1. Introduction

ZnO nanostructures have attracted considerable attention in the past decade owing to their unique optical and electrical properties [1–5]. In particular, the direct wide band gap (~3.40eV) and large excitonic binding energy (~60 meV) as well as relative large refraction index make ZnO nanostructures a good candidate for high efficient ultraviolet (UV) light emitting devices and solar cells [1–5]. Unfortunately, the ZnO nanostructures, particularly in those samples fabricated by hydrothermal synthesis, typically exhibit a near-band-edge UV emission from free exciton recombination and a broad deep-level emission associated with intrinsic or extrinsic defects [6,7]. As of now, the origin of the deep-level emission in ZnO nanorods still remains controversial. However, the surface structure is believed to play a dominant role in the deep-level emission due to the large surface-to-volume ratio of ZnO nanorods. For an uncoated ZnO nanorod or nanowire, the surface defects such as oxygen vacancies would trap free electrons after adsorbing gas molecules, resulting in an upward band bending near the surface. The band bending causes the separation of photo-generated electron-hole pairs. Thus, as a result of the upward band bending, the excess holes at the surface would tunnel into the deep-level defect centers inside the ZnO nanorods, and cause a deep-level defect emission [8,9]. In fact, the unpassivated surface sites may also act as either a nonradiative recombination and/or deep-level defect sites, deteriorating the UV emission. Therefore, engineering of ZnO nanostructures through surface modification has become an important issue for developing efficient ZnO-based devices with suppressed defect-related emission. Some recent experimental results indicate that fabrication of core/shell heterostructures is an efficient method for improving the luminescence properties of ZnO nanostructures [9–14]. However, in most cases the defect-related visible emission can only be suppressed to some extent [9–13].

In the present paper, luminescent ZnO-core/a-SiN_x:H-shell nanorod arrays were synthesized by plasma enhanced chemical vapor deposition (PECVD) of a-SiN_x:H on ZnO nanorod surfaces. A remarkable improvement of photoluminescence (PL) from ZnO-core/a-SiN_x:H-shell nanorod arrays were realized by modulating the bandgap of a-SiN_x:H shell. The a-SiN_x:H shell with a large bandgap is found to significantly enhance ZnO nanorod UV emission by more than 8 times compared with the uncoated nanorods. Moreover, the deep-level defect emission can be almost completely suppressed for all the core–shell structures, which is independent of the bandgaps of a-SiN_x:H shells. The improved PL is mainly attributed to the H and N passivation as well as the screening effect from the a-SiN_x:H shell.

2. Experimental

The vertically aligned ZnO-core/a-SiN_x:H-shell nanorod arrays were synthesized in two steps. Uniform ZnO nanorod arrays were firstly prepared by hydrothermal synthesis using an equimolar (0.01 M) aqueous solution of zinc nitrate hexahydrate (Zn (NO₃)₂·6H₂O) and hexamethylenetetramine (C₆H₁₂N₄) as reported elsewhere [15]. Then, a high frequency (40.68MHz) PECVD technique was used for producing a-SiN_x:H coatings directly on the surfaces of ZnO nanorod arrays. The a-SiN_x:H were prepared by using a gas mixture of SiH₄, NH₃ and H₂ as precursors. The flow rate of SiH₄ was kept at 5 sccm, while the flow rate of NH₃ was varied from 5 to 20 sccm so as to control the optical bandgap of a-SiN_x:H ranging from 2.4 to 3.6 eV. The growth process was conducted at a relatively low temperature of 250 °C. The structure of the nanorods was characterized by a Philips XL30 scanning electron microscope (SEM) and a JEOL high-resolution transmission electron microscope (HRTEM, 2010F). The photoluminescence (PL) measurements were performed at room temperature using a near-field scanning optical microscope (NSOM, Nanonics Cryoview2000) equipped with a 325 nm He-Cd laser. A Fourier transform infrared (FTIR) spectroscope was also employed to record the bonding configurations of the a-SiN_x:H samples.

3. Results and discussion

Figure 1(a) shows a SEM image of the ZnO-core/a-SiN_x:H-shell nanorods fabricated by hydrothermal growth followed by an a-SiN_x:H deposition at 250 °C for 4 min. The core-shell nanorod arrays are highly aligned and structurally uniform in wafer-scale. Figure 1(b) shows the typical TEM image of an individual core/shell nanostructure in which an a-SiN_x:H shell (~6 nm thick) uniformly covers the surface of the ZnO rod. This indicates that a homogeneous a-SiN_x:H shell was grown on the nanorod. HRTEM image shown in Fig. 1(d) reveals that the core-shell interface is atomically abrupt and that the shell is amorphous. The EDS elemental mapping analysis for Zn, O, Si and N, shown in Fig. 1(c), further unambiguously confirms the uniform core-shell structure consisting of ZnO core and a-SiN_x:H shell.

 

Fig. 1 (a) SEM image of aligned ZnO-core/a-SiN_x:H-shell nanorods fabricated by hydrothermal growth followed by an a-SiN_x:H deposition duration of 4 min at 250 °C. (b) TEM image of an individual ZnO-core/a-SiN_x:H-shell nanorod. (c) The EDS elemental mappings for Zn, O and Si. (d) The corresponding HRTEM image taken at the interface.

Download Full Size | PDF

Figure 2 shows the PL spectra taken from the as-synthesized ZnO nanorods and ZnO-core/a-SiN_x:H-shell nanorods, respectively. The optical bandgaps of a-SiN_x:H shell layers are also present in the insets of Fig. 2, which are calculated according to the Tauc equation (αhν)^1/2 = Β (hν−Ε_opt), where α is the absorption coefficient and B is a constant [16]. For the as-grown ZnO nanorods, one can see that the PL spectrum is composed of two typical emission peaks, the sharp UV emission near 380 nm and the broad orange-red emission centered at 600 nm. The broad orange-red emission is considered to be associated with oxygen defects and vacancies [7]. Compared to the as-grown ZnO nanorods, it is found that the UV emission can be significantly enhanced after the ZnO nanorods were coated with a-SiN_x:H layers as shown in Fig. 2(b). This indicates that the UV emission efficiency from the as-grown ZnO nanorods can be effectively increased by a-SiN_x:H decoration. One can see that the enhancement factor of UV emission strongly depends on the optical bandgap (E_opt) of the a-SiN_x:H shells. For the a-SiN_x:H shell with the E_opt of 3.6 eV, the UV emission intensity can be greatly increased by more than 8 times its prior level. With the E_opt of a-SiN_x:H shell decreasing down to 2.4 eV, the UV emission intensity is rapidly reduced and less than its prior level. Since the E_opt (2.4-2.9 eV) of a-SiN_x:H is much smaller than that (3.4 eV) of ZnO, it is believed that the high optical absorption coefficient from a-SiN_x:H shells would seriously deteriorate the UV emission from ZnO nanorods. From Fig. 2, it is interesting to find that the broad orange-red emission (deep-level defect emission) is totally eliminated for all the core/shell structures, which is independent of the E_opt of a-SiN_x:H shells at all. Because the E_opt of a-SiN_x:H shells in our case are larger than the peak energy (2.0 eV) of deep-level defect emission, it can exclude the possibility of the completely illuminated deep-level defect emission resulting from the absorption of the a-SiN_x:H shells. It seems that well surface passivation by a-SiN_x:H shells are responsible for the complete elimination of the deep-level defect emission from the ZnO nanorods. This is different from that observed in the cases of ZnO coated with wide band gap materials or polymers where the defect-related visible emission can only be suppressed to some extent [9–13].

 

Fig. 2 The PL spectra obtained from the ZnO nanorods with and without a-SiN_x:H shells plotted in a log scale: as-synthesized nanorods (a), nanorods coated with different bandgaps of a-SiN_x:H shells (b) 3.6 eV, (c) 2.9 eV, and (d) 2.4 eV. Insets show the tauc plots of the corresponding a-SiN_x:H shells.

Download Full Size | PDF

To analyze the origin of the improved characteristics in the PL spectra, FTIR was employed to examine the local bonding configurations of the a-SiN_x:H shells. Figure 3 shows the FTIR absorption spectra and congruously displays the following absorption bands [17]: the intense band at ~850 cm⁻¹ with a shoulder at ~1170 cm⁻¹, the ~2140 cm⁻¹ band and the ~3350 cm⁻¹ band. These features are characteristics of the hydrogenated a-SiN_x and can be assigned to the Si-N stretching, N-H rocking, Si-H stretching, and N-H stretching mode, respectively. For all the a-SiN_x:H shells, the hydrogen atom concentration estimated according to the number of N-H and Si-H bonds as well as the nitrogen atom concentration evaluated from the number of Si-N and N-H bonds are higher than 1 × 10²² cm⁻³ as shown in Fig. 3. This evidently demonstrates that the coating shells contain a large number of H and N atoms. Therefore, it is considered that the H and N atoms located at the interface between ZnO core and a-SiN_x:H shells would passivate the surface defect states of the ZnO nanorods. In our case, the a-SiN_x:H shells were prepared from the precursors of SiH₄, NH₃ and H₂ by using VHF-PECVD. In the growth process, the depletion of the precursors through electron impact would produce a large amount of H and N radicals flux to the growth surface, which undoubtedly passivates the surface sites that act as either a nonradiative recombination and/or deep-level defect site, causing an remarkable enhancement of UV emission and reduction of the deep-level defect emission as revealed in Fig. 2.

 

Fig. 3 FTIR spectra of the a-SiN_x:H films with different bandgaps, respectively. The inset presents the N and H concentration as a function of the bandgaps of a-SiN_x:H films, which were estimated according to the number of N-H, Si-H as well as Si-N bonds. The number of bonds is determined by integrating the different stretching absorption band using the equation A∫(a(ω)/ω)dω, where a(ω) and ω are the absorption coefficient and the wave number, respectively. A is taken as equal to 6.3 × 10¹⁸, 9.2 × 10¹⁹, and 2.8 × 10²⁰ cm⁻² for the Si-N, Si-H, and N-H stretching bands, respectively [17].

Download Full Size | PDF

To verify the role of H and N atoms on the improved PL characteristics, the ZnO nanorods were also treated by NH₃ plasma. The PL spectra from the NH₃-plasma treated ZnO nanorods were present in the inset of Fig. 4. One can see that the UV emission intensity from the NH₃-plasma ZnO nanorods is increased by about 8 times as compared with as-grown ZnO nanorods. Moreover, the deep-level defect emission is significantly suppressed. These phenomena are the same as that from ZnO nanorods coated with a-SiN_x:H shells, strongly indicating that the improved PL characteristics of the core-shell structure are resulted from the H as well as N passivation [18,19]. We also noted that a weak deep-level defect emission can still be seen in the NH₃-plasma ZnO nanorods. This is different from that observed in ZnO-core/a-SiN_x:H-shell structure where the deep-level defect emission can be totally eliminated as displayed in Fig. 2. These results demonstrate that a-SiN_x:H decorating on ZnO nanorods can more effectively improve the PL characteristics of ZnO nanorods compared with the NH₃ plasma treatment. It is due to the fact that the a-SiN_x:H shell grown on the ZnO core surface can also serve as a dielectric medium and screen the charge carriers located in surface, thus lowing the surface band bending of ZnO core [8,9]. As a consequence, it decreases the probability of the photo-generated holes located at the surface tunneling into the deep-level defect centers inside the ZnO core, and leads to a stronger overlap of the wave functions of electrons and holes in the ZnO cores, which not only promotes the enhancement of the UV emission but also significantly reduces the deep-level defect emission. To further gain more insight on the improved PL characteristics, the NH₃-plasma ZnO nanorods and the ZnO-core/a-SiN_x:H-shell structure were annealed at 400 °C for 30 min in a N₂ atmosphere, respectively, as is shown in Fig. 4. It is found that the UV emission would seriously deteriorate and the deep-level defect emission would recover after the NH₃-plasma ZnO nanorods were annealed at 400 °C. These behaviors are sharply different from that observed in the annealing ZnO-core/a-SiN_x:H-shell structure where strong UV emission with negligible deep-level defect emission can be obtained. It is reported that in the case of NH₃-plasma ZnO nanorods the H and N radicals would desorb from defect states at the ZnO nanorod surface under high annealing temperatures [20]. In contrast, in our case of ZnO-core/a-SiN_x:H-shell nanorods, the H and N atoms located at the surface defect sites of ZnO nanorods would further react with other radicals such as (Si, H, or N) to form a-SiN_x:H. For the a-SiN_x:H, the Si-H bonds as well as the N-H bonds are stable even at an annealing temperature of 400 °C [17]. Therefore, it is very reasonable that PL emission from ZnO-core/a-SiN_x:H-shell nanorods can almost remain unchanged at the annealing temperature of 400 °C. For the appearance of the negligible defect emission after annealing at 400 °C, we consider that it may come from the reconstruction of the interface structure between the ZnO-core and a-SiN_x:H-shell. Further studies are needed in order to obtain a better understanding of this phenomenon.

 

Fig. 4 The PL spectra of ZnO nanorods coated by a large bandgap (3.6 eV) of a-SiN_x:H shell after annealing at 400 °C for 30 min in a N₂ atmosphere. Inset shows the PL spectra of the NH₃-plasma ZnO nanorods before and after annealing at 400 °C for 30 min in a N₂ atmosphere, respectively.

Download Full Size | PDF

4. Conclusion

In summary, luminescent ZnO-core/a-SiN_x:H-shell nanorod arrays were synthesized by direct PECVD coating of a-SiN_x:H on ZnO nanorod surfaces. A remarkable improvement of PL from ZnO-core/a-SiN_x:H-shell nanorod arrays were realized by modulating the bandgap of a-SiN_x:H shell. The a-SiN_x:H shell with a large bandgap is found to significantly enhance ZnO nanorod UV emission by more than 8 times compared with the uncoated ZnO nanorods. Moreover, the deep-level defect emission can be almost completely suppressed for all the core–shell nanostructures, which is independent of the bandgaps of a-SiN_x:H shells. The improved PL is ascribed to the H and N passivation as well as the screening effect from a-SiN_x:H shell. Our results suggest that ZnO nanorods decorated by a-SiN_x:H are promising material for fabricating stable and efficient UV-only emitting devices.