Quantitative Determination of Some Water-Soluble B Vitamins by Kinetic Analytical Method Based on the Perturbation of an Oscillatory Reaction

Um novo procedimento para a determinação da cinética de algumas vitaminas solúveis em água do grupo B (tiamina (B 1 ), riboflavina (B 2 ), niacina (B 3 ) e piridoxina (B 6 )), pelas perturbações da concentração do sistema químico oscilatório de Bray-Liebhafsky (BL), na presença de íons iodeto e de hidrogênio é proposto e validado. O método usa um eletrodo de Pt para o monitoramento potenciométrico das perturbações de concentração da matriz BL em um estado estacionário de não equilíbrio, estável, próximo ao ponto de bifurcação. O método proposto baseia-se na relação linear entre as diferenças de potencial máximas, DE m , causadas pelas quantidades adicionais conhecidas da espécie B. Sob condições analíticas ótimas, curvas de calibração lineares foram obtidas no intervalo de 0,01-1,0; 0,016-0,128; 5,0-50,0 e 0,05-2,5 mmol com limites de detecção de 0,01; 0,018; 2,6 e 0,03 mmol, assim como velocidade analítica de 30, 5, 12 e 20 determinações por hora, para B 1 , B 2 , B 3 e B 6 , respectivamente. A abordagem técnica usada também proporciona um método simples, efetivo e conveniente para o ensaio de formulações farmacêuticas contendo B 1 em conjunto com outros princípios ativos como a nicotinamida e vitamina B 12 , bem como B 3 .


Introduction
A considerable number of publications in which B vitamins were determined individually and simultaneously in pharmaceutical dosage forms using many methods that were developed with largely linear range and lower detection limit including kinetic analytical method having H 2 O 2 -KSCN-CuSO 4 and Mn(II)-BrO 3 --diacetone-H 2 SO 4 oscillatory reaction systems as matrices, 1,2 were found.Thus, on one hand, there is necessity for micro-quantitative analysis of vitamins from B-group, and, on the other hand, the examinations which related to the kinetic method based on oscillatory reactions are the appropriate one for this purpose.Therefore, a current new trend in study of Vol.22, No. 1, 2011   oscillatory reaction system is its application to analytical chemistry, as well as its particularly analytical use to determine various biologically and pharmaceutically important compounds [1][2][3][4][5][6][7][8][9][10][11][12][13][14] which could be of benefit to life sciences in the future.As an analytical method it has been recognized more and more to be useful and convenient.  In s field, the using of analyte pulse perturbation (APP) technique in both oscillatory 21 and stable steady state in vicinity of bifurcation point, [7][8][9][10][11][12][13] the uses of the largest Lyapunov exponent 22 and the high-sensitive oscillating chemical system 17 make the technique almost perfect and consequently favorable to use it in the really routine analysis.
The first above-mentioned method 21 is based on the relationship between the concentrations of analyte and the response of the matrix in the oscillatory state with respect to the main characters of oscillations, such as amplitude, period and others.For both analyzed vitamin B 6 (pyridoxine) 1 and vitamin B 2 , 2 the amplitude of the second response cycle as well as the change of the oscillating period (Dt), were correlated with the amount of analyte added.In the second method [7][8][9][10][11][12][13] based on perturbing the matrix that was in a stable stationary state in the vicinity of a bifurcation point, only relationship between the maximal potential displacement (DE m ) in the moment of perturbation and the logarithm of analyte concentrations would be analyzed.When using the matrix that is in a stable steady state, it is not necessary to test oscillatory phases nor to perturb the matrix always in the same selected oscillatory phase point, which turns out to be a very delicate moment.Compared with the matrix being in the oscillatory state, the regeneration of the matrix being in the stable non-equilibrium stationary state (stable steady state) is considerably shorter.
This paper describes the method for the quantitative determination of some water-soluble B vitamins in the BL matrix generated in continuously fed well stirred tank reactor (CSTR) [23][24][25][26] by electrode potential measurements.For this purpose, the BL matrix, 27,28 in a stable steady state near a bifurcation point is perturbed with variable amounts of vitamins, which results in substantial changes in the potential of the matrix dynamic state that are relevant to the concentration of an analyte.In particular, the aim of this work is to demonstrate that the mentioned kinetic method can be successfully applied for quantitative determination of the mentioned vitamins in bulk drug and pharmaceutical formulations.
0][31][32][33][34] The wide linear range (of about two order of magnitude; occasionally more than two) and low detection limit (ca.10 −6 -10 −7 mol L −1 , occasionally down to 10 −8 mol L −1 ) of the above-mentioned methods satisfy the requirements of most determinations.However, the determination of water-soluble vitamins has always been an unpleasant problem mostly due to instability of these substances and complexity of the matrices in which they usually exist.In addition, some of the proposed analytical methods are cumbersome, time-consuming or not enough accurate.Therefore, rapid methods based on a relatively simple and inexpesive equipment is desirable.In this sense, the described kinetic analytical method for quantitative determination of vitamins provides a promising alternative to some instrumental methods due to its low cost instrumentation and rapid detection procedure.The main advantage of the proposed method is its simplicity of operation, in fact, it requires no derivatization reaction nor any time-consuming extraction procedure.In addition, our method involved neither sophisticate instruments nor uncommon expensive reagents.As results, it can be implemented with the modular equipment available at any laboratory.
Standard stock solutions of vitamins B 1 , B 6 and B 3 were prepared at a concentration 1.0 × 10 -1 mol L -1 (B 1 and B 6 ) and 1 mol L -1 (B 3 ) by dissolving 0.8432 g, 0.4229 g and 3.0533 g, respectively of the pure substances in a 25 mL volumetric flask with water.Just because vitamin B 2 has poor solubility and a slow soluble speed, a vitamin B 2 solution (2.65 × 10 -4 mol L -1 ) was prepared by dissolving 0.0997 g of B 2 in a 1000 mL volumetric flask with water under ultrasonication (Bandelin Sonorex, Germany) for 5 min and thermostated at 60 °C for the same time in order to accelerate dissolving speed.These solutions were kept Quantitative Determination of Some Water-Soluble B Vitamins J. Braz.Chem.Soc.40   in brown volumetric flasks and stored in refrigerator in the dark.Prior to injection, stock solutions of each analyte were appropriately diluted with water before being used as working solutions.
Three pharmaceutical formulations containing vitamins B-group, vitamin B 1 tablets (sample 1) and vitamin B 3 tablets (sample 3) that were obtained from Now Foods, Bloomingdale, IL, 60108, USA as well as vitamin B-complex tablets (sample 2) was obtained from Srbolek, Belgrade, Serbia, bought in Serbian pharmacies, and analyzed by the procedure proposed.
Samples 1 and 3 contain 100 mg of thiamine and 500 mg of niacin, respectively.B-complex vitamin tablet (sample 2) contain 12.3 mg of B 1 , 5 mg of B 6 , 0.05 mg of vitamin B 12 , 0.4 mg of folic acid, 20 mg of nicotinamide and excipients.Aqueous stock solutions of the following substances (compounds) or excipients were also prepared for interference study: ascorbic acid, folic acid, citric acid, uric acid, carbamide, glucose, saccharine, fructose and starch.

Apparatus
The instrumental set-up used to implement the BL reaction, as the matrix, for the determination of analytes is shown schematically by Pejić et al. 10 The oscillating assembly is composed of a 50 mL glass CSTR vessel (Metrohm model 876-20, Herisau, Switzerland) wrapped in a water recirculation jacket connected to a thermostat (series U8, MLW Freital, Germany) with an accuracy of ± 0.1 °C.For homogenization of the reaction mixture, a magnetic stirrer (Combimag RET, Staufen, Germany) was used.Peristaltic pumps (manuel/RS 232-controlled peristaltic pumps, Type 110, Ole Dich Instrumentmakers, Hvidovre, Denmark) controlled the flows (inflow and outflow) of reactants.Three of the channels were used to deliver the reactants (aqueous solutions of potassium iodate, sulfuric acid and hydrogen peroxide) and one channel of the other pump was used to remove the surplus volume of the reaction mixture through a U-shaped glass tube.In this way, the volume of the reaction mixture, keeps constant at 22.2 ± 0.2 mL.
A Pt electrode (Metrohm model 6.0301.100,Herisau, Switzerland) was used as the working electrode and a double-junction Ag/AgCl electrode (Metrohm model 6.0726.100)as the reference electrode against which all potential were recorded.An electrochemical analytical device (PC-Multilab EH4 16-bit ADC) connected with a personal computer was used to record the potential potential-time curves of the BL matrix.Signals were recorded as a function of time with time step 1.0 s.Perturbation of the BL matrix was performed by manual injections.The analyte was introduced using micropipettes (Brand, Wertheim, Germany).A 50 µL shot was estimated to last about 0.5 s.The intensity of the perturbation corresponded to the total amount (in micromoles) of analyte injected in the 50 mL aliquot of standard samples.

Preparation of samples from tablets
To determine the vitamins from B-groups in the pharmaceutical preparation (sample 1, 2 and 3), ten tablets of each pharmaceutical were weighted, pulverized and the average mass of one tablet was evaluated.The sample solution was prepared by quantitatively transferring the average mass of one tablet, equivalent to 0.3809 g (sample 1), 0.3509 g (sample 2) and 0.9108 g (sample 3) in a 25-mL volumetric flasks as well as diluting to the volumes with water (samples 1 and 3); for sample 2, about 20 mL of water were added, the dispersion was shaken for 10 min in ultrasonic bath, the solutions were diluted to volumes with water and finally filtered through the filter paper Whatman No. 1. Perturbations of the matrix were performed with suitable aliquot of the solution.

Procedure for determination of analytes
The start-up procedure was performed in the following way.First, thermostated (T = 56.0 ± 0.1°C) and protected from light, the reaction vessel was filled up with the three separate inflows of the reactants, 5.90 × 10 -2 mol L -1 KIO 3 , 6.47 × 10 -2 mol L -1 H 2 SO 4 and 1.50 × 10 -1 mol L -1 H 2 O 2 , at the maximal flow rate (12 mL min -1 ).Under these conditions, within 3.5 min, about twice the volume of the reaction mixture becomes charged.Then, the inflows were stopped, the stirrer was turned on 900 rpm, and the excess of the reaction mixture was sucked out through the U-shaped glass tube, to achieve the actual reaction mixture volume 22.2 ± 0.2 mL.Hence, the reaction commenced under the bath conditions.After two bath oscillations (after about 20 min) the inflows were turned on at the required specific flow rate 2.95 × 10 -2 min -1 and the inflow concentration of sulfuric acid was varied in an interval, 6.47 × 10 -2 mol L -1 ≤ [H 2 SO 4 ] ≤ 9.00 × 10 -2 mol L -1 .In this way, we examined dynamic behavior of the BL matrix (bifurcation analysis) and found precise locations of operation points in the concentration phase space that was perturbed with analytes.
Once the bifurcation diagram as well as bifurcation point has already been determined, in the subsequent routine analysis it is sufficient to adjust the inflow concentration of sulfuric acid to some of the operation values (8.16 × 10 -2 mol L -1 , 8.44 × 10 -2 mol L -1 or 9.00 × 10 -2 mol L -1 ).Thus, after two bath oscillations obtained under the above-mentioned conditions, the inflow concentration of sulfuric acid was immediately adjusted to the selected operation value.In this way, the validity of the preparatory procedure and the used chemicals was confirmed before the experiments.The preparatory procedure in all cases took about 40 min.
For analytical purposes the appropriate dynamic state, that will be perturbed, ought to be selected; the location of the actual bifurcation point in the parameter phase space, which may vary slightly for different setups in different laboratories, is an analytical start point.With this aim, the investigation of dynamic behaviors of the BL system as a function of the control parameter (bifurcation analysis) must be performed before the actual analysis, but only once.
Under the CSTR conditions characterized by constant parameters such as inflow concentrations of , temperature T = 56.0°C and specific flow rate, j o = 2.95 × 10 -2 min -1 , we examined the dynamics of the BL system by varying the inflow concentration of sulfuric acid from [H 2 SO 4 ] = 6.47 × 10 -2 mol L -1 to [H 2 SO 4 ] = 9.00 × 10 -2 mol L -1 .The bifurcation diagram (Figure 1), showing the envelope of the simple periodic oscillations (for inflow concentration of sulfuric acid in the range 6.47 × 10 -2 mol L -1 ≤ [H 2 SO 4 ] ≤ 7.46 × 10 -2 mol L -1 (zone I)) as well as stable stationary states (for inflow concentration of sulfuric acid: 8.02 × 10 -2 mol L -1 ≤ [H 2 SO 4 ] ≤ 9.00 × 10 -2 mol L -1 (zone III)) is presented in Figure 1.Also, for inflow concentration of sulfuric acid in the range 7.61 × 10 -2 mol L -1 ≤ [H 2 SO 4 ] ≤ 7.88 × 10 -2 mol L -1 (zone II) a region of mode oscillations was provided.The bifurcation point is found at [H 2 SO 4 ] = 7.95 ×10 -2 mol L -1 .Determination of a type of bifurcation point does not have to be performed for the application of the analytical procedure.However, it is important for sensitivity of the system on perturbation, and it is interesting from scientific point of view. 48

Sensitivity of BL matrix to the perturbation
As can be seen from Figure 1, different dynamic states of BL matrix can be achieved with respect to inflow concentration of sulfuric acid as the control parameter.We chose to perform perturbations only in the stable non-equilibrium stationary states.0][11][12][13] Therefore, around the found bifurcation point we analyzed sensitivity of several stable stationary states, indicated by arrows in Figure 1, by perturbations of the matrix with the standard solutions of vitamin B 1 and vitamin B 3 in order to find the maximum response to the analytes, i.e., the optimal injection point.In particular, we perturbed the stationary states that were realized for the inflow concentration of sulfuric acid, [H 2 SO 4 ] = 8.16 × 10 -2 mol L -1 , 8.44 × 10 -2 mol L -1 and 9.00 × 10 -2 mol L -1 (Figure 2).As can be seen from Figure 2, the maximum response of the considered matrix system to selected concentration of B 1 and B 3 was obtained at [H 2 SO 4 ] = 8.44 × 10 -2 mol L -1 .Consequently, the dynamic The bifurcation between stable and unstable steady state occurs at inflow concentration C H 2 SO 4 = 7.95 × 10 -2 mol L -1 .The operations points for perturbations of matrix system with analytes are indicated by arrows.state at this acidity was chosen as optimal injection point for B 1 and B 3 , as well as in the cases of the other examined analytes (vitamin B 2 and vitamin B 6 ).
The influence of a dynamic state obtained for inflow concentration [H 2 SO 4 ] = 8.44 × 10 -2 mol L -1 , on the sensitivity of the matrix after perturbations is illustrated in Figure 3.The typical dynamic profiles obtained before and after perturbation of matrix system by the standard solutions of B 1 (Figure 3a), B 2 (Figure 3b), B 3 (Figure 3c) and B 6 (Figure 3d) under the above described experimental conditions, are presented.Matrix response behavior was investigated by injecting examined analytes into the reaction mixture, when BL reaction was operated under the conditions where stable nonequlibrium stationary states endured.Thus, before perturbation, the system was in a stable stationary state, and the corresponding potential denoted as E s was constant.Injection of vitamin of B-group causes an abrupt change in potential.The potential value denoted as E p is maximal (for B 1 and B 6 ) or minimal (for VB 2 and VB 3 ) value of the potential that is being reached.The change in potential (analytical signal) is defined as maximal potential displacement, DE m = E p -E s .

Determination of vitamins of B-group
Under the selected experimental conditions described above, the response of the matrix to the perturbation was measured by employing the change in potential, DE m = E p -E s .For the examined compounds, as recommended by ICH, 49 a calibration was established over five analyte levels in triplicate.The figures of merit for the calibration graphs (regression equations and regression factor), in addition to other parameters of analytical interest (limit of detection (LOD), limit of quantification (LOQ) and relative standard deviation (RSD) are all summarized in  Table 1.In accordance with ICH guidelines, 49 the LOD is defined as concentration of analyte that was produced signal-to-noise ratio of 3, where LOQ was assessed at a minimum signal-to-noise ratio of 10.These limits were experimentally verified by three injections of analyte at the LOD and LOQ amounts that all give acceptable precision and accuracy under the ICH guidelines.The uncertainty of the estimated values (expressed as relative standard deviation, RSD) of the analyte concentrations arises from uncertainties in the estimated values of maximal potential shift, and it is propagated through all of the analysis using technique found in Bevington and Robinson. 50n the case of vitamins B 1 and B 6 , DE m is linearly proportional to the logarithm of the analyte concentrations.For vitamin B 3 , DE m is linearly proportional to the analyte concentrations whereas in the case of vitamin B 2 , the calibration data obey the second-order polynomial equation (Table 1).Although the mentioned quantitative determination can be performed at all three tested inflow concentration of sulfuric acid, the determination at inflow of the acidity [H 2 SO 4 ] = 8.44 × 10 -2 mol L -1 has a wider linear range, higher instrumental sensitivity (the slope obtained from the regression curve) and higher precision (Table 1).This is an additional reason why mentioned inflow concentration of sulfuric acid was used as the optimum acidity value for quantitative determinations of vitamins B-group.
Sample throughput (sample determinations per hour) is required by the time needed for the system to recover to the initial stationary state after each perturbation.The proposed analytical procedure provided under the selected experimental conditions has an analytical throughput of 30, 5, 12 and 20 determinations per hour, for vitamins B 1 , B 2 , B 3 and B 6 , respectively, when sample volume of 50 mL was used.
In literature we can find the quantitative determination of vitamins B 6 and B 2 based on the perturbation response of the oscillatory matrix.For vitamin B 6 determination, the present method is characterized by lower limit of detection (0.025 mmol) and higher sample throughput (20 samples per hour) than the method based on perturbation H 2 O 2 -KSCN-CuSO 4 -NaOH oscillatory reaction, 1 although the range of linearity of about two orders of magnitude is comparable with the mentioned method.On the other hand, the limit of vitamin B 2 determination achieved in our study (0.018 mmol) is higher than that reported in oscillating matrix Mn(II)-BrO 3 --diacetone-H 2 SO 4 . 2On the other hand, using this matrix system that involves no important kinds of active oxygen, such as superoxide radical, requires UV irradiation setup for B 2 determination; this fact makes the method based on perturbation of this matrix less convinient related to BL matrix in which oxygen is the product of reaction.In addition, here used method has important advantage over the mentioned methods 1,2 that requires the relationship between different oscillation characteristics and perturbation concentrations in the selected oscillatory phase point: (i) procedure is simplified and (ii) time required for a full analysis is shortened considerably.
Under the above experimental conditions, dynamic behaviors of the BL matrix, after perturbations with vitamins B 2 and B 3 , are different from the ones obtained with vitamins B 1 and B 6 ; injection of B 2 and B 3 vitamins causes an abrupt decrease of the potentials to the values denoted as E p (Figure 3).This sudden response was followed by oscillatory return to the initial stationary state (Figure 4).As can be seen from Figures 4a and 4b, the number of full cycles that appear after perturbation of stable steady state is proportional to the concentration of the examined species.Therefore, we may define typical period, t p as the moment in which the BL matrix reverts to the initial stationary state (Figure 4a and 4b), or some new stationary state which is characterized by slightly different potential related to potential of initial state.Characteristic period, t p -t 0 , where t 0 is the moment in which analyte was injected, (Figure 4), depends on perturbation intensity. 10,12hus, the response to the vitamins B 2 and B 3 could be evaluated in one more manner, that is, by using the method based on the relationship between characteristic periods t p -t 0 and intensity of the perturbations.After the complete analysis we concluded that this method compared to the previous one, has a comparable dynamic linear range, lower precision and a higher detection limit.
Although the chemical reaction between examined analytes and matrix system is not necessary to be known for the application of analytical procedure, some discussion about it can be done.
2][53] The B 1 can be oxidized with both H 2 O 2 and I 2 51 and react with iodide anion to form thiamine hydroiodide (THI 2 ) 52 as well as a helical chain structure. 53Having in mind that in acidic media, the protonated thiamin (TH 2+ ) is the electroactive species, 54 the B 1 reduction through interactions with iodide anions is crucial 52 immediately after perturbation of the BL matrix with B 1 .Therefore, after perturbation we see, at the beginning, the sudden decrease of iodide concentration by following equation: In addition, having in mind that potentiometric behavior of B 6 is similar to the one for B 1 , as well as knowing that B 6 can react with hydriodic acid, 55 we assumed that introduced B 6 in BL matrix also first react with iodide ion.
In order to clarify the complex mechanism of interaction between both vitamins B 2 and B 3 , and the matrix, in preliminary experimental examinations, we tested whether analytes reacted with the reactants of the BL matrix.Thus, potential vs. time curves were recorded by using the CSTR and the folloving media: H 2 SO 4 , KIO 3, H 2 O 2 , H 2 SO 4 + KIO 3 , H 2 SO 4 + H 2 O 2 , and KIO 3 + H 2 O 2 .In the case of vitamins B 2 , only significant effect was obtained by injecting the vitamin B 2 into CSTR containing sulfuric acid alone what can be ascribed to its electrochemical behavior.This suggests that B 2 does not react directly with both hydrogen peroxide and iodate nor with the intermediates formed in the BL reaction.According to the obtained experimental results and knowing that reduced form of vitamin B 2 (B 2red ) can be oxidized with O 2 through a complex mechanism resulting in oxidized riboflavin (B 2ox ) and hydrogen peroxide, 56 we conclude that the main active specie in our system is hydrogen peroxide that is got from riboflavin peroxide.Having this in mind, as well as that in BL reaction a continuous production of oxygen occurs, 57,58 we suggest that following reactions play an crucial role in the examined system: Thus, the hydrogen peroxide produced by a complex (reaction 1) can react with hypoiodous acid (reaction 2) changing the ratio between iodine intermediates ([HOI] ss and [I -] ss ) that was established in the stationary state before perturbations.The given explanation is also in accordance with the results obtained after perturbation of the Mn(II)-BrO 3 --diacetone-H 2 SO 4 matrix system with B 2 . 2 Since in this oscillatory system there is no oxygen, direct interaction between B 2 and matrix is not found.As we mentioned, the analytical figure of merit for the determination of vitamin B 2 achieved in our study is worse than that reported in literature. 2 Thus, the oscillating matrix Mn(II)-BrO 3 --diacetone-H 2 SO 4 modified with active oxygen (O 2 •-→ H 2 O 2 ) 2 exhibits very high sensitivity for vitamin B 2 rather than BL matrix being in a stable nonequilibrium steady state.It is not easy to explain why and how to raise the sensitivity of analogous Belousov-Zhabotinskii oscillating reaction, since the mechanism of both matrixes, to which vitamin B 2 is added, is extremely complex as well as kinetically different.There is no doubt that the H 2 O 2 plays a very important role in both cases.Just due to the formation of H 2 O 2 caused by UV illumination of vitamin B 2 the effective concentration of the internal bromine species are change 2 and the strong perturbation can be observed.We do not know what would happen if the same procedure was applied for determination of vitamin B 2 in the BL matrix.In addition, the sensitivity of matrix depends on operating phase point; the determination of vitamin B 2 in the BL matrix is performed in a point where reinitiation of oscilatory state is found.It cannot be easily compared with excitability found in other cases.
In the case of vitamins B 3 the above described experiments were also evaluated.A potentiometrically recordable response to B 3 injections was observed in the mixture of KIO 3 + H 2 O 2 , in H 2 O 2 alone as well as in the mixture H 2 SO 4 + KIO 3 , known as Dushman reaction. 59According to the obtained experimental results and assuming that, B 3 similarly to nicotinamide nucleotides can be oxidized by hypoiodous acid under acidic conditions through a mechanism dominated by the kinetics of the halogenation of pyridine, 60,61 we suggest that B 3 oxidation through interactions with hypoiodous acid denote the first step after perturbation.Consequently, the ratio between iodine intermediates ([HOI] ss and [I -] ss ), established in the stationary state before perturbations is altered resulting in sudden increase of iodide concentration determined by very fast iodine hydrolysis.
where Py and Py/I, denotes pyridine unit of nicotinamide and iodinated Py unit, respectively. 60

Effect of foreign species
An interference study was carried out with aim to determine vitamin B 1 in real samples (pharmaceutical formulations).Thus, a systematic study of the effects of potentially interfering foreign species frequently existing with thiamine in pharmaceuticals, on the monitoring of these vitamins was undertaken.This study was carried out through measuring the response of vitamin B 1 alone and in the presence of some typical active principles, such as other vitamins B group, folic acid, ascorbic acid as well as some typical excipients (starch, glucose, sucrose and talc).In addition, the potential interference of metal cations commonly presented in several drugs, such as Cu(II), Mn(II), Mg(II), Zn(II) and Fe(III) was evaluated.The analytical signal of solutions containing a fixed amount of 6.0 × 10 -3 mol L -1 B 1 was compared with the analytical signal of these solutions spiked with different known concentrations of possible interfering agents.The tolerance limit was taken as ± 5% change in analytical signal.The species examined were found to interfere above tolerable ratios (TR) that are defined as number of micromoles of the interferent to the number of micromoles examined analyte.
The mentioned excipients under experimental conditions do not react with matrix system nor with to examined vitamins and no potential shifts were obtained when they were injected in matrix for concentrations of excipients a twenty-times higher than those of thiamine or niacin.
For vitamin B 1 determination, the following species, when present in amounts for which tolerable ration (TR) is shown in brackets, do not interfere: B 2 [20], B 12 [15], B 6 [12], Ca 2+ We would like to note strong interference of folic acid, as well as Mn 2+ , Mg 2+ and Fe 3+ ions that can be components of some mineral preparations, which all make the proposed method less convenient for quantitative determination of the examined vitamins in this kind of sample.

Analysis of active components (thiamine and niacin) in pharmaceutical dosage forms
In order to study the validity of the proposed method, it was applied to the determination of both B 1 and B 3 in commercially available pharmaceutical preparation (thiamine and niacin in tablets of both B 1 and B 3 as well as thiamine in vitamin B complex tablet) listed in Table 2.The selection of commercial tablets was made in such a way that the method could be applied to samples containing a massive dose of thiamin such as in the sample 1 (100 mg per serving), and to vitamin B-complex containing a normal dose of B 1 (12.3 mg per serving), as well as to the sample with a massive dose (500 mg per serving) of niacin (Table 2).The amounts of both thiamine and niacin obtained by the proposed method are in good agreement with those claimed by the manufactures (average RSD in the range 1.4-5.3%,Table 2).In order to study the reliability and suitability of the proposed method, the additional recovery experiments were carried out with the examined pharmaceuticals from those listed in Table 2.In the cases of thiamine determination (sample 1 and 2) as well as niacin determination (sample 3), the standard addition methods were performed by accurate additions of 0.05 mmol and 0.16 µmol of B 1 (sample 1 and 2), as well as 20 mmol of B 3 (sample 3) in the dilute samples.The Table 2 shows the results obtained; it can be seen that the average recovery varies from 100.7 to 103.2% indicating that the developed method was free from interference, and provided accurate results.In addition, the contents of B 1 and B 3 in vitamin B 1 , B complex and B 3 tablets were determined by using the methods of titrimetry in Yugoslav Pharmacopoeia. 62he obtained results are given in Table 2.It is shown that the results obtained by the proposed method agree with those obtained by the pharmacopoeia method; it is a useful method for quantitative analysis of B 1 and B 3 in pharmaceutical formulations.
In addition, the method employing the characteristic period t p -t 0 for determination of vitamin B 3 was used to analyze this compound in sample 3.This method tested did not provide acceptable results.The method employing maximal potential displacement was the more appropriate, probably as a results of the pharmaceutical containing some excipient that might interact briefly with the matrix, in such a way that its effect was obvious in the time for which the matrix reverts to the initial stationary state, but not in the potential displacement.

Conclusions
Our results demonstrate suitability of the use of the oscillatory reaction matrix in a stable non-equilibrium stationary state near a bifurcation point, for the determination of some water-soluble vitamins (B 1 , B 2 , B 3 and B 6 ) in bulk drug, as well as thiamin and niacin in pharmaceuticals.The developed kinetic method is simple and cheap; it operates without any derivatization reaction and shows good analytical features.The results are accurate and precise, and there are advantages in terms of short time required for each assay.For vitamin B 1 determination, the main limitation of the proposed method is a strong interference with folic acid having concentration larger than 6 × 10 -4 mol L -1 , as well as ions (Mn 2+ , Mg 2+ and Fe 3+ ) that can be components of some mineral preparations if their concentration exceed approximately 6 × 10 -3 mol L -1 .Therefore, it may be applicable for vitamin B 1 determination in pharmaceuticals that contain sufficiently low concentrations of these substances, as it is in our case.On the other hand, for vitamin B 3 determination, only samples with relatively high dose of this vitamin may be analyzed by the proposed method.Anyway, it is demonstrated that the proposed method is very appropriate for routine analysis of pharmaceuticals, without any pretreatment of the samples, apart from its dissolution, and the proposed method could also be used for their quality control.

Figure 1 .
Figure 1.Bifurcation diagrams showing transition from the stable stationary state (solid circles -region III) to the large-amplitude oscillations (open circles -region I) through deterministic chaos formed of aperiodic mixed mode large (open circles) and small amplitude (triangle) oscillations (region II).Open circles and triangles denote minimal and maximal potentials in oscillations, whereas dash doted lines show the region of boundary between different dynamic structures: stable stationary states (III), periodic oscillatory states (I) and deterministic chaos (II).The bifurcation between stable and unstable steady state occurs at inflow concentration C H 2 SO 4 = 7.95 × 10 -2 mol L -1 .The operations points for perturbations of matrix system with analytes are indicated by arrows.

Figure 2 .
Figure 2. Influence of inflow concentration of sulfuric acid on the analytical signal.Curves a and b correspond to B 1 and B 3 , respectively.In the cases a and b the amounts of B 1 and B 3 are 0.15 and 30 mmol, respectively.

Figure 4 .
Figure 4.The typical period, t pt 0 in the system (a) BL-B 2 and (b) BL-B 3 observed when inflow concentration of sulfuric acid was [H 2 SO 4 ] = 8.44 × 10 -2 mol L -1 ; the moment in which analyte was injected is denoted by t 0 .The perturbation strengths are (from left to right: (a) 0.032 and 0.064 mmol; (b) 20 and 30 mmol.

Table 1 .
Analytical parameters of thiamin (B 1 ), riboflavin (B 2 ), niacin (B 3 ) and pyridoxine (B 6 ) potential shift and X, absolute analyte injected concentration (mmol); b Correlation coefficient; c Limit of detection established at a signal-to-noise ratio of 3; d Limit of quantification established at a signal-to-noise ratio of 10; e Average relative standard deviations obtained from three determinations of 0.03 and 0.4 mmol of B 1 , 0.03 and 0.08 mmol of B 2 , 7 and 25 mmol of B 3 as well as 0.06 and 1.0 mmol of B 6 .
a Y, maximal

Table 2 .
62termination of both thiamine and niacin in the commercial vitamin B 1 and B 3 tablets as well as thiamine in vitamin B complex (B 1 , B 6 and B 12 ) tablets Samples 1 and 3 of Vitamin B tablets were produced by Now Foods, Blodmingdale, USA as well as sample 2 of vitamin B complex tablet was produced by Srbolek, Belgrade, Serbia.Samples containing: thiamine and excipients (sample 1); thiamin (12.3 mg), pyridoxine (5 mg), B 12 (0.05 mg), folic acid (0.4 mg), nicotinamide (20 mg) and excipients (sample 2); niacin (500 mg) and excipients (sample 3); b Mean concentration (in mg) of six determination; c Average standard deviation; d Average relative standard deviation; e Performed as accurate addition of 0.05 and 0.16 mmol of B 1 (sample 1 and 2) as well as 20 mmol of B 3 (sample 3) in the dilute samples.The RCV values are mean recoveries (n = 3); f Yugoslav Pharmacopoeia.62 a