Determination of Vanadium in Urine by Electrothermal Atomic Absorption Spectrometry Using Hot Injection and Preconcentration Into the Graphite Tube

É proposto um procedimento para a determinação direta de vanádio em amostras de urina por espectrometria de absorção atômica com atomização eletrotérmica em forno de grafite usando injeção à quente e pré-concentração dentro do tubo de grafite. Três aliquotas de 60 μL foram injetados seqüencialmente dentro do atomizador pré-aquecido a uma temperatura de 110 oC. As etapas de secagem e pirólise foram repetidas após cada injeção. Às amostras de urina foram adicionados o modificador químico fluoreto de bário (100 mg L) e o surfactante Triton X-100 (0,3% v v). A vazão da amostra durante a injeção dentro do tubo pré-aquecido foi de 0,5 μL s. Os limites de detecção e de quantificação sem pré-concentração foram, respectivamente, 0,54 e 1,82 μg L, e com pré-concentração 0,11 e 0,37 μg L. O teste de adição e recuperação foi empregado em amostras de urina para avaliar a exatidão do procedimento proposto. Os valores de recuperação obtidos variaram de 96,0 a 103,0% para as adições de 0,8 a 3,5 μg L V. O procedimento proposto possibilita a determinação de vanádio sem nenhum pré-tratamento e com mínima diluição da amostra.


Introduction
Vanadium is an essential trace element for plants and animals which acts in the synthesis of chlorophyll and promotes the growth of young animals.Vanadium compounds are toxic in high concentrations or after long periods of exposure. 1,2[5][6] However, the investigation of physiological effects caused by vanadium is hindered due to the severe difficulties for its determination in tissues and biological fluids.[9][10][11][12] Rigorous analytical quality control of all steps is mandatory.
A variety of methods have been used for vanadium determination including UV-visible spectrophotometry, fluorimetry, voltammetry, ion-selective electrodes, gas chromatography, neutron activation analysis, X-ray fluorescence spectrometry, emission spectrometry and electrothermal atomic absorption spectrometry (ETAAS). 13,14lectrothermal AAS has been used for determination of vanadium in different types of samples.Owing to the levels (µg L -1 ) of vanadium in urine, it is generally necessary to adopt a preconcentration procedure, such as ionexchange, 15,16 coprecipitation of vanadium with iron(II) hydroxide 17,18 or complexation and extraction with an organic solvent such as ammonium pyrrolidinedithiocarbamate (APDC) in MIBK, 19 cupferron in MIBK, 7,9,20 Ncinnamoyl-N-(2,3-xylyl)hydroxylamine in carbon tetrachloride 21 or 8-hydroxyquinoline in MIBK. 22These methods of sample preconcentration are time consuming, require a relatively large sample volume for each measurement and are prone of losses or contamination.
The direct preconcentration procedure using hot injection in ETAAS has been applied for determining vanadium in water samples, 21,22 urine after extraction of vanadium with cupferron in MIBK, 9 and milk samples. 14hese procedures offer advantages compared to conventional ones including reduced sample preparation time, decreased possibility of losses of analyte during the preconcentration step, and reduced possibility of sample contamination caused by manipulation or added reagents.
In the work here described it was developed a procedure for vanadium direct determination in urine samples using successive injections into a preheated graphite tube.

Instrumentation
Measurements were carried out using a Varian Model 800 atomic absorption spectrometer equipped with a graphite furnace atomizer, GTA 100 autosampler and transverse Zeeman effect background correction (Varian, Mulgrave, Australia).A vanadium hollow cathode lamp was employed as radiation source operating at 20 mA.Absorbance signals were measured using the 318.4 nm line and a spectral resolution of 0.2 nm.Pyrolytic coated graphite tubes (Part Number 63-100011-00, Varian) were used throughout.All signals were measured as integrated absorbance with an integration time of 4.3 s.The temperature program for the atomizer is shown in Table 1 for an injected volume of 60 µL.

Reagents and materials
All dilutions were made using ultra pure water (resistivity 18.2 MΩ cm) obtained in a Milli-Q water purification system (Millipore, Bedford MA, USA).A stock standard solution containing 1000 mg L -1 V (Spex, Metuchen NJ, USA) was used.Triton X-100 (Merck, Darmstadt, Germany) was also used.Throughout experimental work double sub-boiling distilled HNO 3 was used.This acid was prepared by sub-boiling distillation of HNO 3 concentrated (Merck) using a Milestone apparatus model subPUR (Sorisole, Italy).A solution of BaF 2 was used as chemical modifier by dissolving 100 mg of the salt (99.99% purity, Aldrich, Milwaukee WI, USA) in 100 mL of ultrapure water.Argon (99.999% purity, White Martins, Sertãozinho SP, Brazil) was used as purge gas.
All glassware, pipettes, micropipette tips, auto-sampler cups, and polypropylene containers were acid washed with 0.14 mol L -1 HNO 3 for 24 h and thoroughly rinsed 3 times with deionized water before use.

Procedure
Volumes of 600 µL of urine sample, 100 µL 1000 mg L -1 BaF 2 , and 100 µL 3% v v -1 Triton X-100 were transferred to the autosampler cup, and then diluted to 1000 µL with distilled-deionized water.Three successive 60 µL aliquots were injected into the atomizer, inserting drying and ashing steps between each injection.When injecting into a hot graphite tube, the sample flow-rate was 60 µL in 120 s (0.5 µL s -1 ).The graphite furnace was heated according to the program showed in Table 1 and the integrated absorbance signal was measured.

Electrothermal behavior of vanadium
One of the difficulties for determining vanadium by ETAAS is the formation of refractory carbides during the heating cycle.This effect causes a pronounced loss of sensitivity.This problem can be circumvented by using pyrolytic graphite tubes and by adding a chemical modifier.In a previous work on vanadium determination in milk, the use of BaF 2 as chemical modifier was successfully demonstrated.Probably the positive effect on sensitivity is caused by the formation of volatile compounds between vanadium and fluoride. 14n the developed work the electrothermal behavior of V was studied in the following media: 0.014 mol L -1 HNO 3 , urine, and urine in the presence of NH 4 NO 3 and BaF 2 as chemical modifiers.These chemical modifiers were chosen to facilitate the formation of volatile salts with the urine matrix and to improve vanadium atomization.In the absence of these modifiers, it was observed an intense deposition of salts in the tube orifice and graphite electrodes mainly when working with less diluted urine samples.The conventional heating program for the graphite tube atomizer is shown in Table 2 for an injected volume of 30 µL.The effect of NH 4 NO 3 and BaF 2 as chemical modifiers was evaluated.For a sample volume of 30 µL it was added 300 µg NH 4 NO 3 . 1aking into account that the effect caused by BaF 2 on V atomization in urine samples is less trivial, different masses of this salt were added and the results obtained can be observed in Figure 1.Best sensitivity was attained when using 6 µg BaF 2 .Pyrolysis and atomization curves for V without (Figure 2) and with chemical modifiers (Figure 3) were obtained.The optimum pyrolysis and atomization temperatures were 1800 ºC and 2700 ºC without or with BaF 2 .In a medium containing NH 4 NO 3 pyrolysis and atomization temperatures experimentally established were 1500 ºC and 2700 ºC.Background signals obtained for previously cited temperatures were 0.078, 0.056, and 0.052 for urine, either with BaF 2 , or with NH 4 NO 3 , respectively.In spite of the better sensitivity observed without adding any chemical modifier, the deposition of salts in the tube orifice and graphite electrodes caused memory effects and graphite tube lifetime was 120 firings.In the worst situations even partial obstruction of the optical path occurred.Both chemical modifiers eliminated the deposition of salts but lifetime of the graphite tube was 150 firings in a medium containing NH 4 NO 3 .Thus, further experiments were carried out by adding 6 µg BaF 2 to urine samples.

Effect of sample volume
Integrated absorbance signals increased with augment of sample volume.In order to increase sensitivity the introduction of higher volumes of samples in a preheated tube was evaluated.The effect of hot injection into a   graphite tube on the integrated absorbance signals of V was studied from 70 to 120 ºC (Figure 4).Better sensitivity was reached at 110 ºC.The hot injection increases sensitivity and allows a faster drying step since solvent is at least partially volatilized by introduction at a programmed solution flow-rate.The increment in sensitivity is clearly related to the higher mass of analyte introduced and with the use of hot injection.The reaction: V 2 O 3(s) + 3 C (s) → 2 V (s) + 3 CO (g) , proposed by Styris and Kaye, 24 may be occurring when using a preheated tube. 14igure 5 shows the effect of sample volume introduced into a graphite tube at room temperature (10-40 µL) using BaF 2 and Triton X-100 as matrix modifiers and the heating program showed in Table 2 and when solution (30 to 70 µL) was introduced into the heated tube kept at 110 ºC.The repeatability of measurements may be affected when larger volumes are dispensed into the atomizer.Relative standard deviations (RSDs) of the measurements varied from 1.6 to 5.2% and 1.5 to 4.1% without and with preheating for sample volumes of 30 and 70 µL, respectively (n = 3).This procedure can be adopted for improving sensitivity.

Figure 5.
Effect of sample volume introduced into a graphite tube at room temperature (A) and into a graphite tube heated at 110 ºC (B).Integrated absorbance refers to a solution containing 20 µg L -1 V in 0.014 mol L -1 HNO 3 with the use of BaF 2 as matrix modifier.J. Braz.Chem.Soc.

Graphite furnace preconcentration heating program
The typical sensitivity obtained using conventional procedures is not enough for determination of vanadium in urine by ETAAS and the use of a preconcentration step is necessary.A method of sample preconcentration by urine mineralization and vanadium chelation followed by solvent extraction has been described, 7 but despite its effectiveness it requires many steps that can impair accuracy and precision.Thus it was decided to investigate an in situ preconcentration procedure into the graphite tube using multi-injection into a preheated tube.Triton X-100 was added to each urine sample for improving autosampler action and sample distribution into the tube wall.Volumes of 600 µL of urine sample, 100 µL of BaF 2 solution (1000 mg L -1 ), 100 µL of Triton X-100 (3% v v -1 ) were transferred to the autosampler cup, and the volume was made up to 1000 µL with distilled-deionized water.Three 60 µL volumes were sequentially injected into the atomizer kept at 110 ºC.Drying and pyrolysis steps were applied after each sample aliquot (Figure 6).The experimental conditions adopted are presented in Table 1.All signals were measured as integrated absorbance.When injecting into a hot graphite tube, sample flow-rate was 0.5 µL s -1 .For a 60 µL urine aliquot containing 5.0 µg L -1 V without preconcentration (Figure 7), net absorbance and background signals were 0.027 and 0.050, respectively.This same solution 3 times preconcentrated (Figure 7) into a preheated graphite tube (3 x 60 µL) generated signals of 0.090 and 0.123 for net absorbance and background, respectively.The background correction was based on Zeeman effect.The lifetime of the graphite tube under these conditions was about 200 firings.

Analytical performance
The matrix effects were evaluated by comparing the slopes for analytical calibration curves built up from aqueous reference solutions and matrix-matched solutions.Sensitivities obtained varied depending on the medium.Hence, calibration based on reference solutions prepared in water is not feasible and the method of matrix matching was therefore used.The analytical curve was obtained by adding appropriate volumes of BaF 2 and Triton X-100 solutions in matrix-matched reference solutions containing from 1.0 to 10.0 µg L -1 V.The solvent used in the preparation of all the solutions was a sample of urine containing a vanadium concentration below the detection limit.Good linear correlation coefficient (r = 0.9997) was obtained for calibration graph: Q A = 2.1 x 10 -3 + 1.7 x 10 -2 c, where Q A is the integrated absorbance and c is the vanadium concentration.The accuracy of the developed procedure

Figure 4 .
Figure 4. Effect of temperature of the heated tube on absorbance signal (injected sample volume: 30 µL and concentration of V: 20 µg L -1 ).

Table 1 .
Graphite Furnace heating program with preconcentration a a Sample flow-rate: 0.5 µL s -1 .

Table 2 .
Graphite Furnace heating conventional program

Table 3 .
Added-recovery of V in urine (n=3) are thankful to Prof. Dr. Pedro V. Oliveira (IQ/USP, São Paulo) by loan of the radiation source.