Anal. Chem.

XXXX, xxx, 000000

Development of a Low-Cost Optical Sensor for Cupric Reducing Antioxidant Capacity Measurement of Food Extracts
zyu Mustafa Bener, Mustafa O rek, Kubilay Gu c lu , and Res at Apak* Department of Chemistry, Faculty of Engineering, Istanbul University, Avcilar 34320, Istanbul, Turkey A low-cost optical sensor using an immobilized chromogenic redox reagent was devised for measuring the total antioxidant level in a liquid sample without requiring sample pretreatment. The reagent, copper(II)-neocuproine (Cu(II)-Nc) complex, was immobilized onto a cationexchanger lm of Naon, and the absorbance changes associated with the formation of the highly colored Cu(I)-Nc chelate as a result of reaction with antioxidants was measured at 450 nm. The sensor gave a linear response over a wide concentration range of standard antioxidant compounds. The trolox equivalent antioxidant capacity (TEAC) values of various antioxidants reported in this work using the optical sensor-based cupric reducing antioxidant capacity (CUPRAC) assay were comparable to those of the standard solution-based CUPRAC assay, showing that the immobilized Cu(II)-Nc reagent retained its reactivity toward antioxidants. Common food ingredients like oxalate, citrate, fruit acids, and reducing sugars did not interfere with the proposed sensing method. This assay was validated through linearity, additivity, precision, and recovery, demonstrating that the assay is reliable and robust. The developed optical sensor was used to screen total antioxidant capacity (TAC) of some commercial fruit juices without preliminary treatment and showed a promising potential for the preparation of antioxidant inventories of a wide range of food plants.
The accumulation of reactive oxygen species (ROS) in the organism, unless counterbalanced by antioxidants taken in through diet, may cause oxidative damage to DNA and cellular membranes under oxidative stress conditions, eventually giving rise to certain human diseases, especially cardiovascular disease and some types of cancer.1 In this context, the measurement of antioxidant capacity of food and biological samples through development of selective and sensitive new techniques has recently gained importance. Various antioxidant assay methods, extensively spectroscopic techniques, applied to solutions have been developed to determine antioxidant capacity/activity in a wide range of matrixes such as
* To whom correspondence should be addressed. Fax: +90 212 473 7180. E-mail: rapak@istanbul.edu.tr. (1) Gutteridge, J. M. C.; Halliwell, B. Antioxidants in nutrition, health, and disease; Oxford University Press: Oxford; New York, 1994. 10.1021/ac100646k XXXX American Chemical Society

biological uids, food, and plant extracts.2 These methods may be broadly classied as electron transfer (ET)-based assays and hydrogen atom transfer (HAT)-based assays,2 though in some cases, these two mechanisms may not be differentiated with distinct boundaries. In fact, most nonenzymatic antioxidant activity (e.g., scavenging of free radicals, inhibition of lipid peroxidation, etc.) is mediated by redox reactions. HAT-based assays measure the capability of an antioxidant to quench free radicals (generally peroxyl radicals) by H-atom donation. Since both the uorescent probe and antioxidants react with peroxyl radical (ROO) in HATbased assays, antioxidant activity can be determined from competition kinetics by recording the uorescence decay curve of the probe in the absence and presence of antioxidants and integrating the area under these curves to take the difference.2 HAT-based assays basically include oxygen radical absorbance capacity (ORAC)3 and total peroxyl radical-trapping antioxidant parameter (TRAP) assays4 using R-phycoerythrin as the uorescent probe. In most ET-based assays, the antioxidant action is simulated with a suitable redox-potential probe, i.e., the antioxidants react with a uorescent or colored probe (oxidizing agent) instead of peroxyl radicals. Spectrophotometric ET-based assays measure the capacity of an antioxidant in the reduction of a chromogenic oxidant, which changes color when reduced. The degree of color change (either an increase or decrease of absorbance at a given wavelength) is correlated to the concentration of antioxidants in the sample. ABTS: 2,2azinobis(3-ethylbenzothiazoline-6-sulfonic acid)/TEAC (troloxequivalent antioxidant capacity)5 and 2,2-diphenyl-1-picryl hydrazyl6 (DPPH) are decolorization assays, whereas in Folin total phenolics,7 FRAP (ferric reducing antioxidant power)8 and CUPRAC (cupric reducing antioxidant capacity)9 assays, there is an increase in absorbance at a prespecied wavelength as the antioxidant reacts with the chromogenic reagent (i.e., in
(2) Prior, R. L.; Wu, X.; Schaich, K. J. Agric. Food Chem. 2005, 53, 4290 4302. (3) Cao, G.; Verdon, C. P.; Wu, A. H. B.; Wang, H.; Prior, R. L. Clin. Chem. 1995, 41, 17381744. (4) Wayner, D. D. M.; Burton, G. W.; Ingold, K. U.; Locke, S. FEBS Lett. 1985, 187, 3337. (5) Miller, N. J.; Rice-Evans, C. A.; Davies, M. J.; Gopinathan, V.; Milner, A. Clin. Sci. 1993, 84, 407412. (6) Sanchez-Moreno, C.; Larrauri, J. A.; Saura-Calixto, F. J. Sci. Food Agric. 1998, 76, 270276. (7) Singleton, V. L.; Orthofer, R.; Lamuela-Raventos, R. M. Methods Enzymol. 1999, 299, 152178. (8) Benzie, I. F. F.; Strain, J. J. Anal. Biochem. 1996, 239, 7076. zyurek, M.; Karademir, S. E. J. Agric. Food Chem. (9) Apak, R.; Gu c lu , K.; O 2004, 52, 79707981.

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX

the latter two methods, the lower oxidation states of iron and copper, namely Fe(II) and Cu(I), respectively, emerge as a result of the redox reaction with antioxidants form chargetransfer complexes with the ligands). These assays generally set a xed time for the concerned redox reaction and measure thermodynamic conversion (i.e., reduction of the colored species) during that period. Among the ET-based methods, the CUPRAC method of antioxidant measurement, introduced by our research group to world literature,9 is based on the absorbance measurement of the CUPRAC chromophore, Cu(I)-Nc chelate, formed as a result of the redox reaction of antioxidants with the CUPRAC reagent, Cu(II)-Nc, where absorbance is recorded at the maximal light absorption wavelength of 450 nm. The CUPRAC method of total antioxidant capacity (TAC) assay has been successfully applied to the quantication of antioxidants in food plants and human serum.10 Other scavenging assays for reactive oxygen species (ROS) performed with modied CUPRAC methods were determination of OH scavenging rate constants of certain water-soluble compounds11 and of polyphenolics12 and estimation of H2O2 scavenging activities of polyphenolics.13 Simultaneous determination of lipophilic and hydrophilic antioxidants in the same solvent medium utilizing their inclusion complexes formed with 2% methyl- -cyclodextrin14 and measurement of xanthine oxidase inhibition activities of polyphenolics15 were also performed with modied CUPRAC methods. In conclusion, the CUPRAC methodology is evolving into an antioxidant measurement package in biochemistry and food chemistry comprising many assays, and the results are in accordance with those of independent reference methods, having distinct advantages over certain established methods in regard to simplicity, versatility, cost-effectiveness, robustness, and reagent stability. Optical sensors employ optical transduction techniques (absorption, luminesence, or reectivity) to yield analyte information. A type of these sensors, reagent mediated sensors, are particularly suited to rapid and low-cost screening applications with high sensitivity and selectivity and, at the same time, can be provided in the form of inexpensive test kits and strips.15 Many optical sensors utilize redox chromogenic reagents immobilized in suitable polymeric membranes. Naon, a peruorosulfonate ion exchange membrane having R-{-O-CF2-CF(CF3)-}x-O(CF2)2SO3H functional groups, comprises hydrophilic columns composed of anionic sulfonate groups providing ionic conductivity and cation selectivity, hydrophobic columns composed of main chains, and interlayer regions between hydrophilic and hydrophobic columns.16,17 It is thermally stable, mechanically strong, and highly transparent in the UV-vis-NIR (near-infrared)
zyurek, M.; Karademir, S. E.; Altun, M. Free Radical (10) Apak, R.; Gu c lu , K.; O Res. 2005, 39, 949961. zyu (11) Bektas og lu, B.; C elik, S. E.; O rek, M.; Gu c lu , K.; Apak, R. Biochem. Biophys. Res. Commun. 2006, 345, 11941200. zyu (12) O rek, M.; Bektas og lu, B.; Gu c lu , K.; Apak, R. Anal. Chim. Acta 2008, 616, 196206. zyu (13) O rek, M.; Bektas og lu, B.; Gu c lu , K.; Gu ngo r, N.; Apak, R. J. Food Comp. Anal., in press, DOI: 10.1016/j.jfca.200909.005, (2010). zyu (14) O rek, M.; Bektas og lu, B.; Gu c lu , K.; Gu ngo r, N.; Apak, R. Anal. Chim. zyu Acta 2008, 630, 2839. O rek, M.; Bektas og lu, B.; Gu c lu , K.; Apak, R. Anal. Chim. Acta 2009, 636, 4250. (15) Wolfbeis, O. S. Fiber Optic Chemical Sensors and Biosensors; CRC Press: Boca Raton, FL, 1991. (16) Silva, V.; Ruffmann, B.; Vetter, S.; Boaventura, M.; Mendes, A. M.; Madeira, L. M.; Nunes, S. P. Electrochim. Acta 2006, 51, 36993706.

region so as to be chosen as polymeric support for a redox-active spectroscopic sensor. In more recent years, the potential use of Naon as a sensor (energy transfer, pH, oxygen, and metal ions sensor) in various applications has been explored.18 There is very limited study about the usage of optical sensors for molecular spectroscopic TAC assays. Steinberg and Milardovic 19 developed an optical sensor based on immobilized chromogenic radicals, i.e., DPPH (2,2-diphenyl-1-picryl hydrazyl) and galvinoxyl radical (GV), making use of plasticized poly(vinyl chloride) (PVC) matrixes for reacting standard antioxidants, including ascorbic acid, trolox, glutathione, uric acid, and polyphenolics. Measurement of antioxidants was also realized using an optical ber biosensor based on the quenching of the chemiluminescence produced from the reaction of H2O2 with a coimmobilized luminol/hematein reagent phase.20 The electron-transfer behaviors of ruthenium-bipyridyl complexes (Ru(bpy)33+/2+) immobilized on a Naon lm have been electrochemically investigated,21,22 but these materials have not been used as sensors for antioxidant assays. The radical scavenging mediating reversible uorescence quenching of an anionic conjugated polymer has been proposed as a probe for antioxidants, and the uorescence recovery was suggested to be used to probe the processes of a hydrogen-transfer reaction from antioxidants to radicals and the reduction of radical by antioxidants, but the material was only useful for ascorbic acid detection, the sensitivity being much lower for cysteine, glutathione, and uric acid.23 Thinlm organic photodiode (OPD) chemiluminescence (CL)-sensors for antioxidant capacity screening were developed, but CL intensity (as I0/I) with respect to antioxidant concentration did not give a perfectly linear t (correlation coefcients as r2 were low 0.990 and blank levels were high; intercept 1), and only three antioxidants, namely R-tocopherol, -carotene, and quercetin, were tested as model compounds.24 Linearity of responses over a reasonable concentration range together with the additivity of TAC values for constituents of complex mixtures are a prerequisite for meaningful comparison of TAC values of different food samples found with the aid of a sensor. Of the few antioxidant sensors available in literature, electrochemical-based sensors have been found not to yield perfectly linear calibration curves, and the obtained response for a limited number of antioxidants was generally not a simple function of antioxidant concentration. For example, through the use of amperometric sensors based on iron-containing complexes and protein-modied electrodes, the hydrogen peroxide scavenging capacities of antioxidants were quantied by means of a kinetic model, where the reciprocal of the difference of cathodic currents in the absence and presence of antioxidants was presumed to be
(17) Zinger, B.; Shier, P. Sens. Actuators, B 1999, 56, 206214. (18) Misra, V.; Mishra, H.; Joshi, H. C.; Pant, T. C. Sens. Actuators, B 2002, 82, 133141. (19) Steinberg, I. M.; Milardovic , S. Talanta 2007, 71, 17821787. (20) Palaroan, W. S.; Bergantin, J.; Sevilla, F. Anal. Lett. 2000, 33, 17971810. (21) Buttry, D. A.; Anson, F. C. J. Am. Chem. Soc. 1982, 104, 48244829. (22) Zhang, J.; Zhao, F.; Abe, T.; Kaneko, M. Electrochim. Acta 1999, 45, 399 407. (23) Tang, Y.; He, F.; Yu, M.; Wang, S.; Li, Y.; Zhu, D. Chem. Mater. 2006, 18, 36053610. (24) Wang, X.; Amatatongchai, M.; Nacapricha, D.; Hofmann, O.; de Mello, J. C.; Bradley, D. D. C.; De Mello, A. J. Sens. Actuators, B 2009, 140, 643648.

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX

a linear function of the reciprocal of antioxidant concentration.25 The antioxidant concentration found by means of DNA-modied carbon screen-printed electrode sensors through the use of a bipyridine-Ru(II) complex redox mediator again required kinetic modeling; the ratio of the peak currents obtained in the absence and presence of electro-adsorbed DNA oxidation was a linear function of the reciprocal of the square root of antioxidant concentration.26 The intercept of the calibration equation for a butylated hydroxyanisole (BHA) sensor was as large as the slope, and the sensor worked well only for a single antioxidant (BHA).27 The reciprocal of the difference of the reciprocals of reduction currents of methylene blue (used as a DNA intercalant and redox mediator) was measured after a 30 min irradiation (for photooxidation of adsorbed DNA on indium tin oxide-modied TiO2 electrodes) and initially was a curvi-linear function of antioxidant concentration; only glutathione and gallic acid were tested as model antioxidants.28 In the present work, we have studied the feasibility of transforming the solution-based CUPRAC assay to the optical sensing of simple nonenzymatic antioxidant compounds, their synthetic mixtures, and real samples and developed a low-cost Naon membrane-immobilized cupric neocuproine sensor to be used by antioxidant researchers, food scientists, and bioanalytical chemists. In comparison to Fe(III)-based antioxidant assays29 using octahedrally coordinated chromophore complexes (e.g., Fe(phen)32+, Fe(TPTZ)22+), the tetrahedrally coordinated 1:2 cuprous-neocuproine chelate may be more suited to distorted planar geometry enforced by membrane immobilization, and the resulting loss of sensitivity in analyte detection (i.e., through loss of molar absorptivity of the complex within a xed time of antioxidant capacity measurement) would be less.30 The developed method was successfully applied interference free to complex samples such as fruit juices without preliminary treatment. The results obtained from the optical sensor-based CUPRAC assay were correlated to those found by the wellestablished solution-based CUPRAC assay.9 This work offers the potential of future renements of the CUPRAC antioxidant assay to a kit format. EXPERIMENTAL SECTION Instrumentation and Chemicals. The following chemical substances of analytical reagent grade were supplied from the corresponding sources. Neocuproine (2,9-dimethyl-1,10-phenanthroline), morin (MR), quercetin (QR), naringenin (NG), gallic acid (GA), and uric acid (UA) were from Sigma (Steinheim, Germany); trolox (TR), rosmarinic acid (RA), R-tocopherol (TP), L-ascorbic acid (AA), and Naon 115 peruorinated membrane (thickness 0.005 in.) were from Aldrich (Steinheim, Germany). Copper(II) chloride dihydrate, ammonium acetate (NH4Ac), ethanol (EtOH), and bilirubin (BIL) were from Merck (Darm(25) Guo, Q.; Ji, S.; Yue, Q.; Wang, L.; Liu, J. Anal. Chem. 2009, 81, 5381 5389. (26) Liu, J.; Su, B.; Lagger, G.; Tacchini, P.; Girault, H. H. Anal. Chem. 2006, 78, 68796884. (27) Jayasri, D.; Narayanan, S. S. Food Chem. 2007, 101, 607614. (28) Liu, J.; Roussel, C.; Lagger, G.; Tacchini, P.; Girault, H. H. Anal. Chem. 2005, 77, 76877694. (29) Berker, K. I.; Gu c lu , K.; Tor, I.; Apak, R. Talanta 2007, 72, 11571165. (30) Ditzler, M. A.; Pierre-Jacques, H.; Harrington, S. A. Anal. Chem. 1986, 58, 195200.

stadt, Germany); (+)catechin (CT), myricetin (MYR), setin (FS), and kaempferol (K) were from Fluka (Buchs, Switzerland). Lipton green tea (Camellia sinensis) was purchased from Unilever San. Tic. Turk AS (Istanbul, Turkey); orange and cherry juice were from Tamek Gda Konsantre San. Tic. AS (Istanbul, Turkey), and apricot and peach juice were from Aroma Meyve Sular Gda San. AS (Bursa, Turkey). The visible spectra and absorption measurements were recorded in matched quartz cuvettes (using a Varian CARY Bio 100 UV-vis spectrophotometer (Mulgrave, Victoria, Australia)). The optical thickness of the cuvettes was 1 cm for solution phase and 1 mm for Naon solid sensor measurements. Other related apparatus and accessories were a Elektromag vortex stirrer (Istanbul, Turkey) and a BIOSAN Programmable rotator-mixer Bulti Bio RS-24 (Riga, Latvia). Preparation of Solutions. CuCl2 solution, 1.0 10-2 M, was prepared by dissolving 0.4262 g of CuCl2 2H2O in water and diluting it to 250 mL. Ammonium acetate buffer at pH ) 7.0, 1.0 M, was prepared by dissolving 19.27 g of NH4Ac in water and diluting it to 250 mL. Neocuproine (Nc) solution, 7.5 10-3 M, was prepared daily by dissolving 0.039 g of Nc in absolute ethanol and diluting it to 25 mL with EtOH. All phenolic compounds were freshly prepared in EtOH at 1 mM concentration and L-ascorbic acid in water at the same concentration. Uric acid (1 mM) and bilirubin (0.5 mM) were prepared in 0.01 M NaOH, and the excess base was neutralized with 0.01 M HCl. The phenolics stock solutions were stored at +4 C in a refrigerator prior to analysis. The green tea bag (2 g) was dipped into and pulled out of beakers containing 250 mL of freshly boiled water for the rst 2 min and allowed to steep for the remaining 3 min in the covered beakers (total steeping time was 5 min). The bag was removed, and the partly turbid solutions were ltered through a black-band Whatman quantitative lter paper after cooling to room temperature. Recommended Procedure. (i) Solution-Based CUPRAC Method. The CUPRAC method, as described by Apak et al.,9 is based on the reduction of a cupric neocuproine complex (Cu(II)-Nc) by antioxidants to the cuprous form (Cu(I)-Nc). To a test tube were added 1 mL each of Cu(II), Nc, and NH4Ac buffer solutions. Antioxidant standard solution (x mL) and H2O (1.1 - x) mL were added to the initial mixture so as to make the nal volume: 4.1 mL. The tubes were stoppered, and after 1/2 h, the absorbance at 450 nm (A450) was recorded against a reagent blank. The scheme for normal measurement of antioxidants is summarized as 1 mL of Cu(II) + 1 mL of Nc + 1 mL of buffer + x mL of antioxidant soln. + (1.1 - x) mL of H2O; total volume ) 4.1 mL, measure A450 against a reagent blank after 30 min of reagent addition. (ii) Optical Sensor-Based CUPRAC Method. The commercial Naon membrane was sliced into 4.5 0.5 cm pieces, immersed into a tube containing 2 mL of 1.0 10-2 M Cu(II) + 2 mL of 7.5 10-3 M Nc + 2 mL of 1 M NH4Ac + 2.2 mL of H2O, and agitated for 30 min in a rotator. The reagent-impregnated membrane was taken out and immersed in a tube containing 8.2 mL of standard antioxidant or real solutions (x mL of sample + (8.2 - x) mL of EtOH). The tube was placed in a rotator and agitated for 30 min so as to enable color development. The
Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX C

colored membrane was taken out and placed in a 1 mm optical cuvette containing H2O (to prevent sticking of slices to the walls of the cuvette), and its absorbance at 450 nm was read against a blank membrane prepared under identical conditions excluding analyte. The calibration curves (absorbance vs concentration graphs) of each antioxidant were constructed under the described conditions, and their trolox equivalent antioxidant capacities (TEAC coefcients, found as the ratio of the molar absorptivity of each compound to that of trolox in the optical sensor-based CUPRAC method) were calculated. Standard Addition of AA, QR, and TP to Green Tea Extract. A 20 L aliquot of green tea infusion and 25 L of 1 mM QR, 200 L of 1 mM AA, or 50 L of 1 mM TP solution were taken into a tube. AA-, QR-, and TP-added solutions were separately subjected to CUPRAC spectrophotometric analysis. Measurement of Synthetic Mixture Solutions. Synthetic mixtures of the antioxidants in EtOH were prepared in suitable volume ratios, and these mixtures were diluted to 8.2 mL with EtOH and subjected to optical sensor-based CUPRAC analysis. The theoretical trolox equivalent TAC of a synthetic mixture solution (expressed in the units of mM TR) was calculated by multiplying the TEAC coefcient of each antioxidant constituting the mixture with its nal concentration (in mM TR units) and summing up the products. The experimental trolox equivalent TAC of the same mixture was calculated by dividing the observed absorbance (A450) to the molar absorptivity of TR (TR being 2.40 104 L mol-1 cm-1 under the selected conditions). Then, the theoretically found TAC were compared to the experimentally observed ones to test the applicability of Beers law (i.e., the principle of additivity of individual absorbances of constituents making up a mixture). Validity of Beers law for a mixture implies that the observed absorbance is the sum of the individual absorbances of the constituents. TAC expected ) TEAC1concn1 + TEAC2concn2 + ... + TEACnconcnn (1.1) TAC found experimentally ) absorbance(total) ( intercept 103 (1.2) trolox Interference Studies. The interference effects of 1000-fold (as mol/mol) concomitant species commonly found in fruit juices to the determination of 4.88 10-6 M QR in ethanolic solution using optical sensor-based CUPRAC methods were studied. Statistical Analysis. Descriptive statistical analyses were performed using Excel software (Microsoft Ofce 2003) for calculating the means and the standard error of the mean. Results were expressed as the mean standard deviation (SD). Using SPSS software for Windows (version 13), the data were evaluated by two-way analysis of variance (ANOVA).31 RESULTS AND DISCUSSION Rapid and low-cost chemical sensing of antioxidants with the use of sensors has recently gained importance. Optical sensors
(31) Miller, J. C.; Miller, J. N. Statistics for Analytical Chemists, third ed.; Ellis Horwood and Prentice Hall: New York and London, 1993.

Table 1. Precision and Recovery of the Optical Sensor-Based CUPRAC Method antioxidant quercetin addition to green tea extract added conc., M mean, M S.D.a R.S.D., %b REC, %c added conc., M mean, M S.D.a R.S.D., %b REC, %c added conc., M mean, M S.D.a R.S.D., %b REC, %c 3.05 2.83 0.08 3.16 92.79 24.23 24.20 0.53 2.19 99.87 6.09 6.08 0.39 6.54 99.83

ascorbic acid addition to green tea extract

R-tocopherol addition to green tea extract

a Standard deviation. ) 3).

Relative standard deviation. c Recovery (N

have the advantages of exibility and miniaturization and may also be used for remote sensing. Among optical sensors, colorimetric sensors based on light absorption at a specied visible wavelength are the simplest types and can be used for relatively rapid on-site applications. Moreover, such sensors working in the visible range of the electromagnetic spectrum are not affected from potentially interfering plant pigments essentially absorbing in the UV range. Considering the enforced planar (distorted tetrahedral) geometry of cupric chelates when immobilized on membranes, the more benecial coordination number and geometry of copper compared to iron30 may enable the design of a more sensitive antioxidant sensor with cupric-neocuproine. A ferric ion-based antioxidant sensor does not exist in literature, probably due to the difculties encountered in xing the tripositive Fe(tripyridyltriazine)23+ hydrophilic cation (the FRAP reagent) in a hydrophobic polymer matrix and in regionally achieving a low pH (3.6) on a polymer membrane. Thus, an inexpensive colorimetric sensor was developed and was effectively used in a quantitative TAC assay. Analytical Figures of Merit. Table 1 summarizes the precision and recovery of the optical sensor-based CUPRAC assay using AA, QR, and TP as representative antioxidant compounds. The precision, which is expressed as the relative standard deviation (RSD, %) in absorbance measurement within the tested concentration range, was approximately 6.54%. The recovery of the method varied from 92.8 to 99.9% within individual batches covering vitamins, simple phenolic antioxidants, and avonoids. The recoveries for the individual antioxidant compounds were calculated by means of a CUPRAC calibration curve (as absorbance vs concentration) for the specic antioxidant of concern. Figure 1 illustrates the redox reaction between the chromogenic oxidizing reagent used for the CUPRAC assay, i.e., bis(neocuproine)copper(II) chloride (Cu(II)-Nc) and antioxidants. This reagent was useful at pH 7, and the absorbance of the Cu(I)-chelate formed as a result of redox reaction with reducing antioxidants (i.e., polyphenolics, vitamins, synthetic antioxidants) was measured at 450 nm. The color was due to the Cu(I)-Nc chelate formed (see Figure 2, for Cu(I)-Nc spectra obtained by reacting varying concentrations of TR with the CUPRAC reagent).

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX

Figure 1. CUPRAC reaction and chromophore: Bis(neocuproine)copper(I) chelate cation. (Protons liberated in the reaction are neutralized by the NH4Ac buffer.)

Figure 2. Visible spectra of Cu(I)-Nc chelate produced as a result of optical sensor-based CUPRAC reaction with varying concentrations of trolox (nal concentrations).

The linear equation for the calibration graph of TR drawn at the wavelength of 450 nm with respect to the optical sensor-based CUPRAC method was A450 ) 2.40 104CTR + 0.054 (r ) 0.999)

and the molar absorptivity was ) 2.40 104 L mol-1cm-1. The limit of detection (LOD) and limit of quantication (LOQ) for TR in the optical sensor-based CUPRAC assay were calculated using the equation; LOD ) 3 sbl/m and LOQ ) 10 sbl/m, respectively, where sbl is the standard deviation of a blank and m is the slope of the calibration line. The LOD and LOQ for TR were found to be 1.01 and 3.33 M, respectively. The precision, which is expressed as the relative standard deviation (RSD, %) in the tested concentration range, was approximately 5.20%. CUPRAC absorbances of TR were linear within the concentration range of 4.11 10-8-5.18 10-5 M (as nal concentrations in solution), and the method showed excellent linearity (r ) 0.999) over a relatively broad concentration range of analyte, a condition not frequently experienced in solid sensors. To determine the reproducibility of the method, a ruggedness experiment was performed. The optical sensor-based CUPRAC procedure for fruit samples gave intra- and interassay coefcient of variation (CV)s around 1.03% and 5.71%. This large difference between intra- and interassay precision values can also be found in other work.32
(32) Pfaf, M. W. Nucleic Acid Res. 2001, 29, 20022007.

The cupric-neocuproine loaded (single-use) sensor was tested for stability and was shown to lose only 3% signal intensity after 15 days of storage in distilled water kept in a desiccator at room temperature. The correlation between the CUPRAC absorbance (A) for a given antioxidant compound (ascorbic acid, R-tocopherol, trolox, uric acid, bilirubin, and other phenolic antioxidants) and its molar concentration (c) was evaluated using 14 compounds. Table 2 summarizes the linear equations (A ) mC + n), correlation coefcients (r), and linear concentration ranges of these pure compounds. As can be seen from Table 2, all antioxidants could be assayed with the optical sensor-based CUPRAC assay (TEAC coefcients (signicantly different) and (P ) 0.05, Fexp ) 0.681, Fcrit (table) ) 4.667, Fexp < Fcrit (table))). The TEAC coefcients found by the optical sensor-based CUPRAC method correlated well (r ) 0.876) with those of the solution-based CUPRAC method (Table 2). If serum antioxidants are excluded, the correlation coefcient for 12 compounds is raised to r ) 0.957. Charge transport in solid polymer matrixes with redox centers in their ground state may involve various mechanisms such as physical diffusion of redox species,33 and these mechanisms may be quite different from those in aqueous solution. The redox reactions of antioxidants with the Naon membrane-xed reagent, cupric neocuproine, initially requires the physical diffusion of the antioxidant to the membrane surface, decreasing the reaction rate with respect to that in bulk solution. The steric hindrance caused by the bulky substituents of the polyphenol and the immobilized oxidizing reagent is also a slowing factor for electron transfer, whereas the local enrichment of the oxidizing reagent for antioxidant molecules diffused to the membrane surface is a rate enhancing factor compared to that in solution. These conicting factors hint to the fact that there may not be an exact one-to-one correspondence between the TEAC values of antioxidants measured with the solution- and sensor-based CUPRAC procedures (Table 2). The large TEAC value of bilirubin may possibly be attributed to the 450 nm absorption interference of biliverdin emerging as the oxidation product of bilirubin.34 Nevertheless, the slope of the calibration line is still close to unity (i.e., TEACsolution ) 1.033TEACsensor + 0.099). All of the easily oxidized avonoids exhibited standard reduction potentials of e0.2 V, whereas naringenin, having a potential
(33) Kaneko, M. Prog. Polym. Sci. 2001, 26, 11011137. (34) Stocker, R.; Glazer, A. N.; Ames, B. N. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 59185922.

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX

Table 2. Linear Calibration Equations, TEAC Coefcients, and Linear Concentration Range of the Tested Antioxidants with Respect to the Optical Sensor-Based CUPRAC Methoda antioxidants quercetin (QR) morin (MR) setin (FS) myricetin (MYR) catechin (CT) kaempferol (K) rosmarinic acid (RA) gallic acid (GA) naringenin (NG) ascorbic acid (AA) R-tocopherol (TP) uric acid (UA) bilirubin (BIL) A A A A A A A A A A A A A linear equation and correlation coefcient ) ) ) ) ) ) ) ) ) ) ) ) ) 9.88 4.61 7.44 3.31 4.61 2.95 9.19 5.06 1.46 1.71 3.11 3.30 1.10 104c 104c 104c 104c 104c 104c 104c 104c 104c 104c 104c 104c 105c + + + + + + + + + 0.015, 0.018, 0.016, 0.034, 0.054, 0.106, 0.015, 0.002, 0.005, 0.059, 0.010, 0.129, 0.076, r r r r r r r r r r r r r ) ) ) ) ) ) ) ) ) ) ) ) ) 0.998 0.996 0.999 0.999 0.999 0.996 0.999 0.999 0.999 0.999 0.999 0.999 0.995 3.52 6.95 6.20 2.53 1.30 1.35 7.14 9.44 3.05 6.36 1.94 1.51 2.20 linear range (M) 10-7 10-7 10-8 10-6 10-7 10-7 10-7 10-7 10-6 10-6 10-6 10-7 10-8 1.30 2.77 2.05 4.02 2.70 4.04 1.43 2.56 8.86 7.93 4.21 3.54 1.11 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-5 10-6
(table).

TEAC (optical sensorbased CUPRAC) 4.11 1.92 3.10 1.38 1.92 1.23 3.83 2.10 0.60 0.71 1.29 1.38 4.58

TEAC (solutionbased CUPRAC)9,44 4.38 1.88 3.90 1.38 3.09 1.58 5.30 2.62 0.05 0.96 1.10 0.96 3.18

a TEAC coefcients (signicantly different); P ) 0.05, Fexp ) 0.681, Fcrit (r ) 0.876).

(table)

) 4.667, Fexp < Fcrit

TEACsolution ) 1.033 TEACsensor + 0.099

close to that of the Cu(Nc)22+-Cu(Nc)2+ couple,9 underwent a slow reaction with the reagent. Naringenin oxidation could only be forced to completion after 50 C incubation in the solutionbased CUPRAC method (TEAC coefcient for NG ) 0.05). In the optical sensor-based CUPRAC method, NG could be directly assayed without incubation, the corresponding TEAC coefcient being 0.60. The TEACCUPRAC coefcients of NG in pure EtOH and MeOH were 0.05 and 0.57, respectively,35 probably due to facilitated e-transfer in ionizing solvents capable of anion (phenolate) solvation, because MeOH is the alcohol that best supports ionization.36 Compared to aqueous medium, the sensor membrane may also be considered to be a less hydrophilic medium like MeOH, enhancing e-transfer. Moreover, by embedding a cationic redox-active metal complex (Cu(Nc)22+) into a water-insoluble anionic polymer such as Naon, a molecular aggregate can be formed showing a unique and active redox behavior that cannot be achieved with either a homogeneous solution or a neat catalyst.37 Such a connement of the redox-active complex may increase its local concentration (compared to that in homogeneous solution), thereby increasing the redox reaction rate (e.g., for naringenin). As opposed to the enhanced antioxidant power of naringenin, catechin and rosmarinic acid exhibited signicantly lower TEAC coefcients on the sensor membrane (than in solution) (Table 2). Catechin lacks the 2,3-double bond conjugated with the 4-oxo group responsible for electron delocalization, which is considered to be an important prerequisite for high antioxidant power.38 The superior antioxidant ability of quercetin results from the formation of a stable aryloxy radical, due to C2-C3 double bond and the resulting planar geometry39 which delocalizes the radical throughout the entire molecule, whereas A and B rings are perpendicular to each other in catechin.40 When the aryloxy radical produced from 1-e oxidation of a avonoid is stabilized by conjugation, the
zyu (35) C elik, S. E.; O rek, M.; Gu c lu , K.; Apak, R. Talanta, in press, DOI: 10.1016/j.talanta.201002.025, (2010). (36) Litwinienko, G.; Ingold, K. U. J. Org. Chem. 2003, 68, 34333438. (37) Abe, T.; Kaneko, M. Prog. Polym. Sci. 2003, 28, 14411488. (38) Pietta, P.-G. J. Nat. Prod. 2000, 63, 10351042. (39) Rice-Evans, C. A.; Miller, N. J.; Paganga, G. Free Radical Biol. Med. 1996, 20, 933956. (40) Fukuhara, K.; Nakanishi, I.; Kansui, H.; Sugiyama, E.; Kimura, M.; Shimada, T.; Urano, S.; Yamaguchii, K.; Miyata, N. J. Am. Chem. Soc. 2002, 124, 59525953.

redox potential of the avonoid is lowered, increasing its antioxidant power. Considering the enforced planar geometry of copper-neocuproine on the sensor membrane, lack of planarity of catechin is an important drawback playing part in the decreased antioxidant power of CT. On the other hand, in spite of the four phenolic -OH groups and excellent conjugated structure of rosmarinic acid, its TEAC coefcient in the sensor assay decreased, probably due to its large molecular size being an important parameter in optical sensor response41 and its low pKa which is 2.8.42 Likewise, the relatively decreased TEAC capacity of ascorbic acid (with respect to that measured in the solutionphase CUPRAC method) may be attributed to its negative charge at pH ) 7, since ascorbic acid is in the form of monohydrogen ascorbate (pKa1 ) 4.2 and pKa2 ) 11.6) at the working pH of the sensor and should essentially be repelled by the negatively charged Naon membrane.43 TAC Measurement of Synthetic Mixture Solutions. Synthetic mixtures of antioxidants exhibited the theoretically expected antioxidant capacities (TAC) within 6.8% (Table 3), meaning that chemical deviations from Beers law were essentially absent, and the CUPRAC absorbances of constituents of antioxidant mixtures were additive. The original CUPRAC method was previously shown to be free from chemical deviations from Beers law, as demonstrated on synthetic mixtures of hydrophilic phenolic compounds.9 This is a prerequisite for precise estimation of antioxidant capacity, as the capacity of a mixture should be composed of the sum of individual capacities of constituents in order to make reliable comparisons of the antioxidant power of different foodstuffs. The two-way analysis of variance (ANOVA) comparison by the aid of F-test of the mean-squares of between-treatments (i.e., theoretically expected capacity with respect to optical sensor-based CUPRAC method and experimentally found capacities of different
(41) Zhao, S.; Zhang, K.; Bai, Y.; Yang, W.; Sun, C. Bioelectrochemistry 2006, 69, 158163. (42) Rein, M. Copigmentation reactions and color stability of berry anthocyanins, Academic Dissertation, EKT series 1331, University of Helsinki, Department of Applied Chemistry and Microbiology, Food Chemistry Division, Helsinki, 2005. (43) Furbee, J. W., Jr.; Thomas, C. R.; Kelly, R. S.; Malachowski, M. R. Anal. Chem. 1993, 65, 16541657. zyu (44) Apak, R.; Gu c lu , K.; Demirata, B.; O rek, M.; C elik, S. E.; Bektas og lu, zyurt, D. Molecules 2007, 12, 14961547. B.; Berker, K. I.; O

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX

Table 3. Comparison of the Theoretically Expected and Experimentally Found Trolox (TR)-Equivalent Antioxidant Capacities (in mM TR units) of Synthetic Mixtures of Antioxidants (with Respect to the Optical Sensor-Based CUPRAC Assay)a
capacity expected (as mM TR-equivalent) 4.25 10
-2

Table 4. Interference of Various Molecular Species with the Optical Sensor-Based CUPRAC Assay of 4.88 10-6 M Quercetin (QR) in Ethanolic Solutiona presence (+) or absence (-) of interference with the optical sensor-based CUPRAC method +

number 1 2 3 4 25 50 50 60 50 25 25 60 25 50 50 50

composition of mixture L L L L L L L L L L L L of of of of of of of of of of of of 1 1 1 1 1 1 1 1 1 1 1 1 mM mM mM mM mM mM mM mM mM mM mM mM QR GA KM MOR TR RA QR MOR RA GA TR KM

capacity found experimentally (as mM TRequivalent) (4.07 0.30) 10

-2

interfering agent oxalate tartarate citrate glucose malic acid fumaric acid fructose galactose nitrite

interfering agent/ QR mole ratio 1000 1000 1000 1000 1000 1000 1000 1000 <50

3.18 10-2 3.82 10-2 3.61 10-2

(3.22 0.19) 10-2 (4.08 0.13) 10-2 (3.50 0.13) 10-2

a The QR and interfering agent concentrations in nal solution were 4.88 10-6 and 4.88 10-3 M, respectively.

a Samples were analyzed in triplicate; P ) 0.05, Fexp ) 0.001, Fcrit (table) ) 10.13, Fexp < Fcrit (table).

regression analysis of capacity data presented in Figure 3 found with the solution-based CUPRAC assay showed that this assay correlated well with the optical sensor-based CUPRAC assay. Interferences in the Optical Sensor-Based CUPRAC Assay. For the developed sensing method, the possible interference effects of 1000-fold (as mol/mol) concomitant species commonly found in fruit juices on the determination of 4.88 10-6 M QR in ethanolic solution are shown in Table 4. The presence of 1000fold citrate, oxalate, and tartarate ions that may be found in fruit juices and other food plant extracts did not interfere with the CUPRAC reaction forming the basis of sensor measurement. CONCLUSIONS The Cu(II)-Nc reagent was electrostatically immobilized onto a cation exchanger Naon membrane and was demonstrated to retain its reactivity toward various antioxidants after incorporation into the exchanger membrane. The developed optical sensor was validated to quantify antioxidants over a wide concentration range in comparison to solution-phase methods and showed good sensitivity for antioxidant detection in complex samples such as fruit juices. The optical sensor-based CUPRAC assay can be used for quantitative assessment of total antioxidant capacity of complex mixtures in trolox equivalent units. Common food ingredients like oxalate, citrate, fruit acids, and common sugars did not interfere with the proposed sensing method. This method offers advantages over solution-based assays, including the potential of being used as a rapid, low-cost, diversely, and easily applicable screening tool for TAC of plant and food extracts of unknown composition without pretreatment and of future renements of the assay to a kit format. ACKNOWLEDGMENT The authors would like to thank Istanbul University Research Fund, Bilimsel Arastirma Projeleri (BAP) Birimi Koordinatorlugu for the Projects BAP-2724 and YADOP-6849. Received for review March 11, 2010. Accepted April 13, 2010.
AC100646K

Figure 3. TAC values (mM) of some commercial fruit juices using the solution-based and optical sensor-based CUPRAC assays. Data are presented as (mean ( SD) (error bars), N ) 3. (P ) 0.05, Fexp ) 0.131, Fcrit (table) ) 10.13, Fexp < Fcrit (table).)

mixtures in Table 3) and of residuals31 for a number of real samples (consisting of synthetic mixtures of antioxidants) enabled one to conclude that there was no signicant difference between the population means for a given sample. In other words, the experimentally found capacity results and theoretically expected capacity calculations were alike at a 95% condence level (Fexp ) 0.001, Fcrit ) 10.13, Fexp < Fcrit at P ) 0.05). Thus, the proposed methodology was validated for synthetic mixtures of antioxidants of differing lipophilicity. On the other hand, there was signicant difference between samples with respect to composition of mixtures (i.e., the residual mean-square was much greater than between-sample mean-square at 95% condence level). This was natural, as these mixtures were deliberately prepared at different total concentrations of trolox equivalents. TAC Measurement of Fruit Juices. In Figure 3, the optical sensor-based and solution-based TACCUPRAC values of fruit juices were reported as trolox equivalents (mM TR). The hierarchy for TAC of fruit juices with respect to optical sensor-based CUPRAC method was apricot < peach < orange < cherry. Linear

Analytical Chemistry, Vol. xxx, No. xx, Month XX, XXXX