- Email: [email protected]
A novel colorimetric sensor array based on boron-dipyrromethene dyes for monitoring the storage time of rice
Food Chemistry 268 (2018) 300–306
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
A novel colorimetric sensor array based on boron-dipyrromethene dyes for monitoring the storage time of rice Hao Lina, Zhong-xiu Mana, Wen-cui Kanga, Bin-bin Guana, Quan-sheng Chena, Zhao-li Xueb, a b
T
⁎
School of Food and Biological Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, PR China School of Chemistry and Chemical Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, PR China
A R T I C LE I N FO
A B S T R A C T
Chemical compounds studied in this article: 18-Crown-6 (PubChem CID: 28557) 5,10,15,20-Tetraphenyl-21H,23H-porphine (TPP, PubChem CID: 70186) 5,10,15,20-Tetraphenyl-21H,23H-porphine zinc (TPPZn, PubChem CID: 3580039) 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(II) (TPPMn, PubChem CID: 14757041) 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese (III) Chloride (TPPMnCl, PubChem CID: 91864721) 5,10,15,20-Tetrakis (pentafluorophenyl) porphyrin iron (III) Chloride (FTPPFeCl, PubChem CID: 71431176) 5,10,15,20-Tetraphenyl-21H,23H-porphine iron (III) Chloride (TPPFeCl, PubChem CID: 67043998) 8-Phenyl-4,4-difluoro-BODIPY (HBDP, PubChem CID: 46868721) 8-(4-Bromophenyl)-4,4-difluoro-BODIPY (NO2BDP, PubChem CID: 102150779) 8-(4-Nitrophenyl)-4,4-difluoro-BODIPY (BrBDP, PubChem CID: 23631055)
A novel colorimetric sensor array based on boron-dipyrromethene (BODIPY) dyes was developed to monitor the volatile organic compounds (VOCs) of rice at different storage times. The VOCs of rice at different storage times were analyzed through GC–MS combined with multivariate analysis, and the compound 18-crown-6 was found significantly changed during rice aging process. Aimed at 18-crown-6 with particular macrocyclic structure, a series of BODIPYs were targeted synthesized for the selection of sensitive chemically responsive dyes. Four dyes were chosen to construct colorimetric sensor array based on sensitivity to VOCs of aged rice samples. Data acquired from the interactions of dyes and rice VOCs were subjected to the principal components analysis (PCA) and linear discriminant analysis (LDA). The optimal performance obtained by the LDA model was 98.75% in prediction set. Application of BODIPYs in this work has improved the sensitivity and expanded the choices of colorimetric dyes for the specific detection.
Keywords: Rice aging Colorimetric sensor array BODIPYs 18-Crown-6
1. Introduction Rice (Oryza sativa L.) as a staple food feeds about two-thirds of the world’s population, especially in Asia (Choi et al., 2015). Rice plays a crucial role as a primary dietary source of carbohydrates to assure the basic energy and nutrient supply (Tsuzuki et al., 2014). It’s no doubt that the quality of rice greatly affects human health. Vast demand for rice and its apparent seasonality make most of the countries store them to guarantee supplies during its lean periods (Park, Kim, Park, & Kim,
⁎
2012). In the storage process, the chemical and physical properties of rice change which are termed as aging (Thanathornvarakul, Anuntagool, & Tananuwong, 2016; Zhou, Robards, Helliwell, & Blanchard, 2002). The aging of rice was embodied by following three aspects: the decrease in its nutritive value and culinary quality, and the deterioration in its sensory quality. Flavor is a major rice aging criterion which not only directly reflects the sensor quality, but also relates to the changes in nutrient content and culinary quality (Griglione et al., 2015).
Corresponding author. E-mail address: [email protected] (Z.-l. Xue).
https://doi.org/10.1016/j.foodchem.2018.06.097 Received 23 June 2017; Received in revised form 30 November 2017; Accepted 19 June 2018 Available online 20 June 2018 0308-8146/ © 2018 Elsevier Ltd. All rights reserved.
Food Chemistry 268 (2018) 300–306
H. Lin et al.
2.2. Method
Sensory and instrumental analyses have long been used to measure the food flavor (Sung, Kim, Kim, & Kim, 2014). Sensory analyses using human nose as a smell assessment instrument provide unique and direct information on food flavor. However, sensor analyses inevitably are affected by subjective factors, such as emotion, physical condition and environment (Radi, Litananda, Rivai, & Purnomo, 2016). Alternatively, the gas chromatography-mass spectrometry (GC–MS) as the most common instrumental analyses method for flavor determination has been widely applied. Although it has been used for qualitative and quantitative analysis of volatile organic compounds, some typical drawbacks appears to limit its application such as: it requires sample preparation, time-consuming, expensive and requires experts for operation. laborious and time-consuming, high costs of implementation and need of skilled personnel (Feng et al., 2011; Sanaeifar, Mohtasebi, Ghasemi-Varnamkhasti, & Ahmadi, 2016). A new branch of colorimetric sensor array technology proposed by Suslick has been developed and has gained importance since 2000 (Rakow & Suslick, 2000). It is a novel electronic nose system which consisted of a colorimetric sensor array made of chemically responsive dyes with partial specificity and a pattern-recognition system capable of recognizing simple or complex odors (Kim, Li, Lim, Kang, & Park, 2016; Musto & Suslick, 2010). The colorimetric sensor array is composed of chemically responsive dyes that directly influence the detections of volatiles compounds. Nowadays chemically responsive dyes mainly include porphyrins and pH indicators that have been reported to perform intense coloration effect (Huang, Gu, Yao, Teye, & Wen, 2014; Huang, Zou, et al., 2014). Fabricated sensor arrays have been applied to the detection of organic small-molecular compounds such as alcohol (Suslick, Rakow, & Sen, 2004), biogenic amines (Xiaowei et al., 2015), and TVB-N (Huang, Gu, et al., 2014; Huang, Zou, et al., 2014). Therefore, colorimetric sensor array technology has great potential for analyzing the volatile organic compounds (VOCs) of rice with different storage times. A variety of VOCs would be obtained during the storage process of rice. It is reported that some specific organic compounds such as aldehydes and heterocyclic compounds remarkably changed with prolonged storage time. In this case, the sensitivity of conventional chemically responsive dyes should been investigated in these specific volatile compounds. In fact, it is found that colorimetric sensor array fabricated by conventional dyes (TPP porphyrins and pH indicators) find it difficult to discriminate aging of rice at the early stage (Guan, Zhao, Jin, & Lin, 2016). Given this, improving performance of dyes is of utmost importance. Presumably, sensitive chemically responsive dyes synthesized to target specific samples could change this situation. Besides, application of new chemically responsive dyes will increase the choices available and may provide dyes that will ensure the accurate detection of the VOCs in aging rice. In this study, a novel colorimetric sensor array was developed to monitor VOCs in the aging process of rice. The specific organic compounds which were identified during rice aging process were analyzed through GC–MS. Different kinds of chemically responsive dyes were synthesized comparatively to characterize the VOCs of aging rice. The process of VOCs exposure to dyes was simulated, and some dyes with simpler structures were specially synthesized to detect the specific VOCs during rice aging. Furthermore, multivariate analysis was employed to discriminate storage time of rice samples.
2.2.1. The synthesis of chemically responsive dyes Five kinds of general porphyrins were synthesized to determine the characteristic gas of 18-crown-6 and VOCs of rice with different storage times. In view of the special macrocyclic structure of 18-crown-6, eight BODIPYs targeted synthesized with simple chemical structure were also used in this study. Metalloporphyrins: 0.25 mmol of TPP and 1.25 mmol of different metal salts were dissolved in 50 mL of N, N-Dimethylformamide (DMF) solution, respectively, and refluxed for 24 h under N2. After solvent evaporation, the residue was dissolved in CHCl3 and rinsed with water for several times. The solution was dried with anhydrous Na2SO4. The solvent was re-evaporated and recrystallized using CHCl3 and CH3OH to get the pure target metalloporphyrin dyes. BODIPYs: Dyes were synthesized according to the classic Lindsey methodology (Loudet & Burgess, 2007; Wu et al., 2012). All reagent used in the synthesis were purchased from Sigma Chemical (USA). Porphyrins and BODIPYs dyes used are listed as follows: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
5,10,15,20-Tetraphenyl-21H,23H-porphine (TPP); 5,10,15,20-Tetraphenyl-21H,23H-porphine zinc (TPPZn); 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(II) (TPPMn); 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(III) Chloride (TPPMnCl); 5,10,15,20-Tetrakis(pentafluorophenyl)porphyrin iron (III) Chloride (FTPPFeCl); 5,10,15,20-tetraphenyl-21H,23H-porphine iron (III) Chloride (TPPFeCl); 8-phenyl-4,4-difluoro-BODIPY (HBDP); 8-(4-bromophenyl)-4,4-difluoro-BODIPY (NO2BDP); 8-(4-nitrophenyl)-4,4-difluoro-BODIPY (BrBDP); 8-(4-nitrophenyl)-6-bromo-4,4-difluoro-BODIPY (NO2BrBDP); 8-(4-nitrophenyl)-6,6-dibromo-4,4-difluoro-BODIPY (NO2Br2BDP); 8-(4-methoxylphenyl)-4,4-difluoro-BODIPY (OCH3BDP); 8-(6-methoxyl-2-naphthyl)-4,4-difluoro-BODIPY (NaiOCH3BDP); 8-(4-carbazolylphenyl)-4,4-difluoro-BODIPY (pCarBDP);
Besides, bromcresol purple (BCP) and dimethyl yellow (DMY) obtained from Sinopharm Chemical Reagent Co., Ltd were also used to characterize the VOCs of rice. 2.2.2. Colorimetric sensor array system The colorimetric sensor array system is illustrated in Fig. 1. It consists of gas-collecting chamber, vacuum pump, reaction and digital image acquisition system and computer with specific software. In the reaction and digital image acquisition system, a tri-CCD was employed to obtain images of colorimetric sensor array placed inside the reaction chamber whiles the diffuse reflection integrating sphere source provided continuous light with no flashes. All image processing of colorimetric sensor array were executed by the computer with specific software. Before the analysis, two valves were open and vacuum pump was turned on for 5 min to remove the remaining or other unrelated gases. The method of fabricating a colorimetric sensor array has been described by (Guan, Zhao, Lin, & Zou, 2013). The colorimetric sensor array was placed in the reaction chamber and the before image taken by tri-CCD. Then samples were transferred to a gas-collecting chamber and valves were closed to equilibrate for 5 min. Subsequently, two valves were opened and the vacuum pump was turned on again to extract the VOCs of analyte into the reaction chamber. Finally, digital image of the array after exposure to samples was captured by tri-CCD as after image. A series of image processing including filtering, threshold segmentation, morphological processing, center extraction and additional operations were performed for the next analysis. After that, color change files were produced through subtracting of the corresponding before image from the after image.
2. Material and method 2.1. Material Japonica rice (Wuyujing 3) was grown in a paddy of a local farm in Jiangsu Province, eastern part of China, from May to October 2014. After harvesting, the rice samples were obtained by hulling, separation, selection and stored in shaded plastic bags. Rice samples with different storage times (1, 3, 7 and 10 months) were prepared for the study. Standard 18-crown-6 was purchased from J&K scientific Ltd. 301
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 1. Diagram of colorimetric sensor array system.
spectrometer HP6890–5973 equipped with DB-WAX nonpolar column (30 m × 0.25 mm ID × 0.25 μm film thickness, Agilent Technologies, Santa Clara, CA, USA). The oven temperature was held at 35 °C for 5 min following injection, followed by a ramp of 5 °C·min-1 till 60 °C, then raised to 100 °C at 3 °C·min-1, and finally to 220 °C at 10 °C·min-1 with15 min hold; thus culminating to a total run time of 50.33 min. Instrument parameters were set as follows: inlet temperature, 230 °C; injection mode, splitless; carrier gas, helium and flow rate, 1 mL·min-1. The mass spectrometer was operated under electron impact ionization (70 eV) with a range of 33–500 amu. VOCs were identified by comparison of the mass spectrum with the NIST08 mass spectra library.
2.2.3. Selection of sensitive chemically responsive dyes To develop a colorimetric sensor array for identifying rice storage time with high specificity and sensitivity, it is necessary to select chemically responsive dyes. The selection was conducted in two steps as follows and all of these were realized by the colorimetric sensor array system. (1) Chemically responsive dyes sensitive to 18-crown-6 which is the representative gas of rice aging detected by GC–MS were screened. Fourteen types of synthesized dyes contain six porphyrins and eight BODIPYs were used, the volume of 18-crown-6 (density of 0.02 g·L−1) was 8 mL and the temperature and time were 25 °C and15 min, respectively. (2) The sensitive dyes to 18-crown-6 were also used for real application (rice sample) to test their specificity. In view of the light odor of rice, the temperature was raised to 65 °C in water bath heating. The size of gas-collecting chamber was set to 60 cm × 30 cm, the time remained unchanged and 8.0 g of rice samples was used. Chemically responsive dyes capable of characterizing comprehensively the flavor of rice at different storage times were selected. The ability in characterizing rice flavor by two pH indicators (BCP and DMY) was also determined.
2.2.6. Data analysis and processing The data analyses were performed by the SPSS statistical software using one-way ANOVA followed by Duncan’s multiple range tests and the differences were considered significant at P < 0.05. All algorithms were implemented in Matlab R2010a (Mathworks, USA) under Windows XP. Density functional theory (DFT) using the B3LYP functional of the Gaussian 09 program package with SDD basis sets was employed to investigate the geometry structure changes of chemically responsive dyes after binding with 18-crown-6. 3. Results and discussion
2.2.4. Application of the colorimetric sensor array in rice storage time detection Based on the selection results, TPPClMn, NO2BDP, OCH3BDP and BCP were screened and chosen to fabricate the colorimetric sensor array. There are four kinds of rice at different storage times, with 30 samples of each kind used (1-month age, 3-month age, 6-month age, and 1-year age), all making up a total of 120 samples.
3.1. GC - MS detection of rice with different storage times During storage and transportation process under variable conditions, the properties of rice showed a complex forms of changes in VOCs with aging. The most obvious change was degradations of lipids in rice. Analysis has proven that lipid oxidation products such as aldehydes and acids, generate a negative influence on flavor and affects rice acceptability (Aibara et al., 2014). Therefore, VOCs of rice associated with the four different storage times were extracted using a SPME and analyzed by GC–MS. Every kind of rice sample was analyzed for three times. Six kinds of volatile organic compounds were found present in more than three storage times of rice are indicated in Table 1. Three aldehydes, two ethers, one ketone and one heterocyclic were identified in more than three kinds of the sampled aged rice. Although aldehydes and ketone were found in every storage time, they did not show a significant difference (p < 0.05). However, the content of 18-crown-6 decreased from 30.14 to 7.42 and significant difference (p < 0.05) was observed in the process of aging.
2.2.5. SPME-GC–MS analysis conditions 8.0 g milled rice samples were held in an extraction flask (15 mL) to which 10 μL of 2.424 mg·L-1 4-Methyl-2-pentanol as internal standard was added. The extraction flask was then closed with polytetrafluoroethylene faced silicone septum cap and the SPME fiber of 75 μm Carboxen/ Polydimethylsiloxane (CAR/PDMS) (Sigma–Aldrich Australia) was exposed in the headspace of flask at 80 °C for 30 min. Once the isolation step was completed, the fiber was desorbed for 5 min in the GC injection port. The fiber was cleaned at 300 °C for 30 min based on manufacturer’s recommendations before use. GC–MS analysis was performed using a gas chromatograph-mass 302
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Table 1 Volatile organic compounds during different storage times of rice. Compounds
Relative content (%) ± standard deviation 1 month
Hexanal 2-Methylbutanal 3-Methylbutyraldehyde Acetone 18-Crown-6 Pyridine 4-Methyl-2-pentanol (internal standard)
3 months b
7 months a
15.47 ± 0.79 1.31 ± 0.27c 0.39 ± 0.36c 2.26 ± 0.56b 16.96 ± 0.89b 1.63 ± 0.82a 0.71 ± 0.04
7.20 ± 1.56 1.53 ± 0.32b 2.27 ± 1.25c 2.07 ± 0.81b 30.14 ± 0.83a 0.34 ± 0.06b 1.07 ± 0.03
10 months a
16.09 ± 2.36 2.10 ± 0.37a 5.16 ± 1.69a 8.98 ± 1.43a 12.30 ± 0.64c 0.68 ± 0.31
2.34 0.35 0.64 2.01 7.42 0.36 0.46
± ± ± ± ± ± ±
0.30c 0.02c 0.04c 1.01b 0.25d 0.05b 0.05
Variants possessing the same letter are not significantly (P < 0.05) different.
cavitation of 18-crown-6, a set of BODIYs were synthesized as dyes in which half of porphyrins in structure, contained two pyrrole rings. Surprisingly, and somewhat interestingly, all of BODIPYs performed higher values compared to porphyrins after exposure to 18-crown-6 (as seen in G and R components response value). Dyes of HBDP, BrBDP, NO2BDP, NO2Br2BDP, OCH3BDP and NaiOCH3BDP were more sensitive dyes to 18-crown-6 based on their response values. In order to explain the sensitivity of the BODIPYs to 18-crown-6, structures of dyes and binding forms between dyes and 18-crown-6 were analyzed. As observed in Fig. 3a, central metal ion (Zn) of TPPZn no longer was bonding sites as previously (Gu, Huang, Yao, Teye, & Wen, 2014; Zheng, Shan, Yu, & Wang, 2008), while the phenyl of TPPZn bonded with 18-crown-6. Porphyrins have macrocyclic structure locate at the center of the molecule, as well as 18-crown-6, which led to 18-crown-6 was not ligated to the central metal ion of porphyrins. The phenyl of porphyrins out of the plane is easier to enter the center of macrocycle of 18-crown-6. This bonding form confirmed the former assumption that macrocyclic structure of porphyrins appeared steric hindrance with 18crown-6’s. And the opinion maybe should be changed that odorant bonding with the central of porphyrins is intrinsic to the mechanism of action (Janzen, Ponder, Bailey, Ingison, & Suslick, 2006). The previous binding principle was adapted to small organic components, but things are quite different when the VOCs are macromolecular compounds. The compound 18-crown-6 provides a readily available archetype of such unique macrocyclic structure. BODIPYs are comprised of two pyrrole rings retained chromophore partly and own smaller structure, compared to porphyrins’. The smaller structure of BODIPYs allowed them to easily bond with 18-crown-6, as shown in Fig. 3b, corresponding to their response values. However, 2, 8-substituted BODIPY, like NO2BrBDP, exhibits lower sensitivity to 18-crown-6 compared to other BODIPYs, and the sensitivity of pCarBDP also obviously decreased. As shown in Fig. 3c and d, such differences are widely attributed to 2substituent preventing 18-crown-6 binding with 8-substituent group of BODIPY and the relatively larger structure of 8-substituent leading to the steric hindrance effect. To test the capacity of resisting disturbance in real rice samples, six dyes sensitive to 18-crown-6 were applied to detect the four kinds of rice stored for different durations. The color and brightness of points in Fig. 3e expressed color components (R, G and B) responding and sensitivity of dyes. Brighter point means color of dyes changed more after exposure, that is, dyes were more sensitive to VOCs. As displayed in Fig. 3e, all points of HBDP, BrBDP and NO2Br2BDP were dim, which demonstrated that the sensitivity of these dyes decreases dramatically in rice. It means that other VOCs in rice samples disguised the 18crown-6 and thus resulted in low response values from these dyes. The R component of NaiOCH3BDP changed after exposure to rice samples, but it showed similar response value among four kinds of rice. From the digital image of OCH3BDP, it was easy to discriminate the four kinds of rice. Its brightness gradually dimmed following the storage time, just same as the tendency of the concentration of 18-crown-6. Response values of NO2BDP in R, G and B components were similar which
3.2. The selection of chemically responsive dyes corresponding to VOCs of rice As the key material of colorimetric sensor array, chemically responsive dyes directly influence the accuracy of VOCs detection. In previous research, colorimetric sensor array based on universal porphyrins and pH indicators was unable to identify the storage time of rice correctly. In this work, some dyes with simpler structures were synthesized aiming at the special structure of 18-crown-6 to improve the sensitivity of the colorimetric sensor array. 3.2.1. The sensitive chemically responsive dyes to the 18-crown-6 The component 18-crown-6 was the characteristic VOC in the aging as revealed by the results of GC–MS and the minimum concentration was 0.15 g·L−1. For achieving rice storage monition, six porphyrins (numbered 1 to 6, as shown in Fig. 2) were chosen as dyes to detect 18crown-6 with concentration of 0.10 g·L−1. Six porphyrins exhibited similarly and none of them significantly higher in values comparatively with each other after exposure to 18-crown-6. The component 18crown-6 is one of macrocyclic polyethers which can selectively bind with various cations due to the cavitation of the macrocycle. Both porphyrins and 18-crown-6 are macromolecular, and their increasing steric hindrance may decrease their binding degree. Considering the
Fig. 2. Response values of different dyes exposure to volatile 18-crown-6 and the structure of sensitive dyes. 303
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 3. Difference in binding forms between dyes ((a) TPPZn, (b) NO2BDP, (c) NO2BrBDP, (d) pCarBDP and 18-crown-6; (e) Response values of sensitive dyes in rice with different storage time.
demonstrated that some of acid gases changed in the aging process. Consequently, TPPClFe and BCP were screened as sensitive dyes to detect the whole flavor of rice in different storage times.
reflected in response digital image as white points. Here, response values in red component were 38.31 ± 2.03a, 33.87 ± 2.68b, 25.28 ± 1.76c and 20.25 ± 1.13d, respectively, for the four rice storage times with significant difference observed among each group. Based on variance analysis above, NO2BDP and OCH3BDP were considered as efficient dyes for monitoring the storage time of rice.
3.3. The results of a colorimetric sensor array application in rice storage time detection 3.3.1. Colorimetric sensor array responses for different storage times of rice Based on the above selection results, four kinds of sensitive dyes were employed to monitor the storage time of rice. The color differences of sensor arrays were visualizations by the difference maps that were superimposed on differences in the gray values of red, green and blue (RGB). Fig. 5a clearly shows the difference images of colorimetric sensor array after exposure to rice with different storage times. Different storage time rice had their unique different images which can are distinguishable with the naked eye. The results indicated that the developed colorimetric sensor array has the potential for monitoring the storage time of rice.
3.2.2. The chemically responsive dyes characterized VOCs of rice with different storage times Although characteristic gas provided the possibility of differentiating among different storage times of rice, other VOCs with natural aroma and stale flavor formed during aging cannot be underestimated. VOCs of rice during aging are complexed and dynamic. Therefore characterizing the whole information of VOCs facilitates the development of colorimetric sensor array with high specificity. There were ten dyes (excluded six dyes sensitive to 18-crown-6, added two pH indicators (BCP and DMY)) were exposure to rice samples with results showed in Fig. 4. TPP, pCarBDP and NO2BrBDP had lower response values and no significant difference among the four kinds of rice storage times. This demonstrated that they are not sensitive to the flavor of rice in the whole aging process. TPPZn, TPPMn and TPPClMn were observed to generate higher but same response values among the different aged rice. TPPClFe performed good discrimination in rice storage time while the color change files of FTPPClFe just discriminated the first two kinds of rice from others. The fluorine-based of FTPPClFe exerted a negative effect for its sensitivity in this study. Four kinds of rice were also classified through the color change files of BCP which
3.3.2. Principal component analysis (PCA) Principal component analysis (PCA) is widely used in dimensionality reduction, involving the extraction of relevant information from complexed data sets (Lu, Tan, & Wang, 2013). The relevant information was turn into PCs by the process of reconstructing the original data using an orthogonal transformation (Yu & Wang, 2007; Zhang, Tian, & Pei, 2014). When accumulation contribution rate of first several principal components reaches 85%, they can represent data-set of rice sample without loss in variables information. In this work, color files of rice with different storage times were analyzed by PCA. The result showed that the first three principal components had their cumulative reliability summing up to 94.21% (> 85%). So scores of the first three PCs were used to construct 3-dimensional (3D) PCA scatter plot to visualize the cluster trends of samples as shown in Fig. 5b. The data-set performed a good discrimination in most rice samples. Data points of the 1-month-storage-time and the 10-month-storage-time rice clustered slightly; others were separated from each other. Some volatile gases increased first and then subsequently decreased with the aging of rice. Hence they performed similar to each other in the color change files and cluster slightly in 3-dimensional (3D) PCA scatter plot. Furthermore, the sample distribution of the rice with 1 and 3-months-storagetimes were relatively discrete, whereas the distribution of samples with
Fig. 4. Response values of dyes after exposure to rice with different storage time. 304
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 5. (a) Difference images of colorimetric sensor array expose to different storage time of rice; (b) The 3-D scatter plot constructed by the first three PCs for all samples; (c) Discrimination rates of LDA model according to different PCs.
selected in characterizing the flavor of rice in the aging process. NO2BDP, OCH3BDP, TPPClFe and BCP were eventually chosen to construct colorimetric sensor array for monitoring the rice storage time and good characterization was achieved by PCA and LDA analysis.
7 and 10-month-storage-times appeared more clustered. It might be that VOCs of rice changed continuously (aroma decreasing and stale flavor increasing) in the initial aging process. 3.3.3. Determination results of rice samples based on the colorimetric sensor array Linear Discriminant Analysis (LDA) as one of more effective tools was applied to discriminate rice storage time. Unlike PCA, LDA is guaranteed to provide maximum separation between classes and minimum within classes according to their class labels (Martinez & Kak, 2001). In this work, LDA was employed to discriminant rice storage time based on PCA scores from PCA analysis and labels of samples. Total number of 20 samples from each class were applied to training of calibration model, and the remaining 10 samples for testing of the calibration model. Fig. 5c shows the discrimination rates of LDA model according to different principal components (PCs). The LDA model performed good discrimination in rice storage when PCs = 4. The discrimination rates of calibration set and prediction set were respectively 98.75% and 97.50% with the ideal number of PCs, up 12.50% in prediction set from our previous study (Guan et al., 2016). There is only one sample of three months was incorrectly identified as seven month in the prediction set.
Acknowledgements This work has been financially supported by Foundation for the National Key Technology R&D Program of China (Grant No. 2015BAD19B05), China Postdoctoral Natural Science Foundation (2016M601746), Jiangsu province Postdoctoral Natural Science Foundation (1601135B). References Aibara, S., Ismail, I. A., Yamashita, H., Ohta, H., Sekiyama, F., & Morita, Y. (2014). Changes in rice bran lipids and free amino acids during storage. Agricultural and Biological Chemistry, 50(3), 665–673. Choi, S., Jun, H., Bang, J., Chung, S. H., Kim, Y., Kim, B. S., et al. (2015). Behaviour of Aspergillus flavus and Fusarium graminearum on rice as affected by degree of milling, temperature, and relative humidity during storage. Food Microbiology, 46, 307–313. Feng, T., Zhuang, H., Ye, R., Jin, Z., Xu, X., & Xie, Z. (2011). Analysis of volatile compounds of Mesona Blumes gum/rice extrudates via GC–MS and electronic nose. Sensors and Actuators B: Chemical, 160(1), 964–973. Griglione, A., Liberto, E., Cordero, C., Bressanello, D., Cagliero, C., Rubiolo, P., et al. (2015). High-quality Italian rice cultivars: Chemical indices of ageing and aroma quality. Food Chemistry, 172, 305–313. Gu, H., Huang, X., Yao, L., Teye, E., & Wen, Y. (2014). Effects of substitute group, axial ligand and volatile organic compounds on binding ability of colorimetric sensor array. Materials Technology, 29(4), 220–226. Guan, B., Zhao, J., Jin, H., & Lin, H. (2016). Determination of rice storage time with colorimetric sensor array. Food Analytical Methods, 10(4), 1054–1062. Guan, B., Zhao, J., Lin, H., & Zou, X. (2013). Characterization of volatile organic compounds of vinegars with novel electronic nose system combined with multivariate analysis. Food Analytical Methods, 7(5), 1073–1082. Huang, X., Gu, H., Yao, L., Teye, E., & Wen, Y. (2014). A study of the interactions between colorimetric sensor array and volatile organic compounds. Journal of Computational and Theoretical Nanoscience, 11(11), 2304–2309. Huang, X., Zou, X., Zhao, J., Shi, J., Zhang, X., Li, Z., et al. (2014). Sensing the quality parameters of Chinese traditional Yao-meat by using a colorimetric sensor combined with genetic algorithm partial least squares regression. Meat Science, 98(2), 203–210. Janzen, M. C., Ponder, J. B., Bailey, D. P., Ingison, C. K., & Suslick, K. S. (2006).
4. Conclusion A novel colorimetric sensor array was designed to discriminate rice samples with different storage times. A series of BODIPYs were targeted synthesized as chemically responsive dyes to detect the 18-crown-6 determined by GC–MS results of four kinds of rice with different storage times as a characteristic gas associated with aging of rice. It is the first attempt of selecting BODIPYs as dyes for the fabrication of colorimetric sensor array, for which remarkable success achieved vis-à-vis their sensitivities. Meanwhile, the successful application of BODIPYs as chemically responsive dyes was explored as a way to improve the sensitivity of the colorimetric sensor array through changing dyes structure according to different VOCs. Furthermore, ten dyes were also 305
Food Chemistry 268 (2018) 300–306
H. Lin et al.
system for assessing rice quality during storage at different temperatures. Journal of Stored Products Research, 59, 204–208. Suslick, K. S., Rakow, N. A., & Sen, A. (2004). Colorimetric sensor arrays for molecular recognition. Tetrahedron, 60(49), 11133–11138. Thanathornvarakul, N., Anuntagool, J., & Tananuwong, K. (2016). Aging of low and high amylose rice at elevated temperature: Mechanism and predictive modeling. Journal of Cereal Science, 70, 155–163. Tsuzuki, W., Suzuki, Y., Yamada, S., Kano, S., Ohnishi, H., Fujimoto, T., et al. (2014). Effect of oxygen absorber on accumulation of free fatty acids in brown rice and whole grain wheat during storage. LWT – Food Science and Technology, 58(1), 222–229. Wu, W., Zhao, J., Guo, H., Sun, J., Ji, S., & Wang, Z. (2012). Long-lived room-temperature near-IR phosphorescence of BODIPY in a visible-light-harvesting N^C^N PtII–acetylide complex with a directly metalated BODIPY chromophore. Chemistry – A European Journal, 18(7), 1961–1968. Xiaowei, H., Xiaobo, Z., Jiewen, Z., Jiyong, S., Zhihua, L., & Tingting, S. (2015). Monitoring the biogenic amines in Chinese traditional salted pork in jelly (Yao-meat) by colorimetric sensor array based on nine natural pigments. International Journal of Food Science & Technology, 50(1), 203–209. Yu, H., & Wang, J. (2007). Discrimination of LongJing green-tea grade by electronic nose. Sensors and Actuators B: Chemical, 122(1), 134–140. Zhang, L., Tian, F., & Pei, G. (2014). A novel sensor selection using pattern recognition in electronic nose. Measurement, 54, 31–39. Zheng, W., Shan, N., Yu, L., & Wang, X. (2008). UV–visible, fluorescence and EPR properties of porphyrins and metalloporphyrins. Dyes and Pigments, 77(1), 153–157. Zhou, Z., Robards, K., Helliwell, S., & Blanchard, C. (2002). Ageing of stored rice: changes in chemical and physical attributes. Journal of Cereal Science, 35(1), 65–78.
Colorimetric sensor arrays for volatile organic compounds. Analytical Chemistry, 78(11), 3591–3600. Kim, S. Y., Li, J., Lim, N. R., Kang, B. S., & Park, H. J. (2016). Prediction of warmed-over flavour development in cooked chicken by colorimetric sensor array. Food Chemistry, 211, 440–447. Loudet, A., & Burgess, K. (2007). BODIPY dyes and their derivatives: syntheses and spectroscopic properties. Chemical Reviews, 107(11), 4891–4932. Lu, J., Tan, Y. P., & Wang, G. (2013). Discriminative multimanifold analysis for face recognition from a single training sample per person. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(1), 39–51. Martinez, A. M., & Kak, A. C. (2001). PCA versus LDA. IEEE Transactions on Pattern Analysis and Machine Intelligence, 23(2), 228–233. Musto, C. J., & Suslick, K. S. (2010). Differential sensing of sugars by colorimetric arrays. Current Opinion in Chemical Biology, 14(6), 758–766. Park, C.-E., Kim, Y.-S., Park, K.-J., & Kim, B.-K. (2012). Changes in physicochemical characteristics of rice during storage at different temperatures. Journal of Stored Products Research, 48, 25–29. Radi, Ciptohadijoyo, S., Litananda, W. S., Rivai, M., & Purnomo, M. H. (2016). Electronic nose based on partition column integrated with gas sensor for fruit identification and classification. Computers and Electronics in Agriculture, 121, 429–435. Rakow, N. A., & Suslick, K. S. (2000). A colorimetric sensor array for odour visualization. Nature, 406(6797), 710–713. Sanaeifar, A., Mohtasebi, S. S., Ghasemi-Varnamkhasti, M., & Ahmadi, H. (2016). Application of MOS based electronic nose for the prediction of banana quality properties. Measurement, 82, 105–114. Sung, J., Kim, B.-K., Kim, B.-S., & Kim, Y. (2014). Mass spectrometry-based electric nose
306
Contents lists available at ScienceDirect
Food Chemistry journal homepage: www.elsevier.com/locate/foodchem
Analytical Methods
A novel colorimetric sensor array based on boron-dipyrromethene dyes for monitoring the storage time of rice Hao Lina, Zhong-xiu Mana, Wen-cui Kanga, Bin-bin Guana, Quan-sheng Chena, Zhao-li Xueb, a b
T
⁎
School of Food and Biological Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, PR China School of Chemistry and Chemical Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, PR China
A R T I C LE I N FO
A B S T R A C T
Chemical compounds studied in this article: 18-Crown-6 (PubChem CID: 28557) 5,10,15,20-Tetraphenyl-21H,23H-porphine (TPP, PubChem CID: 70186) 5,10,15,20-Tetraphenyl-21H,23H-porphine zinc (TPPZn, PubChem CID: 3580039) 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(II) (TPPMn, PubChem CID: 14757041) 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese (III) Chloride (TPPMnCl, PubChem CID: 91864721) 5,10,15,20-Tetrakis (pentafluorophenyl) porphyrin iron (III) Chloride (FTPPFeCl, PubChem CID: 71431176) 5,10,15,20-Tetraphenyl-21H,23H-porphine iron (III) Chloride (TPPFeCl, PubChem CID: 67043998) 8-Phenyl-4,4-difluoro-BODIPY (HBDP, PubChem CID: 46868721) 8-(4-Bromophenyl)-4,4-difluoro-BODIPY (NO2BDP, PubChem CID: 102150779) 8-(4-Nitrophenyl)-4,4-difluoro-BODIPY (BrBDP, PubChem CID: 23631055)
A novel colorimetric sensor array based on boron-dipyrromethene (BODIPY) dyes was developed to monitor the volatile organic compounds (VOCs) of rice at different storage times. The VOCs of rice at different storage times were analyzed through GC–MS combined with multivariate analysis, and the compound 18-crown-6 was found significantly changed during rice aging process. Aimed at 18-crown-6 with particular macrocyclic structure, a series of BODIPYs were targeted synthesized for the selection of sensitive chemically responsive dyes. Four dyes were chosen to construct colorimetric sensor array based on sensitivity to VOCs of aged rice samples. Data acquired from the interactions of dyes and rice VOCs were subjected to the principal components analysis (PCA) and linear discriminant analysis (LDA). The optimal performance obtained by the LDA model was 98.75% in prediction set. Application of BODIPYs in this work has improved the sensitivity and expanded the choices of colorimetric dyes for the specific detection.
Keywords: Rice aging Colorimetric sensor array BODIPYs 18-Crown-6
1. Introduction Rice (Oryza sativa L.) as a staple food feeds about two-thirds of the world’s population, especially in Asia (Choi et al., 2015). Rice plays a crucial role as a primary dietary source of carbohydrates to assure the basic energy and nutrient supply (Tsuzuki et al., 2014). It’s no doubt that the quality of rice greatly affects human health. Vast demand for rice and its apparent seasonality make most of the countries store them to guarantee supplies during its lean periods (Park, Kim, Park, & Kim,
⁎
2012). In the storage process, the chemical and physical properties of rice change which are termed as aging (Thanathornvarakul, Anuntagool, & Tananuwong, 2016; Zhou, Robards, Helliwell, & Blanchard, 2002). The aging of rice was embodied by following three aspects: the decrease in its nutritive value and culinary quality, and the deterioration in its sensory quality. Flavor is a major rice aging criterion which not only directly reflects the sensor quality, but also relates to the changes in nutrient content and culinary quality (Griglione et al., 2015).
Corresponding author. E-mail address: [email protected] (Z.-l. Xue).
https://doi.org/10.1016/j.foodchem.2018.06.097 Received 23 June 2017; Received in revised form 30 November 2017; Accepted 19 June 2018 Available online 20 June 2018 0308-8146/ © 2018 Elsevier Ltd. All rights reserved.
Food Chemistry 268 (2018) 300–306
H. Lin et al.
2.2. Method
Sensory and instrumental analyses have long been used to measure the food flavor (Sung, Kim, Kim, & Kim, 2014). Sensory analyses using human nose as a smell assessment instrument provide unique and direct information on food flavor. However, sensor analyses inevitably are affected by subjective factors, such as emotion, physical condition and environment (Radi, Litananda, Rivai, & Purnomo, 2016). Alternatively, the gas chromatography-mass spectrometry (GC–MS) as the most common instrumental analyses method for flavor determination has been widely applied. Although it has been used for qualitative and quantitative analysis of volatile organic compounds, some typical drawbacks appears to limit its application such as: it requires sample preparation, time-consuming, expensive and requires experts for operation. laborious and time-consuming, high costs of implementation and need of skilled personnel (Feng et al., 2011; Sanaeifar, Mohtasebi, Ghasemi-Varnamkhasti, & Ahmadi, 2016). A new branch of colorimetric sensor array technology proposed by Suslick has been developed and has gained importance since 2000 (Rakow & Suslick, 2000). It is a novel electronic nose system which consisted of a colorimetric sensor array made of chemically responsive dyes with partial specificity and a pattern-recognition system capable of recognizing simple or complex odors (Kim, Li, Lim, Kang, & Park, 2016; Musto & Suslick, 2010). The colorimetric sensor array is composed of chemically responsive dyes that directly influence the detections of volatiles compounds. Nowadays chemically responsive dyes mainly include porphyrins and pH indicators that have been reported to perform intense coloration effect (Huang, Gu, Yao, Teye, & Wen, 2014; Huang, Zou, et al., 2014). Fabricated sensor arrays have been applied to the detection of organic small-molecular compounds such as alcohol (Suslick, Rakow, & Sen, 2004), biogenic amines (Xiaowei et al., 2015), and TVB-N (Huang, Gu, et al., 2014; Huang, Zou, et al., 2014). Therefore, colorimetric sensor array technology has great potential for analyzing the volatile organic compounds (VOCs) of rice with different storage times. A variety of VOCs would be obtained during the storage process of rice. It is reported that some specific organic compounds such as aldehydes and heterocyclic compounds remarkably changed with prolonged storage time. In this case, the sensitivity of conventional chemically responsive dyes should been investigated in these specific volatile compounds. In fact, it is found that colorimetric sensor array fabricated by conventional dyes (TPP porphyrins and pH indicators) find it difficult to discriminate aging of rice at the early stage (Guan, Zhao, Jin, & Lin, 2016). Given this, improving performance of dyes is of utmost importance. Presumably, sensitive chemically responsive dyes synthesized to target specific samples could change this situation. Besides, application of new chemically responsive dyes will increase the choices available and may provide dyes that will ensure the accurate detection of the VOCs in aging rice. In this study, a novel colorimetric sensor array was developed to monitor VOCs in the aging process of rice. The specific organic compounds which were identified during rice aging process were analyzed through GC–MS. Different kinds of chemically responsive dyes were synthesized comparatively to characterize the VOCs of aging rice. The process of VOCs exposure to dyes was simulated, and some dyes with simpler structures were specially synthesized to detect the specific VOCs during rice aging. Furthermore, multivariate analysis was employed to discriminate storage time of rice samples.
2.2.1. The synthesis of chemically responsive dyes Five kinds of general porphyrins were synthesized to determine the characteristic gas of 18-crown-6 and VOCs of rice with different storage times. In view of the special macrocyclic structure of 18-crown-6, eight BODIPYs targeted synthesized with simple chemical structure were also used in this study. Metalloporphyrins: 0.25 mmol of TPP and 1.25 mmol of different metal salts were dissolved in 50 mL of N, N-Dimethylformamide (DMF) solution, respectively, and refluxed for 24 h under N2. After solvent evaporation, the residue was dissolved in CHCl3 and rinsed with water for several times. The solution was dried with anhydrous Na2SO4. The solvent was re-evaporated and recrystallized using CHCl3 and CH3OH to get the pure target metalloporphyrin dyes. BODIPYs: Dyes were synthesized according to the classic Lindsey methodology (Loudet & Burgess, 2007; Wu et al., 2012). All reagent used in the synthesis were purchased from Sigma Chemical (USA). Porphyrins and BODIPYs dyes used are listed as follows: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
5,10,15,20-Tetraphenyl-21H,23H-porphine (TPP); 5,10,15,20-Tetraphenyl-21H,23H-porphine zinc (TPPZn); 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(II) (TPPMn); 5,10,15,20-Tetraphenyl-21H,23H-porphine manganese(III) Chloride (TPPMnCl); 5,10,15,20-Tetrakis(pentafluorophenyl)porphyrin iron (III) Chloride (FTPPFeCl); 5,10,15,20-tetraphenyl-21H,23H-porphine iron (III) Chloride (TPPFeCl); 8-phenyl-4,4-difluoro-BODIPY (HBDP); 8-(4-bromophenyl)-4,4-difluoro-BODIPY (NO2BDP); 8-(4-nitrophenyl)-4,4-difluoro-BODIPY (BrBDP); 8-(4-nitrophenyl)-6-bromo-4,4-difluoro-BODIPY (NO2BrBDP); 8-(4-nitrophenyl)-6,6-dibromo-4,4-difluoro-BODIPY (NO2Br2BDP); 8-(4-methoxylphenyl)-4,4-difluoro-BODIPY (OCH3BDP); 8-(6-methoxyl-2-naphthyl)-4,4-difluoro-BODIPY (NaiOCH3BDP); 8-(4-carbazolylphenyl)-4,4-difluoro-BODIPY (pCarBDP);
Besides, bromcresol purple (BCP) and dimethyl yellow (DMY) obtained from Sinopharm Chemical Reagent Co., Ltd were also used to characterize the VOCs of rice. 2.2.2. Colorimetric sensor array system The colorimetric sensor array system is illustrated in Fig. 1. It consists of gas-collecting chamber, vacuum pump, reaction and digital image acquisition system and computer with specific software. In the reaction and digital image acquisition system, a tri-CCD was employed to obtain images of colorimetric sensor array placed inside the reaction chamber whiles the diffuse reflection integrating sphere source provided continuous light with no flashes. All image processing of colorimetric sensor array were executed by the computer with specific software. Before the analysis, two valves were open and vacuum pump was turned on for 5 min to remove the remaining or other unrelated gases. The method of fabricating a colorimetric sensor array has been described by (Guan, Zhao, Lin, & Zou, 2013). The colorimetric sensor array was placed in the reaction chamber and the before image taken by tri-CCD. Then samples were transferred to a gas-collecting chamber and valves were closed to equilibrate for 5 min. Subsequently, two valves were opened and the vacuum pump was turned on again to extract the VOCs of analyte into the reaction chamber. Finally, digital image of the array after exposure to samples was captured by tri-CCD as after image. A series of image processing including filtering, threshold segmentation, morphological processing, center extraction and additional operations were performed for the next analysis. After that, color change files were produced through subtracting of the corresponding before image from the after image.
2. Material and method 2.1. Material Japonica rice (Wuyujing 3) was grown in a paddy of a local farm in Jiangsu Province, eastern part of China, from May to October 2014. After harvesting, the rice samples were obtained by hulling, separation, selection and stored in shaded plastic bags. Rice samples with different storage times (1, 3, 7 and 10 months) were prepared for the study. Standard 18-crown-6 was purchased from J&K scientific Ltd. 301
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 1. Diagram of colorimetric sensor array system.
spectrometer HP6890–5973 equipped with DB-WAX nonpolar column (30 m × 0.25 mm ID × 0.25 μm film thickness, Agilent Technologies, Santa Clara, CA, USA). The oven temperature was held at 35 °C for 5 min following injection, followed by a ramp of 5 °C·min-1 till 60 °C, then raised to 100 °C at 3 °C·min-1, and finally to 220 °C at 10 °C·min-1 with15 min hold; thus culminating to a total run time of 50.33 min. Instrument parameters were set as follows: inlet temperature, 230 °C; injection mode, splitless; carrier gas, helium and flow rate, 1 mL·min-1. The mass spectrometer was operated under electron impact ionization (70 eV) with a range of 33–500 amu. VOCs were identified by comparison of the mass spectrum with the NIST08 mass spectra library.
2.2.3. Selection of sensitive chemically responsive dyes To develop a colorimetric sensor array for identifying rice storage time with high specificity and sensitivity, it is necessary to select chemically responsive dyes. The selection was conducted in two steps as follows and all of these were realized by the colorimetric sensor array system. (1) Chemically responsive dyes sensitive to 18-crown-6 which is the representative gas of rice aging detected by GC–MS were screened. Fourteen types of synthesized dyes contain six porphyrins and eight BODIPYs were used, the volume of 18-crown-6 (density of 0.02 g·L−1) was 8 mL and the temperature and time were 25 °C and15 min, respectively. (2) The sensitive dyes to 18-crown-6 were also used for real application (rice sample) to test their specificity. In view of the light odor of rice, the temperature was raised to 65 °C in water bath heating. The size of gas-collecting chamber was set to 60 cm × 30 cm, the time remained unchanged and 8.0 g of rice samples was used. Chemically responsive dyes capable of characterizing comprehensively the flavor of rice at different storage times were selected. The ability in characterizing rice flavor by two pH indicators (BCP and DMY) was also determined.
2.2.6. Data analysis and processing The data analyses were performed by the SPSS statistical software using one-way ANOVA followed by Duncan’s multiple range tests and the differences were considered significant at P < 0.05. All algorithms were implemented in Matlab R2010a (Mathworks, USA) under Windows XP. Density functional theory (DFT) using the B3LYP functional of the Gaussian 09 program package with SDD basis sets was employed to investigate the geometry structure changes of chemically responsive dyes after binding with 18-crown-6. 3. Results and discussion
2.2.4. Application of the colorimetric sensor array in rice storage time detection Based on the selection results, TPPClMn, NO2BDP, OCH3BDP and BCP were screened and chosen to fabricate the colorimetric sensor array. There are four kinds of rice at different storage times, with 30 samples of each kind used (1-month age, 3-month age, 6-month age, and 1-year age), all making up a total of 120 samples.
3.1. GC - MS detection of rice with different storage times During storage and transportation process under variable conditions, the properties of rice showed a complex forms of changes in VOCs with aging. The most obvious change was degradations of lipids in rice. Analysis has proven that lipid oxidation products such as aldehydes and acids, generate a negative influence on flavor and affects rice acceptability (Aibara et al., 2014). Therefore, VOCs of rice associated with the four different storage times were extracted using a SPME and analyzed by GC–MS. Every kind of rice sample was analyzed for three times. Six kinds of volatile organic compounds were found present in more than three storage times of rice are indicated in Table 1. Three aldehydes, two ethers, one ketone and one heterocyclic were identified in more than three kinds of the sampled aged rice. Although aldehydes and ketone were found in every storage time, they did not show a significant difference (p < 0.05). However, the content of 18-crown-6 decreased from 30.14 to 7.42 and significant difference (p < 0.05) was observed in the process of aging.
2.2.5. SPME-GC–MS analysis conditions 8.0 g milled rice samples were held in an extraction flask (15 mL) to which 10 μL of 2.424 mg·L-1 4-Methyl-2-pentanol as internal standard was added. The extraction flask was then closed with polytetrafluoroethylene faced silicone septum cap and the SPME fiber of 75 μm Carboxen/ Polydimethylsiloxane (CAR/PDMS) (Sigma–Aldrich Australia) was exposed in the headspace of flask at 80 °C for 30 min. Once the isolation step was completed, the fiber was desorbed for 5 min in the GC injection port. The fiber was cleaned at 300 °C for 30 min based on manufacturer’s recommendations before use. GC–MS analysis was performed using a gas chromatograph-mass 302
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Table 1 Volatile organic compounds during different storage times of rice. Compounds
Relative content (%) ± standard deviation 1 month
Hexanal 2-Methylbutanal 3-Methylbutyraldehyde Acetone 18-Crown-6 Pyridine 4-Methyl-2-pentanol (internal standard)
3 months b
7 months a
15.47 ± 0.79 1.31 ± 0.27c 0.39 ± 0.36c 2.26 ± 0.56b 16.96 ± 0.89b 1.63 ± 0.82a 0.71 ± 0.04
7.20 ± 1.56 1.53 ± 0.32b 2.27 ± 1.25c 2.07 ± 0.81b 30.14 ± 0.83a 0.34 ± 0.06b 1.07 ± 0.03
10 months a
16.09 ± 2.36 2.10 ± 0.37a 5.16 ± 1.69a 8.98 ± 1.43a 12.30 ± 0.64c 0.68 ± 0.31
2.34 0.35 0.64 2.01 7.42 0.36 0.46
± ± ± ± ± ± ±
0.30c 0.02c 0.04c 1.01b 0.25d 0.05b 0.05
Variants possessing the same letter are not significantly (P < 0.05) different.
cavitation of 18-crown-6, a set of BODIYs were synthesized as dyes in which half of porphyrins in structure, contained two pyrrole rings. Surprisingly, and somewhat interestingly, all of BODIPYs performed higher values compared to porphyrins after exposure to 18-crown-6 (as seen in G and R components response value). Dyes of HBDP, BrBDP, NO2BDP, NO2Br2BDP, OCH3BDP and NaiOCH3BDP were more sensitive dyes to 18-crown-6 based on their response values. In order to explain the sensitivity of the BODIPYs to 18-crown-6, structures of dyes and binding forms between dyes and 18-crown-6 were analyzed. As observed in Fig. 3a, central metal ion (Zn) of TPPZn no longer was bonding sites as previously (Gu, Huang, Yao, Teye, & Wen, 2014; Zheng, Shan, Yu, & Wang, 2008), while the phenyl of TPPZn bonded with 18-crown-6. Porphyrins have macrocyclic structure locate at the center of the molecule, as well as 18-crown-6, which led to 18-crown-6 was not ligated to the central metal ion of porphyrins. The phenyl of porphyrins out of the plane is easier to enter the center of macrocycle of 18-crown-6. This bonding form confirmed the former assumption that macrocyclic structure of porphyrins appeared steric hindrance with 18crown-6’s. And the opinion maybe should be changed that odorant bonding with the central of porphyrins is intrinsic to the mechanism of action (Janzen, Ponder, Bailey, Ingison, & Suslick, 2006). The previous binding principle was adapted to small organic components, but things are quite different when the VOCs are macromolecular compounds. The compound 18-crown-6 provides a readily available archetype of such unique macrocyclic structure. BODIPYs are comprised of two pyrrole rings retained chromophore partly and own smaller structure, compared to porphyrins’. The smaller structure of BODIPYs allowed them to easily bond with 18-crown-6, as shown in Fig. 3b, corresponding to their response values. However, 2, 8-substituted BODIPY, like NO2BrBDP, exhibits lower sensitivity to 18-crown-6 compared to other BODIPYs, and the sensitivity of pCarBDP also obviously decreased. As shown in Fig. 3c and d, such differences are widely attributed to 2substituent preventing 18-crown-6 binding with 8-substituent group of BODIPY and the relatively larger structure of 8-substituent leading to the steric hindrance effect. To test the capacity of resisting disturbance in real rice samples, six dyes sensitive to 18-crown-6 were applied to detect the four kinds of rice stored for different durations. The color and brightness of points in Fig. 3e expressed color components (R, G and B) responding and sensitivity of dyes. Brighter point means color of dyes changed more after exposure, that is, dyes were more sensitive to VOCs. As displayed in Fig. 3e, all points of HBDP, BrBDP and NO2Br2BDP were dim, which demonstrated that the sensitivity of these dyes decreases dramatically in rice. It means that other VOCs in rice samples disguised the 18crown-6 and thus resulted in low response values from these dyes. The R component of NaiOCH3BDP changed after exposure to rice samples, but it showed similar response value among four kinds of rice. From the digital image of OCH3BDP, it was easy to discriminate the four kinds of rice. Its brightness gradually dimmed following the storage time, just same as the tendency of the concentration of 18-crown-6. Response values of NO2BDP in R, G and B components were similar which
3.2. The selection of chemically responsive dyes corresponding to VOCs of rice As the key material of colorimetric sensor array, chemically responsive dyes directly influence the accuracy of VOCs detection. In previous research, colorimetric sensor array based on universal porphyrins and pH indicators was unable to identify the storage time of rice correctly. In this work, some dyes with simpler structures were synthesized aiming at the special structure of 18-crown-6 to improve the sensitivity of the colorimetric sensor array. 3.2.1. The sensitive chemically responsive dyes to the 18-crown-6 The component 18-crown-6 was the characteristic VOC in the aging as revealed by the results of GC–MS and the minimum concentration was 0.15 g·L−1. For achieving rice storage monition, six porphyrins (numbered 1 to 6, as shown in Fig. 2) were chosen as dyes to detect 18crown-6 with concentration of 0.10 g·L−1. Six porphyrins exhibited similarly and none of them significantly higher in values comparatively with each other after exposure to 18-crown-6. The component 18crown-6 is one of macrocyclic polyethers which can selectively bind with various cations due to the cavitation of the macrocycle. Both porphyrins and 18-crown-6 are macromolecular, and their increasing steric hindrance may decrease their binding degree. Considering the
Fig. 2. Response values of different dyes exposure to volatile 18-crown-6 and the structure of sensitive dyes. 303
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 3. Difference in binding forms between dyes ((a) TPPZn, (b) NO2BDP, (c) NO2BrBDP, (d) pCarBDP and 18-crown-6; (e) Response values of sensitive dyes in rice with different storage time.
demonstrated that some of acid gases changed in the aging process. Consequently, TPPClFe and BCP were screened as sensitive dyes to detect the whole flavor of rice in different storage times.
reflected in response digital image as white points. Here, response values in red component were 38.31 ± 2.03a, 33.87 ± 2.68b, 25.28 ± 1.76c and 20.25 ± 1.13d, respectively, for the four rice storage times with significant difference observed among each group. Based on variance analysis above, NO2BDP and OCH3BDP were considered as efficient dyes for monitoring the storage time of rice.
3.3. The results of a colorimetric sensor array application in rice storage time detection 3.3.1. Colorimetric sensor array responses for different storage times of rice Based on the above selection results, four kinds of sensitive dyes were employed to monitor the storage time of rice. The color differences of sensor arrays were visualizations by the difference maps that were superimposed on differences in the gray values of red, green and blue (RGB). Fig. 5a clearly shows the difference images of colorimetric sensor array after exposure to rice with different storage times. Different storage time rice had their unique different images which can are distinguishable with the naked eye. The results indicated that the developed colorimetric sensor array has the potential for monitoring the storage time of rice.
3.2.2. The chemically responsive dyes characterized VOCs of rice with different storage times Although characteristic gas provided the possibility of differentiating among different storage times of rice, other VOCs with natural aroma and stale flavor formed during aging cannot be underestimated. VOCs of rice during aging are complexed and dynamic. Therefore characterizing the whole information of VOCs facilitates the development of colorimetric sensor array with high specificity. There were ten dyes (excluded six dyes sensitive to 18-crown-6, added two pH indicators (BCP and DMY)) were exposure to rice samples with results showed in Fig. 4. TPP, pCarBDP and NO2BrBDP had lower response values and no significant difference among the four kinds of rice storage times. This demonstrated that they are not sensitive to the flavor of rice in the whole aging process. TPPZn, TPPMn and TPPClMn were observed to generate higher but same response values among the different aged rice. TPPClFe performed good discrimination in rice storage time while the color change files of FTPPClFe just discriminated the first two kinds of rice from others. The fluorine-based of FTPPClFe exerted a negative effect for its sensitivity in this study. Four kinds of rice were also classified through the color change files of BCP which
3.3.2. Principal component analysis (PCA) Principal component analysis (PCA) is widely used in dimensionality reduction, involving the extraction of relevant information from complexed data sets (Lu, Tan, & Wang, 2013). The relevant information was turn into PCs by the process of reconstructing the original data using an orthogonal transformation (Yu & Wang, 2007; Zhang, Tian, & Pei, 2014). When accumulation contribution rate of first several principal components reaches 85%, they can represent data-set of rice sample without loss in variables information. In this work, color files of rice with different storage times were analyzed by PCA. The result showed that the first three principal components had their cumulative reliability summing up to 94.21% (> 85%). So scores of the first three PCs were used to construct 3-dimensional (3D) PCA scatter plot to visualize the cluster trends of samples as shown in Fig. 5b. The data-set performed a good discrimination in most rice samples. Data points of the 1-month-storage-time and the 10-month-storage-time rice clustered slightly; others were separated from each other. Some volatile gases increased first and then subsequently decreased with the aging of rice. Hence they performed similar to each other in the color change files and cluster slightly in 3-dimensional (3D) PCA scatter plot. Furthermore, the sample distribution of the rice with 1 and 3-months-storagetimes were relatively discrete, whereas the distribution of samples with
Fig. 4. Response values of dyes after exposure to rice with different storage time. 304
Food Chemistry 268 (2018) 300–306
H. Lin et al.
Fig. 5. (a) Difference images of colorimetric sensor array expose to different storage time of rice; (b) The 3-D scatter plot constructed by the first three PCs for all samples; (c) Discrimination rates of LDA model according to different PCs.
selected in characterizing the flavor of rice in the aging process. NO2BDP, OCH3BDP, TPPClFe and BCP were eventually chosen to construct colorimetric sensor array for monitoring the rice storage time and good characterization was achieved by PCA and LDA analysis.
7 and 10-month-storage-times appeared more clustered. It might be that VOCs of rice changed continuously (aroma decreasing and stale flavor increasing) in the initial aging process. 3.3.3. Determination results of rice samples based on the colorimetric sensor array Linear Discriminant Analysis (LDA) as one of more effective tools was applied to discriminate rice storage time. Unlike PCA, LDA is guaranteed to provide maximum separation between classes and minimum within classes according to their class labels (Martinez & Kak, 2001). In this work, LDA was employed to discriminant rice storage time based on PCA scores from PCA analysis and labels of samples. Total number of 20 samples from each class were applied to training of calibration model, and the remaining 10 samples for testing of the calibration model. Fig. 5c shows the discrimination rates of LDA model according to different principal components (PCs). The LDA model performed good discrimination in rice storage when PCs = 4. The discrimination rates of calibration set and prediction set were respectively 98.75% and 97.50% with the ideal number of PCs, up 12.50% in prediction set from our previous study (Guan et al., 2016). There is only one sample of three months was incorrectly identified as seven month in the prediction set.
Acknowledgements This work has been financially supported by Foundation for the National Key Technology R&D Program of China (Grant No. 2015BAD19B05), China Postdoctoral Natural Science Foundation (2016M601746), Jiangsu province Postdoctoral Natural Science Foundation (1601135B). References Aibara, S., Ismail, I. A., Yamashita, H., Ohta, H., Sekiyama, F., & Morita, Y. (2014). Changes in rice bran lipids and free amino acids during storage. Agricultural and Biological Chemistry, 50(3), 665–673. Choi, S., Jun, H., Bang, J., Chung, S. H., Kim, Y., Kim, B. S., et al. (2015). Behaviour of Aspergillus flavus and Fusarium graminearum on rice as affected by degree of milling, temperature, and relative humidity during storage. Food Microbiology, 46, 307–313. Feng, T., Zhuang, H., Ye, R., Jin, Z., Xu, X., & Xie, Z. (2011). Analysis of volatile compounds of Mesona Blumes gum/rice extrudates via GC–MS and electronic nose. Sensors and Actuators B: Chemical, 160(1), 964–973. Griglione, A., Liberto, E., Cordero, C., Bressanello, D., Cagliero, C., Rubiolo, P., et al. (2015). High-quality Italian rice cultivars: Chemical indices of ageing and aroma quality. Food Chemistry, 172, 305–313. Gu, H., Huang, X., Yao, L., Teye, E., & Wen, Y. (2014). Effects of substitute group, axial ligand and volatile organic compounds on binding ability of colorimetric sensor array. Materials Technology, 29(4), 220–226. Guan, B., Zhao, J., Jin, H., & Lin, H. (2016). Determination of rice storage time with colorimetric sensor array. Food Analytical Methods, 10(4), 1054–1062. Guan, B., Zhao, J., Lin, H., & Zou, X. (2013). Characterization of volatile organic compounds of vinegars with novel electronic nose system combined with multivariate analysis. Food Analytical Methods, 7(5), 1073–1082. Huang, X., Gu, H., Yao, L., Teye, E., & Wen, Y. (2014). A study of the interactions between colorimetric sensor array and volatile organic compounds. Journal of Computational and Theoretical Nanoscience, 11(11), 2304–2309. Huang, X., Zou, X., Zhao, J., Shi, J., Zhang, X., Li, Z., et al. (2014). Sensing the quality parameters of Chinese traditional Yao-meat by using a colorimetric sensor combined with genetic algorithm partial least squares regression. Meat Science, 98(2), 203–210. Janzen, M. C., Ponder, J. B., Bailey, D. P., Ingison, C. K., & Suslick, K. S. (2006).
4. Conclusion A novel colorimetric sensor array was designed to discriminate rice samples with different storage times. A series of BODIPYs were targeted synthesized as chemically responsive dyes to detect the 18-crown-6 determined by GC–MS results of four kinds of rice with different storage times as a characteristic gas associated with aging of rice. It is the first attempt of selecting BODIPYs as dyes for the fabrication of colorimetric sensor array, for which remarkable success achieved vis-à-vis their sensitivities. Meanwhile, the successful application of BODIPYs as chemically responsive dyes was explored as a way to improve the sensitivity of the colorimetric sensor array through changing dyes structure according to different VOCs. Furthermore, ten dyes were also 305
Food Chemistry 268 (2018) 300–306
H. Lin et al.
system for assessing rice quality during storage at different temperatures. Journal of Stored Products Research, 59, 204–208. Suslick, K. S., Rakow, N. A., & Sen, A. (2004). Colorimetric sensor arrays for molecular recognition. Tetrahedron, 60(49), 11133–11138. Thanathornvarakul, N., Anuntagool, J., & Tananuwong, K. (2016). Aging of low and high amylose rice at elevated temperature: Mechanism and predictive modeling. Journal of Cereal Science, 70, 155–163. Tsuzuki, W., Suzuki, Y., Yamada, S., Kano, S., Ohnishi, H., Fujimoto, T., et al. (2014). Effect of oxygen absorber on accumulation of free fatty acids in brown rice and whole grain wheat during storage. LWT – Food Science and Technology, 58(1), 222–229. Wu, W., Zhao, J., Guo, H., Sun, J., Ji, S., & Wang, Z. (2012). Long-lived room-temperature near-IR phosphorescence of BODIPY in a visible-light-harvesting N^C^N PtII–acetylide complex with a directly metalated BODIPY chromophore. Chemistry – A European Journal, 18(7), 1961–1968. Xiaowei, H., Xiaobo, Z., Jiewen, Z., Jiyong, S., Zhihua, L., & Tingting, S. (2015). Monitoring the biogenic amines in Chinese traditional salted pork in jelly (Yao-meat) by colorimetric sensor array based on nine natural pigments. International Journal of Food Science & Technology, 50(1), 203–209. Yu, H., & Wang, J. (2007). Discrimination of LongJing green-tea grade by electronic nose. Sensors and Actuators B: Chemical, 122(1), 134–140. Zhang, L., Tian, F., & Pei, G. (2014). A novel sensor selection using pattern recognition in electronic nose. Measurement, 54, 31–39. Zheng, W., Shan, N., Yu, L., & Wang, X. (2008). UV–visible, fluorescence and EPR properties of porphyrins and metalloporphyrins. Dyes and Pigments, 77(1), 153–157. Zhou, Z., Robards, K., Helliwell, S., & Blanchard, C. (2002). Ageing of stored rice: changes in chemical and physical attributes. Journal of Cereal Science, 35(1), 65–78.
Colorimetric sensor arrays for volatile organic compounds. Analytical Chemistry, 78(11), 3591–3600. Kim, S. Y., Li, J., Lim, N. R., Kang, B. S., & Park, H. J. (2016). Prediction of warmed-over flavour development in cooked chicken by colorimetric sensor array. Food Chemistry, 211, 440–447. Loudet, A., & Burgess, K. (2007). BODIPY dyes and their derivatives: syntheses and spectroscopic properties. Chemical Reviews, 107(11), 4891–4932. Lu, J., Tan, Y. P., & Wang, G. (2013). Discriminative multimanifold analysis for face recognition from a single training sample per person. IEEE Transactions on Pattern Analysis and Machine Intelligence, 35(1), 39–51. Martinez, A. M., & Kak, A. C. (2001). PCA versus LDA. IEEE Transactions on Pattern Analysis and Machine Intelligence, 23(2), 228–233. Musto, C. J., & Suslick, K. S. (2010). Differential sensing of sugars by colorimetric arrays. Current Opinion in Chemical Biology, 14(6), 758–766. Park, C.-E., Kim, Y.-S., Park, K.-J., & Kim, B.-K. (2012). Changes in physicochemical characteristics of rice during storage at different temperatures. Journal of Stored Products Research, 48, 25–29. Radi, Ciptohadijoyo, S., Litananda, W. S., Rivai, M., & Purnomo, M. H. (2016). Electronic nose based on partition column integrated with gas sensor for fruit identification and classification. Computers and Electronics in Agriculture, 121, 429–435. Rakow, N. A., & Suslick, K. S. (2000). A colorimetric sensor array for odour visualization. Nature, 406(6797), 710–713. Sanaeifar, A., Mohtasebi, S. S., Ghasemi-Varnamkhasti, M., & Ahmadi, H. (2016). Application of MOS based electronic nose for the prediction of banana quality properties. Measurement, 82, 105–114. Sung, J., Kim, B.-K., Kim, B.-S., & Kim, Y. (2014). Mass spectrometry-based electric nose
306