Field calibration of a cluster of low-cost available sensors for air quality monitoring. Part A: Ozone and nitrogen dioxide

Abstract

The performances of several field calibration methods for low-cost sensors, including linear/multi linear regression and supervised learning techniques are compared. A cluster of ozone, nitrogen dioxide, nitrogen monoxide, carbon monoxide and carbon dioxide sensors was operated. The sensors were either of metal oxide or electrochemical type or based on miniaturized infra-red cell. For each method, a two-week calibration was carried out at a semi-rural site against reference measurements. Subsequently, the accuracy of the predicted values was evaluated for about five months using a few indicators and techniques: orthogonal regression, target diagram, measurement uncertainty and drifts over time of sensor predictions. The study assessed if the sensors were could reach the Data Quality Objective (DQOs) of the European Air Quality Directive for indicative methods (between 25 and 30% of uncertainty for O3 and NO2). In this study it appears that O3 may be calibrated using simple regression techniques while for NO2 a better agreement between sensors and reference measurements was reached using supervised learning techniques. The hourly O3 DQO was met while it was unlikely that NO2 hourly one could be met. This was likely caused by the low NO2 levels correlated with high O3 levels that are typical of semi-rural site where the measurements of this study took place.

Full text

Accepted Manuscript
Title: Field calibration of a cluster of low-cost available
sensors for air quality monitoring. Part. A: ozone and nitrogen
dioxide
Author: Laurent Spinelle Michel Gerboles Maria Gabriella
Villani Manuel Aleixandre Fausto Bonavitacola
PII: S0925-4005(15)00355-X
DOI: http://dx.doi.org/doi:10.1016/j.snb.2015.03.031
Reference: SNB 18224
To appear in: Sensors and Actuators B
Received date: 25-10-2014
Revised date: 17-3-2015
Accepted date: 19-3-2015
Please cite this article as: L. Spinelle, M. Gerboles, M.G. Villani, M. Aleixandre, F.
Bonavitacola, Field calibration of a cluster of low-cost available sensors for air quality
monitoring. Part. A: ozone and nitrogen dioxide., Sensors and Actuators B: Chemical
(2015), http://dx.doi.org/10.1016/j.snb.2015.03.031
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Page 1 of 18
Accepted Manuscript
1
Highlights:

We have tested a high number of commercial sensors on the same site, applying the same data treatment and
evaluation
 We have modelled sensor responses without the need for applying temperature and humidity corrections with
external sensors.
 We have found a solution to solve the O
3
/NO
2
interference on O
3
sensor without the need for NO
2
reference
measurements
 Even though the prediction period is 6 times longer than the calibration one, the quality of the O
3
neural network
predictions is satisfying.
 The O
3
Data Quality Objective of the Air quality Directive is met allowing a larger diffusion of this type of
measurement techniques for regulatory purposes.

Page 2 of 18
Accepted Manuscript
2
Field calibration of a cluster of low-cost available sensors for air quality
monitoring. Part. A: ozone and nitrogen dioxide.
Laurent Spinelle
a1
, Michel Gerboles
a2
, Maria Gabriella Villani
b
, Manuel Aleixandre
c
, Fausto Bonavitacola
d
a
European Commission, Joint Research Centre (JRC), Institute for Environment and Sustainability (IES), Air and Climate Unit, Via
Enrico Fermi 2749, 21027 Ispra VA, Italy, laurent.spinelle@jrc.ec.europa.eu and michel.gerboles@jrc.ec.europa.eu
b
ENEA, Agenzia Nazionale per le nuove tecnologie, l'energia e lo sviluppo economico sostenibile, Ispra (VA), Italy,
mariagabriella.villani@enea.it
c
Institute for applied Physics, CSIC, Madrid, Spain, manuel.aleixandre@gmail.com
d
Phoenix Sistemi & Automazione s.a.g.l., Muralto (TI), Switzerland, fausto.bonavitacola@ingpec.eu
Abstract
The performances of several field calibration methods for low-cost sensors, including linear/multi linear regression and
supervised learning techniques are compared. A cluster of ozone, nitrogen dioxide, nitrogen monoxide, carbon
monoxide and carbon dioxide sensors was operated. The sensors were either of metal oxide or electrochemical type or
based on miniaturized infra-red cell. For each method, a two-week calibration was carried out at a semi-rural site
against reference measurements. Subsequently, the accuracy of the predicted values was evaluated for about five
months using a few indicators and techniques: orthogonal regression, target diagram, measurement uncertainty and
drifts over time of sensor predictions. The study assessed if the sensors were could reach the Data Quality Objective
(DQOs) of the European Air Quality Directive for indicative methods (between 25 and 30 % of uncertainty for O
3
and
NO
2
). In this study it appears that O
3
may be calibrated using simple regression techniques while for NO
2
a better
agreement between sensors and reference measurements was reached using supervised learning techniques. The hourly
O
3
DQO was met while it was unlikely that NO
2
hourly one could be met. This was likely caused by the low NO
2
levels
correlated with high O
3
levels that are typical of semi-rural site where the measurements of this study took place.
Highlights:
 We have tested a high number of commercial sensors on the same site, applying the same data treatment and
evaluation
 We have modelled sensor responses without the need for applying temperature and humidity corrections with
external sensors.
 We have found a solution to solve the O
3
/NO
2
interference on O
3
sensor without the need for NO
2
reference
measurements

Even though the prediction period is 6 times longer than the calibration one, the quality of the O
3
neural network
predictions is satisfying.

The O
3
Data Quality Objective of the Air quality Directive is met allowing a larger diffusion of this type of
measurement techniques for regulatory purposes.
Keywords:
gas sensors, validation, measurement uncertainty, multivariate linear regression, neural network, air quality Directive
1 Introduction
Compared to the reference methods defined in the Air Quality Directive [1], the use of low-cost gas sensors for
monitoring ambient air pollution would reduce air pollution monitoring costs and would also allow larger spatial
coverage especially in remote areas where monitoring with traditional facilities is cumbersome. However, the
calibration of low-cost sensors for monitoring air quality remains a challenge. The selectivity and stability of sensors
1
Corresponding author: tel: +39 0332 789572, laurent.spinelle@jrc.ec.europa.eu
2
Corresponding author: tel: +39 0332 785652, michel.gerboles@jrc.ec.europa.eu

Page 3 of 18
Accepted Manuscript
3
are generally found problematic [2,3,4]. Consequently, more sophisticated algorithms for quantifying air pollution are
being developed. Among the studied methods, the temperature cycle operation was shown to limit cross sensitivities
and aging of sensors [5] under laboratory conditions. This method is also relevant for the identification of organic
compounds [6]. Kamionka et al. reported the use of several metal oxides (MOx) sensors operated at different heating
temperature [7,8]. Theses multi-sensors were either calibrated against standard gas mixtures or using artificial neural
network under field conditions. The latter method resulted in mixed results either satisfactory for short periods or
generally weak for longer data series. Neural network calibration has been mainly implemented for the identification of
organic compounds and smell [9,10] or for monitoring compounds such as CO or CH
4
at high levels [11]. Few attempts
were made to use neural network for the calibration of sensors for monitoring in the low nmol/mol range [12,13]. One
of these studies looked at neural network calibration for benzene at nmol/mol levels [14]. However, the majority of the
studies mentioned above used the sole MOx-type sensors which are known to suffer from a lack of stability and long
response time [15]. A recent study describes a new real-time field calibration by comparing mobile sensor responses
with reference measurements of existing reference monitoring stations [16].
Recently, within the EURAMET MACPoll project [17], the performance of single commercial sensors has been
evaluated [18,19,20,21] according to a precise protocol [22]. This study produced large datasets of measurements for
several compounds under laboratory conditions and field campaigns. Such datasets were not previously available in
literature, especially considering the number of controlled parameters (NO
x
, O
3
, CO, SO
2
, CO
2
, temperature, relative
humidity, wind and pressure).
In this study, an analysis of the performance of different calibration models over a great number of O
3
and NO
2
sensors
tested in the same conditions is performed. The performances of these methods were compared taking as indicator their
resulting measurement uncertainty. It was then evaluated if the uncertainty could meet the Data Quality Objective
(DQO) of the European Air Quality Directive [1].
2 Material and methods
Experiments were carried out in collaboration with the European Reference Laboratory for Air Pollution (ERLAP) at
the EMEP station of the Joint Research Centre (45°48.881’N, 8°38.165’E). The station is located in a semi-rural area at
the NW edge of the Po valley (Italy) and is equipped with meteorological sensors (temperature, relative humidity, wind
and pressure) and reference gas analysers for NO
x
, O
3
, CO, CO
2
and SO
2
. These reference measurements were used for
data validation, comparison and data treatment of sensor responses.
Based on the evaluation and validation of low-costs sensors [18-21], several sensors were chosen to be grouped in a
clustered system able to detect O
3
, NO/NO
2
, CO and CO
2
. The best performing sensors showing the shortest response
time, the highest sensitivity and the best repeatability were selected.
2.1 Low-cost Sensors
The cluster consisted of 5 NO
2
sensors and 2 CO sensors, both electrochemical and metal oxide type, 1 NO and 2 O
3
electrochemical sensors and 2 infrared CO
2
sensors (see Table 1). For NO
2
, MOx and electrochemical sensors were
used in order to benefit from the different inherent cross-sensitivities of both types of sensors. The list of tested sensors
is presented in Table 1 with manufacturer and models information. All sensors were connected through NI DAQ boards
(NI USB 6009 and NI USB 6018 from National Instruments, USA) to our LabVIEW in-house designed DAQ software.
The periodicity of data acquisition was 100 Hz and measurements were averaged every minute without filtering. No
data treatment was applied during data acquisition. None of the sensors were calibrated before delivery. The sensors

Page 4 of 18
Accepted Manuscript
4
were enclosed into aluminium protective boxes and the evaluation boards were covered with Teflon tape to protect the
electronic and to avoid contamination of the sensor.
Table 1: List of clustered sensors
Manufacturer Sensor models Pollutant Number of sensors
αSense O3B4 O
3
1
Citytech O3_3E1F O
3
2
αSense NO2B4 NO
2
2
NO2_3E50 NO
2
2
Citytech
NO_3E100 NO 2
MICS-2710 NO
2
2
SGX-Sensotech
MICS-4514-NO2 NO
2
2
CairPol CairClip NO2 NO
2
2
Figaro TGS-5042 CO 2
SGX Sensortech MICS-4514-CO CO 2
Edinburgh Sensors Gascard NG CO
2
1
ELT Sensors S-100 CO
2
2
Two CairClip sensors, model NO2 ANA [23] were supplied by CAIRPOL (La Roche Blanche -France). CairClip is an
integrated system that includes an amperometric sensor, a dynamic air sampling, a patented filter, and an electronic
circuit which allows a direct real time display of the measured value and complete status with internal data logging.
Reliability of the measurement is achieved by limiting the effect of humidity variations by using a gas specific inlet
filter combined with dynamic air sampling system.
Citytech sensors (Life Safety Germany GmbH, City Technology, Bonn, Germany) consist of 3 Electrodes amperometric
sensors with organic electrolyte. Two O
3
sensors (model O3_3E1F [24]), two NO
2
sensors (model NO2_3E50 [25]) and
two NO sensors (NO_3E100 [26]) were tested. Each sensor was mounted on a Citytech evaluation board that converts
the raw sensor signal voltage, with the possibility to vary the bias potential, using various load, feedback resistors and
different levels of current amplification. The board was configured to give an output of 1V-100 nA with damping 10.
αSense sensors were supplied by αSense Ltd (Essex - United Kingdom). One O
3
sensor (model O3B4 - 4 electrodes
[27]) and two NO
2
sensors (NO2B4 - 4 electrodes [28]) were tested. The B4 type sensor is a 4 electrodes
electrochemical sensor designed for nmol/mol gas levels. As well as the normal Working, Reference and Counter
electrodes, B4 sensors include a 4th auxiliary electrode, which is used to correct for zero current changes. Each sensor
gave two signals, the 2
nd
one being the background signal of the auxiliary electrode that has to be subtracted to the
sensor raw response of the working electrode. Each sensor was mounted on a αSense test boards (αSense 4-electrodes
Individual Sensor Board (ISB) [29]).
Two models of SGX Sensortech (Neuchâtel-Switzerland) sensors were tested in this study: the MICS 2710 for NO
2
[30]
and the MICS 4514 which is a combined NO
2
and CO sensor [31]. Both of these sensors consist of Metal Oxide
semiconductor sensors. While the MICS 2710 is used to detect NO
2
, the MICS 4514 can detect NO
2
and CO with two
different signal outputs. Three MICS-EK1 gas sensor evaluation kits [32] were used. The MICS 4514 were directly
soldered on the EK1 adapter by the manufacturer. Based on the manufacturer datasheet, the evaluation kit were
operated in manual mode on low heating for the NO
2
sensors (43 mW corresponding to a R
LOAD
of 1 kOhm) and high
heating for the CO sensors (76 mW correspond to a R
LOAD
of 256 kOhm).
TGS 5042-A00 sensors are manufactured by Figaro (Illinois – USA). It consists in a battery like electrochemical sensor
[33]. Two TGS 5042-A00 were mounted on two evaluation modules COM5042 able to convert the sensor output
current into a voltage [34]. .

Page 5 of 18
Accepted Manuscript
5
The carbon dioxide module S-100 is manufactured by TCC ELT (Environment Leading Technology, South Korea) and
is based on the NDIR (Non-dispersive Infrared) technology [35].
The OEM Gascard ® NG infrared gas sensor (0-1000 µmol/mol) is manufactured by Edinburgh Sensors (Lancashire -
UK). It is based on dual wavelength NDIR technology with automatic temperature and pressure corrections using real-
time environmental condition measurements. The CO
2
sensor uses an active sampling with a 1l/min pump.
2.2 Reference measurements
The measuring campaign was performed at the JRC –Ispra station from January 2014 until July 2014. Within this
period, only 5 months (from March to July) have been considered as valid and were taken into account in this study. As
described in [18-21], the mobile laboratory was equipped with routine analysers, meteorological and low cost sensors:

for meteorological parameters: ambient temperature, ambient relative humidity, ambient pressure and a 10m mast
for wind speed and wind direction
 for O
3
, a UV Photometric Analyser Thermo Environment 49C, a chemiluminescence Nitrogen Oxides Analyser
Thermo 42C for NO
2
/NO/NOx, a non-Dispersive Infrared Gas-Filter Correlation Spectroscopy Horiba APMA 370
for CO, a UV Fluorescent Analyser Thermo 43C TL for SO
2
. For CO
2
, we used a differential non dispersive
Infrared gas analyser Li-cor 6262
The gas analysers were calibrated in laboratory before the field tests and then they were checked every month. Field
checks were carried out using filtered zero air and span value. This one consisted of low concentration gas cylinder
certified by the Joint Research Centre which is accredited for these analyses. The gas cylinders used included
concentration levels of 50, 100 and 200 nmol/mol for NO/NOx, 50 nmol/mol for SO
2
, 1.3 µmol/mol for CO and
369 µmol/mol for CO
2
(uncertified). An ozone generator Thermo Environment 49 CPPS II model, delivering
100 nmol/mol of ozone, was used for the calibration checks of the ozone analyser. The highest observed calibration
drift during field tests consisted of 2.5% for NO/NO
2
and O
3
, 4.5% for CO, 2% for SO
2
and 1.5% for CO
2
. These drifts
are consistent with the uncertainty of the working standards used on field (about 3%) and the data quality objective of
reference measurements (15%) given in the European Directive for air quality. Therefore, no corrections of
measurements were made apart from the discarding values during maintenance and calibration checks.
3 Calibration Methods of the sensors and choice of variables
Three calibration methods were tested: simple linear regression (LR), multivariate linear regression (MLR) established
within MACPoll [18-21] and artificial neural networks (ANN) with raw, standardized (scaled) and calibrated sensor
responses.
3.1 Linear regression (LR)
For each sensor a calibration function was established by assuming the linearity of the sensor responses with reference
measurement for each pollutant. Ordinary linear regression was used with the minimization of square residuals of the
sensor responses versus reference measurements. The calibration functions were of the type Rs = a.X + b where Rs
represents the sensor responses and X is the corresponding reference measurements of air pollutant. Finally, the
measuring function, the converse equation X = (Rs – b)/a was applied to all sensor responses in order to predict air
pollutant levels.

Page 6 of 18
Accepted Manuscript
6
Among our dataset, the cases corresponding to the initial two weeks of valid measurements were used for calibration
(about 336 hourly values). The remaining data (about 90 % of the total dataset) were used for validation of the
measuring functions.
3.2 Multivariate linear regression (MLR)
The calibration was carried out using the least square method taking into consideration more than one explanatory
variables Y
i
. Models were established during the MACPoll studies (see Table 2). Coefficients a, b, c, d and e represent
calibration parameters extracted from the multi-linear regression; NO
2
, O
3
and NO stand for the reference
measurements. RH, T and H
2
O are respectively relative humidity, temperature and absolute humidity. As for the LR,
the calibration functions consisted of equations of the type Rs = f(X,Y
i
), where f(X,Y
i
) is a function of multiple
reference measurements. The resulting measuring function, X = f(Rs,Y
i
), was applied to each sensor. The same pattern
of calibration/validation sets as for linear regression was used for the multi linear regression.
Table 2: MLR models of single sensor
Sensor’s model Multivariate linear model
O3B4
O3_3E1F
NO2B4
NO2_3E50
MICS-2710
MICS-4514
CairClip NO2
3.3 Artificial neural network (ANN)
Artificial neural networks (ANN) are very sophisticated modelling techniques able to model extremely complex
functions well suited for the calibration of a cluster of sensors. In this study, two types of ANN architectures were
considered: radial based functions and multilayer perceptron (MLP). The former did not produce good results and is not
presented hereafter. The latter is the most popular network architecture used today, due originally to Rumelhart and
McClelland [36]. It consists of artificial units that receive a number of inputs (either from original data, or from the
output of other units in the neural network) and typically one hidden layer with hidden units. The weighted sum of the
inputs is formed to compose the activation of the unit. The activation signal is passed through an activation function to
produce the output of the unit. With a defined number of layers and number of units in each layer, the network's weights
and thresholds must be set in order to minimize the prediction error made by the network. This is the role of the training
algorithms which uses iterative techniques called backpropagation. We used the BFGS (Broyden-Fletcher-Goldfarb-
Shanno) algorithm, the most recommended techniques for training [36] to automatically adjust the weights and
thresholds in order to minimize this error. The error of a particular configuration of the network can be determined by
running all the training cases through the network, comparing the actual output generated with the desired or target
outputs. The differences are combined by an error function which gives the network error as a sum squared error, where
the individual errors of output units on each case are squared and summed together.
Among the whole dataset, the first week of valid measurements was used for training (about 168 hourly values). This
train period contains input and output variables assuring ten times as many cases as connections in the networks. When

Page 7 of 18
Accepted Manuscript
7
any sensor or reference measurements were missing the whole cases were discarded. The 2
nd
week of the measuring
campaign was used as test dataset in order to limit the over-learning of the backpropagation algorithm. It was used to
check progress against an independent dataset. As training progresses, the training error naturally dropped. However,
when the selection error stopped dropping, or indeed started to rise indicating that the network was starting to over fit
the data, the train was stopped. In this case, the number of hidden units was decreased.
The rest of the dataset (about 85 % of data) was used as a validation set to ensure that the results on the testing and
training set are real, and not artifacts of the training process. Moreover we iteratively conduct a number of experiments
with each configuration, retaining the best networks in terms of error of the testing set since the validation dataset
remains unknown at calibration time. In this way we also avoid being fooled if training locates a local minimum. Once
we had experimentally determined an effective configuration for our networks, we resampled and generated new
networks with that configuration.
In order to select input variables for the ANN, initially all sensors that were both correlated with the air pollutant of
interest and independent between each other were selected. Between sensor pairs of the same brand model, we have
selected the one giving the highest correlation with the pollutant of interest.
Then a sensitivity analysis was performed in order to discard sensors which were not significant for the NN
architectures. The sensitivity analysis used the Sums of Squares residuals (SSR) of the model, computing the ratios of
SSR for the full NN models out of SSR when the respective sensor was eliminated from the neural net. The parameters
that were not found significant were discarded one at a time and the training was repeated until all parameters were
found significant. As far as possible, we tried to avoid selecting meteorological sensor and reference measurements as
input of ANNs in order to only rely on low-cost sensors.
Three studies have been performed by changing the input data: raw, standardized and calibrated by MLR sensors data.
For the standardized values, the numeric data were scaled applying a
z transformation with means of zero and standard
deviation of 1. The output of ANN consisted in a set of at most 100 networks within 10 000 tested networks with
different MLP architectures (see Table 3).
Table 3: Lists of possible and selected inputs for the different ANN
3.4 Evaluation of calibration method
The evaluation of sensor performances took into account hourly values. It was carried out using only values predicted
by each calibration method. For each one, regression and difference-based analysis were conducted to evaluate their
performance. These included the calculation of the coefficient of determination (R²), comparing the slope and intercept
of the regression line with objective values of 1 and 0 respectively. The mean bias error (MBE) and the root mean
squared error (RMSE) standardised with the standard deviation of the reference measurements were used to draw a
target diagram [38].
Possible parameters and sensors Selected inputs after sensitivity analysis Selected architectures of ANN
O
3
O
3
: O3_3E1F; NO
2
: MICS-2710 and MICS-4514-
NO2, NO2_3E50; CO: MICS-4514-CO and TGS-
5042; CO
2
sensors avoided because of correlation
of O
3
and CO
2
; Absolute Humidity
O3_3E1F,MICS-2710
and TGS-5042
Number of networks selected: 100
Number of hidden layer: 3 to 10
Hidden activation: exp, logistic, tanh
Ouput activation: exp, identity, sine
NO
2
O
3
: O3_3E1F; NO
2
: NO2_3E50, MICS-2710,
MICS-4514-NO2 and NO2B4_107;
NO: NO_3E100; CO sensors avoided because of
correlation of NO
2
and CO; CO
2
: CO2_712;
Absolute Humidity
NO2_3E50, MICS-4514-NO2, O3_3E1F,
MICS-2710, MICS-4514-CO, NO2B4_107,
Absolute Humidity
Number of networks selected: 100
Number of hidden layer: 4 to 12
Hidden activation: exp, logistic, tanh, sine
Ouput activation: exp, identity, logistic, sine, tanh

Page 8 of 18
Accepted Manuscript
8
To assess the performance of each calibration method at individual air pollutant levels, we have also calculated the
measurement uncertainty using orthogonal regression of the estimated outputs against reference data. This uncertainty
was compared to the DQO for indicative method that corresponds to a relative expanded uncertainty of 30% for O
3
and
25 % for NO
2
at the limit value set by the European Directive. The estimation method of the uncertainty, which
corresponds to the relative expanded uncertainty U
r
, was carried out using Eq. 1 where
and are the slope and
intercept of the orthogonal regression and RSS the sum of square of residuals is calculated using Eq. 2.The details of
calculation of the orthogonal regression can be found in the Guide for the demonstration of equivalence [39].
Eq. 1
Eq. 2
Finally the drift over time of each calibration methods was plotted in order to evidence general trends. To ease the
detection of possible patterns by filtering noise, the daily residuals were plotted between reference measurements and
sensor predictions rather than the hourly ones.
4 Results
4.1 Presentation of dataset
The dataset was analysed using descriptive statistics. Special attention was paid to outliers, data magnitude and
variability in order to produce a full valid and complete dataset. The JRC EMEP station being a semi-rural site in a
humid region, it shows high relative humidity, low air pollutant levels for NO and NO
2
and relatively high O
3
levels
(see Figure 1). The high value of CO and NO
2
, respectively around 1.3 µmol/mol and 150 nmol/mol, are due to the
provisional location of the mobile laboratory near to a railroad crossing.
An important aspect of the dataset is the lack of independence between parameters. Usually, O
3
is highly correlated
with temperature and anti-correlated with relative humidity and CO
2
and to a lower extent with CO and NO
2
. As a
consequence, it will be difficult to estimate O
3
correctly using temperature, relative humidity and CO
2
as estimators.
Unfortunately, it is well known that temperature and humidity are important factors affecting sensor responses. An
alternative consists in using absolute humidity instead of temperature and relative humidity since absolute humidity is
not correlated with O
3
nor CO
2
. There is always a possibility that any statistical model able to correctly predict O
3
will
in fact predict CO
2
and benefit of the high correlation of CO
2
to predict O
3
. The same type of doubt exists for NO
2
and
CO since they are highly correlated.

Page 9 of 18
Accepted Manuscript
9
T RH AR P O
3
NO
2
NO CO CO
2
T 1.00
RH -0.60 1.00
AH 0.35 0.51 1.00
P -0.06 -0.05 -0.12 1.00
O
3
0.72 -0.72 -0.08 -0.15 1.00
NO
2
-0.38 0.17 -0.17 0.36 -0.50 1.00
NO -0.16 0.04 -0.11 0.19 -0.32 0.52 1.00
CO -0.45 0.22 -0.25 0.45 -0.49 0.80 0.40 1.00
CO
2
-0.61 0.62 0.05 0.28 -0.81 0.43 0.20 0.49 1.00
T: temperature, RH: relative humidity, AH: absolute humidity, P atmospheric pressure
Figure 1: Box whisker plots and correlation table pattern in the reference measurements at the JRC-EMEP station
4.2 Results of calibration methods
Table 4 gives all the parameters for both linear regression (LR) and multi-linear regression (MLR) parameters for every
single sensor. The regression analysis was performed using data of the first two weeks as calibration period. For each
sensor, the measuring equation (X = (Rs – B)/a or X = f(Rs,Y
i
)) was applied to the validation dataset.
Table 4: Performances of linear and multi-linear regression for calibration of single sensors.
Results were observed on the validation set (the quoted values represent standard uncertainties)
Sensors Calibration R² Validation R² Slope ± u Intercept ± u n
O3B4_20 (LR) 0.023 0.212 4.83 ± 0.05 -142.99 ± 1.97 2111
O3_3E1F (LR) 0.845 0.667 1.84 ± 0.02 -26.31 ± 0.86 2037
O3_3E1F (LR) 0.878 0.813 1.65 ± 0.02 -20.29 ± 0.62 1996
O3B4_20 (MLR) 0.503 0.479 2.33 ± 0.03 -37.97 ± 1.19 2070
O3_3E1F (MLR) 0.852 0.584 2.10 ± 0.03 -30.15 ± 1.05 1996
O3_3E1F (MLR) 0.945 0.824 1.49 ± 0.01 -12.03 ± 0.55 1955
NO2Cair1 (LR) 0.465 0.004 -60.94 ± 0.09 438.99 ± 0.74 2091
NO2Cair2 (LR) 0.240 0.036 -29.14 ± 0.12 229.86 ± 0.97 2091
NO2B4_107 (LR) 0.108 0.009 -73.28 ± 0.15 522.71 ± 1.23 2089
NO2B4_113 (LR) 0.230 0.002 -139.52 ± 0.13 991.24 ± 1.08 2089
NO2_3E50 (LR) 0.001 0.068 -247.53 ± 1.37 1727.00 ± 11.04 2055
NO2_3E50 (LR) 0.002 0.051 280.68 ± 1.36 -1977.94 ± 10.95 2053
MICS-2710 (LR) 0.206 0.131 8.13 ± 0.07 -53.83 ± 0.52 2089
MICS-2710 (LR) 0.200 0.126 8.48 ± 0.07 -56.45 ± 0.54 2089
MICS-4514- NO
2
(LR) 0.168 0.016 56.37 ± 0.16 -378.51 ± 1.24 2111
MICS-4514- NO
2
(LR) 0.269 0.203 6.18 ± 0.06 -40.00 ± 0.48 2089
NO2Cair1 (MLR) 0.745 0.021 16.17 ± 0.06 -97.84 ± 0.47 2091
NO2Cair2 (MLR) 0.585 0.004 32.50 ± 0.06 -205.92 ± 0.44 2091
NO2B4_107 (MLR) 0.351 0.026 17.07 ± 0.07 -108.36 ± 0.53 2048
NO2B4_113 (MLR) 0.679 0.086 6.96 ± 0.05 -33.59 ± 0.40 2048
NO2_3E50 (MLR) 0.768 0.078 10.97 ± 0.07 -78.28 ± 0.57 2014
NO2_3E50 (MLR) 0.562 0.062 16.01 ± 0.09 -108.18 ± 0.72 2012
MICS-2710 (MLR)
0.745
0.057 -477.45 ± 2.42 3176.44 ± 19.08 2098
MICS-2710 (MLR)
0.744
0.063 31407.64 ± 166.23 -207890.74 ± 1308.95 2098
MICS-4514- NO
2
(MLR)
0.525
0.010 577.84 ± 1.27 -3970.53 ± 10.01 2098
MICS-4514- NO
2
(MLR)
0.786
0.016 -852.21 ± 2.36 5692.90 ± 18.55 2098
Results have shown that sensors from the same type are slightly different. However, they mainly stay within the same
range. This shows that identical sensors tend to perform in a similar way even if some variance can be observed.

Page 10 of 18
Accepted Manuscript
10
Figure 2: Scatterplot of O3_3E1F calibrated sensor data
using the linear regression against reference measurements
Figure 3: Scatterplot of O3_3E1F calibrated sensor data using
the MLR against reference measurements
For example, Figure 2 gives the scatterplot of the LR predicted sensor values versus the O
3
reference measurements for
the 2
nd
O3_3E1F sensor. Red dots represent the values used during the calibration process and the blue ones represent
the predicted data based on the validation data set. This sensor was selected because it showed the best correlation
factor (R
2
= 0.88) during the calibration period. The scatterplot shows that the strength of association slightly decreased
during the validation period similar to the calibration with R² = 0.81 compared to R² = 0.88. During the MACPoll
project, it was observed that the O3_3E1F sensor was not affected by temperature or humidity but it suffered from a
cross-sensitivity to NO
2
. As needed in the MLR model (see Table 2), the needed inputs, both gaseous and
meteorological, have been selected within the reference measurements to maximize the benefits of the calibration. The
constants a, b and c of the sensor models were fitted during the first two weeks of valid measurements. Subsequently,
the equation was applied to the validation dataset. Figure 3 gives the scatterplot of the orthogonal regression of the
calibrated sensor data using the MLR method against the reference measurements. In this particular case, the use of
NO
2
reference values improved R
2
from 0.88 to 0.95 for the calibration period.
Table 5: Performances of ANN calibrations. Results were observed on the validation set. The quoted
values represent the standard uncertainties and n is the number of data used in the calculation
Sensors R² Slope ± u Intercept ± u n
O3 (ANN raw) 0.915 1.12 ± 0.01 -2.77 ± 0.29 1996
O3 (ANN std) 0.910 1.10 ± 0.01 -2.05 ± 0.29 2020
O3 (ANN MLR) 0.857 1.02 ± 0.01 1.36 ± 0.35 1979
NO
2
(ANN raw) 0.596 0.64 ± 0.01 5.81 ± 0.08 2064
NO
2
(ANN std) 0.560 0.70 ± 0.01 5.34 ± 0.09 2064
NO
2
(ANN MLR) 0.553 0.79 ± 0.01 6.50 ± 0.10 2064

Page 11 of 18
Accepted Manuscript
11
Figure 4: Target diagram for ANNs with raw, scaled and modeled inputs,
LR and MLR for NO
2
and O
3
calibration methods. A few NO
2
LR and MLR
sensor fell outside the limits with values higher than 7
Table 3 gives the lists of all inputs before the sensitivity analysis and the selected set of inputs. Table 5 gives the
regression parameters fitted during the validation period. Results are given for the three types of input data: raw,
standardized (std) and calibrated with MLR data (MLR). For the 3 types of data, the same list of inputs has been
respected in order to be able to compare them. The difference observed in the number of data used for the calculation is
mainly due to the manual validation of data performed in order to remove artefacts and wrong values.
Figure 4 gives the target diagram for LR, MLR and ANNs calibration methods for both gaseous species. This type of
diagram [38] was used to evaluate the time trends of the sensor predictions (E) and the reference measurements (M).
This kind of plot represents on the X-axis the centered root mean square error (CRMSE) normalised by the standard
deviation of reference mea
surements (σ
0
), which is an indicator of the modelled random error. The Y-axis represents
the correlation of the coefficient R normalised by σ
0
, which symbolises the systematic bias. The distance between each
point and the origin represents the root mean square error (RMSE). Finally, the circle area corresponds to the area of
acceptance and stands for points where the model random error is equivalent to the variance of the observations. Data
inside this circle indicate a positive correlation between modelled and observed values.
Based on Eq. 1, the relative expanded uncertainty (U
r
) was plotted against O
3
and NO
2
. Figure 5 shows U
r
versus O
3
reference measurements for the best O3_3E1F calibrated by LR, MLR and by ANN raw data, ANN scaled data or ANN
MLR data. Figure 6 shows U
r
for NO
2
. Only the plots of ANNs methods appear since the ones for LR and MLR fell
outside the y-axis. Finally, Figure 7 gives the times series of the O
3
and NO
2
residuals between reference measurements
and sensor predictions using LR, MLR and ANNs calibration methods.

Page 12 of 18
Accepted Manuscript
12
Figure 5: U
r
of the different calibration models
versus reference data of O
3
Figure 6: U
r
of the different calibration models
versus reference data of NO
2

Page 13 of 18
Accepted Manuscript
13
Figure 7: Drifts of calibration methods for single sensors (at left for linear and multi-liner regression)
and clusters of sensors (at right for artificial neural networks - ANN)
5 Discussion
Considering the best O
3
sensors, the coefficients of determination of the calibration dataset were high for LR calibration
and slightly higher for the MLR method, both methods resulting in a high R² for the validation dataset. However in both

Page 14 of 18
Accepted Manuscript
14
cases, the slope (about 2) and intercept of the orthogonal regression were respectively different from 1 and 0. All these
indicators are noticeably much better for the ANNs methods: higher R² for the validation dataset, slope nearer from 1
and intercept of a few nmol/mol.
Similarly, the target diagram shows that the ANNs result both in a lower bias (shown on the y-axis) and lower unbiased
RMSE (called centered root-mean-square error, CRMSE, shown on the x-axis) than LR and MLR. Moreover, LR and
MLR symbols fall generally outside the target circle, called efficiency score, evidencing RMSE up to 2 fold higher than
the standard deviation of reference measurements.
The use of ANN for calibration purpose appears to be the most efficient in terms of uncertainties. In fact, the O
3
DQO is
only reached with ANNs for concentrations higher than 35 nmol/mol, which is a lower level than the limited value of
the Directive. Moreover, the ANNs based on modelled inputs performed slightly worse at low level and reach the 30%
DQO at about 45 nmol/mol. The ANNs based on raw and standardized inputs showed identical Ur plots. Consequently,
the easier ANNs inputs based on raw or standardised ANNs should be preferred. The
Ur plots for LR and MLR show
higher values exceeding the DQO. These plots also show a positive trend towards high O
3
, showing the effect of large
slopes of the orthogonal regression (see Table 4). Figure 7 gives evidences of a slight drift of the calibration methods
over time of about 5 nmol/mol over nearly 4 months for ANNs, and about 20 nmol/mol for LR and MLR. While the
ANNs with the raw, scaled and MLR input results in similar drifts and constant noise, MLR showed slightly higher drift
and noise than LR.
It should be noted that in Figure 2 and 3 we have observed a slight overestimation of the predicted values. Actually, the
main error is due to the extrapolation of data higher than the maximum value observed in the calibration dataset. Both
methods are suffering from a lack of sensitivity regarding interfering effect such as temperature and relative humidity.
For NO
2
, none of the sensors has given a high R² for LR methods apart from one CairClip NO
2
with the calibration
dataset. Higher values where reached for MLR up to 0.75 for some sensors both of MOx and electrochemical types.
Unfortunately the R² was very low resulting in slop and intercept far from 1 and 0 respectively. It is likely that NO
2
sensors at a semi-rural site are affected by the low NO
2
levels and high correlation between O
3
and NO
2
to which the
sensor are generally sensitive. These observations are corroborated by the target diagram which shows that none of the
NO
2
sensors are within the efficiency score, with LR and MLR calibration being much higher than the ANNs methods.
However, within ANNs, as for O
3
the raw and scaled inputs resulted in lower biases than ANN-MLR.
No high NO
2
was measured during the measuring campaign, making it difficult to correctly apply the calibration
methods. For LR and MLR, Ur was too high to be visible within Figure 6. However, using ANNs, an interesting Ur of
around 20% was reached, this value increasing for NO
2
higher than 20 nmol/mol for the raw and scaled inputs. This
behaviour was not observed with the ANN-MLR which appeared to remain rather constant. One shall remember that
implementing the ANN-MLR requires a set of 7 sensors, of which 2 NO
2
MOx and 2 NO
2
electrochemical sensors, 1 O
3
electrochemical sensors, 1 CO electrochemical sensor and absolute humidity (therefore temperature and relative
humidity sensor). Moreover, all gas sensors were previously calibrated using correction models (Table 2) including
reference measurements for O
3
.
Finally, Figure 7 shows that LR and MLR methods appear to be without drift over time. Nevertheless, enormous noise
in particular for MOx sensors can be observed. As for O
3
, the ANNs methods appears to suffer from a low drift of about
4 nmol/mol in about 4 months for raw and scaled inputs and 6 nmol/mol for MLR inputs.

Page 15 of 18
Accepted Manuscript
15
Based on the requirement of the European Air Quality Directive for indicative methods, calibration would have been
performed only if the uncertainty had exceeded the Data Quality Objective (DQOs). For O
3
the DQO corresponds to an
uncertainty of 30% at the limit value of 60 nmol/mol, which means 18 nmol/mol. For NO
2
, the DQO is 25 nmol/mol,
which represents 25 % of uncertainty at the limit value of 100 nmol/mol. For both O
3
and NO
2
, ANN method shows a
maximum drift on residuals of 6 nmol/mol, three times lower than the DQO of O
3
.
6 Conclusions
Based on the measurement uncertainty estimated by orthogonal regressions of the sensor outputs versus reference data,
the most suitable calibration method appeared to be ANN using raw or scaled sensor inputs. Simple LR and MLR have
shown to produce the highest measurement uncertainty. While ANN with MLR inputs needed reference data for
calibration of most sensors, ANN with raw/scaled data, using only 3 sensors of different types (1 O
3
chemical, 1 NO
2
resistive sensor and 1 CO electrochemical sensor), were able to solve the main interferences of the O
3
sensor.
In general, it was shown that the ANN method increased the strength of association between estimated and reference
data (higher R² and lower CRMSE). Moreover, it also allowed the decrease of the bias to reference data, with the slope
and intercept of orthogonal regression being respectively nearer to 1 and 0.
It is likely that by combining different type of sensors, like electrochemical O
3
and NO
2
MOx sensors for example, the
ANN can solve the cross sensitivity issues from which suffers the major part of sensors. We have also observed that the
humidity/temperature dependence was also corrected, without the needs of such measurements. We suppose that it is
linked with the difference of influence of these parameters on both types of sensors. Finally, we showed that using a
cluster of sensors for calibration purpose, the data quality objectives of the European Directive for indicative methods
could be met for O
3
(uncertainty, U
r
, of 30 %) at semi-rural stations. Formal conclusions on the possibility to meet the
DQO for NO
2
should be evaluated at higher levels typical for urban environments.
Acknowledgements
The authors wish to acknowledge the collaboration of our JRC colleagues C. Grüning and G. Manca for their
contribution with CO
2
measurements and F. Lagler, N. R. Jensen and A. Dell’Acqua for carrying out air pollution
measurements. This study was carried out within the EMRP Joint Research Project ENV01 MACPoll. The EMRP is
jointly funded by the EMRP participating countries within EURAMET and the European Union.

Page 16 of 18
Accepted Manuscript
16
References
[1] Directive 2008/50/EC of the European Parliament and the Council of 21 May 2008 on ambient air quality and cleaner air for Europe
[2] H. Nakagawa, S. Okazaki, S. Asakura, K. Fukuda, H. Akimoto, S. Takahashi, and S. Shigemori, “An automated car ventilation system”, Sensors
and Actuators B: Chemical, vol. 65, no. 1–3, pp. 133–137, Jun. 2000.
[3] M. C. Carotta, G. Martinelli, L. Crema, M. Gallana, M. Merli, G. Ghiotti, and E. Traversa, “Array of thick film sensors for atmospheric
pollutant monitoring”, Sensors and Actuators, B: Chemical, vol. 68, no. 1, pp. 1–8, 2000.
[4] C. Pijolat, C. Pupier, M. Sauvan, G. Tournier, and R. Lalauze, “Gas detection for automotive pollution control”, Sensors and Actuators B:
Chemical, vol. 59, no. 2–3, pp. 195–202, Oct. 1999.
[5] M. Lösch, M. Baumbach, and A. Schütze, “Ozone detection in the ppb-range with improved stability and reduced cross sensitivity”, Sensors
and Actuators B: Chemical, vol. 130, no. 1, pp. 367–373, 2008.
[6] A. Gramm and A. Schütze, “High performance solvent vapor identification with a two sensor array using temperature cycling and pattern
classification”, Sensors and Actuators B: Chemical, vol. 95, no. 1–3, pp. 58–65, Oct. 2003.
[7] M. Kamionka, P. Breuil, and C. Pijolat, “Calibration of a multivariate gas sensing device for atmospheric pollution measurement”, Sensors and
Actuators B: Chemical, vol. 118, no. 1–2, pp. 323–327, Oct. 2006.
[8] M. Kamionka, P. Breuil, and C. Pijolat, “Atmospheric pollution measurement with a multi-materials sensing device”, Materials Science and
Engineering: C, vol. 26, no. 2–3, pp. 290–296, Mar. 2006.
[9] D. B. Reza Nadafi, S. N. Nejad, M. Kabganian, and F. Barazandeh, “Neural network calibration of a semiconductor metal oxide micro smell
sensor”, in 2010 Symposium on Design Test Integration and Packaging of MEMS/MOEMS (DTIP), pp. 154–157, 2010.
[10] M. Baumbach, A. Sossong, H. Delprat, K. Soulantica, A. Schuetze, A. Maisonnat, and B. Chaudret, “New micro machined gas sensors
combined with intelligent signal processing allowing fast gas identification after power-up”, Proceedings Sensor 2005, vol. 2, pp. 91–96.
[11] G. Huyberechts, P. Szecówka, J. Roggen, and B. W. Licznerski, “Simultaneous quantification of carbon monoxide and methane in humid air
using a sensor array and an artificial neural network”, Sensors and Actuators B: Chemical, vol. 45, no. 2, pp. 123–130, Dec. 1997.
[12] S. De Vito, G. Di Francia, and L. Martinotto, “Neural calibration of portable multisensor device for urban atmospheric pollution measurement”,
in Sensors and Microsystems, 0 vols., World Scientific, pp. 283–290, 2008.
[13] K. Brudzewski and S. Osowski, “Gas analysis system composed of a solid-state sensor array and hybrid neural network structure”, Sensors and
Actuators B: Chemical, vol. 55, no. 1, pp. 38–46, 1999.
[14] S. De Vito, E. Massera, M. Piga, L. Martinotto, and G. Di Francia, “On field calibration of an electronic nose for benzene estimation in an urban
pollution monitoring scenario”,
Sensors and Actuators B: Chemical, vol. 129, no. 2, pp. 750–757, Feb. 2008.
[15] L. Spinelle, M. Aleixandre and M. Gerboles, “Comparison between the performances of commercially available low cost sensors for the
monitoring of O
3
and NO
2
in ambient air”. Unpublished.
[16] D. Hasenfratz, O. Saukh, and L. Thiele, “On-the-Fly Calibration of Low-Cost Gas Sensors”, in Wireless Sensor Networks, G. P. Picco and W.
Heinzelman, Eds. Springer Berlin Heidelberg, pp. 228–244, 2012.
[17] http://macpoll.eu
[18] L. Spinelle, M. Gerboles, and M. Aleixandre, “Report of laboratory and in-situ validation of micro-sensor for monitoring ambient O12:
CairClipO3/NO2 of CAIRPOL (F).”, Publications Office of the European Union, Luxembourg, EUR 26373, 2013.
[19] L. Spinelle, M. Gerboles, and M. Aleixandre, “Report of laboratory and in-situ validation of micro-sensor for monitoring ambient air pollution -
NO9: CairClipNO2 of CAIRPOL (F)”, Publications Office of the European Union, Luxembourg, EUR 26394, 2013.
[20] L. Spinelle, M. Gerboles, and M. Aleixandre, “Report of laboratory and in-situ validation of micro-sensor for monitoring ambient air - Ozone
micro-
sensors, αSense, model B4 O3 sensors”, Publications Office of the European Union, Luxembourg, EUR 26681, 2013.
[21] L. Spinelle, M. Gerboles, and M. Aleixandre, “Report of laboratory and in-situ validation of micro-sensor for monitoring ambient air - NO2:
NO2-B4 sensor of αSense (UK)”, unpublished.
[22] L. Spinelle, M. Aleixandre, and M. Gerboles, “Protocol of evaluation and calibration of low-cost gas sensors for the monitoring of air
pollution”, Publications Office of the European Union, Luxembourg, EUR 26112, 2013.
[23] Technical Data Sheet CairClip NO2 (preliminary version)
http://www.cairpol.com/index.php?option=com_content&view=article&id=72&Itemid=151&lang=en
[24] Ozone Sensor, Citytech O3 3E 1 F, Rev. 11/2011, Data sheet, http://www.citytech.com/PDF-Datasheets/o33e1f.pdf
[25] Nitrogen Dioxide Sensor, Citytech NO2 3E 50, Rev. 11/2011, Data sheet, http://www.citytech.com/PDF-Datasheets/no23e50.pdf
[26] Nitric Oxide Sensor, Citytech NO 3E 100, Rev. 11/2011, Data sheet, http://www.citytech.com/PDF-Datasheets/no3e100.pdf
[27] O3B4 Ozone Sensor 4-Electrode, data sheet http://www.alphasense.com/WEB1213/wp-content/uploads/2013/11/O3B4.pdf
[28] NO2B4 Nitrogen Dioxide, 4_Electrode, Data sheet, http://www.alphasense.com/WEB1213/wp-content/uploads/2013/11/NO2B4.pdf
[29] http://www.alphasense.com/WEB1213/wp-content/uploads/2014/02/ISB.pdf
[30] http://www.cdiweb.com/datasheets/e2v/mics-2710.pdf
[31] http://www.sgxsensortech.com/content/uploads/2014/08/MICS-%E2%80%93-4514-Carbon-Monoxide-+-Nitrogen-Dioxide-dual-sensor.pdf
[32] http://www.sgxsensortech.com/content/uploads/2014/07/MICS-EK1-Evaluation-Kit.pdf
[33] http://www.figarosensor.com/products/5042pdf.pdf
[34] http://www.figaro.co.jp/en/product/docs/com5042_product_information_rev03.pdf
[35] http://tccelt.co.kr/2012/kor/pdf/S-100/DS_S-100_Rev2.7.pdf
[36] D. E. Rumelhart, J. L. McClelland, and PDP Research Group, Eds.,
Parallel Distributed Processing: Explorations in the Microstructure of
Cognition. Cambridge, MA, USA: MIT Press, 1986.
[37] C. M. Bishop, Neural Networks for Pattern Recognition. New York, NY, USA: Oxford University Press, Inc., 1995.
[38] J. K. Jolliff, J. C. Kindle, I. Shulman, B. Penta, M. A. M. Friedrichs, R. Helber, and R. A. Arnone, “Summary diagrams for coupled
hydrodynamic-ecosystem model skill assessment”,
Journal of Marine Systems, vol. 76, no. 1–2, pp. 64–82, Feb. 2009.
[39] Guide to the demonstration of equivalence of ambient air monitoring methods, Report by an EC Working, Group on Guidance

Page 17 of 18
Accepted Manuscript
17
Vitae
Laurent Spinelle
Laurent Spinelle is graduated of a PhD in materials for electronics. During his study, he acquired a wide experience in
the analysis and characterization of inorganic materials. He was entrusted with the development of a process making
possible the selective separation of two oxidizing gaseous species. From the elaboration of a gas sensor microsystem, he
is now involved in the development of a validation protocol for commercial sensors in the Institute for Environment and
Sustainability at the Joint Research Centre. This work gave him the opportunity to increase his knowledge in
Environmental Science and acquire skills in statistical analysis and uncertainties.
Michel Gerboles
Michel Gerboles studied Analytical Chemistry at the University of Bordeaux (FR). In 1990, he joined the European
Reference Laboratory for Air Pollution at the Joint Research Centre of the European Commission. His initial activities
were focused on the traceability of measurement techniques for inorganic air pollutants, primary reference methods for
air quality monitoring and the organisation of proficiency testing/intercomparison exercises in support to the European
policy on air quality. His activities moved to the speciation of PM (ions and heavy metals) and the evaluation of
indicative measurement methods like diffusive sampling. His current research interests include the validation of low-
cost gas sensors, and the development of automatic methods for the validation and extraction of quality indicators from
large ambient air monitoring datasets.
Maria Gabriella Villani
Maria Gabriella Villani received her Laurea in Physics in 1998 at Padua University (Italy), Master Degree in
Geography-Meteorology, 2002, at Indiana University, Bloomington Indiana (USA), and her PhD in Atmospheric
Sciences, 2006, at Carlo Bo University, Urbino (Italy). She worked as Research assistant at the National Research
Council, Institute of Atmospheric Sciences and Climate, and, as a Postdoc, at the Joint Research Centre, Institute of
Sustainability, in the Climate Change Unit. Currently, she is a researcher in ENEA, Italian National Agency for New
Technologies, Energy and Sustainable Development, UTTEI-SISP Unit.
Manuel Aleixandre
Manuel Aleixandre received his BSc in physics from the Universidad Autonoma of Madrid, Spain, in 1999 and his PhD
degree from the Universidad Nacional de Educacion a Distancia in 2007. His research interests include neural networks,
fuzzy logic, statistics and multivariate data analysis, calibrations, gas sensors based in fibre optic sensors,
instrumentation and development of electronic noses for several applications.
Fausto Bonavitacola
Fausto Bonavitacola is an expert in data acquisition, control and simulation devices, sensors and measure
instrumentation. He holds a degree in “Electronic engineering” with a specialization in "devices and circuits". He has
been involved for the last 15 years as an IT expert and/or an analyst/programmer in many projects and/or research
activities carried out in various laboratories at the Joint Research Centre in Ispra.

Page 18 of 18
Accepted Manuscript
18
Vitae
Laurent Spinelle
Laurent Spinelle is graduated of a PhD in materials for electronics. During his study, he acquired a wide experience in
the analysis and characterization of inorganic materials. He was entrusted with the development of a process making
possible the selective separation of two oxidizing gaseous species. From the elaboration of a gas sensor microsystem, he
is now involved in the development of a validation protocol for commercial sensors in the Institute for Environment and
Sustainability at the Joint Research Centre. This work gave him the opportunity to increase his knowledge in
Environmental Science and acquire skills in statistical analysis and uncertainties.
Michel Gerboles
Michel Gerboles studied Analytical Chemistry at the University of Bordeaux (FR). In 1990, he joined the European
Reference Laboratory for Air Pollution at the Joint Research Centre of the European Commission. His initial activities
were focused on the traceability of measurement techniques for inorganic air pollutants, primary reference methods for
air quality monitoring and the organisation of pro
ficiency testing/intercomparison exercises in support to the European
policy on air quality. His activities moved to the speciation of PM (ions and heavy metals) and the evaluation of
indicative measurement methods like diffusive sampling. His current research interests include the validation of low-
cost gas sensors, and the development of automatic methods for the validation and extraction of quality indicators from
large ambient air monitoring datasets.
Maria Gabriella Villani
Maria Gabriella Villani received her Laurea in Physics in 1998 at Padua University (Italy), Master Degree in
Geography-Meteorology, 2002, at Indiana University, Bloomington Indiana (USA), and her PhD in Atmospheric
Sciences, 2006, at Carlo Bo University, Urbino (Italy). She worked as Research assistant at the National Research
Council, Institute of Atmospheric Sciences and Climate, and, as a Postdoc, at the Joint Research Centre, Institute of
Sustainability, in the Climate Change Unit. Currently, she is a researcher in ENEA, Italian National Agency for New
Technologies, Energy and Sustainable Development, UTTEI-SISP Unit.
Manuel Aleixandre
Manuel Aleixandre received his BSc in physics from the Universidad Autonoma of Madrid, Spain, in 1999 and his PhD
degree from the Universidad Nacional de Educacion a Distancia in 2007. His research interests include neural networks,
fuzzy logic, statistics and multivariate data analysis, calibrations, gas sensors based in fibre optic sensors,
instrumentation and development of electronic noses for several applications.
Fausto Bonavitacola
Fausto Bonavitacola is an expert in data acquisition, control and simulation devices, sensors and measure
instrumentation. He holds a degree in “Electronic engineering” with a specialization in "devices and circuits". He has
been involved for the last 15 years as an IT expert and/or an analyst/programmer in many projects and/or research
activities carried out in various laboratories at the Joint Research Centre in Ispra.