Temperature-dependence on the structural, optical, and paramagnetic properties of ZnO nanostructures

Applied Surface Science 293 (2014) 62–70

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Temperature-dependence on the structural, optical, and paramagnetic properties of ZnO nanostructures Gugu H. Mhlongo a,∗ , David E. Motaung a,∗∗ , Steven S. Nkosi b , H.C. Swart c , Gerald F. Malgas d , Kenneth T. Hillie a,c , Bonex W. Mwakikunga a a National Centre for Nano-structured Materials, Council for Scientific and Industrial Research, 1-Meiring Naude Road, Brummeria, PO Box 395, Pretoria 000, South Africa b CSIR-National Laser Centre, 626 Meiring Naude Road, Brummeria, Pretoria 0001, South Africa c Department of Physics, University of the Free State, Bloemfontein ZA9300, South Africa d Department of Physics, University of the Western Cape, P/Bag X17, Bellville, South Africa

a r t i c l e

i n f o

Article history: Received 22 July 2013 Received in revised form 13 December 2013 Accepted 14 December 2013 Available online 22 December 2013 Keywords: Microwave-assisted synthesis ZnO nanostructures Annealing effects Optical properties Magnetic properties

a b s t r a c t Violet-blue emitting ZnO nanostructures were synthesized by a microwave-assisted hydrothermal method followed by post-synthesis annealing at different temperatures. Scanning electron microscope analysis revealed a morphological transformation upon increasing annealing temperature from welldefined “flower-like” structure composed of ZnO multi-nanorods to randomly oriented worm-like ZnO nanostructures. Raman analysis showed that the E2 (high) mode became sharper and stronger while the intensity of the phonon peak at 580 cm−1 was gradually enhanced with the increase of annealing temperature. X-ray diffraction and X-ray photoelectron spectroscopy (XPS) measurements showed that all ZnO samples possess a typical wurtzite structure with high crystallinity and no other impurity phases were observed. A decreasing trend in the photoluminescence (PL) intensity of a strong broad violet-blue emission from ZnO nanostructures with increasing annealing temperature was also observed. The electron spin resonance (ESR) signal was also found to gradually decrease with increasing annealing temperature indicating the decrease in the concentration of zinc interstitials (Zni ) and/or zinc vacancies (VZn ) defects in ZnO nanostructures. Moreover, a combination of results from the PL, XPS and ESR suggested that Zn related defects; especially VZn and Zni are the primary source of the paramagnetism observed in the ZnO nanostructures. © 2013 Elsevier B.V. All rights reserved.

1. Introduction ZnO has attracted a lot of research interest because of its wide direct band-gap of 3.37 eV, which is useful for a variety of applications in light emitting diodes [1], gas sensors [2–4], solar cells [5,6] photocatalytic agents [7] and other technological applications [8]. ZnO has a larger exciton binding energy (60 meV) than GaN (26 meV) and ZnSe (20 meV), respectively, which can ensure an efficient exciton emission at room temperature under low excitation energy. Recently, the renewed interest in nanostructured ZnO has been driven by its attractive prospects for applications in room temperature nanolasers [9], p-type doping [10], and room temperature ferromagnetic semiconductor nanomaterials [11]. These prospects

∗ Corresponding author. Tel.: +27 12 841 3935; fax: +27 12 841 2229. ∗∗ Corresponding author. Tel.: +27 12 841 4775; fax: +27 12 841 2229. E-mail addresses: [email protected] (G.H. Mhlongo), [email protected] (D.E. Motaung). 0169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apsusc.2013.12.076

have led to extensive studies on numerous aspects of ZnO nanostructures, including synthesis strategies, physical and chemical properties, and applications. However, some of the basic properties of ZnO are not well understood [12,13] and are still debatable owing to the different intrinsic defects such as oxygen interstitials (Oi ), zinc interstitials (Zni ), oxygen vacancies (VO ), zinc vacancies (VZn ). At room temperature, ZnO typically exhibits one emission peak in the UV region due to the recombination of free excitons and possibly one or more peaks in the visible spectral range which are attributed to defect emissions, but the origin of defect emissions is still unclear. For the most commonly observed green defect emission [14–19], various hypotheses have been proposed such as transition between singly ionized VO and photo-excited holes [20,21], transition between electrons close to the conduction band and deeply trapped holes at VO [19,22]. The orange-red emission is also commonly attributed to the presence of excess oxygen in the samples [17,23] such as Oi defects [19,22,24]. The violet-blue emission has been previously attributed to electronic transitions from Zni to valence band and transitions between VO and Oi [25,26]. Other

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

101

(a)

800000

as prepared o 200 C o 600 C o 900 C

002

100

103

110

400000

112

300000

112

500000

102

Intensity [a.u]

700000 600000

63

200

200000 100000 0 20

30

40

50

60

70

2 theta [degrees] 40 0.0055

35

0.0050

0.0045 30 0.0040

FWHM of (101) peak

Crystallite size [nm]

(b)

25 0

200

400

600

800

1000

o

Calcination temperatureC][ Fig. 1. (a) The XRD patterns of the ZnO nanostructures before and after annealing at different temperatures. (b) The FWHM of (1 0 1) diffraction peaks and estimated crystallite size of the ZnO nanostructures depending on annealing temperature.

researchers attributed this emission to related intrinsic defects such as VO and VZn or interstitials and their complexes [24,27]. It is obvious based on these reports that the origin of defect emissions in ZnO is still an unresolved issue. The proposed explanations for the different visible emissions in ZnO are often contradictory, with different defect types proposed to explain the same emission or the same defect types proposed to explain emission in the different spectral ranges. Therefore, considerable interest is still being shown in exploring the ZnO nanostructures as magnetic semiconductors. In this paper, we report on a hydrothermal synthesis of one-dimensional ZnO nanostructures followed by annealing at different temperatures. The origin of defects responsible for violetblue emissions in ZnO nanostructures has been investigated by the aid of electron spin resonance (ESR), X-ray photo-electron spectroscopy (XPS), and photoluminescence (PL). 2. Experimental ZnO nanostructures were prepared from aqueous solutions of zinc acetate (Zn(CH3 COO)2 ·2H2 O) and hydrazine hydrate (N2 H4 ). These precursor solutions were mixed in a molar ratio of 1:4 in 100 ml of distilled water under vigorous stirring. N2 H4 solution was added drop-wise into zinc acetate solution and it slowly reacted

with zinc acetate resulting in a formation of slurry-like precipitate. The solution was continuously stirred for 15 min, then transferred in a 100 ml Teflon liner and finally subjected to microwave (Perkin Elmer/Anton Paar Multiwave 3000) oven irradiation at a microwave power of 150 W for 10 min. The resulting white precipitate at the bottom was collected, then filtered off and then washed several times with absolute ethanol and distilled water to remove any impurities. For preparation of ZnO nanostructures, the remaining powder was then dried in air for 2 h at 90 ◦ C. Annealing in air was also performed at different temperatures from 200 to 900 ◦ C for 2 h. All the chemicals were purchased from Sigma Aldrich and used as received without further purification. The samples structure was analyzed by X-ray diffraction (XRD) (PANalytical Xpert PRO) using the Cu K␣ radiation source. The morphologies were observed from the JEOL JSM-7500F, field emission scanning electron microscopy (FE-SEM). Optical properties were characterized using room temperature UV-VIS absorption (Perkin-Elmer Lamda 750S UV-VIS) and PL (Jobin-Yvon NanoLog spectrometer) at an excitation wavelength of 325 nm. X-ray photoelectron spectroscopy (XPS) analyses were carried using a PHI 5000 Versaprobe-Scanning ESCA Microprobe. A low energy Ar+ ion gun and low energy neutralizer electron gun were used to minimize charging on the surface. Monochromatic Al K␣ radiation

64

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

Fig. 2. FE-SEM images of ZnO nanostructures (a) before and after annealing at different temperatures (b) 200 ◦ C, (c) 600 ◦ C, and (d) 900 ◦ C.

(h = 1486.6 eV) was used as the excitation source. A 25 W, 15 kV electron beam was used to excite the X-ray beam of 100 ␮m diameter that was used for the analyses. For the higher resolution spectra the hemispherical analyser pass energy was maintained at 11.8 eV (C 1s, O 1s, and Zn 2p) for 50 cycles. Measurements were performed using either a 1 eV/step and 45 min acquisition time (binding energies ranging from 0 to 1400 eV) for survey scans or a 0.1 eV/step and 20–30 min acquisition times for the high resolution scans. The pressure during acquisition was typically below 1 × 10−8 Torr. The same measurements were repeated after the surfaces were sputtered clean for 30 s using an Ar+ ion gun (2 kV energy ions). The Raman spectroscopy measurements were conducted using a Horiba Jobin-Yvon HR800 Raman microscope at room temperature with a 514 nm excitation laser with a spectral resolution of 0.4 cm−1 . The microwave absorption measurements were carried out using JEOL X-band electron spin resonance (ESR) spectrometer JES FA 200 equipped with an Oxford ESR900 gas-flow cryostat and a temperature control (Scientific instruments 9700). The microwave power was kept constant at 30 (5) mW while the frequency was about 9.4 GHz. The DC field was modulated with a superposed ac field whose amplitude was maintained constant. The microwave response was measured as a derivative signal. 3. Results and discussions Fig. 1(a) shows the XRD diffraction patterns of ZnO nanostructures before and after annealing at different temperatures (200, 600 and 900 ◦ C). All the diffraction patterns correspond to the wurtzite structure (hexagonal phase) of ZnO. No additional diffraction peaks due to impurities were detected. The crystallographic phases were identified by comparison with the XRD patterns of pure ZnO JCPDS card No. 79-2205 database. The preferred orientation corresponding to the (1 0 1) plane was the most prominent diffraction peak. It was also noted that the primary diffraction peaks became stronger and sharper after the annealing temperature was increased up to 900 ◦ C indicating improved crystallinity of the ZnO nanostructures. By using the Debye–Scherrer’s equation, the full width at half-maximum (FWHM) of the most intense

diffraction peak (1 0 1) and the average crystallite size were estimated and the results are shown in Fig. 1(b). It can be seen from the plot that, the increase of annealing temperature led to the decrease of the FWHM of the (1 0 1) diffraction peaks while the average crystallite size is increased. However, for the ZnO sample annealed at 600 ◦ C, a slight increase and decrease in FWHM of the (1 0 1) diffraction peak and average crystallite size was noted, respectively. The decrease of the FWHM with increasing annealing temperature could be due to the coalescences of grains at higher temperatures thus leading to increase in the average crystallite size due to improvement of crystallinity of ZnO nanostructures [28]. The FE-SEM images presented in Fig. 2 show the morphological change of ZnO nanostructures before and after annealing at different temperatures. It can be seen from Fig. 2(a) that prior to annealing; the ZnO nanostructures consisted of a bunch of multi-nanorods originating from one nucleating centre to form “flower-like” structures. A slight increase in annealing temperature to 200 ◦ C resulted in the formation of a hexagonal structure with a smooth surface at the bottom of each bunch of multi-nanorods, Fig. 2(b). At 600 ◦ C, these multi-nanorods were clustered together and then deviated into cones as shown in Fig. 2(c). However, it was also noticed that with a further increase of annealing temperature up to 900 ◦ C (Fig. 2(d)), the structures were made-up of small rings packing up together towards the end of the tip, having the shape of a worm with smooth surface. Fig. 3 shows the optical absorption spectra of the as prepared and annealed ZnO nanostructures at different temperatures. The absorption spectra revealed absorption peaks around 378 nm, corresponding to the as-prepared ZnO sample. When annealing the samples to 200 ◦ C and higher temperatures an obvious red-shift in the absorption edge was observed. This red-shift might be due to changes in the ZnO nanostructures’ morphologies and surface microstructure. Also, the presence of shallow levels inside the band gap due to some defects or impurities present in the ZnO lattice cannot be excluded as one of the possible causes of progressive red-shift observed with an increasing temperature on the ZnO nanostructures in this case.

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

65

Fig. 3. UV–vis absorption spectra of ZnO nanostructures for as-prepared and annealed at 200, 600, and 900 ◦ C.

Fig. 5. PL emission spectra ZnO flower-like nanostructures prior and after annealing at 200, 600, and 900 ◦ C after excitation at 325 nm.

To study the vibrational modes of the ZnO nanostructures, Raman scattering measurements were performed at room temperature. According to the group theory predictions, the Raman-active centre modes of ZnO are A1 + 2E2 + E1 , where A1 and E1 are polar phonons and, hence, exhibit different frequencies for the transverse optical (TO) and longitudinal-optical (LO) phonons. Non-polar phonon modes with symmetry E2 have two frequencies: E2 (high) associated with oxygen atoms and E2 (low) associated with Zn sublattice [29]. Fig. 4 depicts the Raman-scattering spectra of the ZnO nanostructures prior and after annealing at different temperatures. For all spectra, the sharp, strong and dominant peak located at about 438 cm−1 believed to be a characteristic scattering peak of the Raman-active dominant E2 (high) mode of wurtzite hexagonal ZnO was observed [30,31]. In addition, two peaks at 330 and 370 cm−1 which can be assigned to E2H –E2L (multi-phonon) A1 (TO) modes were also observed, respectively [22,32]. A broadened and weak peak at 580 cm−1 also appeared in all spectra. Normally, in bulk ZnO, the frequency 574 cm−1 of 1LO phonon peak corresponds to A1 (LO) phonon and it can only be observed in the configuration when the c-axis of wurtzite ZnO is parallel to the sample face [30]. However, if the c-axis is perpendicular to the sample face, the E1 (LO) (591 cm−1 ) phonon is observed, instead [29]. Based on the theory of polar optical phonons in wurtzite nanocrystals [32],

the frequency of 1LO phonon mode in ZnO nanoparticles should be between 574 and 591 cm−1 . Therefore the peak around 580 cm−1 in this study could be due to the superimposition of A1 (LO) and E1 (LO) [33] which is caused by defects in ZnO [30,33–35]. Furthermore, the intensity of this phonon peak was shown to increase with an increasing annealing temperature, implying that the concentration of the defects increases with the annealing temperature increase. It is also important to note that the E2 (high) mode became stronger and sharper with the increase of annealing temperature demonstrating crystallinity improvement upon high temperature annealing which correspond to the XRD results. The structural and morphological changes have a considerable influence on PL analysis. In order to verify this, the PL analyses on both as-prepared and annealed samples were conducted. Fig. 5 illustrates the room temperature PL spectra of the ZnO nanostructured before and after annealing at different temperatures as well as the PL emission peaks by Gaussian fitting. The PL spectra of ZnO nanostructures (Fig. 5) show resolved emission peaks at 384, 402, 430 and 455 nm. It can be seen from this spectra that the Gaussian curves fitted the PL curves perfectly. The wellknown broad green emission in ZnO [14–18] was not observed in this study. Fan et al. [36] reported the single violet emission from ZnO centred at 413–424 nm without any accompanying deeplevel emission and UV-emission. The PL emission peak located at around 384 nm (3.23 eV) corresponds to the excitonic recombination [11,18,19,22,37], while the intense violet emission band around 402 nm (3.08 eV) can be ascribed to the electron transition from the bottom of the conduction band to the VZn level. This transition is positioned approximately 3.06 eV below the conduction band edge. The assignment of this emission peak agrees well with the results of Kale et al. [14] for ZnO flower-like nanorods. The dependence of the most intense emission peak at 402 nm on annealing temperatures is plotted in Fig. 6. The PL intensity decreases with an increase in annealing temperature reaching a minimum at 900 ◦ C. In addition, the shape and position of the emission peaks in the near violet-blue region did not shift with an increasing annealing temperature. The variation in the PL intensity indicated that annealing induces a variation in concentration of the native defects such as VZn , Zni that act as luminescence centres in ZnO. It is well known that the PL properties strongly rely on the type and density of defects in the ZnO nanostructures. However, some of the defects can also be trapped and act as recombination centres for charge carriers. If the density of defects does not reach a limit or not high enough, the defects will contribute mainly to the PL emission resulting to the increase of the PL emission

Fig. 4. Raman spectra of ZnO nanostructures prior and after annealing at different temperatures: 200, 600, and 900 ◦ C.

66

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

PL intensity [a.u.]

4000000

3200000

2400000

1600000

800000

0

200

400

600

800

1000

Annealing temperature [ oC] Fig. 6. Variation of PL intensity of the violet/blue emission peak with calcination temperature.

ZnO nanostructures obtained before performing high resolution scan for Zn and O. As shown in the figure, only photoelectron peaks of the main elements Zn, O, Auger Zn MMN, O KLL and C elements could be detected for both the annealed and un-annealed samples. The detected carbon may be due to the carbon tape used during the measurements and the carbon adsorbed on the surface during the exposure of the sample to the ambient atmosphere. As shown in Fig. 9, the high resolution XPS spectra of the as prepared, 200, 600, and 900 ◦ C annealed ZnO nanostructures in the Zn 2p region, revealed two photoelectron peaks identified as Zn 2p1/2 and Zn 2p3/2 with binding energies of 1045.1 and 1021.4 eV, respectively. The change of oxygen on the surface of ZnO nanostructures prior and after annealing at 200, 600, and 900 ◦ C was well understood by investigating the high resolution XPS spectra of ZnO nanostructures in the O 1s region and its de-convolution result is depicted in Fig. 10. The asymmetric O 1s peak on the surface was coherently fitted by three nearly Gaussian components, centred at 530.06 eV (O1 ), 531.13 eV (O2 ) and 532.15 eV (O3 ), for as prepared as well as the annealed samples, respectively. The three fitted binding energy peaks supplemented the results of Chen et al. [42], Hsieh et al. [43], Li et al. [44], and Wang et al. [45]. A shift in peak position for the O3 peak to higher binding energies (533.15 eV) when the temperature was increased to 600 ◦ C was observed. The O1 component located on the low bingeing energy side of the O 1s spectrum

400

Zn 3s

600

200

Zn 3d O 2s Zn 3p

800

C1 s

1000

900 oC

Zn LMN

1200

O 1s Zn LMN Zn LMN

Zn 2p 1/2

1400

Zn LMN

Zn 2p 3/2 Zn 2p 1/2

As-prep. 200 oC 600 oC

Zn 2s

Intensity[a.u.]

intensity with an increasing density of defects. Whereas, if the density of defects is high enough, the defects will mainly behave as the recombination centres, thus leading to the PL intensity decrease with the increase of defect density. It has also been previously reported that the highly non-equilibrium processes produce a high concentrate of Zni defect states with different charges [25,27]. Heat treating at low temperature provides enough ionization energy and increases the concentration of Zni thus strengthening the blue emissions whereas high temperature heat treatment induces the outward diffusion of Zni resulting in quenching of the blue emissions. In the current study, the lessening of the violet emission intensity upon increasing heat treatment temperature can be associated with the decrease in the concentration or density of defects including VZn and/or Zni acting as luminescent centres in ZnO which then resulted in the reduction of the number of radiative recombination centres. Based on the calculations of the energy levels of various defect emission centres conducted by Xu et al. [38] and Lin et al. [39] using the full-potential linear muffin-tin orbital (FP-LMTO) method, the energy interval from the donor level of Zni to the top of the valence band is about 2.9 eV, which is most close to the value of 2.88 eV for the blue emission in this study. In addition, the Zni level is located at 0.22 eV slightly below the conduction band as previously determined by Bylander et al. [40] experimentally. This level can trap photo-excited electrons, followed by the radiative recombination with holes in the valence band, leading to blue emission. Therefore the broad shoulder emission peak at 455 nm (2.7 eV) can be attributed to the electronic transition from the donor level Zni to the acceptor level VZn which is about 2.6 eV. This result compares well to the results of Wei et al. [40] for ZnO films. In this study, a proposed mechanism for the observed PL emission is demonstrated on an energy band diagram in Fig. 7. When the incident photon energy is enough to excite the electrons up to a sub-band of the conduction band, the electrons first relax non-radiatively to the Zni state and then transit to the valence band by inducing effective transitions from the Zni energy level to the valence band and other deep levels which will then results into several emissions in the violet-blue according to the energy differences. However, since the intensity of the violet emission peak at 402 nm was shown to be more intense compared with the rest of the emission peaks; this indicates that more electrons at the conduction band level recombine with holes at the VZn level. For detailed information about surface chemical compositions of different ZnO nanostructures, the XPS analyses were conducted. Fig. 8 shows the XPS survey spectra of the as prepared and annealed

Fig. 7. Schematic diagram of energy band diagram proposed for visible emissions of ZnO flower-like nanostructure.

0

Binding Energy [eV] Fig. 8. XPS survey spectra of the as prepared ZnO nanostructures and that annealed at different temperatures.

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

Intensity[a.u.]

Zn 2p Zn 2p

3/2

1/2

As-prep. 200 o C 600 o C 900 o C 1050

1040

1030

1020

1010

Binding Energy [eV] Fig. 9. XPS spectra of Zn 2p peaks for as prepared, 200, 600, and 900 ◦ C ZnO nanostructures.

can be attributed to the O2− ions on the wurtzite structure of the hexagonal Zn2+ ion array, which are surrounded by zinc atoms with the full supplement of nearest-neighbour O2− ions [43–46]. The O2 component in the medium binding energy can be associated with the O2− ions in the oxygen deficient regions within the ZnO matrix [42,43,46]. Changes in the intensity of this component may therefore be in connection with the variation in the concentration of oxygen defects such as VO or/and Oi . In addition, it is also important to mention that the intensity of the O2 peak always appears higher than that of the O1 component located at the low binding energy and this indicates the large oxygen-deficient state of the surface layer. The O3 component in the higher binding energy at 532.15 eV could be due to chemisorbed or dissociated oxygen or OH species on the surface of the ZnO nanostructures including CO3 , adsorbed H2 O or adsorbed O2 [41,44]. The variation before and after annealing was understood better by calculating the integrated intensity ratios of O1 /O2 and the O1 /Zn and (O1 + O2 )/Zn as shown in Table 1. It can clearly be seen that the O1 + O2 /Zn ratio was found to be higher than 1.0 except for the sample annealed at 600 ◦ C whose (O1 + O2 )/Zn ratio was less than 1.0. Therefore, ZnO nanostructures annealed at 600 ◦ C are

As-prep.

o

200 C

O1

O1

O2

O2

O3

O3

Intensity [a.u.]

Intensity [a.u.] 536

534

67

532

530

528

526

536

534

532

Binding Energy [eV]

530

528

526

Binding Energy [eV] o

600 C

o

900 C

O1

O1

O2 Intensity [a.u.] 536

534

532

530

Binding Energy [eV]

528

O2

Intensity [a.u.]

O3

526

O3

536

534

532

530

528

526

Binding Energy [eV]

Fig. 10. Oxygen 1s spectra of ZnO nanostructures prior and after annealing at different temperatures (200, 600, and 900 ◦ C) obtained before Ar+ ion sputtering.

68

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

Table 1 XPS data of O and atom ratio of Zn to oxygen on the surface of the Zn nanostructures annealed at different temperature. Annealing temperature (◦ C)

O1 position (eV)

O2 position (eV)

O3 position (eV)

Intensity of O1 (%)

Intensity of O2 (%)

Intensity of O3 (%)

O1 /O2

O1 /Zn

O1 + O2 /Zn

As-prep. 200 600 900

530.06 530.26 530.41 530.26

531.13 531.47 531.69 531.61

532.15 532.39 533.19 532.26

68.31 69.94 3.64 76.87

17.02 13.50 59.91 17.60

14.67 16.55 36.45 5.53

4.01 5.18 0.06 4.36

1.03 1.05 0.05 1.16

1.29 1.26 0.96 1.43

oxygen-deficient while the rest of the samples contain excess oxygen. A similar behaviour was noticed for the O1 /Zn ratio for ZnO nanostructures annealed at 600 ◦ C indicating that most of the oxygen ions are not in the stoichiometric Zn O bonding. This corresponds to the drastic drop in the intensity of the O1 peak which was observed after annealing at 600 ◦ C. The strong Zn O bonding was observed for un-annealed samples and those annealed at 200 and 900 ◦ C as the intensity of their O1 component was shown to exceed that of the O2 and O3 components. On the other hand, the intensity of the O2 peak showed an increase with increasing the annealing temperature up to 600 ◦ C and a significant drop with further increase in annealing temperature up to 900 ◦ C. The increase in the intensity of this component with an increase in annealing temperature up to 600 ◦ C can be associated with the increase in the concentration of defects on the surface of ZnO and this is believed to be due to thermal diffusion which normally occurs at high temperatures. The drop in the intensity of this peak at 900 ◦ C may be related to elimination of some defects on the surface of ZnO nanostructures. For the O3 component, it was noticed that the intensity continuously increases when increasing temperature up to 600 ◦ C and thereafter drastically decrease when increasing the annealing temperature up to 900 ◦ C. This clearly indicates that the chemical adsorption of the surface of the ZnO nanostructures is higher for the samples annealed at 600 ◦ C as compared to those annealed at 200 and 900 ◦ C. According to the XPS results there was excess oxygen present in all the ZnO nanostructures except for the 600 ◦ C annealed sample. However, the variation of the (O1 + O2 )/Zn ratio with increasing temperature suggests the involvement of other defects related to Zn such as VZn and Zni as confirmed by PL analysis. In general, it is believed that the intensity ratio of UV- to visible emission is considered as an indication of excellent crystal quality and low defect concentration. It is known that higher annealing temperatures reduce the trap states in a material, thus improving the crystallinity. In the current study, the crystallinity of the ZnO nanostructures increases with an increase in annealing temperature up to 900 ◦ C. This suggests that the crystal quality of the as-prepared or samples annealed at 200 ◦ C or 600 ◦ C is relatively poor compared to those annealed at 900 ◦ C. Therefore it is believed that annealing temperatures lower than 900 ◦ C induces more defects than those annealed at 900 ◦ C. A possible explanation for this may be that the non-radiative defect concentration is lower for samples annealed below 900 ◦ C hence the samples annealed below 900 ◦ C exhibited higher PL and UV–vis relative intensities as compared to that annealed at 900 ◦ C. A similar trend is also observed on the EPR signal (see Fig. 11) as it reduces with increasing annealing temperature. Also significant change in microstructural features evolved at different temperatures and this has led to variation both in the optical absorption and luminescent characteristics of the samples. In order to further understand the presence of paramagnetic defects in ZnO nanostructures, the ESR measurements were performed at room temperature. The comparison between the ESR responses from ZnO nanostructures prior and after annealing at 200, 600, and 900 ◦ C were conducted. Fig. 11 shows the first order differential ESR spectra of the ZnO nanostructures before and after annealing at different temperatures. The low-field microwave

absorption (LFMA) process which usually originates from the magnetization processes far from the saturation state also supported by bulk magnetization is observed for the as-prepared ZnO nanostructures. The LFMA signal is opposite in phase with respect to that of paramagnetic resonance (PMR). This indicates that the LFMA has microwave absorption minimum at zero fields in contrast to the microwave absorption maximum for the PMR line at its resonance field. LFMA is understood to be connected to the magnetization processes that occur at applied low field. It is therefore assumed that magnetization processes is present at HDC = 0 and such magnetization process is attributed to the interaction of magnetic moments of the ZnO with the magnetic-field component of electromagnetic radiation. It is interesting to note that the LFMA spectra are only significant at room temperature (298 K) and it disappears completely with annealing temperature. The disappearance of the LFMA signal in the ZnO nanostructure above room temperature is in good

Fig. 11. ESR spectra of the as prepared and annealed (200, 600, and 900 ◦ C) ZnO nanostructures.

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

69

Table 2 Lande ‘g’ factor, line width (H) and number of spins (Nspins ) of ZnO nanostructures. Annealing temperature (◦ C)

FMR field (mT)

g-Factor

HEPR (mT)

No. of spins, Nspins (×104 )

Crystallite size (nm)

As-prep. 200 600 900

327.67 328.12 328.46 328.45

1.955 1.945 1.956 1.954

3.41 4.12 3.11 2.89

3.26 2.28 2.32 0.27

26.4 28.9 27.8 38.8

10

10

Intensity (a.u.)

agreement with arguments made by Alvarez et al. [47] which states that it is indicative of disappearance of long range order and is strongly associated with magnetization processes of the magnetic state. Moreover, the violet-blue emitting ZnO samples before and after annealing at 200, 600, and 900 ◦ C exhibited the ESR signals with the following g-factors: 1.955, 1.945, 1.956, and 1.954 which may be associated with complex defects such as VZn and/or Zni , respectively. EPR signals close to those obtained in this study have been previously reported. Li et al. [44], and Morazzoni et al. [48] assigned the ESR signal at g = 1.955 to Zni for green luminescence in ZnO structures. Also, Kursunska et al. [49] assigned this signal to complex defects including Zni for green emitting ZnO single crystals. The VZn and VZn -Zni complex have also been reported for g-factor values around and close to 2.0 [50,51]. Generally, in absence of any foreign material in ZnO (undoped), the magnetism is believed to arise from the change in net spins in the d-orbits of Zn (3d9 ) and this can only be achieved by native defects such as Zni , VZn , and VO . The Zni , are known to completely diffuse at low temperatures (<500 ◦ C) after heating the sample in air. However, based on the PL results, the broad violet-blue emission peak did not shift or vanish even after heating above 500 ◦ C. Instead, its intensity decreased with increasing annealing temperature confirming the involvement of other defects such as VZn and other Zn complexes acting as luminescence centres other than Zni . Correspondingly, the ESR signal did not vanish with increasing annealing temperature. However, the ESR signal intensity gradually decreased with the increase of annealing temperatures as clearly shown in Fig. 11(b). This behaviour can be directly related to the decrease in the concentration of Zni and/or VZn defects in ZnO nanostructures as confirmed by PL analysis. In the current study, it is expected that the paramagnetism (PM) arises from an intrinsic exchange interaction of magnetic moments in undoped ZnO. Although the exact mechanism of intrinsic ferromagnetism in undoped oxides is still under debate, defects have greatly been suggested to play a vital role in the magnetic origin in the un-doped ZnO system as mentioned earlier. The combination of the present defects analyses based on the PL and the ESR measurements provides a good opportunity to clarify the physical nature of the local magnetic moment. Therefore, the Zni and/or VZn defects are responsible for the room temperature PM of the samples. Based on the PL and EPR measurements, we propose that the Zni and/or VZn defects are the primary source of the observed PM in these samples. Furthermore, it is observed from Table 2 that the linewidth (peak-to-peak width) with the annealing temperature decrease which could be due to an exchange narrowing of the ESR signal. It is evident that as the annealing temperature increases, the number of spins decreases due to a decrease in absorption as shown in the UV–vis analysis, Fig. 3. As shown in Fig. 12, spin population by ESR and the intensity of the violet-blue emission peak show a negative linear correlation with annealing temperature. It is however interesting to note that calcination temperature has a negative linear correlation with (1) ZnO stoichiometry via oxygen-to-zinc atomic ration measured from XPS, (2) wavelength of absorption peak in nm via UV–vis–IR spectrophotometry and (3) crystallite size from XRD. The fact that the spin population and violet-blue emission PL intensity both increase with increasing annealing temperature, is an indication that the population of spins in ZnO is linked to the intensity of

spin population crystallite size (nm) O/Zn ratio absorption edge (nm) violet-blue PL intensity

R = -0.85

9

10

R = -0.85

5

10

R = +0.98 R = +0.82

2

10

R = +0.02 -1

10

0

200

400

600

800

1000

Annealing Temperature ( oC) Fig. 12. A plot showing how annealing temperatures of the ZnO nanostructures affect spin population, crystallite size, oxygen-to-zinc ratio, wavelength of absorbed photons and the intensity of the violet-blue emission peak.

R = +0.97

6

10

O/Zn ratio violet-blue PL intensity (arb. units)

R = -0.42 0

10

0

10

1x10

10

2x10

10

3x10

Spin population/mol Fig. 13. A plot of spin population against violet-blue PL intensity and the oxygento-zinc ratio.

the violet-blue emission. The plot of spin population versus the intensity of the violet-blue emission is shown in Fig. 13. In this plot one also sees that the oxygen-to-zinc atomic ratio decreases as spin population increases. This may suggest that most of the electron spins are contributed by Zn rather than O. 4. Conclusion ZnO nanostructures were successfully synthesized by a microwave-assisted hydrothermal method and annealed at different temperatures. Structural and elemental analysis revealed that

70

G.H. Mhlongo et al. / Applied Surface Science 293 (2014) 62–70

the ZnO nanostructures are polycrystalline in nature with no impurity phases. Morphological variation upon increasing annealing temperature from well-defined “flower-like” structure composed of ZnO multi-nanorods to randomly oriented “worm-like” ZnO nanostructures was also observed. The violet-blue emission from ZnO nanostructured associated with VZn and Zni defects was shown to decrease with increasing annealing temperature indicating reduction of the number of radiative recombination centres caused by decrease of concentration of VZn and/or Zni defects acting as luminescent centres in ZnO. The appearance of the phonon peak at 580 cm−1 from Raman also confirmed the presence of defects in ZnO nanostructures. The ESR analysis showed ESR signals with the g-factors of 1.955, 1.945, 1.956, and 1.954 which may be associated with VZn and/or Zni defects, for as prepared, 200, 600, and 900 ◦ C ZnO samples, respectively. The decrease in the oxygen-to-zinc atomic ratio measured from XPS with increasing spin population suggests that most of the electron spins are contributed by Zn rather than O. Based on ESR and PL analyses, it was concluded that VZn and Zni defects are the main cause of the room temperature PM of the ZnO nanostructures. Acknowledgement This project is financially supported by the Department of Science and Technology of South Africa and the Council for Scientific and Industrial Research of South Africa (Project numbers HGER28P and HGER27S). References [1] N.H. Alvi, K. ul Hasan, O. Nur, M. Willander, Nanoscale Res. Lett. 6 (2011) 130. [2] N.F. Hamedani, A.R. Mahjoub, A.A. Khodadadi, Y. Mortazavi, Sens. Actuators B 156 (2011) 737. [3] S.D. Shinde, G.E. Patil, D.D. Kajale, V.B. Gaikwad, G.H. Jain, J. Alloy Compd. 528 (2012) 109. [4] D.E. Motaung, G.H. Mhlongo, I. Kortidisd, S.S. Nkosi, G.F. Malgas, B.W. Mwakikunga, S. Sinha Ray, G. Kiriakidis, Appl. Surf. Sci. 279 (2013) 142. [5] S.H. Ko, D. Lee, H.W. Kang, K.H. Nam, J.Y. Yeo, S.J. Hong, C.P. Grigoropoulos, H.J. Sung, Nano Lett. 11 (2011) 666. [6] D.E. Motaung, G.F. Malgas, C.J. Arendse, S.E. Mavundla, Mater. Chem. Phys. 135 (2012) 401. [7] G. Zhang, X. Shen, Y. Yang, J. Phys. Chem. C 115 (2011) 7145. [8] P. Yang, H. Yan, S. Mao, R. Russo, J. Johnson, R. Saykally, N. Moris, J. Pham, R. He, H.J. Choi, Adv. Funct. Mater. 12 (2002) 2. [9] M.H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, P. Yang, Science 292 (2001) 1897. [10] B. Xiang, P. Wang, X. Zhang, S.A. Dayeh, D.P.R. Aplin, C. Soci, D. Yu, D. Wang, Nano Lett. 7 (2007) 2. [11] Y. Wu, K.V. Rao, W. Voit, T. Tamaki, O.D. Jayakumar, L. Belova, Y.S. Liu, P.A. Glans, C.L. Chang, J.H. Guo, IEEE Trans. Magn. 46 (2010) 6. [12] M.D. McCluskey, S.J. Jokela, J. Appl. Phys. 106 (2009) 071101. [13] L. Schmide-Mende, J.L. MacManus-Driscoll, Mater. Today 10 (2007) 40.

[14] R.B. Kale, Y.J. Hsu, Y.F. Lin, S.Y. Lu, Solid State Commun. 42 (2007) 302. [15] K.H. Tam, C.K. Cheung, Y.H. Leung, A.B. Djurisi, C.C. Ling, C.D. Beling, S. Fung, W.M. Kwok, W.K. Chan, D.L. Phillips, L. Ding, W.K. Ge, J. Phys. Chem. B 110 (2006) 20865. [16] G.H. Mhlongo, O.M. Ntwaeaborwa, H.C. Swart, R.E. Kroon, P. Solarz, W. RybaRomanowski, K.T. Hillie, J. Phys. Chem. C 115 (2011) 17625. [17] A. Djurisi, W.C.H. Chay, V.A.L. Roy, Y.H. Leung, C.Y. Kwong, K.W. Cheah, T.K.G. Rao, W.K. Chan, H.F. Lui, Adv. Funct. Mater. 14 (2004) 9. [18] L.H. Quanga, S.J. Chua, K.P. Loh, E. Fitzgerald, J. Cryst. Growth 287 (2006) 157. [19] C. Jing-wei, X. Jian-ping, Z. Xiao-song, N.I.U. Xi-ping, X. Tong-yan, J.I. Ting, L.I. Lan, Optoelectron. Lett. 8 (2012). [20] Y. Yang, H. Yan, Z. Fu, B. Yang, L. Xia, Y. Xu, J. Zuo, F. Li, Solid State Commun. 138 (2006) 521. [21] R.B.M. Cross, M.M. De Souza, E.M.S. Narayanan, Nanotechnology 16 (2005) 2188. [22] L. Dai, X.L. Chen, W.J. Wang, T. Zhou, B.Q. Hu, J. Phys.: Condens. Matter 15 (2003) 2221. [23] A.B. Djurisi, Y.H. Leung, K.H. Tam, Y.F. Hsu, L. Ding, W.K. Ge, Y.C. Zhong, K.S. Wong, W.K. Chan, H.L. Tam, K.W. Cheah, W.M. Kwok, D.L. Phillips, Nanotechnology 18 (2007) 095702. [24] Y. Zhang, Y. Liu, L. Wu, H. Li, L. Han, B. Wang, E. Xie, Appl. Surf. Sci. 255 (2009) 4801. [25] H. Zeng, G. Duan, Y. Li, S. Yang, X. Xu, W. Cai, Adv. Funct. Mater. 20 (2010) 561. [26] X. Zhang, Y. Xia, T. He, Mater. Chem. Phys. 137 (2012) 622. [27] T.K. Kundu, N. Karak, P. Barik, S. Saha, Int. J. Soft Comput. Eng. 1 (2011) 2231. [28] Y. Gupta, A. Mansingh, J. Appl. Phys. 80 (1996) 1063. [29] N. Ashkenov, B.N. Mbenkum, C. Bundesmann, V. Riede, M. Lorenz, D. Spemann, E.M. Kaidashev, A. Kasic, M. Schubert, M. Grundmann, J. Appl. Phys. 93 (2003) 126. [30] C.J. Youn, T.S. Jeong, M.S. Han, J.H. Kim, J. Cryst. Growth 261 (2004) 526. [31] K.A. Alim, V.A. Fonoberov, M. Shamsa, A.A. Balandin, J. Appl. Phys. 97 (2005) 124313. [32] V.A. Fonoberov, A.A. Balandin, Phys. Rev. B 70 (2004) 195410. [33] J. Yang, X. Liu, L. Yang, Y. Wang, Y. Zhang, J. Lang, M. Gao, B. Feng, J. Alloy Compd. 477 (2009) 632. [34] Y. Zhao, Y. Jiang, Y. Fang, J. Cryst. Growth 307 (2007) 278. [35] J.N. Zeng, J.K. Low, Z.M. Ren, T. Liew, Y.F. Lu, Appl. Surf. Sci. 197 (2002) 362. [36] X.M. Fan, J.S. Lian, L. Zhao, Y.H. Liu, Appl. Surf. Sci. 252 (2005) 420. [37] N.S. Ridhuan, K.A. Razak, Z. Lockman, A.A. Aziz, PLoS ONE 7 (2012) e50405. [38] P.S. Xu, Y.M. Sun, C.S. Shi, F.Q. Xu, H.B. Pan, Nucl. Instr. Meth. Phys. Res. B 199 (2003) 286. [39] B. Lin, Z. Fu, Y. Jia, Appl. Phys. Lett. 79 (2001) 943. [40] E.G. Bylander, Appl. Phys. Lett. 49 (1978) 3. [41] X.Q. Wei, B.Y. Man, M. Liu, C.S. Xue, H.Z. Zhuang, C. Yang, Physica B 388 (2007) 145. [42] M. Chen, X. Wang, Y.H. Yu, Z.L. Pei, X.D. Bai, C. Sun, R.F. Huang, L.S. Wen, Appl. Surf. Sci. 158 (2000) 134. [43] P.T. Hsieh, Y.C. Chen, K.S. Kao, C.M. Wang, Appl. Phys. A 90 (2008) 317. [44] L. Li, L. Fang, X. Ju Zhou, Z.Y. Liu, L. Zhao, S. Jiang, J. Electron Spectrosc. Relat. Phenom. 173 (2009) 7. [45] H. Wang, S. Dong, X. Zhou, X. Hu, Y. Chang, Physica E 44 (2011) 307. [46] S. Wei, J. Lian, H. Wu, Mater. Characterization 61 (2010) 1239. [47] G. Alvarez, H. Montiel, J.F. Barron, M.P. Gutierrez, R. Zamorano, J. Mag. Magn. Mater. 322 (2010) 348. [48] F. Morazzoni, R. Scotti, P. Di Nola, C. Milani, D. Narducci, J. Chem. Soc. Faraday Trans. 88 (1992) 1691. [49] N.O. Korsunska, L.V. Borkovska, B.M. Bulakh, L.Y. Khomenkova, V.I. Kushnirenko, I.V. Markevich, J. Lumin. 102–103 (2003) 733. [50] M.G. Kakazey, M. Vlasova, M. Dominguez-Patin, G. Dominguez-Patin, G. Gonzalez-Rodriguez, B. Salazar-Hernandez, J. Appl. Phys. 92 (2002) 5566. [51] D. Galland, A. Herve, Solid State Comm. 14 (1974) 953.