Analytica Chimica Acta 864 (2015) 64–73

Contents lists available at ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

A single cataluminescence sensor based on spectral array and its use in the identification of vinegars Jiayi Zeng, Xiaoan Cao * , Yonghui Liu, Jinglin Chen, Keke Ren Environmental Science and Engineering Institute, Guangzhou University, 510006 Guangzhou, China

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

 The single CTL sensor for vinegar fast discrimination was simple and stable.  The spectral arrays of vinegars served as their fingerprints.  9 types and 8 brands of vinegar and artificial samples were successfully identified.  The single sensor was capable of discriminating very similar complex mixtures.

The spectral arrays of vinegars here served as their fingerprints. Facile discrimination of vinegars was made possible by these unique patterns.

A R T I C L E I N F O

A B S T R A C T

Article history: Received 19 October 2014 Received in revised form 17 January 2015 Accepted 22 January 2015 Available online 23 January 2015

The discrimination of complex mixtures, especially those with very similar compositions, remains a challenging part of chemical analysis. In this paper, a single cataluminescence (CTL) sensor constructed using MgO nanomaterials in a closed reaction cell (CRC) was used to identify vinegars. It may provide an archetype of this type of highly multicomponent system. By scanning the CTL spectra, which were distributed in 15 wavelengths during the reaction period, the spectral array patterns of the vinegars were obtained. These functioned as their fingerprints. The CTL signals of the array were then normalized and identified through linear discrimination analysis (LDA). Nine types and eight brands of vinegars and an additional series of artificial samples were tested; the new technique was found to distinguish between them very well. This single sensor demonstrated excellent promise for analysis of complex mixtures in real-world applications and may provide a novel method for identifying very similar complex analytes. ã 2015 Elsevier B.V. All rights reserved.

Keywords: Vinegar Sensor Cataluminescence Chemiluminescence Spectra

1. Introduction Sensor technology provides a versatile approach to chemical analysis. During the past several decades, increasing interest has been drawn to the development of different sensors for specific purposes. The artificial olfactory systems, known as electronic noses and artificial noses, are composed of a group of sensor arrays. They are used for the identification of both simple and complex gas

* Corresponding author. Tel.: +86 20 39366937; fax: +86 20 39366946. E-mail address: [email protected] (X. Cao). http://dx.doi.org/10.1016/j.aca.2015.01.035 0003-2670/ ã 2015 Elsevier B.V. All rights reserved.

mixtures [1–3]. They have become a common tool for food quality assessment with the advantages of high portability for in situ and on-site testing. They also tend to be low-cost and reliable [4,5]. Studies addressing the evaluation of foodstuffs such as vinegar [6,7], wine [8,9], coffee [10], and tea [11] using electronic noses that have a few to a dozen sensor elements have reported success. However, multiple sensing units are not conducive to instrument stability [12–14]. Alterations of the characteristics of the sensors unit over time affect the final testing results [15]. In order to improve the reproducibility and reduce the need for frequent calibration, a sensor system with as few sensing units as possible is greatly needed.

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

In recent years, chemiluminescence (CL)-based detection has become quite a useful detecting tool in the field of food analysis [16,17]. This kind of optical system may facilitate collection abundant information simultaneously, including signal intensity, luminescence lifetime, wavelength, spectral shape, and other factors [18,19]. Nanomaterial-assisted cataluminescence (CTL) sensors can provide stable and reversible responses because their solid sensing nanomaterials are essentially non-expendable during the sensing process [20–23]. Analytes can be wiped out through thermal catalytic degradation after detection. In addition, no light source was needed in CTL sensors so simple instrumentation can be made. These sensors are low-cost and easy to operate, making them attractive for many industrial applications [24–26]. For the discrimination of complex and similar mixtures, one may usually consider a sensor-array approach rather than a single sensor. However, a recent report indicated a novel method for the identification of complex mixtures using a single CTL sensor in a closed reaction cell (CRC). First, 12 medicines were graphically recognized based on their characteristic, multiple-peak CTL response profiles [27]. In order to explore the ability of our single sensor to discriminate other complex mixtures, 72 wine samples were examined as a second trail [28]. By extracting the CTL intensities at different points in time in the response curves as variables and adopting an appropriate algorithm, 72 samples were all assigned to the correct groups. In order to further improve discrimination among highly similar complex mixtures, a method of CTL spectral array based on the single sensor in CRC was here developed. Commercially available vinegars served as an archetype of such complex analytes. Vinegar is a popular seasoning worldwide, and its quality directly affects people's health. Many methods of analysis are highly accurate and suitable for assessment of vinegar quality [29–36]. However, most of these methods are costly, require long processing times and focused operators, making them inconvenient for on-site discrimination [37]. A new, fast method of identifying vinegars is necessary. In the present study, the spectral signals of vinegar on nanosized MgO in CRC were obtained by scanning the CTL spectra repeatedly during the reaction period. These were classified into 15 groups according to the signals' distribution at 15 wavelengths. These were named the spectral array. Then they were normalized and subjected to LDA to classify different vinegars. Vinegars of different types and different brands were discriminated with 100% accuracy. The adulterated samples were distinguished from uncontaminated ones using this method. This simple single sensor demonstrates excellent potential for the analysis of truly complex mixtures in real-life applications and provides a novel approach to identification of complex samples of relatively similar compounds. 2. Experimental 2.1. Materials Nano-MgO (purity, 99.99%; average particle size, 50 nm) was purchased from Beijing Nacheng Scientific Trading Co., Ltd. (Beijing, China). Nano-ZrO2 (purity, 99.99%; average particle size, 10 nm) was purchased from Shanghai Zhuerna High-Tech Powder Material Co., Ltd. (Shanghai, China). Nano-Y2O3 (purity, 99.99%; average particle size, 30 nm) was purchased from Guangdong Huizhou Ruier Chemical Technology Co., Ltd. (Guangdong, China).

65

vinegars were divided into 9 types: baoning vinegar, mature vinegar, red vinegar, sweetened vinegar, balsamic vinegar, fumigated vinegar, white vinegar, apple vinegar, and rice vinegar. The 8 brands of apple vinegar were labeled, “Huashengtang apple vinegar”, “Lecufang apple vinegar”, “Jingcufang apple vinegar”, “Tiandiyihao apple vinegar”, “Youyang apple vinegar”, “Pinger vinegar”, “Maijinli apple vinegar”, and “Hongyuan apple vinegar”. The samples were stored in the laboratory at a constant temperature of 25  1  C and used directly from their containers without dilution. Synthesized vinegar consisted of 3.5% acetic acid (analyticalgrade), 1.5% lactic acid (analytical-grade), 3% ammonium sulfate (analytical-grade), 1% caramel, 1% maltose, and 90% distilled water according to the work of Qiu [38]. Three different adulterated baoning vinegar samples were obtained by adding 25, 50, and 75 mL/100 mL of the synthesized vinegar. These adulterations were carried out to produce three different adulterated mature vinegar samples. The 4% acetic acid was obtained by diluting the 36% acetic acid (analytical-grade) with distilled water. 2.3. Sensor fabrication Fig. 1 shows the CTL detection system. The system includes three parts: (1) a CTL chamber (a cylindrical ceramic heater of 5 mm in diameter sintered with 0.5 mm layers of materials was placed in the middle of a quartz tube with an inner diameter of 8 mm. This tube had a sampling port and a gas inlet and outlet with valve switches). (2) A BPCL ultra weak luminescent analyzer (Model BPCL-GP-TGC, Biophysics Institute of Chinese Academy of Science, Beijing, China). There was wheelwork equipped with 15 interference optical filters (1.5 cm in diameter, broadband 24 nm). These filters showed 15 wavelengths (400, 412, 425, 440, 460, 475, 490, 505, 520, 535, 555, 575, 590, 605, and 620 nm) in the detector (Biophysics Institute of Chinese Academy of Science, Beijing, China). As the wheelwork rotated during the measurement process, the CTL intensities of these 15 wavelengths were measured for each sample. The CTL response profiles of vinegars were determined as the wheelwork was set under each fixed wavelength. The resultant CTL signals were detected and recorded using a photomultiplier tube (Model GDB23, Beijing Nuclear Instrument Factory, Beijing, China). (3) A digital temperature controller (Zhihong Electronic Co., Ltd., Beijing, China). The surface temperature of the nanomaterials was controlled by adjusting the voltage of the heating cylindrical ceramic rod. 2.4. Measurement procedures Vinegar samples were sprayed on the nano-MgO surface through the sampling port when the two valve switches were closed and then oxidized at a given temperature. The rotation of the wheelwork continued in a loop during the testing process, and CTL signals under the 15 wavelengths were recorded. In this way, the signals’ spectral array was obtained for 15 wavelengths per sample. The test lasted 500 s, so the number of rotations was set to 100 cycles and the rotation speed to 5 s per cycle (s c 1). The samples’ CTL response profiles were determined for each given wavelength by setting the wheelwork at a fixed wavelength. In order to remove previous residues on the materials, the sensor was heated at 500  C for 15 min in air before each measurement (a steady airstream providing by an air pump (Model GA-2600A, Beijing Zhongxing Huili Co., Ltd., Beijing, China) flew through the quartz tube when the two valve switches were unscrewed).

2.2. Vinegars and adulterated samples 2.5. Data processing and analysis Seventeen commercial vinegars (9 types and 8 brands), which could cover an important range of types of vinegar available on the Chinese market, were purchased from a local supermarket. The

The signals of spectral array for each sample were determined 5 times and the average was used for analysis. The CTL intensities

[(Fig._1)TD$IG]

66

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

Fig. 1. Schematic of the cataluminescence (CTL)-based sensor.

were normalized. The strongest CTL intensity among the 15 wavelengths was converted to 1, and the others were converted into decimals in proportion. These 15 normalized CTL intensities were treated as variables and subjected to linear discrimination analysis (LDA) in SPSS (version 20.0) for discrimination. The data of the CTL response profiles were obtained in triplicate and the mean of the three tests was used for analysis. The corresponding response values of different points in time of the curves were extracted as eigenvalues and then subjected to LDA for identification [28]. 3. Results and discussion 3.1. Design of the spectral array sensor with CRC Previous reports have described a single CTL sensor with CRC for discrimination of analytes, in which the analytes and catalysts repeatedly came into contact with each other [27,28]. The emission periods of the reactants and intermediate products were relatively long, allowing the production of characteristic CTL response profiles (at a fixed wavelength). Both single-component and multicomponent substances have achieved satisfactory classification by the response profiles. However, it has been limited to the recognition of highly similar complex mixtures. In order to collect more information and improve the ability of single sensors to distinguish specific substances, a novel method based on spectral array was developed in this paper. In this method, the CTL spectra of the analytes were rapidly and repeatedly scanned (once every 5 s for 15 wavelengths) during the emission periods (about 500 s), and then all the collected CTL intensities of each wavelength were superimposed. The spectral arrays of 15 wavelengths, which could comprehensively reflect the spectral characteristics of the analytes in the reaction period, were obtained. By dividing the CTL signals into the groups according to different wavelengths, the dimensionality of the sensor data was increased. This has greatly broadened its ability to discriminate among complex samples of relatively similar compounds. The 15 CTL intensities in the spectral array were then normalized and identified through LDA. By normalizing the CTL intensities, the

signals’ little drift in continuous measurement could be compensated. The reaction temperature of 250  C and injection volume of 10 mL were chosen as optimal conditions for vinegar discrimination in the present study. Detailed discussions are shown in the Supporting information.

3.2. Discriminatory ability of the single sensor 3.2.1. Type discrimination of vinegars To evaluate the discriminatory ability of this sensor, 9 types of vinegar (baoning vinegar, mature vinegar, red vinegar, sweetened vinegar, balsamic vinegar, fumigated vinegar, white vinegar, apple vinegar, and rice vinegar) were chosen to be catalyzed by nanoMgO. As described in measurement procedures, the spectral signals of them were measured using this single sensor. The characteristic spectral array of 15 wavelengths in the emission periods is shown in Fig. 2. The 15 normalized CTL intensities of different wavelengths for each vinegar function as a fingerprint for itself. The figure shows these differences graphically. The CTL intensities of most vinegars were the strongest at 535 nm, but the intensities of sweetened and white vinegar were the greatest at 505 and 620 nm, respectively. Mature and apple vinegar showed similar distribution patterns at 400, 412, 425, 440, 460, 475, 490, 505, 520, 535, and 555 nm, but they were very different at 575, 590, 605, and 620 nm; moreover, the CTL intensities of white vinegar were more evenly distributed across all of the wavelengths than other 8 types of vinegars. However, the spectral arrays among mature vinegar, balsamic vinegar, and rice vinegar were more similar to each other than to other kinds of vinegar. However, the CTL intensities of balsamic vinegar at 490 nm and 505 nm were stronger than those of mature vinegar and rice vinegar. The CTL intensities at 400, 412, 425, and 440 nm of mature vinegar were weaker than those of rice vinegar. In this way, the facile discrimination of one vinegar from the others was achieved based on their unique patterns in Fig. 2. Different vinegars exhibited different CTL spectral arrays. They are essentially different complex mixtures, and they are composed of different substrates. For example, baoning vinegar is mainly

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

[(Fig._2)TD$IG]

67

Fig. 2. Spectral arrays of 15 wavelengths for 9 types vinegars on nano-MgO surfaces in the closed reaction cell (working temperature: 250  C; injection volume: 10 mL; rotation speed: 5 s c 1; rotation number: 100 cycles).

fermented from rice and corn, and mature vinegar is mainly produced from sorghum and bran. According to the generally accepted theory of chemiluminescence, energy released by the catalytic reaction can lead to the formation of excited intermediates, and different CTL signals with corresponding wavelengths are produced when different intermediates decay to the ground state. In this highly multicomponent system, when vinegars were catalytically oxidized in the CRC, multiple active intermediates were successively generated over the course of the CTL reactions. The spectra from the excited intermediates were produced and superimposed during the testing process. The superimposed spectra were characteristic of each sample with specific compositions. Different vinegars were identified based on the characteristic spectral array. These results were consistent with previous findings showing that different analytes could exhibited different CTL signals when they reacted on the same nanomaterial [39]. In order to assess the ability of this sensor to distinguish different types of vinegar, the normalized CTL intensities (15 wavelengths  9 vinegars  5 replicates) were subjected to LDA. The first three canonical factors contained 64.1%, 17.0%, and 12.3% of the variation, occupying 93.4% of total variation. As shown in Fig. 3a, the canonical patterns were clustered into 9 different groups at an accuracy level of 100%, which indicated that the samples were well discriminated. In order to determine the model’s reliability, the model achieved was validated by crossvalidation (leave-one-out validation). Results showed 100% success in classification. The present method could be applied to classify vinegars of different types.

The spectral array of CTL signals was found to be better suited to the assessment of this system than that of the CTL response profiles at a fixed wavelength; the CTL response profiles of these 9 vinegars were tested under a detection wavelength of 425 nm (Fig. 4). As expected, each vinegar showed a unique CTL response profile. The extreme points in the curves were marked and the response values that represented these different points in time were extracted. In this way 16 eigenvalues (20, 60, 75, 90, 110, 130, 150, 170, 190, 210, 240, 260, 290, 320, 350, and 400 s) were obtained and then subjected to LDA. The CTL response values were transformed into canonical scores (Fig. 3b), which were visualized as a three-dimensional (3D) plot with the first three functions of the data. It was clustered with the classification accuracy of 96.3%. The results of the cross-validation showed 88.9% success. The model constructed using these response profiles was not as reliable as the one constructed using spectral arrays. This was because the accuracy of classification was relatively low (96.3% and 100%, respectively). In other words, the present method was found to be superior to the previous method with respect to type discrimination of vinegars. 3.2.2. Brand discrimination of the same types of vinegar Apple vinegar is a popular drink for people of all ages and there are many brands on the market. The internal qualities of apple vinegars differ across apple varieties, production technology, and manufacturers. One purpose of this part of the work was identifying apple vinegars of different brands. As shown in Fig. 5, though the strongest CTL intensities of these 8 brands samples all appeared at 535 nm, each sample showed a

68

[(Fig._3)TD$IG]

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

Fig. 3. Canonical score plot of the first three factors analyzed by LDA for the 9 types of vinegars. (a) Spectral arrays of 15 wavelengths. (b) CTL response profiles of 425 nm.

[(Fig._4)TD$IG]

Fig. 4. CTL time-course profiles of 9 types vinegars on nano-MgO surfaces in the closed reaction cell (working temperature: 250  C; detection wavelength: 425 nm; injection volume: 10 mL).

[(Fig._5)TD$IG]

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

69

Fig. 5. Spectral arrays of 15 wavelengths for 8 brands vinegars on nano-MgO surfaces in the closed reaction cell (working temperature: 250  C; injection volume: 10 mL; rotation speed: 5 s c 1; rotation number: 100 cycles).

characteristic pattern. It was evident that the differences in Fig. 2 (different types) were much greater than those shown in Fig. 5 (different brands), these results were consistent with the facts that the compositions of vinegar of the different brand were more similar than those of different types. Similarly, normalized CTL signals for these 8 brands samples were subjected to LDA, whose size of the training matrix was 15 wavelengths  8 vinegars  5 replicates. The 3D discriminant score plot with the first 3 canonical factors (containing 77.3%, 13.3%, and 4.0% variation, respectively) is shown in Fig. 6a. All cases were correctly assigned to their respective groups with 100% accuracy. The results of cross-validation showed 90% classification success. The CTL response profiles under the detection wavelength of 425 nm of these 8 samples are shown in Fig. 7. All of them showed characteristic response profiles. The curves of brands 1, 2, and 6 exhibited an obvious peak at 320 s, but brands 3–5, 7, and 8 all showed peaks at 350 s. Similarly, 16 eigenvalues (20, 40, 60, 90, 120, 150, 180, 200, 230, 250, 270, 290, 300, 320, 350, and 400 s) were extracted and subjected to LDA. The CTL response values were transformed into canonical scores. These were visualized as a 3D plot (Fig. 6b), and the first three canonical factors contained 41.1%, 20.0%, and 18.3% of the variation, occupying 79.4% of total variation. The classification accuracy was 95.8%, but the results of cross-validation only showed 12.5% success. To sum up, this single sensor showed excellent performance, and the method based on the CTL response profile was slightly less effective. The single sensor based on spectral arrays demonstrated a greater classification accuracy and permitted facile discrimination among highly similar complex mixtures, such as these 8 brands of vinegar. 3.2.3. Discrimination of adulterated samples The objective of this part of the project was to determine the feasibility of detecting adulterated vinegar using single sensors. The quality and authenticity of vinegars are of particular

importance because of the large quantities consumed in people’s daily life. However, cheap synthetic vinegar is widely used to adulterate authentic fermented products. The adding of cheap synthetic acetic acid to vinegar is the main form of adulteration in conventional industrial manufacturing of vinegar [40]. Considering this, 25%, 50%, and 75% synthetic vinegar was added to baoning vinegar and mature vinegar for testing. An additional 4% acetic acid solution was measured as pure substrate because any product called “vinegar” should contain at least 4% acidity according to the United States Food and Drug Administration (FDA) [41]. As illustrated in Fig. 8, all samples resulted in a variety of CTL spectral patterns. The CTL pattern of 4% acetic acid was the most unusual. Its optical signals were observed at 400, 412, 425, 440, 460, 475, 490, 505, and 520 nm, but no obvious signals were harvested at 535, 555, 575, 590, 605, or 620 nm. The spectral patterns of other samples in Fig. 8 showed CTL signals with unique characteristics at all of these 15 wavelengths. CTL signals at a longwavelength increased with the ratios of synthetic vinegar added to the authentic samples, finally resulting in the similarities of their CTL patterns to those of the pure synthetic vinegar. It was apparent that we could obtain the different spectral arrays among these adulterated and authentic vinegars. The spectral arrays of vinegars were attributable not only to acetic acid, but also to other coexisting components. The multiple peak phenomenon was due to the combined effects of the mixtures in the CRC. This conclusion was consistent with those of previous works [27,28]. This fingerprints of the vinegars were used for the identification of adulterated and authentic samples when the normalized CTL intensities were analyzed using LDA. The CTL spectral patterns were converted into canonical scores which were visualized as a well-clustered 3D plot (Fig. 9) and the first three canonical factors contained 87.5%, 10.5%, and 0.9% of the variation. The authentic samples (marked by dots) compacted together, and the adulterated ones (marked by rhombus) and the 4% acetic acid (marked by

70

[(Fig._6)TD$IG]

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

Fig. 6. Canonical score plot of the first three factors analyzed by LDA for the 8 brands of vinegar. (a) Spectral arrays of 15 wavelengths. (b) CTL response profiles of 425 nm.

triangles) became distributed in the up side and the lower left corner of the authentic cluster, respectively. This single sensor could discriminate adulterated samples from the authentic ones, revealing its potential to determine adulteration in vinegar.

essentially not consumed in any of these tests, and the analytes were wiped out during a heating process after detection.

3.3. Reversibility and stability of the single sensor

The CTL response of baoning vinegar and mature vinegar to different nanomaterials (nano-MgO, nano-ZrO2, and nano-Y2O3) was tested. Results are shown in Fig. 10. Nanomaterials possessed various catalytic activities and produced diverse light emissions during the catalytic oxidations of vinegar. Different CTL patterns were produced by different samples when catalyzed by the same catalyst, which may be attributable to the characteristics of the different vinegars. For example, when catalyzed by Y2O3, the normalized CTL intensity at 425 nm of baoning vinegar was stronger than that of mature vinegar. However, the CTL patterns of the same analyte on different catalysts were also different, such as the strongest CTL intensity of baoning vinegar was at 535 and

To determine the reversibility and stability of this method, the CTL patterns of baoning vinegar and mature vinegar were individually tested 18 times for three days at a sampling interval about 1.5 h. The normalized CTL intensities of 425, 520, and 590 nm in the spectral array of these two samples are shown in Fig. S3a. In addition, the CTL responses of apple vinegar (brand 7) were tested before and after the entire experiments (about 900 working hours, Fig. S3b). It was evident that there was no obvious difference in CTL intensities, indicating that the method was reversible and stable. This may be because the nanomaterials were solid catalysts and

3.4. Choice of catalyst

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

[(Fig._7)TD$IG]

71

Fig. 7. CTL time-course profiles of 8 brands vinegars on nano-MgO surfaces in the closed reaction cell (working temperature: 250  C; detection wavelength: 425 nm; injection volume: 10 mL).

490 nm when catalyzed by MgO and ZrO2, respectively. To sum up, this system was not limited to only one kind of nanomaterials as sensing element. A wide range of choices were suitable. When there were no visible differences among the CTL patterns of the

analytes on any one catalyst, the CTL patterns of two or more catalysts were combined. MgO is an alkaline-earth compound which was chosen as the sensing element in this study because of its good thermal stability, lack of toxicity, and richness in catalysts.

[(Fig._8)TD$IG]

Fig. 8. Spectral arrays of 15 wavelengths for authentic and synthetic samples on nano-MgO surfaces in the closed reaction cell (working temperature: 250  C; injection volume: 10 mL; rotation speed: 5 s c 1; rotation number: 100 cycles; B: baoning vinegar; M: mature vinegar; S: synthesized vinegar).

72

[(Fig._9)TD$IG]

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

Fig. 9. Canonical score plot of the first three factors in the spectral arrays analyzed by LDA for the authentic and synthetic samples. (B: baoning vinegar; M: mature vinegar; S: synthesized vinegar).

[(Fig._10)TD$IG]

Fig. 10. Spectral arrays of 15 wavelengths for baoning vinegar and mature vinegar on different nano-catalytic materials surfaces (working temperature: 250  C; injection volume: 10 mL; rotation speed: 5 s c 1; rotation number: 100 cycles).

4. Conclusion In conclusion, a single sensor was developed in CRC mode. It can fingerprint vinegars using their characteristic spectral arrays. CTL signals were scanned and classified into 15 groups according to the distribution at fifteen wavelengths. After processing the data matrix using LDA, the discrimination of 9 types and 8 brands of vinegars and a series of artificial samples were performed using this simple sensor. Owing to the high dimensionality of the sensor data, this single sensor showed the ability to discriminate among

highly similar complex mixtures. The slight drift of the signals could be compensated by normalizing the CTL intensities. This method involves simple sensing elements and instrumentation. It has a reversible response and long-term stability. This single sensor shows great promise in real applications. Acknowledgements The authors would like to thank the National Natural Science Foundation of China (No. 21375030 and 21075024), Science and

J. Zeng et al. / Analytica Chimica Acta 864 (2015) 64–73

Technology Project Foundation of Guangzhou Education Bureau (2012A009) for financial support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.aca.2015.01.035. References [1] F. Röck, N. Barsan, U. Weimar, Electronic nose: current status and future trends, Chem. Rev. 108 (2008) 705–725. [2] M. Su, S. Li, V.P. Dravid, Miniaturized chemical multiplexed sensor array, J. Am. Chem. Soc. 125 (2003) 9930–9931. [3] H.W. Lin, K.S. Suslick, A colorimetric sensor array for detection of triacetone triperoxide vapor, J. Am. Chem. Soc. 132 (2010) 15519–15521. [4] S. Deshmukh, K. Kamde, A. Jana, S. Korde, R. Bandyopadhyay, R. Sankar, N. Bhattacharyya, R.A. Pandey, Calibration transfer between electronic nose systems for rapid in situ measurement of pulp and paper industry emissions, Anal. Chim. Acta 841 (2014) 58–67. [5] M. Peris, L. Escuder-Gilabert, On-line monitoring of food fermentation processes using electronic noses and electronic tongues: a review, Anal. Chim. Acta 804 (2013) 29–36. [6] Q. Zhang, S. Zhang, C. Xie, C. Fan, Z. Bai, ‘Sensory analysis’ of Chinese vinegars using an electronic nose, Sens. Actuators B Chem. 128 (2008) 586–593. [7] Y. Yin, H. Yu, B. Chu, Y. Xiao, A sensor array optimization method of electronic nose based on elimination transform of Wilks statistic for discrimination of three kinds of vinegars, J. Food Eng. 127 (2014) 43–48. [8] M.D.L.N. López De Lerma, A. Bellincontro, T. García-Martínez, F. Mencarelli, J.J. Moreno, Feasibility of an electronic nose to differentiate commercial Spanish wines elaborated from the same grape variety, Food Res. Int. 51 (2013) 790–796. [9] M. Liu, X. Han, K. Tu, L. Pan, J. Tu, L. Tang, P. Liu, G. Zhan, Q. Zhong, Z. Xiong, Application of electronic nose in Chinese spirits quality control and flavour assessment, Food Control 26 (2012) 564–570. [10] M. Pardo, G. Niederjaufner, G. Benussi, E. Comini, G. Faglia, G. Sberveglieri, M. Holmberg, I. Lundstrom, Data preprocessing enhances the classification of different brands of Espresso coffee with an electronic nose, Sens. Actuators B Chem. 69 (2000) 397–403. [11] Q. Chen, J. Zhao, Z. Chen, H. Lin, D. Zhao, Discrimination of green tea quality using the electronic nose technique and the human panel test, comparison of linear and nonlinear classification tools, Sens. Actuators B Chem. 159 (2011) 294–300. [12] D.N. Corrado, M. Antonella, P. Roberto, D.A. Arnaldo, Artificial olfaction systems principles and applications to food analysis, Biotechnol. Agron. Soc. Environ. 5 (2001) 159–165. [13] F.D. Francesco, B. Lazzerini, F. Marcelloni, G. Pioggia, An electronic nose for odour annoyance assessment, Atmos. Environ. 35 (2001) 1225–1234. [14] A. Pavlou, A.P.F. Turner, N. Magan, Recognition of anaerobic bacterial isolates in vitro using electronic nose technology, Lett. Appl. Microbiol. 35 (2002) 366–369. [15] T. Okabayashi, N. Matsuo, I. Yamamoto, K. Utsunomiya, N. Yamashita, M. Nakagawa, Temperature-programmed sensing for gas identification using the cataluminescence-based sensors, Sens. Actuators B Chem. 108 (2005) 515–520. [16] M. Liu, Z. Lin, J. Lin, A review on applications of chemiluminescence detection in food analysis, Anal. Chim. Acta 670 (2010) 1–10. [17] M.S. Attia, E. Bakir, A.A. Abdel-aziz, M.S.A. Abdel-mottaleb, Determination of melamine in different milk batches using a novel chemosensor based on the luminescence quenching of Ru(II) carbonyl complex, Talanta 84 (2011) 27–33. [18] B. Li, J. Liu, G. Shi, J. Liu, A research on detection and identification of volatile organic compounds utilizing cataluminescence-based sensor array, Sens. Actuators B Chem. 177 (2013) 1167–1172. [19] J. Han, F. Han, J. Ouyang, L. He, Y. Zhang, N. Na, Low temperature CO sensor based on cataluminescence from plasma-assisted catalytic oxidation on Ag doped alkaline-earth nanomaterials, Nanoscale 6 (2014) 3069–3072.

73

[20] H. Kong, H. Wang, S. Zhang, X. Zhang, A thermochemiluminescence array for recognition of protein subtypes and their denatured shapes, Analyst 136 (2011) 3643–3648. [21] J. Han, F. Han, J. Ouyang, Q. Li, N. Na, Venturi-electrosonic spray ionization cataluminescence sensor array for saccharides detection, Anal. Chem. 85 (2013) 7738–7744. [22] R. Zhang, X. Cao, Y. Liu, X. Chang, Development of a simple cataluminescence sensor system for detecting and discriminating volatile organic compounds at different concentrations, Anal. Chem. 85 (2013) 3802–3806. [23] R. Zhang, Y. Hu, G. Li, Development of a cyclic system for chemiluminescence detection, Anal. Chem. 86 (2014) 6080–6087. [24] L. Zhang, W. Rong, Y. Chen, C. Lu, L. Zhao, A novel acetone sensor utilizing cataluminescence on layered double oxide, Sens. Actuators B Chem. 205 (2014) 82–87. [25] Z. Li, W. Xi, C. Lu, Hydrotalcite-assisted cataluminescence: a new approach for sensing mesityl oxide in aldol condensation of acetone, Sens. Actuators B Chem. 207 (2015) 498–503. [26] L. Zhang, Y. Chen, N. He, C. Lu, Acetone cataluminescence as an indicator for evaluation of heterogeneous base catalysts in biodiesel production, Anal. Chem. 86 (2014) 870–875. [27] R. Zhang, X. Cao, Y. Liu, X. Chang, A new method for identifying compounds by luminescent response profiles on a cataluminescence based sensor, Anal. Chem. 83 (2011) 8975–8983. [28] J. Zeng, X. Cao, Y. Liu, J. Chen, K. Ren, A single cataluminescence sensor for wine identification by luminescent response profiles, Anal. Methods 6 (2014) 2633–2641. [29] G. Papotti, D. Bertelli, R. Graziosi, A. Maietti, P. Tedeschi, A. Marchetti, M. Plessi, Traditional balsamic vinegar and balsamic vinegar of Modena analyzed by nuclear magnetic resonance spectroscopy coupled with multivariate data analysis, LWT-Food Sci. Technol. 60 (2015) 1017–1024. [30] T.M. Guerreiro, D.N.D. Oliveira, M.S. Ferreira, R.R. Catharino, High-throughput analysis by SP-LDI-MS for fast identification of adulterations in commercial balsamic vinegars, Anal. Chim. Acta 838 (2014) 86–92. [31] M. Perini, M. Paolini, M. Simoni, L. Bontempo, U. Vrhovsek, M. Sacco, F. Thomas, E. Jamin, A. Hermann, F. Camin, Stable isotope ratio analysis for verifying the authenticity of balsamic and wine vinegar, J. Agric. Food Chem. 62 (2014) 8197–8203. [32] S. Manzini, C. Durante, C. Baschieri, M. Cocchi, S. Sighinolfi, S. Totaro, A. Marchetti, Optimization of a dynamic headspace–thermal desorption–gas chromatography/mass spectrometry procedure for the determination of furfurals in vinegars, Talanta 85 (2011) 863–869. [33] A. Marrufo-Curtido, M.J. Cejudo-Bastante, E. Durán-Guerrero, R. Castro-Mejías, R. Natera-Marín, F. Chinnici, C. García-Barroso, Characterization and differentiation of high quality vinegars by stir bar sorptive extraction coupled to gas chromatography–mass spectrometry (SBSE–GC–MS), LWTFood Sci. Technol. 47 (2012) 332–341. [34] M.M.S. Junior, L.O.B. Silva, D.J. Leão, S.L.C. Ferreira, Analytical strategies for determination of cadmium in Brazilian vinegar samples using ET AAS, Food Chem. 160 (2014) 209–213. [35] F. Thomas, E. Jamin, 2H NMR and 13C-IRMS analyses of acetic acid from vinegar, 18 O-IRMS analysis of water in vinegar: international collaborative study report, Anal. Chim. Acta 649 (2009) 98–105. [36] R.S. Uysal, E.A. Soykut, I.H. Boyaci, A. Topcu, Monitoring multiple components in vinegar fermentation using Raman spectroscopy, Food Chem. 141 (2013) 4333–4343. [37] F. Liu, Y. He, L. Wang, Determination of effective wavelengths for discrimination of fruit vinegars using near infrared spectroscopy and multivariate analysis, Anal. Chim. Acta 615 (2008) 10–17. [38] Q. Qiu, Discrimination and detection of inferior vinegar, Food Res. Dev. 22 (2001) 43–44 (in Chinese). [39] H. Kong, D. Liu, S. Zhang, X. Zhang, Protein sensing and cell discrimination using a sensor array based on nanomaterial-assisted chemiluminescence, Anal. Chem. 83 (2011) 1867–1870. [40] C. Hsieh, P. Li, J. Cheng, J. Ma, Using SNIF-NMR method to identify the adulteration of molasses spirit vinegar by synthetic acetic acid in rice vinegar, Ind. Crop Prod. 50 (2013) 904–908. [41] J. Shi, X. Zou, X. Huang, J. Zhao, Y. Li, L. Hao, J. Zhang, Rapid detecting total acid content and classifying different types of vinegar based on near infrared spectroscopy and least-squares support vector machine, Food Chem. 138 (2013) 192–199.

A single cataluminescence sensor based on spectral array and its use in the identification of vinegars.

The discrimination of complex mixtures, especially those with very similar compositions, remains a challenging part of chemical analysis. In this pape...
4MB Sizes 1 Downloads 9 Views