, and the measurement
copper ion electrode adopted a commercially available copper ion electrode. The electrode was an
concentration range was 5 × 10−8 M to 5 × 10−1 M. All sensing electrodes were made on a glass substrate
ion selective electrode in the form of a solid membrane, and the measurement concentration range
and three layers of material were formed, whereof the first layer was ITO, the second layer was plated
was 5 × 10−8 M to 5 × 10−1 M. All sensing electrodes were made on a glass substrate and three layers
with platinum in a place other than the pH sensing region, and the third layer was a layer of gold on the
of material were formed, whereof the first layer was ITO, the second layer was plated with platinum
contacts and the temperature sensor. The water quality detecting chip was cemented as shown on the
in a place other than the pH sensing region, and the third layer was a layer of gold on the contacts
left-hand side in Figure 3, and the copper ion selective electrode for monitoring copper ion concentration
and the temperature sensor. The water quality detecting chip was cemented as shown on the left-
is shown on the right-hand side in Figure 3.
hand side in Figure 3, and the copper ion selective electrode for monitoring copper ion concentration
is shown on the right-hand side in Figure 3.
Sensors 2019, 19, 3758 6 of 14
Sensors 2019, 19, x FOR PEER REVIEW 6 of 15

Figure 3. The sensing electrodes of the water quality monitoring module.

3.2. Control Circuit

3.2. Control Circuit Module
Module
The
The temperature
temperature monitoring
monitoring was
was aa gold-plated
gold-plated electrode
electrode in
in the
the form
form of RTD, which
of RTD, which changed
changed its
its
resistance value with temperature. The RTD resistor was combined with the operational
resistance value with temperature. The RTD resistor was combined with the operational amplifier amplifier and
used a 5 V aoperational
and used voltagevoltage
5 V operational for the inverting input. One
for the inverting of the
input. Oneinverting amplifiersamplifiers
of the inverting was converted
was
to the RTD resistor. The output voltage could be changed with temperature, but the
converted to the RTD resistor. The output voltage could be changed with temperature, but the ATmega32U4
could not be measured.
ATmega32U4 could notThe output voltage
be measured. The value
outputofvoltage
an analog-to-digital converter (ADC)converter
value of an analog-to-digital from the
temperature
(ADC) from measurement sub-circuit
the temperature read bysub-circuit
measurement a microcontroller
read byunit (MCU), VTEMP , unit
a microcontroller can be expressed
(MCU), as,
VTEMP
can be expressed as
−RT −RT3
V =− RVTCC−· RT 3 · (2)
VTEMP TEMP
= VCC ⋅ ⋅ R RT2 (2)
RT 1 RT 2 T1
where the R RTT is the resistance value of the RTD, RTD, and and RRT1
T1 is the resistance value of the first-stage

circuit in
operational amplifier circuit in the
thetemperature
temperaturemeasurement
measurementsub-circuit.
sub-circuit.RRT2T2 and R T3 are
T3 are the input
resistance
resistance and feedback resistance of the second-stage operational amplifier circuit in the temperature
respectively.
measurement sub-circuit, respectively.
The pH measurement circuit adopts the architecture architecture of EGFET for circuit design, and the IC was
CD4007.
CD4007. The IC contains three sets of CMOS transistors.InIn
IC contains three sets of CMOS transistors. this
this study,
study, thethe gate,
gate, drain
drain andand source
source of
of the
the three
three sets sets of transistors
of transistors werewere connected
connected in parallel,
in parallel, which which can change
can change the characteristic
the characteristic curve,
curve, so
so that
that the output current increases, and the sensitivity increased when the pH
the output current increases, and the sensitivity increased when the pH value was measured. The pH value was measured.
The pH measurement
measurement output V
output voltage voltage
pH wasVin was in
pHform of:form of:
− RpH 3
VpH = I D ⋅ RpH1 ⋅ −RpH3 (3)
VpH = ID · R RpH 2
pH1 · (3)
where ID is the drain current, RpH1 is the resistance value R ofpH2
the current-to-voltage converter, RpH2 and
RpH3 are the resistance values of the input resistance and feedback resistor of the operational amplifier
where ID is the drain current, RpH1 is the resistance value of the current-to-voltage converter, RpH2 and
circuit in the pH measurement sub-circuit, respectively. The pH measurement sub-circuit was
RpH3 are the resistance values of the input resistance and feedback resistor of the operational amplifier
designed with a second-order low-pass filter before the output. The purpose was to filter out the
circuit in the pH measurement sub-circuit, respectively. The pH measurement sub-circuit was designed
noise that may be caused by the 1 kHz AC signal used in the electrical conductivity measurement.
with a second-order low-pass filter before the output. The purpose was to filter out the noise that
The filtering power of the second-order low-pass filter was closer to the ideal filter than the first-
may be caused by the 1 kHz AC signal used in the electrical conductivity measurement. The filtering
order low-pass filter, which could provide better filtering effect.
power of the second-order low-pass filter was closer to the ideal filter than the first-order low-pass
When the conductivity sensor was filled with solution, the resistance between the two electrodes
filter, which could provide better filtering effect.
would decrease as the conductivity increases, and it was used as the resistance of the inverting input
of the inverting amplifier [15]. The distance between the two electrodes in the solution was fixed, and
the equivalent resistance and electrical conductivity of the solution were reciprocal. The output
voltage of the secondary inverting amplifier Vout was in form of:
Sensors 2019, 19, 3758 7 of 14

When the conductivity sensor was filled with solution, the resistance between the two electrodes
would decrease as the conductivity increases, and it was used as the resistance of the inverting input of
the inverting amplifier [15]. The distance between the two electrodes in the solution was fixed, and the
equivalent resistance and electrical conductivity of the solution were reciprocal. The output voltage of
the secondary inverting amplifier Vout was in form of:

−REC3
Vout = Vin_EC · (4)
REC

where REC is the equivalent resistance between the A and B terminals of the conducting electrode, and
REC3 is the feedback resistance of the second stage of the conducting measuring circuit. The Vout is
an AC voltage signal, which must be converted to a DC signal to supply the analog input pin of the
MCU. In this study, a single-phase half-wave rectifier was used to convert AC to DC and remove noise
through a low-pass filter. The DC voltage Vdc output from the single-phase half-wave rectifier was the
input AC voltage, and the average value was in form of:
Z π
1 Vmax h Vmax
− cos(ωt)π0 =
 i
Vdc = sin(ωt) d(ωt) = (5)
2π 0 2π π

where V max is the peak voltage of the AC voltage, Vdc is the average voltage of the DC voltage output
by the single-phase half-wave rectifier, and Vdc is obtained by the first-order low-pass filter to obtain
the ADC voltage value (VEC ) of the conductivity measurement sub-circuit read by the MCU.
The copper ion measurement used the ion selective electrode, and the signal was transmitted to
the measurement circuit by the BNC connector. The theoretical formula for calculating the output
voltage of the copper ion measuring circuit, VCu , is in the form of:
! !
RCu1 + RCu2 − Rv RCu5
VCu = Vin_Cu · 1 + · 1+ (6)
RCu3 + Rv RCu4

where Vin_Cu is the ion-selective electrode measuring voltage signal, RCu1 and RCu3 are the resistances
of the first-stage amplifier of the copper ion measuring sub-circuit, respectively, RCu4 and RCu5 are the
resistances of the second-stage amplifier of the copper ion measuring sub-circuit, respectively, and
RCu2 is the total resistance of the variable resistor, and Rv is the resistance of the variable resistance
between sliding contact and the fixed contact of the RCu3 terminal.
The wireless communication test of the LoRa module utilized the LoRa module from Dragino to
communicate with the LoRa gateway LG-01. Due to the lack of a suitable cloud platform, this paper
did not test the cloud platform upload program for the time being. The LG-01 was set to the WiFi
receiver network mode. The network sent the wireless network signal through the router as the AP,
and the LG01 acted as the receiver of the wireless network and connected through the router. Since
the LG01 could not be used as both the receiving end and the AP, it could only be connected to the
personal computer via the local area network, and the personal computer was connected to the Internet
using the Ethernet route. The LG-01 and the personal computer were also connected to the same
group of WiFi area networks. The LoRa module was connected to the personal computer through
the UNO module below to compose and burn the firmware. The LoRa gateway was the server end.
After receiving the request of the LoRa communication module, it could reply the command to the
communication module indicating that it had been received, and the LoRa communication module
sent the request to the LoRa gateway. If there was a gateway receiving the signal, it could receive
a response from the LoRa gateway. If there was no error, the wireless communication between the
two used the A-type transmission mechanism, and the message could be received at any time, which
has the most power-saving advantages. Although the transmission waiting time was relatively long,
the system did not require too much transmission efficiency. When it is placed outdoors for a long
time, it can save power, which can increase the time that the whole system can operate, and reduce
 Sensors 2019, 19, 3758 8 of 14

the repair numbers and cost of maintenance. Only the wireless communication between the module
and the gateway was tested in this study. If the appropriate cloud platform or application program
could be obtained during the application, the data could be uploaded and monitored as the wireless
communication node.

4. Experimental Results
Three various local water sources were sampled for testing in this study, called Samples A, B, and C,
respectively. The sampled water characteristics were tested by the instrument, and then monitored by
Sensors 2019, 19, x FOR PEER REVIEW 8 of 15
the miniaturized WQMS. The continuous monitoring time was 180 min, and the measuring rate was
five times per 20 min.
4.1. Monitoring of Temperature
4.1. Monitoring of Temperature
The verification of temperature electrode of the miniaturized WQMS depended on the values
The verification of temperature electrode of the miniaturized WQMS depended on the values
between between
the AVR voltage of the MCU on the measured motherboard and real temperature.
the AVR voltage of the MCU on the measured motherboard and real temperature. Temperature
Temperature measured
measured ranged
ranged from 15 ◦ Cfrom
to 50 ◦15
C, °C
andto 50modified
was °C, andbywas
usedmodified by while
of a hot plate, used monitoring
of a hot plate,
the while
monitoring the temperature
temperature by simultaneously.
by a thermocouple a thermocouple simultaneously.
The experimental setup forThe experimental
measuring temperature setup
is for
shown in Figure 4.
measuring temperature is shown in Figure 4.

Figure 4. The
Figure experimental
4. The experimentalsetup formeasuring
setup for measuring temperature.
temperature.

The temperature electrode of the miniaturized WQMS showed a sensitivity of 0.0193 ◦ C/mV in
The temperature◦ electrode of the miniaturized WQMS showed a sensitivity of 0.0193 °C/mV in
the range of 15 C to 50 ◦ C, and a good linearity (coefficient of determination RT 2 = 0.9995) was found.
the rangeThe
of accuracy
15 °C toof50temperature
°C, and a was
good linearity
±0.193 (coefficient
◦ C. During heating,of
thedetermination RT2 = was
temperature electrode 0.9995) was found.
put into
The accuracy of temperature
the water, and it recordedwas ±0.193
the AVR °C. of
voltages During
the MCU heating, the temperature
on the measured electrode
motherboard, was put into
and monitored
the water,
the and
heatingit temperature
recorded the AVR
of the voltages
water, of the As
simultaneously. MCU on the
the results measured tests,
of temperature motherboard,
an X–Y and
scatter diagram showing the calibration curve is shown in Figure 5.
monitored the heating temperature of the water, simultaneously. As the results of temperature tests,
The experimental results of temperature by using Samples A, B, and C are shown in Figure 6.
an X–Y scatter diagram showing the calibration ◦ curve
◦ is shown◦ in Figure 5.
The errors of temperature were 0.311 C, 0.252 C, and 0.304 C, respectively. The temperature
measured by the water quality testing system in both Sample A and Sample B was consistent with the
temperature at which the laboratory had air conditioning. Sample C was tested in a non-air-conditioned
state, and the error was not greater than 0.5 ◦ C.
the range of 15 °C to 50 °C, and a good linearity (coefficient of determination RT2 = 0.9995) was found.
The accuracy of temperature was ±0.193 °C. During heating, the temperature electrode was put into
the water, and it recorded the AVR voltages of the MCU on the measured motherboard, and
monitored the heating temperature of the water, simultaneously. As the results of temperature tests,
Sensors
an X–Y 2019, 19, 3758
scatter diagram showing the calibration curve is shown in Figure 5. 9 of 14

Sensors 2019, 19, x FOR PEER REVIEW 9 of 15

the Sensors
temperature
2019, 19, x at
FORwhich the laboratory had air conditioning. Sample C was tested in a9 non-air-
PEER REVIEW of 15
conditioned state, and the error was not greater than 0.5 °C.
the temperature at which the laboratory
Figure had air curve
5. The calibration conditioning. Sample C was tested in a non-air-
of the temperature.
Figure
conditioned state, and the error 5. The
was calibration
not curve
greater than 0.5 of
°C.the temperature.

The experimental results of temperature by using Samples A, B, and C are shown in Figure 6.
The errors of temperature were 0.311 °C, 0.252 °C, and 0.304 °C, respectively. The temperature
measured by the water quality testing system in both Sample A and Sample B was consistent with

Figure 6. 6.The
Figure Theexperimental
experimentalresults of the temperature byusing
usingSamples
SamplesA,A,B,B,and
and C.
Figure 6. The experimentalresults
results of
of the temperatureby
the temperature byusing Samples A, B, and C. C.
4.2. Monitoring of pH Value
4.2. 4.2.
Monitoring of pH
Monitoring of pHValue
Value
To test the potential of the ITO film, an Arduino Leonardo was utilized in this study to record
To To
testtest
thethe
potential
potential ofofthe
theITO
ITOfilm,
film, an
an Arduino Leonardowas
Arduino Leonardo wasutilized
utilized
in in this
this study
study to record
to record
ADC values and save the monitoring data into the personal computer. The experimental setup for
ADC ADCvalues and
values andsave
savethethemonitoring
monitoringdata data into
into the personal
personalcomputer.
computer.The
The experimental
experimental setup
setup for for
measuring pH values
measuring is shown in in
Figure
Figure7.7.
measuring pHpH values
values is is shown
shown in Figure 7.

Figure 7. The experimental setup for measuring pH values.

Figure 7. The experimental setup for measuring pH values.
Figure 7. The experimental setup for measuring pH values.
The verification of pH calibration compared the AVR voltages of the MCU on the measured
motherboard and real pH values. According to the standard solution pH 4, pH 7, and pH 10,
The verification of pH calibration compared the AVR voltages of the MCU on the measured
respectively, the measured ranges were from pH 4.0 to pH 10.0. The accuracy of pH value was
motherboard
±1.0917. Theand
pH real pH values.
electrode According WQMS
of the miniaturized to the standard solution pH
showed a sensitivity 4, pH pH/mV
of –0.0642 7, and and
pH 10,
respectively,
had a good linearity (coefficient of determination RpH = 0.9954). The pH tests used the Arduinowas
the measured ranges were from pH 4.0 2to pH 10.0. The accuracy of pH value
Sensors 2019, 19, 3758 10 of 14

The verification of pH calibration compared the AVR voltages of the MCU on the measured
motherboard and real pH values. According to the standard solution pH 4, pH 7, and pH 10,
respectively, the measured ranges were from pH 4.0 to pH 10.0. The accuracy of pH value was ±1.0917.
The pH electrode of the miniaturized WQMS showed a sensitivity of −0.0642 pH/mV and had a good
linearity (coefficient of determination RpH 2 = 0.9954). The pH tests used the Arduino Leonardo to
Sensors
acquire 2019,
the19,
AVRx FOR PEER REVIEW
voltages 10 of 15
of the MCU on the measured motherboard, which were the electric potential
Sensors 2019, 19, x FOR PEER REVIEW 10 of 15
of the ITO film; the calibration curve of pH values is shown in Figure 8.

Figure
Figure 8.
8. The
The calibration
calibration curve
curve of
of pH
pH values.
values.
Figure 8. The calibration curve of pH values.
The experimental results of pH values for Samples A, B, and C are shown in Figure 9. The
The errors
errors
The experimental results of pH values for Samples A, B, and C are shown in Figure 9. The errors
of pH were 0.68, 0.87, and 0.56, respectively. The water quality detection system measured mostly
of pH were 0.68, 0.87, and 0.56, respectively. The water quality detection system measured mostly
weakly acidic values. Only
Only a small number of data points were weakly
weakly alkaline.
alkaline. The measurement
weakly acidic values. Only a small number of data points were weakly alkaline. The measurement
results were stable.
results were stable.

The experimental
Figure 9. The experimental results
results of
of the pH values by using Samples A, B, and C.
Figure 9. The experimental results of the pH values by using Samples A, B, and C.
4.3. Monitoring
4.3. Monitoring of
of Electrical
Electrical Conductivity
Conductivity
4.3. Monitoring of Electrical Conductivity
The experimental
The experimental setup
setup for
for measuring
measuring electrical
electrical conductivity
conductivity isis shown
shown in in Figure
Figure 10. The titration
10. The titration
The
method wasexperimental
was adopted
adopted to setup for
to verify measuring
verify the electrical
the conductivity conductivity
conductivity calibration
calibration tests. is shown
tests. Using
Using the in Figure
the method, 10. The
method, deionized titration
deionized water
water
method
method
(DI was
water) wasadopted
added to verify
into the conductivity
saturated salt water calibration
(electrical tests. Using= the
conductivity 5.07 method,
mS/cm) deionized
until the water
voltage
(DI water) was added into saturated salt water (electrical conductivity = 5.07 mS/cm) until the voltage
conductivity =
(DI water)
was close was added into saturated salt water (electrical conductivity = 5.07 mS/cm) untilthe
theportable
voltage
was close to
to zero
zero (electrical
(electrical conductivity = 0.005
0.005 mS/cm). During the
mS/cm). During the titration,
titration, using
using the portable
was close to zero (electrical conductivity = 0.005 mS/cm). During the titration, using the portable
conductivity device for testing, the values of the electrical conductivity and the AVR voltage of the
conductivity device for testing, the values of the electrical conductivity and the AVR voltage of the
MCU on the measured motherboard were recorded simultaneously. The experimental setup for
MCU on the measured motherboard were recorded simultaneously. The experimental setup for
measuring electrical conductivity is shown in Figure 10.
measuring electrical conductivity is shown in Figure 10.
 Sensors 2019, 19, 3758 11 of 14

conductivity device for testing, the values of the electrical conductivity and the AVR voltage of the MCU
Sensors
on 2019, 19, x FOR PEER
the measured REVIEW
motherboard 11 of 15
were recorded simultaneously. The experimental setup for measuring
electrical conductivity is shown in Figure 10.
Sensors 2019, 19, x FOR PEER REVIEW 11 of 15

Figure 10. The experimental setup for measuring electrical conductivity.

Taking the DI Figure 10.as

Figure
water The
10. experimental
The
a sample,experimentalsetup
the saturatedfor measuring
setup for waterelectrical
saltmeasuring conductivity.
electrical
was conductivity.
utilized by the titration method
until the voltage reached the limit of 4 V. When the voltage reached 4 V, the DI water was adopted
Taking
Taking the the
DI DI water
water as aassample,
a sample, the
thevoltagesaturated
saturated saltsalt
waterwater was
waslimit utilized
utilized by the titration method
by the titration method again until the returned to the of 0 byV. the titration
During method
the titration
until
until the the voltage
voltage reached
reached the limit of 4 V. When the voltagereached
reached 4V,V,the theDI DIwater
waterwas wasadopted
adopted by
process, the voltages and the limit
electrical of 4 V. When
conductivity the
valuesvoltage
were recorded4 simultaneously. The accuracy
by the titration
titration method
method again
again until
until the
the voltage
voltage returned
returned to to
thethelimit of 0ofV.0During
limit V. the titration
During the process,
titration
of the electrical conductivity was ±0.0087 mS/cm from 0 mS/cm to 2 mS/cm. The accuracy of the
the
process,voltages and electrical conductivity values were recorded simultaneously. The accuracy of the electrical
electricalthe voltages and
conductivity waselectrical
±0.0327conductivity
mS/cm fromvalues 2 mS/cm wereto recorded
5.07 mS/cm. simultaneously.
As the resultsThe of accuracy
tests, the
of conductivity
the electrical was ±0.0087 mS/cm
conductivity was from 0 mS/cm
±0.0087 mS/cm to 2 mS/cm.
from 0 mS/cmThe to
accuracy
2 mS/cm.of theTheelectrical
accuracy conductivity
of the
electric conductivity had a discontinuity at 2 mS/cm. We obtained two approximations for the electric
was ±0.0327
electrical mS/cmwas from±0.0327
2 mS/cm to 5.07 mS/cm. As the results of tests,Asthetheelectric conductivity had
conductivity. One had a sensitivity of 1.1008 mS/V·cm from 0 mS/cm to 2 mS/cm with the
conductivity mS/cm from 2 mS/cm to 5.07 mS/cm. results of tests, the
a discontinuity
electric conductivity at 2 mS/cm. We
had a discontinuity obtained two approximations for the electric conductivity. One had
determination coefficient of its linearityat R 2EC1
mS/cm. We obtained
2 = 0.9926; and thetwo approximations
other had a sensitivity for the ofelectric
1.1975
a sensitivityOne
conductivity. of 1.1008
had mS/V·cm
a from 0of
sensitivity mS/cm
1.1008 to 2mS/V·cm
mS/cm with fromthe determination
0 mS/cm to coefficient
2 mS/cm of its linearity
with the
mS/V·cm from 2 mS/cm to 5.07 mS/cm with
2 = 0.9926; and the other had a sensitivity the determination coefficient of its linearity R EC22 = 0.5792,
R
determination
EC1 coefficient of its linearity REC1 = 0.9926; and the other had a sensitivity of 1.1975 the
2 of 1.1975 mS/V·cm from 2 mS/cm to 5.07 mS/cm with
as shown in Figure 11. 2 = 0.5792, as shown in Figure 11.
determination
mS/V·cm coefficient
from 2 mS/cm of mS/cm
to 5.07 its linearity
withRtheEC2determination coefficient of its linearity REC22 = 0.5792,
as shown in Figure 11.

Figure 11. The calibration curve of electric conductivity.

Figure 11. The calibration curve of electric conductivity.
The experimental results of electrical conductivity by using Samples A, B, and C are shown in
The experimental Figureof
results 11.electrical
The calibration curve of by
conductivity electric
usingconductivity.
Samples A, mS/cm,
B, and Cand
are 0.092
shown in
Figure 12. The errors of electrical conductivity were 0.051 mS/cm, 0.106 mS/cm,
Figure 12. The errors
respectively. of electrical
The pollution conductivity
ofelectrical
Sample A waswere
lower0.051
than mS/cm,
Sample 0.106
B and mS/cm, and 0.092lower
mS/cm,
The experimental
respectively. results
The pollution of of
Sample conductivity
A was lower thanbySample
using Samples A,Sample
B and Sample B, and C. The
C are
C. The shown
lower
value
in
value
Figure 12. The errors of electrical conductivity were 0.051 mS/cm, 0.106 mS/cm,
meant an upstream source and the higher value meant downstream. The values could get indicateand 0.092 mS/cm,
respectively. Thefrom
higher pollution pollution of Sample
factories A was lower
for downstream thanand
sources Sample
lowerBpollution
and SamplefromC.upstream
The lower value
sources.
meant an upstream source and the higher value meant downstream. The values could get indicate
higher pollution from factories for downstream sources and lower pollution from upstream sources.
Sensors 2019, 19, 3758 12 of 14

meant an upstream source and the higher value meant downstream. The values could get indicate
higher pollution
Sensors 2019, from
19, x FOR factories
PEER REVIEW for downstream sources and lower pollution from upstream sources.
12 of 15

Sensors 2019, 19, x FOR PEER REVIEW 12 of 15

Figure
Figure 12. The experimental
12. The experimental results
results of
of the
the electrical
electrical conductivity
conductivity by
by using
using Samples
Samples A,
A, B,
B, and
and C.
C.

4.4. Monitoring of Copper Ion Concentration

4.4. Monitoring
Figure 12.ofThe
Copper Ion Concentration
experimental results of the electrical conductivity by using Samples A, B, and C.
The copper ion selective electrode was adopted for the measuring experiment of copper ion
The copper ion selective electrode was adopted for the measuring experiment of copper ion
4.4. Monitoring as
concentration, of shown
Copper Ion Concentration
in Figure 13. The calibration curve of copper ion concentration is shown in
concentration, as shown in Figure 13. The calibration curve of copper ion concentration is shown in
Figure 14. The verification
The of the calibration curve of for copper measuring
ion concentrations wereofrespectively
Figure 14.copper ion selective
The verification electrode
of the calibrationwas adopted
curve of copperthe ion concentrations experiment copper10.0
were respectively ion
10.0 ppm, 7.5
concentration, ppm,
as 5.0
shown ppm,
in 2.5
Figure ppm,
13. and
The 1.0 ppm,
calibration so
curvethe measured
of copper ranges
ion were
concentrationfrom is 1shown
ppm toin
ppm, 7.5 ppm, 5.0 ppm, 2.5 ppm, and 1.0 ppm, so the measured ranges were from 1 ppm to 10 ppm.
10 ppm.
Figure The
14. Thecopper ion selective electrode for measuring copper ion concentration showed a sensitivity
The copper ionverification of the calibration
selective electrode curve of
for measuring copperion
copper ionconcentration
concentrations were respectively
showed
2 = 0.9255. a sensitivity 10.0
of
of 0.0111
ppm, 7.5 ppm/mV
ppm, 5.0 and the
ppm, 2.5 determination
ppm, and 1.0 coefficient
ppm, theitsof
so of its linearity2 RCuwere
measured 1 The
ppmaccuracy of
0.0111 ppm/mV and the determination coefficient linearity Rranges
Cu = 0.9255. from
The accuracy toof10 ppm.
copper
copper ion detection
The detection
copper ion was ±0.1846 ppm.
selective During measuring, ion the signal of the copper ion selective
ion was ±0.1846electrode
ppm. Duringfor measuring
measuring, copper
the signal concentration
of the coppershowed a sensitivity
ion selective of
electrode
electrode
0.0111 was transmitted
ppm/mV and to the copper
the determination ion measuring
coefficient circuit;
of its linearity furthermore,
RCu = the the
0.9255. monitoring
The accuracy data from
was transmitted to the copper ion measuring circuit; furthermore, 2
monitoring data of copper
from the
the
ion copper
detectionionwas
measuring
±0.1846 circuitDuring
ppm. via themeasuring,
measuring the motherboard
signal of were
the recorded
copper ion into the electrode
selective personal
copper ion measuring circuit via the measuring motherboard were recorded into the personal
computer simultaneously.
was transmitted to the copper ion measuring circuit; furthermore, the monitoring data from the
computer simultaneously.
copper ion measuringresults
The experimental circuitofvia
copper
the ion concentration
measuring by using were
motherboard Samples A, B, and
recorded intoC are
the shown in
personal
Figure 15. The errors
computer simultaneously. of copper ion concentration were 0.051 ppm, 0.058 ppm, 0.050 ppm, respectively.

Figure 13. The experimental setup for measuring copper ion concentration.

Figure 13.
Figure 13. The experimental setup for measuring copper ion
ion concentration.
concentration.
Sensors 2019, 19, 3758 13 of 14
Sensors 2019, 19, x FOR PEER REVIEW 13 of 15

Figure 14. The calibration curve of copper ion concentration.

The experimental results of copper ion concentration by using Samples A, B, and C are shown
in Figure 15. The errors of copper ion concentration were 0.051 ppm, 0.058 ppm, 0.050 ppm,
respectively.
It is speculated that there may have been no copper ion component or that the copper ion
concentration was too low, possibly lower than the detection limit of the instrument. Disturbances
may occur during the measurement process, affecting the readings of the water quality detection
system, but most of the measurement values were zero. The miniaturized WQMS was designed to
replace the portable instrument. Even though two peaks of copper ion concentration up to 3.0 ppm
were detected by of the miniaturized WQMS, they might have been just some shot noise or
disturbances. They were all under the regulation standard of the National Environmental Water
Quality Monitoring Information Network of Environmental Protection Administration, Executive
Figure 14. The calibration curve of copper ion concentration.
Yuan, R.O.C. (Taiwan).
Figure 14. The calibration curve of copper ion concentration.

The experimental results of copper ion concentration by using Samples A, B, and C are shown
in Figure 15. The errors of copper ion concentration were 0.051 ppm, 0.058 ppm, 0.050 ppm,
respectively.
It is speculated that there may have been no copper ion component or that the copper ion
concentration was too low, possibly lower than the detection limit of the instrument. Disturbances
may occur during the measurement process, affecting the readings of the water quality detection
system, but most of the measurement values were zero. The miniaturized WQMS was designed to
replace the portable instrument. Even though two peaks of copper ion concentration up to 3.0 ppm
were detected by of the miniaturized WQMS, they might have been just some shot noise or
disturbances. They were all under the regulation standard of the National Environmental Water
Quality Monitoring Information Network of Environmental Protection Administration, Executive
Yuan, R.O.C. (Taiwan).

Figure 15. The experimental results of copper ion concentration by using Samples A, B, and C.
Figure 15. The experimental results of copper ion concentration by using Samples A, B, and C.

It is speculated that there may have been no copper ion component or that the copper ion
5. Conclusions.
concentration was too low, possibly lower than the detection limit of the instrument. Disturbances may
occur duringA multifunctional
the measurementminiaturized water the
process, affecting quality monitoring
readings of the system (WQMS)
water quality has beensystem,
detection developed
but most of the measurement values were zero. The miniaturized WQMS was designed to replace ion
in this study to simultaneously measure temperature, pH, electrical conductivity, and copper
concentration,
the portable instrument.andEven
theirthough
sensitivities are 0.0193
two peaks °C/mV,
of copper –0.0642
ion pH/mV, up
concentration 1.1008 mS/V·cm
to 3.0 ppm were(from 0
mS/cm to 2 mS/cm) and 1.1975 mS/V·cm (from 2 mS/cm to 5.07 mS/cm),
detected by of the miniaturized WQMS, they might have been just some shot noise or disturbances.and 0.0111 ppm/mV,
They were respectively. Furthermore,
all under the regulation three different
standard of theplace-to-place comparisonsWater
National Environmental had been sampled
Quality to be water
Monitoring
Information Network of Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan).

5. Conclusions
A multifunctional miniaturized water quality monitoring system (WQMS) has been developed
in this study to simultaneously measure temperature, pH, electrical conductivity, and copper ion
concentration, and their sensitivities are 0.0193 ◦ C/mV, −0.0642 pH/mV, 1.1008 mS/V·cm (from 0 mS/cm
Figure 15. The experimental results of copper ion concentration by using Samples A, B, and C.
to 2 mS/cm) and 1.1975 mS/V·cm (from 2 mS/cm to 5.07 mS/cm), and 0.0111 ppm/mV, respectively.
Furthermore, three different place-to-place comparisons had been sampled to be water quality samples
5. Conclusions.
(i.e., Samples A, B, C) in this study. The continuous sampling period is 180 min, and the measured rate
is fiveAtimes
multifunctional
per 20 min.miniaturized
The feedbackwater quality
signals of themonitoring
WQMS and system (WQMS)
measured hasofbeen
values developed
the instrument
in this
were study to
obtained to simultaneously measureIntemperature,
compare the difference. the measuredpH, electrical
results at threeconductivity, and copper
different locations, ion
the errors
concentration,
of and their are
electrical conductivity sensitivities
0.051 mS/cm,are 0.0193 °C/mV,and
0.106 mS/cm, –0.0642
0.092 pH/mV, 1.1008 mS/V·cm
mS/cm, respectively. (fromof0
The errors
mS/cm to 2 mS/cm) and 1.1975 mS/V·cm (from 2 mS/cm to 5.07 mS/cm), and 0.0111 ppm/mV,
respectively. Furthermore, three different place-to-place comparisons had been sampled to be water
Sensors 2019, 19, 3758 14 of 14

pH are 0.68, 0.87, and 0.56, respectively. The errors of temperature are 0.311 ◦ C, 0.252 ◦ C, and 0.304 ◦ C,
respectively. The errors of copper ion concentration are 0.051 ppm, 0.058 ppm, 0.050 ppm, respectively.
The performances are adequate for environmental water quality monitoring applications.

Author Contributions: H.-C.Y. conceived and designed the experiments; M.-Y.T., Y.-C.T., and J.-J.Y. performed
the experiments mainly; C.-L.C., J.-H.W., and S.-J.L. contributed analysis and investigation; All authors analyzed
sequencing data and simulation; H.-C.Y. wrote the paper, and all authors read and approved the manuscript.
Funding: This research was funded by the Industrial Technology Research Institute and Environmental Protection
Administration, Executive Yuan, R.O.C. (Taiwan), grant number 107A50112 and 107A50116.
Acknowledgments: This work was supported by the Industrial Technology Research Institute and Environmental
Protection Administration, Executive Yuan, R.O.C. (Taiwan).
Conflicts of Interest: The authors declared that they have no conflicts of interest to this work.

References
1. Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan). Available online: https:
//wq.epa.gov.tw/Code/Business/DrinkWaterMetal.aspx (accessed on 8 May 2017).
2. Taipei Water Department. Available online: https://www.water.gov.taipei/cp.aspx?n=B96B3009F7951CD4
(accessed on 26 July 2019).
3. Zhou, B.; Bian, C.; Iong, J.; Xia, S. Fabrication of a Miniature Multi-Parameter Sensor Chip for Water Quality
Assessment. Sensors 2017, 17, 157. [CrossRef] [PubMed]
4. Moparthi, N.R.; Mukesh, C.; Sagar, P.V. Water Quality Monitoring System Using IOT. In Proceedings of the
2018 Fourth International Conference on Advances in Electrical, Electronics, Information, Communication
and Bio-Informatics (AEEICB), Chennai, India, 27–28 February 2018; pp. 1–2.
5. Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan). Available online: http:
//wq.epa.gov.tw/Code/Business/Purpose.aspx (accessed on 10 March 2017).
6. Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan). Available online: https:
//oaout.epa.gov.tw/law/LawContent.aspx?id=GL002557#lawmenu (accessed on 2 February 2007).
7. Environmental Protection Administration, Executive Yuan, R.O.C. (Taiwan). Available online: https:
//oaout.epa.gov.tw/law/LawContent.aspx?id=FL015489 (accessed on 29 April 2019).
8. Alahi, M.E.E.; Xie, L.; Mukhopadhyay, S.; Burkitt, L. A Temperature Compensated Smart Nitrate-Sensor for
Agricultural Industry. IEEE Trans. Ind. Electron. 2017, 64, 7333–7341. [CrossRef]
9. Bergveld, P. Development of an Ion-Sensitive Solid-State Device for Neurophysiological Measurement.
IEEE Trans. Biomed. Eng. 1970, 17, 70–71. [CrossRef] [PubMed]
10. Chen, D.Y.; Chan, P.K. An Intelligent ISFET Sensory System with Temperature and Drift Compensation for
Long-Term Monitoring. IEEE Sens. J. 2008, 8, 1948–1959. [CrossRef]
11. Van der Spiegel, J.; Lauks, I.; Chan, P.; Babic, D. The Extended Gate Chemically Sensitive Field Effect
Transistor as Multi-Species Microprobe. Sens. Actuators B Chem. 1983, 4, 291–298. [CrossRef]
12. Ghusoon, M.A.; Ra’ad, H.D.; Abdullateef, A.A. pH Sensing Characteristics of EGFET Based on Pd-Doped
ZnO Thin Films Synthesized by Sol-Gel Method. In Proceedings of the 2015 Third International Conference
on Technological Advances in Electrical, Electronics and Computer Engineering (TAEECE), Beirut, Lebanon,
29 April–1 May 2015; pp. 234–238.
13. Yuwono, T.; Pramono, W.B.; Ardi, I. Design of the Remote Sensing Circuit for Water Conductivity. In Proceedings
of the 2015 International Conference on Space Science and Communication (IconSpace), Langkawi, Malaysia,
10–12 August 2015; pp. 84–88.
14. Cloete, N.A.; Malekian, R.; Nair, L. Design of Smart Sensors for Real-Time Water Quality Monitoring.
IEEE Access 2016, 4, 3975–3990. [CrossRef]
15. Deyuan Technology, 2017, DFRobot Water Conductivity Sensor Module Circuit Diagram. Available online:
http://www.dfrobot.com.cn/images/upload/File/20140504162158jqhepx.pdf (accessed on 7 January 2014).

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).