Flow injection colorimetric method using acidic ceric nitrate as reagent for determination of ethanol

Accepted Manuscript Title: Flow Injection Colorimetric Method Using Acidic Ceric Nitrate as Reagent for Determination of Ethanol Authors: Piyanut Pinyou, Napaporn Youngvises, Jaroon Jakmunee PII: DOI: Reference: S0039-9140(11)00131-7 doi:10.1016/j.talanta.2011.01.078 TAL 11892 To appear in: Talanta Received date: Revised date: Accepted date: 8-11-2010 29-1-2011 31-1-2011 Please cite this article as: P. Pinyou, N. Youngvises, J. Jakmunee, Flow Injection Colorimetric Method Using Acidic Ceric Nitrate as Reagent for Determination of Ethanol, Talanta (2010), doi:10.1016/j.talanta.2011.01.078 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. *Revised Manuscript Flow Injection Colorimetric Method Using Acidic Ceric Nitrate as Reagent for us cr ip t Determination of Ethanol a an Piyanut Pinyou a, Napaporn Youngvises b, Jaroon Jakmunee a,* Department of Chemistry, and Center of Excellence for Innovation in Chemistry, Faculty of Department of Chemistry, Faculty of Science and Technology, Thammasat University, ep t Pathum Thani 12121, Thailand ed b M Science, Chiang Mai University, Chiang Mai 50200, Thailand *Corresponding author at: Department of Chemistry, and Center of Excellence for Innovation Ac c in Chemistry, Faculty of Science, Chiang Mai University, Chiang Mai 50200, Thailand. Fax: +66 53 941 910. E-mail address: scijjkmn@chiangmai.ac.th (J. Jakmunee). ABSTRACT Ceric ammonium nitrate has been used for qualitative analysis of ethanol. It forms an intensely colored unstable complex with alcohol. In this work, a simple flow injection (FI) Page 1 of 30 colorimetric method was developed for the determination of ethanol, based on the reaction of ethanol with ceric ion in acidic medium to produce a red colored product having maximum absorption at 415 nm. Absorbance of this complex could be precisely measured in the FI system. A standard or sample solution was injected into a deionized water donor stream and t flowed to a gas diffusion unit, where the ethanol diffused through a gas permeable membrane us cr ip made of plumbing PTFE tape into an acceptor stream to react with ceric ammonium nitrate in nitric acid. Color intensity of the reddish product was monitored by a laboratory made LED based colorimeter and the signal was recorded on a computer as a peak. Peak height obtained was linearly proportional to the concentration of ethanol originally presented in the injected an solution in the range of 0.1-10.0 % v/v (r2 = 0.9993), with detection limit of 0.03 %v/v. With the use of gas diffusion membrane, most of the interferences could be eliminated. The M proposed method was successfully applied for determination of ethanol in some alcoholic ed beverages, validating by gas chromatographic method. Ac c ep t Keywords: Flow injection; Ethanol; Ceric ammonium nitrate; Gas diffusion; Beverages 1. Introduction Ethyl alcohol or ethanol is the most publicly known alcohol since it is contained in many products such as alcoholic beverages, antiseptic and fuel for vehicles. Determination of ethanol content is important for quality control of the products. Various methods have been used for quantification of ethanol such as densitometry, gas chromatography, high performance liquid chromatography, infrared spectroscopy, spectrophotometry and enzymatic biosensor. Although densitometry is recommended as a standard method for Page 2 of 30 quality control of beverage products, it concerns the tedious and slow distillation procedure [1]. Flow injection (FI) pervaporation system with in-line densiotometric detection was developed to increase the degree of automation [2]. Gas chromatography is approved as a standard method for ethanol determination in beverages and gasohol fuel [3]. t Chromatographic techniques lack portability and need long analysis time, although they us cr ip provide simultaneous determination of several alcohols. Biosensor has advantages in terms of selectivity and portability but life-time of the sensor is not long and fabrication of the sensor is rather complicated and expensive [4]. Infrared spectroscopy is also limited in selectivity. FI with Near IR detection is proposed in order to increase degrees of automation of the method an [5]. Spectrophotometric measurement is usually performed after separation of ethanol from M other matrices by distillation. Chemical oxidation of ethanol by dichromate followed by ed spectrophotometric detection of the produced Cr(III) or the remaining Cr(VI) is the commonly used chemistry. This detection principle is also widely adapted to flow based ep t techniques for determination of ethanol [6-11], which offer improved analytical performance such as being less tedious, having faster analysis, and having more automatic than batch wise method. However, dichromate is a carcinogenic substance, so it is of interest to reduce the Ac c amount needed, such as by using sequential injection (SI) [9], microfluidics [11] and microplate reader [12]. Enzymatic methods for determination of ethanol and glycerol in wines were introduced in FI [13] and SI[14] systems. These systems employed gas diffusion separation to increase selectivity. Electrochemical oxidation of ethanol on copper electrode under alkaline medium has been introduced in flow injection amperometric system [15]. However, this reaction is not very selective and the electrode is prone to fouling, which leads to low reproducibility. Since the reactions used either in spectrophotometric or electrochemical analysis are not selective, in-line separation of ethanol from sample matrices Page 3 of 30 is needed, which could be done by gas diffusion [8,10,13,15] or pervaporation [6,7] technique, with the use of gas permeable membrane or recently with membraneless gas diffusion device[8,10]. In this work, we developed a simple FI colorimetric system based on the reaction of us cr ip t ethanol and acidic ceric nitrate to produce a red colored complex of Ce(IV)-ethanol. Ce(IV) has been used as a colorimetric reagent for qualitative detection of alcohols for a long time [16]. The formation of 1:1 complex of Ce(IV)-alcohol was found, of which the change of color from yellow to red was observed [17]. Ce(IV) is present in the anionic form as a hexanitrato ceric species or other complexes with combination of H2O, OH- and NO3- as an ligands, with 12-fold coordination sphere of the Ce(IV). Alcohol replaces one of the Ce-O bonds of bidentate nitrate (or other monodentate ligands) leading to a complex composed of M five bidentate nitrate groups, one monodentate nitrate group and one alcohol molecule [18]. ed This complex is not stable, and the alcohol will be oxidized to aldehyde and then to carboxylic acid while Ce(IV) is reduced to a colorless Ce(III) [18]. Apart from alcohol, ep t Ce(IV) can form complexes with α-hydroxy acids such as lactic, malic and tartaric acids [19]. Despite this, the reagent is also not specific to ethanol but it is more selective than a strongly oxidizing agent, dichromate. Moreover, ceric ammonium nitrate is less hazardous than Ac c dichromate. In order to improve selectivity of the method, incorporation of a simple gas diffusion unit employing a PTFE plumbing tape as a membrane, the FI system was investigated for in-line separation of ethanol from the sample matrices. The proposed system provided good analytical performance and was successfully demonstrated for determination of ethanol in some alcoholic beverages. 2. Experimental Page 4 of 30 2.1. Chemicals and reagents All chemicals used were of analytical reagent grade and deionized water (obtained from a system of Milli-Q, Millipore, Sweden) was used throughout. Reagent solution containing 0.04 M ceric ammonium nitrate and 0.3 M nitric acid was prepared by dissolving 2.20 g of us cr ip t ceric ammonium nitrate (BDH, England) in water and making up to a volume of 100 mL in a volumetric flask. Working standard solution of ethanol (0.5-10.0 % v/v) was freshly prepared by diluting the 99.5 % v/v ethanol (Merck, Germany) with water. an 2.2. FI colorimetric system FI colorimetric system for determination of ethanol is shown in Fig.1. It consisted of a M peristaltic pump (Lachat, USA), a six port injection valve (Upchurch, USA), a planar gas diffusion unit [20], a lab-built colorimeter based on light emitting diode as a light source and ed light dependent resistor as a light sensor, equipped with a flow through cell of 10 mm path ep t length (Hellma, Germany), a data acquisition unit [20] and a personal computer. Mixing coils and tubing for assembling the system were a PTFE tube of 0.5 mm diameter, except pump tubing for propelling the solution is a Tygon tube (Fisher Scientific, USA). Ac c A home-made planar gas diffusion unit [20] was made of two acrylic plates (15 cm long, 4.8 cm wide and 1.0 cm thick), engraved for donor and acceptor channels (each 300 mm long, 1.5 mm wide, 0.75 mm deep). A PTFE membrane (teflon tape used for plumbing work) was placed in between the two plates to form channels for liquid to flow on each side. _____Fig.1______ Page 5 of 30 2.3. Determination of ethanol by FI colorimetric method Employing the FI set up as described above, a standard / sample solution (100 µL) was injected into a water carrier stream and flowed through mixing coil C1 to a gas diffusion unit, where the analyte diffused through a PTFE membrane into an acceptor stream of acidic ceric us cr ip t ammonium nitrate. Ethanol reacted with the reagent to produce a red Ce(IV)-ethanol complex which could be monitored for a color change by using a colorimeter with a blue LED as light source. A commercial spectrophotometer could be used by setting to measure absorbance at 415 nm. Output signal from the detector was recorded as FIA peak on a personal computer. Peak height was directly proportional to concentration of ethanol in the injected solution. A M was used for quantification of ethanol in sample. an calibration graph was constructed by plotting peak height versus ethanol concentration and it 2.4. Determination of ethanol by gas chromatographic method ed Gas chromatograph (HP Aligent 6890 GC model, HEWLETT PACKARD company) ep t The following GC condition was used: carrier gas was helium, flow rate was 1.2 mL min-1, injection volume was 1.00 µL, type of separation column was capillary column (Aligent 19091 N-113 HP – INNOWAX Polyethylene glycol), and detector was flame ionization. The Ac c method of internal standard calibration was utilized by adding n-Propanol as an internal standard to sample or standard solution of ethanol to obtain a concentration of 2 % v/v nPropanol. Ethanol standard solutions were prepared by dilution of the 99.5 % v/v ethanol to obtain working standard solutions in the concentration range of 0.1-4.0% v/v. All samples were diluted 20 times with DI water before analysis. Ethanol contents in samples were calculated from a calibration graph constructed from area ratio of ethanol to n-propanol and concentration of the ethanol. Page 6 of 30 2.5. Sample preparation Twenty three samples of commercially available alcoholic beverages were purchased from local stores in Chiang Mai, Thailand. According to the label, ethanol contents in t samples ranged from 5.0 - 40% v/v. Samples were properly diluted with DI water (10 folds us cr ip for distillated liquor, 20 folds for beer and 40 folds for wine) prior to injection into the FI system. an 3. Results and discussion M 3.1. Preliminary study of the reaction Absorption spectra of 0.02 M ceric ammonium nitrate (CAN) and 0.02 M CAN with ed 10% v/v (0.017 M) ethanol in 0.15 M nitric acid solutions recording versus water, and of the CAN-ethanol complex versus CAN reagent blank were recorded in the range of 350-700 nm ep t as depicted in Fig. 2. It was observed that the absorption of the complex occurred at a longer wavelength than that of the reagent. The maximum absorption wavelength of the complex Ac c versus reagent blank was observed at 415 nm. _____Fig.2______ The reaction is fast and is the first order reaction with respect to ceric and alcohol [21]. The rate equation is as follows: rate = k[Ce(IV)][alcohol], with k of about 0.75 L mol-1 for ethanol. Although the complex had high intensity of color, the complex is not stable. Color of the complex faded rapidly which could be observed with naked eyes. In contrast, the reagent itself is stable. Stability of the complex was investigated by monitoring the absorbance of Page 7 of 30 0.02 M CAN-1%v/v ethanol solution at 415 nm versus reagent blank after the solutions were mixed together. It was found that the absorbance decreased dramatically and become zero within about 10 min (Fig. 3). The unstability of this complex might limit the use of this reagent for determination of ethanol in batch wise method but there was no problem in FIA t because of a short and precisely controlled time interval between an injection of sample and us cr ip the detection of the resulting product. _____Fig.3______ 3.2. Optimization of FI colorimetric method an Employing the FI system as shown in Fig. 1, the optimum conditions for determination of ethanol were investigated. The following preliminary conditions were used: concentration M of ceric ammonium nitrate of 0.05 M, flow rates of carrier and reagent solution of 1.0 mL min-1 each, sample volume of 100 µL, length of coils C1, C2, C3 and C4 of 50, 100, 50 and ep t 3.2.1 Effect of nitric acid ed 50 cm , respectively and blue LED as a light source. Effect of nitric acid concentration in range of 0.0-0.6 M was studied by injecting a series Ac c of standard ethanol (0.1 – 10.0 % w/v) into the system while other parameters were kept constant as described above. Plotting peak height obtained versus concentration of ethanol yielded linear calibration graphs with r2 higher than 0.999. It was found that sensitivity (slope of the calibration graph) increased with the increase of nitric acid concentration and leveled off at concentration higher than 0.2 M as shown in Fig. 4. The high cationic charge and smaller ionic size of the ceric ion may make ceric salts much more hydrolyzed in aqueous solution than other trivalent lanthanides, leading to the formation of hydroxide ceric complex (Ce Lm (OH-)(n-1)+) and the releasing of hydrogen ion [22]. In highly acidic solution the hydrolysis is suppressed and aquo complex (Ce Lm (H2O)n+) should be predominant. Ethanol Page 8 of 30 should replace H2O ligand easier than the hydroxide ligand, hence the red colored Ce(IV)Ethanol complex was easily formed in acidic solution. Nitric acid concentration of 0.3 M was selected for further studies. _____Fig.4______ us cr ip t 3.2.2 Effect of ceric ammonium nitrate concentration Effect of concentration of ceric ammonium nitrate on sensitivity was investigated as similar to above. Calibration equations of y = 0.067 x +0.001, y= 0.072x + 0.008, y= 0.073x + 0.009 and y=0.071x + 0.022 were obtained for ceric nitrate concentration of 0.02, 0.03, an 0.04 and 0.05 M, respectively. In all case r2 of the calibration graphs were higher than 0.998. As expected, too low of a concentration of ceric ammonium nitrate gave low sensitivity M because of the slow reaction rate and the degradation of the complexed product. Ceric ammonium nitrate of 0.04 M was chosen as it provided high sensitivity and low background ed signal. ep t 3.2.3 Effect of mixing coil length Effect of length of mixing coil C1 was investigated over the range of 0-50 cm. It was Ac c found that sensitivity slightly decreased with the increase of mixing coil C1 length, due to the increased dispersion of the injected solution. However, with mixing coil length of 50 cm , a more reproducible peak profile was obtained than without mixing coil, so it was selected for further studies. Effect of the length of mixing coils C2 and C3 was studied in range of 0-100 cm. The same length of coils C2 and C3 was considered in order to maintain the same pressure on both sides of the gas diffusion membrane, thus preventing deformation or breakage of the membrane. With increasing length of the coils, a slight increase in sensitivity was observed. A length of 75 cm each was chosen for coils C2 and C3 since having longer coils may cause Page 9 of 30 breakage of the membrane. Coil C4 was incorporated in the system in order to increase pressure in the detection stream, preventing the evolution of bubbles in the line. According to the results, the length of this coil did not have affect on the sensitivity of the method. Coil C4 of 50 cm length was selected as it could maintain pressure in the line and provide us cr ip t reproducible peak profiles. 3.2.4 Effect of sample volume Effect of sample volume was investigated by injecting a series of ethanol standard solutions in the concentration range of 0.0-10.0 %v/v, with different volumes of an injection loop. an Sensitivity (slope of calibration graph) increased linearly with the injection volume and leveled off at about 150 µL volume. However, at higher volume a narrow linear range of 3.2.5 Reagent stability ed up to 10 %v/v ethanol was chosen. M calibration graphs were obtained. Thus sample volume of 100 µL, which gave a linear range ep t Stability of the reagent was examined by using the same reagent solution for construction of calibration graphs in the concentration range of 0.0-10.0 %v/v ethanol at different times. Ac c Slopes of the calibration graphs of 0.078, 0.077, 0.075, 0.075, 0.075, 0.076 and 0.076 V/%v/v for the age of reagent of 0, 3, 6, 9, 12, 15 and 18 h, respectively were obtained, with r2 > 0.999 for all the calibration graphs. This indicated that the reagent was stable within at least 18 h. Although the reagent might be stable up to 5 days based on observation of its color, it was daily prepared in all of our experiments. 3.3. Optimization of gas diffusion system Five commercially available PTFE plumbing tapes with two different thicknesses, 0.75 (tape #1-#3) and 1.00 (tape #4-#5) mm, were tested to be employed as a gas diffusion Page 10 of 30 membrane. Morphology of the membrane was examined by scanning electron microscope as illustrated in the SEM images of the membranes in Fig. 5. The thicker membrane has smaller pore size and the distribution of the pores on the membrane was more regular. Pore sizes of the membrane were in the range of about 0.3x1 to 5x25 µm. The ratio of area of pore per total t area of the membrane was observed as follows: #4 Oishi < #5 Joint < #1 At Indy < #2 us cr ip Protape < #3 Blue. As expected, PTFE tape of smaller thickness and larger pore size provided higher diffusion efficiency of ethanol, hence resulting in higher sensitivity as shown in Fig. 6. However, too thin and large pore membrane was easily broken, so the membrane #2 which provided medium sensitivity was selected. an _____Fig.5-6______ M Flow rates of donor and acceptor streams were identically maintained in order to avoid the building up of pressure difference between two sides of the membrane that may cause ed membrane deformation or breakage. Effect of flow rate in the range of 0.5-2.0 mL min-1 on sensitivity of the system with using various gas permeation membranes was investigated as ep t shown in Fig. 6. As expected, an increase of flow rate resulted in a decrease in sensitivity because of the reduction in gas diffusion efficiency at high flow rate. As could be seen by the Ac c trend from the graph, one could select suitable flow rate to obtain sensitivity and sample throughput required for their application. Flow rate of 1.0 mL min-1 with membrane #2 was selected as it provided high sensitivity and adequate sample throughput for our application. Geometry of gas diffusion unit also plays a role on diffusion efficiency [23]. Long, wide and shallow channel on the GDU provided high diffusion efficiency because it had high diffusion area per volume. However, too long of a channel caused high dispersion of the injected solution, thus reducing sensitivity and sample throughput. In this work, a gas diffusion unit 30.0 cm long, 1.5 mm wide and 0.75 mm deep was utilized and concurrent Page 11 of 30 flow of the donor and acceptor streams was employed. Under the above selected conditions, the diffusion efficiency was found to be about 15%. Temperature also affected the diffusion of ethanol through the membrane. According to the previous studies, high sensitivity was achieved with the increase of temperature [8, 24- t 26]. However, using of relatively high temperature caused evolving of bubbles in the system us cr ip and enough sensitivity was obtained with the operation at room temperature, so that the room temperature was recommended in literature [8]. According to our observation, with the experiment performed at room temperature of about 25 oC in an air conditioned room, the variation of less than 1% in concentration found for 5%v/v ethanol was observed. Slopes of an the calibration graphs obtained during 18 h operation of the system were not significantly different as described in section 3.2.5. A thermostat control unit might be needed if the M system was operated at a place with high variation of temperature. ed 3.4. Analytical characteristics Using the FI system as shown in Fig. 1 and a set of selected conditions: reagent solution ep t consisting of 0.04 M ceric ammonium nitrate and 0.3 M nitric acid as an acceptor stream, water carrier as a donor stream, flow rate of each stream of 1.0 mL min-1, sample volume of Ac c 100 µL, the length of coils C1, C2, C3 and C4 of 50, 75, 75 and 50 cm, respectively, and blue LED (max = 470 nm) as a light source, analytical characteristics of the system were evaluated. A linear calibration graph (y =0.0873x-0.0093, r2= 0.9993) for the concentration range of 0.1-10.0 % v/v % v/v ethanol was obtained as shown with the FIA peaks in Fig. 7. Detection limit calculated from three times standard deviation of blank/slope of the calibration graph [26] was 0.03 % v/v. Relative standard deviations for 11 replicate injections of 1.0, 5.0 and 10.0 % v/v ethanol were 1.2, 0.5 and 0.3, respectively. Sample throughput of 20 h-1 was achieved and each injection consumed 3 mL each of reagent and carrier solutions. Page 12 of 30 Sensitivity of the system using ceric ammonium nitrate as reagent was comparable to those using dichromate as reagent as summarized in Table 1. In addition, ceric ammonium nitrate is more environmentally friendly, since dichromate is a well known carcinogenic substance. _____Fig.7______ us cr ip t _____Table 1______ 3.5. Interference study It has been reported that the ceric ammonium nitrate reagent could react with other alcohols and α-hydroxy acids. However, by incorporating a gas diffusion unit (with an hydrophobic membrane) to the FI system, the hydrophilic substances could be retained on the donor side of the membrane so they did not react with the reagent. Some substances which M can affect the diffusion of ethanol through the membrane were also potential interferences. Interference was investigated by adding different concentrations of the potential interfering ed substances into 2% v/v ethanol standard solution. The solutions were injected into the FI systems with and without a gas diffusion unit. Table 2 shows the tolerance limit (maximum ep t concentration of each substance that did not interfere or did not cause a change in peak height of 2% v/v ethanol more than 5%) of some substances. It was found that methanol and Ac c propanol seriously interfere at low concentration (0.1 %v/v), either without or with the use of GDU. However, beverage samples contained very low concentration of these alcohols so they should not significantly affect the ethanol detection. Other long chain alcohols are not easily dissolved in water, so they should not interfere. Sugars showed negative interferences in both systems, except sucrose in the case of FI system without GDU. The presence of sugar at high concentration may lead to interference due to change of refractive index in the case of the system without GDU or decreasing of diffusion efficiency in the FI system with GDU because of the increase of boiling point of the solution (colligative property). It was clearly Page 13 of 30 observed that the GDU helped reduce interferences from tartaric acid, glycerol and tannic acid, which gave positive interference in the system without GDU at low concentration. The membrane helped prevent diffusion of ionic species and large organic molecules to the acceptor side. Some substances did not interfere even at the maximum concentrations tried, us cr ip _____Table 2______ t which were the concentrations higher than those expected to be found in real samples. However, on the application of the developed method to real samples of beer and wine, it was found that there were some substances in sample which can adsorb on the membrane, prevent the transfer of ethanol through the membrane and hence cause reduction in M an sensitivity. Fortunately, this problem could be solved by extensive dilution of the samples. 3.6. Application to real samples ed The proposed FI system was applied for determination of ethanol in some beverages purchased from a local supermarket. The samples were prepared as described in section 2.5 ep t before injecting the solutions into the FI system. The same solutions were also analyzed by gas chromatographic method [3] as described in section 2.4 for comparison. Contents of Ac c ethanol obtained from both of the methods were in good correlation (EtOHFIA= 1.029EtOHGC – 0.509, r2= 0.9971) as summarized in Table 3. According to t-test at 95% confidence level [23], both the results are not significantly different (tcalculated = 0.57, ttable= 2.07). _____Table 3______ By spiking standard ethanol into the samples #1 to #10, percentage recoveries were obtained in the range of 98.2-102.7%. It was found that the matrices of beer and wine samples affect a lot on the transfer of ethanol through the membrane. Standard addition method was firstly tried, by spiking ethanol standard solution into sample of different folds Page 14 of 30 dilution (2, 5 and 10 folds), to obtain added concentration of ethanol of 0.5, 1.0 and 1.5% v/v. However, this method cannot provide accurate and precise results. Moreover, sensitivity of the method decreased dramatically after several samples were injected. Fortunately, this problem could be solved by injecting the diluted samples (20 fold dilution for beer and 40 t fold dilution for wine) and employing a lower range calibration graph (0.1-0.5% v/v ethanol). us cr ip The proposed method provided advantages such as low chemical consumption and fast analysis time. an 4. Conclusions M The FI method for determination of ethanol was developed based on the reaction between ethanol and ceric ion to form a reddish colored product which has a maximum ed absorption at 415 nm. The detection by using blue LED as a light source was proposed. The reagent had higher selectivity and negligible toxicity as compared to dichromate, which was ep t commonly used as a reagent for determination of ethanol. By incorporating a gas diffusion unit into the FI system, most of the interferences could be eliminated. This reaction also Ac c provides good sensitivity. Despite using a simple colorimeter as a detector, a linear calibration graph in the range of 0.1-10.0% v/v ethanol, with detection limit of 0.03% v/v was obtained. Sensitivity of the method could be increased by employing lower carrier flow rate, which provided higher mass transfer efficiency. On the other hand, if high sensitivity is not a priority, then high flow rate could be used to increase sample throughput of the method. The proposed method was successfully applied for determination of ethanol in some alcoholic beverages. Further development of the system to be used for the monitoring of ethanol in a fermentation tank is being carried out in our laboratory. Page 15 of 30 Acknowledgements We thank the Center for Innovation in Chemistry: Postgraduate Education and Research Program in Chemistry (PERCH-CIC), the Commission on Higher Education (CHE) and the us cr ip t Thailand Research Fund (TRF) for financial support. Science Achievement Scholarship of Thailand (SAST) is gratefully acknowledged for providing scholarship to PP. an References M [1] W. Horwitz, Official Methods of Anaysis of AOAC International, 17th ed., AOAC International, Gaithersburg, (2000) vol.2, chap. 26, pp.2-3. J. Gonzalez-Rodrıguez, P. Perez-Juan, M.D. Luque de Castro, Talanta 59 (2003) 691- ed [2] [3] ep t 696. W. Horwitz, Official Methods of Anaysis of AOAC International, 17th ed., AOAC Ac c International, Gaithersburg, (2000) vol.2, chap. 27, pp.4-5. [4] A. Guzmán-Vázquez de Prada, N. Peña, M. L. Mena, A. J. Reviejo, J. M. Pingarrón, Biosens. Bioelectron. 18 (2003) 1279-1288. [5] P. Tipparat, S. Lapanantnoppakhun, J. Jakmunee, K. Grudpan, Talanta 53 (2001) 1199-1204. [6] E. Mataix, M.D. Luque de Castro, Talanta 51 (2000) 489-496. [7] I.L. Mattos, E.A.G. Zagatto, Anal. Sci. 15 (1999) 63-66. Page 16 of 30 [8] N. Choengchan, T. Mantim, P. Wilairat, P.K. Dasgupta, S. Motomizu, D. Nacapricha, Anal. Chim. Acta 579 (2006) 33-37. [9] P.J. Fletcher, J.F. van Staden, Anal. Chim. Acta 499 (2003) 123-128. t [10] S. Muncharoen, J. Sitanurak, W. Tiyapongpattana, N. Choengchan, N. us cr ip Ratanawimarnwong, S. Motomizu, P. Wilairat, D. Nacapricha, Microchim. Acta 164 (2009) 203-210. [11] L. Lei, I. L. Mattos, Y. Chen, Microelectron. Eng. 85 (2008) 1318-1320. [12] H. B. Seo, H. J. Kim, O. K. Lee, J. H. Ha, H. Y. Lee, K. H. Jung, J. Ind. Microbiol. an Biotechnol. 36 (2009) 285-292. M [13] A.O.S.S. Rangel, I.V. Tóth, Anal. Chim. Acta 416 (2000) 205-210. [14] M.A. Segundo, A.O.S.S. Rangel, Anal. Chim. Acta 458 (2002) 131-138. ed [15] T.R.L.C. Paixão, D. Corbo, M.Bertotti, Anal. Chim. Acta 472 (2002) 123-131. ep t [16] F.R. Duke, G.F. Smith, Ind. Eng. Chem. Anal. Ed. 12 (1940) 201-203. [17] L.B. Young, W.S. Trahanovsky, J. Amer. Chem. Soc. 91 (1969) 5060-5068. Ac c [18] V. Briois, D. Lutzenkirchen-Hecht, F. Villain, E. Fonda, S. Belin, B. Griesebock, R. Frahm, J. Phys. Chem. A 109 (2005) 320-329. [19] N. Datt, R.R. Nagori, R.N. Mehrotra, Can. J. Chem. 64 (1986) 19-23. [20] J. Junsomboon, J. Jakmunee, Anal. Chim. Acta 627 (2008) 232-238. [21] M.P Doyle, J. Chem. Educ. 51 (1974) 131-132. [22] Cerium: A guide to its role in chemical technology, Molycorp, Inc., CA, USA. Page 17 of 30 [23] S.M. Oliveira, T.I.M.S. Lopes, I.V. Toth, A.O.S.S. Rangel, Anal. Chim. Acta 600 (2007) 29-34. [24] I. L. Mattos, R. P. Sartini, E. A. G. Zagatto, B. F. Reis, M. F. Giné, Anal. Sci. 14 (1998) 1005-1008. us cr ip t [25] H. Ohura, T. Imato, Y. Asano, S. Yamasaki, N. Ishibashi, Anal. Sci. 6 (1990) 541545. [26] S. Vicente, E. A. G. Zagatto, P. C. A. G. Pinto, M. L. M. E. S. Saraiva, J. L. F. C. Lima , E. P. Borges, An. Acad. Bras. Ciênc. 78 (2006) 23-29. Ac c ep t ed M an [27] G. D. Christian, Analytical Chemistry; 6th ed., Wiley: New York, USA, 2004. Page 18 of 30 Captions for Figures Fig. 1. FI manifold of flow injection system for determination of ethanol; P = peristaltic pump, I = injection valve, S = standard/sample, GDU = gas diffusion unit, W = waste, C1-C4 us cr ip t = mixing coils, D = colorimeter, ADC = analog to digital converter unit and PC = personal computer. Fig. 2. Absorption spectra of (a) acidic ceric reagent (0.02 M CAN in 0.15 M HNO3), (b) Ce(IV)-ethanol complex (0.02 M CAN and 1%v/v (0.017 M) ethanol in 0.15 M HNO3) with M an respect to water, and (c) solution of (b) with respect to solution of (a). Fig. 3. Absorbance-time profile of 0.02 M CAN-1.0 % v/v ethanol in 0.15 M HNO3 solution. ep t ed Absorbance at 415 nm was measured versus reagent blank. Fig. 4. Effect of nitric acid concentration on sensitivity of the method. (condition: 0.04 M CAN, flow rate 1.0 mL min-1, sample volume 100 μL, length of coils, C1 50 cm, C2 100 cm, Ac c C3 50 cm, C4 50 cm). Fig. 5. SEM images of various PTFE membranes (plumbing tapes); (a) #1 At Indy, (b) #2 Protape, (c) #3 Blue, (d) #4 Oishi and (e) #5 Joint; Images were recorded by using scanning electron microscope (JSM-5910LV, JEOL, Japan), accelerating voltage: 15 kV, magnification: 750x, signal: secondary electron images. Page 19 of 30 Fig. 6. Effect of flow rate on sensitivity of the method, using different membranes as described in Fig. 5. (condition: 0.04 M CAN, 0.3 M HNO3, sample volume 100 μL, length of Ac c ep t ed M an us cr ip Fig. 7. FIAgram of ethanol standard solutions and some samples. t coils, C1 50 cm, C2 75 cm, C3 75 cm, C4 50 cm). Page 20 of 30 Ac ce pt ed M an us cr i Figure 1 Click here to download high resolution image Page 21 of 30 Ac ce pt ed M an us cr i Figure 2 Click here to download high resolution image Page 22 of 30 Ac ce pt ed M an us cr i Figure 3 Click here to download high resolution image Page 23 of 30 Ac ce pt ed M an us cr i Figure 4 Click here to download high resolution image Page 24 of 30 Ac ce pt ed M an us cr ip t Figure 5 Click here to download high resolution image Page 25 of 30 Ac ce pt ed M an us cr i Figure 6 Click here to download high resolution image Page 26 of 30 Ac ce pt ed M an us cr i Figure 7 Click here to download high resolution image Page 27 of 30 us cr ip t Table 1 Table 1 Analytical characteristics of some spectrophotometric/colorimetric methods for the determination of ethanol FI Microfluidic in-line membraneless gas diffusion none off-line solvent extraction FI permeation through concentric silicon tubular membrane in-line gas diffusion FI FI on-line gas diffusion membrane probe FI in-line gas diffusion Linear range (% v/v) Detection limit (% v/v) Precision (%R.S.D.) Sample throughput (h-1) Ref. 600 Wines 1-20 0.5 3 6 [6] 600 Molasses fermentation Distilled liquors Gasohol fuel 1-10 - 1.5 20 [7] up to 6 0.09 <1 19 [9] 3-80 0.9 1-4.9 26 [10] Distilled spirits and wines up to 6 - - - [11] 600 590 600 100 g/L K2Cr2O7 in 5M H2SO4; detection performed in 96 well plate) 0.15M K2Cr2O7 in 6M H2SO4 595 Yeast culture broth up to 8 ~0.1 - 192 [12] 600 Distilled spirits and wines 1-20 0.5 3.7 20 [24] Reduction of excess K2Cr2O7 (0.1 M) in 8M H2SO4 with ferrous ion and detection of ferric ion formed by potentiometry 0.3M K2Cr2O7 in 4M H2SO4 - Distilled spirits, beers and wines 5-40 - 0.8 25 [25] 600 Distilled spirits and red wines up to 50 - <2 30 [26] blue LED (~470 nm) Distilled spirits, beers and wines 0.1-10 0.03 < 1.3 20 This work Ac ce p Batch microplate reader 0.15M K2Cr2O7 in 6M H2SO4; fiber optic detection Sample an SI 16 g/L K2Cr2O7 in 8M H2SO4 17 g/L K2Cr2O7 in 6M H2SO4 0.2M K2Cr2O7 in 4M H2SO4 0.2M K2Cr2O7 in 4M H2SO4 Detection wavelength (nm) M FI in-line pervaporation in-line pervaporation none Reagent d FI Separation te System 0.04M ceric ammonium nitrate in 0.3M HNO3 Page 28 of 30 Table 2 Table 2 Tolerance limits of some substances Substance Unit of concentration Tolerance limit FI with GDU FI without GDU %v/v 0.1, + 0.1, + Propanol %v/v 0.1, + 0.1, + Glucose mg L-1 1000, - 1000, - Fructose mg L-1 1000, - 100, - Sucrose mg L-1 1700, - 1700, + Acetic acid mg L-1 500* 500* Citric acid mg L-1 200, - 100, - Tartaric acid mg L-1 100, - 50, + Tannic acid mg L-1 1000* 50, + Glycerol mg L-1 Acetaldehyde mg L-1 Ethyl acetate mg L-1 an us cr ip t Methanol 5, + 400* 400* 1000* 1000* M 4000, - Ac c ep t ed + positive interference, - negative interference, * maximum concentration tested Page 29 of 30 Table 3 Table 3 Ethanol contents in some alcoholic beverages, determined by the proposed FIA and gas chromatographic methods, and labeled values t 0.0 -3.8 -0.6 3.4 1.7 1.3 1.8 4.4 4.5 1.9 0.0 -1.8 -4.1 -6.4 -13.4 -5.5 13.4 2.1 -2.2 -8.0 -7.0 -7.8 0.0 Ac ce pt e d M 1. Blended spirit 1 2. Blended spirit 2 3. Blended spirit 3 4. Scotch whisky 1 5. Scotch whisky 2 6. White spirit 7. Gin 1 8. Gin 2 9. Vodka 10. Rum 11. Red wine 1 12. Red wine 2 13. Red wine 3 14. White wine 1 15. White wine 16. White wine 17. Lynchee wine 18. Beer 1 19. Beer 2 20. Beer 3 21. Beer 4 22. Beer 5 23. Black beer * result from duplicate injections ** result from triplicate injections *** % different = 100*(FIA – GC)/GC %different** us cr ip Ethanol content (% v/v) Label GC* FIA* 35 34.3+0.1 34.3+0.6 35 34.7+0.1 33.4+0.3 40 40.1+0.1 39.9+0.3 40 38.4+0.1 39.8+0.6 40 36.0+0.1 36.6+0.2 35 30.8+0.1 31.2+0.4 40 38.5+0.1 39.2+0.2 40 37.2+0.1 38.9+0.2 40 38.3+0.1 40.1+0.8 40 38.0+0.1 38.7+0.2 13.0 13.0+0.1 13.0+0.1 12.5 11.1+0.1 10.9+0.2 12.5 12.3+0.1 11.8+0.2 11.0 10.9+0.1 10.2+0.2 11.5 11.9+0.1 10.3+0.2 13.0 12.7+0.1 12.0+0.1 12.5 11.9+0.1 13.5+0.3 5.0 4.7+0.1 4.8+0.1 5.0 4.6+0.1 4.5+0.1 5.0 5.0+0.1 4.6+0.1 6.4 5.7+0.1 5.3+0.1 5.4 5.1+0.1 4.7+0.1 5.6 4.9+0.1 4.9+0.1 an Sample 1 Page 30 of 30