Selective stopped-flow sequential injection method for the spectrophotometric determination of titanium in dental implant and natural Moroccan phosphate rock

Talanta 71 (2007) 1405–1410

Selective stopped-flow sequential injection method for the spectrophotometric determination of titanium in dental implant and natural Moroccan phosphate rock Fotini S. Kika, Demetrius G. Themelis ∗ Laboratory of Analytical Chemistry, Department of Chemistry, Aristotle University of Thessaloniki, 541 24 Thessaloniki, Greece Received 4 April 2006; received in revised form 12 June 2006; accepted 6 July 2006 Available online 1 September 2006

Abstract The present work reports the first sequential injection (SI) method for the spectrophotometric determination of Ti(IV). The method is based upon the reaction of Ti(IV) with chromotropic acid (CA) in acidic medium to form a water-soluble complex (λmax = 420 nm). The chemical and instrumental variables of the system that affected the reaction were studied. Selectivity was greatly enhanced using ascorbic acid. A linear calibration graph was obtained in the range 0.2–10.0 mg l−1 Ti(IV) at a sampling frequency of 24 h−1 . The precision was satisfactory (sr = 1.5% at 5.0 mg l−1 Ti(IV), n = 12) and the 3σ limit of detection, cL , was 0.7 mg l−1 (n = 10). The developed method proved to be adequately selective and was applied successfully to the analysis of real samples (dental implant and natural Moroccan phosphate rock) giving accurate results based on recovery studies (98–105%). © 2006 Elsevier Ireland Ltd. All rights reserved. Keywords: Sequential injection; Titanium; Spectrophotometry; Chromotropic acid; Rocks; Dental implant

1. Introduction Metallic titanium is well known for its excellent corrosion resistance, being able to withstand attack by dilute sulphuric and hydrochloric acid, most organic acids, moist chlorine gas and chloride solutions. It is as strong as steel, but much lighter and it is heavier than aluminum, but twice as strong. These properties make titanium very resistant to the usual kinds of metal fatigue [1]. Approximately 95% of titanium production is consumed in the form of titanium dioxide, TiO2 , an intensely white permanent pigment with good covering power in paints, paper, toothpaste and plastics. Titanium alloys are principally used for aircrafts and missiles where lightweight strength and ability to withstand extremes of temperature are important. It also has potential use in desalination plants for converting sea water in fresh water. Being judged completely inert and immune to corrosion by all body fluids and tissue and thus totally bio-compatible, titanium

∗

Corresponding author. Tel.: +30 2310 997804; fax: +30 2310 997719. E-mail address: [email protected] (D.G. Themelis).

0039-9140/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.talanta.2006.07.018

is one of the few materials that naturally match the requirements for implantation in the human body. Its capacity for joining with bone and other tissue—osseo integration makes it suitable for medical applications, such as total replacement of arthritis hips, knee joints, facial treatments and dental implants [2]. Therefore, the need for the routine determination of titanium either in geological samples, such as silicate rocks or in dental implants have been receiving increasing attention. As regards the automated methods of analysis, flow injection (FI) is an important technique for this purpose as it involves low consumption of reagents and sample, exhibits rapid response and is easy to manipulate. Several methods for the determination of titanium(IV) by FI have been reported using various detection systems, such as spectrophotometry [3–9], chemiluminescence [10,11], inductively coupled plasma atomic emission spectrometry [12], inductively coupled plasma–mass spectrometry [13,14] and isotope dilution high resolution inductively coupled plasma–mass spectrometry [15]. The majority of the above methods were applied to the determination of Ti(IV) in silicate rocks [3–7,15], while some of them were applied to the determination of Ti(IV) in water [8,14], brines [8], aluminum alloy [9] and microalloyds [13].

1406

F.S. Kika, D.G. Themelis / Talanta 71 (2007) 1405–1410

However, FI has some inherent drawbacks that have been addressed by the development of sequential injection analysis, SI [16]. Briefly, the advantages of SI over FI are the simpler flow manifold, the reduced consumption of samples and reagents, the easier and move convenient variation of the experimental parameters and the greater potential for fluidic handling. To the best of our knowledge, no SI method has been reported so far for the spectrophotometric determination of Ti(IV). The present work reports the first SI method for the spectrophotometric determination of Ti(IV) and it is based upon the reaction of Ti(IV) with chromotropic acid, CA, in acidic medium to form a water-soluble complex (λ = 420 nm). The proposed method is simple, rapid, cost-effective, adequately selective and was applied successfully to the accurate and precise analysis of dental implant and natural Moroccan phosphate rock. 2. Experimental 2.1. Apparatus A schematic diagram of the SI setup is depicted in Fig. 1. The SI manifold consists of a micro-electrically actuated 10port selection valve (Valco, Switzerland), a double-beam spectrophotometer (Tecator 5023 FI star, Sweden) with a 1 cm path length flow cell, a detector controller (Tecator 5032 FI star, Sweden), and a peristaltic pump (Gilson Minipuls3, Villiers-le-Bel, France) [17]. The absorbance was monitored at 420 nm. The flow system used 0.7 mm i.d. Teflon tubing throughout. Tygon pump tubes were used for aspirating-delivering the solutions. 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). The response signal of the detector was acquired digitally and the data were saved in ASCII for-

mat for further manipulation (peak height measurement, digital filtering, etc.). 2.2. Chemicals and reagents All chemicals were of analytical-reagent grade and were provided by Merck (Darmstadt, Germany), unless stated otherwise, and all the solutions were made up by doubly de-ionized water. The standard solution of CA [γ(CA) = 8000 mg l−1 ] was prepared daily by dissolving 0.2 g of chromotropic acid disodium salt dehydrate in 25 ml of water. All the CA solutions were covered with aluminum foil during storage and measurements as CA is sensitive to light. The standard solution of ascorbic acid (AsA) (γ = 5000 mg l−1 ) was prepared daily by dissolving 1 g of solid AsA in 200 ml doubly de-ionized water. Working solutions of CA were prepared by dilution with buffer solution, while the AsA working solutions were prepared by dilution with double de-ionized water immediately before use. Working Ti(IV) solutions were made daily by appropriate dilutions of the stock Ti(IV) solution (γ = 1000 ± 2 mg l−1 in 5 mol l−1 HCl, Merck) and appropriate amounts of HCl solution were added in order all the resulted solutions have a final amount concentration of 0.05 mol l−1 . The addition of the HCl solution both prevents the hydrolysis of the Ti(IV) solutions and the standard solutions have the acidity that the real samples have after their pretreatment (see Section 2.4). Constant ionic strength (I = 0.2 mol l−1 ) acetate buffer solution (pH 4.80) was prepared from CH3 COOH and NaOH (1 mol l−1 each). For buffers with higher ionic strength values in the optimization procedure, the ionic strength was adjusted by the addition of appropriate amounts of solid KCl. Appropriate concentration of H3 BO3 (w = 4%) was used for the pretreatment of the dental implant sample. All solutions were degassed with purified nitrogen before use. All stock solutions were stored in polyethylene containers. All the polyethylene containers and glassware used for aqueous solutions containing metal ions were cleaned in (1 + 1) HNO3 , while the rest of the glassware was cleaned in w = 3% Decon 90. All were rinsed with de-ionized water before use. 2.3. Procedure for aqueous solutions

Fig. 1. SI manifold for the determination of titanium. S: sample; R1 : buffered 3500 mg l−1 chromotropic acid (pH 4.8); R2 : 4000 mg l−1 ascorbic acid; C: carrier (H2 O); PP: peristaltic pump; W: waste; AW: auxiliary waste; HC: holding coil; numbers above coils denote length/i.d. (cm/mm) ratio.

The sequence for the determination of Ti(IV) by the proposed method is shown in Table 1. Fifty microlitres of 4000 mg l−1 AsA solution, 100 ␮l of the standard/sample, 50 ␮l of 4000 mg l−1 AsA solution and 150 ␮l of 3500 mg l−1 CA solution in buffer pH 4.8 were aspirated in this order in the holding coil (HC), through ports 1, 2 and 3 of the selection valve, respectively. The four zones were propelled to the detector through port 5 at a flow rate of 1.0 ml min−1 . When the reaction mixture entered the flow cell of the detector, the flow was stopped and the reaction was allowed to proceed for 40 s, in order to gain maximum sensitivity. The cycle time was 150 s (24 determinations per hour). When changing between different solutions of samples, standards or reagents an additional wash-

F.S. Kika, D.G. Themelis / Talanta 71 (2007) 1405–1410

1407

Table 1 Sequence steps of a complete SI measurement cycle for the UV method Time (s)

Pump action

Flow rate (ml min−1 )

Valve position

Action description

0 5 1 10 1 5 1 15 1 22 40 49 0

Off Aspirate Off Aspirate Off Aspirate Off Aspirate Off Deliver Off Deliver Off

– 0.6 – 0.6 – 0.6 – 0.6 – 1.0 – 1.2 –

1 1 2 2 1 1 3 3 5 5 5 5 5

Selection of AsA port Aspiration of AsA in holding coil Selection of sample port Aspiration of sample in holding coil Selection of AsA port Aspiration of AsA in holding coil Selection of CA, buffer port Aspiration of CA, buffer in holding coil Selection of detector port Propulsion of reaction mixture to detector Stopped-flow step Propulsion of reaction mixture to waste End of measuring cycle

ing step was performed in order to avoid carryover effects; the new solution was aspirated to the HC for 10 s at 2.0 ml min−1 , and then flushed through port 4 to the auxiliary waste (AW) for 20 s at 2.0 ml min−1 . Note that in order to avoid overpressure and/or bubble formation in the valve, the pump was stopped for 1 s between changing ports. The transient signal from the detector was recorded as a peak, the height of which was proportional to Ti(IV) concentration in the samples, and was used for all measurements. The recorded peaks were sharp and the base-line stable. The blank signal was recorded by replacing the sample/standard solution with 0.05 mol l−1 HCl solution. Five replicates per sample or standard were made in all instances. 2.4. Pretreatment and determination of titanium in dental implant and natural Moroccan phosphate rock An accurately weighed 0.1 g dental implant was transferred to a beaker. Four and a half microlitres of doubly de-ionized water were added and the sample was dissolved with 0.5 ml of 40% mass fraction of HF. When the sample was completely dissolved, 0.5 ml concentrated HNO3 was added drop wise until the titanium was completely oxidized. When the oxidation of titanium was complete, decoloration of the solution was observed. Afterwards, 2.6 ml of 4% mass fraction of H3 BO3 solution were added and the solution was transferred to a 100-ml volumetric flask with 2 mol l−1 HCl [18]. Then, 0.25 ml of the resulted solution was further diluted with doubly de-ionized water to 100 ml and final amount concentration of 0.05 mol l−1 HCl and analyzed using the above-described SI procedure for aqueous solutions. Five replicates were made in all instances. The analysis of Moroccan rock was carried out by fusing 1 g of rock [Community Bureau of Reference (BCR) No. 32] with a mixture of Na2 CO3 and K2 CO3 of mass fraction ratio of 6:1 for 3 h at 800–1000 ◦ C. The melted mass was cooled and treated with 1% volume fraction of HCl and evaporated to dryness. The residue was dissolved in 10% volume fraction of HCl and filtered. The filtrate was diluted to 100 ml and analyzed using the above-described SI procedure for aqueous solutions [6]. The

resulted solution was analyzed using the above-described SI procedure for aqueous solutions. Five replicates were made in all instances. 3. Results and discussion 3.1. Preliminary studies Preliminary experiments showed that the reaction product could be formed under SI conditions, having an absorbance maximum at 420 nm. The SI setup consisted of two zones, the sample and the buffered reagent zone. The initial values of the chemical and SI variables were: γ(CA) = 5000 mg l−1 in acetate buffer pH 5.2, γ(Ti(IV)) = 5 mg l−1 in 0.05 mol l−1 HCl, I = 0.1 mol l−1 , V(sample) = 100 ␮l, V(CA in acetate buffer) = 100 ␮l (10 s aspiration time at 0.6 ml min−1 ), while the reaction mixture was propelled to the detector at a flow rate of 1.2 ml min−1 . The order of the aspiration of the sample and the reagent solution proved not to be critical. The almost negligible differences in the signals were caused by slightly different dispersion effects on the sample zone according to the order of its aspiration. In addition, stopped-flow experiments showed that the reaction was completed in 40 s. 3.2. Study of chemical and SI variables The chemical and SI variables were studied using an assay which comprises of the sample and the buffered reagent zone. In all cases, calibration curves for Ti(IV) were constructed in the range of 1.0–10.0 mg l−1 in order to study the sensitivity, the determination range and the linearity of the method. The starting SI variables were: V(sample) = 100 ␮l, V(CA in acetate buffer) = 100 ␮l, l(RC) = 0 cm and qv = 1.2 ml min−1 . The influence of pH on the reaction was studied in the range 4.2–5.6, using a series of standard acetate buffers. The results showed that the reaction pH has a marked effect on the linearity for pH values varying between 4.2 and 4.8 without affecting significantly the sensitivity of the method. The best linearity was achieved at a pH value of 4.8 and this value was selected for the subsequent experiments.

1408

F.S. Kika, D.G. Themelis / Talanta 71 (2007) 1405–1410

The effect of the ionic strength was studied in the range 0.1–0.7 mol l−1 . The results showed that the reaction was not influenced by the increase of the ionic strength. Therefore, the value of 0.2 mol l−1 was selected for the following experiments for simplicity reasons. The effect of the CA mass concentration was investigated in the range 1000–5000 mg l−1 . The signal intensity increased by increasing the CA mass concentration in the range of 1000–3500 mg l−1 and then decreased. The maximum signal intensity was achieved for CA mass concentration of 3500 mg l−1 . Additionally, at this value, the maximum working range and linearity of the calibration curve were achieved. Thus, a CA mass concentration of 3500 mg l−1 was chosen for subsequent work. The influence of the HCl amount concentration in the sample solution was studied in the range 0.05–0.2 mol l−1 . At HCl amount concentrations of 0.1 and 0.2 mol l−1 there was a significant decrease in signal intensity due to pH change. Therefore, HCl amount concentration of 0.05 mol l−1 was selected for further experiments. The effect of the stop-flow period was investigated from 20 to 60 s. The results showed that the stop-flow period has a noticeable influence on the sensitivity of the method. For time period lower than 40 s, the signal intensity was increased linearly, leveled off up to 50 s and then decreased. Thus, the stop-flow period of 40 s was selected as higher sensitivity was achieved while the working range and the linearity did not differ in both cases. Using the preferred chemical parameters mentioned above, the SI variables were studied. Experiments were carried out to determine the optimum volumes of the aspirated sample and reagent solution. The effect of the buffered CA solution volume was studied in the range 50–200 ␮l. In this case, the peak heights increased up to 100 ␮l and then leveled off. Maximum linearity was achieved for CA solution volumes above 150 ␮l. Additionally, the signal of the blank solution increased linearly with increasing buffered CA solution volume and therefore the volume of 150 ␮l was selected. The effect of the aspirated sample volume was studied in the range 50–200 ␮l. The peak heights increased non-linearly with increasing sample volume and a volume of 100 ␮l was selected for subsequent experiments as it is preferable the volume of the aspirated acidic sample solution to be lesser than the volume of the aspirated reagent solution in order the buffer to maintain its capacity and the pH of the reaction not to be altered. With the selected sample and buffered reagent volumes and at a propulsion flow rate of 1.2 ml min−1 , the influence of the reaction coil length was examined in the range 0–100 cm. The length of the reaction coil determined both the degree of overlapping of the sample and reagent zones and the period of time that the reaction was allowed to proceed. The results showed that using the necessary PTFE connection tubes between the selection valve and the detector produced adequate overlapping of the sample and reagent zones. Since the stopped-flow step was adopted, no additional reaction coil was used prior to the detector.

Table 2 Maximum tolerance mole ratio of various ions on the determination of 1.0 mg l−1 Ti(IV) under the selected chemical and SI conditions for aqueous solutions Ion added

Tolerable ratio of mass concentrationsa

Cd(II), Ca(II) SO4 2− PO4 3− , CO3 2− Cu(II), Mg(II) Co(II), Ni(II), Mn(II) B(III) Zn(II), V(V), Al(III) Fe(III) Cr(VI)

1000 600 400 150 100 50 20 5 2

a

The tolerable ratio is calculated for a ±5% relative error in the absorbance.

3.3. Study of interferences The effect of foreign ions on the absorbance signal was studied under the selected conditions described above. The criterion for interference was fixed at a relative error, er , of less than ±5% of the average absorbance signal at a Ti(IV) mass concentration of 1.0 mg l−1 . The results are summarized in Table 2. As can be seen, the only serious interferences were found to be Fe(III) and Cr(VI) as they caused a positive error (increase of the absorbance intensity). However, Cr(VI) could not interfere in the samples analyzed in this work. NaF, hydroxylammonium chloride and ascorbic acid (AsA) were used in order to counter the Fe(III) interference. The masking reagents were added in the buffered reagent solution and calibration curves were recorded in each case while the masking of Fe(III) was also checked. When AsA was used the linearity and the dynamic working range were retained and simultaneously the best masking efficiency was obtained. Subsequently, this reagent was selected in order to study and improve further the selectivity of the method. 3.4. Selectivity of the method Several SI manifolds were tested at AsA mass concentrations namely 1000, 2000 and 4000 mg l−1 in order to improve the masking efficiency of AsA. The most effective SI manifold was consisted of three zones by punctuating the zone of AsA solution (100 ␮l) by the zone of the sample solution, having in this way a system with a “sandwich” look. The sequence steps are described in Table 1. Using this manifold calibration curves were recorded at three AsA mass concentrations of 1000, 2000 and 4000 mg l−1 . The results proved that the linearity and the working range were independent of the AsA mass concentration. In addition, the masking efficiency of the AsA reagent at the above three mass concentrations of AsA was checked. The results showed that an AsA mass concentration of 4000 mg l−1 gave the best results. Apart from eliminating the interference of Fe(III) the use of the “sandwich” manifold and of AsA as masking reagent improved significantly the tolerance of the assay against all the ions that were tested for interference effects in Section 3.3, except for Ni(II), Mn(II) and PO4 3− .

F.S. Kika, D.G. Themelis / Talanta 71 (2007) 1405–1410

1409

Table 3 Recovery experiments SI method Natural Moroccan phosphate rock [(BCR) No. 32]

Dental implant [3i (IOSS515)]

a b

Ti(IV) added (mg l−1 )

Ti(IV) founda (mg l−1 )

Recovery (%)

– 1.00 2.00 5.00

1.80 2.85 3.8 7.0

± ± 0.09 ± 0.1 ± 0.1

105 100 104

– 1.00 2.00 5.00

2.30 3.35 4.4 7.2

± ± ± ±

105 105 98

0.06b

0.07 0.06 0.1 0.2

Mean of five results. Standard deviation.

Table 4 Accuracy of the proposed method SI method

Natural Moroccan phosphate rock [(BCR) No. 32] Dental implant [3i (IOSS515)] a b c d e f

Ti(IV) found g−1 samplea (␮g) Proposed method

Reference valueb

er c (%)

180d ± 6e 92f ± 2e

171d 89f

−5.3 +3.4

Mean of five results. Value of the certified reference materials. Relative error. ␮g g−1 . Standard deviation. Mass fraction (%).

3.5. Features of the proposed method Under the selected chemical and SI variables and using the SI setup shown in Fig. 1, a calibration curve for Ti(IV) was recorded. The calibration curve was linear in the range 0.2–10.0 mg l−1 and was described by the regression equation, A = [(67.9 ± 0.2) × 10−3 ] × γ(Ti(IV)) + [(28.1 ± 3.0) × 10−3 ] where A is the absorbance and γ(Ti(IV)) is the mass concentration of the analyte, with a relative standard deviation, sr , of 1.5% at 5 mg l−1 Ti(IV) (n = 12), a correlation coefficient, r, of 0.9999, and a 3σ detection limit of 0.7 mg l−1 (n = 10). All the standards were run in 5 replicate injections, while for the detection limit calculation the blank was run in 10 replicate injections. 3.6. Determination of titanium in natural Moroccan phosphate rock and dental implant

recovery experiments by spiking the samples solutions with known amounts of Ti(IV). The experimental results were satisfactory in all cases, with recoveries ranging between 100 and 105% for the natural Moroccan phosphate rock and 98–105% for the dental implant (Table 3). 4. Conclusions This article describes the first SI spectrophotometric method for the selective determination of Ti(IV) with application in dental implant and natural Moroccan phosphate rock. The developed method is simple, rapid, uses commercially available reagents and its recovery results for the determined samples are satisfactory. Acknowledgement The project was partially funded by “Herakleitos”, EPEAEK, Greek Ministry of Education. References

The proposed SI assay was applied to the determination of titanium in natural Moroccan phosphate rock [(BCR) No. 32] and dental implant [3i (IOSS515)]. In the natural Moroccan phosphate rock, the titanium was calculated as 180 ␮g g−1 and in the dental implant 92% (Table 4). Both values were in good agreement with the theoretical value of the certified reference materials. The accuracy of the SI method was checked with

[1] [2] [3] [4] [5]

http://en.wikipedia.org/wiki/Titanium. http://www.titaniuminfogroup.co.uk/. T. Mochizuki, R. Kuroda, Analyst 107 (1982) 1255. S. Kozuka, K. Saito, K. Oguma, R. Kuroda, Analyst 115 (1990) 431. M.G.M. Andrade, S.L.C. Ferreira, B.F. Santos, A.C.S. Costa, Talanta 39 (1992) 1229. [6] R. Purohit, S. Devit, Analyst 117 (1992) 1175.

1410 [7] [8] [9] [10] [11] [12]

F.S. Kika, D.G. Themelis / Talanta 71 (2007) 1405–1410 R.E. Santelli, R.C. dos Santos Araujo, Analyst 117 (1992) 1519. M. Munoz, J. Alonso, J. Bartroli, M. Valiente, Analyst 115 (1990) 315. J. Wang, R. He, Anal. Chim. Acta 276 (1993) 419. A.A. Alwarthan, A. Townshend, Anal. Chim. Acta 196 (1987) 135. A.A. Alwarthan, A. Almuaibed, A. Townshend, Anal. Sci. 7 (1991) 623. S. Hirata, Y. Umezaki, M. Ikeda, Anal. Chem. 58 (1986) 2602.

[13] [14] [15] [16] [17] [18]

A.G. Coedo, M.T. Dorado, Appl. Spectrosc. 49 (1995) 115. K.L. Yang, S.J. Jiang, T.J. Hwang, J. Anal. At. Spectrom. 11 (1996) 139. A. Makishima, E. Nakamura, J. Anal. At. Spectrom. 15 (2000) 263. C.E. Lenehan, N.W. Barnett, S.W. Lewis, Analyst 127 (2002) 997. A. Ivaska, J. Ruzicka, Analyst 118 (1993) 885. D. Myers, At. Absirp. Newsl. 6 (1967) 89.