Square wave stripping voltammetry of titanium based on adsorptive accumulation of its hydroxynaphthol blue (HNB) complex at the static mercury drop electrode

A new method is described for the determination of titanium based on the square wave adsorptive stripping voltammetry of Ti(IV) complexed with Hydroxynaphthol blue (HNB) at the static mercury drop electrode. Optimal conditions were found to be: preconcentration potential, −0.140 V vs. Ag/AgCl (KCl 3M); preconcentration time, 30 s (with stirring solution); pulse height, 75 mV;frequency, 100 Hz; scan increment, 4 mV; step time, 0.010 s; supporting electrolyte, acetate buffer (0.1 M, pH 3.2), and concentration of hydroxynaphthol blue, 1.0 × 10⁻⁵ M. The response of the system was found to be linear in a range of Ti(IV) concentrations from 5.0 to 20.0 μg/L. The limit of detection was found to be 0.18 μg/L and the limit of determination to be 1.09 μg/L, both using 30 s of preconcentration time. The effects of various potential interferences were also studied including a variety of cations, anions, and organic surfactants. The merits of the procedure were demonstrated in the analysis of certified NIST SRM 1633b and 1646a samples.

Keywords: Titanium; Hydroxynaphthol blue (HNB); Square wave stripping voltammetry; Certified samples

Titanium is the ninth most abundant element, corresponding to 0.6% of the earth's crust. Titanium and its compounds can be produced from ilmenite (FeTiO3) and rutile (TiO2). Because of its whiteness, opacity, inertness, nontoxicity, and relative cheapness, TiO2 is now the most widely used white pigment for paints. TiO2 is also used as a paper whitener and in glass, ceramics, floor coverings, and cosmetics. Extensive production of titanium is in continuous development, spurred by needs of the military and the aircraft industry. Titanium is also present in some natural samples, such as seawater, at very low concentrations in the range from 2 to 300 pM [1]. Because of the importance of this metal and its presence in low concentrations in different samples, it was necessary to develop a simple and highly sensitive analytical method for its determination. Actually, titanium can be determined by several methodologies, such as atomic absorption spectrometry [2] or inductively coupled radio frequency plasma (ICP) spectrometry [3]. However, these analyses have practical difficulties such as the necessary atomization at high temperature or prohibitive costs of instrumentation. In electroanalysis the use of polarography for determination of titanium was initially reported by Adams [4] in 1948. The electrochemical behavior of titanium was an impediment to advances in its polarographic determination. The principal reduction observed, from Ti(IV) to Ti(III), is not useful for trace determinations of this element. Improvement in its determination (down to 1.5 × 10−8 M) can be obtained by coupling the use of the waves of titanium complexes with differential pulse measurements [5] and by using other sensitive voltammetric techniques such as stripping analysis [6]. From the 1980's on, a huge progress in voltammetric analysis has occurred through the use of new electrodes and computerized analyzers, which allow the rise of sensitive and selective methods. Recently, the adsorption characteristics of complexes have been used in the determination of traces of titanium. Then the complexes of titanium are accumulated on the surface of a hanging mercury drop electrode (HMDE) by adsorption. Subsequently, the adsorbed complexes are reduced or oxidized by different potential sweeps. During the last years adsorptive voltammetry, a new branch of voltammetry, the so-called cathodic adsorptive stripping voltammetry, has been attracting considerable attention to the determination of metal ions at the trace and ultratrace levels. This is due to its high sensitivity, the high accuracy achievable, and low instrumental costs. Several complexants for titanium were reported, such as mandelic acid [7], solochrome violet RS [8], beryllon III [9], and other heterocyclic azo compounds [10]. In our laboratory we have been investigating various adsorptive voltammetric techniques employing hydroxynaphthol blue (HNB) as complexing agent for nickel [11] and cobalt [12]. The HNB is an azo dye (o,o′-dihydroxyarylazo compound) of the same class as eriochrome black T, calcon carboxylic acid, calmagite. These azo dyes are widely accepted as metal indicators in the chelatometry. Other reagents with related structure are the gallion, spadns, berryllon II, magon, and magneson [13]. The square-wave voltammetry is a fast method with good discrimination against background currents. The aim of this work was the development of a new square wave adsorptive stripping voltammetric (SWAdSV) procedure for the determination of titanium using HNB as the complexing agent. The studies were carried out at a static electrode (HMDE). The method was performed in acetate buffer (0.1 M, pH 3.2) media. The influence of a number of experimental parameters is presented. The utility of the method was efficiently demonstrated by the recoveries for titanium in the case of the analysis of certified samples (National Institute of Standards and Technology, Standard Reference Materials Program, USA).

Voltammetric determinations were applied with an EG&G PAR model 384-B Polarographic Analyser, together with a Houston Ametek-DMP-40 series digital plotter. The working electrode was a PAR Model 303 static mercury-drop electrode. A large-sized static mercury drop electrode (SMDE) with a surface area of 0.032 cm2 was used. The sample cell (PAR Model 0057) was fitted with an Ag/AgCl (KCl 3M) reference electrode and platinum wire auxiliary electrode. Dissolved oxygen was removed from the test solutions by purging with prepurified nitrogen. A magnetic stirrer EG&G PAR model 305 and a stirring bar (Nalgene Cat. No. 6600-0010) were used to achieve convective transport during the accumulation step.

Purified water, produced in a "Milli-Q" water purification system (Millipore), was used for all dilutions and to prepare solutions. All chemicals were of analytical-reagent grade. A 1000-μg/mL stock solution (Merck, Titrisol) was used to prepare titanium solutions. A 3.2 × 10−3 M stock solution of HNB (J.T.Baker) was prepared in water. As supporting electrolyte a 1.0-M acetate buffer (pH = 3.2) was used. Next, 0.0542 g of Estuarine Sediment SRM 1646a certified samples (supplied by National Institute of Standards and Technology, Standard Reference Materials Program, Gaithersburg, MD, USA) was fused with lithium metaborate in a platinum crucible. The melt was dissolved in dilute nitric acid and made up to 25 mL. Finally, 0.0533 g of Coal Fly Ash SRM 1633b samples were also dissolved by lithium metaborate and made up to 25 mL with 1 M nitric acid.

Ten mL of acetate buffer supporting electrolyte solution (9.0 mL of purified water + 1.0 mL of acetate buffer) containing 1.6 × 10−5 M HNB was purged with nitrogen for 8 min. The preconcentration potential (usually 0.140 V) was applied to a fresh mercury drop while the solution was stirred. Following the accumulation period (usually 30 s) the stirring was stopped and after a 30 s equilibration time, the square wave voltammogram was recorded with a scan rate of 200 mV/s until −0.800 V was reached, while a stream of nitrogen passes over the surface. After the background square wave stripping voltammogram had been obtained, suitable aliquots of titanium standard solution were introduced into the cell. All data were obtained at (23 ± 1)°C.

Figure 1 shows the square wave adsorptive stripping peak current, for a solution of acetate buffer (0.1 M, pH 3.2) containing only 1.6 × 10−5 M HNB (A) and for the same solution after the addition of 20 μg/L titanium (IV) (B). The accumulation potential of 0 V vs. Ag/AgCl (KCl 3M) was applied to the working electrode while the solution was stirred for 30s. At that time the stirring was stopped and after 30 s the square wave voltammograms were recorded in the negative direction with a scan rate of 200 mV/s. Two cathodic peaks resulted due to the reduction of the HNB complexing agent (at −0.22 V) and the titanium/HNB complex (at −0.32 V), respectively. The effect of the pH on the square wave stripping peak current of the Ti-HNB complex (Fig. 2A) and on the peak potential of the free HNB and Ti-HNB complex (Fig. 2B) were studied using acetate buffer 0.1-M solution as supporting electrolyte. The signal current peak remains constant in the range from 0.05 to 0.5 M of sodium acetate as electrolyte. A concentration of 0.1 M was used for the subsequent determinations. Over a pH range of 2.5 to 6.5 and with an accumulation potential of −0.100 V as well as of 30 s of preconcentration time, well-defined peaks were obtained. Higher current values were observed close to pH 3.2. After this pH the current decreases almost linearly. A similar linearity was observed with decreasing peak potential when the pH is changed from 2.7 to 6.2. A 0.1-M acetate buffer at pH 3.2 was chosen for subsequent determinations. The effect of the square-wave frequency in the range of 10 to 120 Hz (see Fig. 3A) was verified by observing the better sensitivity of the current signal of the Ti-HNB complex. The current signal increases until 100 Hz. the best sensitivity and peak resolution were obtained when setting the frequency at 100 Hz with a pulse height at 75 mV. The effect of the accumulation potential was studied from 0 to −0.24 V (see Fig. 3B). In the range from −0.10 to −0.18 V the square wave adsorptive stripping peak showed the same values. The influence of variations of the preconcentration time from 0 to 240 s on the square wave adsorptive stripping peak current also was observed, when using titanium concentrations of: 5, 10, 15, and 20 μg Ti(IV)/L. Figure 4A shows this effect when the titanium concentration is 20 μg/L. Both curves initially showed linear behavior. Depending on the concentration, the drop saturation occurred at different accumulation times: The curve corresponding to the higher concentration showed saturation at a lower accumulation time than the lower concentrated one. The possible stoichiometry of the complex was also investigated (Fig. 4B). The titanium concentration was held constant at 1.7 × 10−7 M and the concentration of the HNB was varied from 6.6 × 10−8 to 5.3 × 10−7 M. The square wave stripping peak current was plotted as a function of the molar ratio of ligand to the metal ion in solution. The "end point" of the determination in this case is not entirely clear and hence does not yield a conclusive result for the Ti:HNB ratio. The inflection point lies between a ratio of 1:1 and 1:2, which is in accordance with published results on the ratio of metal to o,o′-dihydroxyarylazo complexes13. The limitation of this new method for the determination of the composition of a complex is the saturation of the surface mercury drop.

Graph: Figure 1. Square wave voltammograms for (A) 1.6 × 10−5 M HNB and (B) 20 μg Ti (IV)/L complexed with HNB. Experimental conditions: acetate buffer 0.1 M (pH = 3.2) supporting electrolyte; stirred solution; preconcentration time 30 s at 0 V; equilibrium time 30 s; purge time 8 min; drop area 0.032 cm2; drop time 0.010 s; scan increment 2 mV; frequency 100 Hz; pulse height 75 mV; scan rate 200 mV/s.

Graph: Figure 2. Effects of pH on the square wave adsorptive stripping peak current (A) and potential (B) for free HNB and titanium-HNB complex. Experimental conditions: preconcentration time 30 s at −0.100 V. Other conditions same as Figure 1.

Graph: Figure 3. Effects of frequency (A) and accumulation potential (B) on the square wave adsorptive stripping peak current of the titanium-HNB complex. Experimental conditions: acetate buffer 0.1 M (pH = 3.2) as supporting electrolyte containing 1.6 × 10−6 M HNB and 10 μg Ti (IV)/L; preconcentration time 30 s at −0.140 V (A) and frequency of 100 Hz (B); pulse height 50 mV. Other conditions same as Figure 2.

Graph: Figure 4. Effects of accumulation time (A) and HNB concentration (B) on the square wave adsorptive stripping peak current of the titanium-HNB complex. Experimental conditions: acetate buffer 0.1 M (pH = 3.2) as supporting electrolyte containing 1.6 × 10−6 M complexed with 20 μg Ti (IV)/L in (A) and different concentrations of HNB with 10 μg Ti (IV)/L in (B); preconcentration time, 30 s (B). Other conditions same as Figure 3.

Influences of other instrumental parameters such as step potential (2–10 mV), equilibrium time, drop area (0.009–0.032 cm2), and the duration of the nitrogen purge were also studied. It was concluded that the best sensitivity for the square wave peak resolution was obtained at 2 mV, 30 s, 0.032 cm2, and 8 min.

A possible mechanism for the overall electrode process in adsorption voltammetry of the Ti(VI)-HNB system can be divided into three steps:

Where HNB′ is the product of the reduction of the azo group in HNB.

The complex was found to be stable for at least 180 min, when checking its current value each 10 min. Eleven successive measurements yielded a mean square wave peak current of 1094 nA for 10 μg Ti(IV)/L with a deviation of 57.1 nA and a relative standard deviation of 5%. Such behavior is attributed to the reproducibility of the adsorption step and to the use of a new drop of reproducible area in each measurement.

Under the recommended conditions, square wave voltammograms were obtained for solutions of increasing titanium concentration (2–10 μg/L) with a 30 s preconcentration time. Well-defined stripping peaks were observed over this concentration range. In contrast, the corresponding solution-phase response was not useful for quantitative work at this level. The square wave stripping peak current increases linearly with the concentration. The least-square analysis of this plot yielded a slope of 139.2 nA/μgL−1 and a correlation coefficient of 0.999. Different calibration curves were determined with Ti(IV) concentrations of 5 to 20 μg/L employing 30, 60, 120, 180, and 240 s accumulation steps. The resulting calibration curves are linear with respect to the titanium concentration at accumulation times from 30 to 120 s. The same results for the signal current peaks were obtained when the accumulation time used was higher than 120 s. The limit of detection was found to be 0.18 μg/L and the limit of determination1.09 μg/L, both using 30 s (the accumulation time most commonly used in this study) of preconcentration time.

The major sources of interference are likely to be the coexisting ions and organic surfactants. These species could result in either new reduction peaks or could overlap with the Ti-HNB signal current peak, thus obscuring its measurement. The various ions that could interfere were also investigated. The determination of 20 μg Ti(IV)/L was not affected by the addition of up to 20000 μg/L of CDTA (trans-1,2-diaminocyclohexaneN,N,N′,N′tetraacetic acid), 2000 μg/L of sodium tartrate, chloride, sulphate, tetraborate, DTPA (diethylenetriaminepentaacetic acid), NTA (nitrilotriacetic acid), TTHA (triethylenetetraminehexaacetic acid), fluoride, sodium oxalate, nitrate, o-phenanthroline, Au3+, Sm3+, Dy3+, Gd2+, Pr4+, Cr3+, Co2+, Cu2+, Fe3+, Sb3+, La3+, Ca2+, Mg2+, T1+, Sb3+, Se4+, Cd2+, Mn2+, and W6+ or by the addition of up to 200 μg/L of sodium citrate and Pd2+ or by the addition of up to 20 μg/L of Cr3+, Al3+, Eu2+, Sc3+ and Ga3+. The more serious interferents are the Mo6+, Zr4+, U6+, V5+, Ni2+, and EDTA (ethylene diamine tetra acetic acid). The interference by Ni2+, V5+, Mo6+, Zr4+, and Al3+ may be minimized by an adequate choice of the accumulation potential. A destruction of organic surfactants by ultraviolet irradiation is advisable for samples suspected to contain high levels of these materials.

The suitability of this method for the determination of titanium in Estuarine Sediment SRM 1646a certified samples is illustrated in Figure 5. Four successive standard additions to the samples resulted in well-defined square wave adsorptive stripping peaks. The Ti-HNB peak in the original sample (curve a) can thus be quantified based on the resulting standard addition (2.0 to 8.0 μgTi(IV)/L) plot (also shown in Fig. 5). The least-squares analysis of this plot yielded a slope of 125.39 nA/μg.L−1 and a correlation coefficient of 0.999. Five consecutive analyses yielded an average value of 0.48 wt.-% of titanium with a standard deviation of 0.02%. This value is similar to results obtained with graphite-furnace atomic absorption spectrometry (0.53 wt.-%) and is in agreement with the reference value of 0.51 wt.-% Titanium in Coal Fly Ash SRM 1633b samples was also determined (9.0 mL of purified water + 1.0 mL of acetate buffer 1M + 50 μL of HNB 3.2 × 10−3 M + 10 μL of sample; other conditions and procedures as for Fig. 5). Least-squares analysis of this plot yielded a slope of 74.79 nA/μg.L−1 and 0.999 for the correlation coefficient. The average value for six consecutive analyses was 1.03 wt.-% of titanium with a standard deviation of 0.07%. This value is in agreement with the one of graphite-furnace atomic absorption spectrometry analysis (1.06 wt.-% and comes near to the reference value of 0.8 wt.-%. Principal interferents in this method for the sample SRM 1633b are: Ni 127 μg/g, Mo 29 μg/g, Al 14,3 wt.-%, U 10,2 μg/g (Zr not present), and for the Estuarine Sediment SRM 1646a: Ni 32 μg/g, Mo 2,0 μg/g, Al 6,25 wt.-%, (U and Zr not present). The easy hydrolysis of titanium is a disadvantage that would be minimized by determination of this metal within the same day of sample digestion [14].

Graph: Figure 5. Measurements of titanium (IV) in the standard sample SRM 1646a (Estuarine Sediment). (9.0 mL of purified water + 1.0 mL of acetate buffer 1 M + 50 μL of HNB 3.2 × 10−3 M + 10 μL of sample). Square wave adsorptive stripping voltammograms for the sample (a) and after standard addition (2 (b) - 8 (e) μg Ti(IV)/L) following 30 s preconcentration at −0.140 V. Other conditions as in Figure 4. Also shown, the resulting standard addition plots.

The present study describes an effective way for the determination of titanium. Square wave voltammetry is a fast technique with good discrimination against background currents. The sensitivity of this new method described is equivalent or better than those utilizing as complexing agents beryllon, cupferron (adsorptive method), and mandelic acid (catalytic adsorptive method). Nevertheless, the use of these others ligands strongly depends on the kind of samples analyzed. With the HNB method, several masking agents such as CDTA, fluoride, TTHA, and NTA can be utilized to minimize possible interferences. This approach in particular offers advantages such as the possibility to determine titanium in a number of samples in the presence of organic surfactants and of several ions such as Al3+ and Fe3+. The merits of this new method were confirmed by the determination of titanium in SRM 1633b and Estuarine Sediment SRM 1646a samples without the use of any masking agents. The data indicates that the matrix effects, are low. Further studies are in progress to find an alternative to the mercury electrode involving the construction and behavior of a mixed binder carbon paste electrode containing HNB. This method will be investigated with respect to the simultaneous determination of nickel, titanium, and cobalt.

The authors gratefully acknowledge the MCT, PADCT, CNPq and CAPES of the Government of Brazil and PUC-Rio for support of this work. We are indebted to Dr. Silvia M. Sella for revision of the manuscript. The experimental assistance of Anselmo B. Neves is appreciated.