Figure 1a shows cyclic voltammograms of a Fe-Tmp modified electrode in acetate buffer solution (pH 4.0, 0.1 M) in absence and presence of hydrogen peroxide. A defined redox couple (E^o’ = −0.09 V), corresponding to Fe (III)’/Fe (II)’ system can be observed. As shown (Figure 1b–i) an increment in the concentration of H₂O₂ generates an increment in the cathodic peak height while in the anodic peak a decrease is observed. The catalytic currents increased linearly with the H₂O₂ concentration of the range of 0.1 to 2.6 mM using cyclic voltammetry. Lower concentration ranges can be achieved using amperometric detection. The reversibility, stability, reproducibility and fast electron transfer kinetics of the Fe-Tmp modified electrode are appropriate for its use as mediator during detection of hydrogen peroxide at low concentration levels at applied potential of −0.20 V.

The effect of pH on the electrocatalytic response was studied using Britton-Robinson buffer solution (0.1 M) in the concentration range from 2.0 to 10.0 containing 1.0 mM of H₂O₂. The cathodic current increased with increasing pH up to 4.0, and then remained constant between 4.0 and 5.0. At higher pH values a decrease in the cathodic peak is observed likely due to the iron hydrated oxide formation. Therefore, an optimal pH value of 4.0 (acetate buffer solution) was chosen. This pH value corresponds to the value used during enzymatic determination of oxalate using OxOx [9]. The influence of the applied potential to the working electrode on the catalytic response was evaluated using amperometric detection. The amperometric signal increased using applied potentials from 0.0 to −0.2 V and then remained constant at lower values (−0.3 to −0.5 V). An applied potential of −0.2 V was therefore selected in an effort to minimize the interference produced using lower detection potential.

Oxalate is determined according the following sequence of reactions (Figure 2). Several factors affecting the proposed system required optimization [20,21]. A Taguchi parameter design (TPD) was selected as the robust optimization method since this provides the necessary information with the minimal experimentation. This design discriminates between control factors and experimental noise through orthogonal arrays in which the columns (factors) and rows (trials) are arranged in a conveniently fixed manner. These matrices allow the simultaneous evaluation of several variables with the minimum number of trials. The results obtained were analyzed statistically to obtain the optimum of each variable.

Optimization of the system with TPD involves five steps, (a) identifying the output variable to optimize; (b) identifying and selecting factors that affect the system; (c) selecting the appropriate orthogonal array and assigning adequate settings to the chosen variables; (d) analyzing the data and determining the optimum settings and (e) conducting a confirmatory experiment under the optimal conditions obtained.

In analytical flow techniques with amperometric detection, the desired response is the maximum signal height (current, µA). The flow variables optimized in sequential injection systems are the enzyme concentration (aspirated volume of enzyme magnetic suspension), the flow rate of the carrier solution, the aspirated sample volume and the reactor length from the selection valve to the detector. The selected orthogonal array must have a number of columns equal to or higher than the number of degrees of freedom of the system; thus an L₉(3⁴) array was used. The three settings selected for each factor were chosen taking into account preliminary analyzes.

The flow rate (Q) has been shown in enzymatic flow systems that control the H₂O₂ production during the enzymatic reaction. SIA with enzymatic bead injection requires the use of flow rates <1.0 mL·min⁻¹ in order to increase the H₂O₂ generation [10]. The reactor length (RL) must be sufficiently long for conditioning the sample pH. With respect to the aspired volumes for sample (SV) and enzyme magnetic suspension (MV), they are critical in flow systems; therefore, those selected must be large enough to allow the correct dispersion of the sample with the carrier solution and OxOx concentration on the modified electrode.

Table 1 shows factorial design matrix, the settings for each experiments used for the optimization experiments and the results of the peak height for each trial. All experiments were performed in triplicate in order to calculate the residual error; the total number of experiments was therefore 27 (9 experiments × 3 replicates). Measurements were performed with solutions containing 50.0 mg·L⁻¹ of oxalic acid.

The results were analyzed statistically to adjust each variable to its optimum with least variable possible. All calculations were made using ANOVA-TM v2.5 software. Table 2 shows the results for this analysis. The values of the variance ratio (F) and the critical variance ratio (4.56, α = 0.05) show that all the factors taken into account were critical (F_calculated > F_critical). The factor with a greatest influence on the response was the enzyme magnetic suspension volume, which contributes 77.34% of the total variance, followed by the sample volume (16.82%). The contribution of the residual error was 0.01%, this indicates the correct selection of experimental parameters.

Figure 3 shows the effects of the control factors on the output variable, among which the aspired volumes are the most important (longer lines in comparison with the other variable). The combination of settings that allowed the highest output variable (highest response in Figure 3) was, MV, 10 µL; Q, 1.00 mL·min⁻¹; RL, 50 cm; SV, 200 µL. The flow rate used to aspirate the solutions (30 mL·min⁻¹) helps to promote a turbulent flow, which contributes the mixing of the sample and the carrier solution, increasing the sampling rate of the method.

Using the SIAgram obtained under optimal conditions (Figure 4), a standard curve for oxalate was constructed using the mean peak height (µA) values of each solution. Table 3 shows the regression parameters taken from this standard curve. The limit of detection was calculated according to IUPAC criteria [22] i.e., 3.29S_e/b1, where S_e is the square root of the residual variance of the standard curve, and b₁ is the slope. The intermediate precision of the procedure, expressed as the relative standard deviation (%RSD), for three determinations (made on different days) using synthetic samples with analyte concentration of 30 mg·L⁻¹ was 2.2%. The SIAgram was used to calculate the repeatability of the determinations; the RSD was below 2% for all standard solutions and samples-under optimum conditions 30 determinations per hour can be analyzed. The modified electrode was used for a week and the enzyme magnetic solid was renewed daily.

The interference possible compounds present in the urine samples such as: urea, uric acid, ascorbic acid and salicylic acid was studied. Solutions containing 20.0 mg·L⁻¹ of oxalate and the foreign compound at higher concentrations (maximum 100:1) were analyzed. The interference concentration of each compound was considered as the one which caused a variation in the response greater than or equal to ±5% compared to the response obtained in its absence. The results demonstrate that none of these compounds interfered in the determination. The limit of detection and the linear range confirmed that SIA-biosensor provides adequate sensitivity to determine oxalate at the normal interval (0.0–20.0 mg·L⁻¹ in healthy patients and >20.0 mg·L⁻¹ for stone formers) of concentration in urine samples [1,23].

The proposed method was used to determine oxalate in urine samples from apparently healthy male and female in three age groups: children (1−20 year), young (21−45 year) and old (>45 year), (n = 3 in each group). The results were compared to the flow injection analysis with spectrophotometric detection (FIA-Vis). For each sample, the mean oxalate concentrations (n = 3) obtained with the two methods were compared using the Student t test, assuming comparable variances (confirmed by an F test). The values of t_experimental were then compared to a t_tabulated at 95% confidence interval (t = 2.78, α = 0.05, n = 4). No significant differences were seen between the results provided by each method. Results show that urinary oxalate values were in range 9.2–42.8 mg·L⁻¹, in healthy male and female, respectively. These values are comparable to those reported in earlier studies [24].

Magnetite (0.80 g) synthesized by co-precipitation [25] is added to a reactor containing tetramethoxysilane (0.25 g) and 3-aminopropylmethoxysilane (1.25 g). The monomers were dissolved in 24 mL of a solution containing 2% (w/v) Triton X-100, 12.5% (v/v) methanol, 0.02% (w/v) cetyltrimethylammonium bromide and 200 µL NH₃ (28% w/v) as catalyst. The mixture was heated and refluxed for 16 h with stirring. The NH₂ modified magnetic solid was washed with 10.0 mL of each of the following solvents: ethanol and water. The obtained solid was dried at 60 °C for 24 h.

The OxOx was immobilized on the magnetic support as follows: the NH₂ modified magnetic solid (50 mg) was mixed with glutaraldehyde solution (2.0%, 5 mL). The mixture was stirred and after 2 h, the solid was washed with deionized water (previously deoxygenated, 20 mL) followed by phosphate buffer solution (pH 7.0, 0.1 M, 20 mL), also previously deoxygenated. The support was suspended in phosphate buffer solution (5.0 mL) containing OxOx (3.5 U, 500 µL) and the dispersion was stirred for 6 h. The magnetic suspension obtained was used in the sequential injection system.

Figure 5 shows the diffraction pattern (obtained in a Phillips PW1710 diffractometer, Almelo, Holland) for the magnetic solid shows reflections corresponding to magnetite (“m” in Figure 1). In addition, the presence of a wide band at 2θ angles from 15° to 25° is consistent with the amorphous phase of the silica [17]. The overall morphology of the magnetic solid was studied by SEM (Jeol JSM-820, Peabody, MA, USA). The solid has a spherical morphology (Figure 5b). The size distribution of the particles (dried at 100 °C for 24 h) was determined to be in the range of 50 to 150 nm based on the measurements of the sizes of particles.

A methanolic solution of FeCl₃·6H₂O (25 mg, 0.09 mmol, 5 mL) was added to a methanolic solution of K⁺[hydrotris(2-thiopyridone)borate]⁻ (68 mg, 0.09 mmol, 5 mL). A yellow brown precipitate was formed immediately [11]. The suspension was stirred for 12 h, the precipitate was extracted twice with chloroform (5 mL) and the organic phase was evaporated at room temperature. The resulting solid was washed with hexane (5 × 4 mL) [26]. Yield: 59 mg (0.07 mmol, 78%). Infrared characteristic vibrations: nB-H = 2,510; nC-H₃ = 2,850; nC-H_th = 2,920 cm⁻¹. The NMR studies suggested a paramagnetic nature for this complex.

The sequential injection system (Figure 6) consists of a burette multisyringe with programmable speed (S, microBu 2030, Crison, Barcelona, Spain) used to aspire and dispense the reagent solutions, an eight-way selection valve (V, Crison) and a bipontentiostat (D, Drop Sens μSTAT 200) as detector. The different components of the flow manifolds were connected using Omnifit Teflon tubing (0.8 mm i.d.) and Gilson connectors. The instrumental devises were controlled by Autoanalysis 5.0 software.

The analytical cycle began with the aspiration of enzyme magnetic suspension (MS, 10 µL) onto the loading reactor (R1, 100 cm) at a flow rate of 30 mL·min^–1. This was then directed towards the detector, subjected to an external magnetic field, by the carrier solution (CS). These steps were repeated to renew the recognition agent on the electrode surface. A neodymium magnet (Nd-Fe-B, 3.0 mm × 4.0 mm d.e., 1.3 T) was placed at the bottom of the working electrode. The magnet was removed during renew of the recognition agent.

Once the biosensor has been formed, sample solution (200 µL) was aspired. The sample solution (S) was mixed with the carrier solution in the reactor 2 (R2, 50 cm) and detected at −0.2 V using the modified electrode at 1.00 mL·min⁻¹. The screen printed modified carbon electrode (℘ = 4 mm) was performed by impregnating 20 μg of Fe(III)-Tpm (in acetonitrile) on the surface of working electrode (Drop Sens 110A) and drying overnight at room temperature. A wall-jet electrochemical cell (Drop Sens DRP-FLWCL) was used as detector.

The concentration of oxalate in the analyzed samples was also determined for comparative purposes using flow injection analysis with spectrophotometric detection (FIA-Vis), this method was previously validated to analyze oxalate in urine samples. The FIA system consists of a four-channel Gilson Minipuls 2 peristaltic pump fitted with propulsion tubes of the same brand used to propel sample and carrier solutions. A sample volume of 250 µL was injected into the Fe (III) solution (1.0 × 10^–3 M, in H₂SO₄ 0.01 M) using a Rheodyne 5020 four-way rotary valve. The reduction of the Fe (III)-oxalate complex takes place in the photochemical reactor (250 cm × 0.8 mm i.d.). The Fe (II) produced reacts with the ferrozine solution (1 × 10⁻³ M, in acetate buffer solution pH 4.0, 0.2 M) in a reaction coil (25 cm × 0.8 mm i.d.). The Fe (II)-ferrozine complex was detected at 562 nm is an UV-visible spectrophotometer (Lambda 40, Perkin-Elmer, Norwalk, CT, USA) with a Hellma 178.712QS 18 µL flowthrough cell. The flow rate of each individual channel was 0.6 mL·min⁻¹ [27].

Urine samples (5.0 mL) were mixed with saturated CaCl₂ solution (2.0 mL). The mixture was centrifuged 5 min at 2,500 rpm. The liquid phase was discarded and the precipitate was dissolved in H₂SO₄ 0.1 M (1.0 mL). The solution obtained was injected into the FIA-Vis manifold for its analysis. Oxalate concentration was determined by interpolation using an oxalate standard curve constructed in the interval from 10.0 to 50.0 mg·L⁻¹.