Fingerprinting Green Curry: an Electrochemical Approach to Food

May 29, 2018 - Fingerprinting Green Curry: an Electrochemical Approach to Food Quality Control. Thanyarat Chaibun , Chan La-o-vorakiat , Anthony Peter...
0 downloads 0 Views 3MB Size
Download PDF
Article Cite This: ACS Sens. XXXX, XXX, XXX−XXX

pubs.acs.org/acssensors

Fingerprinting Green Curry: An Electrochemical Approach to Food Quality Control Thanyarat Chaibun,† Chan La-o-vorakiat,‡ Anthony P. O’Mullane,*,§ Benchaporn Lertanantawong,*,‡ and Werasak Surareungchai‡,∥ †

Sensor Technology, PDTI, ‡Nanoscience and Nanotechnology Graduate Program, and ∥School of Bioresources & Technology, King Mongkut’s University of Technology Thonburi, Bangkok 10140, Thailand § School of Chemistry, Physics and Mechanical Engineering, Queensland University of Technology, Brisbane, Queensland 4001, Australia S Supporting Information *

ABSTRACT: The detection and identification of multiple components in a complex sample such as food in a cost-effective way is an ongoing challenge. The development of on-site and rapid detection methods to ensure food quality and composition is of significant interest to the food industry. Here we report that an electrochemical method can be used with an unmodified glassy carbon electrode for the identification of the key ingredients found within Thai green curries. It was found that green curry presents a fingerprint electrochemical response that contains four distinct peaks when differential pulse voltammetry is performed. The reproducibility of the sensor is excellent as no surface modification is required and therefore storage is not an issue. By employing particle swarm optimization algorithms the identification of ingredients within a green curry could be obtained. In addition, the quality and freshness of the sample could be monitored by detecting a change in the intensity of the peaks in the fingerprint response. KEYWORDS: electrochemical fingerprint, food fingerprint, differential pulse voltammetry, composition analysis, food quality control

T

mass spectrometry is also utilized. However, there are a few drawbacks with these gold standard methods. Specific sample preparation steps are required often involving extraction, separation, and filtration from turbid solutions which are not compatible with this approach. The disadvantages are that the times to sampling can be too long and the cost involved can be prohibitive as it requires special equipment operated by highly trained professionals. Therefore, is a constant need to develop alternative techniques that can alleviate these issues. Electrochemical techniques have been employed very successfully in health8−10 and environmental monitoring.11−17 Therefore, electroanalytical chemistry is now emerging as a promising alternative for food analysis,18−21 due to reduced costs, rapid sampling, and easier data collection and interpretation. Recent advances have shown that electrochemical techniques can measure toxins in food using biosensors,19 determining the antioxidant capacity of food and beverages,18 the strength of garlic,23 and the heat of chili peppers.24 In many cases surface modification is required to enhance sensitivity or selectivity, or generation of a redox active species such as bromine is required to interact with the species

he ability to detect and discriminate multiple components in a complex sample matrix is an ongoing challenge in analytical chemistry. One route to solve this problem is using sensor arrays; however, the task becomes even more challenging if a single sensing layer/device is to be employed to ensure simplicity of operation and reduce sensor costs. One area where these concerns are consistently encountered is the food industry which requires sophisticated instrumentation to handle the often complex nature of foodstuffs. Compositional analysis is an indispensable control for this industry to monitor the quality and quantity of ingredients as per regulatory requirements and maintain freshness.1−5 It is also becoming increasingly important to monitor food safety in terms of contamination as well as food adulteration.6 A variety of analytical techniques have been employed in the food industry such as vibrational spectroscopies, nuclear magnetic resonance spectroscopy, mass spectrometry, and chromatographic methods. An excellent review of these technologies applied to fingerprinting food, and in particular the identification of contamination has been previously published.6 In this particular area the analytical technique employed should provide a reliable signature of the constituents and define a “fingerprint” to compare with known standards. For the food as well as the pharmaceutical industries, the gold standard for quantification is liquid chromatography coupled to triple quadrupole mass spectrometers.3,6,7 In addition, gas chromatography coupled to © XXXX American Chemical Society

Received: March 2, 2018 Accepted: May 29, 2018 Published: May 29, 2018 A

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors

Figure 1. (a) Simple approach to obtain the compositional signature of green curry by use of a standard three-electrode electrochemical setup; (b) typical differential pulse voltammogram of a commercial Thai green curry (gray dotted dashed line) showing five prominent peaks which functions as a fingerprint of the sample. The background from coconut milk (green dashed line) is subtracted for the data analysis and the remaining four peaks provide the information about the constituent ingredients.

can be utilized for food quality control in a complex sample by providing a reliable electrochemical identification of the ingredients and is suitable for other products.

of interest. Generally, only one component is identified or if composition is required then individual components are identified separately. Another approach is the concept of the electronic tongue using a sensory array of electrodes that may employ potentiometric (ion selective electrodes), voltammetric, or conductometric sensing elements as well as electrochemical biosensors. To provide multicomponent information chemometrics are employed to analyze the data.25 With regard to food analysis it has been demonstrated that an electronic tongue based on an array of electrochemical sensors can be used for example to monitor the freshness of milk,26 analyze beer and wine,27 as well as classify olive oils.28 However, using electrochemical fingerprinting of multicomponent systems at a single electrode is significantly more challenging and therefore there are fewer examples. Recently, however, Wang et al. demonstrated that an electrochemical sensor could detect the presence of cocaine and cutting agents in street samples,29 while Jakubowska et al. were able to demonstrate a voltammetric method to identify 5 different types of Polish ciders.30 In the latter case partial least-squares discriminant analysis was required to discriminate between samples, whereas in Wang’s work clearly discriminated peaks for the various redox active components in the drug samples were found. In this work, we employ differential pulse voltammetry (DPV), a standard electrochemistry technique, with an unmodified glassy carbon electrode for a fingerprint analysis of curry paste. We chose Thai curries as the test samples for our demonstration as the taste and fragrance from curry pastes are provided by multiple ingredients and it is known that the composition of an identically named dish may differ from home to home or region to region.31 The electrochemical signal displays multiple peaks which allow for the identification and quantitative measurement of the key ingredients in the curry paste. We show that the electrochemical signal is strongly sensitive to the condition of the ingredients which is determined by the suppliers or harvest season as well as exposure of the curry paste to ambient conditions over time. This indicates that this sensor could be used as a quality control check for curry pastes in the food industry prior to packaging and shipment. We demonstrate that a simple sensing platform

■

EXPERIMENTAL SECTION

Chemical and Reagents. All chemicals and reagents were purchased from Sigma/Aldrich and used without purification. Solutions were prepared using deionized water. The electrolyte was a buffer solution comprising 0.1 M sodium acetate-acetic acid which was prepared with sodium acetate trihydrate (CH3COONa·3H2O) and mixed with the glacial acetic acid (HOAc) to optimized the required pH to 4.5. Electrochemical Detection. Electrochemical analysis and data collection were performed with an Autolab PGSTAT 12 computercontrolled potentiostat (Eco Chemie) with GPES software. Electrochemical experiments were set up as a three-electrode cell (Figure 1), with an unmodified 3-mm-diameter glassy carbon working electrode (area 0.0962 cm2), Pt wire counter, and Ag/AgCl reference electrode (Basi, USA). Before each experiment, the glassy carbon electrode was polished with an alumina−water slurry and then sonicated in deionized water. Electrochemical measurements were performed using differential pulse voltammetry (DPV) at room temperature. We followed standard electrochemical practice to optimize the DPV parameters to maximize current signal, peak separation, and background current suppression. The optimized experimental parameters were as follows: applied potential modulation amplitude of 5 mV, modulation time of 0.05 s, interval time of 0.1 s, initial potential of −0.5 V, and end potential of 1.5 V with a step potential of 5 mV. Sample Preparation. We used commercial instant Thai curries and coconut milk purchased as test samples. Then, we cooked the curries according to the package description (heat until boiling and then boil for 5 min, and then the sample was left to stand for 15 min at room temperature). The curry solutions were then filtered through a fabric. A 1.5 mL sample of each filtrate was then mixed with the buffer electrolyte for electrochemical detection. We also prepared our own Thai style green curry based on a typical recipe (presented in Table 1) and the ingredients were prepared by the same method. For characterization of individual ingredients, we mixed the single ingredient in coconut milk in the proportion displayed in Table 1 and performed the same sample preparation method as for the curry samples. After cooling down, we centrifuged at 12 000 rpm for 15 min to remove undissolved components. The supernatants were then stored for future measurements. B

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors Table 1. Standard Recipe for Green Curry

■

ingredients

quantity (gram)

%w/v

chili shallot garlic galangal lemongrass kaffir lime leaf coriander coriander root pepper cumin shrimp paste salt

20 20 20 5 10 2 5 10 3 5 5 5

18.2 18.2 18.2 4.5 9.1 1.8 4.5 9.1 2.7 4.5 4.5 4.5

RESULTS AND DISCUSSION A typical differential pulse voltammogram (DPV) of a commercial green curry (Roithai) (black dashed line) measured using a glassy carbon electrode is shown in Figure 1b and displays five distinct peaks. The first peak was found to be due to coconut milk which was run as the background. Therefore, for all subsequent experiments the data was background subtracted using the coconut milk signal. This resulted in four distinct peaks being observable in the potential region of study (green line in Figure 1b). The peaks are located at 0.4, 0.6, 0.9, and 1.1 V where only a minor shift in peak position is observed after the background subtraction. Here we assume that the remaining four peaks are due to redox active species present in the green curry. Given the nature of the ingredients employed in green curry pastes, identifying the origin of the exact species responsible for the observed redox activity is extremely challenging. It has been shown that the heat in chili peppers is due to capsaicinoids which are redox active.24 However, green curry contains complex ingredients such as shrimp paste, for instance, which is a fermented product and therefore has numerous components such as carotenoids, peptides, protein, and salt.32 Therefore, the aim of this work is not to identify the redox active species responsible for the observed voltammetry but quantify the contribution of each ingredient to the DPV response observed in Figure 1b and determine if a sensor can be developed to identify key ingredients in the curry paste. To probe the consistency of the response we measured the response of green curry sourced from different batches of the same brand (Roitai) from different locations used in Figure 1b. The data shown in Figure 2a shows that there is excellent consistency between batches (n = 9) of the same curry and in principle implies that there is consistency in the ingredients used in the curry that contribute to an electrochemical signal. The next step was to measure samples of green curry from 4 different brands (Knorr, Lobo, Maeploy, Rosdee) as well as 2 samples from street stores (located in Tha Kham, Bangkok) which provided their own formulation of green curry (labeled as unbranded A and B). A list of all the key ingredients making up these curry pastes is provided in tabular form in the Supporting Information. It can be seen that the responses are clustered at the 4 peak positions identified in Figure 1b. Some samples only show 3 distinct components, such as Rosdee with an absence of peak 3 and Lobo and Maeploy not exhibiting a distinct process in the peak 4 position. There is also a significant amount of variation over the potential range from 0.5 to 0.8 V which covers the potential region of peak 2 (0.6 V).

Figure 2. DPVs from (a) batches of the same commercial Thai green curry, (b) Thai green curries from different brands, and (c) various curry types from the same commercial brand with the slightly different ingredient composition.

These results suggest that there is a significant composition difference between the different brands of green curry as expected. Interestingly, in contrast, the study undertaken among different curry types of the same commercial brand (Roithai) are significantly more similar as the ingredients are expected to come from the same sources but at different concentration levels. These three results indicate the possibility of using DPV at an unmodified glassy carbon electrode as an analytical tool for compositional analysis. The next aspect to this study was to investigate the electrochemical response for each of the individual ingredients (Figure 3a) typically used in a green curry paste as detailed in Table 1. For the majority of the ingredients one distinct peak can be observed along with some other minor responses which in principle makes ingredient identification more feasible. This is similar to previous work where the electrochemical detection of cocaine was performed in the presence of a variety of cutting agents. In that study distinct peaks were also seen for the individual components on a screen printed carbon electrode.29 For the region of peak 1, there are strong signals from coriander and lemongrass, peak 2 has major contribution from cumin and shallot and less so from chili and kaffir lime leaf, peak 3 has strong contributions from fish sauce and shrimp paste and a minor contribution from coriander and shallot, and peak 4 has major contributions from garlic and pepper and a minor contribution from galangal. All of the ingredients in Table 1 were used to make a standard curry paste and the resultant DPV of that sample is shown in Figure 3b and compared with a commercial product (Rothai). It can be seen that a reasonable correlation can be achieved between the commercial product and one formulated in the laboratory. It is believed that there is likely to be a mismatch in the exact composition (unknown due to the trade secret) which leads to the discrepancy in the peak C

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors

Figure 3. (a) Distinctive DPV responses of the ingredients that contribute to the total signal of the green curry (constant offsets are added for visibility), (b) mixture of all ingredients shows similar peak locations to a commercial product, and (c) quantitative analysis of the DPV peak height showing the distinctive signature of each ingredient (surrounding small subplots with ingredient name on the top) that are normalized to the most pronounced peak.

heights. In addition, the individual responses for the main ingredients compared to the signal observed for a green curry sample are also overlaid in Figure S1 which indicates that interfering effects between the ingredients are not extensive when mixed into a paste and cooked. However, it should be noted that there is likely to be some chemical interaction between the individual components. To further identify the contribution from each component beyond the visual inspection of DPVs, we employed a simple analysis based on a linear combination and global optimization methods. By using a global optimization method, the ingredient

peak signatures allow for the breakdown of the green curry peaks to the individual contributions from each ingredient. By assuming that the four DPV peaks of green curry are independent, we take them as a basis of the linear combination leading to the measured DPV of each ingredient (Φj = ∑i 4= 1cijei) where cij is the amplitude of the ith peak of jth ingredient. Due to the small effects of the sample matrix, the peak position are shifted slightly as noted above and we define the peak with respect to a range of potentials (as indicated by the label on top of Figure 3a). By normalizing to the highest peak value, the DPV peak structure of each ingredient can be D

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors represented (Figure 3c, surrounding small plots) by four numbers. Next the composition of the fingerprints could be broken down by assuming that the DPV response of the curry can also be modeled as a linear combination of the individual ingredient DPV curve (Φcurry = ∑j 12= 1CjΦj). The peaks of green curry are shown in the central figure of Figure 3c and displayed as percent contributions to total peak heights. The problem now becomes an optimization problem to find the set of coefficients (Cj) of each ingredient such that the root-mean-square deviation for the model from experimental data is minimized. We chose a global optimization method based on a particle swarm optimization algorithm to avoid the false solution from local minima by use of random initial guesses.33,34 The optimized coefficients provide a small residue on the order of 10−7 and are independent from the random initial guesses (Figure S2). The optimized results show the dominant contributions to peak 3 and 4 are from shrimp paste and garlic, respectively, while peak 1 has an equal contribution from coriander and lemongrass and peak 2 is dominated by contributions from cumin and shallot and less so from chili and kaffir lime leaf. For the quantitative determination of an individual ingredient in terms of mass or concentration, the DPV signal should scale linearly with the ingredient quantity. Therefore, as an example we constructed a calibration curve for shrimp paste in both coconut milk as well as in green curry soup; note all DPV data is corrected for the coconut milk signal. The concentration of ingredients was prepared from lower to higher concentrations of the labeled amount in Table 1. In coconut milk a welldefined peak at 0.85 V was observed (Figure 4b) where only a slight shift to 0.89 V was observed for the peak in the green curry sample (Figure 4a). Upon increasing the amount of shrimp paste a clear increase in the intensity of the peak was observed in both cases. The standard deviation is only ca. 1.2% at each concentration value (n = 3) which indicates high reproducibility. Excellent calibration plots were achieved with R2 = 0.991 and 0.987 for the coconut milk and green curry samples, respectively, and shows that the response is sensitive only to a change in the ingredient regardless of any matrix effect that is occurring. This approach was taken for the other main ingredients in green curry which is highlighted for garlic, cumin, and green chili in Figure 4d−f which also gave excellent linear responses with R2 values of 0.925, 0.974, and 0.996, respectively. This data is tabulated in Table 2 showing the sensor sensitivity in green curry, coconut milk, linear concentration ranges, and % w/v. A major advantage of this approach is that an unmodified GC electrode is required as the sensing element and therefore storage and stability of the electrode are not an issue. This is an advantage over biosensors which use modified electrode surfaces, in particular, with biological entities which have a limited shelf life. In addition, excellent reproducibility was achieved as GC electrodes can be polished to ensure good surfaces. Although the sensing element is straightforward it shows that unmodified carbon surfaces can be applied to complex samples such as that investigated here and by Wang for cocaine detection. This suggests that possibly other complex sensing environments could be studied using this easy to use approach. Another key aspect in food quality control is identification of possible spoilage of the food product. It must be stated that there are sensors, in particular, biosensors, developed to detect toxins and pathogens;19,22,35,36 however, we employed this

Figure 4. DPVs recorded after the addition of shrimp paste to (a) commercial green curry, (b) coconut milk, and calibration curves showing current signal obtained in green curry and coconut milk of (c) shrimp paste, (d) garlic, (e) cumin, and (f) green chili.

technique to observe if any significant differences in the electrochemical response could be detected with time which may provide another route to detect food spoilage. Illustrated in Figure 5 are the DPV responses for green curry at various times after opening. It can be seen that a change can only be observed after 1 day which becomes more evident after 3 days. There is a significant reduction in the peak intensities for peaks 3 and 4 with time which are associated with garlic and shrimp paste. This may offer insights into the mechanism of degradation of such food products but is not the focus of this study. However, it does provide an alternative marker on the freshness of curry pastes. Often additives such as sodium benzoate, known as E211, are included in food products as a preservative.37 Therefore, it was added to a green curry sample to see its effect on the DPV response (Figure S3). An additional peak was observed at ca. 1.6 V which is more positive than the potential range where the main DPV peaks are recorded for green curry and therefore does not interfere with the sensor. In fact this may be advantageous and offer a way to detect the presence of E211 in a green curry sample.

■

CONCLUSIONS We have demonstrated that a simple electrochemical setup using an unmodified glassy carbon electrode as the sensing layer can be used to determine the principal ingredients of Thai green curries. Differential pulse voltammetry was employed to ensure a good signal-to-noise ratio. Four distinct peaks are E

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors

Table 2. Sensor Data for Individual Ingredients in Green Curry Where Uncertainties Were Evaluated at 95% Confidence Levels ingredients

sensitivity in coconut milk

sensitivity in green curry

linear range (g)

concentration range (% w/v)

% w/v in green curry (Table1)

shrimp paste garlic cumin green chili

2.05(22) μA g−1 155(33) nA g−1 1.67(51) μA g−1 136(15) nA g−1

1.03(14) μA g−1 124(41) nA g−1 0.85(16) μA g−1 95(7) nA g−1

0.01−0.30 0.10−1.20 0.01−0.30 0.10−1.20

1−30 10−120 1−30 10−120

4.5 18.2 4.5 18.2

Benchaporn Lertanantawong: 0000-0003-0086-8371 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the KMUTT 55th Anniversary Commemorative Fund and B.L. gratefully acknowledges the Thailand Research Fund (Grant number: RSA6080050) for financial support.

■

Figure 5. Decay of DPV signals to signify the decay of a green curry sample as a function of exposure time in ambient atmosphere.

observed for different types of green curry as well as different curry types from the same manufacturer which provides a fingerprint signal. With a distinct fingerprint response it allows the composition and quality of the product to be determined. When a particle swarm optimization algorithm is employed the composition of a sample can be obtained. Therefore, using a set of common ingredients the response for a commercial green curry could be attained. In addition, the quality of the product could be monitored, as with exposure time there is a decrease in the intensity of the peaks which could be used to monitor freshness if desired. The reproducibility of the response was excellent with a standard deviation of 1.2%, while storage issues are not a concern as the surface is simply carbon which can be easily polished between sampling. This work shows that a glassy carbon electrode is applicable to the analysis of a complex sample such as that found in green curry and could be of use for other food samples.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssensors.8b00176. Differential pulse voltammograms (DPV) of selected ingredients before subtracting out the coconut milk (background) contribution, the reliability of optimized coefficients over random sets of initial conditions, DPV responses in the presence of sodium benzoate, and the ingredients of all curries discussed in the work (PDF)

■

REFERENCES

(1) Karoui, R.; Downey, G.; Blecker, C. Mid-Infrared Spectroscopy Coupled with Chemometrics: A Tool for the Analysis of Intact Food Systems and the Exploration of Their Molecular Structure−Quality Relationships − a Review. Chem. Rev. 2010, 110, 6144−6168. (2) Smyth, H.; Cozzolino, D. Instrumental Methods (Spectroscopy, Electronic Nose, and Tongue) as Tools to Predict Taste and Aroma in Beverages: Advantages and Limitations. Chem. Rev. 2013, 113, 1429− 1440. (3) Cacciola, F.; Donato, P.; Sciarrone, D.; Dugo, P.; Mondello, L. Comprehensive Liquid Chromatography and Other Liquid-Based Comprehensive Techniques Coupled to Mass Spectrometry in Food Analysis. Anal. Chem. 2017, 89, 414−429. (4) Nagana Gowda, G. A.; Raftery, D. Recent Advances in NMRBased Metabolomics. Anal. Chem. 2017, 89, 490−510. (5) Su, W.-H.; Sun, D.-W. Fourier Transform Infrared and Raman and Hyperspectral Imaging Techniques for Quality Determinations of Powdery Foods: A Review. Compr. Rev. Food Sci. Food Saf. 2018, 17, 104−122. (6) Ellis, D. I.; Brewster, V. L.; Dunn, W. B.; Allwood, J. W.; Golovanov, A. P.; Goodacre, R. Fingerprinting Food: Current Technologies for the Detection of Food Adulteration and Contamination. Chem. Soc. Rev. 2012, 41, 5706−5727. (7) Lutter, P.; Savoy-Perroud, M. C.; Campos-Gimenez, E.; Meyer, L.; Goldmann, T.; Bertholet, M.-C.; Mottier, P.; Desmarchelier, A.; Monard, F.; Perrin, C.; Robert, F.; Delatour, T. Screening and Confirmatory Methods for the Determination of Melamine in Cow’s Milk and Milk-Based Powdered Infant Formula: Validation and Proficiency-Tests of Elisa, HPLC-UV, GC-MS and LC-MS/MS. Food Control 2011, 22, 903−913. (8) Gong, J. R. Label-Free Attomolar Detection of Proteins Using Integrated Nanoelectronic and Electrokinetic Devices. Small 2010, 6, 967−973. (9) Xin, Q.; Liu, Q.; Shah, H.; Gong, J. R. Electron spin resonance and fluorescence imaging assisted electrochemical approach for accurate and comprehensive monitoring of cellular hydrogen peroxide dynamics. Analyst 2017, 142, 316−325. (10) Vijian, D.; Chinni, C. S.; Lee, S. Y.; Lertanantawong, B.; Surareungchai, W. Non-protein coding RNA-based genosensor with quantum dots as electrochemical labels for attomolar detection of multiple pathogens. Biosens. Bioelectron. 2016, 77, 805−811. (11) Gumpu, M. B.; Sethuraman, S.; Krishnan, U. M.; Rayappan, J. B. B. A Review on Detection of Heavy Metal Ions in Water − an Electrochemical Approach. Sens. Actuators, B 2015, 213, 515−533. (12) Arduini, F.; Cinti, S.; Scognamiglio, V.; Moscone, D.; Palleschi, G. How Cutting-Edge Technologies Impact the Design of Electrochemical (Bio)Sensors for Environmental Analysis. A Review. Anal. Chim. Acta 2017, 959, 15−42.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Anthony P. O’Mullane: 0000-0001-9294-5180 F

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX

Article

ACS Sensors (13) Maduraiveeran, G.; Jin, W. Nanomaterials Based Electrochemical Sensor and Biosensor Platforms for Environmental Applications. Trends Environ. Anal. Chem. 2017, 13, 10−23. (14) Ben Messaoud, N.; Ghica, M. E.; Dridi, C.; Ben Ali, M.; Brett, C. M. A. Electrochemical Sensor Based on Multiwalled Carbon Nanotube and Gold Nanoparticle Modified Electrode for the Sensitive Detection of Bisphenol A. Sens. Actuators, B 2017, 253, 513−522. (15) Sarfo, D. K.; Izake, E. L.; O’Mullane, A. P.; Ayoko, G. A. Molecular Recognition and Detection of Pb(II) Ions in Water by Aminobenzo-18-Crown-6 Immobilised onto a Nanostructured SERS Substrate. Sens. Actuators, B 2018, 255, 1945−1952. (16) Yu, H. A.; DeTata, D. A.; Lewis, S. W.; Silvester, D. S. Recent Developments in the Electrochemical Detection of Explosives: Towards Field-Deployable Devices for Forensic Science. TrAC, Trends Anal. Chem. 2017, 97, 374−384. (17) Hussain, G.; Silvester, D. S. Detection of Sub-ppm Concentrations of Ammonia in an Ionic Liquid: Enhanced Current Density Using “Filled” Recessed Microarrays. Anal. Chem. 2016, 88, 12453−12460. (18) Hoyos-Arbeláez, J.; Vázquez, M.; Contreras-Calderón, J. Electrochemical Methods as a Tool for Determining the Antioxidant Capacity of Food and Beverages: A Review. Food Chem. 2017, 221, 1371−1381. (19) Bunney, J.; Williamson, S.; Atkin, D.; Jeanneret, M.; Cozzolino, D.; Chapman, J.; Power, A.; Chandra, S. The Use of Electrochemical Biosensors in Food Analysis. Curr. Res. Nutr. Food Sci. 2017, 5, 183− 195. (20) Maduraiveeran, G.; Ramaraj, R. Gold Nanoparticle-Based Sensing Platform of Hydrazine, Sulfite, and Nitrite for Food Safety and Environmental Monitoring. J. Anal. Sci. Technology 2017, 8, 14. (21) Plowman, B. J.; Bhargava, S. K.; O’Mullane, A. P. Electrochemical Fabrication of Metallic Nanostructured Electrodes for Electroanalytical Applications. Analyst 2011, 136, 5107−5119. (22) Rapini, R.; Marrazza, G. Electrochemical Aptasensors for Contaminants Detection in Food and Environment: Recent Advances. Bioelectrochemistry 2017, 118, 47−61. (23) Martindale, B. C. M.; Aldous, L.; Rees, N. V.; Compton, R. G. Towards the Electrochemical Quantification of the Strength of Garlic. Analyst 2011, 136, 128−133. (24) Kachoosangi, R. T.; Wildgoose, G. G.; Compton, R. G. Carbon Nanotube-Based Electrochemical Sensors for Quantifying the ’Heat’ of Chilli Peppers: The Adsorptive Stripping Voltammetric Determination of Capsaicin. Analyst 2008, 133, 888−895. (25) del Valle, M. Electronic Tongues Employing Electrochemical Sensors. Electroanalysis 2010, 22, 1539−1555. (26) Winquist, F.; Krantz-Rülcker, C.; Wide, P.; Lundström, I. Monitoring of Freshness of Milk by an Electronic Tongue on the Basis of Voltammetry. Meas. Sci. Technol. 1998, 9, 1937. (27) Nery, E. W.; Kubota, L. T. Integrated, Paper-Based Potentiometric Electronic Tongue for the Analysis of Beer and Wine. Anal. Chim. Acta 2016, 918, 60−68. (28) Peres, A. M.; Veloso, A. C. A.; Pereira, J. A.; Dias, L. G. Electrochemical Multi-Sensors Device Coupled with Heuristic or Meta-Heuristic Selection Algorithms for Single-Cultivar Olive Oil Classification. Procedia Eng. 2014, 87, 192−195. (29) de Jong, M.; Sleegers, N.; Kim, J.; Van Durme, F.; Samyn, N.; Wang, J.; De Wael, K. Electrochemical Fingerprint of Street Samples for Fast on-Site Screening of Cocaine in Seized Drug Powders. Chem. Sci. 2016, 7, 2364−2370. (30) Górski, Ł.; Sordoń, W.; Ciepiela, F.; Kubiak, W. W.; Jakubowska, M. Voltammetric Classification of Ciders with PLS-DA. Talanta 2016, 146, 231−236. (31) Siripongvutikorn, S.; Penseng, N.; Ayusk, S.; Usawakesmanee, W. Development of Green Curry Paste Marinade for White Shrimp (Litopenaeus Vannamei). J. Sci. Technology 2008, 30, 35−40. (32) Pongsetkul, J.; Benjakul, S.; Sampavapol, P.; Osako, K.; Faithong, N. Chemical Composition and Physical Properties of Salted Shrimp Paste (Kapi) Produced in Thailand. Internat. Aquatic Res. 2014, 6, 155−166.

(33) Kennedy, J.; Eberhart, R. Particle Swarm Optimization. Proceedings of IEEE International Conference on Neural Networks. IV 1995, 1942−1948. (34) Shi, Y.; Eberhart, R. A Modified Particle Swarm Optimizer. Proceedings of IEEE International Conference on Evolutionary Computation 1998, 69−73. (35) Rotariu, L.; Lagarde, F.; Jaffrezic-Renault, N.; Bala, C. Electrochemical Biosensors for Fast Detection of Food Contaminants − Trends and Perspective. TrAC, Trends Anal. Chem. 2016, 79, 80−87. (36) Fiddes, L. K.; Chang, J.; Yan, N. Electrochemical Detection of Biogenic Amines During Food Spoilage Using an Integrated Sensing RFID Tag. Sens. Actuators, B 2014, 202, 1298−1304. (37) World Health Organization & International Programme on Chemical Safety. (2000). Benzoic acid and sodium benzoate. WHO, Geneva. Available from: [http://www.who.int/iris/handle/10665/ 42310].

G

DOI: 10.1021/acssensors.8b00176 ACS Sens. XXXX, XXX, XXX−XXX