Graphic Art

A novel electronic nose as adaptable device to judge microbiological quality and safety in foodstuff

A novel electronic nose as adaptable device to judge microbiological quality and safety in foodstuff
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Research Article  A Novel Electronic Nose as Adaptable Device to JudgeMicrobiological Quality and Safety in Foodstuff   V. Sberveglieri, 1,2 E. Nunez Carmona, 1,3 Elisabetta Comini, 2,4  Andrea Ponzoni, 2 Dario Zappa, 2,4 Onofrio Pirrotta, 5 and A. Pulvirenti 1,2 󰀱 Department of Life Sciences, University of Modena and Reggio Emilia, Via Amendola,  󰀴󰀲󰀱󰀲󰀲  Reggio Emilia, Italy  󰀲 CNR-INO Sensor Lab, Via Valotti  󰀹  ,  󰀲󰀵󰀱󰀳󰀳  Brescia, Italy  󰀳 CNR IBF, Via Ugo La Malfa  󰀱󰀵󰀳  ,  󰀹󰀰󰀱󰀴󰀶   Palermo, Italy  󰀴 Department of Information Engineering, University of Brescia, Via Valotti,  󰀲󰀵󰀱󰀳󰀳  Brescia, Italy  󰀵 University of Modena and Reggio Emilia, DISMI, Via Amendola,  󰀴󰀲󰀱󰀲󰀲  Reggio Emilia, Italy  Correspondence should be addressed to V. Sberveglieri; veronica.sberveglieri@unimore.itReceived  2󰀷  November  2013 ; Accepted  30  January   2014 ; published  24  March  2014 Academic Editor: Moreno BondiCopyright ©  2014  V. Sberveglieri et al.  T is is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the srcinal work is properly cited. T is paper presents di ff  erent applications, in various foodstu ff  s, by a novel electronic nose (EN) based on a mixed metal oxidesensors array composed of thin  󿬁 lms as well as nanowires.  T e electronic nose used for this work has been done, starting fromthe commercial model EOS 󰀸35  produced by SACMI Scarl.  T e SENSOR Lab (CNR-INO, Brescia) has produced both typologiesof sensors, classical MOX and the new technologies with nanowire.  T e aim of this work was to test and to illustrate the broadspectrum of potential uses of the EN technique in food quality control and microbial contamination diagnosis.  T e EN techniquewas coupled with classical microbiological and chemical techniques, like gas chromatography with mass spectroscopy (GC-MS)withSPMEtechnique. T reedi ff  erentscenariosarepresented:(a)detectionofindigenousmouldingreenco ff  eebeans,(b)selectionof microbiological spoilage of Lactic Acid Bacteria (LAB), and (c) monitoring of potable water. In each case, the novel EN was ableto identify the spoiled product by means of the alterations in the pattern of volatile organic compounds (VOCs), reconstructed by principal component analysis (PCA) of the sensor responses.  T e achieved results strongly encourage the use of EN in industriallaboratories. Finally, recent trends and future directions are illustrated. 1. Introduction Aromaisoneofthemostsigni 󿬁 cantparametersoffoodsfromthesensorypointofview. T echaracteristic 󿬂 avourofVOCs,so called  󿬁 ngerprint, may o ff  er information about safety andquality of food, performing sometimes as an indicator of process mistakes as well [ 1 ].Indeed, some volatile compounds can be originatedfrom biochemical processes of food, as a consequence of technological food chain or product storage.Unwanted smell, so-called o ff  - 󿬂 avour, may involve sub-stancessrcinatingfromthemetabolismofspoilagemicroor-ganisms,bacteria,andfungithatadulteratenaturallyorunin-tentionally the products before or during its production [ 2 ].In the last decade, electronic noses (EN) have become very popular as monitoring tools in evaluating food quality and safety [ 3 ].In this paper, three important applications of EN in foodcontrolwereexamined,concealingthreerelevantissuesinthefood  󿬁 eld of food quality and control.Another main target of this work was to illustrate thebroad spectrum of potential uses of sensor technology inthis  󿬁 eld and to show the potential of the new Nanowiretechnology.At SENSOR laboratory, the studies on chemical sensorsstarted at  1󰀹󰀸󰀸  with the improvement of thin  󿬁 lms and thenof a new technique for the planning of thin  󿬁 lms with anextremely porous structure [ 4 ]. Hindawi Publishing CorporationBioMed Research InternationalVolume 2014, Article ID 529519, 6 pages  2  BioMed Research International F 󰁩󰁧󰁵󰁲󰁥 1 : ZnO nanowires SEM picture, at  50 k magni 󿬁 cation. In  2001  a f er the  󿬁 rst publications demonstrating thepossibility of preparing metal oxide in forms of nanowiresand nanobelt, SENSOR demonstrates the ability of metaloxide nanowires in detecting variety of chemical species [ 5 ].Itiswellknowfromsixdecadesthatmetaloxideelectricalproperty depends on the surrounding atmosphere.Quasi-one-dimensional metal oxide nanostructures(Figure  1 ) have several advantages with respect to thin andthick  󿬁 lmcounterpartssuchaslargesurface-to-volumeratio,lateral dimensions comparable to the surface charge region,andsuperiorstabilitywheninthesinglecrystalstructure[ 󰀶 ].Singlecrystalnanostructures(Figure  2 )oftinoxideshavebeen fabricated and characterized as sensing materials to beimplemented in an electronic nose. T ese nanowires exhibit remarkable crystalline quality and a very high length-to-width ratio, resulting in enhancedsensing performances as well as long-term stability for sus-tained operation [ 󰀷 ]. 2. Materials and Methods (a) All the co ff  ee, produced and consumed, belongs tothe genus  Co  ff  ea  that focuses principally two species: C. arabica  and  C. canephora .  T ose species generally matureintheequatorialzone,wheretheenvironmentconditionsofhumidity,temperature,wind,rains,andaltitudepermittheharvestofmanydi ff  erentvarieties,each one with particular requirements.Most of the producers are developing countries, so thestorage conditions and the methods used for harvesting andshipping of green co ff  ee beans depend on the productioncountry.For this reason, it can be easily contaminated by mouldthroughout the food chain.Usuallythecontaminationappearsduetomouldsbelong-ing to the genus  Aspergillus . T eselectionoftherawmaterialoccursintheearlystagesof the processing chain by visual inspection. Parameters thatdeterminate the quality of green co ff  ee beans are shape,colour, and size.Frequently the raw material is already contaminatedwhen this selection occurs and thus the detection of thecontamination became very di ffi cult. F 󰁩󰁧󰁵󰁲󰁥 2 : Sketch of the conductometric device. Alumina substrateiswhitecoloredinthemiddle,TiWpadsareinbrown,andplatinumheater and contacts are in metallic gray. Rose-Bengal Chloramphenicol Agar (Sigma-AldrichChemical Co., St. Louis, MO, USA) was used, as a selectivemedium for the enumeration of yeast and moulds from awide variety of distinctive food matrix. T e Chloramphenicol works as suppressor of bacteriagrowth, although the Rose-Bengal acts as a limiting agentof the mould growth. In order to assist the enumeration of small colonies, the media have neutral pH.  T e same kind of medium was used to perform all the analysis.In the case of GC-MS-SPME and EN analysis,  5 mL of medium was placed into sterile  20 mL chromatographic vialand le f  to solidify.All the samples, for microbiological analysis, CG-MS-SPME and EN were prepared on the  󿬁 rst day of the analysis,taking this day as zero-time. Once prepared, all the sampleswere stored and incubatedunder the same conditionsduringall the duration of analysis, namely,  2󰀸 ∘ C for a total of   󰀷  days.During the  󰀷  days of the experiment, analysis was doneat zero-time, inoculation day, at T 3  ( 3  days a f er the inocula-tion), at T 4  ( 4  days a f er the inoculation), at T 󰀶  ( 󰀶  days a f erthe inoculation), and at T 󰀷  ( 󰀷  days a f er the inoculation).For each day of GC-MS analysis,  10  vials of every type of samples were prepared (control, Honduras, Indonesia, andIndia): one of each type was used for the GC-MS-SPMEanalysis, while the other  3󰀶  for the EN analysis.For each day of analysis, the vials were crimped andincubated in an oven thermostatically regulated at  40 ∘ C for 15  minutes, to create the headspace equilibrium.In order to extract the volatile compound from the sam-ples, a DVB/carboxen/PDMS stable  󿬂 ex ( 50 / 30  m) (SupelcoCo., Bellefonte, PA, USA) SPME  󿬁 ber was used.  BioMed Research International  3 To furnish the adsorption of volatile compounds, theSPME  󿬁 ber was exposed to the headspace of the vials for 15  minutes at room temperature. For desorption of thecompounds, the  󿬁 ber was placed in the injector of the heatedGC for  󰀶 min. T e ramping of temperature in the column was per-formed in the following way:  󰀶0 ∘ C for  2 min to  100 ∘ C at 5 C ∘ /min, followed by a rise from  100 ∘ C to  240 ∘ C at  5 ∘ C/minand then this temperature was kept for  5 min. Chromato-graphicanalysiswasaccomplishedusingaHP 󰀶󰀸󰀹0 seriesGCsystem, 5󰀹󰀷3 massselectivedetectorwithaDB-WAXcapillary column.  T e injection was veri 󿬁 ed in splitless mode at  240 ∘ Cusing helium as gas carrier with a setting  󿬂 ow of   1 . 5 mL/min.For the electronic nose the sample headspace ( 4 mL) wasthen extracted from the vial in static headspace path andinjected into the carried  󿬂 ow (speed  4 mLmin) through aproperly modi 󿬁 ed gas chromatography injector (with theconnection tube to the EN kept at  40 ∘ C to prevent any condensation).A synthetic chromatographic air with a continuous  󿬂 ow rate of   10 mL/min was used to recover the sensor baseline,resulting in a recovery time of   2󰀸 min.(b) In the case of LAB analysis, the samples were treatedusing spoilage lactic acid micro 󿬂 ora isolated fromthe chicken meat.  T e used procedure for samplingmethodwasconductedasitisdescribedbelow.Understerile conditions,  10 g of chopped chicken meat wasplaced in a stomacher bag with  󰀹0 mL of sterilephysiologicalsolutionandshackedo ff  inLabBlenderStomacher  400  (Type BA  󰀷021  Seward, London) for 1 min at normal speed ( 200  paddles/min).One mL of the supernatant was inoculated in Man,RogosaandSharpeAgarmedium(MRSA)(OXOID)[ 󰀸 ]Petridish following the inclusion method. Once the  󿬁 rst layerwas solid, a second layer was added in order to create themicroaerobic environment conditions for the LAB growth.MRSA medium was considered to be supporting the growthof lactobacilli.  T e plates, inside jar with the gas generatingkit (OXOID), were incubated during  4󰀸  hours at  30 ∘ C. T en, the  4󰀸 -hour  3  colonies were randomly picked upand inoculated in a MRS liquid tube in order to obtain liquidcultures.  T e  3  typologies of tubes were incubated for  4󰀸 hours at  30 ∘ C. T e samples for EN and GC-MS analysis were preparedusing the same procedure. Sterilized chromatographic vials( 20 mL) containing  2 mL of MRSA media were inoculatedindependently with  100  L of the number  3  of McFarlandstandards of the  3  kinds of cultures prepared before.  T esestandards are used as a reference to adjust the turbidity of bacterial suspensions in order to have a number of bacteriawithin a given range. Number  3  of McFarland standardmatches a bacterial concentration of   󰀹  ×  10 8 CFU/mL.Analysis with EN and GC-MS was done at  0  time,inoculation day, and  24  hours later a second cycle of GC-MSwas performed to make a control of the head space changesat the end of the EN analysis.(c) Regarding the incidence of coliforms, an aliquot of water from wc and a well was dispersed in  2  Petridishes with Violet Red Bile Agar (VRBA) (OXOID)[ 󰀹 ]. VRBA is a selective medium used for the detec-tion and enumeration of coliform bacteria in waterandotherfooddairyproducts. T egoalwastoisolatethe single colonies that were used later to inoculateliquidtubesofBrilliantGreenBilemedium(OXOID)[ 10 ] to obtain pure liquid cultures. Brilliant GreenBile medium is a modi 󿬁 cation of MacConkey’s liquidmedium for the isolation of Enterobacteriaceae andhas been formulated to obtain maximum recovery of bacteria of the coli-aerogenes group, while inhibitingmost gram-positive bacteria.For the  2  kinds of samples the followed procedure wasthe same. Once the Petri dishes were inoculated, they wereconserved at room temperature for  2  days.  T en, singlecolonies were selected and inoculated in liquid tubes of Brilliant Green Bile media and incubated for  24  hours at theoptimal growth temperature for coliforms  35 ∘ C.A f er  24  hours, the turbidity of the tubes was evident,and it was adjusted (diluted, using sterile Brilliant Green Bilemedium) until the turbidity was the same as the number  3  of the McFarland standards.SamplesGC-MS-SPMEandENwerepreparedonthe 󿬁 rstdayoftheanalysistakingthisdaylike 0 time.Onceprepared,all the samples were stored and incubate under the sameconditions during all the lasting of analysis, namely,  35 ∘ C fora total of   24  hours. T eanalysesweredonethesamedayofinoculation,andasecond cycle of GC-MS was made  24  hours later. In this case,the samples were not crimped and the ensemble was coverwith aluminium foil in order to keep vial and cap combinedto preserve the sterility inside and, at the same time, a ff  ordthe aerobic condition for the bacterial growth.Principal component analysis (PCA) performed explo-rative data analysis. Data were processed by EDA so f ware,at-home-writtenso f waredevelopedinMATLABatSENSOR laboratory [ 11 ].Exploratory data analysis (EDA) is a fundamental step inthe data analysis cycle (the cycle consists of data acquisition,data preprocessing, exploratory data analysis, and classi 󿬁 ca-tion). T e aims of explorative analysis are as follows: maximizeinsight into a data set, uncover underlying structure, extractimportant features, and detect outliers.  T e most valuableoutcome of EDA is to check for prior assumptions anddetermine optimal experimental settings. 3. Results and Discussion (a) A f er  󰀷  days of incubation (classical microbiologicalanalysis) the di ff  erences among the inoculated plateswere manifested. Each sample corresponds to oneof the  3  di ff  erent provenances of co ff  ee selected forthe experiments. It’s perfectly shown the di ff  erencesin number and typology of colonies between thesamples.  4  BioMed Research International 012301234PC1 46.71    P   C   2   2   1 .   2   3 IndonesiaIndonesia-day T0Indonesia-day T3Indonesia-day T4Indonesia-day T6 − 5  − 4  − 3  − 2  − 1 − 3 − 2 − 1 F 󰁩󰁧󰁵󰁲󰁥3 :PCAscoreplotaboutIndonesiagreenco ff  eebeansduring 󰀶  days of analysis (T 0  to T 󰀶 ). 01234501234PC1 48.79    P   C   2   2   5 .   0   8 Honduras-IndiaIndia-day T0India-day T3India-day T4India-day T6India-day T7Honduras-day T0Honduras-day T3Honduras-day T4Honduras-day T6Honduras-day T7 − 5  − 4  − 3  − 2  − 1 − 4 − 3 − 2 − 1 F 󰁩󰁧󰁵󰁲󰁥 4 : PCA score plot about the comparison of two di ff  erentsrcins, Honduras and India green co ff  ee beans (T 0  to T 󰀷 ). T ese results suggest that co ff  ee from di ff  erent prove-nances have qualitative and quantitative di ff  erence in indige-nous contamination.Results obtained with GC-MS showed to be in perfectcorrelation with those obtained with the other two tech-niques. T ese di ff  erences are both qualitative and quantitative,showing in all cases, except the control, compounds thatcorresponded with the metabolites produced by mouldsduring their growth.In particular, it can be highlighted the formation of car-bon dioxide, ethanol, and compounds belonging to chemicalindole group [ 12 ].In the 󿬁 gures(Figures 3  and 4 )areshownthedatarelatedto Indonesia, Honduras, and India, from day T 0  (day of thesample preparation) to days T 󰀶  and T 󰀷  (six and seven daysa f er sample preparation). LABPC1 70.29    P   C   2   2   0 .   5   6 Day 1-kind 3Day 1-kind 2Day 1-kind 1Day 0-kind C − 0.8 − 0.75 − 0.7 − 0.65 − 0.6 − 1.15  − 1.1  − 1.05  − 1  − 0.95  − 0.9  − 0.85  − 0.8 F 󰁩󰁧󰁵󰁲󰁥 5 : PCA score plot about control in a separate cluster and  3 di ff  erent kinds of LAB. In the blue circle the kind  1 , in the red circlethekind 3 ,ingreencirclethekind 2 ,andinblackcircleinthebottomthe control sample. It is clearly visible a separation in the PC 1  axis among thedays of analysis, showing development of the growth of themoulds that in both case has more statistical signi 󿬁 cance.In thepreviousliterature[ 12 ],alltheresultswerereferredto an array of   󰀶  MOX thin  󿬁 lm sensors that provide a realindividualization of the samples only at day T 󰀶 . T e new array composed of   4  MOX [ 󰀷 ] thin  󿬁 lm sensorsand other  2  made up with MOX nanowire technology; thedi ff  erencesbetweensamplesareevidentalreadyatT 3 ,halvingthe response time and hence increasing the instrumentthreshold.(b) InFigure  5 areshowedtheresultfromthePCAanaly-sis of LAB. It can be separate  2  principal clusters. Onedenotes to the control (black circle), and the secondone is formed for  3  typologies, samples belonging todi ff  erent colonies. T e creation of cluster like this one can be due alsobecause the VOCs that the EN is able to detect were presentin the total pathway of the LAB and to the similarity of thecolonies because of the same provenience of the samples, allthem indigenous contaminated of the chicken meat.Regarding the kind  1  (blue circles), samples belonging tocolony   1  it worth to think that can be bacteria from the samegroup but perhaps to di ff  erent species so it will explain thatsome of the samples start to move away from the generalcluster. T e result obtained in LAB, with GC-MS (Figure  󰀶 ),case seems to be in perfect correlation with those obtainedwith the EN. Actually, there are evident qualitative andquantitative di ff  erences within the components of samples of the same group.(c) In the case of PCA (Figure  󰀷 ) from the EN analysis of coliforms, a separate cluster formed by the samples of the bacteria isolate from the well and those belongingto wc can be observed. Very interesting results cameout concerning the samples proceeded from theisolate of the wc. In this case the di ff  erent measures  BioMed Research International  5 02468101214161820    A  c  e   t  o  n  e   2  -   M  e   t    h  y    l    b  u   t  a  n  a    l   3  -   M  e   t    h  y    l    b  u   t  a  n  a    l   2  -   I  m   i    d  a   z  o    l   i    d  o  n  e   3  -   M  e   t    h  y    l  -   1  -    b  u   t  a  n  o    l   C  y  c    l  o    b  u   t  a  n  o    l   D  o    d  e  c  a  n  e   1  -   P  e  n   t  a  n  o    l   P  y  r  a   z   i  n  e   M  e   t    h  y    l   p  y  r  a   z   i  n  e   P   i  r  a  n  o  s   i  n  e  g  r  o  u   p   2 ,   5  -   M  e   t    h  y    l   p  y  r  a   z   i  n  e   2 ,   6  -   M  e   t    h  y    l   p  y  r  a   z   i  n  e   2  -   N  o  n  a  n  o  n  e   C  a  r    b  o  n  y    l  s  u    l         󿬁     d  e   P  r  o   p  a  n  o   i  c  a  c   i    d   2  -   M  e   t    h  y    l   p  r  o   p  a  n  o   i  c  a  c   i    d   2  -   U  n    d  e  c  a  n  o  n  e   3  -   M  e   t    h  y    l    b  u   t  a  n  o   i  c  a  c   i    d CT01 T01 T1 × 10 6 F 󰁩󰁧󰁵󰁲󰁥󰀶 :HistogramobtainedfromGC-MS,withSPMEtechniquesabout VOCs produced by control and one kind of LAB. 0 1 2 3 4 500.511.5 Wc and wellPC1 93.16    P   C   2   5 .   9   1 Micro-wellMicro-wcMicro-control − 2  − 1 − 1 − 0.5 F 󰁩󰁧󰁵󰁲󰁥 󰀷 : PCA score plot analysis of microorganism isolated fromwc and well (pozzo). form a curve perfectly correlated with the time inwhichthemeasurewasdone,showingakindofcurvefrom bacterial growth.  T is growth correspond to  20 hours of EN analysis and the curve showed by thePCA analysis can be compared with the  󿬁 rst steps of thetypicallogarithmicgrowthcurveofthisspeciesof microorganism.In the case of coliforms, it can be observed, in the resultsobtained with GC-MS (Figure  󰀸 ), an increase of the acidcompounds like benzoic, hexanoic acid,  3  methyl butanol,andalsosomecompoundswithanalcoholgrouplikenonanoland acetone.In some cases, EN is able to detect di ff  erences betweensampleswhiletheGC-MSisnotabletorevealaproductionof metabolites created by the growing bacteria.  T at is owing totheconsumption,bythebacteria,ofsomemetabolitespresent 05101520253035    A  c  e   t  o  n  e   2  -   F    l  u  o  r  o   p  r  o   p  a  n  e   3  -   M  e   t    h  y    l    b  u   t  a  n  o    l   F  o  r  m   i  c  a  c   i    d   H  e   p   t  a  n  o    l   B  e  n   z  a    l    d  e    h  y    d  e   O  c   t  a  n  o    l   4  -   M  e   t    h  y    l   t    h   i  a   z  o    l  e 2.453 6.208 10.975 14.923 16.781 17.991 18.230 19.292T0T2T3 × 10 5 F 󰁩󰁧󰁵󰁲󰁥󰀸 :HistogramobtainedfromGC-MS,withSPMEtechniquesabout VOCs produced by wc samples (T 0  to T 2 ). in the medium, making the EN respond in di ff  erent way andrevealing di ff  erences between the samples. 4. Conclusions In this work, some signi 󿬁 cant applications of an electronicnose, based partially on metal oxide nanowires technology,to microbiological food quality control have been review.Literature review has been accompanied by some signi 󿬁 cantcase studies previously presented in this  󿬁 eld by the sameauthors, in order to provide the reader with an enhancedperceptiveness of the EN application.All the described case studies showed promising results,thus con 󿬁 rming that our EN could represent a rapid meanfor controlling and improving the microbiological quality of food.Keeping in mind advantages and limitations, EN doesnot allow replacing human panels or analytical techniquesyet.  T eir ability to smell odours rather than detecting andquantifying speci 󿬁 c volatiles (VOCs) can still be improved.However, they can be used in parallel to those techniques, oreverconsideredasvaluablealternative.Currentmethodology involves conventional technique such as classical microbiol-ogy, visual techniques, or molecular techniques and requiresin most of cases a big amount of time, not always available. T is work attests that the electronic nose, once trained, isa potentialand useful (rapid and economic) tool for the early detection of microbial grows. A kind of sensor technology like a novel EN provides a faster response of the detection of contaminations in food matrix than the conventional tech-niques (also compared with the commercial EN equippedonly with traditional MOX sensors).In some cases the novel EN is able to anticipate thedetection in a few days with respect to the commercial one[ 12 ]. It is possible thanks to the dimensional structure of 
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks