A digital image-based micro-flow-batch analyzer

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy Microchemical Journal 106 (2013) 238–243 Contents lists available at SciVerse ScienceDirect Microchemical Journal journal homepage: www.elsevier.com/locate/microc A digital image-based micro-flow-batch analyzer Marcelo B. Lima, Stéfani Iury E. Andrade, Inakã S. Barreto, Luciano F. Almeida, Mário César U. Araújo ⁎ Universidade Federal da Paraíba, Departamento de Química, João Pessoa, PB, Brazil a r t i c l e i n f o Article history: Received 15 May 2012 Received in revised form 18 July 2012 Accepted 22 July 2012 Available online 27 July 2012 Keywords: Micro-flow-batch analyzer Urethane-acrylate resin Digital images RGB color model Tannins Green tea a b s t r a c t This paper introduces for the first time the combination of digital images taken with a webcam and a microflow-batch analyzer as a novel strategy for implementing quantitative chemical analysis. The digital imagebased micro-flow-batch analyzer (DIB-μFBA) was formed using urethane-acrylate resin, glass slides and ultraviolet lithography. The glass slides were used as sealant layers on both sides of the substrate urethane-acrylate and provide the necessary transparency to the microsystem to conduct studies using digital images. DIB-μFBA uses digital images obtained from a webcam with CCD sensor, based on the RGB (red–green–blue) color model. The analyzer was used for determining the tannins in green tea employing the well-known ferrous tartrate method. All standard solutions were prepared in-line, and all analytical processes were completed by simply changing the operational parameters in DIB-μFBA control software. The paired t test, at a 95% confidence level, showed no statistically significant differences between results obtained by DIB-μFBA and the spectrophotometric reference method. The proposed microsystem presented satisfactory physical and chemical properties while keeping the flexibility, versatility, robustness and multi-task characteristics of conventional flow-batch analysis. Therefore, it was possible to build a low-cost device with high sample throughput (about 190 h−1) and reduced reagent consumption (about 300 times less than the reference method), contributing to the basic principles of green chemistry and the advancement of micro-analytical procedures. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The automatic flow-batch system was developed in 1999 by Honorato et al. [1]. This system is an instantaneous stop chamber flow system, which integrates batch and flow analysis methods, through the use of programmed multi-commutation [2,3]. The main component is the mixing chamber where the whole analytical process, including fluid addition, sample pretreatment, homogenization, precipitation, extraction, preparation of standard solutions and detection, takes place under the total control of the software [4]. The sample is processed with less: manipulation, consumption of reagents and samples, waste and chance of human error. Classical (discrete) methods can be performed with precision, accuracy and sampling throughput to other flow analysis methods [5]. Recently, the flow-batch analyzer was miniaturized (μFBA) and applied to determine Fe(II) in iron-based supplements (oral solutions) using the 1,10-phenanthroline method [6]. Micro-flow-batch analyzer (μFBA) was also used for the determination of phosphorus in biodiesel, employing the molybdenum blue method [7]. The microfabrication used deep ultraviolet lithography and photopolymerizable urethane- ⁎ Corresponding author at: Departamento de Química, CCEN, Universidade Federal da Paraíba, Caixa Postal 5093, CEP 58051‐970, João Pessoa, PB, Brazil. Tel.: +55 83 3216 7438; fax: +55 83 3216 7437. E-mail address: [email protected] (M.C.U. Araújo). 0026-265X/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.microc.2012.07.010 acrylate photoresist. It contained an integrated photometer based light emitting diode (LED) and phototransistor for detection, and a nylon wire homogenizing system. Digital image-based methods have frequently been used as an alternative for quantitative determinations in analytical sciences [8–12]. These methods usually employ the RGB (red–green–blue) color model whose individual values vary from 0 up to 255 (8 bits). The data may be obtained from a scanner, a webcam and digital cameras along with charge-coupled devices (CCDs) or complementary metal-oxide semiconductor (CMOS) sensors [13]. Mathematical models for digital images and RGB data have also been employed to construct analytical curves in quantitative determinations [14–16]. Lyra et al. [17] proposed a method using the norm for position vectors in an RGB three-dimensional space concept. This method was applied to the determination of lithium, sodium and calcium in anti-depressive drug, physiological serum and water, respectively, using the flame emission spectrometry (FES). This same method was applied to the determination of sodium diclofenac, sodium dipyrone and calcium gluconate in injectable forms [18], and for determining the total acidity in red wines using acid–base titration without an indicator [19]. Recently, Andrade et al. [20] developed a digital image-based flow-batch analyzer. This work, a mixing chamber, built in acrylic (volume 800 μL), was used in the determination of aluminum(III) and chromium(VI) in natural waters. The digital images were obtained employing a webcam and the analytical curve was obtained Author's personal copy M.B. Lima et al. / Microchemical Journal 106 (2013) 238–243 using one of the components of the RGB color model, regarding the complementary color of each complex analyzed. Tannins are a polyphenol group, astringent, with high molecular weights, and have a natural ability to complex with metals. Tannins in general have antioxidant, anticarcinogenic, antimutagenic and antimicrobial activities and can assist in the prevention and treatment of diseases [21]. Tannins are found in common foods and drinks, but especially in green tea, which are consumed worldwide. Green tea is derived from the infusion of twigs and dried leaves of Camellia sinensis [22]. Quality control is certainly important for the world's tea consumers, but the quantification of tannins has always been of particular interest to the industry, as well as to the governmental regulators [23]. The official method in Japan (Official Chemical Analysis of Tea) uses the ferrous tartrate method as the reference method for the determination of tannins in tea samples [24]. This spectrophotometric method is based on the formation, in a buffered medium (pH 6.8), of a complex between ferrous tartrate and tannins in the tea, which absorbs light at around 560 nm. The ferrous tartrate method is more selective than the Folin–Ciocalteau method [25] and is unaffected by the coexistence of reducing agents, such as ascorbic acid [26]. It is being used by researchers for the spectrophotometric determination of tannins in tea [27–32]. In this study, a microfabrication technique, based on deep ultraviolet lithography, photocurable urethane-acrylate resin and glass slides, was used to develop a digital image-based micro-flow-batch analyzer (DIB-μFBA). The glass slides were used as sealant layers on both sides of the substrate urethane-acrylate in order to provide the necessary transparency to the microsystem to conduct studies using digital images. DIB-μFBA uses digital images obtained from a webcam with a CCD sensor, based on the RGB (red–green–blue) color model. It also includes a nylon thread inserted into the micro-chamber to implement simpler, faster and efficient stirrings. The analyzer was used for determining the tannins in green tea using the well-known ferrous tartrate method. 2. Experimental 2.1. Reagents and solutions All reagents were of analytical grade and freshly distilled and deionized water (>18 MΩ cm −1) was used to prepare all solutions. A stock solution of 1000 mg L −1 of tannic acid was prepared with 0.1000 g of tannic acid (Labsynth) diluted to 100.0 mL with water. A ferrous tartrate reagent solution was obtained by mixing 1.0000 g of heptahydrated ferrous sulfate (Vetec) with 2.0000 g of potassium sodium tartrate (Vetec) and 0.1000 g of sodium bisulfite (Reagen) diluted in water to 100.0 mL. The phosphate buffer pH 6.8 was prepared by mixing 0.10 mol L −1 heptahydrated sodium phosphate (Reagen) with 0.10 mol L −1 sodium hydroxide (Vetec). The commercial urethane-acrylate photoresist used for the fabrication of the μFBA was acquired by Carimbos Medeiros Ltda, Brazil (MacDermid, flex-light trademark M050). 2.2. Sample preparation Samples were obtained from six infusions of green tea purchased from different manufacturers and lots. The bags of green tea were purchased in local shops in João Pessoa City, Paraíba State, Brazil. For the sample preparation, an amount of 0.5000 g of tea was heated for 10 min at 90 °C in about 50 mL of deionized water, and the mixture was filtered, washing the residues and completing to 100.0 mL with deionized water. 239 2.3. Apparatus and software To fabricate the micro-chamber (μCH) in urethane-acrylate resin, we used a commercial UV light source (Fotolight-MD2-A4, Carimbos Medeiros Ltda, Brazil), with two sets of mercury lamps (BLB-15W-T8, SCT black light). For the layout design of the μCH, the CorelDraw® X5 program was used. The layout printing was on polyester transparency films for laser printing using an HP LaserJet P2014. After UV exposure, channels on the substrate were revealed by the removal of the non-exposed resin with an ultrasonic bath (UltraCleaner 800, Unique, Brazil). A Hewlett-Packard diode array (model 8453) UV–vis spectrophotometer equipped with a cuvette (inner volume of about 4.0 mL, and optical path of 1.0 cm) was used for absorbance measurements when employing the spectrophotometric reference method. A Philips Webcam SPC900NC VGA and a CCD were used in conjunction with LabVIEW® 7.1 (National Instruments ®) software to control the DIB-μFBA. The images were captured by means of the software written in Delphi (version 3.0). 2.4. Assembly of the DIB-μFBA system A peristaltic pump (model 78001‐12, 8 channels, Ismatec) for fluid propulsion was used, which operated at 10 rpm. Minisolenoid valves (model LHDA 0531415H, Lee Company) were used for controlled fluid additions to the micro-mixing chamber. A 0.4 mm nylon wire was used for agitation within the micro-mixing chamber to ensure the mixing of the added products. The nylon was adapted to a CD/ DVD-ROM motor drive (model MDN3GT3CPAC, 2000 rpm, 5 Vdc). Teflon® tubes with 0.5 mm internal diameter were used for fluid transport. All tasks, such as data acquisition, and valve and drive motor activation, were done using a USB interface (USB6009, National Instruments®), which activated a lab made controller module. The software was developed in LabVIEW® 7.1 (National Instruments ®). The DIB-μFBA was kept in a wooden box measuring 17 cm× 11 cm× 9 cm. Six white high intensity LEDs (NPE, Thailand) were placed at the ceiling of the box to give a constant light intensity throughout the experiment. The interior walls of the box were covered with white paper in order to provide uniform illumination and reduce glare [8,16,19]. 2.5. Lithographic micro-chamber fabrication process As outlined in Fig. 1, the μFBA was fabricated based on the methodology described by S.S. Monte-Filho et al. [6]. Initially the required layout or template was developed and printed. The layout was set between a 2.0 mm acrylic plate, and a 3.4 mm rubber frame which formed the mold, as shown in Fig. 1a. The urethane-acrylate resin was then deposited on the mold and immobilized by another acrylic plate, as sketched in Fig. 1b and c, respectively. The thickness of the rubber framing (in this case 3.4 mm), allows you to define the volume of the system. The layout was engraved in the resin by exposure to UV radiation in one step. The bottom was cured by exposure for 250 s, as illustrated in Fig. 1d. The uncured resin layer was removed with the aid of an ultrasonic bath in a 10% (v/v) detergent solution for 15 min. This layer was then dried in a nitrogen flow. The exposure for the polymerization of only one side of the layer (underside) allows radiation to exceed the resin layer, not allowing the resin to polymerize in the parts printed with toner. Thus, it is possible to obtain a satisfactory resin layer to be sealed with another material in the next step. In the sealing step, the resin layer and the glass slides were superimposed appropriately to form the required channels (Fig. 1e). Tubes of Teflon® were carefully inserted into their respective Author's personal copy 240 M.B. Lima et al. / Microchemical Journal 106 (2013) 238–243 The glass slides provide transparency of the micro-chamber, which allows efficient capturing of the digital images. 2.6. Analytical procedure Before starting the analytical procedure, working solutions for each channel are pumped and re-circulated to their respective reservoirs (Fig. 2a). Then the mini-valves VS, VR1, VR2 and VC are switched on for 3.0 s and the working solutions (S, R1, R2, and C) are pumped to the micro-chamber to fill the channels between the valves and the chamber. Then, immediately, the waste valve VW is switched on for 5.0 s and then any solution inside the micro-chamber (μCH) is emptied using the peristaltic pump for aspiration. This channel filling procedure is very important and must be carried out whenever there is a change of the reservoir liquids. Fig. 2a represents the schematic diagram of the DIB-μFBA used for the determination of tannins by the ferrous tartrate method. They have in common every preparation step for mixing/homogenization in the μCH, capturing of digital images, waste and cleaning. Homogenization is performed by the drive motor (DM) coupled to a nylon wire (NW), the digital image is captured and the μCH is emptied. Afterwards the μCH is cleaned by switching on valve VC, adding water (C). Then, VW is switched on to waste the contents of the μCH. This cleaning and waste procedure must be done twice to effectively clean the μCH. The timing diagrams that illustrate the analytical procedures for both photometric methods are described in Fig. 2b. The time intervals chosen for each in-line preparation are for the drive motor (TDM), switching valves (TS TR1, TR2, TC and TW), and the capture of the digital images (TWC). All preparations used a peristaltic pump (PP) with a flow of 34.0 ± 0.2 μL s −1 (n = 20) for all channels. For in-line blank preparation, valves VR1 (phosphate buffer solution) and VR2 (ferrous tartrate solution) are simultaneously switched on for 1.5 s, then mixing took place for 2.0 s and the capture of the digital images is realized for 1 s at 10 images per second. Fig. 1. Schematic diagram illustrating the procedure for the construction of DIB-μFBA. (a) Mold mounting with the required layout, (b) depositing the urethane acrylic resin over the mold, (c) immobilizing the resin with the second acrylic plate, (d) UV radiation exposure for 250 s, (e) positioning of the layers: glass slide/substrate/glass slide, and (f) layer sealing with inserted Teflon® tubes and glass slides, and UV radiation exposure for 900 s. channels. The system was again exposed to UV radiation in a single step of 900 s on both sides, as shown in Fig. 1f. This exposure time allows an efficient and irreversible seal. Teflon® tubes (0.8 mm diameter) fixed in the channels allow micro-chamber communication with the external environment, making possible the insertions of fluids as well as the agitator shaft. Fig. 2. (a) DIB-μFBA diagram with its dimensions. (b) Determination time diagram for the samples. Micro-chamber (μCH), peristaltic pump (PP), drive motor (DM), nylon wire (NW), webcam (WC), minisolenoid valves (VS, VR1, VR2, VC and VW), working solution or sample (S), phosphate buffer pH 6.8 (R1), ferrous tartarate (R2), water (C) and waste. The time intervals (in seconds) TS, TR1, TR2, TC and TW, correspond to the minisolenoid valves (VS, VR1, VR2, VC and VW), TDM, drive motor and TWC is the digital images' capture time. Author's personal copy M.B. Lima et al. / Microchemical Journal 106 (2013) 238–243 The in-line preparations of standard solutions from 10 to 100 mg L −1 were performed using the working solution at 110 mg L −1 prepared from the appropriate dilution of stock solution at 1000 mg L −1. In these preparations, the valves VS, VR1 and VR2 are activated simultaneously. A working solution (S), phosphate buffer pH 6.8 (R1) and ferrous tartrate reagent (R2) are added in the μCH and homogenized and digital images are captured. VR2 is switched on for 1.0 s and the on times of VS and VR1 vary proportionately with the concentration of the standard solution being prepared. For the in-line preparation of the sample, the procedure is similar to the preparation of standard solutions. The prepared samples as described in Section 2.2 were used without further dilution. The difference here is also that the samples are used instead of the working solution. The time intervals used for this analysis are shown in Fig. 2b. This method used deionized water for cleaning the μCH. 2.7. Reference method The results obtained with the proposed mini-system were compared with the official Japanese ferrous tartrate method [24,26]. Standard solutions from 10 to 100 mg L −1 of tannic acid were prepared by adding appropriate volumes of 1000 mg L −1 stock solution, 10.0 mL of ferrous tartrate solution, 50.0 mL of phosphate buffer (pH 6.8) and water until a final volume of 100.0 mL in a volumetric flask was achieved. For the determination of tannins in green tea we added 10.0 mL aliquots of the sample, 10.0 mL of the reagent and 50.0 mL of the buffer using the same conditions as were used in determining the standard solutions. After 10 min of reaction, the absorbance of all solutions, both standard and sample was measured at 550 nm. 3. Results and discussion 3.1. Characterization of DIB-μFBA Lithographic microfabrication with ultraviolet urethane-acrylate resin and glass slides is reproducible. The sealing of the partially cured layer with UV exposure was performed with both glass slides and Teflon® tubes already properly positioned in their respective channels. Due to the risk of clogging, the technique requires caution, and proper control of the exposure time. A time of 900 s is satisfactory to ensure effective sealing of the resin layer with the glass slides and Teflon® tubes. The transparent surface of the micro-chamber allows the capture of digital images by webcam reproducibly. Homogeneous illumination of the environmental analysis was performed using six white LEDs (with an approximate light output of 0.2 W per LED) which were arranged in the ceiling of the box between the mixing chamber and the webcam. In conventional flow-batch analysis (FBA) the mixing chamber contains a stirring bar that allows fluid agitation. In the DIB-μFBA system agitation is performed using a nylon wire with a shovel tip. Complete mixing of the solutions in the micro-chamber was obtained in less than 2 s, due to the high speed of the diver motor (2000 rpm). 241 3.2. The captured digital images and mathematical model The treatment of the captured digital images was made by means of a second software also written in Delphi (version 3.0). The routine with the working stages of this software is similar to that used elsewhere [17–19]. Initially the user selects the most homogeneous region in the image which will define the coordinates of the selected region, and will also be used for all other images. Then the software scans all the pixels column by column to extract the RGB component for each pixel and calculate a mean integer value of each RGB component. These mean values are used in the RGB-based value calculation (analytical response) as described below. The RGB-based values were calculated by means of a mathematical model developed from the concept of vector norm “‖v‖” [17], calculated as: kvk ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 þ B 2 þ G 2 R s−b s−b s−b ð1Þ     ,G where R s−b s−b and B s−b result from the difference between the R s ,   G s and B s average values obtained from digital images of the standard  and B  ,G  from the blank. solutions and samples and R b b b As a result, a linear relationship was observed between the analyte concentration (C) in the standard solution (or sample) and RGB data, for which the following equation is valid: kvk ¼ k C: ð2Þ As demonstrated in previous work [17–19] Eq. (2) provides the basis for building DIB-μFBA analytical curves, establishing a linear relationship between ‖v‖ (RBG data value adopted as analytical response) and analyte concentration in standard solutions. Furthermore, the vectors associated with the digital images from each analyte should be positioned on the same support line in the RGB three-dimensional space [14]. Fig. 3 shows the digital images obtained from six standard solutions with different concentrations. The first image in the sequence is a blank solution. All images represent a selected area equal to 55 × 35 pixels. As can be seen in Fig. 3, the images present an increase in the intensity for the color dark purple, which is proportional to the tannin concentration in each standard solution. Other linear relationships between the complex color and the ratios B/R, B/G, B/Itot (Itot = RGB), and log(B) were also studied to possibly maximize precision, as already done in other papers [8,20]. However, in all cases the results were poor when compared to using a vector norm for concentration of the complex formed. The statistical quality of the regression was verified by the residuals and analysis of variance (ANOVA). They confirmed the homoscedastic distribution of residuals and a significant linear regression. 3.3. DIB-μFBA application The equations for the analytical curves obtained using DIB-μFBA and the spectrophotometric reference method are respectively: A = 0.6309 + 0.4796C (r 2 = 0.998, with n = 5), and A = 0.0034 + 0.0135C (r 2 = 0.9996, with n = 5), where A is the analytical response and C is the concentration of the analyte in mg L −1 of tannic acid. Both Fig. 3. Digital image captured and selected area (35 × 55 pixels), the complex color in some standard solution (mg L−1) and blank. Author's personal copy 242 M.B. Lima et al. / Microchemical Journal 106 (2013) 238–243 Table 1 Results for tannin determinations in green tea (mg L−1) using colorimetric method (DIB-μFBA) and the reference method. Samples DIB-μFBA Reference 1 2 3 4 5 6 37.51 ± 0.05 59.34 ± 0.02 63.09 ± 0.03 58.13 ± 0.05 49.25 ± 0.05 52.03 ± 0.02 36.98 ± 0.02 59.29 ± 0.01 62.85 ± 0.03 57.65 ± 0.02 49.10 ± 0.02 52.16 ± 0.01 curves are linear in the 10.0 to 100.0 mg L −1 of tannic acid. Its analytical curves were statistically validated by the analysis of variance showing no lack of fit at a 95% confidence level. The limit of detection (LOD) was defined in 2.74 mg L −1 and 4.26 mg L −1 for DIB-μFBA and the reference method, respectively. The LOD for the method was calculated based on the criteria established by IUPAC, and the LOD was evaluated as three times the standard deviation of the blank measurement [33]. The LOD value, although relatively high may be acceptable for the analyte under study, considering the high concentrations of tannin in green tea [34]. Table 1 presents the results obtained with DIB-μFBA and the spectrophotometric reference method. No statistically significant differences at a confidence level of 95% were observed between the results when applying the paired t-test. DIB-μFBA presented an analytical frequency of about 190 h −1 for the determination of tannins in green tea, with a waste generation of the 306 μL per analysis while the spectrophotometric reference method [24,26] provides an analytical frequency lower than 10 h −1 and a waste generation of about 300 times greater than the proposed method. 4. Conclusion This study proposed the fabrication of DIB-μFBA, which was accomplished by employing deep UV-lithography with urethane-acrylate resin and glass slides. The glass slides were used as sealant layers on both sides of the substrate and provided the necessary transparency to conduct studies using digital images. When compared to the conventional flow-batch system previously developed in acrylic [20], DIB-μFBA presents better chemical solvent resistance [35] and smaller volume of the mixing chamber (about 8 times smaller). As a consequence, the analysis may be carried out with less: consumption of reagents and sample, cost and chemical waste. The proposed DIB-μFBA was successfully developed for the analysis of tannins in green tea. We obtained a high speed of analysis, thanks to the optimizations in the addition times and volumes of the fluids, in a microsystem that allowed a homogenization efficient and discrete measurements. Although the proposed system has been applied only to samples of green tea, the ferrous tartrate method can be also applied, in the same DIB-μFBA, for other types of tea, such as black tea, oolong tea, and white tea, where the tannin may vary between 10 and 100 mg L −1 [26–31]. By the use of an inexpensive webcam for analytical detection, the proposed DIB-μFBA offers an alternative to traditional photometry. The use of the webcam dispenses with wavelength selection, which both reduces costs, and simplifies the instrumentation. With the webcam it is also possible to implement chemometric treatments due to the trivariate nature of the detection as well as the spatial-resolution characteristics inherent in digital images. Acknowledgments The authors would like to thank the Brazilian agencies (CNPq and CAPES) for the research fellowships and scholarships. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.microc.2012.07.010. References [1] R.S. Honorato, M.C.U. Araújo, R.A.C. Lima, E.A.C. Zagatto, R.A.S. Lapa, J.L.F.C. Lima, A flow-batch titrator exploiting a one-dimensional optimisation algorithm for end point search, Anal. Chim. Acta 396 (1999) 91–97. [2] I.S. Barreto, S.I.E. Andrade, M.B. Lima, E.C. Silva, M.C.U. Araújo, L.F. Almeida, A monosegmented flow-batch system for slow reaction kinetics: spectrophotometric determination of boron in plants, Talanta 94 (2012) 111–115. [3] E.A.G. Zagatto, J.M.T. Carneiro, S. Vicente, P.R. Fortes, J.L.M. Santos, J.L.F.C. Lima, Mixing chambers in flow analysis: a review, J. Anal. Chem. 64 (2009) 524–532. [4] P.H.G.D. Diniz, L.F. Almeida, D.P. Harding, M.C.U. Araújo, Flow-batch analysis, Trends Anal. Chem. 35 (2012) 39–49. [5] M.B. Lima, M. Insausti, C.E. Domini, M.F. Pistonesi, M.C.U. Araújo, B.S.F. Band, Automatized flow-batch method for fluorescent determination of free glycerol in biodiesel samples using on-line extraction, Talanta 89 (2012) 21–26. [6] S.S. Monte-Filho, M.B. Lima, S.I.E. Andrade, D.P. Harding, Y.N.M. Fagundes, S.R.B. Santos, S.G. Lemos, M.C.U. Araújo, Flow batch miniaturization, Talanta 86 (2011) 208–213 (418). [7] M.B. Lima, I.S. Barreto, S.I.E. Andrade, M.S.S. Neta, L.F. Almeida, M.C.U. Araújo, Photometric determination of phosphorus in mineralized biodiesel using amicro-flow-batch analyzer with solenoid micro-pumps, Talanta (in press), http://dx.doi.org/10.1016/j.talanta.2012.06.056. [8] A. Lopez-Molinero, D. Liñan, D. Sipiera, R. Falcon, Chemometric interpretation of digital image colorimetry. Application for titanium determination in plastics, Microchem. J. 96 (2010) 380–385. [9] Y.L. Shishkin, S.G. Dmitrienko, O.M. Medvedeva, S.A. Badakova, L.N. Pyatkova, Use of a scanner and digital image-processing software for the quantification of adsorbed substances, J. Anal. Chem. 59 (2004) 102–106. [10] A.W. Martinez, S.T. Phillips, E. Carrilho, S.W. Thomas, H. Sindi, G.M. Whitesides, Simple telemedicine for developing regions: camera phones and paper-based microfluidic devices for real-time, off-site diagnosis, Anal. Chem. 80 (2008) 3699–3707. [11] A.K. Ellerbee, S.T. Phillips, A.C. Siegel, K.A. Mirica, A.W. Martinez, P. Striehl, N. Jain, M. Prentiss, G.M. Whitesides, Quantifying colorimetric assays in paper-based microfluidic devices by measuring the transmission of light through paper, Anal. Chem. 81 (2009) 8447–8452. [12] N. Maleki, A. Safavi, F. Sedaghatpour, Single-step calibration, prediction and real samples data acquisition for artificial neural network using a CCD camera, Talanta 64 (2004) 830–835. [13] L. Byrne, J. Barker, G. Pennarun-Thomas, D. Diamond, Digital imaging as a detector for generic analytical measurements, Trends Anal. Chem. 19 (2000) 517–522. [14] E.N. Gaião, V.L. Martins, W.S. Lyra, L.F. Almeida, E.C. Silva, M.C.U. Araújo, Digital image-based titrations, Anal. Chim. Acta 570 (2006) 283–290. [15] S.K. Kohl, J.D. Landmark, D.F. Stickle, Demonstration of absorbance using digital color image analysis and colored solutions, J. Chem. Educ. 83 (2006) 644–646. [16] W. Wongwilai, S. Lapanantnoppakhun, S. Grudpan, K. Grudpan, Webcam camera as a detector for a simple lab-on-chip time based approach, Talanta 81 (2010) 1137–1141. [17] W.S. Lyra, V.B. Santos, A.G.G. Dionízio, V.L. Martins, L.F. Almeida, E.N. Gaião, P.H.G.D. Diniz, E.C. Silva, M.C.U. Araújo, Digital image-based flame emission spectrometry, Talanta 77 (2009) 1584–1589. [18] W.S. Lyra, F.A.C. Sanches, F.A.S. Cunha, P.H.G.D. Diniz, S.G. Lemos, E.C. Silva, M.C.U. Araújo, Indirect determination of sodium diclofenac, sodium dipyrone and calcium gluconate in injection drugs using digital image-based (webcam) flame emission spectrometric method, Anal. Methods 3 (2011) 1975–1980. [19] A.R. Tôrres, W.S. Lyra, S.I.E. Andrade, R.A.N. Andrade, E.C. Silva, M.C.U. Araújo, E.N. Gaião, A digital image-based method for determining of total acidity in red wines using acid–base titration without indicator, Talanta 84 (2011) 601–606. [20] S.I.E. Andrade, M.B. Lima, I.S. Barreto, W.S. Lyra, L.F. Almeida, M.C.U. Araújo, E.C. Silva, A digital image-based flow-batch analyzer for determining Al(III) and Cr(VI) in water, Microchem. J. (in press), http://dx.doi.org/10.1016/j.microc. 2012.03.029. [21] A.A. Bele, V.M. Jadhav, V.J. Kadam, Potential of tannins: a review, Asian J. Plant Sci. 9 (2010) 209–214. [22] H. McKinley, M. Jamieson, Handbook of Green Tea and Health Research, Nova Science Publishers, New York, 2009. [23] Y. Clement, Can green tea do that? A literature review of the clinical evidence, Prev. Med. 49 (2009) 83–87. [24] Official methods of analysis of tea, Jap. Tea Res. Station Bull. 6 (1970) 167–172. [25] Official methods of analysis of AOAC International, official method 952.03, in: 382 Section 26.1.37, 8th, AOAC International, Gaithersburg, 2005. [26] Y.T. Hung, P.C. Chen, R.L.C. Chen, T.J. Cheng, Sequential determination of tannin and total amino acid contents in tea for taste assessment by a fluorescent flow-injection analytical system, Food Chem. 118 (2010) 876–881. [27] R.L.C. Chen, C.H. Lin, C.Y. Chung, T.J. Cheng, Determination of tannin in green tea infusion by flow-injection analysis based on quenching the fluorescence of 3-aminophthalate, J. Agric. Food Chem. 53 (2005) 8443–8446. [28] Y.T. Hung, P.C. Chen, R.L.C. Chen, T.J. Cheng, Determining the levels of tannin in tea by amperometry of ferricyanide pre-reaction with a sample in a flow-injection system, Sens. Actuators B 130 (2008) 135–140. Author's personal copy M.B. Lima et al. / Microchemical Journal 106 (2013) 238–243 [29] T.J. Cheng, H.Y. Hsiao, C.Y. Chung, Determination of tannic acid after precipitation with bovine serum albumin by visible light scattering in a flow injection system, Microchim. Acta 169 (2010) 117–122. [30] M. Nakayama, N. Shigemune, T. Tsugukuni, H. Tokuda, T. Miyamoto, A simple and rapid turbidimetric method for determining catechins in beverages, Int. J. Food Sci. Technol. 45 (2010) 2071–2079. [31] M.B. Lima, S.I.E. Andrade, D.P. Harding, M.F. Pistonesi, B.S.F. Band, M.C.U. Araújo, Turbidimetric and photometric determination of total tannins in tea using a micro-flow-batch analyzer, Talanta 88 (2012) 717–723. [32] M.B. Lima, S.I.E. Andrade, I.S. Barreto, L.F. Almeida, E.C. Silva, M.C.U. Araujo, Quantitative spot test analysis of soluble tannin in green tea using a portable diffuse reflectometer, Anal. Methods (2012), http://dx.doi.org/10.1039/C2AY25361E. 243 [33] A.D. McNaught, W. Andrew, IUPAC Compendium of Chemical Terminology, 2nd ed. Royal Society of Chemistry, Cambridge, 1997. [34] E. Piccin, H.J. Vieira, O. Fatibello-Filho, Flow-injection turbidimetric determination of tannins in tea samples using copper(II)/acetate as precipitant reagent, Anal. Lett. 38 (2005) 511–522. [35] J.C.B. Fernandes, L.O.S. Ferreira, Manufacturing of miniature fluidic modules for lab-on-a-chip using UA photoresin from, flexographic platemaking process, J. Braz. Chem. Soc. 17 (2006) 643–647.