⇦Chinese Physics Letters, 2018, Vol. 35, No. 2, Article code 027501 Fe$^{3+}$-Doped Anatase TiO$_{2}$ Study Prepared by New Sol-Gel Precursors Bahram Khoshnevisan1, Mohammad Bagher Marami1, Majid Farahmandjou2** Affiliations 1Department of Physics, Faculty of Science, University of Kashan, Kashan, Islamic Republic of Iran 2Departments of Physics, Varamin Pishva Branch, Islamis Azad University, Varamin, Iran Received 4 November 2017 **Corresponding author. Email: farahmandjou@iauvaramin.ac.ir Citation Text: Khoshnevisan B, Marami M B and Farahmandjou M 2018 Chin. Phys. Lett. 35 027501 Abstract Fe$_{x}$Ti$_{1-x}$O$_{2}$ ($x=0.00$, 0.05, 0.10) nanocomposites are synthesized using a sol-gel method involving an ethanol solvent in the presence of ethylene glycol as the stabilizer, and acetic acid as the chemical reagent. Their structural and optical analyses are studied to reveal their physicochemical properties. Using the x-ray diffractometer (XRD) analysis, the size of the nanoparticles (NPs) is found to be 18–32 nm, where the size of the NPs decreases down to 18 nm when Fe impurity of up to 10% is added, whereas their structure remains unchanged. The results also indicate that the structure of the NPs is tetragonal in the anatase phase. The Fourier transform infrared spectroscopy analysis suggests the presence of a vibration bond (Ti–O) in the sample. The photoluminescence analysis indicates that the diffusion of Fe$^{3+}$ ions into the TiO$_{2}$ matrix results in a decreasing electron–hole recombination, and increases the photocatalytic properties, where the best efficiency appears at an impurity of 10%. The UV-diffuse reflection spectroscopy analysis indicates that with the elevation of iron impurity, the band gap value decreases from 3.47 eV for the pure sample to 2.95 eV for the 10 mol% Fe-doped TiO$_{2}$ NPs. DOI:10.1088/0256-307X/35/2/027501 PACS:75.50.Tt, 75.75.Cd, 81.16.-c, 81.07.-b, 73.63.Bd © 2018 Chinese Physics Society Article Text In recent years, titanium dioxide (TiO$_{2}$) has been used extensively for applications in different industries due to its suitable physicochemical properties. This has been shown in bulk form, or in micro/nano sizes.[1-9] TiO$_{2}$ has a tetragonal structure and exhibits different phases including rutile, anatase and brookite. The crystal structure of every phase depends on the way the octahedral blocks are connected in the TiO$_{2}$ crystal network.[6] Therefore, each of them shows different properties and applications. Currently, the rutile phase is used as a ferromagnetic material[4] as well as a catalyst in gas sensors,[5] and the anatase phase is employed as an anode in lithium-ion batteries,[7] photocatalysts,[6] filters for ultraviolet waves,[8] production of hydrogen gas,[9] and solar cells.[6] Due to its chemical stability, lack of toxicity, low price, electric and optical characteristics, and high optical activities (photo activity), titanium dioxide has been used as one of the most common semiconductor photocatalysts.[10-12] Recently, we reported a successful synthesis of TiO$_{2}$ NPs using a low price sol-gel method[13] for use in photocatalytic and optoelectronic systems. Apart from the advantages mentioned above, this compound has the following disadvantages: less utilization of photons (less than 4% of sunlight), a relatively high rate of recombination of electrons and holes produced by light, and activity limited to a specific wavelength range (less than 400 nm). Doping TiO$_{2}$ with transition metal ions, such as Fe$^{3+}$, Co$^{3+}$, Ni$^{2+}$ and Cu$^{+}$, is one of the modifications for these catalysts to extend the wavelength into the visible region.[14,15] Studies have shown that transition metal elements including Fe$^{3+}$ can be used to enhance photocatalytic activity.[16] The Fe$^{3+}$ metal ions can easily be replaced in the TiO$_{2}$ network structure due to its electron configuration. In addition, its ionic radius is close to that of Ti$^{4+}$, thus it can enhance the photocatalytic properties within the visible light region.[10-12] Therefore, the Fe$^{3+}$ ion has been considered as an effective and considerable doping agent compared with the other transition metals.[17] There are various methods for preparing Fe-doped TiO$_{2}$ NPs, which include wet impregnation, co-precipitation, sol-gel, hydrothermal, and other methods.[18-20] Out of these, the sol-gel method has been widely used.[17] In this work, Fe$^{3+}$-TiO$_{2}$ NPs have been synthesized with different percentages of iron impurity using the sol-gel method, and their structural, optical, morphological and photocatalytic properties have been examined using the x-ray diffractometer (XRD), FESEM, EDS, the Fourier transform infrared spectroscopy (FTIR), UV-diffuse reflection spectroscopy (UV-DRS), and photoluminescence (PL) analyses. Pure and Fe-doped TiO$_{2}$ NPs were synthesized using TiCl$_{4}$ and FeSO$_{4}\times$8H$_{2}$O as new precursors by the sol-gel method. To synthesize pure TiO$_{2}$, firstly 10 mL of TiCl$_{4}$ was dissolved in pure ethanol using a magnetic stirrer at room temperature. After 10 min, 5 mL of acetic acid was added to the solution, and then the temperature was elevated to 70$^{\circ}\!$C. At this temperature, 2 mL of ethylene glycol (EG) stabilizer was added. The solution was completely evaporated, and then dried. The powder thus obtained was washed with distilled water and ethanol several times to remove the impurities in the sample. Eventually, the product was incubated at 600$^{\circ}\!$C for 4 h to form a crystal structure. For fabricating an impure sample, a method similar to the one used for the preparation of the pure TiO$_{2}$ sample was adopted. Firstly, the FeSO$_{4}$.8H$_{2}$O iron precursor was dissolved in pure water to obtain a 1 M solution. It was then added to the TiCl$_{4}$ solution in ethanol with 5% and 10% molar percentages. Similarly, it was allowed to dry at 70$^{\circ}\!$C to obtain a powder. The solution pH was adjusted to 3 with an acetic acid reagent.
cpl-35-2-027501-fig1.png
Fig. 1. (a) Compared XRD spectra of pure and Fe-doped TiO$_{2}$ nanocrystals for different Fe concentrations. (b) Variation of crystal size and unit cell volume versus Fe dopant.
XRD analysis was carried out to determine the crystal structure, the phases in the sample, and the lattice parameters of the pure and Fe-doped TiO$_{2}$ NPs. All the samples were incubated at 600$^{\circ}\!$C for 4 h, and then were analyzed. Eventually, their crystal structures in the pure and impure states of Fe (including 5% and 10% molar percentages) were examined. The results indicated that the pure sample of TiO$_{2}$ had a tetragonal structure with a mixed anatase phase, and some percentage of the rutile phase (8.4%). For the samples with impurity percentages of 5% and 10%, the rutile phase decreased by up to 5.7% and 2.25%, respectively, and the samples almost entered the anatase phase. The results also showed that in spite of the iron impurities within the TiO$_{2}$ structure, the structure remains unchanged (with slight changes in the lattice constants), signifying that the iron impurity does not influence the TiO$_{2}$ matrix. Figure 1(a) reveals the XRD spectra of the pure and impure samples of Fe. As shown in the figure, the major peaks of the anatase phase at a $2\theta$ angle emerge at 25.37, 37, 37.85, 38.62, 48.11, 53.94, 55.12, 62, 72, 68.84, 70.33, 75.09 and 76.13, which correspond to the crystal planes of (101), (103), (004), (112), (200), (105), (211), (118), (116), (220), (125) and (031), respectively. This suggests that the anatase phase has a tetragonal structure. Further, the weak peaks emerge at $2\theta$ angles of 27.49, and 41.30, corresponding to (110) and (210) crystal planes, respectively, suggesting that the rutile phase occurs in the pure TiO$_{2}$ sample. By applying 10% Fe impurity, the rutile level diminishes, and the sample changes into the anatase phase. It can also be obtained that the ionic radii of Fe$^{3+}$ and Ti$^{4+}$ are very close to each other ($a_{\rm Fe}^{3+}=0.64$ Å, and $a_{\rm Ti}^{4+}=0.68$ Å), thus Fe$^{3+}$ ions can easily replace the Ti$^{4+}$ ions in the TiO$_{2}$ crystal structure, resulting in a constancy of the TiO$_{2}$ structure. The particle size for pure and Fe-doped TiO$_{2}$ samples was calculated using the Debye–Scherrer formula.[21] Figure 1(b) indicates the changes in the size of the NPs in terms of the impurity percentage of Fe. The size of the pure TiO$_{2}$ NPs is 32 nm, and the sizes of the impure Fe$_{x}$Ti$_{1-x}$O$_{2}$ NPs have been measured to be 23 and 18 nm for $x=0.05$ and $x=0.10$, respectively. In addition, as the ionic radius of Fe$^{3+}$ is smaller than that of Ti$^{4+}$, with the entrance of the Fe$^{3+}$ ions into the TiO$_{2}$ matrix, the particle size decreases. In Fig. 1(b), the unit cell volume also increases with the elevation of impurity by up to 5%, whereas it is almost constant for 10% impurity. The elevation of the impurity percentage increases the unit cell volume due to an over-aggregation of the iron atoms in the TiO$_{2}$ network structure, and increases the lattice constant of network $a$.
cpl-35-2-027501-fig2.png
Fig. 2. Plot of diffraction peak of Fe-doped TiO$_{2}$ (a) $d(101)$, $2\theta(101)$; (b) $d(004)$, $2\theta (004)$; and (c) $d(200)$, $2\theta (200)$. (d) Lattice parameters $a$ and $c$ versus Fe dopant (mol%).
Figure 2 reveals the changes in the lattice parameters in terms of the iron impurity percentage. As can be seen from the analysis, an elevation of the impurity percentage of up to 5% leads to an increase in the distance between (101) and (200) planes, whereas along (400), this distance decreases (Figs. 2(a)–2(c)). Indeed, the Fe$^{3+}$ atoms have replaced the Ti$^{4+}$ ions in the TiO$_{2}$ matrix, mostly in the aforementioned crystal direction. Further, as they have a smaller ionic radius, the distance between the planes within this direction declines. In addition, increasing the iron impurity by up to 10% causes the distance between the planes to decrease due to the over-aggregation of iron atoms in the TiO$_{2}$ matrix, which has the maximum percentage of the anatase phase in the crystal structure. An alteration of the $2\theta$ angle causes the planes of (100) and (200) to incline towards smaller angles with the elevation of the impurity percentage up to 5%, whereas the plane of (004) inclines towards larger angles. The reason for these changes is due to the phenomenon that all Fe$^{3+}$ ions cannot be replaced in the TiO$_{2}$ matrix completely, and hence, a distortion is developed at the $2\theta$ angle sites in these crystal planes. Figure 2(d) represents the changes in the lattice constants of $a$ and $c$ in terms of the iron impurity percentage. Evidently, the elevation of iron impurity causes the lattice constant $a$ to increase. On the other hand, the lattice constant $c$ decreases, which corresponds to the changes in the distance between the (200) and (004) planes. Indeed, the very trivial difference in the ionic radii of Fe$^{3+}$ and Ti$^{4+}$ (around 0.04 Å) results in a diminished lattice constant $c$. The changes in the lattice parameters for the different iron impurity percentages are listed in Table 1. The alterations of particle size, unit cell volume, particle volume, the number of unit cells in each particle, band gap energy, and FWHM with variations in the iron impurity percentage have been listed in Table 2.
Table 1. Lattice parameters and Bragg angles ($2\theta$) of pure and Fe-doped TiO$_{2}$ nanocrystals.
Fe dopant 2$\theta$(deg) $d$ (Å) 2$\theta$(deg) $d$ (Å) 2$\theta$(deg) $d$ (Å) $a=b$ $c$ $c/a$
(mol%) (101) (101) (200) (200) (004) (004) (Å) (Å)
JCPDS 25.304 3.51685 48.037 1.89250 37.797 2.37825 3.7850 9.5130 2.51334
Pure 25.374 3.50721 48.112 1.88969 37.857 2.37462 3.70942 9.49873 2.5607
5 25.365 3.50863 48.089 1.89054 37.868 2.37395 3.78181 9.4961 2.51099
10 25.4 3.50375 48.105 1.88998 37.905 2.37175 3.77736 9.48703 2.51155
Table 2. FWHM, particle size, volume of the NPs, volume of the unit cell, number of unit cells, band gap energy and lattice strain of pure and Fe-doped TiO$_{2}$ samples.
Fe $D$ $V$ NPs $v$ Tetragonal $V/v$ Number Band gap Lattice strain
dopant FWHM Particle Volume unit-cell volume of unit cell per energy $\xi$ along
(mol%) size (nm) (nm$^{3}$) (10$^{-3}$nm$^{3}$) particle (10$^{3}$) (eV) the $a$-axis (10$^{-4}$)
Pure 0.335 32 17148 130.7006 131.2 3.47 20
5 0.347 23 6367 135.814 46.9 3.30 20
10 0.441 18 3052 135.5652 32.5 2.95 28
cpl-35-2-027501-fig3.png
Fig. 3. FESEM images of (a) pure, (b) 5% and (c) 10% Fe-doped TiO$_{2}$.
The appearance and mean diameter of the NPs were analyzed by a field emission scanning electron microscope (FESEM). As shown in Fig. 3, elevation of the iron impurity decreases the homogeneity of the Fe-doped TiO$_{2}$ NPs in comparison with pure TiO$_{2}$, which is due to the reduction of the particle size. Evidently, with the reduction of size of the particles, inter-atomic attraction and molecular forces grow, thereby the tendency of the NPs is enhanced to approach each other, which causes agglomeration.[22] Figure 3(a) shows the pure TiO$_{2}$ NPs with a mean size of 33 nm, and Figs. 3(b) and 3(c) reveal TiO$_{2}$ NPs with iron impurity of 5% and 10% with a mean size of 24 nm and 20 nm, respectively, which are in congruence with the results of the XRD analysis. The EDX analysis has been used to determine the elemental percentage in the sample. Figure 4 demonstrates the x-ray energy diffraction spectrum for iron impurity of 5%. Clearly, only O, Ti and Fe elements exist in the sample, indicating the purity of the sample. In the sample, the weight percentage of iron has been measured as 2.3%, which is less than the 5% molar percentage used in the experiment. This difference in the percentages is due to the fact that in the EDX analysis, the surface of the sample is analyzed and scanned, whereas during the fabrication of the sample, the molar percentage is calculated per sample volume.
cpl-35-2-027501-fig4.png
Fig. 4. EDX analysis of 5 mol% of Fe-doped TiO$_{2}$ samples. The inset shows the image of Fe-doped TiO$_{2}$ NPs.
The TEM analysis of Fe-doped TiO$_{2}$ with a 5% level of impurity is displayed in Fig. 5. The TEM study confirms that the particle's size varies in the range 18–28 nm. Uniformity of the particles is desirable and the average size of particles fitting the Gaussian function is about 23.12 nm. According to XRD and FESEM analyses, the average sizes of the particles are measured to be 23 nm and 24 nm, respectively. The FTIR analysis (within 400–4000 cm$^{-1}$) has been used to identify the bonds in the sample. Figure 6 shows the FTIR on the sample in terms of wavenumber. The absorption peaks developed in the spectrum represent the existence of vibration bonds in the sample. The comparison of the spectrum of the pure TiO$_2$ sample and the 5% impurity sample demonstrates that when the iron impurity is incorporated into the TiO$_{2}$ matrix, an absorption peak emerges at the wavenumber of 2992 cm$^{-1}$, which belongs to the O–H vibration bond. The sharp absorption peak at 440 cm$^{-1}$ is attributed to the TiO$_{2}$ sample, and the one developed at 465 cm$^{-1}$ is for the 5% impure sample, which is associated with the Ti–O group. Within 650–1800 cm$^{-1}$, a very sharp peak is developed due to the entrance of environmental oxygen in the sample during the synthesis and fabrication of the NPs.
cpl-35-2-027501-fig5.png
Fig. 5. (a) TEM analysis of 5 mol% of Fe-doped TiO$_{2}$ samples. (b) Size distribution of the samples, fitted with the Gaussian curve.
cpl-35-2-027501-fig6.png
Fig. 6. FTIR spectrum of the Fe-doped TiO$_{2}$ nanocomposite for different Fe dopants.
cpl-35-2-027501-fig7.png
Fig. 7. (a) Reflection spectrum of the Fe-doped TiO$_{2}$ samples. (b) Plot of $[F(R)h\nu]^{1/2}$ versus $E(h\nu)$ with different Fe contents.
Figure 7(a) represents the UV reflective spectrum within the wavelengths of 280–300 nm. To determine the absorption wavelength and bandgap energy ($E_{\rm g}$) of the sample, the UV-DRS optical analysis has been used. The analysis shows that as the impurity increases, the maximum wavelengths ($\lambda_{\rm edge}$) are attained at 357, 375 and 420 nm, for the $x=0$, $x=0.05$ and $x=0.10$ samples, respectively. This reduction in the $\lambda_{\rm edge}$ observed as the impurity percentage increases leads to the development of a redshift, and consequently, a diminished bandgap energy. For calculating $E_{\rm g}$ of the samples, the Tauc plot $(F(R)h\nu)^{1/2}$ has been plotted in terms of the photon energy $E(h\nu)$, where $F(R)$ is the Kubelka–Munk function, and $R$ is the percentage.[23] Figure 7(b) represents the Tauc curve in terms of the iron impurity. The maximum slope of the line tangential to the curve, and its intersection with the energy axis give the bandgap energy value. The results show that as the percentage of impurity increases, the bandgap energy diminishes, such that its values are obtained to be 3.47, 3.30 and 2.95 eV for the $x=0.00$, $x=0.05$ and $x=0.10$ samples, respectively. Moreover, factors such as the size of the impure particles, defects in the lattice, and alteration of the lattice parameters influence the value of the bandgap energy.[24] Researchers have also reported that a reduction of $E_{\rm g}$ is a result of the $sp$–$d$ exchange in the interaction between the band electrons of TiO$_{2}$ and $d$ electrons of the Fe$^{3+}$ impurity bands.[24]
cpl-35-2-027501-fig8.png
Fig. 8. Photoluminescence spectroscopy of Fe-doped TiO$_{2}$ samples for different levels of Fe dopant.
Figure 8 reveals the photoluminescence spectrum of the Fe$_{x}$Ti$_{1-x}$O$_{2}$ sample. To determine the recombination of the sample due to the behavior of the electron–hole pair, the PL optical analysis has been used. As shown in Fig. 8, as the impurity of Fe increases, the intensity of photoluminescence decreases, suggesting that the extent of recombination of the induced electron–hole pairs diminishes, and thus the photocatalytic property of the sample grows. Some peaks are developed at the wavelengths of 342, 363, 376, 390, 425, 447 and 495 nm, which correspond to the energy bandgaps of 3.62, 3.41, 3.29, 3.18, 2.91, 2.77 and 2.70 eV, respectively. Moreover, with the addition of Fe$^{3+}$ impurity, entrapment of the holes by the Ti$^{4+}$ band in the conduction band is prevented by developing the iron impurity level in the bottom of the conduction band, thereby preventing recombination of the electrons and holes. Accordingly, the pairing of electrons and holes in the bottom of the conduction band in the Fe impurity level occurs, causing diminished intensity of PL, and thus leads to the reduced bandgap energy.[25] In conclusion, Fe-doped TiO$_{2}$ NPs have been successfully synthesized using the sol-gel method. The XRD results indicate that in pure TiO$_{2}$ NPs, there is a minor percentage of the rutile phase, whereas with the addition of Fe impurity by up to 10%, a tetragonal structure emerges in the anatase phase. Furthermore, the size of the nanocrystals decreases with the addition of iron impurity by up to 18 nm, without any change in the crystal structure. The PL optical analysis shows that with elevation of impurity percentage by up to 10%, the photoluminescence intensity decreases, causing a reduction of the recombination of electrons and holes, and elevation of the photocatalytic activity. Eventually, the UV-DRS optical analysis shows that with increasing the iron impurity percentage, the maximum wavelength inclines towards larger wavelengths, whereby a redshift state is reached, causing the reduction of bandgap energy by up to 2.95 eV. References Angle-Dependent Optical Effects Deriving from Submicron Structures of Films and PigmentsLow-temperature and one-step synthesis of rutile TiO2 aqueous sol by heterogeneous nucleation methodC-V characteristics of metal-titanium dioxide-silicon capacitorsRoom-Temperature Ferromagnetism in Transparent Transition Metal-Doped Titanium DioxidePhotoresponse of a titanium dioxide chemical sensorPhotoinduced reactivity of titanium dioxideLithium Insertion into Mesoscopic and Single-Crystal TiO[sub 2] (Rutile) ElectrodesDoping of a nanostructured titania thick film: structural and electrical investigationsGold, silver and copper catalysts supported on TiO2 for pure hydrogen productionPreparation and its application of TiO2/ZrO2 and TiO2/Fe photocatalysts: A perspective studyEnhanced photocatalytic water-splitting performance using Fe-doped hierarchical TiO 2 ball-flowersPhotocatalytic performance of Fe-doped TiO 2 nanoparticles under visible-light irradiationControlled Nitrogen Doping and Film Colorimetrics in Porous TiO 2 Materials Using Plasma ProcessingDirectional nickel-induced fielded aided lateral crystallization of amorphous siliconStructural and Magnetic Properties of Fe-Doped Anatase TiO 2 Films Annealed in VacuumPhotocatalytic performance of Fe-doped TiO 2 nanoparticles under visible-light irradiationDoping and transformation mechanisms of Fe 3+ ions in Fe-doped TiO 2Surfactants Effect on the Photoactivity of Fe-Doped TiO2Effect of high manganese substitution at ZnO host lattice using solvothermal method: Structural characterization and propertiesInteractions Between Photoinduced and Dark Charge Transfer across n-TiO[sub 2]-Aqueous Electrolyte InterfaceRole of Nanoparticles in Photocatalysis
[1] Pfaff G and Reynders P 1999 Chem. Rev. 99 1963
[2] Bin Q, Wu L, Zhang Y, Zeng Q and Zhi J 2010 J. Colloid Interface Sci. 345 181
[3] Brown W D and Granneman W W 1978 Solid-State Electron. 21 837
[4] Matsumoto Y, Murakami M, Shono T, Hasegawa T, Fukumura T, Kawasaki M, Ahmet P, Chikyow T, Koshihara S and Koinuma H 2001 Science 291 854
[5] Kumazawa N, Islam M R and Takeuchi M 1999 J. Electrochem. Chem. 472 137
[6] Carp O, Huisman C L and Reller A 2004 Prog. Solid State Chem. 32 33
[7] Kavan L, Fattakhova D and Krtil P 1999 J. Electrochem. Soc. 146 1375
[8] Bonini N, Carotta M C, Chiorino A, Guidi V, Malagù C, Martinelli G, Paglialonga L and Sacerdoti M 2000 Sens. Actuators B 68 274
[9] Boccuzzi F, Chiorino A, Manzoli M, Andreeva D, Tabakova T, Ilieva L and Iadakiev V 2002 Catal. Today 75 169
[10] Foura G, Chouchou N, Soualah A, Kouachi K, Guidotti M and Robert D 2017 Catalysts 7 344
[11] Wang F, Shen T, Fu Z, Lu Y and Chen C 2018 Nanotechnology 29 035702
[12] Ali T, Tripathi P, Azam A, Raza R, Ahmed A S, Ahmed A and Muneer M 2017 Mater. Res. Express 4 015022
[13]Ramazani M, Farahmandjou M and Firoozabadi T P 2015 Phys. Chem. Res. 3 293
[14] Pulsipher D I, Martin I T and Fisher E R 2010 ACS Appl. Mater. Interfaces 2 1743
[15] Izmajlowicz M A T, Flewitt A J, Milne W I and Morrison N A 2003 J. Appl. Phys. 94 7535
[16] Xu J P et al 2009 Chin. Phys. Lett. 26 097502
[17]Hu X, An T, Zhang M, Sheng G and Fu J 2007 Res. J. Chem. Environ. 11 13
[18] Ali T, Tripathi P, Azam A, Raza W, Ahmed A S, Ahmed A and Muneer M 2017 Mater. Res. Express 4 015022
[19] Yang Y, Yu Y, Wang J, Zheng W and Cao Y 2017 Cryst. Eng. Commun. 19 1100
[20]Ramalingam R J, Arunachalam P, Radhika T, Anju K R, Nimitha K C and Al-Lohedan C H 2017 Int. J. Electrochem. Sci. 12 797
[21]Scherrer P 1918 Göttinger Nachr. Gesell. 2 98
[22] Chen Q, Xue C, Li X and Wang Y 2013 Mater. Sci. Forum 743 367
[23] Khatoon S, Wani I A, Ahmed J, Magdaleno T, Al-Hartomy O A and Ahmad T 2013 Mater. Chem. Phys. 138 519
[24] Siripala W and Tomkieviez M 1982 J. Electrochem. Soc. 129 1240
[25] Beydoun D, Amal R, Low G and McEvoy S 1999 J. Nanopart. Res. 1 439