Characterization of p-type ZnO films

Recently, several thin-film growers have demonstrated that nitrogen doping was an effective method to realize p-type ZnO films because N has the smallest ionization energy in group V-dopants and similar radii to that of oxygen atoms. In this session, we used nitrogen-implanted and Si3N4 buffer layer to fabricate p-type ZnO thin films.

Figure 4.6 shows the XRD pattern of nitrogen-implanted ZnO thin films on Si and Si3N4/Si, and then annealed at 850^oC in N2 atmosphere. Only a sharp diffraction peak of (002) at 2θ = 34.5˚ can be detected for all the ZnO films, indicating that these ZnO films were highly c-axis oriented. However, as increasing the nitrogen-implanted concentration form 5×10¹² to 5×10¹⁵ cm^-2, the intensity of the (002) XRD peak in ZnO films grown on Si substrate decreases as shown in inset of Figure 4.6. In contrast, when the ZnO films were grown on Si3N4/Si substrates, it was found that (002) diffraction peak becomes stronger with the increase of the implanted nitrogen ionsfrom 5×10¹² to 1×10¹⁴ cm^-2 (Figure 4.6). This reveals that the occupation of the implanted nitrogen ionson oxygen vacancies can improve the crystallinity of ZnO films on Si3N4/Si can be improved. However, above that, the peak intensity was rapidly decreased, implying that more implanted nitrogen ions probably induces extra defects and this would lead to the lattice distortion.

Figure 4.6: XRD of ZnO films sputtered at Si (inset), and Si3N4/Si substrates with or without nitrogen-implanted various doses.

The electrical conduction type of nitrogen-implanted ZnO films as a function of doping doses is shown in Figure 4.7. The nitrogen-implanted ZnO films on Si substrate show n-type conduction independent of the implanted nitrogen ions concentration. In sharp contrast, the nitrogen-implanted ZnO films on Si3N4/Si substrate exhibit p-type conduction and the carrier concentration increases up to 7.3×10¹⁷ cm^-3 with an increase of nitrogen-implanted concentration from 5×10¹² to 1×10¹⁴ cm^-2. According to our previous study, as the ZnO filmswere deposited on the Si3N4/Si structure, x-ray photoelectron spectroscopy analysis demonstrates that a lower oxygen vacancies concentration and thinner interface layer was detected for ZnO on Si3N4/Si structure compared to that on Si [107]. Furthermore, the film stoichiometry was improved because of the reduction in oxygen vacancies, indicating the ratio of Zn:O was decreased [108].In this condition, the concentration of active acceptors may exceed the donor concentration so that the conduction type was changed from n to p type. The dependence of conduction type on the implanted nitrogen ions concentration in ZnO films in Figure 4.7 suggests the nitrogen-implanted process can product more holelike carries to transform original conduction (compensate the native carriers). However, a further increase in the implanted dose of the nitrogen ionsup to 5×10¹⁵ cm^-2 leads to a decrease in hole concentration of p-ZnO and the conduction type of ZnO films would approach the intrinsic conductor.

Resistivity and Hall mobility as a function of implanted nitrogen ions doses were measured and are shown in Figure 4.8 for p-type conduction ZnO films. As increasing nitrogen ions doses from 5×10¹² to 1×10¹⁴ cm^-2, both hole concentration and Hall mobility increase, but the resistivity decreases. The p-type ZnO films grown on Si3N4/Si show a hole concentration of 7.3×10¹⁷ cm^-3, a mobility of 6.02 cm²/Vs, and a low resisitivity of 10.3 Ω cm. Above that (1×10¹⁴ cm^-2), both hole

concentration and Hall mobility decreases, but an increase in resistivity was observed.

The initial increase in the hole concentration is due to a decrease in oxygen vacancy as the implanted nitrogen ions doses increase. The decrease in hole concentration after the maximum value is caused by the formation of more defects due to excess nitrogen ions that can compensate for a hole carrier which may correspond to the degradation of the crystal quality in p-type ZnO films as supported by the decrease in the peak intensity of (002) XRD peak in Figure 4.6 for the nitrogen-implanted ZnO film with the dose of 5×10¹⁵ cm^-2.

Figure 4.7: Electrical conduction type of nitrogen-implanted ZnO films as a function of various doses.

Figure 4.8: Variation of resistivity and Hall mobility as a function of different nitrogen-implanted doses.

Recently, many research groups announced that p-type ZnO were fabricated by doping group-V elements and proposed a reasonable explanation for the mechanism of atom substitution in crystal structure, such as NO. However, almost no evidence was presented for the behavior of atom occupation in p-type ZnO fine structure.

Therefore, in order to further understand the local structure of p-type ZnO films, the extended x-ray absorption fine structure was investigated. Figure 4.9 shows the pseudoradial distribution functions obtained from the k³-weighted Fourier transforms at Zn K edge for the nonimplanted and nitrogen-implanted ZnO films on a Si3N4/Si structure, where pure ZnO powder was used as standards for comparison. The first peak in the Fourier transforms corresponds to the nearest-neighbor distance around zinc atoms. The peak position of zinc to oxygen for both bulk powder standards (ZnO:

99.999%) and nonimplanted ZnO thin films is equal to 0.1945 nm. For the ZnO film implanted with 1×10¹⁴ cm^-2 N⁺ dose, the Zn-O bond length in ZnO/Si3N4/Si is very similar to that of nonimplanted ZnO films and bulk powder standards. However, for the second nearest-neighbor distance around Zn²⁺, it was found that the Zn-Zn bond length for nonimplanted ZnO films is shorter than that of nitrogen-implanted ZnO films. It could be due to the existence of oxygen vacancies in the ZnO lattice matrix that causes the secondary nearest-neighbor distance to become shorter. In comparison with bulk powder standards (0.327 nm), a little larger Zn-Zn bond length (3.28Å) was also detected for N⁺ -implanted ZnO films. These observations suggest that the implantation of nitrogen ionsinto ZnO films could affect the local structure of ZnO film, such as bond distance, but a limited range of 1×10¹⁴ cm^-2 nitrogen ions dose can effectively improve the crystal quality of ZnO films.

Figure 4.10 illustrates the room-temperature PL spectrum of ZnO films implanted with various nitrogen ions doses on Si3N4/Si substrates. UV emission with peaks at 3.31, 3.28, 3.29, and 3.30 eV is dominantly observed for the ZnO films

implanted with 0, 5

×

10¹², 1

×

10¹⁴, and 5

×

10¹⁵ cm^-2 nitrogen ions dose, respectively.

According to the spectrum, it is noticed that the peak intensity of the UV emission depends markedly on the nitrogen-implanted dose. The ZnO film implanted with a 1×10¹⁴ cm^-2 nitrogen ions dose not only shows a stronger peak intensity but also has a narrower full width at half maximum of 95 meV than that (110 meV) of nonimplanted ZnO films. Furthermore, the relative intensity ratio of UV emission (IUV) to that of deep-level emission (IDLE) are measured about 13(nonimplanted sample), 55-56 (implanted with 5

×

10¹²-1

×

10¹⁴ cm^-2 nitrogen ionsdose), and 25(implanted with 5

×

10¹⁵ cm^-2 nitrogen ionsdose), at room temperature, respectively. This high ratio implies that the nitrogen-implanted ZnO film is of high optical quality.

Figure 4.9: Fourier transforms at Zn K-edge for the ZnO powder standards, non-implanted, and nitrogen-implanted ZnO/Si3N4/Si samples.

4.4 Summary

Phosphorus-implanted in ZnO films tends to react with zinc and oxygen elements to form the clusters and this would induce defects as revealed from PL spectra.

the formation of phosphide compounds. Therefore, in this condition, high resistive but not p-type ZnO film is obtained by phosphorus doping.

However, in another study, we have fabricated the reproducible p-type ZnO films grown on Si3N4/Si by rf magnetron sputtering, implanted with 5

×

10¹²-1

×

10¹⁴ cm^-2 nitrogen ionsdose and then annealed at 850 ^oC in N2 ambient. The EXAFS analysis reveals the local structural variation of the p-type ZnO films due to the substitution of nitrogen ions for oxygen ions in p-type ZnO films. The hole concentration, carrier mobility, and resistivity of p-type ZnO films were 5.0

×

10¹⁶-7.3

×

10¹⁷ cm^-3, 2.51-6.02 cm^-2Vs, and 10.11-15.3 Ωcm, respectively. PL spectra of the nitrogen-implanted ZnO/Si3N4/Si showed a sharp UV emission and invisible deep-level transition at room-temperature measurement. The high value of IUV/IDLE in nitrogen-implanted ZnO/Si3N4/Si also shows better optical quality relative to nonimplanted and nitrogen-implanted into ZnO/Si samples. These results suggest that nitrogen-implanted ZnO films deposited on a Si buffer with Si3N4 show electrical and optical behaviors that make them excellent candidates for a good p-type layer for ZnO-based optoelectronic device on a Si-based substrate.

Figure 4.10: Room-temperature PL spectra of non-implanted and nitrogen-implanted ZnO/Si3N4/Si samples.

Chapter 5

Physical characteristics of ZnO nanorods via soft chemical process

5.1 Introduction

The ZnO nanorods have been thought to be the most suitable for ultraviolet (UV) laser device because of its large exciton binding energy of 60 meV compared to that (26 meV) of GaN at room temperature. Furthermore, many reports in the related literature have demonstrated that directionally grown ZnO nanorods can effectively decrease the threshold power to achieve the UV lasing emission at room-temperature.

Strong efforts have been made to fabricate one-dimensional ZnO, including thermal decomposition[109], hydrothermal synthesis[110], vapor phase transport[111], and metal-organic chemical vapor deposition (MOCVD) [112]. Recently, many wet-chemical approaches have been used for large oriented arrays of ZnO nanorods on polycrystalline (or single-crystalline) substrates from aqueous solutions [113,114].

However, it is worth noting that without suitable treatment on the substrates, highly oriented ZnO nanorods grown on a Si wafer has been rarely achieved.

5.2 ZnO nanorods array on organic substrate

Recently, many research groups used ZnO nanoparticles as seeds to grow the large-scale and well-oriented ZnO nanorods on Si substrates via low temperature

process [115]. However, it is potentially more important to synthesize 1D nanoscale materials on organic substrates for the applications of flexible display and photoelectronic devices. Therefore, we wanted to investigate the nucleation and growth behavior of ZnO nanorods on organic substrates in aqueous solutions. Figure 5.1(a) shows the surface images of large-scale arrayed ZnO nanorods grown on polystyrene (PS)/polycarbonate (PC) substrates. It was found that the ZnO nanorods have a well-defined hexagonal plane with a homogeneous diameter of approximately

~60 nm due to uniform growth rate. The cross-sectional SEM image in Figure 5.1(b) shows that the ZnO nanorods are directionally and densely grown over the entire PS surface of the substrates. Furthermore, it was noted that the well-aligned ZnO nanorods (Figure 5.1(c)) can be developed after the removal of the PS beads from the specimen fabricated using the same conditions as Figure 5.1(b). It implies that the PS layer can supply the appropriate environment to increase the nucleation sites for ZnO nanorods.

Figure 5.1: SEM images of large arrays of oriented ZnO nanorods grown on polystyrene (PS)/polycarbonate (PC) substrates for 8 h. (a) Low magnification, face-on view. (b) Cross-sectional SEM image of ZnO nanorods grown on PS/PC substrates. (c) SEM image of ZnO nanorods grown on PS/PC substrates after removing the PS beads.

Figure 5.2: SEM images of ZnO nanorods grown on PS/PC substrates as a function of reaction time at 75^oC. (a)1 h, and (b) 5 h. (c) High-resolution TEM image of ZnO nanorods grown on organic substrates. (d) High-resolution TEM image of the interface region between ZnO nanorods and ultra-thin ZnO monolayer/ PC.

To understand how the well-aligned ZnO nanorods were developed on the organic substrates, the sample was subjected to grow at 75^oC in a variety of time. The first initial stage (t < 0.5 h) can be considered as the induction time, during which an ultra-thin monolayer was slowly generated under the PS beads in the solution. It is difficult to observe the ultra-thin monolayer by SEM, but this layer can be justified to be ZnO by chemical analysis such as x-ray photon spectroscopy after removing the PS beads. Later, in the second stage, the ZnO nanorods start to nucleate from the concave regions of PS layers as evidenced from Figure 5.2(a). It was believed that the promotion of heterogeneous nucleation on the concave regions is ascribed to the high affinity of the ZnO nuclei to the ZnO layer under PS beads. To investigate the

nucleation behavior of the aligned ZnO nanorods, HRTEM was used to study the interface between ZnO nanorods and ZnO monolayer (~5 nm thicker) shown in Figure 5.2(d )after removing the PS beads. It was observed that the well-aligned ZnO nanorods are preferentially nucleated from the concave region in ZnO monolayer.

Recently, it was also reported that the aligned ZnO nanorods can be grown on the Si substrates with ZnO nanoparticles or crystalline films as buffered layers [115]. After nucleation, in the third period, a preferential growth along longitudinal direction (c-axis) for the oriented ZnO nanorods was expected because the growth in width (or lateral) direction is suppressed due to the size of concave shape in PS layers as shown in Figure 5.2(b). The high resolution TEM image of the well-aligned ZnO nanorods in Figure 5.2(c) demonstrates that <001> direction is the preferred growth direction for the well-aligned ZnO nanorods grown on PS/PC substrates. It indicates that the lattice fringes are perpendicular to the longitude direction of the ZnO nanorods, and the singular fringe spacing is about 0.51nm, which is nearly consistent with the c-axis parameter in hexagonal ZnO structure (c= 0.521nm in Wurtzite ZnO). A detail investigation for the growth behavior of ZnO nanorods grown on inorganic substrates (Si) with ZnO coated can be referred to our previous study [116]. This demonstrates that the well-aligned ZnO nanorods can be grown at a lower temperature from aqueous solutions on both organic and inorganic substrates.

Figure 5.3: (a) Face-on view SEM image of well-aligned ZnO nanorods grown on PS/PC substrates for long-term growth (24 h). A <001> zone-axis TEM bright-field (BF) image of the coalescent ZnO nanorods (its scale bar is 100 nm in the inset), and (b) a corresponding diffraction pattern for the inset of Fig. 3a. (c) Cross-sectional TEM (BF) and (d) dark-field (DF) images of the coalescent ZnO nanorods. (e) A high magnification BF TEM image of (c), showing the interface (marked with dash line) of the coalescent couple ZnO nanorods.

After long-term growth, i.e., 24h, the SEM surface image of ZnO nanorods in Figure 5.3(a) reveals that the ZnO nanorods have started on coalescence process with other adjacent nanorods. As compared with Figure 5.1(a), the ZnO rods become much thicker (more than 0.5 µm) in diameter but no more grow longer (about 1.7 µm) in length. Therefore, in this condition, the growth rate of ZnO nanorods in the <001>

orientation would obviously decrease. The TEM bright-field (BF) image (inset in Figure 5.3(a)) of the (cross-sectional) ZnO nanorods clearly demonstrates the merged grains with side crystal plane attached that can be further confirmed by the <001>

zone-axis selected-area electron diffraction (SAED) pattern of ZnO nanorods in Figure 5.3(b). The split diffraction spots in the edge region (marked with arrows) of the SAED pattern suggest that the merged ZnO nanords are not perfectly aligned in both a and b directions. The TEM BF and dark-field (DF) images of the merged ZnO nanorods in Figure 5.3(c) and (d), respectively, clearly shows that more than two ZnO nanorods were aggregated in a coplanar manner using their side planes to form a larger ZnO nanorod. A magnified picture of Figure 5.3(c) was illustrated in Figure 5.3(e) where the SAED patterns of single and coalescent nanorods are also shown in the left and right insets, respectively. This further explains the coalescence of ZnO nanorods with a slight misalignment between nanorods. Therefore, the growth behavior of the ZnO nanorods in the later long-term growth stage can be considered as a direct combination of a small number of individual nanorods that was similar to oriented attachment [117].

Figure 5.4 illustrates the room-temperature photoluminescence (PL) spectra of the ZnO nanorods grown on PS/PC substrates. For ZnO nanorods grown with a short period of time, i.e. 0.5 h, only a weak PL peak around 378 nm is detected which can be connected to the ultraviolet (UV) emission of ZnO with a bandgap of 3.27 eV. As

corresponding to the 3.26 eV bandgap transition of ZnO, was detected. However, a broad yellow emission around 575 nm was also observed that is due to the deep levels.

The value of the relative PL ratio (Ultraviolet (UV) emission (IUV) to that of deep level emission (IDLE)) is estimated to be about ~ 5.1. Generally, the UV emission peak of ZnO is generally attributed to an exciton-related activity [50,51], and the deep level emission may be due to the transitions of native defects such as oxygen vacancies and zinc interstitials [54,55]. It is well understood that PL spectra depend on the stoichiometry and the microstructure defects of the materials. As confirmed and reported in our previous results [117], the ZnO nanorods grown in the aqueous solution at a lower temperature may induce unstable surface status to trap impurities and further damage the optical properties. Especially for the ZnO nanorods with a long-term growth, i.e. 24 h, it was observed that the broad orange emission in spectrum c becomes stronger compared to that of ZnO nanorods grown for 5 h (spectrum b). The relative PL ratio ((IUV) / (IDLE)) is reduced from 5.1 to 1.2. It is suggested that the native defects due to imperfect boundaries and misalignment between ZnO nanorods were increased for the merged ZnO nanorods. Therefore, it implies that both native defects and optical quality of ZnO nanorods grown on a flexible substrate could be controlled by changing growth conditions.

Figure 5.4: Room-temperature PL spectra of ZnO nanorods grown on PS/PC substrates for growth time at 0.5, 5, and 24 h.

5.3 Post-annealing effects on ZnO nanorods

It is well known that both physical characterization and optoelectronic properties are strongly influenced by the defect concentration of native defect in ZnO and this can be modified via thermal treatment under different atmospheres and annealing conditions [118]. Therefore, it is important to investigate the effect of post-annealing on the crystallinity and optical properties of ZnO nanorods. The variation of structural and optical properties of ZnO nanorods with thermal annealing conditions will be detailed discussed in this work. Figure 5.5 plots the room-temperature photoluminescence spectra of the ZnO nanorods annealed under various conditions.

Only two emission peaks at 377nm (UV emission) and 595 nm (visible emission) were observed. It was found that UV emission intensities increase with annealing

suggesting that the native defects or non-radiative recombination can be reduced by post-annealing treatment in O2 and N2 atmospheres as shown in Figure 5.5(a) and (b).

However, the ZnO nanorods annealed in N2 show stronger visible emission peaks compared to that annealed in oxygen atmosphere because the oxygen vacancies become the predominant point defects in N2 atmosphere. In addition, a unique phenomenon was observed for the sample annealed in H2/N2 atmosphere, as presented in Figure 5.5(c). An optimal UV emission occurs at the 600^oC. It was believed that H2/N2 treatment is able to passivate native defects or impurities that contribute to visible transition, because the hydrogen atoms can be situated in various lattice positions. However, when the sample was exposed to H2/N2 at 800^oC, the UV emission almost disappeared. This marked change can be attributed to the fact that a high-temperature reduction environment of H2/N2 could damage the crystal structure of the ZnO nanorods by surface etching [119]. It was also observed that the peak position of the longer-wavelength visible emission band shifts with different treatments. The peak position of the visible (deep-level) emission is related to the predominant defects in the ZnO nanorods and both defects of zinc interstitials and oxygen vacancies are strongly modified by changing annealing temperature and using different atmospheres [120]. The relative PL ratios (IUV/IDLE) of the samples as a function of various atmospheres can be further summarized and presented in Figure 5.6, revealing that the improvement in the optical quality of post-annealed ZnO nanorods is not only dominated by the annealing temperature but also the annealing atmosphere.

Figure 5.5: Room-temperature PL spectra of ZnO nanorods annealed at various temperatures in (a) O2, (b) N2 and (c) H2/N2 atmospheres.

Figure 5.7(a) shows that the as-grown ZnO nanorods are perpendicular to the substrate with a uniform length of 900-950 nm. As annealed at 600^oC in H2/N2 , the

Figure 5.7(a) shows that the as-grown ZnO nanorods are perpendicular to the substrate with a uniform length of 900-950 nm. As annealed at 600^oC in H2/N2 , the

在文檔中 一維暨二維氧化鋅奈米光電特性及結構研究 (頁 66-0)