See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/337968267

Ammonia concentration monitoring using Arduino

Article in Environmental Engineering · July 2019

DOI: 10.37023/ee.6.1.4

CITATIONS READS

2 2,791

6 authors, including:

Filip Dodigovic Mateo Topic

University of Zagreb University of Zagreb
9 PUBLICATIONS 13 CITATIONS 1 PUBLICATION 2 CITATIONS

SEE PROFILE SEE PROFILE

Lucija Radetić Petar Mrakužić

University of Zagreb University of Zagreb
8 PUBLICATIONS 32 CITATIONS 2 PUBLICATIONS 2 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Special Issue "Photocatalytic Nanoporous Thin Films" View project

Nanocomposites comprising perovskites for photovoltaics, photo-catalysis and sensing (IP-01-2018) View project

All content following this page was uploaded by Ivana Grčić on 29 June 2020.

The user has requested enhancement of the downloaded file.

Petric et al. Ammonia concentration monitoring using Arduino platform

AMMONIA CONCENTRATION MONITORING USING ARDUINO

PLATFORM
Marko Petric 1, 2*, Filip Dodigović 1*, Ivana Grčić 1, Petar Markužić 1, Lucija Radetić 1 Mateo Topić 1
1
University of Zagreb Faculty of Geotehnical Engineering Hallerova aleja 7, 42000 Varaždin, Croatia
2
Jožef Stefan Institute, Jamova cesta 39, SI-1000 Ljubljana, Slovenia
*E-mail of corresponding authors: mpetric@gfv.hr, fdodigov@gfv.hr

Abstract: In order to evaluate the efficacy of ammonia photocatalysis, it is necessary to continuously monitor its concentration
during the process. This paper presents ammonia concentration monitoring using a low cost (non-professional) electronic
device. The device is based on an open source electronic platform, and it is composed of the MQ-137 ammonia detection
sensor, connected to the Croduino Basic 2 board. Sensitive material of the sensor is a metal oxide (SnO2). The change in the
ammonia concentration causes a change in the electrical resistance, which is measured and used for the ammonia
concentration estimation. Four measurements, with ammonia concentration ranging from 0-500 ppm were performed in the
controlled environment. With the purpose of the device calibration and repeatability test, along with a low-cost device, the
monitoring was carried out also with a professional device, Geotech GA5000 gas analyser. Using the information provided in
the sensor datasheet, together with the measured data, the measurement error of the low-cost device was assessed. After
calibration, ammonia concentrations measured using a low-cost device, are consistent with those obtained from the
professional device.
Keywords: MQ-137 sensor, Arduino, ammonia concentration, photocatalyst

Received: 06.06.2019. / Revised: 05.07.2019.

Accepted: 08.07.2019.
Published online: 15.07.2019.

Original scientific paper

1. INTRODUCTION

Over the past several years, open hardware concepts and portable sensors have undergone a period of rapid
development. During the time, open hardware concepts, i.e. Arduino platform, are becoming gradually accepted
within scientific and educational community. Availability, low price, steep learning curve and abundance of
available examples makes this concept very convenient for a non-professional to enter the world of hardware and
software development. The focus of this paper is put on ammonia concentration monitoring during the
photocatalysis process, using a low-cost MQ-137 sensor. Different examples of MQ-137 sensor application in
various scientific areas are available, such as, spoiled meat classification (Kartika et al. 2018), construction of
electronic nose for characterization of saffron aroma (Kiani et al. 2017) or in detection of seasonal allergic
rhinitis in exhaled breath (Saidi et al. 2015). The MQ-137 sensor is also used for monitoring organic abundance
in soil to advocate food security and sustaining the farming system (Dorji et al. 2017).
Increasing awareness of harmful consequences of persistent organic pollutants in water and air makes
development of photocatalysis for their degradation one of significant goals in environmental engineering. A
specific problem of air pollution over agricultural land is ammonia (NH3) and other nitrogen gases. Increasing
the use of pesticides and fertilizers increases the concentration of VOCs and NH3 in the air as agricultural
areas are recognized as collective emission sources. In accordance with the applicable IPPC Directive
(Brinkmann et al. 2016), photocatalytic oxidation is cited as an alternative best available technique (NRT),
effective for the removal of gases from the air. With this type of technology, it is possible to remove the
following pollutants: VOC, unpleasant smells, H2S, NH3, amines and mercaptans, NOX, SOX, CO and O3.
(Gaya and Abdullah 2008). It is a process that uses the synergistic interaction of light radiation, catalysts -
semiconductors and oxidative species to oxidize ammonia to N2 and to break down organic pollutants in fluids
to H2O and CO2 other inert products depending on the chemical composition of the pollutant. The basic
advantage of this method is the necessary conditions that are easy to satisfy, which are the presence of UV or
VIS electromagnetic radiation, atmospheric oxygen and water. All of these conditions already exist in the
environment, it is only necessary to provide photocatalytically active material and photocatalytic oxidation
can begin. It is applied for the purification of water and air in the industry sector, i.e. in waste treatment plants,
fermentation processes, coating applications and the food industry (Boyjoo et al. 2016).
The application of this indoor air purification technology has a much higher potential for
commercialization, especially for the removal of VOCs. and several recents patents can be found to confirm
the applicability (Wang et al. 2007; Pingjian 2019) For the purification of polluted air, photocatalytic reactors

22 EnvEng-IO (2019) / Vol.6 / No.1

Petric et al. Ammonia concentration monitoring using Arduino platform

with the immobilized photocatalyst applied to the carrier (surface) must be properly designed and placed
directly on the source of emission. Due to variations in emissions, concentration of pollutants in air should be
continuously monitored before and after the photocatalytic purification to control the process efficiency and
to estimate the positive impact on the air quality; hence the necessity for the development of simple analytical
platform based on portable sensors and open hardware concepts.

2. EXPERIMENTAL SECTION

2.1. Experimental setup for measurement concentration of the NH3

Low-cost sensor was calibrated in a controlled environment that is made of a plastic container connected to the
ammonia supply, Figure 1. Calibration was conducted using a professional device Geotech GA5000, with the
declared ammonia concentration measurement accuracy of 10 % (Geotech manual). For this purpose, the MQ-137
sensor was placed inside the container, and professional device was connected to it. Along with concentration
measurements, temperature and relative humidity inside container was monitored by DHT-22 sensor. The MQ-
137 sensor, which was used in these experiments is fabricated by Zhengzhou Winsen Electronics Technology Co.,
Ltd. Measurable ammonia concentration is within a range between 5 and 500 ppm. The load resistance is
adjustable, and heating resistor is approx. 29 3 Ω. Measurement read outs were conducted using Arduino based
system, Croduino Basic 2 electronic board (https://e-radionica.com/). The output voltage from the sensor module
MQ-137 is converted to digital signal using a 10-bit analog to digital converter (ADC) which is a part of the board.
Ammonia solution, 110 𝜇𝐿 of ammonia (25 %, Kemika) diluted in 250 mL distilled water, is inside of
erlenmeyer flask that is directly connected to bottle. With help of air pump (Fluval Q2, HAGEN, 240 L/min)
ammonia solution is pumped to the bottle. Some time is needed to achieve balanced atmosphere inside of a system.
After achieving balance, readings are taken with both Geotech and Arduino system. Next, 110 𝜇𝐿 of ammonia is
added to flask and same process repeats until desired number of data is taken. The measured data were fitted by
means of least squares fitting program using Python 3.6 program language and limfit package (Newville et al
2014).

a) b) c) d) f)

Figure 1. The photo of the setup and schematic picture of the calibration setup showing a) fan b) erlenmeyer
flask c) controlled environment made of plastic container with sensors d) Croduino board f) computer.

2.2. Working principles of the MQ-137

MQ-137 is a semiconductor gas sensor which use the sensitivity of the metal oxide SnO2 to the concentration of
ammonia in the air. A detailed description principle of the work sensitive material of the sensor can be found in
the paper (Barsan & Weimar 2003). In this work, focus was on the complete board with the sensor, which can be
purchase on the market, the sensor module used in this work is presented in Figure 2. We found the examples of

EnvEng-IO (2019) / Vol.6 / No.1 23

Petric et al. Ammonia concentration monitoring using Arduino platform

using MQ sensor board in determining ammonia (Aswinth 2018) and butane (https://jayconsystems.com)
concentration in air. The starting point in understanding MQ sensors is the schematic circuit, Figure 2. The
variable resistor that changes its value according to the concentration of gas, in this case ammonia, it is built
between A and B pin and it is connected in serial with load resistor (RL). Between H and H pins it is built-in
another resistor used for the heater of the sensitive material. This is necessary for proper work of the sensor.

Figure 2. (left) The picture of Croduino Basic 2 board and MQ-137 sensor module. (right) The schematic
representation of the sensor circuit. Download from https://jayconsystems.com.

The voltage 𝑉 is measured on the 𝑅 resistor, but it was needed to derive a formula for variable resistor
(𝑅 ) between A and B pins. Using Ohm's Law, the current 𝐼 is written through 𝑅 and 𝑅 resistors as:

𝑉
𝐼 , (1)
𝑅 𝑅

where 𝑉 is the supply voltage. Again, using Ohm's Law the voltage on 𝑅 resistor is written as 𝑉 𝐼𝑅 and
by inclusion of an expression for the current from Equation 1, the formula for 𝑅 is gotten as:

𝑉
𝑅 1 𝑅 . (2)
𝑉

The sensitivity characteristic of the gas concentration in parts per million (ppm) according to the resistance of
the sensor is provided by the producer, Figure 3. The graph is on the log-log scale, where on the y-axis is shown
𝑅
the ratio of the resistors 𝑅 . The 𝑅 is the resistance of the variable resistor in the fresh air and according to
Equation 2 it can be rewritten as:

𝑉
𝑅 ℎ
1 𝑅 𝑘𝑜𝑛𝑠𝑡 𝑅 . (3)
𝑉

Figure 3. The sensitivity characteristics of the MQ-137 sensor provided by producer.

24 EnvEng-IO (2019) / Vol.6 / No.1

Petric et al. Ammonia concentration monitoring using Arduino platform

In the Figure 3, the ratio and the concentration are linearly related in the log-log graph. Therefore, by using
the equation for the line and changing variable 𝑥 and 𝑦 with log functions, the following equation is gotten:

𝑅
log 𝑎 log 𝑝𝑝𝑚 𝑏, (4)
𝑅

where constant 𝑎 and 𝑏 are slope of the line and y-axis intercept, respectively. The constant 𝑎 and 𝑏 can be
calculated from the Figure 3. In general, the idea is to measure voltage 𝑉 and calculate 𝑅 by using Equation
2, after Equation 4 is rewritten to get results in 𝑝𝑝𝑚. In this work, the concentration is independently measured
with Geotech, therefore 𝑉 dependence on change of the concentration is studied. For this reason, 𝑉 can be
derived from Equation 2 and 4 as follow

𝑘𝑜𝑛𝑠𝑡
𝑉 (5)
𝑝𝑝𝑚 𝑘𝑜𝑛𝑠𝑡

where 𝑘𝑜𝑛𝑠𝑡 and 𝑘𝑜𝑛𝑠𝑡 are combination of the constant 𝑉 , 𝑘𝑜𝑛𝑠𝑡 , and 𝑏.

3. RESULTS AND DISCUSSION

In order to estimate standard error of the read out voltage from the sensor module few measurements was taken
ℎ
of the voltage on the fresh air and 𝑉 0.4203 0.0006 𝑉 is gotten. In the next step, focus was on the
estimation of the error in the calibration measurements.
The measured concentration of ammonia from 10 to 500 ppm and corresponding sensor voltage are presented
in Figure 4 (left). The data are fitted with obtained function by Equation 5 and such fitting model reproduces
almost perfectly the measured data with the coefficient of determination 𝑅 0.999. The extracted constants form
fits are listed in Table 1. In photocatalysis experiment the concentration of ammonia goes until 200 𝑝𝑝𝑚, for this
reason measurements and analysis were repeated three times in this concentration range, Figure 4 (right). The
coefficient of determination in all three fitting procedure was higher than 𝑅 0.99. From the Figure 4 (right)
the small deviations between calibration curve are noticed for ammonia concentration below 50 𝑝𝑝𝑚 and excellent
repeatability for the concentration above 100 𝑝𝑝𝑚.

Figure 4. The voltage read from low cost sensor as a function of the concentration of ammonia measured by
Geotech. (left)The concentration of ammonia in range 0 -500 ppm. (right) Three measurements in range 0-
200 ppm of ammonia concentration. The insert figure represents the residual from the fitting line.

Table 1. The data extracted from the fitting procedure

MEASURMENT 𝑘𝑜𝑛𝑠𝑡 𝑘𝑜𝑛𝑠𝑡 𝑎

zero 0.225 0.006 0.074 0.002 0.752 0.009
first 0.17 0.03 0.06 0.01 0.96 0.06
second 0.09 0.03 0.03 0.01 1.1 0.1
third 0.04 0.01 0.017 0.006 1.3 0.1
all 0.23 0.03 0.080 0.009 0.76 0.05

EnvEng-IO (2019) / Vol.6 / No.1 25

Petric et al. Ammonia concentration monitoring using Arduino platform

If all measurements are treated together and analysis of the model function repeated, Equation 5 still
reproduces very good measurement data, with the coefficient of determination 𝑅 0.987. The experimental data
are represented in Figure 5. The error of the calibration measurements can be calculated using partial derivative
of Equation 5 and multiplied by the error of the Geotech concentration measurements as:

𝑘𝑜𝑛𝑠𝑡 ∙ 𝑎 ∙ 𝑝𝑝𝑚
𝛿𝑉 𝛿𝑝𝑝𝑚 ∙ (6)
𝑝𝑝𝑚 ∙ 𝑝𝑝𝑚 𝑘𝑜𝑛𝑠𝑡

The Equation 6 can be evaluated for certain concentration; for example, in case of 100 𝑝𝑝𝑚 the error is
𝛿𝑉 0.04 𝑉. This is two order of magnitude higher than the read-out error, but it is still relatively low.

Figure 5. The voltage read from low cost sensor as a function of the concentration of ammonia measured by
Geotech. The error bars of the calibration procedure are added on the fitting model.

When experimental data are compared with the data given from producer (Figure 2), certain deviation is
noticeable as it is shown on Figure 6. The average absolute deviation is ∆ 0.04. The resistor values around
100 𝑝𝑝𝑚 is 0.12, which is 33 %. Therefore, without calibration procedure at ammonia concentration of
100 𝑝𝑝𝑚 the deviation can be estimated to the 33 𝑝𝑝𝑚, which can be the main error in experiments with MQ-137
sensors. The cause of the deviation is difficult to locate, but some explanation can be proposed. The producer
reported data for the sensor not for the module which include few additional components (resistors, diode,
potentiometer) and this components produced systematic deviation.

Figure 6. Comparison of the experimental data and data given from producer

26 EnvEng-IO (2019) / Vol.6 / No.1

 Petric et al. Ammonia concentration monitoring using Arduino platform

4. CONCLUSION

As shown in this paper, results reproducibility in measuring ammonia concentration in air can be achieved
using a low-cost MQ-137 sensor. The sensor was calibrated by a professional device Geotech GA5000, with
declared full scale accuracy of 10% for ammonia concentration measurement. Using measured values, along with
professional device accuracy, a measurement error analysis was performed. Calculated error of 0.04 V at voltage
corresponding to the concentration of 100 ppm was obtained. Considering requirements of ammonia photocatalysis
process efficiency evaluation, calculated error is acceptable. It was also shown that sensor calibration parameters
declared by the producer should be verified prior measurement and calibrated if necessary. Despite the low price,
the MQ-137 sensor showed good overall performance during ammonia concentration measurements and with
Arduino system can be used to monitor photocatalysis of ammonia.

5. ACKNOWLEDGMENTS

This work has been fully supported by Croatian Science Foundation under the project IP-2018-01-5246 (dr.
Andreja Gajović, Institute Ruđer Bošković).

6. REFERENCES

Aswinth Raj (2019) Measuring PPM from MQ gas sensor using Arduino (MQ-137 Amonia)
https://circuitdigest.com/microcontroller-projects/arduino-mq137-ammonia-sensor
Barsan N and Weimar U (2003) Understanding the fundamental principles of metal oxide based gas sensors;
the example of CO sensing with SnO2 sensors in the presence of humidity. J. Phys.: Condens. Mat. 15 R813
Brinkmann T., Santonja G., Yükseler H., Roudier S., Sancho L. D. (2016) Best available techniques (BAT)
reference document for common waste water and waste gas treatment/management systems in the chemical
sector
Boyjoo Y., Sun H., Liu J., Pareek V. K., Wang S. (2016) A Review on Photocatalysis for Air Treatment :
From Catalyst Development to Reactor Design. Chem. Eng. J. 310(2), 537-559
Dorji U, Pobkrut T, Siyang S, Chaiyasit T, Choden P and Kerdcharoen T, (2017) Development of real time
smart electronic nose system for monitoring organic abundance in soil, 14th International Conference on Electrical
Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), Phuket,
2017, 163-166.
Gaya U I, Abdullah A H (2008), Heterogeneous photocatalytic degradation of organic contaminants over
titanium dioxide: A review of fundamentals, progress and problems, J Photoch Photobio C: Reviews, 9(1), 1-12
Geotech manual (2019), Available via https://www.geotechuk.com/site/wp-
content/uploads/2016/03/GA5000-datasheet-1-1.pdf Cited 1 May 2019
Jaycon Systems (2019), Available via https://jayconsystems.com/blog/understanding-a-gas-sensor Cited 1
May 2019
Kartika V S, Rivai M and Purwanto D (2018) Spoiled meat classification using semiconductor gas sensors,
image processing and neural network, Intern. Conf. Inf. and Com. Tec. (ICOIACT), pp. 418-423.
Kiani S and Minaei S and Ghasemi-Varnamkhasti M and Ayyari M (2017) An original approach for the
quantitative characterization of saffron aroma strength using electronic nose. Int J Food Prop, 20, S673-S683
Newville M, Stensitzki T, Allen D B, Ingargiola A (2014) LMFIT: Non-Linear Least-Square Minimization
and Curve-Fitting for Python. DOI:10.5281/zenodo.11813
Pingjian Z (2019) CN208389639 (U) - A glass fiber filter equipment for photocatalysis exhaust -gas treatment
equipment
Python Software Foundation. Python Language Reference, version 3.6. Available at http://www.python.org
Saidi T, Tahri K, Bari N E, Ionescu R and Bouchikhi B (2015) Detection of seasonal allergic rhinitis from
exhaled breath VOCs using an electronic nose based on an array of chemical sensors, 2015 IEEE SENSORS,
Busan, 1-4.
Wang S, Ang H M, Tade M O (2007) Volatile organic compounds in indoor environment and photocatalytic
oxidation: State of the art. Environ. Int 33(5), 694–705

EnvEng-IO (2019) / Vol.6 / No.1 27

View publication stats