In this article, the ALD technique was employed to deposit ZnO seed layers on various substrates. The substrates included B-doped ZnO (BZO) films, indium tin oxides (ITO) films, single crystal (111) silicon, and glass sheets. Diethylzinc [DEZ, Zn(C2H5)2] and deionized water were used as the precursors for ZnO deposition. Pure N2 gas (99.999%) was used to carry and purge gas. The reaction is carried out as follows:
The reaction chamber was pumped down to 1 to 2
Torr before deposition. The operating environment of ZnO deposition was maintained at 3
Torr and 200°C. Each deposition cycle consisted of four steps, which included DEZ reactant, N2
O reactant, and N2
purge. The typical pulse time for introducing DEZ and H2
O precursors was 0.5
s, and the N2
purge time was 10
s. The deposition cycles of 11, 22, 33, 44, 55, 110, and 275 were chosen to produce ZnO seed layers with the various thickness of 2, 4, 6, 8, 10, 20, and 50
nm. The deposition rate at the above conditions approaches 0.182
The subsequent hydrothermal growth was carried out at 90°C in a sealed kettle by immersing the deposited substrates in aqueous solution (80
mL) containing zinc nitrate (Zn(NO3
mol/L) and hexamethylenetetramine (HMTA; C6
mol/L). ZnO NWs were fabricated according to the following reactions.
Finally, the samples were washed with deionized water and dried in air before characterization. The morphology of the NWs was characterized by scanning electron microscopy (SEM, Philips FEIXL30 SFEG, Amsterdam, Netherlands) and transmission electron microscopy (TEM, Hitachi HF-2000, Chiyoda, Tokyo, Japan). TEM samples were prepared by gently dragging the holey (400 mesh Cu, SPI supplies, West Chester, PA, USA) carbon grids along the surface of the samples. X-ray diffraction (XRD) analysis was performed with a Rigaku Dmax-2000 diffractometer using CuKa radiation (Rigaku Corporation, Tokyo, Japan). The morphology of the seed layers and roughness was characterized by an atomic force microscope (AFM, Park Systems XE-100, Santa Clara, California, USA). The photoluminescence (PL) spectroscopy is performed on an Olympus BX51 microscope with Hg illumination and UV filter cube (U-MWU2, excitation, Olympus Shinjuku, Tokyo, Japan).
b shows the typical hexagonal cylinder shaped ZnO NWs grown on Si and glass substrates with the 10-nm-thick ZnO seed layers. The morphology is different from that of the NWs grown on the BZO (Figure
d) and ITO substrates (Figure
f) although the thickness of the seed layers is the same of 10
d shows the inclined NWs grown on the BZO substrate that have short and thick geometry morphologies. The diameters of the NWs range from several tens of nanometers to hundreds of nanometers. It is noteworthy that the diameters of the NWs are consisted with the size of the grains on the surface of BZO film (shown in Figure
c). Analogously, the nonuniform rough ITO surface with the several tens of nanometers grain size (shown in Figure
e) produced anomalous NWs with the average diameter of about 200
nm. Different from the Si and glass substrates, the BZO and ITO films have obvious grain boundaries on the film surfaces. Grains can be the site of nucleate for the growth of NWs and ZnO seed layer with a 10
nm thickness is too thin to shield the morphologies of BZO and ITO films. So the morphologies of BZO and ITO films have a great influence on NWs, which results in the NWs having similar geometric morphology with the substrate surface. So we get the conclusion that the NWs prepared by hydrothermal reaction were influenced greatly by surface topography of substrates when substrates are covered with ultra-thin seed layers.
AFM three-dimenional images. AFM three-dimensional images of (a) Si substrates, (c) BZO substrates, and (e) ITO substrates SEM images of NWs grown on Si substrates (b), on BZO substrates and (d), on ITO substrates (f).
To learn more about these ZnO NWs, TEM was used to characterize the ZnO NW structures. Figure
shows high-resolution TEM images taken from ZnO NWs grown on various substrates. The insert figures show the corresponding low-resolution TEM images and selected-area electron diffraction patterns, which indicates that the ZnO nanorods are single-crystalline in structure. The HRTEM images of ZnO nanorods grown on various substrates reveal clear lattice spacing of 0.52 or 0.25
nm correspond to the inter-planar spacing of the wurtzite ZnO (001) or (002) face, which indicate that the ZnO nanorod growth occurs preferentially along the  direction. So the crystal structures of NWs prepared by hydrothermal reaction were not influenced by surface topography of substrates.
Figure 2 The high- and low-resolution images and the selected area diffraction patters. High-resolution TEM images, low-resolution TEM images, and the selected area electron diffraction patterns (see inset) of NWs grown on Si substrate (a), on BZO substrate and (more ...)
In order to understand the relationship between the thin seed layer and the NWs, a more systematic structural investigation was carried out. Figure
shows SEM photographs of NWs grown on glass substrates pre-coated with ZnO seed layers. The ZnO seed layers with thickness from 2 to 50
nm were deposited by ALD method. It can be clearly found that ZnO nanoparticles grow out of the 2-nm-thick seed layer, whereas NWs grow out of the seed layers whose thickness is above 4
a). Moreover, the NWs on 6- and 8-nm-thick seed layers have the best alignment, with an average rod diameter of 100
c,d). However, the relatively sparse and poorly aligned NWs are obtained on the seed layers with the thickness greater than 10
nm, and their orientation gets worse and worse with the increase of the seed layers' thickness (Figure
e,f,g). For example, the NWs on the 20-nm-thick seed layer are more disordered than those on the seed layers with the thickness of 6 and 8
nm. When the seed layer's thickness reaches 50
nm, almost all the NWs are slanting as shown in Figure
g. Therefore, the threshold thickness of the seed layers for the conversion between the well-aligned and poorly-aligned NWs is 8
a, the crystal structure of the NWs was examined. All the diffraction peaks can be indexed to the wurtzite structure of ZnO (36–1451). The peak intensity ratio of (101) to (002) according to the different thickness of the seed layers is shown in Figure
b. As discussed above, the NWs grown on the seed layer with the thickness of 6 or 8
nm exhibit a strong peak intensity ratio, indicating good orientation of the NWs. Consequently, the XRD results are consistent with SEM results in Figure
. It should be noted that the peak intensity ratio of the nanoparticles grown on the 2-nm-thick seed layer exhibits the third largest value, which indicates that the nanoparticles prefer the growth along the c-axis direction even if it is a failure to generate NWs due to very thin seed layer.
Figure 4 The XRD spectra and ZNO peak ratios. (a) XRD spectra of NWs grown on pre-coated substrates with different thickness (from top to bottom, 2, 4, 6, 8, 10, 20, and 50nm). (b) ZnO peak ratios for (1 0 1) to (0 0 2) as a function of substrate thickness. (more ...)
We suggest that one of the important reasons for the alignment variation according to different thickness of the seed layers is the ZnO seed roughness, which is also reported by previous research
]. The images of the ZnO seed films with different thickness deposited on glass substrates were characterized by AFM. As shown in Figure
, their roughness increases from 0.479 to 1.37
nm with their thickness (from 2 to 50
The dependence of roughness on seed thickness.
The reason why roughness affects orientation of NWs has been hypothesized and proved. It is well known that roughness represents fluctuation amplitude and frequency of substrate or film surface, which plays an important role in nucleation and growth of NWs
]. However, the fluctuation amplitude and frequency of roughness also determine the orientation. Figure
shows three-dimensional images of the seed layers with the thickness of 6 and 50
nm. As shown in the edges indicated by the circle in Figure
, it could be found that compared with that on 6-nm-thick seed layer, the fluctuation amplitude and frequency of the roughness for the 50-nm-thick seed layer are larger and smaller than those for the 6-nm-thick seed layer, respectively, which may be caused by stack of the ZnO nanoparticles. The augment of spacing of local peaks weakens interaction among ZnO nanorods, which leads to free growth and slant of some nanorods. This relationship is shown schematically in Figure
Figure 6 AFM three-dimensional images of (a) 6nm seed layers and (b) 50nm seed layers.
The schematic model. Schematic model for the effect of fluctuate amplitude and frequency of seed layer roughness on the alignment of NWs.
Another convincing evidence that the fluctuation amplitude and frequency of roughness affect orientation of NWs is shown in Figure
a, c gives AFM photos of 6-nm-thick seed layers before and after annealing. Comparing Figure
b,d, it can be found that the alignment of NWs obtained on the annealed seed layer becomes poor. Although annealing usually can improve the crystallinity of the seeds, the peak spacing of the seed layers increases after annealing, resulting in poor alignment of the NWs. This result shows that the fluctuation amplitude and frequency of roughness determine the orientation.
AFM and SEM images. (a)AFM image of 6-nm-thick seed layer. (b)SEM image of NWs grown on 6-nm-thick seed layer. (c) AFM image of 6-nm-thick seed layer after annealing. (d)SEM image of NWs grown on annealed 6-nm-thick seed layer.
The texture of ZnO seed layers was also reported to be another factor which affects the ZnO NWs' orientation
]. However, in the present paper, it is found that the texture of ZnO seed layers does not affect the alignment. The XRD data for the seed layers with different thickness are shown in Figure
. The ZnO seed layers with thickness under 10
nm do not show any reflection peak due to ultra-thin thickness. On the other hand, 10-, 20-, and 50-nm-thick seed layers appear the same diffraction peaks, indicating that the seed layers deposited at the same condition have the same texture. So, we suggest that the seed layers with different small thickness exhibit almost the same texture and do not have the major change with increase of thickness. Given the analysis above, we suggest that the texture of the ZnO seeds does not directly determine the ZnO NWs orientation in our experiments.
Figure 9 XRD spectra of ZnO seed layers. The XRD spectra of ZnO seed layers with different thickness (from top to bottom, 2, 4, 6, 8,10, 20, and 50nm).
PL spectroscopy is an effective technique for evaluating the optical properties and defects of semiconductor materials. Figure
shows typical room-temperature PL spectra of the ZnO NWs grown on glass substrates with different seed thickness. The PL spectra from all samples exhibit the same profile with a dominant emission peak centered at 383
nm, which corresponds to the ultraviolet emission of ZnO with a band gap of 3.24
]. In addition to the UV emission, two weak emissions at 450 and 468
nm also can be observed for the as-grown samples. The weak peaks in the blue-green band result from an electronic transition from the level of the ionized oxygen vacancies to the valence band
]. We can see clearly that no obvious change of PL spectroscopy is occurred as the increase of seed thickness, which means that there is no relation between crystal defects and seed thickness.
Figure 10 The PL micrographs of ZnO NWs. PL micrographs of ZnO NWs grown on pre-coated glass substrate with different thickness (from top to bottom, 2, 4, 6, 8,10, 20, and 50nm).