solidi
status
pss
physica
phys. stat. sol. (a) 205, No. 10, 2371 – 2375 (2008) / DOI 10.1002/pssa.200779444
a
www.pss-a.com
applications and materials science
Electrochemical growth of ZnO nanowires inside nanoporous alumina templates. A comparison with metallic Zn nanowires growth Daniel Ramirez*, 1, 2, Thierry Pauporte2, Humberto Gomez1, and Daniel Lincot2 1 2
Instituto de Quimica, Pontificia Universidad Catolica de Valparaíso, Avenida Brasil 2950, Codigo postal 4059 Valparaiso, Chile Laboratoire d’Electrochimie et Chimie Analytique (UMR CNRS 7575), Ecole Nationale Supérieure de Chimie de Paris (ENSCP), 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France
Received 2 January 2008, revised 25 July 2008, accepted 25 July 2008 Published online 11 September 2008 PACS 62.23.Hj, 81.05.Bx, 81.05.Dz, 81.15.Pq, 81.16.Be *
Corresponding author: e-mail
[email protected], Phone: + 56-32-2273173, Fax: + 56-32-2273422
Molecular oxygen reduction in presence of Zn(II) allowed to get arrays of zinc oxide (ZnO) nanowires (NWs) inside thick anodic alumina membranes (AAM), previously made by anodization in oxalic acid. A Zinc (Zn) NWs array made by electrochemical deposition (ED) was employed as a metal reference in order to compare its growth behavior with respect to that of ZnO, a semiconducting material. Chronoamperometric measurements for Zn NWs showed a typical response for metallic growth inside porous templates while ZnO NWs did not show this behavior. We observed that Zn NWs grew faster than ZnO NWs as checked by the electrical charge. SEM images showed the presence of ZnO NWs only after long deposition time. X-ray diffraction data confirmed the ZnO structure.
A SEM image of a ZnO NWs array after partial dissolution of the AAM.
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction Template-assisted nanostructuring methods can be considered as convenient route to control the shape of any material. For NWs synthesis, ion track etched [1] and AAM [2] serve as good templates, but AAM result more advantageous due high ordering degree of pores, uniform pore diameter and well defined thickness. These characteristics make them ideal to be combined with ED due to the control of the amount of electricity passed that the latter offers and, the AAM porous similarity with a recessed nanoelectrodes array [3]. This combination has more advantages because it is cheap alternative which strictly follows the bottom-up approach compared with other methods such as thermal evaporation and lithography (top-down based) [4, 5]. In addition, ED enables to prepare metals, conductive polymers, composites and oxide films [6–9]. ZnO, a semiconducting material, is a very interest-
ing material due to its versatile applications in piezoelectric, polarized light emitting devices, gas and chemical sensors, catalysis and photovoltaics [10–13]. Many researchers have achieved the formation of ZnO NWs inside both, AAM or track etched polycarbonate membranes by electrochemical deposition, with good photoluminescence properties in some cases [14–19]. The ED synthesis method can be itself a power tool for providing basic information concerning the NWs growth and properties. For instance, the growth of a metal shows a well defined current–time curve with three current stages: a charging nucleation zone at short times, a steady state zone at intermediate times and, a progressive increase at longer times where NWs are completely filling the template. The latter stage gives evidence about when to stop the deposition. However, for ZnO such a similar behaviour can be modi© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
D. Ramirez et al.: Electrochemical growth of ZnO nanowires inside nanoporous alumina templates
2 Experimental AAM were prepared by a two step anodization of high purity (99.99%) Al foil, carried out at 40 V in oxalic acid as reported elsewhere [21]. Once anodizing is finished, a chemical etching in 5% H3PO4 at 30 °C was applied at different times in order to open the barrier oxide layer at the bottom side of the AAM leaving open the two sides. Once pores were open at the bottom, a gold layer was evaporated by RF sputtering in the top side of the AAM to form a conductive contact. The Electrochemical setup was made with a mercury/sulphate reference electrode (MSE, 0.65 V vs. SHE) and a platinum mesh counter electrode. The working electrode was prepared with a contact between the gold backside of the AAM and a copper wire through a silver conductive paste and, then the AAM was mounted on a glass slide and insulated with epoxy resin leaving a exposed area accessible to the electrolyte. The copper wire, the reference and counter electrode were connected to a potentiostat Autolab and controlled by software (PGSTAT). In order to wet the AAM exposed area from the mouth to the bottom where gold film is located, the samples were ultrasonically treated in the electrolyte before each experiment. The electrodeposition of ZnO NWs was made in 0.1 M Zn(ClO4)2, 0.1 M LiClO4 solution with molecular oxygen bubbled to assure saturation at 70 °C. The deposition potential was –1.0 V. The same solution was employed to form Zn NWs, but with bubbling Argon instead oxygen at a potential of –1.7 V. To obtain free-standing ZnO and Zn NWs, a two solution kits for a selective etching of AAM was employed. These kits were a 5% NaOH solution and a pH = 9 borate buffer. One minute of etching in each solution and deionizated water in between each dip, repeated 5 times, was enough to partially expose the ZnO and Zn NWs. Field emission scanning electronic microscopy images (FE-SEM) were recorded in a Zeiss system model Ultra 55, while X-Ray diffraction analysis were made in a Siemens D5000 difractometer with Co Kα radiation (λ = 1.78897 Å). 3 Results and discussion 3.1 Template preparation Figure 1 shows the AAM thickness and pore’s diameter as function of time. From these data both, the AAM growth rate and the opening rate of the bottom barrier layer to leave exposed the pores can be estimated. The inset in each figure shows an AMM, indicated by the arrows, corresponding to a thickness of 20 µm with a pore diameter of 50 nm, respectively. In order to estimate the AAM effective area, Seff, the surface density of pores, N, needs to be calculated. There are two ways to obtain this parameter. The first one considers the pore array geometry as formed by an ideal hex© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
80
100
60 40 20 0 0
a 100 200 300 400 500 Anodization time / min.
Diameter of the pore / nm
solidi
fied by the semiconducting properties, even if the NWs completely filled the template. As has been reported for hydrogen peroxide, another expected difference can arise from the poor catalytic surface of bulk ZnO films to reduce oxygen precursors [20].
Membrane thickness / mm
2372
status
physica
pss
a
80 60 40 20 0 0
b 15 30 45 60 75 90 H3PO4 etching time / min.
Figure 1 (a) Determination of AAM growth rate in oxalic acid 0.3 M. The slope obtained by the linear fit was 0.156 µm min–1. The inset shows an AAM corresponding to the thickness determined by the arrow. (b) Variation of pore diameter with the time of chemical etching in phosphoric acid 5% at 30 °C. Slope: 1.174 nm min–1. The inset shows an AAM corresponding to the diameter of the pore determined by the arrow.
agonal array of pores cells (Fig. 2a). For an ensemble of microelectrodes in an hexagonal array, N is related to the intermicroelectrode distance, d, by [22]: N = 2/ 3 d 2 .
(1)
The second way is by simple statistical counts. In our case, by the geometrical method we found a value of N equal to 1.25 × 1010 cm–2, whereas by simple statistical counts the value was 1.10 × 1010 cm–2. We consider the last value more realistic, because includes deviations caused by defaults in the AAM (Fig. 2b). 3.2 Electrochemical deposition of ZnO and Zn Figure 3 shows the voltammetric curve recorded for an AAM in Zn(ClO4)2, LiClO4 and molecular oxygen. The letters A, B and C in Fig. 3a correspond to the following sequence of reactions: A. (O 2 ) b Æ (O 2 ) m ,
(1)
(O 2 ) m Æ (O 2 )e ,
(2)
(O 2 )e + 2H 2 O + 4e - Æ 4(OH - )e ,
(3)
(Zn 2+ ) b Æ (Zn 2+ ) m ,
(4)
(Zn 2+ ) m Æ (Zn 2+ )e ,
(5)
(Zn 2+ )e + (OH - )e Æ ZnO(s) + H 2 O ,
(6)
B. (Zn 2+ )e + 2e - Æ Zn (s) ,
(7)
C. Zn (s) Æ (Zn 2+ )e + 2e - .
(8)
According to this sequence, ZnO formation inside of AAM involves firstly diffusion of molecular oxygen and Zn(II) species from bulk (subscript b) towards the pore mouth (subscript m), then inside towards the bottom of www.pss-a.com
Original Paper phys. stat. sol. (a) 205, No. 10 (2008)
pores where oxygen reduction takes place on gold surface (subscript e). After hydroxide ions are generated, they inmediately precipitate in presence of Zn(II) to form Zn(OH)2 which is easily dehydrated at 70 °C to form ZnO NWs. The drawing in the inset of Fig. 3 is useful to understand this phenomenology, which describes the concentration profile inside a single pore developed after the application of the deposition potential to the gold film. Here, ce, cm and cb correspond to the concentrations in the bottom, mouth of the pore and the bulk of the electroactive species, respectively. However, according to Bartlett’s group findings, we can estimate no remarkable differences between cm and cb because the pores are deeply recessed, and hence no diffusion layer is expected to develop outside the AAM pores [23]. Figure 4 shows the current density–time (J–t) curves recorded for Zn (Fig. 4a) and ZnO NWs (Fig. 4b). The Zn NWs J–t characteristic presents the classical shape corresponding to a metallic growth behaviour. Depending on the elapsed time three zones are observed: (i) at short times, the high current density corresponds to the charging zone where Zn nucleation occurs; (ii) at 200 s the system
Figure 3 Voltammetric curve of AAM in 0.1M Zn(ClO4)2, 0.1 M LiClO4 and dissolved molecular oxygen at 70 °C. Scan rate: 20 mV s–1. The inset schema shows a single pore and the concentration profile developed inside and around it. L is the membrane thickness, a is the pore radius and d is the diffusion layer length outside the pore. www.pss-a.com
40
40 20 0
0
500
1000
1500 2000
time / s
20 0 0
0.20 -2
-2
60
60
200 400 600 800 1000
time / s
J / mA cm
-2
-2
Figure 2 SEM images of a AAM showing: (a) the geometrical model to determinate N and, (b) low magnification, showing some defects indicated by arrows. The letter a indicates the pore radius and d the interpore distance.
J / mA cm
a
2.0
80
1.5
0.16
J / mA cm
80
b
J / mA cm
d
a
2373
0.12 0.08
1.0
0.04 0.00 0
5k
10k 15k 20k 25k
time / s
0.5 0.0 0
5k
10k 15k 20k 25k
time / s
Figure 4 Experimental chronoamperograms for Zn (left side) and ZnO (right side) NWs showing (in little sketches) the different stages of growth inside the AAM.
achieves a steady state; and (iii) at 300 s the steady state starts to evolve to a continuous increase of current density, indicating that the Zn NWs are arriving to the pores mouth. At this point, each individual diffusion field determined by each pore, start the overlapping process. The theoretical electrical charge value, QTEO, to completely fill the AAM was estimated by the following equation [24]: QTEO = zF ρ LSeff /M.M .
(9)
Where z is the number of exchanged electrons (2), F is the Faraday constant (96487 C mol–1), ρ is the density of the deposited material (g cm–3) and L is the AAM thickness (cm). The Seff value is given by: Seff = πa 2 NA .
(10)
Where A is the AAM exposed area (cm2), N is the surface density of pores (cm–2) and a is the pore radius (cm). This parameter is very important because allows to show current densities instead of currents in voltammetric and chronoamperometric curves. The QEXP /QTEO (%) ratio for Zn NWs, where QEXP is the experimentally exchanged quantity of electricity directly calculated from the chronoamperograms, was 55%. This means that the AMM filling is not complete. In spite of applying a negative potential step and, taking into account the results of other authors during the electrodeposition of different metals, this percentage is low [24]. Due to its thickness and small pore size, all the pores of the AAM were not wet by the electrolyte. Another explanation could result from pore to pore variations in the cathode interface, which may have an influence on the nucleation rate. In fact, a 50 nm diameter pore generates an area equal to 2500 nm2. The latter could be enough to completely suppress the nucleation of a wire either in presence of very small amounts of adsorbed impurities or by heterogeneities caused by grain boundaries. This point of view has just been proposed by Prieto et al, who tried to explain similar difficulties in the growth of Bi2Te3 nanowires, reaching a 10–20% of efficiency in the filling of the pores [25]. In the case of ZnO NWs (Fig. 4b) the J–t curve has a different characteristic compared to that obtained for Zn NWs. First, the current density value was much lower than © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.3 Characterization After the selective etching of the AAM, good quality SEM images of ZnO NWs were obtained. Experimental observations (not shown) have
a
b
c
d
Figure 5 (online colour at: www.pss-a.com) SEM images of ZnO NWs after partial chemical etching treatment at low (a) and high (b) magnification. (c) Four Zn nanowires and (d) cross sectional view of Zn NWs. © 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
36
37
(101)
(002)
solidi
that of Zn NWs which is simply due to the low concentration of molecular oxygen as compared to that of Zn(II) (the solubility of molecular oxygen is 8 × 10–4 M in the same media and temperature). As shown in previous work [20], perchlorate ions do not participate to the formation of zinc oxide, due to the great stability of these species making the kinetics of the corresponding reaction very slow, contrary to what is observed when nitrate ions are used. Second, instead of reaching a steady state, the current density tends to decrease as time increases, even for long time values (see the J–t curve at high magnification in the inset). A model to explain this behaviour can be based in a continuous growth of the resistance in each ZnO NW. On ZnO covered Pt electrodes, in presence of Zn(II) ions, Goux et al., reported a more irreversible behaviour for oxygen reductions. They also found lower values in kinetic current for oxygen reduction when compared with bare Pt [26]. They associated this behaviour with a low exchange current density value, J0, and hence, to a greater charge transfer resistance, Rtc. In order to explain their results, they did not discard a possible surface blockage due to reaction of hydroxide with Zn ions. This would only explain the resistance itself, whereas its continuous increase (and the corresponding J decrease) can be attributed to the thickness effect. An increase in the ohmic drop can influence the oxygen reduction due to a surface potential increase. This was adequately described in the case of the deposition of thick CdTe films [27]. In addition, after 7 h of deposition, the overall QEXP /QTEO (%) ratio was only 5%. This value is much lower than in the case of metallic zinc deposition, which means that the blocking of the growth in some sections of the AAM is even more severe.
(100)
D. Ramirez et al.: Electrochemical growth of ZnO nanowires inside nanoporous alumina templates
Intensity / a. u.
2374
status
physica
pss
a
38
39
40
41
42
43
2q / ° Figure 6 X-Ray diffraction of as deposited ZnO NWs corresponding to Fig. 5.
proved that at pH 9.2 the rate of dissolution of amorphous alumina was not fast enough to reveal ZnO NWs. At more alkaline pH values the amorphous alumina dissolution is both thermodynamically and kinetically favoured without a significant lost of ZnO NWs by chemical dissolution. ZnO NWs isolated by this procedure are presented in Fig. 5a and b. In fact, these observations were only made in specific AAM zones. No growth of ZnO NWs was evidenced in other places. This confirms our previous observations about the J–t curves and the low value of the QEXP/QTEO (%) ratio. Many of the pores do not present any growth. However, on a small fraction of them proceed well towards the outer surface. From SEM analysis a mean diameter of 66 nm was measured, which is consistent with the original pore diameter in the AAM. The aspect ratio can be estimated as high as 400. On other hand Zn NWs SEM images are displayed in the Fig. 5c and d. In this case we found that, as for ZnO NWs, there is non uniform distribution of these NWs over the AAM. In addition, the Fig. 5c shows that the Zn NWs have a non homogeneous length, which may indicate some differences in the nucleation induction time from pore to pore due to interfacial structure differences between them, as said previously, and it will be kept during the growth due the different NW length will cause different potential drop. Figure 6 shows the X-ray diffraction pattern for an ensemble of ZnO NWs arrays inside the AAM, after 7 hours of electrodeposition. The main peaks for ZnO are present, showing its polycrystalline character but without a preferential orientation. The estimated sizes of the crystallites were 22 nm, 50 nm and 29 nm for the (100), (002) and (101) planes, respectively. 4 Conclusion Zn and ZnO nanowires were electrochemically synthesized from perchlorate solutions in AAM. A preparation method for AAM was set up. The difference found between J–t characteristics were related to the metallic/semiconductor nature of the materials as well as the different electrochemical reactions involved. Currently, the main difficulty found is the strong lateral heterogeneity of www.pss-a.com
Original Paper phys. stat. sol. (a) 205, No. 10 (2008)
the deposition. This could be the consequence of sufficient wetting or activation of the gold contact at the bottom of the pores. Acknowledgements D.R. acknowledges Stephan Borensztajn for SEM images, to ECOS-CONICYT convention for financial support number C05E07, to CONICYT for the project No. AT-24060121 and to CENAVA for the project No. PBCTACT027.
References [1] P. B. Price and R. M. Walker, J. Appl. Phys. 33, 3407 (1962). [2] F. Keller, M. Hunter, and D. Robinson, J. Electrochem. Soc. 100, 411 (1953). [3] M. Granström, J. Carlberg, and O. Inganäs, Polymer 36, 3191 (1995). [4] P. Gao and Z. Wang, J. Phys. Chem. B 106, 12653 (2002). [5] A. Morales and C. Lieber, Science 279, 208 (1998). [6] G. Possin, Rev. Sci. Instrum. 41, 772 (1970). [7] M. Granström, J. Carlberg, and O. Inganäs, Polymer 36, 2867 (1995). [8] C. Huber, T. Huber, M. Sadoqi, J. Lubin, S. Manalis, and C. Prater, Science 263, 800 (1994). [9] J. Oh, Y. Tak, and J. Lee, Electrochem. Solid-State, Lett. 7, C27 (2004). [10] R. Könenkamp, R. C. Word, and C. Schlegel, Appl. Phys. Lett. 85, 6004 (2004). [11] H. Nanto, T. Minami, and S. Takata, J. Appl. Phys. 60, 482 (1986).
www.pss-a.com
2375
[12] M. H. Sarvari and H. Sharghi, Tetrahedron 61, 10903 (2005). [13] T. Yoshida and H. Minoura, Adv. Mater. 12, 1219 (2000). [14] Q. Wang, G. Wanga, Bo Xua, J. Jiea, X. Hana, G. Lia, Q. Lic, and J. G. Houa, Mater. Lett. 59, 1378 (2005). [15] M. J. Zheng, L. D. Zhang, G. H. Li, and W. Z. Shen, Chem. Phys. Lett. 363, 123 (2002). [16] Min Lai and D. Jason Riley, Chem. Mater. 18, 2233 (2006). [17] B. Y. Leprince-Wang, A. Yacoubi-Ouslim, and G. Y. Wang, Microelectron. J. 36, 625 (2005). [18] Sh. U. Yuldashev, S. Woo Choi, T. Won Kang, and L. A. Nosova, J. Korean Phys. Soc. 42, S216 (2003). [19] Y. Leprince-Wang, G. Y. Wang, X. Z. Zhang, and D. P. Yu, J. Cryst. Growth 287, 89 (2006). [20] T. Pauporte and D. Lincot, J. Electroanal. Chem. 517, 54 (2001). [21] S. Green, J. A. Badan, M. Gilles, A. Cortes, G. Riveros, D. Ramirez, H. Gomez, E. Quagliata, E. A. Dalchiele, and R. Marotti, phys. stat. sol. (c) 4, 618 (2007). [22] B. Scharifker, J. Electroanal. Chem. 240, 61 (1988). [23] P. N. Bartlett, J. Electroanal. Chem. 453, 49 (1998). [24] G. Riveros, H. Gómez, A. Cortés, R. E. Marotti, and E. A. Dalchiele, Appl. Phys. A 81, 17 (2005). [25] A. Prieto, M. Sander, M. Martin-Gonzalez, R. Gronsky, T. Sands, and A. Stacy, J. Am. Chem. Soc. 123, 7160 (2001). [26] A. Goux, T. Pauporte, and D. Lincot, Electrochim. Acta 51, 3168 (2006). [27] J. Rousset and D. Lincot, J. Electrochem. Soc. 154, D310 (2007).
© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
