Photoluminescence properties of anodic alumina membranes with ordered nanopore arrays

Photoluminescence properties of anodic alumina membranes with ordered nanopore arrays

ARTICLE IN PRESS Journal of Luminescence 121 (2006) 588–594 Photoluminescence properties of anodic alumina membranes ...

213KB Sizes 0 Downloads 31 Views


Journal of Luminescence 121 (2006) 588–594

Photoluminescence properties of anodic alumina membranes with ordered nanopore arrays Xiuyu Sun, Faqiang Xu, Zongmu Li, Wenhua Zhang National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei, 230029, National Center for Nanoscience and Nanotechnology, PR China Received 14 July 2005; received in revised form 14 December 2005; accepted 21 December 2005

Abstract Photoluminescence (PL) of anodic alumina membranes (AAMs) with ordered nanopore arrays fabricated in oxalic acid has been investigated under different annealing temperatures. The PL intensity firstly increases, and at 500 1C reaches a maximum value, then decreases. The structural transition from amorphous to g-Al2O3 in AAMs has been confirmed by X-ray diffraction. Thermogravimetric analysis results and electron paramagnetic resonance measurements revealed that the PL band of alumina membranes could be attributed to the oxygen-related defect centers (F+ centers) rather than the luminescent centers transformed from oxalic impurities. r 2006 Elsevier B.V. All rights reserved. PACS: 78.55 Mb; 78.66 Nk Keywords: Photoluminescence; Alumina membrane; Oxalic impurities; Oxygen vacancies

1. Introduction Porous anodic alumina membrane (AAM) has stimulated considerable interests as a nanostructural template and has been widely used to fabricate various one-dimensional nanomaterials, such as metals [1,2], polymers [3], semiconductors [4] and even heterostructures [5]. Furthermore, Corresponding author. Tel.: +86 551 3602127;

fax: +86 551 5141078. E-mail address: [email protected] (F. Xu).

porous anodic alumina membranes have been used as photonic crystals [6], as humidity sensors [7,8], or as cathodes for organic light emitting diodes [9]. Therefore, to study the photoluminescence properties of the AAM itself for practical applications is important. The optical property of anodic alumina membranes with ordered nanopore arrays prepared by electrochemical anodization attracted considerable attention recently [10–13]. The interests for this topic are mainly focused on the relationship of the photoluminescence (PL) property with the multi

0022-2313/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jlumin.2005.12.057

ARTICLE IN PRESS X. Sun et al. / Journal of Luminescence 121 (2006) 588–594

phase of alumina and the mechanism evolving the light emission. There are mainly two opinions about the origin of the blue photoluminescence band of AAMs formed in oxalic acid solution. The first one is the defect centers related to oxygen vacancies that are used almost universally in the explanation of the PL of solid materials besides AAMs [10,11,14–17]. The second opinion says that oxalic impurities are responsible for the luminescence because the AAM is produced in oxalic acid solution and large amount of oxalic ions are incorporated into the bulk of AAMs [18–20]. However, up to now, the PL nature of AAMs is still ambiguous because of the difficulty of qualitative detection of oxalic ions and quantitative measurement of oxygen vacancy defects. In this paper, we investigated the PL properties of AAMs formed in oxalic acid. A blue photoluminescence band was observed for the asanodized sample and those annealed under different temperatures. Thermogravimetric analysis (TGA) and electron paramagnetic resonance (EPR) measurements suggest that the luminescent band of AAM arises from the oxygen-related defect centers rather than the luminescent centers transformed from oxalic impurities.

2. Experimental The AAMs were prepared electrochemically in oxalic acid solution by a two-step anodization process as described previously [21]. The high purity (99.999%) aluminium foils annealed at 500 1C for 5 h in air were degreased in acetone and electropolished in a mixture of sulfuric (5 vol%), phosphoric (95 vol%) and chromic acids (20 g L1). The treated aluminum sheets were anodized in 0.3 mol L1 oxalic acid solutions under constant voltage (40 V). After anodization for 6 h, the formed alumina was removed by a mixture solution of phosphoric (6.0 wt%) and chromic acid (1.8 wt%). Then the aluminium foil was anodized again for 10 h under the same anodization conditions as the first step. After removing the remaining aluminium in HCl and CuCl2 mixed solution, the transparent AAMs were obtained. A series of samples were prepared by


annealing at different temperatures in high-purity (99.999%) Ar for 6 h, respectively. The crystallographic structures of the samples were determined by X-ray powder diffraction (XRD) using a MXPAHF X-ray diffractometer ˚ with CuKa radiation (l ¼ 1:54056 A). The PL spectral measurements were taken on a Hitachi 850 fluorescence spectrophotometer with a Xe lamp as the excitation light source. The thermogravimetric analysis (TGA) was conducted on Shimadzu TGA-50 H. The weight of sample is ca. 8 mg. The heating rate is 10 1C/min and N2 is used as the carrier gas with the flow rate of 20 ml/min. The electron paramagnetic resonance (JEOL JESFA200) spectra were also measured. The weight of the sample is 1 mg for each measurement. The frequency is 9.069 GHz and the range of magnetic field is 0.179–0.479 T. All the measurements were carried out at room temperature.

3. Results and discussion Fig. 1 shows the PL spectra of the as-prepared AAM and the annealed AAMs at different temperatures, respectively, measured with an excitation of the 348 nm line of a Xe lamp. It can be seen that an intensive and broad PL emission band appears at about 445 nm [10]. The intensity of this band increases with elevated annealing temperature (T a ) and reaches a maximum for the sample annealed at about 500 1C, but drastically decreases with further increase of annealing temperature [15]. It is well known that the physical properties of AAMs are directly associated with their structure, which can be changed by further increasing the annealing temperature. Therefore, the XRD measurement of the samples was carried out. The XRD patterns in Fig. 2 clearly show that the asprepared and low-temperature annealed AAMs are amorphous [12]. As the annealing temperature increases to 800 1C, five obviously diffraction peaks indexed to g-Al2O3 phase appear indicating the crystallization of AAMs [22]. When the Ta increases to 900 1C, the diffraction peaks are similar to that of the samples annealed at 800 1C, but the crystallization of AAMs is increased. At

ARTICLE IN PRESS X. Sun et al. / Journal of Luminescence 121 (2006) 588–594


40 500°C


PL Intensity (a.u.)

30 25 400°C 20 525°C 300°C


200°C 10 600°C


700°C 0


As prepared

-5 400





Wavelength (nm) Fig. 1. PL spectra of the as-anodized AAM and the AAMs annealed at different temperatures for 6 h in Ar.


400 222



900°C 311 Intensity (a.u.)

800°C 700°C 600°C 500°C 300°C as-prepared







2θ (Degree)

Fig. 2. XRD patterns of the as-anodized AAM and the AAMs annealed at different temperatures for 6 h in Ar.

this temperature, the AAMs belong to g-Al2O3 phase. The thermogravimetric analysis can be used to reveal the change of AAM in the process of annealing. Fig. 3 shows the TGA result of the as-

prepared AAM. Three weight loss regions were observed in the TGA curve. The weight loss for the first section from room temperature to 410 1C is ca. 4.65% and this is mainly attributed to the desorption of weakly bound water from the

ARTICLE IN PRESS X. Sun et al. / Journal of Luminescence 121 (2006) 588–594


8.6 section I 8.4 4.65% Weight (mg)

8.2 section II 1.00%


7.8 5.04% 7.6 section III 7.4 100









Temperature (°C) Fig. 3. Thermogravimetric analysis curve for the as-prepared AAM.

surface and the inner walls of nanopores. The water is referred to the physically absorbed and chemisorbed water molecules. In the second process from 410 to 510 1C, the weight loss is ca. 1.00%, which may come from the decomposition of oxalic impurities because the decomposition of metal oxalates is believed to occur at temperature higher than 350 1C [23–26]. The small weight loss and short temperature range for this section indicate that a small quantity of oxalic ions are incorporated within the structure of the porous AAMs. In the range of 510–800 1C, no obvious weight loss can be observed. The third section from 800 to 900 1C is a typical TGA feature indicating a strong phase transition [27]. Taking this in combination with the XRD patterns of the samples, it can be concluded that the AAM experiences a transition from amorphous to g phase. During this transformation, the –OH groups in crystalline structure of g-Al2O3 [28] will lose, which result in the weigh loss. The EPR spectroscopy has been established as a universal, sensitive and quantitative method to detect and characterize radicals in solid materials. Here, we examined the defect structures of alumina films using EPR technique (Fig. 4). For the as-prepared and the heat-treated AAMs at

various temperatures, the obvious EPR signals can be observed, though the signal intensity and shape were different. Generally, there are two obvious EPR signals (signals I and II) in Fig. 4. The g factors of signal I (1.9765) and signal II (2.0253) are different from that of free electrons (2.0023) suggesting that two kinds of paramagnetic defect centers exist in porous alumina films. The g value of signal II is slightly larger than that of the free electron indicating the formation of an oxygencentered radical i.e. Al–O–O, namely, the oxygen hole center is in AAM [29,30]. Ishizaka et al. [29] reported that the EPR signals of the oxygencentered radicals were attributed to the trapped electron at oxygen vacancy-F+ like center when studying thermally treated alumina films. Therefore, it is rational to conclude from the EPR results that the as-prepared and annealed AAMs contain oxygen-related defect centers (F+ centers). It has been demonstrated that the PL intensity of AAM is different when it is annealed in O2 and inert atmosphere [11,31], that is, the annealing atmosphere has an important effect on the PL property of the AAM. Additionally, the electrolyte ions such as oxalic impurities in AAMs [19,20] would be decomposed during the calcinations of AAM as demonstrated by the TGA results in


X. Sun et al. / Journal of Luminescence 121 (2006) 588–594

Fig. 4. EPR spectra measured from the as-anodized AAM and the AAMs annealed at different temperatures for 6 h in Ar. The inset gives the dependence of the spin density corresponding to signal II on the temperature of the annealing of the samples.

Fig. 3. Therefore, the oxalic impurities could not be responsible for the PL emission in AAMs. According to the previous reports on the luminescence of F center in crystalline alumina, the luminescent behavior had been mainly attributed to oxygen vacancies [32,33]. The existence of the F+ centers has been observed in an as-received AAM [34]. Li et al. [16] also reported that the F+ centers in AAMs caused the optical properties of AAMs in the wavelength range of 200–500 nm. Our EPR results further disclosed that the asprepared and annealed AAMs have contained F+ centers. Consequently, the photoluminescence band in AAMs originates from the singly ionized oxygen vacancies (F+ centers). As we see in the XRD patterns in Fig. 2, the AAM is amorphous under the annealing temperature of 700 1C. Undoubtedly, the crystallization of AAMs would occur with the increase of annealing temperature, but the process may not take place completely. So the as-anodized and heat-treated samples should have different local structures due to crystallization of amorphous alumina membrane, that is to say, the different local environment of the oxygen vacancies result in the distinct

EPR features. The spin density can be obtained approximately by calculating the area of absorption signal. The inset of Fig. 4 gives the dependence of the spin density corresponding to signal II on the temperature of the annealing of the samples. From the inset, it can be observed that the spin density corresponding to signal II increases with raising temperature, and reaches a maximum at about 600 1C, but drastically decreases with further increase of temperature. While the intensity of signal I has the trend of decrease with an increase of heat-treating temperature. The correspondence of the dependences between the luminescence and EPR signal II indicates that the F+ centers are responsible for the luminescent centers in AAMs. During the growth of AAMs, the electrochemical anodization of Al may be insufficient leaving some residual metal Al which will be oxidized in the subsequent heating treatment [35]. Because the annealing process was carried out under such an oxygen-poor atmosphere, the oxygen reacting with the remaining Al comes mainly from the porous alumina membrane itself. This might promote the intensive increase of oxygen vacancies in the

ARTICLE IN PRESS X. Sun et al. / Journal of Luminescence 121 (2006) 588–594

AAMs. In addition, the alumina membrane itself would lose a great deal of water during annealing as proved by the TGA experiment and this process also increased the formation of oxygen vacancies. Therefore, the striking increase of PL intensity with annealing temperature when Tap500 1C could be attributed to the increase of F+ centers density. When Ta4500 1C, the PL intensity of AAMs decreased drastically. The appearance of g-Al2O3 after annealing at 800 1C for 6 h proves that the crystallization definitely took place during the heat treatment. A long period heat treatment rearranges the cells and reduces the internal stress and the number of defects. That is to say, it is the ordering process of alumina lattice that reduces the F+ centers density, which inversely leads to the decrease of PL intensity. When the annealing temperature is up to 900 1C, the crystallization of AAMs is further increased and this process will further reduce the F+ centers density. Conse-

PL Intensity (a.u.)


450 nm

Annealing for 6 hour Annealing for 24 hour

20 445 nm

4. Conclusion


0 400






Wavelength (nm)

12 440 nm

10 PL Intensity (a.u.)

quently, the optical properties of AAMs are tightly correlated with the defect structures of the alumina. In the work, we took 300 and 600 1C as typical temperatures to anneal the AAMs for 6 and 24 h respectively to further examine the effect of the defect structures in AAMs on PL properties. For the case of 300 1C as shown in Fig. 5(a), it is found that the increase of annealing time intensifies the intensity of PL, while it is inverse for the 600 1C (Fig. 5(b)). As we know, the AAMs annealed at 300 1C always have amorphous structure no matter how long the annealing process is. The longer the annealing time is, the more the desorbed water is and the more the oxygen vacancies are. Consequently, the PL intensity increased with the annealing time. When the sample is annealed at 600 1C, however, the structure transformation of alumina from amorphous to g phase took place and the g-Al2O3 phase was observed after 24 h (not shown here). The phase transition process of alumina may lead to the change of band gap in addition to the increase of crystallinity, which decreases the concentration of oxygen vacancies. Hence the intensities of PL become weaker.



Annealing for 6 h Annealing for 24 h

8 6

The AAMs with ordered nanopore arrays formed in oxalic acid exhibit very special and interesting optical properties. The PL phenomenon is intimately related to the temperature-induced structural transitions. The experimental results and discussion presented herein reveal that the oxygen-related defect centers, F+ centers, are responsible for the photoluminescence of AAMs.

400 nm

References 4 2 0 300








Wavelength (nm)

Fig. 5. PL spectra of AAMs annealed at 300 and 600 1C in Ar for 6 and 24 h, respectively: (a) 300 1C and (b) 600 1C.

[1] K. Nielsch, F. Mu¨ller, A.P. Li, U. Go¨sele, Adv. Mater. 12 (2000) 582. [2] G. Sauer, G. Brehm, S. Schneider, K. Nielsch, R.B. Wehrspohn, J. Choi, H. Hofmeister, U. Go¨sele, J. Appl. Phys. 91 (2002) 3243. [3] C.R. Martin, Science 266 (1994) 1961. [4] H.Q. Cao, Y. Xu, J.M. Hong, H.B. Liu, G. Yin, B.L. Li, C.Y. Tie, Z. Xu, Adv. Mater. 13 (2001) 1393.


X. Sun et al. / Journal of Luminescence 121 (2006) 588–594

[5] D.J. Pen˜a, J.K.N. Mbindyo, A.J. Carado, T.E. Mallouk, C.D. Keating, B. Razavi, T.S. Mayer, J. Phys. Chem. B 106 (2002) 7458. [6] A.P. Li, F. Mu¨ller, A. Briner, K. Nielsch, U. Go¨sele, Adv. Mater. 11 (1999) 483. [7] R.K. Nahar, V.K. Khanna, Sens. Actuaors B 46 (1998) 35. [8] G. Sberveglieri, R. Murri, N. Pinto, Sens. Actuaors B 23 (1995) 177. [9] A.V. Kukhta, G.G. Gorokh, E.E. Kolesnik, A.I. Mitkovets, M.I. Taoubi, Y.A. Koshin, A.M. Mozalev, Surf. Sci. 507–510 (2002) 593. [10] G.S. Huang, X.L. Wu, Y.F. Mei, X.F. Shao, G.G. Siu, J. Appl. Phys. 93 (2003) 582. [11] J.H. Wu, X.L. Wu, N. Tang, Y.F. Mei, X.M. Bao, Appl. Phys. A: Mater. Sci. Process. 72 (2001) 735. [12] T. Li, S.G. Yang, L.S. Huang, J.R. Zhang, B.X. Gu, Y.W. Du, J. Phys.: Condens. Matter 16 (2004) 2463. [13] W.L. Xu, M.J. Zheng, S. Wu, W.Z. Shen, Appl. Phys. Lett. 85 (2004) 4364. [14] G.S. Huang, X.L. Wu, Y. Xie, X.F. Shao, S.H. Wang, J. Appl. Phys. 94 (2003) 2407. [15] Y. Du, W.L. Cai, C.M. Mo, J. Chen, L.D. Zhang, X.G. Zhu, Appl. Phys. Lett. 74 (1999) 2951. [16] Y. Li, G.H. Li, G.W. Meng, L.D. Zhang, F. Phillipp, J. Phys.: Condens. Matter 13 (2001) 2691. [17] Z.L. Wang, J. Phys.: Condens. Matter 16 (2004) R829. [18] T. Gao, G.W. Meng, L.D. Zhang, J. Phys.: Condens. Matter 15 (2003) 2071.

[19] Y. Yamamoto, N. Baba, S. Tajima, Nature 289 (1981) 572. [20] G.E. Thompson, G.C. Wood, Nature 290 (1981) 230. [21] X.Y. Sun, F.Q. Xu, Z.M. Li, W.H. Zhang, Mater. Chem. Phys. 90 (2005) 69. [22] R. Krishnan, S. Dash, C.B. Rao, R.V.S. Rao, A.K. Tyagi, B. Raj, Scr. Mater. 45 (2001) 693. [23] E.D. Macklen, J. Inorg. Nucl. Chem. 30 (1968) 2689. [24] D. Dollimore, D. Nicholson, J. Chem. Soc. (1962) 960. [25] D. Dollimore, D.L. Griffiths, J. Thermal. Anal. 2 (1970) 229. [26] J. Paulik, F. Paulik, L. Eroley, Anal. Chim. Acta. 44 (1969) 153. [27] P.P. Mardilovich, A.N. Govyadinov, N.I. Mukhurov, A.M. Rzhevskii, R. Paterson, J. Membrane. Sci. 98 (1995) 131. [28] J.A. Wang, X. Bokhimi, A. Morales, O. Novaro, J. Phys. Chem. B 103 (1999) 299. [29] T. Ishizaka, S. Tero-Kubota, Y. Kurokawa, T. Ikoma, J. Phys. Chem. Solids 64 (2003) 801. [30] U. Roland, F. Holzer, A. Po¨ppl, F.-D. Kopinke, Appl. Catal. B: Environ. 58 (2005) 227. [31] J. Chen, W.L. Cai, J.M. Mou, J. Inorg. Mater. 16 (2001) 677. [32] K.H. Lee, J.H. Crawford Jr., Phys. Rev. B 19 (1979) 3217. [33] B.D. Evans, M. Stapelbroek, Phys. Rev. B. 18 (1978) 7089. [34] F.A. KrO¨ger, The Chemistry of Imperfect Crystals, NorthHolland Publishing Co, Amsterdam, 1964. [35] J. Siejka, C. Ortega, J. Electrochem. Soc. 124 (1997) 883.