Hybrid sequential injection–flow injection manifold for the spectrophotometric determination of total sulfite in wines using o-phthalaldehyde and gas-diffusion

Talanta 77 (2009) 1614–1619

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Hybrid sequential injection–flow injection manifold for the spectrophotometric determination of total sulfite in wines using o-phthalaldehyde and gas-diffusion Paraskevas D. Tzanavaras, Eleni Thiakouli, Demetrius G. Themelis ∗ Laboratory of Analytical Chemistry, Department of Chemistry, Aristotle University of Thessaloniki, 54124 Thessaloniki, Greece

a r t i c l e

i n f o

Article history: Received 24 June 2008 Received in revised form 24 September 2008 Accepted 29 September 2008 Available online 14 October 2008 Keywords: Sulfite Determination Gas-diffusion Sequential injection Spectrophotometry o-Phthalaldehyde Wine

a b s t r a c t A new automated spectrophotometric method for the determination of total sulfite in white and red wines is reported. The assay is based on the reaction of o-phthalaldehyde (OPA) and ammonium chloride with the analyte in basic medium under SI conditions. Upon on-line alkalization with NaOH, a blue product is formed having an absorption maximum at 630 nm. The parameters affecting the reaction – temperature, pH, ionic strength, amount concentration and volume of OPA, amount concentration of ammonium chloride, flow rate and reaction coil length – and the gas-diffusion process – sample and HCl volumes, length of mixing coil, donor flow rate – were studied. The proposed method was validated in terms of linearity (1–40 mg L−1 , r = 0.9997), limit of detection (cL = 0.3 mg L−1 ) and quantitation (cQ = 1.0 mg L−1 ), precision (sr = 2.2% at 20 mg L−1 sulfite, n = 12) and selectivity. The applicability of the analytical procedure was evaluated by analyzing white and red wine samples, while the accuracy as expressed by recovery experiments ranged between 96% and 106%. © 2008 Elsevier B.V. All rights reserved.

1. Introduction Sulfite is a naturally occurring by-product of the fermentation process in winemaking. Sulfites are added to wine and many other food products in order to preserve their freshness and shelf life. It plays two important roles. Firstly, it is an anti-microbial agent, and as such is used to help curtail the growth of undesirable fault producing yeasts and bacteria. Secondly, it acts as an antioxidant, safeguarding the wine’s fruit integrity and protecting it against browning. In winemaking, this usually takes the form of sulphur dioxide. If SO2 is not added, the aging process of wine is greatly accelerated [1,2]. Since the 1980s, the use of sulfites has come under increased scrutiny due to potential health concerns. United States Food and Drug Administration regulations require food and wine producers to indicate “contains sulfites” on the label of any product that has at least 10 mg L−1 . These regulations, which went into effect in 1986, were instituted because sulfite-sensitive individuals who are deficient in the natural enzyme to metabolize it can experience allergic reactions [3,4]. It is therefore important to develop reliable methods for the efficient quality control of sulfite in wines both during production and storage after bottling. Recently reported analytical meth-

∗ Corresponding author. Tel.: +30 2310997804; fax: +30 2310997719. E-mail address: [email protected] (D.G. Themelis). 0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2008.09.051

ods (2000–2008) for the determination of sulfite apply a variety of instrumental approaches: electroanalytical approaches [5–10] including amperometric biosensors [11–13] are popular among scientists due to the properties of the analyte, capillary electrophoresis coupled to either UV [14] or conductivity detection [15,16], fluorescence/chemiluminescence [17–19], vapour generation ICP-OES [20], FTIR [21], piezoelectric sensors [22] and light scattering detection [23]. UV–vis spectrophotometry is an interesting alternative for routine applications as it offers cost-effective and widely available instrumentation, low operational costs and simple handling. On this basis, two batch spectrophotometric methods [24,25] and two optical sensors [26,27] have been reported recently for the determination of sulfites. Although they offer adequate analytical figures of merit they suffer for the necessity of synthesizing the analytical reagent [24] and the sensing membranes [26,27] making them rather unattractive to routine analyses. Flow injection techniques have offered invaluable analytical tools for over thirty years in terms of rapidity, cost-effectiveness, precision and accuracy. Recent reports on FI to the analysis of sulfites include a variety of approaches based on amperometry [28,29], chemiluminescence [30], fluorescence [31], conductivity [32] and most commonly spectrophotometric detection [33–36]. A potential disadvantage of FI is the complicated, multi-channelled manifolds that are often proposed by analytical scientists. This is particularly evident in the case of sulfite, where a gas-diffusion step is

P.D. Tzanavaras et al. / Talanta 77 (2009) 1614–1619

also incorporated in the majority of the assays [28,29,31–36]. Fourchannelled [32,35], five-channelled [31] or even six-channelled setups [33,36] have therefore been recommended. To overcome this limitation Ruzicka and Marshall introduced sequential injection analysis, where all necessary analytical steps are carried out in a practically single-channelled mode [37]. To our knowledge, only two SI methods have been reported so far for the determination of sulfite. The first is based on the well-known formaldehyde–pararosaniline system with photometric detection [38] and the second on the use of a boron-doped diamond electrode and amperometric detection [39]. Both assays employ on-line gasdiffusion while the former utilizes three pumps [38] and the latter a continuously flowing acceptor stream to avoid disturbances of the electrochemical detector from the discontinuous operation of SI [39]. The scope of the present study is to propose a new automated method for the determination of total sulfite in wines. The method is based on a promising batch chemical system in terms of both selectivity and sensitivity [40]. The analyte reacts with ophthalaldehyde (OPA) in the presence of ammonia in phosphate buffer medium. Upon addition of NaOH to the reaction mixture, a deep blue product is formed having an absorption maximum at 630 nm. In order to accomplish this two stage process, a hybrid SI–FI system was utilized. Sulfite was separated from the wine matrix through an on-line gas-diffusion process incorporated in the SI manifold, followed by reaction with OPA in the presence of ammonia. The reaction mixture was merged on-line with a continuously flowing of NaOH prior to detection. The proposed method offers considerable advantages for routine analysis: automation through a practically two-channelled manifold; simple and low cost instrumentation including photometric detection; commercially available reagents; no complicated sample pretreatment prior to analysis; adequate sensitivity and linearity for this kind of determination; selectivity against potential interfering species and acceptable sampling rate. 2. Material and methods 2.1. Instrumentation A schematic diagram of the hybrid SI–FI manifold used is shown in Fig. 1. It was comprised of the following parts: a micro-

1615

electrically actuated 10-port valve (Valco, Switzerland); a Tecator (Hoganas, Sweden) 5023 FIAstar double-beam spectrophotometer equipped with a 1-cm long flow cell of 18-␮l internal volume; a Gilson (Minipuls3, France) peristaltic pump for the SI steps; a peristaltic pump of a Tecator 5010 FI analyzer for delivering the NaOH solution; a ChemifoldTM Type V gas-diffusion unit (Tecator, Sweden). The hardware was interfaced to the controlling PC through a multi-function I/O card (6025 E, National Instrument, Austin, TX). The control of the system and the data acquisition from the detector were performed through a special program developed in house using the LabVIEW 5.1.1 instrumentation software package (National Instrument, Austin, TX). PTFE gas-diffusion membranes were also provided by Tecator and were typically replaced on a weekly basis. The flow system used 0.5 mm i.d. Teflon tubing throughout, while Tygon pump tubes were used for aspirating/delivering the solutions. 2.2. Reagents All reagents were of analytical grade and provided by Merck (Darmstadt, Germany) unless stated otherwise. All other reagents used for interferences experiments were purchased from either Sigma–Aldrich or Merck and were salts of analytical grade as well. Ultra-pure quality water, produced by a Milli-Q system (Millipore, Bedford, US) was used for preparation of all the solutions. The stock sulfite solution ( (SO3 2− ) = 1000 ␮g mL−1 ) was prepared daily in 0.1 mol L−1 NaOH. Working standards were prepared by appropriate dilutions of the stock solution also in 0.1 mol L−1 NaOH in order to match the alkalinity of the samples after treatment. The OPA working solution (c (OPA) = 1 × 10−2 mol L−1 ) was prepared by dispersing an accurately weighed amount of the reagent in 100 ␮L MeOH and diluting to 10 mL with de-ionized water. The resulting solution was ultra-sonicated for 5 min to facilitate complete dissolution. The reagent was found to be stable for 1 week if kept refrigerated and protected from the light. The stock NH4 Cl solution (c (NH4 Cl) = 0.1 mol L−1 ) was prepared in de-ionized water. Working solutions were prepared by dilution of the stock in 7.5 × 10−2 mol L−1 phosphate buffer (pH 8.5). A 1 mol L−1 NaOH stock solution was prepared by dissolving the appropriate amount of NaOH pellets in de-ionized water. HCl stock solution (3 mol L−1 ) was prepared by proper dilution of the concentrated acid ( = 1.19, w = 37%) in water. 2.3. SI–GD procedure for aqueous solutions

Fig. 1. Schematic depiction of the hybrid SI–FI setup: C: water as carrier; PP: peristaltic pump; HC: holding coil; MPV: multi-position valve; S: sample; HCl: 1.0 mol L−1 HCl solution; R1 : OPA (c = 1 × 10−2 mol L−1 ); R2 : ammonium chloride (c = 2 × 10−2 mol L−1 )/phosphate buffer (pH 8.5, 7.5 × 10−2 mol L−1 ); W: waste; MC: mixing coil (30 cm/0.5 mm i.d.); GDU: gas-diffusion unit; TS: thermostat; RC1 and RC2 : reaction coils (60 cm/0.5 mm i.d.) and D: spectrophotometric detector (max = 630 nm).

The optimum SI sequence is shown in Table 1. In brief, each cycle began with filling of the acceptor stream of the gas-diffusion unit with the ammonium chloride–buffer mixture (steps 1–6). Sampling involved delivering of a “sandwiched” acidified sample zone through the donor line (steps 7–14). During the detection step, OPA reagent was aspirated and the reaction mixture was propelled at the confluence point where it was merged with the continuous-flowing NaOH stream and driven towards the spectrophotometric detector (steps 15–18). The duration of the cycle was 360 s, corresponding to a sampling rate of 10 h−1 . At the beginning and end of a working day all ports and lines of the SI manifold were flushed with 3 mL of de-ionized water. It should be noted that when changing between samples, an additional washing step was performed in order to avoid carryover effects; 3 × 200 ␮L of the new sample/standard was aspirated to the HC, and then flushed through port 7 to waste (W).

1616

P.D. Tzanavaras et al. / Talanta 77 (2009) 1614–1619

Table 1 SI sequence for the determination of sulfite. Valve position

Pump action

Flow rate (mL min−1 )

Action description

1 10 1 10 1 20

6 6 1 1 7 7

Off Aspirate Off Deliver Off Deliver

– 1.8 – 0.9 – 1.8

Selection of acceptor port Aspiration of acceptor (R2 ) in the HC Selection of detector port Filling the acceptor groove of the GDU Selection of waste port Flushing of the HC

Sampling 7 8 9 10 11 12 13 14

1 5 1 10 1 5 1 100

4 4 3 3 4 4 2 2

Off Aspirate Off Aspirate Off Aspirate Off Dispense

– 0.6 – 1.2 – 0.6 – 0.4

Selection of HCl reagent port Aspiration of HCl in the HC Selection of sample port Aspiration of sample in the HC Selection of HCl reagent port Aspiration of HCl in the HC Selection of donor port Propulsion of mixture to the donor channel

Detection 15 16 17 18

1 7.5 1 180

5 5 1 1

Off Aspirate Off Deliver

– 0.6 – 0.2

Selection of OPA port (R1 ) Aspiration of OPA in the HC Selection of detector port Propulsion of mixture to the detector

Step a/a 1 2 3 4 5 6

Time (s)

2.4. Analysis of wine samples The method was applied to the analysis of white and red wines. 5 mL of each sample was treated with an equal volume of 1 mol L−1 NaOH in order to liberate bound sulfites from the wine matrix, since due to their nucleophilic character they tend to react with endogenous wine components such as aldehydes [41–43]. The mixture was diluted to a final volume of 50 mL with de-ionized water prior to injection in the SI system. 3. Results and discussion 3.1. Preliminary experiments Preliminary experiments were carried out in order to investigate whether the chemical system proposed by Abdel-Latif for the determination of sulfite [40] could be automated using sequential injection analysis. The experiments confirmed that the sulfite–OPA–ammonium reaction mixture had to be merged with a highly alkaline NaOH solution in order to achieve adequate sensitivity in the visible region (max = 630 nm). However, effective alkalization could not be performed under sequential injection conditions, in a practically single-channelled mode. It was therefore necessary to develop a hybrid sequential–flow injection manifold, by incorporating a continuously flowing NaOH stream prior to spectrophotometric detection, as can be seen in Fig. 1. Although the method proposed by Abdel-Latif was claimed to be highly selective, a gas-diffusion step was introduced in the automated setup in order to avoid potential interferences from the wine samples matrix, including color and amino acids. It should be noted that Abdel-Latif in order to compensate the effect of such interferences used the first derivative of the absorption spectrum [40]. 3.2. Study of the color-forming reaction The various instrumental and chemical variables that affect the chemical reaction were investigated without the gas-diffusion step – using the univariate approach – at 10 mg L−1 sulfites. A hybrid SI–FI manifold similar shown in Fig. 1 was used throughout the study, adopting a three-zones configuration. In brief, sample, ammonium-buffer and OPA zones were aspirated sequen-

tially in the holding coil and propelled towards the confluence point where they were merged with the NaOH stream. The final reaction product was monitored spectrophotometricaly at 630 nm. The starting values of the studied variables were: T = 25 ◦ C; V (sample) = V (ammonium-buffer) = V (OPA) = 50 ␮L; qV (detector) = 0.6 mL min−1 ; c (OPA) = 5 × 10−3 mol L−1 ; l (RC2 ) = 60 cm/ 0.5 mm i.d.; c (NH4 Cl) = 1 × 10−2 mol L−1 ; pH 8.5 (5 × 10−2 mol L−1 phosphate buffer) and c (NaOH) = 0.5 mol L−1 . It should be noted that the length of the thermostated reaction coil (RC1 ) was kept at 60 cm/0.5 mm i.d. and the flow rate of the NaOH stream at 0.2 mL min−1 throughout this study. The effect of temperature on the reaction was investigated in the range of 25–70 ◦ C. The experimental results are depicted in Fig. 2A. As can be seen, an almost linear increase in the signals was observed within the studied range. Higher temperatures produced irreproducible results due to bubbles formation even if a restriction coil was used at the detector waste-line. 70 ◦ C was therefore selected for further studies. The effect of the flow rate of the reaction mixture towards the confluence point prior to detection was studied in the range of 0.6–0.2 mL min−1 . The experimental results are shown in Fig. 2B. The signals increased by ca. 165% by decreasing the flow rate within the studied range. The value of 0.2 mL min−1 was selected for subsequent experiments. The reaction coil RC2 determines the contact-time between the reaction mixture and the NaOH solution. Its effect was studied between 30 and 90 cm (0.5 mm i.d. in all cases). Maximum absorbance was observed at 60 cm, while longer coils resulted in lower absorbance values due to pre-domination of dispersion effects. The effect of the amount concentration of NaOH was studied in the range of 0.25–1.0 mol L−1 . No significant variations were recorded within this range. The value of 0.5 mol L−1 was selected for further experiments. The amount concentration of ammonium chloride had a more marking effect on the sensitivity of the developed method. An almost linear increase in absorbance was obtained between 0.5 and 2 × 10−2 mol L−1 , while the phenomenon was less pronounced in the range of 2–5 × 10−2 mol L−1 . The value of 2 × 10−2 mol L−1 was preferred for subsequent studies. The effect of the pH on the reaction was investigated in the range of 7.0–10.0 using 5 × 10−2 mol L−1 phosphate buffer in all

P.D. Tzanavaras et al. / Talanta 77 (2009) 1614–1619

1617

3.3. Study of the gas-diffusion step

Fig. 2. Study of instrumental and chemical variables: (A) effect of temperature; (B) effect of the acceptor flow rate and (C) effect of the amount concentration of OPA. For experimental details see Section 3.2.

cases. A plateau was obtained in the range of 8.0–9.0, while the signals decreased for lower and higher pH values. This behavior is an advantage of the proposed assay, since no strict pH control is required. The ionic strength of the buffer had a moderate effect on the signals in the range of 1–5 × 10−2 mol L−1 and practically no effect up to 0.1 mol L−1 . The value of 7.5 × 10−2 mol L−1 was selected for further experiments. Variation of the amount concentration of OPA between 0.2 and 1 × 10−2 mol L−1 resulted in a non-linear increase in the measured absorbance (Fig. 2C). Higher OPA amount concentrations resulted in the formation of blue precipitates within the flow lines, a phenomenon that has also been previously observed in flow injection manifolds [31]. These precipitates could be removed effectively only by flushing the system with acetone. The amount concentration of 1 × 10−2 mol L−1 was selected for subsequent investigations. Finally, the effect of the volume of the OPA reagent was studied in the range of 25–100 ␮L. Steady state in absorbance was observed within the range of 50–100 ␮L, with a small decrease at 25 ␮L. A volume of 75 ␮L was preferred for further investigations.

The gas-diffusion step was studied using the hybrid SI–FI setup shown in Fig. 1 and under the selected values of the variables mentioned in the previous section. The investigated variables were the flow rate of the donor stream, the sample injection volume, the volume and amount concentration of the HCl solution and the length of the mixing coil (MC). It should be noted that in order to achieve efficient acidification of the sample, the latter was “sandwiched” between two HCl zones of equal volumes. The starting values of these variables were: qV (donor) = 0.6 mL min−1 ; V (sample) = 50 ␮L; V (HCl) = 2 × 50 ␮L; c (HCl) = 1.0 mol L−1 and l (MC) = 60 cm/0.5 mm i.d. The flow rate of the donor stream is a critical parameter in on-line gas-diffusion procedures since it determines the time contact between the sample and the semi-permeable membrane and therefore the efficiency of the diffusion process. Its effect was investigated in the range of 0.2–0.6 mL min−1 . Higher sensitivity was obtained at a flow rate of 0.2 mL min−1 , while no significant differences were observed in the range of 0.3–0.6 mL min−1 . A flow rate of 0.4 mL min−1 was selected as a compromise between sensitivity and sampling frequency. The effect of the sample injection volume was studied in the range of 50–300 ␮L. The experiments showed an almost linear dependence between the volume and the measured absorbance. The value of 200 ␮L was preferred in terms of sensitivity and sampling rate. On the other hand, no significant variations were obtained for HCl volumes in the range from 100 (2 × 50) to 200 (2 × 100) ␮L. The volume of 100 ␮L was therefore selected for subsequent experiments. Another potentially critical parameter affecting the gasdiffusion process is the amount concentration of HCl. Since the standards and wine samples contain a final NaOH amount concentration of 0.1 mol L−1 , a sufficiently acidic environment must be ensured in order to promote the formation of gaseous sulfur dioxide. The experiments showed a ca. 40% increase on sensitivity in the range of 0.2–1.0 mol L−1 HCl, while leveling-off was obtained thereafter. The latter value was selected for further studies. Finally, the effect of the length of the MC was studied in the range of 30–90 cm. The length of the mixing coil determines the extent of overlapping between the sample and HCl zones and, therefore, the efficiency of the formation of sulfur dioxide and gas-diffusion step. However, the experiments showed no variations within the investigated range. 30 cm was selected as it provided adequate overlapping of the zones. 3.4. Method validation The developed method was validated in terms of linearity, limits of detection and quantitation, precision, selectivity and accuracy using the SI setup shown in Fig. 1. 3.4.1. Linearity Linearity was obeyed in the range of 1–40 mg L−1 sulfite following the regression equation A = 13.7 (±2.3) + 12.6 (±0.1) (sulfite) where A is the absorbance as measured by the flow through detector in mA U and  (sulfite) is the mass concentration of the analyte in mg L−1 . The validity of the calibration curve was evaluated by the value of the regression coefficient, the distribution of the residuals and response factor (R.F.) test [44]. The acceptance limit for the regression coefficient was a value of >0.999, while for the response factor test was set to be within ±5% of the experimental slope as

1618

P.D. Tzanavaras et al. / Talanta 77 (2009) 1614–1619 Table 3 Accuracy of the proposed assay.

Table 2 Study of interferences. Interfering species

Specie/analyte mass concentration ratio

Percent recovery, R (%)a (20 mg L−1 SO3 2− )

Glycine Tartaric, citric, oxalic, malic, malonic, ascorbic, boric SO4 2− , NO3 − , F− , SCN− , CH3 COO− , CO3 2− Ca(II), Mg(II), Zn(II), Fe(III), Cu(II) CN− Ethanol

100 100

101 97–102

100

96–102

50

97–103

a b

10 20%b

Wine samplea , b

Sulfite added (mg L−1 )

Sulfite found (mg L−1 )

Recovery, R (%)

White A

5.0 10.0 15.0

5.3 9.9 15.3

106 99 102

White B

5.0 10.0 15.0

5.0 10.2 15.5

100 102 103

Red A

5.0 10.0 15.0

4.8 10.1 14.7

96 101 98

Red B

5.0 10.0 15.0

5.1 9.7 15.0

102 97 100

102 99

The acceptance limit is 95–105%. Corresponds to a volume fraction of 20% in the test sample. a

given by the equation: R.F. =

peak height (mA U) − intercept  (sulfite)

The regression coefficient was 0.9996. The values of the R.F. ranged between −4.8% (at 2.0 mg L−1 sulfite) and +1.3% (at 25 mg L−1 sulfite), verifying the validity of the regression line. The plot of the residuals indicated random distribution around the “zero” line. 3.4.2. Limits of detection and quantitation The limits of detection and quantitation (cL and cQ ) of the proposed assay were calculated according to the IUPAC recommendations, as 3.3sb /m and 10sb /m, respectively, where sb is the standard deviation of the blank measurements (n = 8), and m is the slope of the calibration graph [45]. On this basis, the cL and cQ of the proposed method were calculated to be 0.3 and 1.0 mg L−1 sulfite, respectively. These values are adequate for the determination of the analyte in real wine samples. 3.4.3. Precision The within-day precision of the proposed method was evaluated by repeated injections (n = 12) of a sulfites standard solution at the 20 mg L−1 level. The relative standard deviation was 2.2%. The between-day precision for five consecutive at the same concentration level was less than 8%. 3.4.4. Study of interferences The selectivity of the assay against several potentially interfering species was examined by analyzing mixtures of the species with the analyte at a sulfite mass concentration of 20 mg L−1 . The study of interferences was focused on metals, anions and organic acids that may co-exist with sulfites in real samples. The experimental results are shown in Table 2. As OPA is a well-known reagent for amino acids, glycine was investigated as a model compound at a maximum mass ratio of 100:1. The incorporation of the gas-diffusion step in the method, offered sufficient selectivity against such compounds. On the other hand, cyanide ions react with OPA [46] and are volatile under acidic conditions forming HCN. However, the experiments showed that cyanides are adequately tolerated at a 10:1 ratio. Finally, ethanol was found to be tolerable at satisfactory volume fraction of 20% that is above the usual percentage in wine samples. It should be noted that the acceptance criterion for non-interference was set at percent recoveries in the range of 95–105% compared to a sulfite standard solution at the same mass concentration. 3.4.5. Accuracy The accuracy of the developed method was validated by recovery experiments after spiking wine and red wine samples (diluted

b

Mean of three results. Spiked after 1:10 sample dilution.

Table 4 Determination of total sulfites in white and red wine samples. Wine sample

Sulfite found by SI (mg L−1 )a ± SD

White A White B White C White D Red A Red B Red C

114 132 101 112 63 91 88

a b

± ± ± ± ± ± ±

3 5 2 4 2 2 3

Reference methodb (mg L−1 ) 108 136 105 116 60 89 92

Mean of three results. Based on iodimetric titration [47].

1:10) with known amounts of sulfite standards, in the range of 5–15 mg L−1 . These experiments included in Table 3, verified the accuracy of the proposed assay since the calculated percent recoveries ranged between 96% and 106%. 3.5. Analysis of wine samples The applicability of the developed SI method was evaluated by analyzing white and red wine samples for total sulfite. Samples were treated as described in Section 2.4 and analyzed using the sequence as shown in Table 1. The experimental results are included in Table 4. The total sulfite content ranged between 63 and 132 mg L−1 . The validity of the results was confirmed by comparison of the found values to those derived by the EU recommended iodine-based titrimetric approach [47]. Good agreement was observed in all cases, as the relative errors were in the range of −4.5% to +5.2%. As expected, higher total sulfite values were found in white wines, which is in accordance with the literature [1,2]. 4. Conclusions The present study reports a new automated assay for the determination of total sulfite in white and red wines by incorporation of a gas-diffusion step in a hybrid SI–FI manifold and spectrophotometric detection. The proposed method offers adequate sensitivity (cL = 0.3 mg L−1 ) and linear determination range (up to 40 mg L−1 ), no complicated procedures prior to analysis, acceptable sampling rate of 10 h−1 and readily available instrumentation and reagents. Thorough validation in terms of precision, accuracy and selectivity confirmed the applicability of the developed analytical assay. The percent recoveries from the analysis of real samples were in the range of 96–106%.

P.D. Tzanavaras et al. / Talanta 77 (2009) 1614–1619

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

L. Pizzoferrato, G. Lullo, E. Quattucci, Food Chem. 43 (1998) 275. C.M.N. Azevedo, K. Araki, H.E. Toma, L. Agnes, Anal. Chim. Acta 387 (1999) 175. A.F. Gunnison, D.W. Jacobsen, CRC Crit. Rev. Toxicol. 17 (1987) 185. H. Valley, P.J. Thompson, Thorax 56 (2001) 763. U.T. Yilmaz, G. Somer, Anal. Chim. Acta 603 (2007) 30. S. McLeod, D.E. Davey, Anal. Chim. Acta 600 (2007) 72. M.H. Pournaghi-Azar, R.E. Sabzi, Electroanalysis 16 (2004) 860. T. Garcia, E. Casero, E. Lorenzo, F. Pariente, Sens. Actuator B 106 (2005) 803. D.R. Shankaran, K.-I. Iimura, T. Kato, Electroanalysis 16 (2004) 556. R. Carballo, V.C. Dall’Orto, A. Lo Balbo, I. Rezzano, Sens. Actuator B 88 (2003) 155. M. Zhao, D.B. Hibbert, J.J. Gooding, Anal. Chim. Acta 556 (2006) 195. A.K. Abass, J.P. Hart, D. Cowell, Sens. Actuator B 62 (2000) 148. M.K. Sezginturk, E. Dinckaya, Talanta 65 (2005) 998. G. Jankovskiene, Z. Daunoravicious, A. Padarauskas, J. Chromatogr. A 934 (2001) 67. M. Masar, M. Dankova, E. Olvecka, A. Stachurova, D. Kaniansky, B. Stanislawski, J. Chromatogr. A 1084 (2005) 101. M. Masar, M. Dankova, E. Olvecka, A. Stachurova, D. Kaniansky, B. Stanislawski, J. Chromatogr. A 1026 (2004) 31. X.-F. Yang, X.-Q. Guo, Y.-B. Zhao, Anal. Chim. Acta 456 (2002) 121. Z. Xin-qi, L. Dong-hui, Z. Hong, X. Jin-gou, Anal. Chim. Acta 448 (2001) 71. R.L. Bonifacio, N. Coichev, Anal. Chim. Acta 517 (2004) 125. J. Cmelik, J. Machat, E. Niedobova, V. Otruba, V. Kanicky, Anal. Bioanal. Chem. 383 (2005) 483. S.K. Verma, M.K. Deb, J. Agric. Food Chem. 55 (2007) 8319. B. Palenzuela, B.M. Simonet, A. Rios, M. Valcarcel, Anal. Chim. Acta 535 (2005) 65. R.C. Rodriguez-Diaz, M.P. Aguilar-Caballos, A. Gomez-Hens, J. Agric. Food Chem. 52 (2004) 7777.

1619

[24] S.S.M. Hassan, M.S.A. Hamza, A.H.K. Mohamed, Anal. Chim. Acta 570 (2006) 232. [25] M.M.S. Abdul Galil, M.S. Yogendra Kumar, M.A. Sathish, G. Nagendrappa, J. Anal. Chem. 63 (2008) 239. [26] K.R.B. Silva, I.M. Raimundo, I.F. Gimenez, O.L. Alves, J. Agric. Food Chem. 54 (2006) 8697. [27] S. de Marcos, N. Alcubierre, J. Galban, J.R. Castillo, Anal. Chim. Acta 502 (2004) 7. [28] T.J. Cardwell, M.J. Christophersen, Anal. Chim. Acta 416 (2000) 105. [29] D. Corbo, M. Bertotti, Anal. Bioanal. Chem. 374 (2002) 416. [30] D. He, Z. Zhang, Y. Huang, Anal. Lett. 38 (2005) 563. [31] H. Mana, U. Spohn, Anal. Chem. 73 (2001) 3187. [32] S. Faldt, B. Karlberg, W. Frenzel, Fresenius J. Anal. Chem. 371 (2001) 425. [33] J. Carinhanha, C. Santos, M. Korn, Microchim. Acta 153 (2006) 87. [34] D. Melo, E.A.G. Zagatto, I.L. Mattos, N. Maniasso, J. Braz. Chem. Soc. 14 (2003) 375. [35] M.A. Segundo, A.O.S.S. Rangel, A. Cladera, V. Cerda, Analyst 125 (2000) 1501. [36] G.T. Atanassov, R.C. Lima, R.B.R. Mesquita, A.O.S.S. Rangel, I.V. Toth, Analusis 28 (2000) 77. [37] J. Ruzicka, G.D. Marshall, Anal. Chim. Acta 237 (1990) 329. [38] M.A. Segundo, A.O.S.S. Rangel, Anal. Chim. Acta 427 (2001) 279. [39] C. Chinvongamorn, K. Pinwattana, N. Praphairaksit, T. Imato, O. Chailapakul, Sensors 8 (2008 1846). [40] M.S. Abdel-Latif, Anal. Lett. 27 (1994) 2601. [41] R. Walker, Food Addit. Contam. 2 (1985) 5. [42] B.L. Wedzicha, Chem. Br. 27 (1991) 1030. [43] J.B. Adams, Food Chem. 59 (1997) 401. [44] J.M. Green, Anal. Chem. 68 (1996) 305A. [45] J. Mocak, A.M. Bond, S. Mitchell, G. Scollary, Pure Appl. Chem. 69 (1997) 297. [46] Y. Nomura, K. Nagakubo, H.S. Ji, A. Watanabe, T. Akimoto, S. Mcniven, K. Hayashi, Y. Arikawa, I. Karube, Environ. Sci. Technol. 34 (2000) 2618. [47] E. Mataix, M.D. Luque de Castro, Analyst 123 (1998) 1547.