Biosensors & Bioelectronics 16 (2001) 363– 369 www.elsevier.com/locate/bios Real-time detection of L-ascorbic acid and hydrogen peroxide in crude food samples employing a reversed sequential differential measuring technique of the SIRE-technology based biosensor K. Kriz a, M. Anderlund b, D. Kriz a,b,* a Department of Pure and Applied Biochemistry, Lund Uni6ersity, P.O. Box 124, SE-221 00 Lund, Sweden b Chemel AB, Research Park IDEON, SE-223 70 Lund, Sweden Received 11 September 2000; received in revised form 17 April 2001; accepted 26 April 2001 Abstract Detection of the common electrochemical interferents, ascorbic acid and hydrogen peroxide, using a SIRE (Sensors based on Injection of the Recognition Element) technology based biosensor in reverse mode operation is reported. The differential measuring principle employed in the SIRE biosensor during operation in reverse mode is such that the sample is measured first in the presence of enzyme (yielding matrix signal only), and then measured again in the absence of enzyme (yielding signal from matrix+ analyte). Subtraction of the signal obtained in the presence of enzyme from the signal obtained in the absence of enzyme gives a specific signal for the analyte only and correlates directly to its concentration in solution. The linear range for the determination of ascorbic acid and hydrogen peroxide was 0 – 3 mM and 0 – 2 mM, respectively, with an enzyme concentration of 25 U ascorbate oxidase/ml and 1000 U catalase/ml. The reproducibility was 5% for ascorbic acid (R.S.D. n= 15) and 10% for hydrogen peroxide (R.S.D. n= 18). The cost per measurement was 0.28 USD for ascorbic acid analysis and 0.0008 USD for hydrogen peroxide analysis. The degradation of ascorbic acid in cereal was followed in real-time, as was the stabilization of low pH on the degradation process. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Biosensor; Ascorbic acid; Hydrogen peroxide; SIRE; Differential measurement 1. Introduction The use of biosensors in food analysis has sparked the interest of many researchers due to the stringent demands on specificity, sensitivity, rapidity, stability and accuracy set by the food industry for analytical analysis of food products (Wagner and Guilbault, 1994). Hydrogen peroxide is commonly used as a food preservative in milk and cheeses (Fox and Kosikowski, 1962) or as a sterilant of packaging materials (Toledo, 1975) due to its inherent sporicidal and bactericidal properties (Juven et al., 1996) while ascorbic acid (vitamin C) has been used as a pharmaceutical, food preservative and a food supplement (Bauernfeind, 1982) either to restore the loss of vitamin during processing or to enhance the nutritious qualities of the food * Corresponding author. Tel.: + 46-46-18-2230; fax: + 46-46-2862499. product. Detection and quantification of hydrogen peroxide in the food industry often involves spectrometric, fluorescence or chemiluminescence assays which employ horseradish peroxidase (HRP) and follow the oxidation of different commercial substrates in the presence of H2O2 (Juven et al., 1996). Detection of ascorbic acid has involved biological, chemical and chromatographic methods (Iqbal, 1995), where HPLC has been the most extensively used for the determination of vitamins in food (Kall and Andersson, 1999), plasma and urine (Rumelin et al., 1999). Electrochemical biosensors may also be used for the determination of both hydrogen peroxide and ascorbic acid providing fast and accurate results, but often require the use of an electrochemical mediator to shuttle electrons from the active site of the enzyme or biomolecule to the electrode surface (Gorton et al., 1991; Pandey et al., 1998). Most mediated electrochemical biosensors used in the detection of hydrogen perox- 0956-5663/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S0956-5663(01)00151-8 364 K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 ide have involved the immobilization of a peroxidase, such as horse radish peroxidase to the electrode surface and the use of ferrocene as the electroactive mediator (Mulchandani et al., 1995); however, other mediators such as, N-methyl phenazine methosulfate (Liu et al., 1997), osmium redox polymer (Park, 1999), polypyrrole (Lovovich and Scheeline, 1997), and tetrathiafulvalenep-tetracyanoquinodimethane (Korell and Spichiger, 1994) have been successfully employed, among others. Tissue-based bioelectrodes employing asparagus tissue and ferrocene mediation (Oungpipat et al., 1995) have also been used for H2O2 determination as well as artificial electrodes using Prussian Blue polycrystals as the artificial recognition element which, circumvents problems normally associated with biosensor instability due to biomolecule denaturation (Karyakin and Karyakina, 1999). Detection of ascorbic acid in food products such as beverages and vitamin tablets has involved electrochemical analysis in combination with flow injection analysis and gold microelectrodes (Matos et al., 1998), batch injection analysis and a mercury drop electrode (De Donato et al., 1999) and capillary electrophoresis (Olsson et al., 1998). Our group has previously suggested the use of SIRE (Sensors based on Injection of the Recognition Element) technology based biosensors for the detection of metabolites in crude samples (Kriz and Johansson, 1996). The SIRE biosensor uses a reaction chamber (integration of an electrochemical transducer and enzyme solution) in combination with a flow-injection principle where a small amount of the recognition element is introduced into the reaction chamber, and is only used once before being discarded. Consequently, this circumvents sensor stability problems associated with the instability of the recognition element. A differential measuring technique is employed where the sample is measured both in the presence and absence of enzyme giving control over the matrix signal without the prerequisite of requiring sample pre-treatment. The versatility of the SIRE biosensor has been demonstrated by the detection of several different metabolites (Kriz et al., 1998a), and its industrial and medical applications have been shown by the determination of the glucose concentration in an on-going fermentation process (Kriz et al., 1998b) and whole blood (Johnson and Kriz, 1998), respectively. However, it is the SIRE biosensor’s differential measuring technique, which gives it the flexibility to measure a wide range of analytes using the same transducer, including the common electrochemical interferents of amperometric measurements, ascorbic acid and hydrogen peroxide, when operated in a reversed mode. At the same time the SIRE biosensor also offers certain advantages for industrial analysis such as, its ability to operate under extreme conditions of thermal and mechanical stress, to operate in the presence of high concentration of interfering substances and to stand direct sterilization. The enzymes ascorbate oxidase (E.C. 1.10.3.3, Lascorbate:oxygen oxidoreductase) and catalase (E.C.1.11.1.6, hydrogen peroxide: hydrogen peroxide oxidoreductase) catalyse the following reactions, L-ascorbic acid+ 12 O2 = \dehydroascorbic acid+H2O 2H2O2 = \ 2H2O+O2 and can be employed for the determination of L-ascorbic acid and hydrogen peroxide, respectively, present in aqueous solution. It is here, in this paper, we report new ascorbate and peroxide sensors based on the SIRE-technology for food analysis. The effect of the enzyme concentration, with the respect to cost and sensor sensitivity has been investigated. The influence of interfering substances present in crude samples of cereal food products containing fruit juice concentrate, vegetable oil, skim milk powder, iron, and vitamin was also studied. The stability of the sensors was investigated in buffer systems with a well-defined composition and in crude samples containing complex electroactive substances. In addition, we have studied the effect of using strong acid to prevent enzymatic destruction of ascorbic acid in food products containing fruit juice concentrate. 2. Materials and methods 2.1. Material Ascorbate oxidase (E.C. 1.10.3.3) and catalase (E.C. 1.11.1.6) from Cucurbita species and Bovine liver, respectively, were obtained from SIGMA (St Louis, MO, USA). Bovine Albumin, hydrogen peroxide (30 wt.% in water) and L-ascorbic acid were also obtained from SIGMA (St Louis, MO, USA). The SIRE-Biosensor P100 and phosphate buffer pH 5.8 and pH 7.4 were obtained from Chemel AB (Lund, Sweden). All solutions were prepared from distilled water. 2.2. Preparation of the enzyme solutions The enzymes, ascorbate oxidase and catalase, were dissolved in 0.1 M phosphate buffer, pH 5.8 and 7.4, respectively (Chemel AB, Lund, Sweden). 2.3. Stabilization of ascorbic acid in samples containing fruit juice concentrate and metal ions Meta-phosphoric acid (20 ml, 20%) was added to approximately 5 mg food products containing fruit juice concentrate during vigorous stirring. Before the analysis, 40 ml distilled water was added and the pH adjusted to 5.8 with NaOH. Finally, phosphate buffer K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 pH 5.8 was added, giving a total sample volume of 100 ml and analysis of the ascorbic acid was performed with the SIRE-biosensor. 2.4. The measuring procedure The SIRE-biosensor was equilibrated with a continuous internal buffer flow (0.1 ml/min) until the current had stabilised at the applied potential (+450 mV vs silver wire reference electrode). The reaction chamber is positioned in front of the amperometric transducer composed of a potentiostatic three-electrode configuration, with platinum wires acting as the working and auxiliary electrode, and an internal silver wire as the reference electrode. The biosensor probe was immersed into the sample solution containing the target analyte and 0.1 ml of the enzyme solution, ascorbate oxidase (0– 30 U/ml) or catalase (0– 1200 U/ml), was injected into the system. The buffer flow was stopped when the enzyme had reached the reaction chamber and the measurement initiated. The analyte, in this case an electroactive compound such as ascorbic acid or peroxide, passed through the membrane, which is located in 365 front of the reaction chamber, and reacted with the enzyme causing non-electroactive products to be formed, such as, H2O and dehydroascorbate and H2O, respectively. After a reaction time of 15–120 s, the current was measured giving the matrix signal caused by interfering compounds that had passed the membrane. After the matrix current was recorded the reaction chamber was washed automatically with buffer (1 ml) to regenerate the sensor and refilled with buffer (without enzyme). After 15–120 s the current was measured giving the value caused by the analyte and interfering compounds in the matrix (see Fig. 1). Thus, subtracting these two values yields a differential biosensor signal, which is compensated for. These processes, which were automatic, and microprocessor controlled, were repeated for different samples containing various concentrations of analyte at room temperature (273 K). Temperature compensation for fluctuations in sample temperature was also carried out automatically using a temperature sensor, which is placed permanently in the biosensor probe and a computer program, which accounts for temperature fluctuations. Fig. 1. An illustration of the electrode configuration and the reversed sequential differential measuring principle of the SIRE biosensor P100. 366 K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 Fig. 2. Influence of different ascorbate oxidase concentrations on the differential response value (solid curve) and on the matrix value (dotted curve) obtained with the SIRE-biosensor P100. The reaction time was 120 s. The measurements were done in a 3.0 mM ascorbic acid solution containing 100 mM sodium phosphate buffer pH 5.8. 3. Results 3.1. The effect of the enzyme concentration Since the enzyme used in the SIRE-biosensor is not reused, the amount of the enzyme needed for the measurement was optimized, with respect to cost and to the sensitivity of the sensor. The influence of the ascorbate oxidase and catalase concentration on the matrix value and differential response value was investigated from 0 –30 U ascorbate oxidase/ml with 3.0 mM ascorbic acid present in the samples and from 0– 1200 U catalase/ml with 2.0 mM hydrogen peroxide present in the samples. The differential response value for ascorbic acid (matrix value subtracted from the matrix+analyte value) increased with increasing ascorbate oxidase concentration up to a saturation level around a concentration of 10–15 U/ml (Fig. 2, solid curve). Higher enzyme concentrations did not increase the differential response value. The matrix value (signal caused by interfering compounds in the sample) decreased with increasing enzyme concentration (0– 10 U/ml) down to a level obtained for pure buffer containing no ascorbic acid (247 biosensor response units) (Fig. 2, dotted curve). Higher enzyme concentrations (above 10 U/ml) did not affect the matrix value. An excess of enzyme was chosen as a working concentration (25 U/ml) so that the enzyme concentration would not be the limiting factor for analyte conversion. The differential response value for hydrogen peroxide increased with increasing catalase concentration up to a saturation level around a concentration of 800 U/ml (Fig. 3, solid curve). Higher enzyme concentrations did not increase the differential response value. The matrix value decreased with increasing enzyme concentration (0 –800 U/ml) down to a level obtained for pure buffer containing no hydrogen peroxide (350 biosensor re- sponse units) (Fig. 3, dotted curve). Higher enzyme concentrations (above 800 U/ml) did not affect the matrix value. An excess of enzyme was chosen as a working concentration (1000 U/ml). 3.2. Sensor response to ascorbic acid and hydrogen peroxide The linear range for the determination of ascorbic acid was 0–3 mM with an ascorbate oxidase concentration within 10– 50 U/ml and using a 120 s reaction time (Fig. 4). The detection limit for ascorbic acid was determined to be 5 mM. The linear range could not be extended for higher ascorbic acid concentrations by performing the measurement using higher enzyme concentrations (50 U/ml), thus, indicating that perhaps enzyme inhibition by dehydroascorbate was occurring or there was a diffusional limitation of oxygen through the membrane to the enzyme. The linear range for the determination of hydrogen peroxide was 0–2.0 mM with a catalase concentration of 1000 U/ml using a 15– 50 sec reaction time (Fig. 5). The detection limit for hydrogen peroxide was 4 mM. A cross-reactivity study was carried out where detection of ascorbic acid was carried out using catalase as the recognition element and detection of hydrogen peroxide was carried out using ascorbate oxidase and the recognition element. No detectable signal was obtained in either of the two cases, thus, further illustrating the selectivity of the biosensor. 3.3. Biosensor accuracy The repeatability was studied for four different concentrations of ascorbic acid (0.2, 0.5, 0.7, and 1.0 mM) and five different concentrations of hydrogen peroxide (0.05, 0.08, 0.1, 0.12, 0.15 mM) and the relative stan- K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 367 Fig. 3. Influence of different catalase concentrations on the differential response value (solid curve) and on the matrix value (dotted curve) obtained with the SIRE-biosensor P100. The reaction time was 15 s. The measurements were done in a 2.0 mM hydrogen peroxide solution containing 100 mM sodium phosphate buffer pH 7.4. dard deviation (R.S.D.) of five independent measurements was calculated. The repeatability for ascorbic acid and for hydrogen peroxide was 3.5% (n= 20) and 8.0% (n=25), respectively. The reproducibility was investigated in the same manner except that the water, buffer, enzyme stock solution, and samples were changed between each subsequent measurement. Three measurements on five different concentrations of ascorbic acid (0.0, 0.5, 1.0, 2.0, and 3.0 mM) and six different concentrations of hydrogen peroxide (0.0, 0.05, 0.08, 0.1, 0.12, 0.15 mM) were performed. The relative standard deviation (R.S.D.) was determined to be 5.0% (n =15) for ascorbic acid and 10% (n =18) for hydrogen peroxide. 3.4. Influence of interfering substances present in cereal products The effect of interfering substances on ascorbic acid analysis was investigated in cereal products containing fruit juice concentrate, vegetable oil, skim milk powder, iron, and vitamins. Cereal product samples (5 gram in 100 ml buffer solution) were spiked with ascorbic acid in the concentration range of 0–3 mM and the obtained values were compared with values obtained in a calibration curve performed in buffer solution. In all cases, nearly 100% of the spike was recovered indicating no significant matrix interferences. The result is shown in Fig. 6. The ascorbic acid concentration in the food sample was determined to be 0.17 mM by the standard addition method. Ascorbic acid in the sample was known (given by the manufacturer) to have a concentration of 0.163 mM. 3.5. Real-time analysis of ascorbic acid degradation in cereal products containing fruit juice concentrate Ascorbic acid is oxidised to dehydroascorbic acid by oxygen, metal ions such as copper (II) or iron (III), and by enzymatic action (Yuan and Chen, 1998). For amperometric detection of ascorbic acid it is important to prevent degradation because measurement of the electroinactive dehydroascorbic acid is not possible. In cases where analysis of ascorbic acid can not be done, for practical reasons, directly or immediately after sampling, the ascorbic acid has to be stabilised and its degradation stopped. The future on-line measuring possibilities offered by the SIRE biosensor for food analysis was demonstrated by following the stabilising effect of low pH on ascorbic acid in cereal products containing fruit juice concentrate and metal ions in real-time. It was shown that samples pre-treated with meta-phosphoric acid show a considerably lower degradation rate of ascorbic acid than samples not treated with acid (Fig. 7). All measurements were performed in real-time allowing for accurate assessment of the ascorbic acid degradation as it was happening. Fig. 4. Calibration curve for ascorbic acid using 25 U ascorbate oxidase/ml and a reaction time of 120 s. Sample solutions contained 0 – 3.0 mM ascorbic acid in 100 mM sodium phosphate buffer pH 5.8. 368 K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 Fig. 5. Calibration curve for hydrogen peroxide using 1000 U/ml catalase and a reaction time of 15 s. Sample solutions contained 0– 2.5 mM hydrogen peroxide in 100 mM sodium phosphate buffer pH 7.4. Fig. 7. The figure shows the degradation of ascorbic acid in an untreated cereal product containing fruit juice concentrate (lower curve), and in a sample that has been pre-treated with meta-phosphoric acid (upper curve). The enzyme (ascorbate oxidase) concentration was 25 U/ml. 4. Discussion In this study we have demonstrated the advantages of using the SIRE technology for the analysis of ascorbic acid and hydrogen peroxide in the food industry. The differential measuring technique allows for fast and accurate detection of ascorbic acid and hydrogen peroxide, which are common interferents of amperometric detection due to the fact that they are easily oxidized at the electrode surface, without the use of mediators or two-electrode systems. Additionally, the SIRE-biosensor circumvents the problem of enzyme instability normally associated with biosensor technology as it uses new and freshly prepared enzyme for each measurement and furthermore, an excess of enzyme is used so that the sensor response is dependent on the diffusion limitation of the analyte. Transducer fouling is prevented by cleaning the transducer in a buffer stream after each measurement. Since the enzyme used in the SIRE-biosensor is not reused, the amount of enzyme needed for each measurement was optimised, costing 0.28 USD/measure- Fig. 6. Ascorbic acid determination in a sample of cereal product containing fruit juice concentrate measured by the standard addition method. Cereal samples spiked with ascorbic acid in the concentration range of 0 – 3 mM and the obtained values (upper curve) are compared with the calibration curve (lower curve) obtained for buffered samples. ment for ascorbic acid analysis and 0.0008 USD/measurement for hydrogen peroxide analysis. The linear range for the determination of ascorbic acid and hydrogen peroxide were 0– 3 mM and 0 – 2 mM, respectively, with an enzyme concentration of 25 U ascorbate oxidase/ml and 1000 U catalase/ml. The linear range could not be extended for higher ascorbic acid and hydrogen peroxide concentrations simply by performing the measurement using higher enzyme concentrations or by using a shorter or longer reaction time. One reason for this maybe that the ascorbate oxidase could be inhibited by its product dehydroascorbate. Another possible explanation to a limitation of the linear range is that the activity of the enzyme could be decreasing due to some diffusional limitation of oxygen through the membrane. These problems can probably be circumvented by changing the membrane or by modifying the shape of the reaction chamber. One of the key advantages of the SIRE biosensor is its ability to perform real-time analysis which was demonstrated by following in real-time the stabilising effect of low pH on samples containing fruit juice concentrate and metal ions. The fact that low pH stabilized the degradative process indicates that enzymatic action was probably responsible for the destruction of ascorbic acid. The active proteins that play a part in the destruction could be deactivated by hydrolysis at low pH and high salt concentration. It is also important to note that in spite of the high salt concentration in the pre-treated samples, measurement of ascorbic acid was still possible using the SIRE biosensor. However, even though the differential measuring procedure does compensate for nearly all types of matrix effects (fluctuating and non-fluctuating), if the total matrix signal of a sample would be so large that it is above the upper limit of the measuring range, a signal could not be calculated using the SIRE bionsensor. Another advantage associated with the SIRE biosensor is the automatic temperature compensation, which K. Kriz et al. / Biosensors & Bioelectronics 16 (2001) 363–369 could be useful when the probe is placed in a process line where there are large temperature variations. In thick and viscous samples which contain large amount of cells or grains, clogging of the probe membrane could be a problem and hence, sensor response could be inhibited. However, monitoring of on-line processing was not studied in this paper. 5. Conclusions We believe that SIRE technology based biosensors can revolutionize how current analysis is performed in the food industry. The SIRE biosensor is versatile, allowing for the detection of a wide range of analytes using the same instrument, in addition to being sensitive, stable and accurate. Its ability to withstand mechanical and thermal stress (due to the fact the recognition element or enzyme is not immobilized on the sensor probe) and to follow processes in real-time allows for the possibility of on-line monitoring which would allow the food industry to follow the composition of raw materials through a processing method. References Bauernfeind, J. Christopher, 1982. Ascorbic acid technology in agricultural, pharmaceutical, food and industrial applications. Adv. Chem. Ser. 200, 395 – 497. De Donato, A., Pedrotti, J.J., Gutz, Ivano, G.R., 1999. A batch injection analysis system for ascorbic acid determination using amperometric detection on a sessile mercury drop electrode, Electroanalysis 11 (15), 1124 – 1129. Fox, P.F., Kosikowski, F.V., 1962. Heat-treated and hydrogen peroxide treated milks for cheddar cheese. J. Dairy Sci. 45, 648. Gorton, L., Bremle, G., Csöregi, E., Jönsson-Pettersson, G., Persson, B., 1991. Amperometric glucose sensors based on immobilized glucose-oxidizing enzymes and chemically modified electrodes. Anal. Chem. Acta. 249, 43 – 54. Iqbal, Zafar, 1995. Methods for the determination of ascorbic acid, Sci. Int. (Lahore) 7 (3), 357 – 360. Johnson, K.A., Kriz, D., 1998. SIRE-technology. Part II. Glucose tolerance monitoring, after a peroral intake, employing small volume whole blood measurements with an amperometric biosensor. Instru. Sci. and Technol. 26 (1), 59 – 67. Juven, Benjamin J., Pierson, Merle D., 1996. Antibacterial effects of hydrogen peroxide and methods for its detection and quantification, J. Food Prot. 59 (11), 1233 – 1241. Kall, M.A., Andersson, C., 1999. Improved method for simultaneous determination of ascorbic acid and dehydroisoascorbic acid in food and biological samples. J. Chromatogr. B: Biomed. Sci. Appl. 730 (1), 101 – 111. . 369 Karyakin, A.A., Karyakina, E.E., 1999. Prussian blue-based ‘‘artifical-peroxidase’’ as a transducer for hydrogen peroxide detection. Application to biosensors. Sens. Actuators, B. 57 (1 – 3), 268 – 273. Korell, U., Spichiger, U.E., 1994. Novel membranes amperometric peroxide biosensor based on a tetrathiafulvalene-p-tetracyanoquinodimethane electrode. Anal. Chem. 66 (4), 510 – 515. Kriz, D., Johansson, A., 1996. A preliminary study of a biosensor based on flow injection of the recognition element. Biosens. and Bioelectron. 11, 1259. Kriz, D., Berggren, C., Johansson, A., Ansell, R.J., 1998a. SIREtechnology. Part I. Amperometric biosensor based on flow injection of the recognition element and differential measurements. Instru. Sci. and Technol. 26 (1), 45 – 47. Kriz, D., Berggren, C., Palmqvist, E., 1998b. SIRE-technology. Part III. Glucose monitoring during fermentation of lignocellulosic hydrolysate by saccharomyces cere6isiae employing a differential amperometric biosensor. Instru. Sci. and Technol. 26 (1), 69 – 79. Liu, H., Ying, T., Sun, K., Li, H., Qi, D., 1997. Reagentless amperometric biosensors highly sensitive to hydrogen peroxide, glucose and lactose based on N-methyl phenazine methosulfate incorporated in a Nafion film as an electron transfer mediator between horseradish peroxidase and an electrode. Anal. Chim. Acta. 344, 187 – 199. Lovovich, V., Scheeline, A., 1997. Amperometric sensors for simultaneous superoxide and hydrogen peroxide detection. Anal. Chem. 69 (3), 454 – 462. Matos, R.C., Augelli, M.A., Pedrotti, J.J., Lago, C.L., Angnes, L., 1998. Amperometric differential determination of ascorbic acid in beverages and vitamin C tablets using a flow cell containing an array of gold microelectrodes modified with palladium. Electoanalysis 10 (13), 887 – 890. Mulchandani, A., Wang, C.L., Weetall, H.H., 1995. Anal. Chem. 67, 94. Olsson, J., Nordstrom, O., Nordstrom, A., Karlberg, B., 1998. Determination of ascorbic acid in isolated pea plant cells by capillary electrophoresis and amperometric detection. J. Chromatogr. 826 (2), 227 – 233. Oungpipat, W., Alexander, P.W., Southwell-Keely, P., 1995. A reagentless amperometric biosensor for hydrogen peroxide determination based on asparagus tissue and ferrocene mediation. Anal. Chim. Acta. 309, 35 – 45. Pandey, P.C., Upadhyay, S., Pathak, H.C., Pandey, C.M.D., 1998. Sensitivity, selectivity and reproducibility of some mediated electrochemical biosensors/sensors. Anal. Lett. 31 (14), 2327 – 2348. Park, Tae-Mung, 1999. Amperometric determination of hydrogen peroxide by utilizing a sol-gel-derived biosensor incorporating an osmium redox polymer as mediator, Anal. Lett. 32 (2), 287 – 298. Rumelin, A., Fauth, U., Halmagyi, M., 1999. Determination of ascorbic acid in plasma and urine by high performance liquid chromatography with ultraviolet detection. Clin. Chem. Lab. Med. 37 (5), 533 – 536. Toledo, R.T., 1975. Chemical sterilants for aseptic packaging. Food Technol. 29 (5), 102 – 112. Wagner, G., Guilbault, G.G. (Eds.), 1994. Food Biosensor Analysis. Marcel Dekker Inc., New York. Yuan, Jian-Ping, Chen, Feng, 1998. Degradation of ascorbic acid in aqueous solution, J. Agric. Food Chem. 46 (12), 5078 – 5082.