Compositional Analysis of Electrodeposited Bismuth Telluride

Aug 2, 2008 - Bismuth telluride (Bi2Te3) is a benchmark material for thermoelectric power generation and cooling applications. Electrodeposition is a ...
0 downloads 10 Views 189KB Size
Download PDF
Anal. Chem. 2008, 80, 6724–6730

Compositional Analysis of Electrodeposited Bismuth Telluride Thermoelectric Thin Films Using Combined Electrochemical Quartz Crystal Microgravimetry-Stripping Voltammetry Sunyoung Ham,† Soyeon Jeon,‡ Ungki Lee,‡ Minsoon Park,‡ Ki-Jung Paeng,† Noseung Myung,*,‡ and Krishnan Rajeshwar*,§ Department of Chemistry, Yonsei University, Wonju Campus, Wonju, Kangwondo 220-710, Korea, Department of Applied Chemistry, Konkuk University Chungju Campus, Chungju, Chungbuk 380-701, Korea, and Center for Renewable Energy Science & Technology (CREST), The University of Texas at Arlington, Arlington, Texas 76109-0065 Bismuth telluride (Bi2Te3) is a benchmark material for thermoelectric power generation and cooling applications. Electrodeposition is a versatile technique for preparing thin films of this material; however, it affords films of variable composition depending on the preparation history. A simple and rapid assay of electrodeposited films, therefore, has both fundamental and practical importance. In this study, a new protocol for the electroanalysis of Bi2Te3 thin films is presented by combining the two powerful and complementary techniques of electrochemical quartz crystal microgravimetry (EQCM) and stripping voltammetry. First, any free (and excess) tellurium in the electrodeposited film was reduced to soluble Te2- species by scanning to negative potentials in a 0.1 M Na2SO4 electrolyte, and the accompanying frequency increase (mass loss) was used to determine the content of free tellurium. The film was again subjected to cathodic stripping in the same medium (to generate Bi0 and soluble Te2- from the Bi2Te3 film component of interest), and the EQCM frequency change was used to determine the content of chemically bound Te in the Bi2Te3 thin film and thereby the compound stoichiometry. Finally, the EQCM frequency change during Bi oxidation to Bi3+ and the difference between total Bi and Bi in Bi2Te3 resulted in the assay of free (excess) Bi in the electrodeposited film. Problems associated with the chemical/electrochemical stability of the free Bi species were circumvented by a flow electroanalysis approach. Data are also presented on the sensitivity of electrodeposited Bi2Te3 film composition to the electrodeposition potential. This newly developed method can be used for the compositional analysis of other thermoelectric thin-film material candidates in general. * To whom correspondence should be addressed. E-mail: [email protected]; [email protected]. † Yonsei University. ‡ Konkuk University. § The University of Texas at Arlington.

6724

Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

Bismuth telluride (Bi2Te3) is a benchmark thin-film material for applications in thermoelectric power generation and cooling.1–14 This material has been synthesized by a variety of methods including evaporation, sputtering, metal organic chemical vapor deposition, and electrodeposition.1–14 Among these synthesis candidates, electrodeposition has several attractive features relative to other approaches, including versatility, simplicity, low cost, mild preparative conditions in terms of temperature, etc., and the ability to afford interesting morphologies for the end product, e.g., nanowires,2–10 nanotubes,11 nanocrystals,12–14 etc. However, contamination of target materials with impurity phases and variable composition (that is often exceedingly sensitive to the electrodeposition history and conditions) is often a problem with the electrosynthesis approach.15–17 It is well-known that the properties and morphology of Bi2Te3 thin films change with composition. For example, Seebeck coefficient, diffraction patterns, and electrical conductivity are (1) Wen, S.; Corderman, R. R.; Seker, F.; Zhang, A.-P.; Denault, L.; Blohm, M. L. J. Electrochem. Soc. 2006, 153, C595. (2) Sander, M. S.; Gronsky, R.; Sands, T.; Stacy, A. M. Chem. Mater. 2003, 15, 335. (3) Sander, M. S.; Prieto, A. L.; Gronsky, R.; Sands, T.; Stacy, A. M. Adv. Mater. 2002, 14, 665. (4) Sapp, S. A.; Lakshmi, B. B.; Martin, C. R. Adv. Mater. 1999, 11, 402. (5) Trahey, L.; Becker, C. R.; Stacy, A. M. Nano Lett. 2007, 7, 2535. (6) Menke, E. J.; Brown, M. A.; Li, Q.; Hemminger, J. C.; Penner, R. M. Langmuir 2006, 22, 10564. (7) Prieto, A. L.; Sander, M. S.; Martin-Gonzalez, M. S.; Gronsky, R.; Sands, T.; Stacy, A. M. J. Am. Chem. Soc. 2001, 123, 7160. (8) Wang, W.; Huang, Q.; Jia, F.; Zhu, J. J. Appl. Phys. 2004, 96, 615. (9) Jin, C.; Xiang, X.; Jia, C.; Liu, W.; Cai, W.; Yao, L.; Li, X. J. Phys. Chem. B 2004, 108, 1844. (10) Wang, W.; Jia, F.; Huang, Q.; Zhang, J. Microelectron. Eng. 2005, 77, 223. (11) Xiao, F.; Yoo, B.; Lee, K. H.; Myung, N. V. J. Am. Chem. Soc. 2007, 129, 10068. (12) Deng, Y.; Nan, C.-W.; Wei, G-D.; Guo, L.; Lin, Y-h. Chem. Phys. Lett. 2003, 374, 410. (13) Purkayastha, A.; Kim, S.; Gandhi, D.; Ganesan, P. G.; Tasciuc, T. B.; Ramanath, G. Adv. Mater. 2006, 18, 2958. (14) Zheng, Y. Y.; Zhu, T. J.; Zhao, X. B.; Tu, J. P.; Cao, G. S. Mater. Lett. 2005, 59, 2886. (15) Myung, N.; Wei, C.; Rajeshwar, K. J. Electroanal. Chem. 1993, 356, 281. (16) Myung, N.; Wei, C.; Rajeshwar, K. Anal. Chem. 1992, 64, 2701. (17) Myung, N.; de Tacconi, N. R.; Rajeshwar, K. Electrochem. Commun. 1999, 1, 42. 10.1021/ac8008127 CCC: $40.75  2008 American Chemical Society Published on Web 08/02/2008

dependent on the electrodeposited film composition.18–20 In addition, n- or p-type thermoelectric materials can be produced depending on the experimental conditions.1 Therefore, development of methods for compositional analysis of electrodeposited Bi2Te3 thin films is both fundamentally and practically important. Up until now, the film composition has been analyzed by atomic spectroscopies, which yield only the total species content.9,18,21 However, our electroanalysis methodology presented below is chemical-state selective; i.e., our approach can differentiate between free Te, free Bi, and Bi2Te3 in the electrodeposited thin film. We have extensively used electrochemical quartz crystal microgravimetry (EQCM) for the compositional analysis of electrosynthesized semiconductor and metal alloy thin films.17,22–24 While side reactions such as hydrogen evolution can interfere with amperometric or coulometric analysis procedures, EQCM is free from such interferences since this technique utilizes only mass change (or frequency change). Reviews of this technique are available.25,26 In this report, we describe the compositional analysis of Bi2Te3 thin films electrodeposited at different potentials by combined EQCM and stripping voltammetry. Both a static and a flow electroanalysis methodology are demonstrated below. To our knowledge, the present study constitutes the first example of this hybrid approach (and indeed, one of the first examples of the use of electroanalytical methods in general) for the compositional analysis of electrodeposited Bi2Te3 thin films. We note a recent report27 of in situ analysis of Bi2Te3 electrodeposition using combined spectroscopic ellipsometry and EQCM; however, differentiation of the Bi2Te3 film content from any excess Bi and Te deposited was not addressed in this previous study. As mentioned earlier, it is worth emphasizing that our electroanalysis methodology presented below can differentiate between free Te, free Bi, and Bi2Te3 in the electrodeposited thin film. EXPERIMENTAL SECTION Materials. All chemicals were from commercial sources: bismuth nitrate pentahydrate (purity 99.99 +%), tellurium dioxide (purity 99 +%), sodium sulfate (purity 99 +%), and nitric acid (purity 70%) were from Aldrich. All chemicals were used as received. Methods. Details of the electrochemical instrumentation and the EQCM setup are given elsewhere.28 For voltammetry and film deposition, an EG&G Princeton Applied Research 263A instrument equipped with model M250/270 electrochemistry software was (18) Heo, P.; Hagiwara, K.; Ichino, R.; Okido, M. J. Electrochem. Soc. 2002, 153, C213. (19) Wang, W. L.; Wan, C. C.; Wang, Y. Y. Electrochim. Acta 2007, 52, 6502. (20) Yoo, B. Y.; Huang, C.-K.; Lim, J. R.; Herman, J.; Ryan, M. A.; Fleurial, J.-P.; Myung, N. V. Electrochim. Acta 2005, 50, 4371. (21) Wang, W. L.; Wang, Y. Y.; Wan, C. C. J. Electrochem. Soc. 2006, 153, C400. (22) Wei, C.; Myung, N.; Rajeshwar, K. J. Electroanal. Chem. 1993, 347, 223. (23) Zhou, M.; Myung, N.; Chen, X.; Rajeshwar, K. J. Electroanal. Chem. 1995, 398, 5. (24) Zhou, M.; Myung, N.; Rajeshwar, K. Electroanalysis 1996, 8, 1140. (25) Buttry, D. A. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker and Academic: New York, 1991; Vol. 17, pp 1-85. (26) Hepel, M. In Interfacial Electrochemistry. Theory, Experiment, and Applications; Wieckowski, A., Ed.; Marcel Dekker and Academic: New York, 1999; pp 599-630. (27) Zimmer, A.; Stein, N.; Johann, L.; Beck, R.; Boulanger, C. Electrochim. Acta 2007, 52, 4760. (28) Ham, S.; Choi, B.; Myung, N.; de Tacconi, N. R.; Rajeshwar, K. J. Electroanal. Chem. 2007, 601, 77.

Figure 1. Schematic diagram of the combined FI-EQCM setup: (a) 0.1 M Na2SO4, (b) 1 M HNO3 containing 1 mM Bi(NO3)3 and 1 mM TeO2, (c) 1 M HNO3, (d) four-way valve, (e) reference electrode, (f) counterelectrode, (g) EQCM electrode, (h) strip chart recorder, and (i) waste.

used. For EQCM, a Seiko EG&G model QCA 917 instrument consisting of an oscillator module (QCA 917-11) and a 9-MHz, AT-cut, gold-coated quartz crystal (geometric area, 0.2 cm2) working electrode, a Pt counter electrode, and a Ag/AgCl/3 M NaCl reference electrode was used. All potentials below are quoted with respect to this reference electrode. Prior to the electrodeposition experiments, the cleanliness of the gold electrode was checked by cyclic voltammetry in 0.1 M H2SO4. The potential was cycled between -0.8 and 0.7 V until the voltammetric and frequency signals were stable.28,29 Flow Injection (FI)-EQCM Electroanalyses. Figure 1 contains a schematic diagram of the combined FI-EQCM setup. An EG&G Princeton Applied Research model 273A system equipped with model M 250/270 electrochemistry software was used. The EQCM setup in this case (Seiko EG&G model QCA 907) consisted of an oscillator module (QCA 927) and a 9-MHz, AT-cut, Au quartz crystal (QA AM9-AU). Frequency changes were displayed on a Kipp and Zonen model BD111 single pen chart recorder or a personal computer. Both sides of the quartz crystal were sputtered with a 50-nm-thick Ti film and subsequently a 300-nm-thick Au film (geometric area, 0.2 cm2). Solution to the FI-EQCM system was delivered by a Gilson Minipuls 3 peristaltic pump. A Valco model six-way slider valve equipped with 0.15-mm-i.d. tubing was employed to switch and inject solutions at a nominal flow rate of 0.23 mL/min. The sensitivity or conversion factor (k) was obtained by potentiostatic deposition of silver from a silver nitrate solution and analyzing the coulometry and EQCM data using a combination of Faraday’s law and the Sauerbrey equation (see eq 1 below).30 A value for k of 0.98 ng/Hz was thus obtained under the conditions pertinent to our EQCM setup. Other Instrumentation. Film morphology and composition were obtained on a JEOL model 6700F field emission scanning electron microscope equipped with an energy-dispersive X-ray analysis probe. RESULTS AND DISCUSSION Figure 2A shows cyclic voltammograms (solid lines) accompanied with frequency changes (dashed lines) for the Au (29) Ham, S.; Choi, B.; Paeng, K.; Myung, N.; Rajeshwar, K. Electrochem. Commun. 2007, 9, 1293. (30) Myung, N.; Jun, J. H.; Ku, H. B.; Chung, H.-K.; Rajeshwar, K. Microchem. J. 1999, 62, 15.

Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

6725

Figure 2. Cyclic voltammograms (solid line) and the corresponding EQCM frequency changes (dash line) for the Au electrode in (A) 1 mM Bi(NO3)3 in 1 M HNO3, (B) 1 mM TeO2 in 1 M HNO3, and (C) 1 mM Bi(NO3)3, 1 mM TeO2 in 1 M HNO3. Scan rate 10 mV/s.

electrode in 1 M HNO3 electrolyte containing 1 mM Bi(NO3)3. The main cathodic and anodic peaks at -0.06 (C2) and +0.09 V (A2) represent bismuth reduction and oxidation, respectively, and the two peaks are accompanied by frequency changes in the corresponding EQCM traces. From the frequency changes, it can be seen that deposited bismuth is completely stripped during the oxidation scan. Two small peaks at +0.05 (C1) and +0.21 V (A1) result from the deposition and stripping of Bi monolayers.31 Frequency changes during the deposition (+0.05 V) and stripping (+0.21 V) are exactly the same which underline the fact that the two peaks are related to each other. A prior study on Bi deposition and stripping using EQCM32 is noted in this regard. Tellurium electrochemistry was investigated using an Au electrode in 1 M HNO3 electrolyte containing 1 mM TeO2 (Figure (31) Jeffrey, C. A.; Harrington, D. A.; Morin, S. Surf. Sci. 2002, 512, L367. (32) Hepel, M.; Bruckenstein, S.; Kanlge, K. J. Chem. Soc., Faraday Trans. 1993, 89, 251.

6726

Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

2B). A peak at ∼-0.1 V (C1) accompanied by EQCM frequency decrease represents the deposition of tellurium on the Au electrode, and the peak at +0.55 V (A1) accompanied with frequency increase represents the stripping of electrodeposited tellurium on the Au surface. New features are obtained when 1 M HNO3 electrolyte contains both bismuth and tellurium. As shown in Figure 2C, a cathodic peak at -0.08 V (C1) results from the deposition of Bi2Te3. Deposition rates vary with potential as seen from the frequency changes in the corresponding EQCM trace. Initially, the frequency changed rapidly with the sharp cathodic peak and decreased slowly after -0.1 V. Also, the magnitude of frequency change is ∼4 times compared with the bismuth (Figure 2A) or tellurium (Figure 2B) only cases. It should be noted that the peaks due to monolayer deposition/stripping disappeared in this case. This may be due to the fact that the first step in the electrodeposition of Bi2Te3 involves the reduction of tellurium adsorbed on the electrode, which triggers the reduction of bismuth as discussed by previous authors.33 The stripping behavior of electrodeposited Bi2Te3 thin films was investigated next with an aim to develop a method for the compositional analysis of thin films (Figure 3). First, Bi2Te3 thin films electrodeposited (at -1.0 V using 1 M HNO3 electrolyte containing 1 mM Bi(NO3)3 and 1 mM TeO2 for 60 s) were subjected to an anodic scan in 0.1 M HNO3 blank electrolyte (Figure 3A). The first peak at 0.125 V (A1) is due to the oxidation of free bismuth: Bi f Bi3+ + 3e, which is clearly supported by the data in Figures 2A and 3C. A major peak in Figure 3C is from the dissolution of bulk bismuth. Here, we found that a peak due to Bi oxidation was shifted somewhat when the films contained other than neat Bi. Another peak at 0.45 V (Figure 3A, A2) results from the oxidation of Bi2Te3 to Bi3+ and Te4+ (reverse reaction of (2)), which will be discussed in detail below. In order to further assign the peaks in Figure 3A, the number of electrons transferred (n) was calculated during the reactions. For this, charge and frequency were sampled at different potentials and the n value was determined by the approach combining the Sauerbrey equation30 and Faraday’s law (as further elaborated in ref 22): Q ) -(nFk ⁄ M)∆f

(1)

In eq 1, n is the electron stoichiometry, F is the Faraday constant, ∆m is the mass change, and M is the molar mass of the deposit. The n value thus obtained from the slope of a coulometry-EQCM plot (cf. Figure 4A) for a scan (10 mV/s) from 0.43 (A2) to 0.48 V (A3) was found to be 18.2, which clearly shows that the peak is due to the oxidation of Bi2Te3, i.e., the reverse reaction of (2): 3HTeO2+ + 2Bi3+ +18e + 9H+ f Bi2Te3(s) + 6H2O

(2)

The stripping voltammogram in Figure 3B supports that the peak at 0.50 V (A3) in Figure 3A originated from the oxidation of free tellurium to Te4+. The data in Figure 3B were obtained during the anodic (positive potential-going) scan of the Te thin films in (33) Martin-Gonzalez, M. S.; Prieto, A. L.; Gronsky, R.; Sands, T.; Stacy, A. M. J. Electrochem. Soc. 2002, 149, C546.

Figure 4. (A) Charge-mass plot obtained from coulometry-EQCM data for a scan from 0.43 to 0.48 V. Scan rate 10 mV/s. (B) Linear sweep voltammograms (solid line) and the corresponding frequency changes (dashed line) obtained during the stripping of free Te. The free Te was obtained after complete stripping of Bi2Te3, which was electrodeposited as in Figure 2. See text for details.

Figure 3. Linear sweep voltammograms (solid line) and the corresponding EQCM frequency changes (dash line) for (A) the stripping of Bi2Te3 in 1 M HNO3 solution, (B) the stripping of Te in 1 M HNO3 solution, and (C) the stripping of Bi in 1 M HNO3 solution. Bi2Te3 was electrodeposited using a electrolyte in Figure 1C at -1.0 V for 60 s. Te was electrodeposited at -0.3 V for 120 s using 1 mM TeO2 in 1 M HNO3. Bi was electrodeposited at -0.1 V for 120 s using 1 mM Bi(NO3)3 in 1 M HNO3.

1 M HNO3 blank electrolyte. The tellurium film was preelectrodeposited on the Au electrode at -0.3 V for 120 s using a 1 M HNO3 containing 1 mM TeO2 for this purpose. Another piece of evidence showing that the peak is due to Te oxidation is shown in Figure 4B. For this, a Bi2Te3 film was electrodeposited at -0.1 V for 60 s in 0.1 M HNO3 containing 1 mM Bi(NO3)3 and 1 mM TeO2. Then, the film was potentiostatically oxidized at 0.4 V for 600 s in 1 M HNO3 to remove Bi2Te3. After thoroughly cleaning the cell, the film was subjected to anodic stripping in the blank electrolyte again (Figure 4B). The peak at 0.50 V (A1) is clearly identical to the Te stripping feature in Figure 3B. Therefore, the anodic stripping of Bi2Te3 films containing Bi, Bi2Te3, and Te is believed to proceed as follows: (a) free Bi oxidation followed by

the oxidation of Bi2Te3 to Bi3+ and Te4+ and (b) oxidation of free Te to Te4+species. The anodic scan (cf. Figure 2A) cannot be used for the compositional analysis of Bi2Te3 thin films due to the following facts: (a) stripping of free Te and Bi2Te3 is not fully resolved and (b) free Bi in electrodeposited Bi2Te3 films is not stable enough. It should be noted that free Bi is immediately corroded in a blank electrolyte as soon as the electrode is immersed in the electrolyte. To avoid corrosion problems, we applied a potential of -0.1 V and poured distilled water into the cell followed by concentrated HNO3. Finally, the stripping voltammogram was obtained as above in 1 M HNO3 electrolyte. Due to the above problems, a cathodic stripping approach was deployed. Figure 5 shows the cathodic stripping scheme for the compositional analysis of Bi2Te3 thins films. First, Bi2Te3 thin films were electrodeposited on the Au electrode at -0.3 V for 60 s using a 1 M HNO3 containing 1 mM Bi(NO3)3 and 1 mM TeO2 and the electrode was subjected to cathodic stripping in 0.1 M Na2SO4 blank electrolyte of pH ∼3.9 (Figure 5A). A first peak at ∼-1.0 V (C1) results from the free Te reduction (Te + 2e f Te2-), which is supported by the data in Figure 5B. Figure 5B was obtained during the cathodic stripping of a Te-modified Au electrode. Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

6727

Figure 6. Composition of Bi2Te3 films electrodeposited at different potentials (wt %).

Figure 5. Linear sweep voltammograms (solid line) and the corresponding frequency changes (dashed line) for (B) the stripping of free Te in 0.1 M Na2SO4 solution. (The Te film were pre-electrodeposited by holding the Au electrode at -0.3 V for 120 s in 0.1 M H2SO4.) (A) contains data for the cathodic stripping of a Bi2Te3 thin film in 0.1 M Na2SO4 solution. (C) Stripping of total Bi in 1 M HNO3 solution. The Bi2Te3 thin film was electrodeposited using the electrolyte in Figure 1C at -0.3 V for 60 s.

Another peak in Figure 5A at -1.23 V (C2) results from the reduction of Bi2Te3 to Bi + Te2-. The electroreduction of Bi2Te3 is completed beyond -1.3 V. Finally, the remaining Bi (free Bi + Bi in Bi2Te3) is anodically stripped in 1.0 M HNO3 (Figure 5C); this scan furnishes the total Bi content of the film. The peak positions for Bi stripping slightly depended on the amount of Bi while the onset potential of the corresponding EQCM frequency change was relatively insensitive to this variable. The underlying origins for these trends are being investigated and are beyond the scope of this particular study. Compositional analysis of Bi2Te3 thin films according to this newly developed combined EQCM-stripping voltammetry was 6728

Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

performed using the sequence described above. First, the content of free Te and Bi2Te3 in the electrodeposited film was determined by its cathodic stripping in 0.1 M Na2SO4 blank electrolyte and monitoring the EQCM frequency changes accompanying the stripping peaks. The Te content, established thus, in the Bi2Te3 film can be used for the determination of the Bi2Te3 content of the film based on assumption of the 2:3 compound stoichiometry. Then the total Bi content (free Bi and Bi in Bi2Te3) was determined via anodic stripping (cf. Figure 5C). The results are summarized in Figure 6. As shown in Figure 6, the content of free (excess) Te in the film is almost constant at all deposition potentials investigated. However, the free Bi content increases as deposition potential is decreased (as also observed in ref 29), and the Bi2Te3 content shows a reverse trend. This behavior can be clearly understood based on the electrodeposition mechanism. First, Bi2Te3 is electrodeposited exclusively up to -0.6 V via an 18 e- process. However, at more negative potentials, free Bi electrodeposition takes over to an increasing extent and Bi2Te3 content is concomitantly decreased (Figure 6). This is further manifested by the data in Figure 7. As the deposition potential is decreased (i.e., made more negative), the anodic stripping peak (at ∼0.1 V, A1) due to free Bi oxidation increases as shown in the figure. FI-EQCM Electroanalyses of Electrodeposited Bi2Te3. The less-than-optimal stability of the free Bi species in the electrodeposited Bi2Te3 films prompted us to evaluate a flow injectionelectroanalysis approach using the setup schematically shown in Figure 1. The crucial feature with this approach is that the working electrode is always under potential control during switching of solutions. This then avoids the problems encountered above with leachout of any Bi or (even) Te species from the electrode surface. The various species (Bi2Te3, free Bi, free Te) in the electrodeposited film was assayed in the same manner as described for the “static” analysis approach described above: stripping free Te cathodically at -0.9 V in 0.1 M Na2SO4 and anodic stripping of free Bi and the Bi2Te3 compound in 1 M HNO3 at 0.1 and 0.5 V, respectively. Figure 8 contains EQCM frequency (mass) changes during the electrodeposition and stripping stages at two different film

Figure 7. Cyclic voltammograms (solid lines) and the corresponding frequency changes (dashed lines) for an Au electrode in 1 M HNO3 containing 1 mM Bi(NO3)3 and 1 mM TeO2. Note that potential windows were varied in the two frames. Scan rate 10 mV/s.

deposition potentials, (A) -0.1 and (B) -1.0 V, and Figure 9 maps compositional trends as a function of the deposition potential (cf. counterpart of Figure 7 for the static electroanalysis approach). Finally, Table 1 summarizes these compositional trends as a function of deposition potential. All quantatative data were based on the EQCM measurement of frequency changes during the stripping step, and the reprocibility was checked by running at least five replicates. As seen earlier, the free Bi content of the film systematically increases as the deposition potential is made more negative. Careful examination of the quantitative trends in Figures 6 and 9 reveal systematic differences that can be ascribed to Bi and Te loss from the film for reasons mentioned earlier. Qualitatively, however, the trends are similar. CONCLUSIONS A new electroanalytical method for the compositional analysis of electrodeposited Bi2Te3 thin films has been described combining stripping voltammetry and EQCM. This method is species selective to free Bi, free Te, and these elements present in the electrodeposited film as the chemically bound Bi2Te3 compound. The amounts of these three film constituents (especially free Bi and Bi2Te3 compound) varied with the deposition potential. The corrosion of Bi in HNO3 is a nuisance for the compositional analysis of Bi2Te3 thin films, which can be circumvented by always maintaining potential control on the film as in the flow injection electroanalysis approach presented above. The electroanalytical

Figure 8. EQCM frequency (mass) changes during the electrodeposition and stripping steps. Deposition potential: (A) -0.1 V; (B) -1.0 V. Species in films were consecutively analyzed by applying -0.9 (in 0.1 M Na2SO4, free Te determination), 0.1 (in 1 M HNO3, free Bi determination), and 0.5 V (in 1 M HNO3, Bi2Te3 determination).

Figure 9. As in Figure 6 but for the FI-EQCM assay case.

method as developed in this proof-of-concept study can be applied to the compositional assay of other thermoelectric materials and alloy materials in general. The only constraint is that the films Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

6729

Table 1. Compositional Analysis of Bi2Te3 Thin Films Electrodeposited at Different Potentials (N ) Replicates)a potential (V)

N

free Bi (wt %)

free Te (wt %)

Bi2Te3 (wt %)

0.0 -0.2 -0.4 -0.6 -0.8 -1.0

5 5 5 5 5 5

6.01 (±2.12) 6.13 (±2.14) 6.52 (±1.74) 6.93 (±0.51) 22.16 (±0.61) 32.52 (±2.05)

4.24 (±2.36) 4.95 (±2.19) 4.75 (±1.04) 4.49 (±1.40) 4.53 (±0.67) 4.82 (±0.40)

89.74 (±0.37) 88.93 (±1.52) 88.67 (±1.33) 88.86 (±1.37) 73.30 (±2.30) 72.15 (±2.50)

a

Numbers in parentheses, mean standard deviations.

must be electrodeposited on an electronically conductive substrate (supported by the quartz EQCM sensor) and that these films are

6730

Analytical Chemistry, Vol. 80, No. 17, September 1, 2008

thin enough for the applicability of the Sauerbrey viscoelastic model regime.25 ACKNOWLEDGMENT This research was supported by the Program for the Training of Graduate Students in Regional Innovation, which was conducted by the Ministry of Knowledge and Industry of the Korean Government. S.H. is the recipient of the Seoul Science Fellowship in 2007. K.R. thanks the University of Texas System for a STARS grant. We thank the two anonymous reviewers for constructive criticisms of an earlier manuscript version. Received for review April 22, 2008. Accepted July 2, 2008. AC8008127