alexa Application of MOOSY32 eNose to Assess the Effects of Some Post Harvest Treatments on the Quality of Salustiana Orange Juice | Open Access Journals
ISSN: 2155-6210
Journal of Biosensors & Bioelectronics
Like us on:
Make the best use of Scientific Research and information from our 700+ peer reviewed, Open Access Journals that operates with the help of 50,000+ Editorial Board Members and esteemed reviewers and 1000+ Scientific associations in Medical, Clinical, Pharmaceutical, Engineering, Technology and Management Fields.
Meet Inspiring Speakers and Experts at our 3000+ Global Conferenceseries Events with over 600+ Conferences, 1200+ Symposiums and 1200+ Workshops on
Medical, Pharma, Engineering, Science, Technology and Business

Application of MOOSY32 eNose to Assess the Effects of Some Post Harvest Treatments on the Quality of Salustiana Orange Juice

Cupane M1,2, Pelegri-Sebastia J3*, Climent E3, Guarrasi V2, Sogorb T3 and Germana MA1

1Department of Agricultural and Forestry Science (SAF), Università degli Studi di Palermo, Viale delle Scienze 11, Ed. 4, 90128, Palermo, Italy

2Istituto di BioFisica, Consiglio Nazionale delle Ricerche, Via La Malfa 153, 90146, Palermo, Italy

3Deparmtent of Electronic Engineering, Universitat Politecnica de Valencia, C/ Paranimf 1, 46730, Grao de Gandia, Valencia, Spain

*Corresponding Author:
Pelegri-Sebastia J
Deparmtent of Electronic Engineering
Universitat Politecnica de Valencia
C/Paranimf 1, 46730
Grao de Gandia, Valencia, Spain
Tel: 34 962849404
E-mail: jpel[email protected]

Received Date: September 21, 2015; Accepted Date: October 07, 2015; Published Date: October 17, 2015

Citation: Cupane M, Pelegri-Sebastia J, Climent E, Guarrasi V, Sogorb T, et al. (2015) Application of MOOSY32 eNose to Assess the Effects of Some Post Harvest Treatments on the Quality of ‘Salustiana’ Orange Juice. J Biosens Bioelectron 6:184. doi:10.4172/2155-6210.1000184

Copyright: © 2015 Cupane M, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Visit for more related articles at Journal of Biosensors & Bioelectronics


A new prototype of Electronic Nose instrument, Multisensory Odor Olfactory System MOOSY32, with a processing method based on a multivariate classification analysis was used to assess different postharvest and storage treatments effects to Salustiana oranges. The analysis method is based on the measurement of the volatile compounds produced under different environmental and operational conditions. The Electronic Nose system revealed that orange juice flavor changes even when juices are analyzed right after each treatment and fruits are stored under refrigerated conditions. The instrument was able to detect even small changes in the aromatic pattern of the juices, confirming that the packing line itself is able to cause perceptible changes in the flavor. This can be a new and important finding in the Salustiana orange treatment that can lead to a significant improvement of fruits quality on the markets.


Electronic nose; Orange; Flavor; Ethylene; Coating; Storage; WEKA; ANN


eNose: Electronic nose; MOOSY32: Multisensory Odor Olfactory System with 32 sensors; MOS: Metal Oxide Semiconductors; VOC: Volatile Organic Compound; WEKA: Waikato Environment for Knowledge Analysis; GNU: Operating system free software; ANN: Artificial Neuron Networks; MLP: Multi-Layer Perceptron; RBF: Radial Basis Function network; IB1: Instance-Based learning; J48: Open source java implementation of the C4.5 algorithm


The electronic nose (eNose) instruments are able to simulate the human nose, replicating the four fundamental functions of the smell sense: detection, recording, memory search and identification. The first two functions are simulated by the use of chemical sensors; the other can be simulated by Artificial Intelligence software [1]. The most common eNoses are based on the use of an array of electronic chemical sensors, with partial or no specificity, coupled with an appropriate system of pattern recognition. The statistical treatments of the data use complex algorithms to extract all the information that can be useful for the different applications [2]. ENose instrument provides a fast comparative measure of patterns of odors, representative of compounds disengaged by a substratum [3], offering a fast nondestructive alternative to sense aroma. In the last decade there have been several reports on electronic sensing in environmental control, medical diagnostics and food industry [4-9].

Regarding citrus several applications of commercial eNose instruments can be found in literature. In most cases the application of these kinds of instruments concerned the authentication, classification and characterization of Citrus spp. [10], or for quality monitoring of mandarins [11] and processed orange juices [12].

The most common sensors used to achieve these objectives are usually Metal Oxide Semiconductors (MOS) sensors. This kind of sensors is based on the interaction with the volatile compounds induces mechanisms of adsorption and desorption, taking place on the surface of sensors and provoking the modification of its electrically measurable properties by a variation of resistance versus time [13]. The general operating principle of a MOS gas sensor is based on the changes that occur in the sensing material when it is heated. In fact when a metal oxide crystal, such as SnO2, is heated, the oxygen is adsorbed on the crystal surface causing a negative charge, and the donor electrons are transferred to the adsorbed oxygen. As a result of this process, a positive charge is formed in a space charge layer forming a surface potential that serve as a potential barrier against electron flow. Inside the sensor, at the level of SnO2 micro crystals, electric current flows and at grain boundaries, the adsorbed oxygen forms a potential barrier that prevents carriers from moving freely. This potential barrier determines the electrical resistance of the sensor. In this way when a deoxidizing gas arrives on the sensor, the surface density of the negatively charged oxygen decreases, so the barrier height in the grain boundary is reduced. The reduced barrier height decreases sensor resistance and increases the electrical conductivity.

The opportunity to apply this kind of technology and instruments directly on line to monitoring the complete process of citrus processing could be of fundamental importance to improve quality of citrus products, ensuring the maintenance of quality throughout all the production chain.

The objectives of this study were: a) to evaluate the quality of ‘Salustiana’ juice squeezed from fruits previously submitted to postharvest treatments with fungicide, degreening and wax coating, by monitoring the changes in the aroma fingerprint; b) to evaluate the effect on juices aromatic patterns of different treatments during two different fruits storage conditions; c) to evaluate the capacity of prototype designed of eNose, named MOOSY32, to monitoring the change in volatile composition of ‘Salustiana’ oranges juices obtained from fruits submitted to different postharvest treatments and storage conditions. Moreover, different methods of data classification, such as Bayesian nets, Artificial Neuron Networks and classification tree were applied to analyze the results.

Materials and Methods

In order to test the ability of the whole eNose system in the detection of the aromatic changes occurring during citrus processing and storage, this investigation was carried out on fruits of ‘Salustiana’ orange (Citrus sinensis L. Osbeck), purchased a farm located in Xeraco, in the region of Valencia, Spain. Fruits were harvested at the commercial maturity stage and were passed through to the packing line. Different postharvest treatments were conducted picking up the samples in subsequent step of the packing line, in order to assess the effect of each treatment. Precisely choosen steps were: Control, fruits collected right after harvest; after washing with chemical products (Chemical treatment); after the degreening process with application of ethylene (Ethylene treatment); and after waxing process (Wax treatment). After the treatments, the fruits were transported to the laboratory for applying the different storage tests. Specifically, the samples were analyzed in the first 24 - 48 h after harvest (T0). Treated and control fruits were stored at 5±1°C and 85% RH for eight weeks and analyzed after four (T1), and after eight weeks (T2) to assess the changes occurring during the storage. After the storage period, a sample of fruits was kept at 20°C for 7 days to simulate the shelf-life conditions (T3). Experimental design is showed in Table 1.

Treatment Storage
Control T0 (0 week)
T1 (4 weeks at 5°C)
T2 (8 weeks at 5°C)
T3 (8 weeks at 5°C + 1 week at 20°C)
Chemical treatment T0 (0 week)
T1 (4 weeks at 5°C)
T2 (8 weeks at 5°C)
T3 (8 weeks at 5°C + 1 week at 20°C)
Ethylene degreening
T0 (0 week)
T1 (4 weeks at 5°C)
T2 (8 weeks at 5°C)
T3 (8 weeks at 5°C + 1 week at 20°C)
Wax treatment T0 (0 week)
T1 (4 weeks at 5°C)
T2 (8 weeks at 5°C)
T3 (8 weeks at 5°C + 1 week at 20°C)

Table 1: Experimental design.

For each applied treatment and for each storage time 10 fruits were hand squeezed to extract the juice for the analysis. The aromatic pattern was analyzed with MOOSY32 instrument, using a method based on the measurement of the volatile compounds produced by the fruits under different processing and storage conditions.

The instrument, which is still in the development phase, utilizes 32 commercial MOS sensors of 5 different types, all produced by Figaro Engineering Inc (Table 2). The use of different types of sensors combined with the selected operating temperature leads to a wide range of different responses toward volatile organic compounds with a wide variety of applications (Table 2).

Model Target Gas Typical detection Range
TGS2600 General Air Contaminants 1 – 30 ppm
TGS2610 – C00 LP gas 500 – 10.000 ppm
TGS2610 – D00 LP gas 500 – 10.000 ppm
TGS2611 Methane 500 – 10.000 ppm
TGS2620 Alcohol, solvents vapor 500 – 5.000 ppm

Table 2: MOS sensor array configuration of the MOOSY32. Specificity from Figaro.

The whole system composition and operation was already described by Del Cueto Belchi [14]. Briefly, it consists of two parts: the electrical part and the processing of the sample part (prechamber). The prechamber is linked with a clean air pump which air flow is splitted in two streams, one that goes directly to the sensors and is used as a reference line, while the other passes through the sample chamber to carry the volatile molecules to the sensors chamber. The sensor signal are recorded continuously until the signal of each sensor reaches a steady state, and acquired by a board of National Instruments. Thereafter, the output signals from the sensors are digitized and stored. It is possible to analyze this data in different ways, using different algorithms for the extraction of different features. The classification features utilized for this study were: the minimum (vA) and the maximum (vB) values of the rising part of the curve, the minimum value of the descending part of the curve (vC), and the differences among these values, as difA= vBvA and difB=vB-vC (Figure 1).


Figure 1: Typical response of sensors to sample air and classification features: vA) minimum value of the ascending curve; vB) maximum value of the ascending curve; vC) minimum value of the descending curve.

For the analysis, 10 mL of juice was kept in a Petri dish maintained at 25°C. Total acquisition time was 3,5 minutes and the air flow was settled at 0.5 m/s. After 2 minutes of equilibration time, in which the sample was placed in the acquisition chamber, for each analysis 5 measurements were performed with a delay between repetitions of 10 sec.

The e-nose sensors responses were analyzed used a multivariate classification analysis with the software WEKA, using different algorithms. WEKA is an open source software issued under the GNU General Public License [15]. The different algorithms applied were: two probabilistic models, Bayes Net and Naïve Bayes, that calculate a set of probabilities by counting the frequency and combination values on a given data set; two ANN, MLP and RBF, that permit to create a model, during a training test, that is able to classify the data [16]; a classification tree, J48 that is a support system that uses a tree- like graph decisions and their possible after effect; and a lazy algorithm named IB1, that uses the technique of the ‘Nearest Neighbour’ to classify new instances using a similarity function to calculate the similarity between the training instance and the instances of the data set [17,18].

The classification analysis consists in the organization of data in classes, using given class labels to order the objects in the data collection. Classification approaches normally use a training set where all objects are already associated with known class labels. The classification algorithms learn from the training set to build a model. The model is used to classify new objects [19]. ANNs learn from examples through iteration, without requiring a priori knowledge of the relationship among variables under investigation [20].

The results are illustrated in the confusion matrix that represents the accuracy of the solution of the classification problem. It allows the visualization of the performance of an algorithm. Each column of the matrix represents the instances in a predicted class, while each row represents the instances in the actual class. The ideal result is to have all the samples end up on the diagonal cells of the matrix [20]. Figure 2 represent the typical graphical response of the WEKA software for the analysis time T2, for the analysis parameters vB and vC.


Figure 2: Graphical response of the WEKA software for the analysis time T2. It is possible to observe the analysis parameters vB and vC.

Results and Discussion

At the beginning of the data analysis all the responses from the eNose system were used as inputs to the model that was build. The classification was performed separating one subset of samples for training and another for testing, using the cross-validation method. Figure 2 and Table 3 shows the results obtained by the different algorithms to classify fruits from different treatment in each time. It was necessary to use algorithms to test the variables that most strongly influenced the classification. Variable selection consists in the selection of a subset of variables that are the most discriminating [17] (Figure 2).

For this purpose we used two different algorithms: CfsSubsetEval and GreedyStepWise. The stepwise selection is based on a greedy search that sequentially adds or deletes variables from the pool of total variables [17]. The addiction or deletion of a variable is determined based on the largest improvement in the classification, until the search finds the most influencing variables (Table 3).

Sample type Classifier algorithm % accuracy
Bayes Net Naive Bayes MLP RBF J48 IB1
T0 92,5 95 95 97,5 88,75 97,5
T1 100 100 100 100 96,25 100
T2 97,5 97,5 98,75 98,75 91,25 97,5
T3 87,5 93,75 87,5 92,5 85 90

Table 3: Results of the used classification algorithms for the MOOSY32.

The results show high classification accuracy, being able to distinguish among the different juice sample even at T0, right after each treatment. It is interesting to notice that the applied algorithms proved different performances showing a general decrease of classification accuracy over time and being better in storage time T1, after 4 weeks of storage at 5°C, and worst in storage time T3, after the additional storage week at 20°C. The algorithm J48 permitted the worse classification, and the percentage of accuracy decreased with time.

So, the instrument allowed a good classification of the analyzed juice samples according to the time of storage, and it is possible to highlight a trend in the observed data:

T0: This time gave the worst classification. The aromatic patterns were similar right after the treatments. The lower percentage of classification is caused by confusion between ‘Control’ and ‘Chemical’;

T1 and T2: Showed the greater percentage of correctly classified instances with some accuracy difference due to the different instruments and algorithms;

T3: This time permits the worst classification with all the algorithms.

These results are consisting with the hypothesis that initially the aromatic patterns were quite similar (T0), but with slight differences that permitted the classification. Since fruits belonging to the same lot, the differences at T0 can only be due to the applied treatments. Major mistakes in classification were mainly due to un-correct classification of ‘Control’ and ‘Chemical’ treated samples.

With advancing of time the aromatic patterns became more different, resulting in an increase of the correctly classified instances. After 4 and 8 weeks of storage (T1 and T2) the instruments were able to better classify the samples, meaning that the major changes occurred during this period. It is possible to suppose that the different treatments caused variations in the internal atmosphere of fruits, probably due to metabolic changes and variations in the respiration rate, as reported in literature [21-23], that led to different VOCs production and consequently to changes in the aromatic pattern.

The last analysis time (T3) gave the worst results. There were no more detectable differences in the aromatic patterns that became similar for all the treatments after a week of storage at 20°C. All the changes that were stimulated by the applied treatments were amplified by the storage, and overall by the temperature of storage.


Electronic nose revealed different aromatic pattern of the fruit juices. The MOOSY32 with the processing method used was able to detect variation in the juice aroma even when juices were analyzed right after each treatment (T0). Actually, treatments caused changes in the ‘flavor’ of orange fruit juices that are easily detected by electronic nose. This confirming that even slight changes in the chemical composition of the volatile fraction can lead to changes in the perceived aroma.

Moreover, the eNose results showed that something changed in the aromatic pattern of orange fruits due to the packing line, and that the pattern of variation is different on the basis of the applied storage conditions. Furthermore, there is a major effect due to ethylene degreening and coating, but also control fruits changed their aromatic pattern in response to the storage time and temperature. Storage amplified the effect of each treatment, making possible to distinguish among them.

The eNose results showed that not only coating or degreening caused changes in the aromatic pattern, but also volatile composition is altered as well by passage through the drenching system.

This study suggests that the packing line itself is able to affect the aromatic pattern of oranges, and that this effect is more pronounced when fruits are stored. Particularly, major changes were caused by ethylene and wax applications. Take care in the manipulation of fruits and to minimize the other treatments and the storage times would be an appropriate suggestion in trying to preserve flavor quality of Salustiana orange juice.


Select your language of interest to view the total content in your interested language
Post your comment

Share This Article

Relevant Topics

Article Usage

  • Total views: 11643
  • [From(publication date):
    December-2015 - Oct 22, 2017]
  • Breakdown by view type
  • HTML page views : 7869
  • PDF downloads :3774

Post your comment

captcha   Reload  Can't read the image? click here to refresh

Peer Reviewed Journals
Make the best use of Scientific Research and information from our 700 + peer reviewed, Open Access Journals
International Conferences 2017-18
Meet Inspiring Speakers and Experts at our 3000+ Global Annual Meetings

Contact Us

Agri, Food, Aqua and Veterinary Science Journals

Dr. Krish

[email protected]

1-702-714-7001 Extn: 9040

Clinical and Biochemistry Journals

Datta A

[email protected]

1-702-714-7001Extn: 9037

Business & Management Journals


[email protected]

1-702-714-7001Extn: 9042

Chemical Engineering and Chemistry Journals

Gabriel Shaw

[email protected]

1-702-714-7001 Extn: 9040

Earth & Environmental Sciences

Katie Wilson

[email protected]

1-702-714-7001Extn: 9042

Engineering Journals

James Franklin

[email protected]

1-702-714-7001Extn: 9042

General Science and Health care Journals

Andrea Jason

[email protected]

1-702-714-7001Extn: 9043

Genetics and Molecular Biology Journals

Anna Melissa

[email protected]

1-702-714-7001 Extn: 9006

Immunology & Microbiology Journals

David Gorantl

[email protected]

1-702-714-7001Extn: 9014

Informatics Journals

Stephanie Skinner

[email protected]

1-702-714-7001Extn: 9039

Material Sciences Journals

Rachle Green

[email protected]

1-702-714-7001Extn: 9039

Mathematics and Physics Journals

Jim Willison

[email protected]

1-702-714-7001 Extn: 9042

Medical Journals

Nimmi Anna

[email protected]

1-702-714-7001 Extn: 9038

Neuroscience & Psychology Journals

Nathan T

[email protected]

1-702-714-7001Extn: 9041

Pharmaceutical Sciences Journals

John Behannon

[email protected]

1-702-714-7001Extn: 9007

Social & Political Science Journals

Steve Harry

[email protected]

1-702-714-7001 Extn: 9042

© 2008-2017 OMICS International - Open Access Publisher. Best viewed in Mozilla Firefox | Google Chrome | Above IE 7.0 version