Analytica Chimica Acta 823 (2014) 2531

Contents lists available at ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Electrochemical monitoring of citric acid production by

Aspergillus niger
Anna Kutya-Olesiuk, Urszula E. Wawrzyniak, Patrycja Ciosek, Wojciech Wrblewski *
Department of Microbioanalytics, Faculty of Chemistry, Warsaw University of Technology, Noakowskiego 3, 00-664 Warsaw, Poland

H I G H L I G H T S

G R A P H I C A L A B S T R A C T

 Citric acid fermentation process

(production) by Aspergillus niger.
 Qualitative/quantitative monitoring
of standard culture and culture
infected with yeast.
 Electronic tongue based on potentiometric and voltammetric sensors.
 Evaluation of the progress and the
correctness of the fermentation process.
 The highest classication abilities of
the hybrid electronic tongue.

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 20 January 2014
Received in revised form 24 March 2014
Accepted 25 March 2014
Available online 26 March 2014

Hybrid electronic tongue was developed for the monitoring of citric acid production by Aspergillus niger.
The system based on various potentiometric/voltammetric sensors and appropriate chemometric
techniques provided correct qualitative and quantitative classication of the samples collected during
standard Aspergillus niger culture and culture infected with yeast. The performance of the proposed
approach was compared with the monitoring of the fermentation process carried out using classical
methods. The results obtained proved, that the designed hybrid electronic tongue was able to evaluate
the progress and correctness of the fermentation process.
2014 Elsevier B.V. All rights reserved.

Keywords:
Citric acid production
Fermentation process
Aspergillus niger
Process monitoring
Electrochemical sensors
Hybrid electronic tongue

1. Introduction
Citric acid (2-hydroxy-1,2,3-propanetricarboxylic acid) is involved in the metabolic process of energy conversion such as the
citric acid cycle (Krebs cycle, tricarboxylic acid cycle) [1,2]. Due to
its acidic and antioxidant properties, the citric acid is widely
applied as an acidity regulator, preservative and avoring agent in
food industry [3] but also in chemical industry, pharmacy,

* Corresponding author. Tel.: +48 22 2345631; fax: +48 22 2345631.

E-mail address: wuwu@ch.pw.edu.pl (W. Wrblewski).
http://dx.doi.org/10.1016/j.aca.2014.03.033
0003-2670/ 2014 Elsevier B.V. All rights reserved.

medicine [46]. Although lemon juice is a rich source of citric

acid [7], its industrial-scale production by fermentation has been
established for decades. Many microorganisms were exploited to
obtain citric acid, however, lamentous fungus Aspergillus
(especially Aspergillus niger) are the main producers [8]. The
biochemical mechanism by which A. niger accumulates citric acid
has attracted much scientic attention [9]; nevertheless, further
methods of biosynthesis of this acid by other fungi such as Candida
lipolytica, Yarrowia lipolytica were also developed [10,11].
The fermentation production can be carried out by the surface
method (on trays), submerged method, on solid media and in
dynamic conditions [12]. Fermentation process of a substrate to

26

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

this carboxylic acid is related directly to the quality and quantity of

the sugar source [13]. The basic raw material for the production of
citric acid is molasses, which contains sucrose [14]. The addition of
inorganic salts is required for fungal growth and to reach high yield
of citric acid [15]. The possibility of cultivating the mycelium on
ethanol [16], whey [17] and glycerol [18] was also investigated.
The production of citric acid by A. niger is strongly inuenced by
the composition of the fermentation medium, dependent on
culture method and on the selection of optimal parameters
(optimum pH range, oxygen content, cultivation temperature and
time) [19]. Since the formation of small amounts of others
carboxylic acids (especially oxalic and gluconic acid) is evidenced
during the synthesis of citric acid, the side-products are removed
in a purication step. It should be taken into account that the
contamination by other microorganisms may limit the efciency of
citric acid production. Probably, this effect results from an
inadequately sterilized medium, which is contaminated by airspread microorganisms or contaminated equipment and water
[20]. Therefore, appropriate analytical methods of culture monitoring are selected to assure the proper breeding of Aspergillus
niger (pH, acidity, CO2 and O2 are the basic parameters controlled
during the fermentation [21]).
The quantication of citric acid produced during the fermentation process is commonly performed by HPLC [22], GCMS/29Si NMR
[23] and in some cases by pyrolysis mass spectrometry [24].
However, such methods require time-consuming steps of derivatization and sample pre-treatment. The multi-wavelength uorescence spectroscopy was proposed for the on-line monitoring of the
most relevant process variables in fungal cultures i.e. changes of CO2
concentration, cell dry mass and sugar concentration. Moreover, this
approach enabled the observation of spore germination, metabolic
activity, and quantitative product formation [25].
On the other hand, simple devices based on chemical sensor
arrays coupled with chemometric analysis electronic tongues
(ET) were also employed for the monitoring of biotechnological
processes [26], eliminating the necessity of sample preparation. As
an example, system based on an array of potentiometric sensors
has been proposed for the monitoring of batch fermentation
process of starting culture for light cheese production. The
reliability of such device was compared with classic HPLC
technique [27]. The control of alcoholic fermentation was
conducted using near and mid infrared spectroscopies combined
with electronic nose and electronic tongue. It was observed that
the electronic tongue enabled to detect the evolution of taste and
aroma prole [28]. A potentiometric electronic tongue was also
used as a screening tool for the analysis of different types of beer
(prediction of real extract, alcohol and polyphenol content, and
bitterness) [29]. Flow-through array of miniaturized ion-selective
electrodes was developed for the monitoring of periodic anaerobic
digestion on the basis of chemical oxygen demand and volatile
fatty acid analysis [30]. Another potentiometric sensor array was
tested during the batch fermentations of Escherichia coli i.e. the
increase in organic acids concentration, changes in media
composition were monitored and correlated with biomass growth
[31]. Hybrid electronic tongues, combining several types of (bio)
sensors (including potentiometric, voltammetric, conductivity,
optical), were also designed to analyse biotechnological samples.
Potentiometric and voltammetric electronic tongue was demonstrated to be a promising tool for the qualitative and quantitative
analysis of samples collected periodically during the beer
fermentation process [32]. Other hybrid ET systems were introduced for the quantication of grape variety in red wines [33] as
well as for the differentiation of fermented milks [34].
Finally, it should be emphasized, that only a single attempt was
made to control the fermentation process involving A. niger using
ET system. Nevertheless, the proposed potentiometric electronic

tongue has been applied for the analysis of simulated fermentation

complex media, where a simultaneous determination of ammonium, citrate and oxalate was achieved with good precision [35].
In this work, a novel hybrid electronic tongue system was
developed for the monitoring of the citric acid fermentation
process by A. niger. A system based on potentiometric and
voltammetric sensors was applied to assess the progress and
correctness of the fermentation process carried out in standard
fungal culture and culture infected by yeast.
2. Experimental
2.1. Preparation of samples
Mycelium of A. niger cultures were cultivated on synthetic
medium. Medium consisted of sucrose (20%), ionic additives:
NH4NO3 (0.3%), KH2PO4 (0.15%), MgSO4 (0.1%) in deionised water
(total volume 1000 mL). The initial pH of the medium was 3.6. The
medium was sterilized by triple sterilization at 100  C for 30 min.
Before vaccination, samples contained only the medium were
collected. Then the medium was inoculated with mycelium A. niger
(before the inoculation, the mycelium was proliferated on standard
Petri dish on agar medium). Breeding was carried out for three
weeks. After four days one of the cultures had been infected with
bakers yeast (Saccharomyces cerevisiae). The samples were
collected in 3 successive weeks, one sample per day during 5
days a week.
All samples were frozen and stored in 20  C. Just before the
measurement, the samples were subsequently thawed and ltered
through double lter paper and then analyzed using the hybrid
electronic tongue. All experiments were performed in bulk
solutions at room temperature. All studied samples were analyzed
in ve replicates.
2.2. Chemicals, membrane materials and potentiometric sensor array
preparation
All inorganic salts used, 1-morpholinoethanesulfonic acid
(MES), sulfuric acid and sodium hydroxide were of analytical
grade and were purchased from Fluka. The stock solutions of salts
(0.1 M) were prepared in deionised water.
High-molecular-weight poly(vinyl chloride) (PVC), plasticizers:
o-nitrophenyl octyl ether (o-NPOE), bis(2-ethylhexyl) sebacate
(DOS), lipophilic salts: potassium tetrakis [3,5-bis(triuoromethyl)
phenyl] borate (KTFPB) and tridodecylmethyl-ammonium chloride
(TDMAC), ionophores: 4-tert-butylcalix [4] arene-tetraacetic acid
tetraethyl ester (sodium ionophore X), tridodecylamine (hydrogen
ionophore I), were purchased from Fluka (Selectophore). The
ionophore dichloro(5,10,15,20-tetra-phenylporphyrinato) zirconium(IV) (Zr(IV)[TPP]Cl2) was synthesized via a metallation of the
free porphyrin (Porphyrin Products) using ZrCl4 (Fluka) [36].
Freshly distilled tetrahydrofuran (Fluka) was used as a solvent for
the membrane components.
The method of the membranes preparation and the electrodes
conditioning were the same as for the standard ISEs. The
membranes contained: 12 wt% appropriate ionophores, 20
50 mol% versus ionophore lipophilic salt, 6466 wt% plasticizer,
and 3133 wt% high-molecular-weight PVC (the membranes based
on an ion-exchanger contained 3.5 wt% TDMAC or 1 wt% KTFPB, see
Table 1). The membrane components (200 mg in total) were
dissolved in 2 mL of THF.
Flow-through sensor array consisted of 10 miniaturized ionselective electrodes (two electrode specimens were prepared for
each membrane composition). A detailed architecture of the
miniaturized ion-selective electrodes compatible with a single
ow-through module was presented in Ref. [30]. The constructed

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

27

Table 1
Components used for preparation of potentiometric sensors.
Electrode type Ionophore
+

Na

CH3COO

1.7 wt% sodium

ionophore X
1.0 wt% Zr(IV)[TPP]Cl2

H+

2.0 wt% tridodecylamine

Cationselective
Anionselective

Lipophilic salt

Plasticizer

Internal lling solution

Conditioning solution

0.15 wt%
KTFPB
0.15 wt%
KTFPB
1.75 wt%
KTFPB
1.00 wt%
KTFPB
3.50 wt%
TDMAC

65 wt% DOS

0.01 M NaCl

0.01 M NaCl 0.025 M KH2PO4/Na2HPO4 0.0001 M

CH3COONa

66 wt% oNPOE
64 wt% DOS

0.01 M NaCl/0.001 M CH3COONa

66 wt% DOS
64 wt % oNPOE

0.1 M NaCl/0.5 M KH2P04/0.25 M

Na2HPO4
0.01 M KCl
0.01 M NaCl

sensors were preconditioned for at least 24 h (the components of

internal lling and conditioning solutions were presented in
Table 1). To verify the performances of the electrodes, potentiometric selectivity coefcients were determined by the Separate
Solution Method (SSM), using 0.1 M solutions of corresponding
salts [37]. Satisfactory results, consistent with previous studies
[32], were obtained for all the electrodes.
2.3. Potentiometric and conductometric measurements
All measurements were carried out in ow-through mode with
cells of the following type: Ag, AgCl; KCl 3 MCH3COOLi 1 Msample
solutionmembrane internal lling solution; AgCl, Ag. Potentiometric multiplexer (EMF 16 Interface, Lawson Labs Inc., Malvern,
USA) was used for EMF measurements. The analyzed sample
solutions (2.5 mL) were diluted twice and pumped by a peristaltic
pump to the ow-through sensor array. The steady-state responses
of the sensors were recorded.
ORP and pH values of sample solutions (5 mL) were measured
using laboratory pH-meter (Metler Delta 350, Metler Toledo,
Switzerland). pH electrode (Cole Parmer EW-5991-61, Cole Parmer,
USA) and two redox electrodes: Au and Pt (ERAu-13 and ERPt-13,
Hydromet, Poland) were applied for such purpose. The measurements of the electrical conductivity were performed using a
portable conductivity meter (CPC-551, Elmetron, Poland) with a
conductivity cell (EPS-2, Eurosensor, Poland).
2.4. Voltammetric measurements
Voltammetry measurements were carried out using a CHI 1040
potentiostat (CH Instrument, Austin, USA) in the three electrode
arrangement, with silver/silver chloride (Ag/AgCl) as the reference,
platinum rod as the counter electrode and the glassy carbon disk
electrode (BASi, USA) as the working electrode. The reference
electrode potential was calibrated by using a ferricyanide electrode
process in 0.1 M phosphate buffer solution (pH 7.4). Argon was
used to deaerate the solutions. The working electrode was
sequentially mechanically polished with 1.0, 0.3 and 0.05 mm
alumina powder, and rinsed thoroughly with methanol and water.
In order to remove remaining powder, the electrode has been
sonicated for 5 min in methanol and water respectively. After this
procedure, the substrates were transferred to 1 mL sample
solutions diluted to 5 mL with 0.1 M phosphate buffer solution
(pH 7.4). Cyclic voltammograms were registered in the potential
range from 0.1 V to 1.5 V (scan rate 0.1 V s 1). Only the forward
scans were considered for further data processing.
2.5. Titrimetric analysis
1 mL sample solutions, collected from the fermentation culture
according to the scheme presented above, were diluted to 5 mL and

titrated with 0.1 M solution of sodium hydroxide (in the presence

of phenolphthalein) in order to quantify the acidity changes during
the process.
2.6. Data analysis
Data analysis and calculations were carried out in MatLab (The
MathWorks, Inc., Natick, USA) and Origin (Microcal Software, Inc.,
Northampton, USA) software. Sensors signals were recorded in two
different fermentation processes by fungus A. niger (standard
fungal culture and culture infected after four days by addition of
yeast). Chemical images of the samples were processed using
Principal Component Analysis (PCA), Partial Least Squares
Discriminant Analysis (PLS-DA) and Partial Least Squares (PLS).
3. Results and Discussion
3.1. Conducting the fermentation process
The fermentation process was carried out according to the
scheme described in the Section 2.1. Two types of A. niger cultures
i.e. the standard and infected culture were studied. The fermentation process can be divided into three phases:
 Phase I proliferation of mycelium and start of citric acid

production (after this phase, one of the culture was infected with
yeast),
 Phase II stabilization of citric acid production during the
standard process or interruption of the fermentation process in
the case of the infected culture,
 Phase III termination of the citric acid production in the
standard culture.
A classical breeding monitoring, involving the measurement of
the pH changes and the determination of the total acidity of the
sample solutions collected from both fermentation cultures, was
carried out. Successive production and accumulation of citric acid
resulted in the gradual decrease of pH (Fig. 1) and, after the rst
phase, almost linear growth of the total acidity of the samples
(Fig. 2) during the standard process. Moreover, according to our
expectations, the measured values of pH and total acidity were
similar for both cultures before the yeast infection. The introduction of the bakers yeast into the culture (4 days after starting the
culture), caused an abrupt increase of acid concentration in the
samples, whereas the pH was stabilized at a level of 2.5 (higher
than for standard culture). The results suggested that another
organic acid was produced after the culture infection, inuencing
the total acidity of the medium. However, a weaker acid than the
citric acid was probably produced since the pH remained
unchanged after the infection (pH was maintained by the citric
acid formed during the I phase).

28

[(Fig._1)TD$IG]

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

[(Fig._3)TD$IG]

phase I

phase II

phase III

-5

8.0x10

3.5
-5

3.0

infection

pH

infected culture
2.5

Current [A]

6.0x10

phase III

-5

4.0x10

phase II

-5

2.0

standard culture

2.0x10

phase I
0.0
0.1

1.5
0

100

200

300

400

Fig. 1. pH measurements of the samples collected during the standard and infected
fermentation process.

3.2. Monitoring of the citric acid production by an electronic tongue

An electronic tongue system based on miniaturized ionselective electrodes and classical 3-electrode electrochemical cell
with glassy-carbon electrode as working electrode was proposed
for the monitoring of the fermentation process. The steady-state
signals recorded for the ion-selective electrode array were used to
create a data matrix of the potentiometric electronic tongue PotET (i.e. based only on potentiometric data).
The forward scans of cyclic voltammograms registered in the
potential range from 0.1 to 1.5 V comprised 1400 points (exemplary
voltammograms recorded in the samples collected during the
standard culture were plotted in Fig. 3). Due to the large number of
data points, initial voltammetric signal compression involving
discrete wavelet transform was applied to reduce the amount of
data, while maintaining a waveform characteristics. Since the
results of the preprocessing depend on the selected parameters of

[(Fig._2)TD$IG]
phase I

phase II

phase III

infected culture

cHA [mol/L]

0.3

0.2

standard culture

infection
0.1

0.0

100

200

0.5

0.7

0.9

1.1

1.3

1.5

Potential [V]
Fig. 3. Exemplary cyclic voltammograms (forward scans) registered in the samples
collected during the standard fermentation process.

Time [h]

0.4

0.3

300

400

Time [h]
Fig. 2. Determination of the total acidity of the samples collected during the
standard and infected fermentation process.

the wavelet compression, ve wavelets from different families

(bior3.5, coif2, db6, rbio3.5, sym6) and different decomposition
levels were employed for data compression (as an example, the
level of decomposition a = 5 led to a reduction of the data from
1400 points to 54 coefcients, whereas 32 coefcients were
obtained after decomposition at a = 6). On the basis of visual
inspection of clusters separation on PLS plots, the parameters of
the nal data compression method (rbio3.5 wavelet, a = 5) were
selected for further calculations. The extracted coefcients (used as
inputs of the PLS model) allowed to form a data matrix of the
voltammetric electronic tongue Volt-ET (i.e. based only on
voltammetric data).
Finally, the data recorded by various electrochemical techniques were merged in a data matrix of the hybrid electronic tongue
Hybrid-ET (i.e. based on the mentioned above potentiometric
and voltammetric measurements as well as on pH and ORP values).
3.2.1. Qualitative monitoring of the fermentation process
Principal Component Analysis (PCA) technique enabled the
preliminary evaluation of the recognition ability of the hybrid
electronic tongue. The chemical images of the samples collected
during standard fermentation process created separated clusters.
Moreover, the clusters representing consecutive fermentation
samples were arranged according to the phase order (Fig. 4). On the
other hand, a clear linear separability between the clusters
corresponding to different phases of the infected culture was
not observed (Fig. 5). Especially, the chemical images of the
infected samples from the II and III phase were located close
together. Such results were well correlated with those gained by
the classical breeding control (see Figs. 1 and 2), where in contrast
to the standard culture the pH and total acidity changes were
negligible during the II and III phase of the infected culture.
Nevertheless, the proper classication shown in Fig. 4 did not
result solely from the pH measurement, since in the PCA model the
loadings of the outputs of all potentiometric sensors were
comparable, as well as the loadings of the coefcients obtained
after discrete wavelet transform of voltammetric data.
The results presented in Figs. 4 and 5 indicated the potential
usefulness of the developed hybrid electronic tongue for the
differentiation between standard and infected fermentation
process. Additionally, the hybrid system was able to estimate
the progress of the fermentation in the case of standard culture.
Next, the qualitative monitoring of citric acid production
was attempted using Partial Least Squares-Discriminant Analysis

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

[(Fig._4)TD$IG]

29

[(Fig._6)TD$IG]

10

8
6

LV 2 (9.1%)

PC 2 (18.3%)

4
2
0
-2

-6
-15

-10

-5

10

15

20

4
2
0
-2

phase I
phase II
phase III

-4

standard culture - phase I

standard culture - phase II+III
infected culture - phase I
infected culture - phase II+III

-4
-6
-20

25

-15

PC 1 (67.3%)

-10

-5

10

15

LV 1 (74.7%)

Fig. 4. PCA plot of chemical images of the samples collected during standard fungal
culture (data obtained for Hybrid-ET).

Fig. 6. PLS-DA plot of chemical images of the samples collected during the standard
and infected fermentation process (data obtained for Hybrid-ET).

(PLS-DA) technique. PLS-DA established the correlation of data

matrix with target matrix. In our case, a data matrix containing
chemical images of samples and its corresponding four-column
target matrix (coded information on sample classication) were
created. The rst two columns indicated the process duration,
whereas the two subsequent columns informed about the
occurrence of infection. On the basis of minimization of RMSE
value, 14 latent variables were chosen to create the PLS-DA model.
The results obtained by processing of the data by PLS-DA were
presented in Fig. 6 (for the sake of clarity, the chemical images
representing the phase II and III were plotted by the same marker).
In accord with the results presented in Figs. 1 and 2, the samples
corresponding to the rst phase formed one cluster for both
fermentation processes. Moreover, the chemical images of the
samples from the II + III phase exhibited different changes of the
position in the pattern space related to the occurrence of infection.
It should be also stressed, that the larger scattering of the chemical
images representing the II + III phase of the standard culture was in
good agreement with the previous conclusions. The obtained
results conrmed the suitability of the hybrid ET to assess the
correct course of fermentation.

monitoring of the studied fermentation process. Therefore, it

was necessary to build a proper target matrix embracing the
duration of the process (counted from the inoculation moment of
A. niger mycelium, enabling to estimate the progress of the
fermentation process) and the value of the total acidity (enabling
to estimate the correctness of the process). The obtained
measurement data were divided into train and test matrixes,
serving as inputs for PLS model (the data representing 4 replicates
for each sample formed the train matrix, whereas the fth sample
replicates were used to create the test matrix). The model
performance was characterized after building comparison graphs
and performing the linear tting of the expected data versus the
data predicted by the PLS model. Exemplary results provided by
the hybrid electronic tongue (Hybrid-ET) during the standard
culture, comparing the determined total acidity and the values
provided by the PLS model, were presented in Fig. 7. The values of
the parameters: slope (a), intercept (b) and determination
coefcient (R2) indicated that the hybrid system permitted a
correct quantitative modeling of the fermentation run. Moreover,
comparable linear trends and correlation coefcients were
attained for the train and the test set.

3.2.2. Quantitative monitoring of the fermentation process

The developed electronic tongue combined with Partial Least
Squares (PLS) technique was employed for the quantitative

[(Fig._7)TD$IG]
0.30

[(Fig._5)TD$IG]

0.25

PRED CHA [mol/L]

PC 2 (9.5%)

4
2
0

0.20
0.15
train
2
y=0.98x+0.04, R =0.978
test
2
y=0.94x+0.07, R =0.980

0.10

-2
phase I
phase II
phase III

-4
-6
-15

-10

-5

10

0.05
0.05

15

Fig. 5. PCA plot of chemical images of the samples collected during infected fungal
culture (data obtained for Hybrid-ET).

0.10

0.15

0.20

0.25

0.30

REAL CHA [mol/L]

Fig. 7. Model performances characterized after linear tting of the real values of the
total acidity for standard culture to the predicted data by the PLS model (data
obtained for Hybrid-ET).

30

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

Table 2
Parameters of linear tting of real and PLS-predicted total acidity and process duration for standard culture.
Total acidity

Process duration

Pot-ET

a
b
R2

Volt-ET

Hybrid-ET

Pot-ET

Volt-ET

Hybrid-ET

Train

Test

Train

Test

Train

Test

Train

Test

Train

Test

Train

Test

0.97
0.05
0.967

0.91
0.12
0.970

0.87
0.20
0.875

0.61
0.68
0.770

0.98
0.04
0.978

0.94
0.07
0.980

0.98
4.90
0.976

0.92
14.2
0.975

0.88
24.7
0.880

0.99
1.35
0.993

0.99
12.7
0.996

0.92
16.2
0.921

Table 3
Parameters of linear tting of real and PLS-predicted total acidity and process duration for infected culture.
Total acidity

Process duration

Pot-ET

a
b
R2

Volt-ET

Hybrid-ET

Train

Test

Train

Train

Test

0.99
0.02
0.991

1.01
0.00
0.993

0.89
0.27
0.886

0.83
0.66
0.765

0.99
0.02
0.993

Pot-ET
Train
1.02
0.02
0.992

Finally, the classication properties of the developed electronic

tongue were analyzed in details, determining the contribution of
the individual electrochemical techniques to the nal result. For
this purpose, data matrixes composed of chemical images of
samples created on the basis of separate potentiometric (Pot-ET)
and voltammetric (Volt-ET) measurements as well as the data
matrix corresponding to the hybrid system (Hybrid-ET) used above
were correlated with the target matrix. Tables 2 and 3 summarize
the obtained regression parameters (a, b and R2) for the standard
and infected culture, respectively. It should be noticed, that the
processing of the separate voltammetric data (Volt-ET) did not
ensure a proper quantitative monitoring of the fermentation
process the values of a, b, and R2 were not acceptable for the test
samples. Signicantly better results were obtained for Pot-ET (see
e.g. the total acidity determination for standard culture). In
general, the fusion of various electrochemical techniques, i.e. the
use of the hybrid electronic tongue, improved the classication
ability of the system for both standard and infected culture; similar
dependences were notices in the case of the prediction of process
duration and total acidity of the samples.
4. Conclusions
The qualitative and quantitative monitoring of the citric acid
fermentation process by A. niger was reported in this work. A
hybrid electronic tongue combining various electrochemical
sensors was exploited for the recognition of the samples collected
during standard and infected A. niger culture. The highest
classication abilities were achieved in the case of the hybrid
system (Hybrid-ET) i.e. when the chemical images recorded
applying various electrochemical techniques were processed.
The presented results led to the conclusion, that the developed
device, based on automated measurements with sensor array
systems, could be a simple and useful tool for the monitoring of the
progress and the correctness (e.g. detection of infection occurrence) of the fermentation process. Therefore, the proposed
approach can replace the commonly performed off-line process
monitoring, using high-performance liquid chromatography and/
or gas chromatography.
It should be also stated, that to date, the electronic tongue
systems were not applied in the monitoring and especially in the
detection of microbial infection during the citric acid production
by A. niger. The single approach described in Ref. [35] reported the
use of a potentiometric sensor array for the analysis of simulated

Volt-ET

Hybrid-ET

Test

Train

Test

Test

Train

Test

0.99
1.35
0.993

0.99
12.7
0.996

0.92
16.2
0.921

0.94
20.8
0.718

1.00
0.61
0.997

1.02
2.39
0.996

fermentation samples, which could not reect the processes

occurring in the real conditions and therefore the complex
composition of the medium.
Acknowledgement
This work has been nancially supported by the project LIDER/
17/202/L-1/09/NCBiR/2010.
References
[1] J.M. Lowenstein, Methods in Enzymology: Citric Acid Cycle, vol. 13, Academic
Press, Boston, 1969.
[2] P.A. Mayes, D.A. Bende, The citric acid cycle: the catabolism of Acetyl-CoA, in:
R.K. Murray, D.K. Granner, P.A. Mayes, V.W. Rodwell (Eds.), Harpers Illustrated
Biochemistry, The McGraw-Hill Companies, Inc., 2003, pp. 130135.
[3] F.H. Verhoff, Citric Acid, Ullmanns Encyclopedia of Industrial Chemistry,
Wiley-VCH, Weinheim, 2005.
[4] M. Berovic, M. Legisa, Citric acid production, Biotechnology Annual Review 13
(2007) 303343.
[5] C.R. Soccoll, L.P.S. Vandenberghe, C. Rodrigues, A. Pandey, New perspectives for
citric acid production and application, Food Technology and Biotechnology 44
(2006) 141149.
[6] H. Iwase, Use of citric acid as mobile phase and sample medium for the
determination of ascorbic acid by liquid chromatography/electrochemistry,
Current Separations 20 (2004) 127131.
[7] K.L. Penniston, S.Y. Nakada, R.P. Holmes, D.G. Assimos, Quantitative assessment of citric acid in lemon juice, lime juice, and commercially available fruit
juice products, Journal of Endourology 22 (2008) 567570.
[8] D.B. Xu, C.P. Madrid, M. Rohr, C.P. Kubicek, The inuence of type and
concentration of the carbon source on production of citric acid by Aspergillus
niger, Applied Microbiology & Biotechnology 30 (1989) 553558.
[9] M. Papagianni, Advances in citric acid fermentation by Aspergillus niger:
biochemical aspects, membrane transport and modeling, Biotechnology
Advances 25 (2007) 244263.
[10] H.S. Grewal, K.L. Kalra, Fungial production of citric acid, Biotechnology
Advances 13 (1995) 209234.
[11] S.K. Yalcin, M.T. Bozdemir, Z.Y. Ozbas, Citric acid production by yeasts:
fermentation conditions, process optimization and strain improvement, in: A.
Mendez-Vilas (Ed.), Current Research, Technology and Education Topics in
Applied Microbiology and Microbial Biotechnology, World Scientic, London,
2009, pp. 1381374.
[12] G.S. Dhillon, S.K. Brar, M. Verma, R.D. Tyagi, Recent advances in citric acid bioproduction and recovery, Food and Bioprocess Technology 4 (2011) 505529.
[13] A.R. Angumeenal, D. Venkappayya, An overview of citric acid production, LWT
Food Science and Technology 50 (2013) 367370.
[14] M. Pazouki, P.A. Felse, J. Sinha, T. Panda, Comparative studies on citric acid
production by Aspergillus niger and Candida lipolytica using molasses and
glucose, Bioprocess Engineering 22 (2000) 353361.
[15] A.A. Abou-Zeid, M.A. Ashy, Production of critic acid: a review, Agricultural
Wastes 9 (1984) 5176.
[16] G. Xie, T.P. West, Citric acid production by Asperigillus niger on the ethanol dry
milling coproduct thin stillage, Research Journal of Microbiology 2 (2007)
678683.

A. Kutya-Olesiuk et al. / Analytica Chimica Acta 823 (2014) 2531

[17] M.A. El-Holi, K.A. Al-Delaimy, Citric acid production from whey with sugars
and additives by Aspergillus niger, African Journal of Biotechnology 2 (2003)
356359.
[18] S. Papanikolaou, L. Muniglia, I. Chevalot, G. Aggelis, I. Marc, Yarrowia lipolytica
as a potential producer of citric acid from raw glycerol, Journal of Applied
Microbiology 92 (2002) 737744.
[19] T. Roukas, Citric and gluconic acid production from g by Aspergillus niger
using solid-state fermentation, Journal of Industrial Microbiology and
Biotechnology 25 (2000) 298304.
[20] S. Anastassiadis, I.G. Morgunov, S.V. Kamzolova, T.V. Finogenova, Citric acid
production patent review, Recent Patents on Biotechnolology 2 (2008) 107
123.
[21] S.J.A. Hesse, G.J.G. Ruijter, C. Dijkema, J. Visser, Measurement of intracellular
(compartmental) pH by 31P NMR in Aspergillus niger, Journal of Biotechnology
77 (2000) 515.
[22] J. Pintado, B.K. Lonsane, I. Gaime-Perrau, S. Roussos, On-line monitoring of
citric acid production in solid-state culture by respirometry, Process
Biochemistry 33 (1998) 513518.
[23] A. Ghassempour, S. Nojavan, Z. Talebpour, A.A. Amiri, N.M. Naja, Monitoring
of the fermentation media of citric acid by the trimethylsilyl derivatives of the
organic acids formed, Journal of Agricultural and Food Chemistry 52 (2004)
63846388.
[24] A. Ghassempour, N.M. Naja, A.A. Amiri, Determination of citric acid in
fermentation media by pyrolysis mass spectrometry, Journal of Analytical and
Applied Pyrolysis 70 (2003) 251261.
[25] M. Ganzlin, S. Marose, X. Lu, B. Hitzmann, T. Scheper, U. Rinas, In situ multiwavelength uorescence spectroscopy as effective tool to simultaneously
monitor spore germination, metabolic activity and quantitative protein
production in recombinant Aspergillus niger fed-batch cultures, Journal of
Biotechnology 132 (2007) 461468.
[26] M. Perisa, L. Escuder-Gilabert, On-line monitoring of food fermentation
processes using electronic noses and electronic tongues: a review, Analytica
Chimica Acta 804 (2013) 2936.

31

[27] K. Esbensen, D. Kirsanov, A. Legin, A. Rudnitskaya, J. Mortensen, J. Pedersen, L.

Vogensen, S. Makarychev-Mikhailov, Y. Vlasov, Fermentation monitoring using
multisensor systems: feasibility study of the electronic tongue, Analytical and
Bioanalytical Chemistry 378 (2004) 391395.
[28] S. Buratti, D. Ballabio, G. Giovanelli, C.M. Dominguez, A. Moles, S. Benedetti, N.
Sinelli, Monitoring of alcoholic fermentation using near infrared and mid
infrared spectroscopies combined with electronic nose and electronic tongue,
Analytica Chimica Acta 697 (2011) 6774.
[29] E. Polshin, A. Rudnitskaya, D. Kirsanov, A. Legin, D. Saison, F. Delvaux, F.R.
Delvaux, B.M. Nicola, J. Lammertyn, Electronic tongue as a screening tool for
rapid analysis of beer, Talanta 81 (2010) 8894.
[30] E. Witkowska, A. Buczkowska, A. Zamojska, K.W. Szewczyk, P. Ciosek,
Monitoring of periodic anaerobic digestion with ow-through array of
miniaturized ion-selective electrodes, Bioelectrochemistry 80 (2010) 8793.
[31] C. Turner, A. Rudnitskaya, A. Legin, Monitoring batch fermentations with an
electronic tongue, Journal of Biotechnology 103 (2003) 8791.
[32] A. Kutya-Olesiuk, M. Zaborowski, P. Prokaryn, P. Ciosek, Monitoring of beer
fermentation based on hybrid electronic tongue, Bioelectrochemistry 87
(2012) 104113.
[33] M. Gutierrez, C. Domingo, J. Jordi Vila-Plana, A. Ipatov, F. Capdevila, S.
Demming, S. Bttgenbach, A. Llobera, C. Jimnez-Jorquera, Hybrid electronic
tongue for the characterization and quantication of grape variety in red
wines, Sensors and Actuators B: Chemical 156 (2011) 695702.
[34] F.S. Winquist, S. Holmin, C. Krantz-Rlcker, P. Wideb, I. Lundstrm, A hybrid
electronic tongue, Analytica Chimica Acta 406 (2000) 147157.
[35] A. Legin, D. Kirsanov, A. Rudnitskaya, J.J.L. Iversen, B. Seleznev, K.H. Esbensen, J.
L.P. Mortensen Houmller, Y. Vlasov, Multicomponent analysis of fermentation
growth media using the electronic tongue (ET), Talanta (2004) 766772.
[36] E. Malinowska, . Grski, M.E. Meyerhoff, Zirconium(IV)-porphyrins as novel
ionophores for uoride-selective polymeric membrane electrodes, Analytica
Chimica Acta 468 (2002) 133141.
[37] E. Bakker, E. Pretsch, P. Buhlmann, Selectivity of potentiometric ion sensors,
Analytical Chemistry 72 (2000) 11271133.