Available online at www.sciencedirect.com

Journal of Chromatography A, 1187 (2008) 34–39

Preparation and characterization of new solid-phase microextraction fibers obtained by sol–gel technology and zirconium oxide electrodeposited on NiTi alloy Dilma Budziak, Edmar Martendal, Eduardo Carasek ∗ Departamento de Qu´ımica, Universidade Federal de Santa Catarina, Florian´opolis, SC 88040-900, Brazil Received 6 December 2007; received in revised form 3 February 2008; accepted 5 February 2008 Available online 8 February 2008

Abstract This study describes the use of zirconium oxide electrolytically deposited onto a NiTi alloy as a new substrate for sol–gel reactions. Polydimethylsiloxane (PDMS) was used to coat the fiber after activation of the NiTi–ZrO2 surface with sodium hydroxide solution followed by hydrochloric acid solution. Micrographs obtained by scanning electron microscopy (SEM) showed good uniformity of the PDMS coating on the proposed substrate and also permitted the evaluation of coating thickness, being approximately 25 ␮m. Thermal stability of the coating on the NiTi–ZrO2 surface was evaluated, showing excellent stability up to 320 ◦ C. The applicability of the proposed NiTi–ZrO2 –PDMS fiber was evaluated through extraction of benzene, toluene, ethylbenzene and o-xylene (BTEX) from the headspace of aqueous samples. Some parameters affecting the extraction efficiency such as the salting-out effect, extraction temperature and extraction time were optimized by two consecutive two-level full-factorial experimental designs. This optimization allowed the experimental domain of maximum response to be attained and also the robustness range for the variables. Excellent detection limits in the range of 0.6–1.6 ␮g L−1 were obtained as well as correlation coefficients higher than 0.99994. Precision for one fiber (n = 7) was in the range of 1.4–4.0% and fiber-to-fiber reproducibility (n = 5) was in the range of 3.9–6.7%. Comparison of the extraction profile of the proposed fiber with those of commercially available 100, 30 and 7 ␮m PDMS fibers showed that NiTi–ZrO2 –PDMS had a better response compared to that of the 7 and 30 ␮m fibers. Such characteristics of the NiTi–ZrO2 –PDMS fiber suggest that this fiber represents an excellent alternative for gas chromatography sample preparation. © 2008 Elsevier B.V. All rights reserved. Keywords: Sol–gel reaction; PDMS; NiTinol; Zirconium oxide; BTEX; SPME

1. Introduction Solid-phase microextraction (SPME) represents an important advance in the extraction efficiency of many organic contaminants at trace levels [1]. It is a technique based on the partitioning of the organic analytes between the sample matrix and the extracting phase, commonly with immobilization onto a fused silica rod. SPME presents advantages such as simplicity, low cost and fast sample preparation compared to liquid–liquid extraction, solid-phase extraction and supercritical fluid extraction [2]. Due to the fragility of the silica rods traditionally used as supports for coatings in SPME, several studies have proposed

∗

Corresponding author. Fax: +55 48 37216845. E-mail address: [email protected] (E. Carasek).

0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.02.003

different materials to replace silica rods, where emphasis has been given to metallic supports, such as: platinum wire [3–5], anodized aluminum wire [6], gold wire [7,8], titanium wire [9], stainless steel [10–12] and copper wire [13]. However, the absence of an adequate interaction between the substrate and the coating can result in low chemical and thermal stability [14–17]. Sol–gel chemistry, first introduced in SPME by Malik and co-workers [14] offers a simple and convenient manner to synthesize advanced materials for application as surface coatings. This technique is commonly used to coat organic components onto inorganic polymeric structures under mild thermal conditions [14]. The sol–gel coating technique combines fiber surface treatment, coating and immobilization of the extracting phase into a single and simple step. The porous structure of the sol–gel coating offers a high surface area; allowing high extraction efficiency and the coating composition can be altered with a relative ease to

D. Budziak et al. / J. Chromatogr. A 1187 (2008) 34–39

give different selectivity characteristics. Strong adhesion of the coating onto the support due to chemical bonding is a very important characteristic which increases the coating stability toward organic solvents and high desorption temperatures [14–17]. The advantages of the NiTi alloy in terms of the shape memory, superelasticity, durability, corrosion resistance and biocompatibility [18], along with the advantages of zirconium oxide coatings, including strong adhesion to metallic surfaces, excellent biocompatibility, high thermal stability and wear and corrosion resistance [19,20], suggest that NiTi electrolytically coated with zirconium oxide could offer a promising alternative as a new support for the SPME technique. In this study, a new SPME fiber is proposed, combining the advantages of NiTi coated with ZrO2 previously reported in the literature [21,22] with those of sol–gel chemistry. NiTi–ZrO2 is used as a substrate (to replace the silica rod) for polydimethylsiloxane (PDMS) coating using sol–gel technology. The new fiber was characterized and evaluated in the extraction of BTEX from an aqueous matrix. 2. Experimental 2.1. Instrumentation Chromatographic analysis was performed on a Shimadzu GC-14B gas chromatograph, equipped with split/splitless injector and flame ionization detector. Chromatographic separation was carried out in an OV-5 capillary column (30 m × 0.25 mm, 0.25 ␮m in film thickness; OV Specialty Chemical, Marietta, OH, USA). Ultrapure nitrogen was used as the carrier and make-up gas at 1.0 and 35 mL min−1 , respectively. Column oven temperature was 40 ◦ C (1 min), 10 ◦ C min−1 to 100 ◦ C, 15 ◦ C min−1 to 180 ◦ C. Injector and detector temperatures were fixed at 300 and 280 ◦ C, respectively. The morphology and composition of the fibers, with and without PDMS coating, were evaluated by scanning electron microscopy (SEM), using a Philips XL-30 microscope, and semi-quantitative microanalysis using energy dispersion spectroscopy (EDS). 2.2. Chemicals BTEX standards including benzene, toluene, ethylbenzene and o-xylene (Sigma–Aldrich, Milwaukee, WI, USA) were prepared in HPLC-grade methanol (Tedia, Fairfield, OH, USA). Sodium chloride (Merck, Darmstadt, Germany) was used for the modification of the ionic strength of the samples. Sodium hydroxide (Vetec, Rio de Janeiro, Brazil) and hydrochloric acid (Merck) were both used for activation of the NiTi–ZrO2 surface before sol–gel coating. The sol–gel materials were methyltrimethoxysilane (MTMS) (UCT, Bristol, PA, USA), trimethylmethoxysilane (TMMS), hydroxy-terminated polydimethylsiloxane (PDMS-OH), polymethylhydrosiloxane (PMHS) (Sigma–Aldrich,) and trifluoroacetic acid (TFA) (Vetec).

35

NiTi wires were donated by Nano Endoluminal (Florian´opolis, SC, Brazil). 2.3. Electrodeposition Wires of NiTi with 0.2 mm thickness and approximately 2 cm length were used as supports for the electrodeposition. Only 1 cm was electrodeposited with ZrO2 . The optimum coating conditions were obtained from Giacomelli et al. [23]. More details on the electrodeposition procedure employed for fiber manufacturing are presented in our previous publications [21,22]. 2.4. Preparation of sol–gel fibers NiTi alloy previously electrodeposited with ZrO2 was dipped into NaOH 1.0 mol L−1 solution for 1 h for activation of the fiber surface. After this time, the fibers were washed with deionized water and immersed in an HCl 0.1 mol L−1 solution for 30 min, washed in deionized water and used for sol–gel coating in up to 5 h. PDMS sol solution was prepared by mixing 300 ␮L of methyltrimethoxysilane, 180 mg of hydroxy-terminated PDMS, 30 mg of PMHS, and 200 ␮L of 95% TFA (containing 5% water) in glass capillary tubes. One centimeter of the activated NiTi–ZrO2 fibers was vertically dipped into the clear sol solution for approximately 20 min. This procedure was repeated three times, always with a freshly prepared sol solution. Fibers were mounted in SPME commercial assemblies and conditioned for 30 min in the GC injector port at 250 ◦ C. Three fibers were also prepared in exactly the same way, but without the ZrO2 film on the NiTi alloy. 2.5. Optimization of the method for extraction of BTEX The performance of the new SPME fiber, named NiTi–ZrO2 –PDMS, was evaluated through the extraction of BTEX compounds from aqueous samples. Headspace sampling, magnetic stirring, 20 mL of aqueous sample in a 40 mL vial were used. Parameters influencing extraction, such as saltingout effect (sodium chloride was used), extraction temperature and extraction time were optimized using two consecutive fullfactorial experimental designs. 3. Results and discussion The importance of suitably activating the fused silica when it is used as a substrate for sol–gel reactions is described in the literature [24,25]. This activation leads to a greater exposure of the silanol groups to the silica rod, favoring the chemical bonding between the support and the polymeric coating, which in turn leads to a higher thermal stability of the fiber. Therefore, for sol–gel reactions on metallic surfaces, the presence of hydroxyl groups is also required. Azenha et al. [9] obtained good results by activating the native titanol groups of a titanium wire used as a support to produce sol–gel fibers.

36

D. Budziak et al. / J. Chromatogr. A 1187 (2008) 34–39

Fig. 1. Scanning electron micrographs of the NiTi–ZrO2 –PDMS fiber (A) before and (B) after thermal conditioning.

The approximately equiatomic NiTi alloy electrodeposited with zirconium oxide was successfully applied as an SPME fiber for the analysis of alcohols, BTEX and trihalomethanes from gaseous samples [21], as well as in the determination of halophenols in river water [22]. However, its small thickness and hydrophilic character are two features which limit the applicability of this fiber for determination of compounds in water samples that can be detected with a very sensitive detector, such as an electron capture detector. By activating the NiTi–ZrO2 substrate with 1.0 mol L−1 NaOH followed by 0.1 mol L−1 HCl, we may obtain a greater exposure of Zr–OH groups on the surface of the support, allowing its application as a new substrate for sol–gel reactions. 3.1. Characterization of NiTi–ZrO2 –PDMS fiber The surface of the fiber obtained after three cycles of sol–gel reactions on the activated NiTi–ZrO2 was characterized by SEM and EDS. Fig. 1 illustrates two micrographs of the fiber coated with PDMS. Fig. 1A shows a fiber without previous conditioning in the GC injector port, where points with an excess of polymeric coating can be seen and confirmed to be PDMS by punctual analysis through EDS. In Fig. 1B, the fiber was conditioned for 30 min at 250 ◦ C in the GC injector port. The surface is now more uniform and homogeneous, without the points with an excess of PDMS coating. Six fibers were coated by sol–gel procedure; three of them were thermally conditioned and presented lower amount of PDMS coating on its surface and better homogenization compared to the other three which were not conditioned,

Fig. 2. NiTi–ZrO2 surface-bonded sol–gel PDMS coating.

and presented the same point with an excess of PDMS coating than that of Fig. 1A. In this way, it is suggested that thermal conditioning aids on the surface homogenization. Zr–OH groups on the surface of the substrate can also participate through condensation reactions and thus promote chemical linkage to the polymeric chain. A schematic of the fiber coated with PDMS is given in Fig. 2. Following the suggestion of Azenha et al. [9], only the metallic support (NiTi without the ZrO2 film) was activated with 1.0 mol L−1 NaOH for 1 h, followed by 0.1 mol L−1 HCl, water and methanol, and used as a substrate for the sol–gel coating procedure. The micrographs of such fibers are shown in Fig. 3.

Fig. 3. SEM of the NiTi fibers coated with PDMS without the ZrO2 film (A) before and (B) after thermal conditioning.

D. Budziak et al. / J. Chromatogr. A 1187 (2008) 34–39

37

Fig. 4. Cross-section of the polymeric coating of the NiTi–ZrO2 –PDMS fiber.

The PDMS coating without ZrO2 on the NiTi did not result in satisfactory adhesion onto the support, mainly after thermal conditioning (Fig. 3B), where areas can be seen in which there is no polymer attached to the substrate. The micrographs correspond to two different fibers. When there is no ZrO2 film on the NiTi alloy, the polymeric coating is not uniform and in some cases the coating thickness is not reproducible, leading to the difference in the thickness, as presented in Fig. 3A and B. Thus, for a higher uniformity and thermal stability of the sol–gel coating, the ZrO2 electrodeposited onto the NiTi alloy was shown to be necessary. A cross-section of the coating of a previously conditioned fiber was performed to evaluate the coating thickness. This micrograph is illustrated in Fig. 4. The coating thickness was also evaluated by comparing the micrographs of fibers with and without the PDMS. Both approaches lead to the same result, that is, coating thickness of approximately 25 ␮m. A chemical evaluation of the fiber-coating surface was carried out through EDS, as shown in Fig. 5. The spectrum had a peak at 0.2 keV and a peak at 1.8 keV, which were attributed to the K␣ emission line of carbon and silicon, respectively. These elements are characteristics of PDMS coatings. At 2.2 keV, the peak is attributed to the M␣ emission line of gold, which was used to coat the fibers so that EDS could be performed.

Fig. 5. EDS of the surface of the NiTi–ZrO2 –PDMS fiber.

Fig. 6. Evaluation of thermal stability of PDMS coating on the NiTi–ZrO2 support.

3.2. Evaluation of the NiTi–ZrO2 –PDMS fiber: thermal stability and BTEX extraction of aqueous samples The applicability of the superelastic SPME fiber, NiTi–ZrO2 –PDMS, was demonstrated through the extraction of BTEX compounds from aqueous samples. Some parameters which affect the extraction efficiency were optimized, and the main figures of merit were estimated. Before applying the fibers, the thermal stability of the PDMS coating on the NiTi–ZrO2 support was evaluated in the temperature range of 240–320 ◦ C. Five replicates of each evaluation were performed and the desorption time was fixed at 10 min. The result can be verified in Fig. 6. There was no loss of extraction efficiency when increasing the desorption temperature to 320 ◦ C, that is, there was no loss of the polymeric coating. This result suggests that the polymeric coating is chemically bonded to the ZrO2 film electrodeposited on the NiTi alloy. This characteristic is very positive when compared to the thermal stability of commercial PDMS fibers (100 and 30 ␮m), where the maximum recommended temperature is 280 ◦ C. A higher thermal stability is highly desirable, since the desorption time is thus reduced, improving the shape of chromatographic peaks and eliminating carry-over effects. Thus, the desorption temperature was fixed at 300 ◦ C and the desorption time at 3 min, with no carry-over effects observed. The main variables influencing the extraction process by the fiber were simultaneously optimized through a two-level fullfactorial experimental design with inclusion of a four-replicate center point. The following variables were studied: salting-out effect (0.0 and 6.0 g NaCl per 20 mL sample), extraction temperature (3.0 and 35.0 ◦ C) and extraction time (5.0 and 20.0 min), resulting in 12 experiments. The Pareto chart in Fig. 7 shows the results. The most significant effect observed in Fig. 7 corresponds to the salt addition, with a positive value (32.49). This indicates higher extraction efficiency due to the salting-out effect. The interaction effect between ionic strength and extraction temperature was also significant, indicating that these variables should be

38

D. Budziak et al. / J. Chromatogr. A 1187 (2008) 34–39

Fig. 7. Pareto chart obtained by the full-factorial design for evaluation of saltingout effect (NaCl), extraction time and extraction temperature.

Fig. 8. Pareto chart obtained for the optimization of time and extraction temperature through full-factorial design.

studied. However, this interaction effect must be interpreted with the aid of the main effects of these variables, which indicate that better responses are obtained with high ionic strength and low extraction temperature. Thus, as the high level of ionic strength studied was the sample saturation (30%), we decided to fix this variable in its high level and evaluate extraction time and temperature in the next factorial design. The extraction temperature effect was also statistically significant, but with a negative value, indicating higher extraction efficiency at the lower level studied (3.0 ◦ C). This behavior in relation to temperature is related to the exothermic nature of the sorption process and the high volatility of the compounds even at low temperature. In other words, the determining step in the kinetic of mass transfer of the analytes from the aqueous sample to the fiber coating is the diffusion of the compounds through the fiber, and not their evaporation from the aqueous sample. Extraction time was also significant at the levels studied, demonstrating that the equilibrium is not reached within 5 min of extraction. The final optimization of the parameters found to be statistically significant according to the Pareto chart is usually performed using a response surface methodology. Alternatively, a second factorial design can be carried out, where the minimum and maximum levels of each variable are rearranged according to the results of the first factorial design. Thus, the variable’s extraction time and extraction temperature were once again optimized using a new full-factorial design with a triplicate center point, resulting in 7 experiments. Extraction time was evaluated at the levels 12.0 and 20.0 min and extraction temperature at 3.0 and 9.0 ◦ C. The Pareto chart is shown in Fig. 8. No main or interaction effect of the variables was statistically significant at the interval studied. Thus, the interval for each variable corresponds to that at which the maximum responses and also the robustness range for these variables at 95% confidence level occur. Therefore, the extraction time used was 16 ± 4 min and extraction temperature was 6 ± 3 ◦ C. Some analytical features of the method were evaluated and the results are shown in Table 1. Excellent correlation coefficients were obtained for the analytical curve in the linear range studied (R > 0.99994). Detection

Table 1 Linear range, correlation coefficients, detection and quantification limits obtained for the method proposed to determine BTEX in water samples using the NiTi–ZrO2 –PDMS fiber Compounds

Linear rangea

Rb

LODc (␮g L−1 )

LOQd (␮g L−1 )

Benzene Toluene Ethylbenzene Xylene

5–200 5–200 2–200 2–200

0.99997 0.99994 0.99999 0.99999

1.6 1.6 0.6 0.6

5.2 5.2 1.9 2.1

a b c d

Linear range studied in ␮g L−1 . Correlation coefficient of calibration curve. Limit of detection. Limit of quantification.

limits (S/N = 3) were in the range of 0.6–1.6 ␮g L−1 for all compounds. Repeatability of the method (using the same SPME fiber), obtained for 7 replicates was in the range of 1.4–4.0%. Reproducibility (using 5 different fibers) was also evaluated, and the relative standard deviation was in the range of 3.9–6.7% for all compounds.

Fig. 9. Comparison of NiTi–ZrO2 –PDMS fiber with commercially available PDMS fibers.

D. Budziak et al. / J. Chromatogr. A 1187 (2008) 34–39

3.3. Comparison of the NiTi–ZrO2 –PDMS fiber with commercial PDMS Extraction efficiency of the NiTi–ZrO2 –PDMS fiber was compared to that of commercially available PDMS fibers (100, 30 and 7 ␮m). According to Chong et al. [14] the extraction efficiency of PDMS fibers coated by sol–gel reactions may slightly differ from those of commercial PDMS fibers. One of the reasons is that individual molecules of hydroxy-terminated PDMS used in the preparation of the sol solution contain terminal silanol groups, which are absent in PDMS molecules used in conventional coatings. As can be observed in Fig. 9, the NiTi–ZrO2 –PDMS fiber has an extraction capability superior to that of a 30 ␮m commercial PDMS fiber, even with a lower thickness of 25 ␮m. 4. Conclusions The results presented in this work show that the NiTi electrolytically coated with zirconium oxide is an excellent substrate for sol–gel reactions. The NiTi–ZrO2 –PDMS fiber presented excellent thermal stability up to 320 ◦ C, which enables the use of a high desorption temperature without loss of extraction efficiency. The fiber also has an excellent sensitivity toward the compounds evaluated, demonstrating its applicability. With an extraction profile superior to that of the 30 ␮m PDMS fiber and higher thermal stability, together with the superelastic and shape memory characteristics of the NiTi alloy, the proposed fiber is a reliable alternative to the use of commercial fibers. NiTi–ZrO2 –PDMS fiber has been applied for our research group for the determination of trace pesticides in herbal infusions, and it has maintained its extraction efficiency up to this moment, for about 250 extraction cycles.

39

Acknowledgment The authors thank the Brazilian National Counsel of Technological and Scientific Development (CNPq) for financial support and Nano Endoluminal Company for the supplying of NiTi wires. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]

J. Pawliszyn, R. Belardi, Water Pollut. Res. J. Can. 24 (1989) 179. T.P. Gbatu, K.L. Sutton, J.A. Carus, Anal. Chim. Acta 402 (1999) 67. H. Bagheri, A. Mir, E. Babanezhad, Anal. Chim. Acta 532 (2005) 89. A. Mohammadi, Y. Yamini, N. Alizadeh, J. Chromatogr. A 1063 (2005) 1. J. Wu, W.M. Mullett, J. Pawliszyn, Anal. Chem. 74 (2002) 4855. D. Djozan, Y. Assadi, S.H. Haddadi, Anal. Chem. 73 (2001) 4054. Dj. Djozan, S. Bahar, Chromatographia 59 (2004) 95. D. Djozan, Y. Assadi, Chromatographia 58 (2003) 637. M. Azenha, P. Nogueira, A. Fernando-Silva, Anal. Chem. 78 (2006) 2071. J.-G. Hou, Q. Ma, X.-Z. Du, H.-l. Deng, J.-Z. Gao, Talanta 62 (2004) 241. Y. Liu, Y. Shen, M.L. Lee, Anal. Chem. 69 (1997) 190. H. Minjia, T. Chao, Z. Qunfang, J. Guibin, J. Chromatogr. A 1048 (2004) 257. M.A. Farajzadeh, N.A. Rahmani, Talanta 65 (2005) 700. S.L. Chong, D. Wang, J.D. Hayes, B.W. Wilhite, A. Malik, Anal. Chem. 69 (1997) 3889. J. Yu, L. Dong, C. Wu, L. Wu, J. Xing, J. Chromatogr. A 978 (2002) 37. L. Yun, Anal. Chim. Acta 486 (2003) 63. Z. Zeng, W. Qiu, Z. Huang, Anal. Chem. 73 (2001) 2429. L. Tan, W.C. Crone, Acta Mater. 50 (2002) 4449. I. Zhitomirsky, A. Petric, Mater. Lett. 46 (2000) 1. S.K. Yen, M.J. Guo, H.Z. Zan, Biomaterials 22 (2001) 125. D. Budziak, E. Martendal, E. Carasek, J. Chromatogr. A 1148 (2007) 131. D. Budziak, E. Martendal, E. Carasek, Anal. Chim. Acta 598 (2007) 254. F.C. Giacomelli, C. Giacomelli, A.G. Oliveira, A. Spinelli, Mater. Lett. 59 (2005) 754. Y.-l. Hu, Y.-l. Fu, G.-K. Li, Anal. Chim. Acta 567 (2006) 211. M. Liu, Z. Zeng, H. Fang, J. Chromatogr. A 1076 (2005) 16.