Characterization of Crystalline Zinc Oxide in the Form of Hexagonal

Dec 7, 2009 - ZnO particles with a rod-like shape and with a regular bipod structure ... prismatic defects obtained in the hexagonal ZnO bipod particl...
0 downloads 0 Views 3MB Size
DOI: 10.1021/cg901193g

Characterization of Crystalline Zinc Oxide in the Form of Hexagonal Bipods

2010, Vol. 10 830–837

Marko Bitenc,† Goran Drazic,‡ and Zorica Crnjak Orel*,† †

National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia and Department for Nanostructured Materials, Jo zef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia



Received September 29, 2009; Revised Manuscript Received November 13, 2009

ABSTRACT: We present a solution-phase preparation method for the synthesis of ZnO in an open reactor (with constant stirring) or in a closed reactor (autoclave). The method is based on a combination of the polyol method and the homogeneous precipitation of zinc nitrate with urea. ZnO particles with a rod-like shape and with a regular bipod structure were prepared in an open reactor, while needle-like shapes were prepared in the autoclave. A detailed characterization of the planar base-plane and prismatic defects obtained in the hexagonal ZnO bipod particles with electron microscopy revealed that in the middle of the bipod the planar defects contain a small amount of silicon. A rough estimation indicated that there was a single atomic layer of silicon at the defect. From convergent-beam electron-diffraction experiments, we found that these planar defects were inversion domain boundaries (IDBs) with a head-to-head orientation of the polar axis. In the needle-like particles, besides IDBs on the basal planes, prismatic IDBs with a small amount of silicon were also found. The formation mechanism for the prismatic IDB was thought to be the overgrowth of one part of the bipod over another. In this case, not just the head-to-head arrangement of the polar axis, but also the tail-to-tail arrangement, must be able to occur.

Introduction The solution-phase preparation of nano- to submicrometer particles represents a low-cost preparation technique that has recently attracted a lot of interest due to its low energy consumption, since the temperatures needed for the synthesis are low (85-95 °C). In addition, it is much more favorable for the large-scale synthesis of ZnO nanoparticles with a controllable morphology. In the literature, different methods have been presented for the preparation of ZnO nanostructures. Zinc oxide nanocrystals were prepared by a homogeneous precipitation method using urea and zinc nitrate as the raw materials, and the orientation adhesion of the nanocrystallites formed by the connection of crystallites was discussed using the growth-unit model of anionic coordination polyhedrons. It was shown that the growth of zinc oxide nanocrystals was more likely to occur along the c-axis.1 Furthermore, the microwave irradiation of solutions of Zn(NO3)2 and urea provides a straightforward route to producing crystals with a needle-like and microjavelins morphology.2,3 It was found that due to supersaturation of Zn2þ and Zn(OH)þ species under the moderately basic pH condition induced by microwave, initial growth started through the oxygen terminated (0001) facet. These facets were identified as the key steps responsible for the formation of microjavelins.2 Not only does the microwave procedure yield ZnO microparticles much more rapidly than conventional heating methods, it also produces ZnO with a quite different, welldefined needle-like morphology based on an unusual and unexpected a-axis.3 The large-scale synthesis of uniformly sized, hexagonal, pyramid-shaped ZnO nanocrystals can be performed by the thermolysis of a Zn-oleate complex, which was prepared from *Corresponding author: Zorica Crnjak Orel National Institute of Chemistry Slovenia Hajdrihova 11 SI-1001 Ljubljana. Phone: þ386 (0)1 476 02 36. Fax: þ386 (0)1 476 03 00. E-mail: [email protected]. Web: www.ki.si. pubs.acs.org/crystal

Published on Web 12/07/2009

the reaction of inexpensive and environmentally friendly reagents such as zinc chloride and sodium oleate.4 ZnO microspheres and hexagonal microrods with sheet-like and plate-like nanostructures were prepared by the hydrothermal synthesis approach by using trisodium citrate, which plays a key role in directing the formation of these microstructures. By increasing the reaction time, these microspheres gradually dissolved to form short hexagonal microrods with a stacked nanoplate or nanosheet structure.5 Disk-like, flower-like, and nanorod flower-like ZnO nanostructures have been controllably fabricated by a citric-acid-assisted hydrothermal process,6,7 while different shapes of ZnO microcrystals have been controllably prepared by a capping-molecule-assisted hydrothermal process. Furthermore, flower-like, disk-like, and dumbbell-like ZnO microcrystals of the hexagonal phase have been obtained using ammonia, citric acid, and poly(vinyl alcohol) as the capping molecules.8 Apart from the different morphologies in which ZnO can be synthesized, as well as the influence of the additives or process parameters on the final size and shape of ZnO particles, it has many times been reported that the ZnO has a twinned structure. An unusual feature of the prepared ZnO samples, visible as a crack perpendicular to the length, in the middle of the particle, and sometimes also along the length, with rare occurrences of a well-separated splitting, was reported by Oliviera et al.9 Wang et al.10 reported the linkage of ZnO46tetrahedral units in a weak basic solution that leads to the formation of the crystal nucleus of a twinning crystal. The mechanism of the formation of that type of crystal was based on an already existing nucleus in solution, and the growth of crystallite twins takes place along the polar c-axis by the incorporation of growth units on the growth interfaces. Hu et al.11 reported the formation of linked ZnO rods by sonochemical and microwave-assisted solution-phase routes. In the article, they presented preparation methods without templates, seeds, or surfactants for the large-scale production r 2009 American Chemical Society

Article

Crystal Growth & Design, Vol. 10, No. 2, 2010

of linked ZnO rods with various morphologies. They found that ZnO rods grow along the [0001] direction and the formation of linked ZnO rods can be controlled by an oriented attachment mechanism. The second-harmonic (SH) far-field scattering patterns from single-twinned and twin-free ZnO rods were presented in a paper by Liu et al.12 They successfully correlated the distributions of the interference fringes in the pattern with the twinning structure in the rods. They also determined that at the scattering angle of the fringe it was dark, but for the twinned rod it was bright because of the destructive or constructive interference, respectively. Moreover, they presented the possibility of identifying the polarities in more complicated ZnO nanostructures using these SH scattering patterns. In this paper, the solution-phase preparation method for the synthesis of ZnO is presented, based on a combination of the polyol method13-17 and a homogeneous precipitation with urea.18-20 In our previous articles,18,19 we presented the influence of different types of added polyols on the morphology and size of the prepared ZnO particles. The samples were prepared with the method of homogeneous precipitation by using zinc nitrate and urea in a mixture of water/polyol with constant stirring at a low temperature of 90 °C. Here we report on the fabrication of hexagonal ZnO with a regular twinning structure by a facile-template-free method in a mixture of water (W) and ethylene glycol (EG). The structure and morphology were controlled by changing the ratio of W/EG, the concentration of the starting chemicals, and the reaction conditions. Experimental Methods In a typical experimental procedure, zinc nitrate hexahydrate (Zn(NO3)2 3 6H2O, Sigma-Aldrich, 98%) and urea (Sigma-Aldrich, 99%) were dissolved in Milli-Q water, as a fresh stock solution to avoid hydrolysis upon storage. The chemicals were analytical reagents and used without any further purification. Two types of sample preparation were performed. The first sets of experiments (marked as Samples A1, A2, and A3) were carried out in the 40-mL Teflon-lined autoclave, which was placed in an oven preheated at 90 °C. The experiments were performed in a mixture of water (W) and ethylene glycol (EG, Merck) with concentrations of Zn2þ (0.001 M) and urea (0.005 M) after 72 h (Sample A1: volume ratio W/EG = 1/3 and Sample A2: W/EG = 1/1). After 24 h the A3 sample was prepared at concentrations of Zn2þ (0.01 M) and urea (0.05 M) with a volume ratio W/EG = 1/3. The as-prepared solids were cooled to room temperature, filtered off, washed with water, and dried in air. The second sets of experiments (marked as Sample B) were carried out under constant stirring (200 rpm) in 250-mL open reactors at 90 °C for 0.5 h. The concentrations of Zn2þ ions and urea were 0.01 and 0.05 M, respectively. A reaction mixture of W/EG with the volume ratio 1/1 was used. The as-prepared solids were filtered off, washed with water, and dried in air.

Characterization The structures of the samples were studied using the Jeol 2010 FEG STEM and Jeol 2100 TEM microscopes operating at 200 kV, equipped with EDX detectors for the compositional analyses. In the case of small particles (up to 100 nm), a fraction of the particles in the form of a suspension in ethyl alcohol was transferred to a lacy, carbon-coated Ni or Cu grid and examined in the microscope. In the case of larger particles (in the micrometer range), the crystals were crushed in a mortar and the resulting powder was subsequently embedded in phenol-formaldehyde resin. A total of 3-5 wt % of phenol-formaldehyde powder was added to the ZnO powder

831

and thoroughly mixed. The mixture was placed in a die with a suitable diameter, pressed with a pressure of 0.5 MPa, and cured at 150 °C for a few minutes. The resulting pellet was mechanically very strong, so subsequent grinding, dimpling, and ion erosion were used without any problems. Because of the large quantity of powder particles, no liquid-nitrogen cooling during the ion-erosion process was necessary and no charging effects were observed in the transmission electron microscope. Highresolution TEM (HRTEM), selected-area electron diffraction (SAED), and convergent-beam electron diffraction (CBED) were employed to examine the structure of the materials. The simulated CBED patterns, used for the polar-axis orientation, were calculated with the EMS program code.21 The morphologies of the samples were characterized by scanning field-emission electron microscopy (FE-SEM, Zeiss Supra 35 VP with an EDS analyzer). The X-ray diffraction analyses (XRD) were carried out on a Siemens D-500 X-ray diffractometer. The IR spectra were obtained on an FT-IR spectrometer (Perkin-Elmer 2000) in the spectral range between 4000 and 400 cm-1 with a spectral resolution of 2 cm-1 in the transmittance mode. The KBr pellet technique was used for the sample preparation. Results and Discussion During the preparation of the ZnO several parameters were identified as important factors in controlling the morphology, size, and homogeneity of the ZnO structures. We found that varying the reaction conditions, that is, the closed reactor (autoclave) or the open reactor (steering), while keeping the temperature and the solvent (the ratio between W and EG) concentration constant, had an influence on the final size and shape of the prepared ZnO particles.22 By changing only one of the specified parameters in the open reactor, we got a series of ZnO nanostructures, the morphologies of which were presented in our previous papers.18,19 Similar morphologies were obtained in the closed reactor. In some cases, besides needle-like particles, hollow rod-like hexagonal ZnO bipods, flower-like hydrozincite, and mixtures of these two were observed. The details of reaction conditions (Table 1) and FE-SEM micrographs (Figures 1-4) of these morphologies are presented as Supporting Information. In this experiment, no additional chemicals were added and always the relation between urea and zinc nitrate was constant. The aging time, reaction media (volume ratio of water/EG), and concentration were changed. Similar homogeneous precipitation method of preparation of ZnO was proposed by Tsuchida and Kitajima,23 where different shapes of ZnO were obtained. By using ZnSO4 in the presence of urea, they obtained hexagonal prismatic morphology, while by using Zn(NO3)2, ZnCl2, and Zn(CH3COO)2 spindleshaped morphology were obtained. Figure 1 shows low-magnification FE-SEM images of the products prepared at 90 °C at two different concentrations of Zn2þ and urea and different volume ratios of W/EG. Sample A1 was prepared after 72 h with a concentration of Zn2þ (0.001 M) and urea (0.005 M) at a volume ratio W/EG = 1/3 on the basis of the standard conditions of preparation (Figure 1a). If the volume ratio W/EG was 1/1 (Sample A2), much bigger, needle-like particles were synthesized (Figure 1b). At a 10 times higher concentration, at a volume ratio W/EG = 1/3 and at a shorter reaction time (24 h), needle-like particles (Sample A3) were also synthesized, but they were not so homogeneous, as presented in Figure 1c.

832

Crystal Growth & Design, Vol. 10, No. 2, 2010

Bitenc et al.

By altering the reaction conditions (concentration, time, and volume ratio of W/EG mixture), we established that the most homogeneous needle-like shapes of the ZnO nanostructures were prepared when the concentrations of the Zn2þ ions and the urea were 0.001 and 0.005 M, respectively, in a W/EG 1/3 mixture (Figure 1a). The samples were characterized by Fourier-transform infrared spectroscopy (FT-IR) and X-ray powder diffraction (XRD) in order to confirm the presence of some amorphous phase and their crystal structures, respectively. The FT-IR spectrum of Sample A1, prepared in the autoclave, is presented in Figure 2a. All the samples (the other spectra are not presented in this paper) confirmed the formation of pure ZnO, since typical Zn-O stretching bands in the 600-400 cm-1 range are present in the IR spectrum. The band at 3423 cm-1 is attributed to the stretching of the hydroxyl group (OH) of water molecules adsorbed on the surface of nanostructures. The XRD patterns of all the samples (A1, A2, A3, and B) match the wurtzite structure (space group P63mc) of the pure ZnO reported in the JCPDS card No. 00-036-1415. As an example, the XRD pattern of Sample B, prepared in the open reactor, is presented in Figure 2b. A detailed inspection of the FE-SEM picture, presented in Figure 1a, of the prepared ZnO needle-like particles shows that the crystals were a few micrometers in size and with a high

Figure 1. FE-SEM micrographs of samples prepared in an autoclave: (a) Sample A1, (b) Sample A2, and (c) Sample A3.

Figure 3. FE-SEM image of Sample B prepared in the open reactor.

Figure 2. (a) FTIR spectra of Sample A1 prepared in the autoclave and (b) XRD spectra of Sample B prepared in the open reactor.

Article

Crystal Growth & Design, Vol. 10, No. 2, 2010

833

Figure 4. Sample B prepared in the open reactor: (a) TEM image (bright field) of ZnO particles, (b) SAED of indicated particle, oriented in the [1120] zone axis.

Figure 5. (a) HRTEM image of Sample B prepared in open reactor; the middle part of the particle with a planar defect. (b) EDXS spectrum of the bulk particle (spt 1) of Sample B prepared in the open reactor. (c) EDXS spectrum of the planar defect (spt 2) of Sample B prepared in the open reactor.

aspect ratio (15-20 and higher) and that they were composed of two needle-like, hexagonal parts divided by a planar defect, so we can describe the crystals as hexagonal, needle-like bipods. The SEM and TEM examinations of Sample B, prepared in an open reactor under constant stirring (200 rpm) after 0.5 h (Zn2þ 0.01 M and urea 0.05 M), are presented in Figures 3 and 4. The obtained samples were in the form of elongated hexagonal prisms, with an aspect ratio of 3.2. At the middle of every particle, planar defects were observed (Figure 3). The prismatic facets were well developed, compared to the pyramidal, and both parts of the bipods were, in most cases, symmetrical. The preliminary TEM analyses revealed that the particles were monocrystalline, hexagonal ZnO (wurtzite). From the SAED pattern, we can conclude that both parts of the particles are crystallographically identical. The particles are

elongated in the [0001] direction and the planar defects (the boundary between both parts of the bipods) are parallel to these basal planes (Figure 4a,b). To determine the character and the origin of the planar defects, we performed an EDXS analysis across the defect using a 5-nm electron-beam diameter. For comparison, an analysis was also made in the vicinity of the planar defect. Figure 5 shows a HRTEM image of an edge of the particle, and an intersecting planar defect is shown with labeled areas where the EDXS analyses were made. In Figures 5b,c, EDXS spectra collected from the labeled areas are shown. We found that all the planar defects in the basal planes contained a certain amount of silicon. A rough estimation of the Si concentration, taking into account the beam diameter and the d value of the (0001) planes, shows that there is around one atomic layer of silicon.

834

Crystal Growth & Design, Vol. 10, No. 2, 2010

Bitenc et al.

Figure 7. TEM image of Sample B prepared in the open reactor: (a, b) amorphous particles (S) containing a large amount of silicon and (c) EDXS spectrum of this phase.

Figure 6. Sample B prepared in the open reactor: (a) HRTEM image across the IDB with labeled areas where the CBED patterns were recorded, (b, c) experimental CBED patterns at the indicated areas, (d) calculated CBED pattern where the polar axis (head) is indicated with a triangle (simulated using a 50-nm-thick foil).

On the basis of the literature data for ZnO-based ceramics doped with various valent ions (Sb, Sn, In, Ti, etc.),24-26 we assumed that those planar defects are inversion domain boundaries (IDBs). To confirm this assumption and to find how the domains are oriented (head-to-head or tail-to-tail), we performed convergent-beam electron-diffraction (CBED) experiments,25 presented in Figure 6. Two CBED patterns were collected from the opposite sides of the IDB and compared with a calculated pattern. On the basis of the distribution of the intensities of the different diffraction discs, we could estimate the polar axis. From this, we found that the polar axis was in a “head-to-head” orientation, meaning that the [0001] directions of both parts of the particle (pods) are pointing to the IDB. This means that the bipods were terminated with two (0001) planes. To determine the possible origin of the silicon, a chemical analysis of the starting materials was performed. Using inductively coupled plasma atomic emission spectroscopy (ICP-AES), we determined this to be