Open Access
Add to BibSonomy
Abstract
The anti-reflection functions of a surface nanostructure–including transparent conductive Ga-doped ZnO (GaZnO) nanoneedles (NNs), a GaZnO thin film, and buried Ag nanoparticles (NPs), on GaN and Si templates through the combination of the effects of gradient effective refractive index, index matching, and the surface plasmon (SP) resonances in the visible and infrared range–are studied by measuring its reflection, transmission, and scattering behaviors. The NNs are grown under different molecular beam epitaxy conditions with the low-temperature vapor-liquid-solid mode by using Ag NPs as growth catalyst. Based on the crystal structure study, it is found that the c-axis of a GaZnO NN is controlled by the local Ag (111) orientation of the un-melted portion of an Ag NP, which is influenced by the crystal structure of the growth template. On c-plane GaN, by using small and separate Ag NPs as catalysts, the alignment of GaN (002), Ag (111), and ZnO (002) leads to the growth of mostly vertical NNs for producing a strong anti-reflection effect. On Si (100), no crystal matching condition can be used such that the grown NNs are randomly oriented, leading to a relatively weaker anti-reflection effect. The GaZnO thin film and buried Ag NPs also make contributions to the anti-reflection function through the effects of index matching and SP resonance.
© 2017 Optical Society of America
1. Introduction
Besides the structure of single- or multiple-layer surface coating [1, 2], surface nanostructures have been widely used for generating the anti-reflection effects in photovoltaic devices. Various Si surface nanostructures on Si solar cells have been fabricated to show strong anti-reflection effects [3–10]. Based on the effective refractive-index model in a surface Si nanostructure, the gradient change of effective refractive index can lead to very low reflection in a Si solar cell. Such a low reflection level is partly attributed to the same surface nanostructure material as the solar cell material. However, with a Si nanostructure on the surface of a Si solar cell, it is usually difficult to fabricate a metal contact on it. Also, normally a transparent conductive oxide (TCO) or passivation layer is needed. Recently, surface metal nanoparticles (NPs) for inducing surface plasmon (SP) resonance have been used for enhancing the sunlight absorption of solar cells [9, 10]. Also, nanostructured TCOs, including ZnO [11–14] and TiO2 [15], have been widely applied to solar cell for generating the anti-reflection effects. In particular, due to its low growth temperature, low growth selectivity, and high growth speed along its c-axis, ZnO nanowires have been widely used for this purpose [16–19]. For increasing its conductivity, Ga or Al has been doped into ZnO in fabricating nanowire structures. By using such highly transparent and conductive nanowires for generating the anti-reflection effects, the fabrication process of a solar cell can be much simplified. Also, with the electron concentration as high as 1021 cm−3 in highly Al-doped ZnO or Ga-doped ZnO (GaZnO), its SP resonance behaviors in the spectral range beyond 1000 nm have also been considered for solar cell applications [20, 21]. In this regard, efforts have been made for transferring the energies of SP-resonance induced GaZnO absorption and strong nearby electromagnetic field into photocurrent [22, 23]. Such an effect equivalently increases the anti-reflection function in a solar cell.
For fabricating the Al-doped ZnO or GaZnO nanowires, the vapor-liquid-solid (VLS) growth mode is a commonly used technique [24–26]. In this technique, certain metal NPs are used as growth catalyst. At a high growth temperature, a metal NP is melted for absorbing the constituent atoms of the nanowire to be grown. When the supersaturation condition is reached, the constituent atoms precipitate at the bottom of the melted metal NP for forming the nanowire. To grow GaZnO nanowires of a high Ga-doping concentration for increasing electron concentration and conductivity, normally the growth temperature needs to be lower than 450 °C [27–29]. At such a low growth temperature, an Ag NP of a couple tens nm in size cannot be completely melted for serving as growth catalyst. Only the upper portion of such an Ag NP is melted for absorbing the constituent atoms of GaZnO. In this situation, GaZnO precipitates onto the un-melted lower portion of an Ag NP. Therefore, the crystal structure and orientation of a GaZnO nanowire are not directly influenced by the growth template. Because the atoms of the catalytic Ag NP portion mix into the precipitated GaZnO, this Ag NP portion shrinks its size as it is lifted up along the nanowire growth. In this process, the cross section of the nanowire becomes smaller and smaller along the growth until the catalytic Ag is exhausted. Hence, the geometry of a nanoneedle (NN) is formed. Such an NN array can produce an effective anti-reflection function based on the model of effective refractive index. However, the effectiveness of the anti-reflection function depends on the NN geometries and more importantly the NN orientations.
In this paper, we study the anti-reflection function of a surface nanostructure, including transparent conductive GaZnO NNs, a GaZnO thin film, and buried Ag NPs, on GaN and Si templates through the combination of the effects of gradient effective refractive index, index matching, and the SP resonances in the visible and infrared ranges. The GaZnO thin film at the NN bottom is formed simultaneously with the growth of NNs. The buried Ag NPs on the template surface correspond to the un-melted portions of the originally prepared Ag NPs. We first demonstrate the relations between the anti-reflection behaviors and the morphologies, including the orientation, length, and width of GaZnO NNs grown under different conditions of molecular beam epitaxy (MBE) on c-plane GaN and Si (100) templates. A sample with thinner and mostly vertical NNs with respect to the template surface leads to a stronger anti-reflection effect. Then, we investigate the microscopic mechanism of determining the orientation of a GaZnO NN under a certain growth condition. It is found that a GaZnO NN is always grown along its c-axis [(002) direction], which is always aligned with the local Ag (111) orientation of the un-melted Ag NP portion. The Ag (111) orientation of the un-melted Ag NP portion is influenced by the crystal structure of the used template and the planar density of Ag NPs. The roles of the GaZnO thin film and buried residual Ag NPs in the anti-reflection function are also elucidated. In section 2 of this paper, we describe the MBE growth conditions for forming GaZnO NNs. Then, in section 3, we describe the morphologies of the GaZnO NNs in the samples under study. The behaviors of reflection, transmission, and scattering of those GaZnO NN samples are presented in section 4. In section 5, the study results of the microscopic mechanism determining the NN orientation is reported. Discussions about the results are made in section 6. Finally, the conclusions are drawn in section 7.
2. Sample growth conditions and characterization methods
GaZnO NNs are formed based on the VLS growth mode using Ag NPs as growth catalyst on c-plane, Ga-polar GaN and Si (100) templates in an RF-plasma assisted MBE reactor. In a GaN template, a 3-μm GaN layer is deposited on c-plane sapphire substrate with metalorganic chemical vapor deposition. GaZnO is grown under the Zn-rich conditions of 320 °C in Zn effusion cell temperature, 900 °C in Ga effusion cell temperature, 1 sccm in O2 flow rate, 350 W in RF-plasma power, and 350 or 450 °C in substrate temperature. During the GaZnO growth, besides the formation of NNs, a GaZnO thin film covering the residual Ag NPs is deposited through the vapor-solid growth mode. The Ag NPs are formed by first depositing an Ag layer of 1-1.6 nm in thickness on a template followed by a thermal annealing process at 160-300 °C for 30 min with ambient nitrogen. The technique of selected area electron diffraction (SAED) based on transmission electron microscopy (TEM) observation is used for identifying the local crystalline orientations of Ag NP, GaZnO NN, GaN, and Si templates to understand their relations.
3. Morphologies of GaZnO nanoneedles and sample designations
Figure 1(a1) shows the tilted scanning electron microscopy (SEM) image of Ag NPs on a GaN template formed by first depositing an Ag layer presumably of ~1.6 nm in thickness. Without any thermal annealing process, Ag NPs of 3.13 x 1011 cm−2 in planar particle density are formed. Figures 1(a2) and 1(a3) show the SEM images of GaZnO NNs with different magnifications grown at 350 °C for 80 min on the Ag NP template shown in Fig. 1(a1). This GaZnO NN sample on GaN grown at 350 °C is referred to as sample GaN/NN-A. Figures 1(b1)-1(b3) show the SEM images similar to Figs. 1(a1)-(a3), respectively, with the GaZnO NNs grown based on a different Ag NP template, which is formed by depositing Ag of ~1.6 nm in thickness and then thermal annealing at 160 °C for 30 min with ambient nitrogen. As shown in Fig. 1(b1), the planar particle density of the Ag NPs (2.83 x 1011 cm−2) becomes lower after thermal annealing. This GaZnO NN sample on GaN template grown at 350 °C is referred to as sample GaN/NN-B. Figures 1(c1)-1(c3) show the SEM images similar to Figs. 1(b1)-(b3), respectively, with the GaZnO NNs grown on an Ag NP template formed by depositing Ag of ~1.0 nm in thickness and then thermal annealing at 200 °C for 30 min with ambient nitrogen. With the thinner Ag deposition and the higher annealing temperature, the planar density of the Ag NPs in Fig. 1(c1) becomes even smaller (2.60 x 1011 cm−2). This GaZnO NN sample on GaN grown at 350 °C is referred to as sample GaN/NN-C. Figures 2(a1)-2(a3) show the SEM images similar to Figs. 1(c1)-(c3), respectively, with the GaZnO NNs grown on an Ag NP template formed under the same conditions, but the MBE growth temperature is increased to 450 °C. This GaZnO NN sample on GaN grown at 450 °C is referred to as sample GaN/NN-D. Figures 2(b1)-2(b3) show the SEM images similar to Figs. 1(c1)-(c3), respectively, with the GaZnO NNs grown on an Ag NP template, which is formed on a double-side polished Si (100) substrate by depositing Ag of ~1.4 nm in thickness and then thermal annealing at 300 °C for 30 min with ambient nitrogen. The substrate temperature for growing GaZnO NNs is 350 °C in this case. This GaZnO NN sample on Si grown at 350 °C is referred to as sample Si/NN.
Fig. 1 (a1): Tilted SEM image of Ag NPs on a GaN template for growing sample GaN/NN-A. (a2) and (a3): SEM images of GaZnO NNs of sample GaN/NN-A with different magnifications. (b1)-(b3) [(c1)-(c3)]: SEM images similar to parts (a1)-(a3), respectively, for sample GaN/NN-B (GaN/NN-C).
Fig. 2 (a1)-(a3) [(b1)-(b3)]: SEM images similar to Figs. 1(c1)-(c3), respectively, for sample GaN/NN-D (Si/NN).
By comparing the NN morphologies between Figs. 1(a2)-1(c2) or between Figs. 1(a3)-1(c3), one can see that when the planar density of Ag NPs is reduced, more GaZnO NNs are vertically oriented with respect to the template surface. Many NNs in sample GaN/NN-A tilt almost along the template surface. Although it is difficult to estimate the lengths and base widths of the NNs in those samples, we can still see that generally the NN length increases and the base width decreases with decreasing Ag NP density in samples GaN/NN-A, GaN/NN-B, and GaN/NN-C. By increasing the growth temperature from 350 to 450 °C, the NNs in sample GaN/NN-D become thinner and longer, when compared with those in sample GaN/NN-C. Although some of the thin NNs bend in sample GaN/NN-D, they are essentially vertically oriented with respect to the template surface. In sample Si/NN, the orientations of grown NNs are randomly distributed although many of them are vertically oriented with respect to the template surface.
During the growth of NNs, a GaZnO thin film of ~150 nm in thickness is simultaneously deposited at the bottom of the NNs. Also, residual (un-melted) Ag NPs lie at the interface between GaZnO and the used template. To observe the effects on the anti-reflection function of the GaZnO thin film and residual Ag NPs, we prepare a sample of a ~150-nm GaZnO thin film on GaN with a growth temperature at 350 °C, which is designated as sample GaN/GZO, for comparison. Meanwhile, on Si (100), we prepare a sample with surface Ag NPs covered by a ~150-nm GaZnO thin film. To avoid the formation of NNs, the GaZnO growth temperature is reduced to 250 °C. This sample of Ag NPs covered by a GaZnO thin film is designated as sample Si/NP/GZO. In addition, another sample of a ~150-nm GaZnO thin film without Ag NP, which is also grown at 250 °C, is prepared and designated as sample Si/GZO. In the following optical behavior studies, GaN and Si templates are also used and designated as samples GaN and Si, respectively.
4. Optical behaviors of GaZnO nanoneedles
For understanding the behaviors of reflection, transmission, and scattering of those GaZnO NN samples, we measure the line-of-sight transmission (transmittance), the specular reflection (reflectance), the total scattered power integrated over the incidence half-space, and the total scattered power integrated over the transmission half-space. In all measurements, the incident angle is 5 degrees. Figure 3 shows the spectral variations of transmittance of those NN samples grown on GaN. For comparison in this figure, we also plot the transmittance spectra of samples GaN and GaN/GZO. Here, the slight oscillation in the transmittance curve for the GaN template is caused by the Fabry-Perot effect. After depositing the GaZnO thin film on the GaN template, the transmittance is slightly increased, particularly on the short-wavelength side. Among the NN samples, sample GaN/NN-A of highly tilted NNs does not show enhanced transmission, when compared with that of the GaN template. The transmittance is enhanced in the spectral range between 700 and 1400 nm in sample GaN/NN-B. In samples GaN/NN-C and GaN/NN-D, the spectral windows of enhanced transmittance are even larger and the maximum transmittance levels are further increased. In other words, the transmittance is generally higher when light passes through more vertically oriented GaZnO NNs. In sample GaN/NN-D, the transmittance can reach 85% in the spectral range between 1000 and 1400 nm. The decreasing transmittance beyond 1200 nm in samples GaN/NN-A, GaN/NN-B, and GaN/NN-C is due to the SP resonance induced absorption of GaZnO NNs and thin film because of the relatively lower growth temperature (350 °C) and hence higher electron concentration. Because the growth temperature is higher (450 °C) and hence the electron concentration is lower in sample GaN/NN-D, the SP resonance in the concerned spectral range is weaker such that the transmittance is higher beyond 1200 nm. Figure 4 shows the spectral variations of reflectance corresponding to the transmittance data in Fig. 3. Here, we can see that by depositing a GaZnO thin film or forming GaZnO NNs on GaN, the reflectance can be more or less reduced. Among the NN samples, the reflectance is generally lower when light is reflected from more vertically oriented GaZnO NNs.
Fig. 3 Spectral variations of transmittance of samples GaN, GaN/GZO, GaN/NN-A, GaN/NN-B, GaN/NN-C, and GaN/NN-D.
Fig. 4 Spectral variations of reflectance of samples GaN, GaN/GZO, GaN/NN-A, GaN/NN-B, GaN/NN-C, and GaN/NN-D.
Figure 5 shows the power percentages integrated over the transmission half-space corresponding to the transmittance data shown in Fig. 3. Here, one can see that the spectral variations among different samples are similar to the corresponding curves in Fig. 3. However, by including the scattered light power in the directions deviated from the line-of-sight, the collected light power is enhanced in each sample. Figure 6 shows the power percentages integrated over the incidence half-space corresponding to the data shown in Fig. 5. Again, the spectral variations among different samples are similar to the corresponding curves in Fig. 4. However, by including the scattered light power in the directions deviated from the specular angle, the collected light power is enhanced in each sample. With the power percentages collected in the transmission and incidence half-spaces in Figs. 5 and 6, respectively, we can evaluate the absorbance spectra of those samples to give the results in Fig. 7. Here, the increased absorbance below 600 nm in the NN samples are due to the SP-resonance induced absorption of the residual Ag NPs. The increased absorbance beyond 1400 nm in those samples is caused by the SP-resonance induced absorption of GaZnO.
Fig. 5 Spectral variations of the power percentage collected over the transmission half-space of samples GaN, GaN/GZO, GaN/NN-A, GaN/NN-B, GaN/NN-C, and GaN/NN-D.
Fig. 6 Spectral variations of the power percentage collected over the incidence half-space of samples GaN, GaN/GZO, GaN/NN-A, GaN/NN-B, GaN/NN-C, and GaN/NN-D.
Fig. 7 Spectral variations of absorbance of samples GaN, GaN/GZO, GaN/NN-A, GaN/NN-B, GaN/NN-C, and GaN/NN-D.
Figure 8 shows the spectral variations of transmittance (T) and reflectance (R) of samples Si/NN, Si/NP/GZO, Si/GZO, and Si. Because of the strong Si absorption below 900 nm, the transmittance of either sample is almost zero below this wavelength. Due to the absorption caused by the SP resonance of GaZnO, the transmittance of sample Si/NN is lower than that of the Si template beyond 1000 nm. The even stronger SP-resonance induced absorptions in samples Si/NP/GZO and Si/GZO result in the even lower transmittance levels beyond 1100 nm. The stronger SP resonance in either sample Si/NP/GZO or sample Si/GZO is attributed to the lower GaZnO growth temperature. At the lower growth temperature of 250 °C, the electron concentration of GaZnO becomes higher and hence the SP resonance becomes stronger, when compared with the cases of 350 and 450 °C in growth temperature [27, 28]. With the anti-reflection effect of GaZnO NNs in sample Si/NN, the reflectance of this sample is significantly reduced, when compared with sample Si. The reflectance levels of samples Si/NP/GZO and Si/GZO lie between those of samples Si/NN and Si below 1300 nm. They increase fast with wavelength beyond 1000 nm. The reflectance level of sample Si/NP/GZO is lower than that of sample Si/GZO, indicating that the Ag NPs contribute significantly to the anti-reflection function. In Fig. 9, we show the power percentages collected over the incidence (R) and transmission (T) half-spaces corresponding to the reflectance and transmittance results in Fig. 8. The spectral variations of the power percentages collected over the incidence and transmission half-spaces are similar to those of reflectance and transmittance, respectively, shown in Fig. 8. However, the power levels are all increased. In particular, beyond 1000 nm, the power percentage collected over the transmission half-space in sample Si/NN is significantly higher than the transmittance level, indicating that the strong scattering of GaZnO NNs deviates light propagation from the line-of-sight when light passes through the sample. Figure 10 shows the spectral variations of sample absorbance obtained from the data in Fig. 9. Below 1000 nm, due to the reduced returned power percentage, the Si absorption is significantly increased in sample Si/NN. In this spectral range, the Si absorbance levels in samples Si/NP/GZO and Si/GZO are also enhanced due to their anti-reflection functions. Beyond 1100 nm, the high absorbance levels of samples Si/NP/GZO and Si/GZO are caused by the SP-resonance induced absorption of the GaZnO thin films.
Fig. 8 Spectral variations of transmittance (T) and reflectance (R) of samples Si, Si/NP/GZO, Si/GZO, and Si/NN.
Fig. 9 Spectral variations of the power percentages collected over the incidence (R) and transmission (T) half-spaces of samples Si, Si/NP/GZO, Si/GZO, and Si/NN.
5. Microscopic mechanism determining GaZnO nanoneedle orientation
Figure 11(a) shows the TEM image of an NP-like body on a GaN template in a sample similar to sample GaN/NN-C, but the MBE growth duration is only 10 min. Here, we plot the horizontal (pink) dotted line to show the top boundary of GaN. Figures 11(b1), 11(b2), and 11(d1)-11(d4) [11(c1)-11(c4)] show the SAED patterns of the circled regions in Fig. 11(a) together with the dashed-line indicators of the Ag (111) [ZnO (002)] orientation. Figure 11(e) shows the similar result of a GaN region with the GaN (002) orientation. Here, we can see that the NP-like body consists of three layers with the major contents of Ag, ZnO, and Ag. In all layers, the Ag (111) and ZnO (002) orientations are the same as that of GaN (002) in the template. The crystal structure in this NP-like body shows an early-stage condition of the VLS growth for forming a GaZnO NN. The top layer corresponds to the melted Ag portion for serving as growth catalyst and absorbing Ga and Zn atoms. Under the supersaturation condition, Ga and Zn atoms precipitate at the bottom of the melted Ag portion to interact with oxygen atoms coming from the triple-phase line (the boundary between melted and un-melted Ag and vapor) for forming GaZnO on the un-melted, lower Ag portion [30]. Therefore, a layer of GaZnO is sandwiched by the top and bottom Ag layers. For further understanding the composition distribution of this NP-like body, we perform line-scan energy-dispersive X-ray spectroscopy (EDX) measurement along the three vertical lines, 1-3, in Fig. 12(a), which is duplicated from Fig. 11(a). The EDX signal profiles of Ga, Zn, and Ag along lines 1-3 are shown in Figs. 12(b)-12(d), respectively. Along all lines, Ga content increases essentially monotonically along depth. Along each line-scan, there is a Zn distribution peak roughly in the 15-30 nm range in depth. Also, we can see two Ag distribution portions roughly in the ranges of 0-18 and 28-35 nm in each line-scan. The Ag content is lower in the range of 0-18 nm along line-scan 3 because of the thin NP structure around this corner. In each line-scan, small amounts of Ga and Zn atoms exit in the range of 1-15 nm. From these EDX line-scan results, we can conclude that GaZnO is formed in the range around 15-30 nm, as indicated in Fig. 12(a). In the layer above the GaZnO layer, residual Ga and Zn atoms mix with Ag atoms, confirming that this is the melted portion of the Ag NP for absorbing Ga and Zn to form GaZnO below. The range of 35-40 nm in depth of each line-scan corresponds to the GaN template.
Fig. 11 (a): Cross-sectional TEM image of an NP body on a GaN template. The pink dotted line plots the boundary between Ag and GaN. (b1), (b2), (c1)-(c4), (d1)-(d4), and (e): SAED patterns together with the dashed-line indicators of the Ag (111), ZnO (002), and GaN (002) orientations in the circled regions of part (a).
Fig. 12 (a): TEM image duplicated from Fig. 11(a) with three vertical green arrows drawn for EDX line scans. (b)-(d): Line-scan EDX signal profiles of Ga, Zn, and Ag along arrows 1-3, respectively, shown in part (a).
Figure 13(a) shows the cross-sectional TEM image of the GaZnO NNs on GaN in a sample similar to sample GaN/NN-C, but the growth duration is only 10 min. Here, one can see essentially vertical NNs with an Ag NP at the top of each NN. The bottom Ag clusters come from the mixture of the un-melted portions of Ag NPs. Figure 13(b) shows the TEM image of a single NN, which is essentially vertical to the template surface. We magnify the circled top and bottom portions of the NN to give the TEM images in Figs. 13(c) and 13(d), respectively. The top Ag NP has a single-crystal structure with the SAED pattern and Ag (111) orientation shown in Fig. 13(c1). Near its top, the grown GaZnO NN shows a quasi-single-crystal structure with its SAED pattern and ZnO (002) orientation shown in Fig. 13(c2). In Fig. 13(d), one can see that an Ag layer of 15-20 nm in thickness separates the GaZnO NN and GaN template. Here, we demonstrate the SAED patterns and the key crystalline orientations of the GaZnO NN, Ag layer, and GaN template in Figs. 13(d1)-13(d3), respectively. One can see that the orientations of Ag (111) in the top Ag NP, ZnO (002) in the GaZnO NN near its top, ZnO (002) in the GaZnO NN near its bottom, Ag (111) in the bottom Ag layer, and GaN (002) in the template are all the same and vertically aligned.
Fig. 13 (a): Cross-sectional TEM image of the GaZnO NNs on GaN in a sample similar to sample GaN/NN-C, but the growth duration is only 10 min. (b): TEM image of a single NN. (c) and (d): Magnified TEM images in the circled top and bottom portions, respectively, of the NN in part (b). (c1), (c2), and (d1)-(d3): SAED patterns of the designated areas and the corresponding Ag (111), ZnO (002), and GaN (002) orientations.
Figure 14(a) shows the TEM image of a few GaZnO NNs on GaN in a sample similar to sample GaN/NN-A, but the growth duration is only 10 min. Here, we can see two NNs of different orientations grown from a cluster of Ag NP. Figure 14(b) shows the magnified TEM image of the circled bottom portion of the two NNs. As shown in the SAED patterns of Figs. 14(c1)-14(c3), the ZnO (002) orientation of the vertical NN on the left and the Ag (111) orientations in the Ag portion close to the bottom of this NN and in the bottom Ag portion close to GaN all align with GaN (002) orientation in the template [see Fig. 14(e)]. On the other hand, as shown in Figs. 14(d1) and 14(d2), the ZnO (002) orientation of the tilted NN on the right and the Ag (111) orientation in the Ag portion close to the bottom of this NN are the same and are different from those of other image portions. Such results indicate that the growth direction of a GaZnO NN follows the local crystalline orientation of an Ag NP by aligning the ZnO (002) direction with the Ag (111) orientation of the growth base. Because of the high growth speed of ZnO along its (002) direction, the NN growth follows the direction of Ag (111) of the un-melted residual Ag portion used as the base for NN growth. Figure 15(a) shows the TEM image of the GaZnO NNs on Si (100) in sample Si/NN. Here, we focus our analysis on the tilted NN circled by the pink square, particularly its bottom portion. Figure 15(b) shows the magnified TEM image around the bottom of this NN. Here, the dark portion near the center of this image corresponds to an Ag NP. Figures 15(c)-15(e) show the SAED patterns together with the individual orientation indications of Zn (002), Ag (111), and Si (400), respectively, in the circled regions. The Si (400) orientation is the same as that of Si (100), confirming the crystal structure of the used Si template. Although the Ag NP directly contacts the Si template, the Ag (111) orientation deviates from the Si (100) direction. Influenced by this Ag (111) orientation, the growth of the GaZnO NN tilts along a direction close to the Ag (111) orientation.
Fig. 14 (a): Cross-sectional TEM image of the GaZnO NNs on GaN in a sample similar to sample GaN/NN-A, but the growth duration is only 10 min. (b): Magnified TEM image of the circled bottom portion of the two NNs. (c1)-(c3), (d1), (d2), and (e): SAED patterns of the designated areas and the corresponding Ag (111), ZnO (002), and GaN (002) orientations.
Fig. 15 (a): Cross-sectional TEM image of the GaZnO NNs on Si in sample Si/NN. (b): Magnified TEM image around the bottom of the circled NN in part (a). (c)-(e): SAED patterns together with the individual orientation indications of ZnO (002), Ag (111), and Si (400), respectively, in the circled regions.
In the hexagonal structure of ZnO or GaN, its (002) plane has seven closely packed Zn or Ga atoms (O or N atoms). The nearest atomic distance in such a hexagonal, closely packed ZnO (GaN) structure, i.e., the a constant, is 0.325 nm (0.318 nm). Ag crystal has a face-centered cubic structure. On its (111) plane, it also shows a hexagonal, closely packed structure of seven Ag atoms with the nearest atomic distance at 0.288 nm. Although the nearest atomic distances of Ag (111) and ZnO (002) [or GaN (002)] are slightly different, their similar lattice structures of hexagonal, closely packed arrangements can lead to the minimum energy by aligning the Ag (111) and ZnO (002) [or GaN (002)] orientations. Therefore, on a c-plane GaN template, Ag NPs with Ag (111) orientation aligned with GaN (002) are formed when the template temperature is high enough for Ag crystal reorganization. Similarly, the ZnO (002) orientation of the grown GaZnO NN is the same as the Ag (111) direction of the Ag portion, onto which GaZnO is formed. On a Si (100) template, because there is no crystalline matching condition between Si and Ag, the Ag NPs have poly-crystalline structures of random orientations with respect to Si (100) direction. Following the local Ag (111) orientations, the grown GaZnO NNs become randomly oriented.
6. Discussions
The anti-reflection effect of a surface nanowire structure is usually interpreted with the effective refractive-index model, i.e., the gradual change of the laterally average dielectric constant along the nanowire length. Based on this model, a sample of vertically oriented GaZnO NNs can result in a gradual decrease of effective refractive-index along depth. Therefore, the reflectance (transmittance) and power percentage collected in the incidence (transmission) half-space are the lowest (highest) in sample GaN/NN-C, followed by sample GaN/NN-B, and then sample GaN/NN-A, among the three samples grown at 350 °C on GaN. With the growth temperature increased to 450 °C, the thinner and longer NNs of essentially vertical orientations lead to even lower (higher) reflectance (transmittance) and power percentage collected in the incidence (transmission) half-space in sample GaN/NN-D. When GaZnO NNs are grown on a Si (100) template, the random distribution of NN orientation reduces the anti-reflection effect. However, the overall anti-reflection effect is also related to the refractive-index difference between the template and NN materials. In a GaZnO NN sample, a GaZnO thin film is simultaneously deposited at the bottom of NNs. In other words, light penetrating into the NNs can be reflected at the interface between the GaZnO thin film (refractive-index at ~1.8) and the template (refractive-index at ~2.4 in GaN and >3 in Si). Therefore, the power percentage collected over the incidence half-space in sample Si/NN (either below or beyond 1000 nm) is higher, when compared with samples GaN/NN-C and GaN/NN-D, because of the larger refractive-index difference between GaZnO and Si.
As shown in Figs. 13 and 14, a less dense Ag NP distribution with a larger NP separation on GaN can lead to the growth of vertically oriented GaZnO NNs for a stronger anti-reflection effect. The vertical orientation of an NN on c-plane GaN is due to the crystalline matching between ZnO (002), Ag (111), and GaN (002). On a Si template, because of the lack of such a crystalline matching condition between Si and Ag, the orientation of an NN is controlled by the local Ag crystal orientation, which is randomly distributed. Therefore, the NNs on Si are randomly oriented. One possible solution for making the GaZnO NNs vertical as possible is to first deposit a buffer layer, which has the crystal structure similar to GaZnO with its c-axis in the vertical direction, on Si. On such a buffer layer, the Ag (111) orientation of the formed Ag NPs can be aligned in the vertical direction such that the GaZnO NNs grown on them can be vertically aligned. For Si photovoltaic applications, this buffer layer needs to be transparent and conductive. A GaZnO thin film can be a good choice for this buffer layer. However, it needs efforts for finding the optimized growth condition to form such a favored buffer layer on Si.
As shown in Figs. 4, 6, 8, and 9, a GaZnO thin film on the template in either sample GaN/GZO or sample Si/GZO can reduce the reflectance and the power percentage collected over the incidence half-space based on the effect of index matching except in the long-wavelength range, in which the SP resonance of GaZnO changes the reflection behavior. Under the SP resonance condition of GaZnO, both reflection and absorption increase. The absorption of GaZnO can contribute to the photocurrent of a solar cell through the process of hot carrier generation [21, 22]. The strong electromagnetic field produced near a GaZnO nanostructure at SP resonance can enhance the solar cell absorption beyond 1000 nm [20]. This is particularly useful for a Si solar cell because the Si absorption coefficient decreases significantly beyond 1000 nm. The buried Ag NPs can also induce SP resonance in the visible range. It has been shown that such an SP resonance behavior can scatter incident sunlight into the forward direction such that the solar cell absorption can be enhanced [23]. The SP resonance and non-coherent scattering of Ag NPs lead to the enhanced anti-reflection function of sample Si/NP/GZO, when compared with sample Si/GZO, as shown in Figs. 8 and 9. The residual Ag NPs in sample Si/NN and the NN samples on GaN should also make the similar contributions to the anti-reflection function even though their Ag NP densities can be lower, when compared with that in sample Si/NP/GZO.
7. Conclusions
In summary, we have compared the morphology, particularly the orientation, and the anti-reflection behaviors of GaZnO NNs formed under different MBE growth conditions and with different catalytic Ag NP distributions on c-plane GaN and Si (100) templates based on the VLS growth method. For understanding the microscopic mechanism of controlling the GaZnO NN orientation, we analyzed the relations of crystal structure between the GaZnO NN, catalytic Ag NP, and growth template. It was found that the c-axis direction of a GaZnO NN was controlled by the local Ag (111) orientation of the un-melted portion of an Ag NP, which was influenced by the crystal structure of the growth template. On c-plane GaN, by using small and separate Ag NPs as catalyst, the alignment of GaN (002), Ag (111), and ZnO (002) could lead to the growth of mostly vertical NNs for producing a strong anti-reflection effect. On Si (100), no crystal matching condition could be used such that the grown NNs were randomly oriented, leading to a weaker anti-reflection effect. Besides the effects of NNs, the contributions of the simultaneously grown GaZnO thin film and buried Ag NPs through SP resonances to the anti-reflection function were also illustrated.
Funding
Ministry of Science and Technology, Taiwan, The Republic of China (MOST 104-2622-E-002-031-CC2, MOST 105-2221-E-002-159-MY3, MOST 106-2221-E-002-163-MY3, and MOST 105-2221-E-002-118) National Taiwan University (105R89095A); US Air Force Office of Scientific Research (AOARD-14-4105).
References and links
1. Y. F. Makableh, R. Vasan, J. C. Sarker, A. I. Nusir, S. Seal, and M. O. Manasreh, “Enhancement of GaAs solar cell performance by using a ZnO sol-gel anti-reflection coating,” Sol. Energy Mater. Sol. Cells 123, 178–182 (2014).
2. D. Li, D. Wan, X. Zhu, Y. Wang, Z. Han, S. Han, Y. Shan, and F. Huang, “Broadband antireflection TiO2-SiO2 stack coatings with refractive-index-grade structure and their applications to Cu(In,Ga)Se2 solar cells,” Sol. Energy Mater. Sol. Cells 130, 505–512 (2014).
3. S. K. Srivastava, D. Kumar, P. K. Singh, M. Kar, V. Kumar, and M. Husain, “Excellent antireflection properties of vertical silicon nanowire arrays,” Sol. Energy Mater. Sol. Cells 94, 1506–1511 (2010).
4. J. Zhu, Z. Yu, G. F. Burkhard, C. M. Hsu, S. T. Connor, Y. Xu, Q. Wang, M. McGehee, S. Fan, and Y. Cui, “Optical absorption enhancement in amorphous silicon nanowire and nanocone arrays,” Nano Lett. 9(1), 279–282 (2009). [PubMed]
5. P. Pignalosa, H. Lee, L. Qiao, M. Tseng, and Y. Yi, “Graded index and randomly oriented core-shell silicon nanowires for broadband and wide angle antireflection,” AIP Adv. 1, 032124 (2011).
6. A. Rahman, A. Ashraf, H. Xin, X. Tong, P. Sutter, M. D. Eisaman, and C. T. Black, “Sub-50-nm self-assembled nanotextures for enhanced broadband antireflection in silicon solar cells,” Nat. Commun. 6, 5963 (2015). [PubMed]
7. H. Fang, X. Li, S. Song, Y. Xu, and J. Zhu, “Fabrication of slantingly-aligned silicon nanowire arrays for solar cell applications,” Nanotechnology 19(25), 255703 (2008). [PubMed]
8. L. Zeng, X. Yu, Y. Han, and D. Yang, “Performance of silicon nanowire solar cells with phosphorus-diffused emitters,” J. Nanomater. 2012, 156986 (2012).
9. S. H. Baek, S. B. Kim, J. K. Shin, and J. Hyun Kim, “Preparation of hybrid silicon wire and planar solar cells having ZnO antireflection coating by all-solution processes,” Sol. Energy Mater. Sol. Cells 96, 251–256 (2012).
10. X. Liang, L. Shu, H. Lin, M. Fang, H. Zhang, G. Dong, S. Yip, F. Xiu, and J. C. Ho, “Inverted silicon nanopencil array solar cells with enhanced contact structures,” Sci. Rep. 6, 34139 (2016). [PubMed]
11. T. H. Wu, R. W. Chuang, C. Y. Huang, C. Y. Cheng, C. Y. Huang, Y. C. Lin, and Y. K. Su, “ZnO nanoneedles/ZnO:Al film stack as an anti-reflection layer for high efficiency triple junction solar cell,” Electrochem. Solid-State Lett. 15, H208–H210 (2012).
12. R. J. Chung, Z. C. Lin, C. A. Lin, and K. Y. Lai, “Study of an antireflection surface constructed of controlled ZnO nanostructures,” Thin Solid Films 570, 504–509 (2014).
13. R. Dalvand, S. Mahmud, J. Rouhi, and C. H. R. Ooi, “Well-aligned ZnO nanoneedle arrays grown on polycarbonate substrates via electric field-assisted chemical method,” Mater. Lett. 146, 65–68 (2015).
14. J. Y. Chen and K. W. Sun, “Growth of vertically aligned ZnO nanorod arrays as antireflection layer on silicon solar cells,” Sol. Energy Mater. Sol. Cells 94, 930–934 (2010).
15. J. W. Leem, J. Su Yu, D. H. Jun, J. Heo, and W. K. Park, “Efficiency improvement of III-V GaAs solar cells using biomimetic TiO2 subwavelength structures with wide-angle and broadband antireflection properties,” Sol. Energy Mater. Sol. Cells 127, 43–49 (2014).
16. Y. F. Yao, C. H. Shen, W. F. Chen, P. Y. Shih, W. H. Chou, C. Y. Su, H. S. Chen, C. H. Liao, W. M. Chang, Y. W. Kiang, and C. C. Yang, “Void structures in regularly patterned ZnO nanorods grown with the hydrothermal method,” J. Nanomater. 2014, 756401 (2014).
17. M. Sakurai, Y. G. Wang, T. Uemura, and M. Aono, “Electrical properties of individual ZnO nanowires,” Nanotechnology 20(15), 155203 (2009). [PubMed]
18. Z. Zheng, Z. S. Lim, Y. Peng, L. You, L. Chen, and J. Wang, “General route to ZnO nanorod arrays on conducting substrates via galvanic-cell-based approach,” Sci. Rep. 3, 2434 (2013). [PubMed]
19. S. Xu and Z. L. Wang, “One-dimensional ZnO nanostructures: Solution growth and functional properties,” Nano Res. 4, 1013–1098 (2011).
20. G. V. Naik, V. M. Shalaev, and A. Boltasseva, “Alternative plasmonic materials: beyond gold and silver,” Adv. Mater. 25(24), 3264–3294 (2013). [PubMed]
21. C. Clavero, “Plasmon-induced hot-electron generation at nanoparticle/metal-oxide interfaces for photovoltaic and photocatalytic devices,” Nat. Photonics 8, 95–103 (2014).
22. F. Wang and N. A. Melosh, “Plasmonic energy collection through hot carrier extraction,” Nano Lett. 11(12), 5426–5430 (2011). [PubMed]
23. F. J. Tsai, J. Y. Wang, J. J. Huang, Y. W. Kiang, and C. C. Yang, “Absorption enhancement of an amorphous Si solar cell through surface plasmon-induced scattering with metal nanoparticles,” Opt. Express 18(Suppl 2), A207–A220 (2010). [PubMed]
24. V. Bhosle, A. Tiwari, and J. Narayan, “Electrical properties of transparent and conducting Ga doped ZnO,” J. Appl. Phys. 100, 033713 (2006).
25. H. M. Chiu, H. J. Tsai, W. K. Hsu, and J. M. Wu, “Experimental and computational insights in the growth of gallium-doped zinc oxide nanostructures with superior field emission properties,” CrystEngComm 15, 5764–5775 (2013).
26. C. H. Hsu and D. H. Chen, “Synthesis and conductivity enhancement of Al-doped ZnO nanorod array thin films,” Nanotechnology 21(28), 285603 (2010). [PubMed]
27. Y. F. Yao, C. G. Tu, T. W. Chang, H. T. Chen, C. M. Weng, C. Y. Su, C. Hsieh, C. H. Liao, Y. W. Kiang, and C. C. Yang, “Growth of highly conductive Ga-doped ZnO nanoneedles,” ACS Appl. Mater. Interfaces 7(19), 10525–10533 (2015). [PubMed]
28. Y. F. Yao, C. H. Lin, C. Hsieh, C. Y. Su, E. Zhu, S. Yang, C. M. Weng, M. Y. Su, M. C. Tsai, S. S. Wu, S. H. Chen, C. G. Tu, H. T. Chen, Y. W. Kiang, and C. C. Yang, “Multi-mechanism efficiency enhancement in growing Ga-doped ZnO as the transparent conductor on a light-emitting diode,” Opt. Express 23(25), 32274–32288 (2015). [PubMed]
29. C. H. Lin, Y. F. Yao, C. Y. Su, C. Hsieh, C. G. Tu, S. Yang, S. S. Wu, Y. W. Kiang, and C. C. Yang, “Thermal annealing effects on the performance of a Ga-doped ZnO transparent-conductor layer in a light-emitting diode,” IEEE Trans. Electron Dev. 62, 3742–3749 (2015).
30. H. Simon, T. Krekeler, G. Schaan, and W. Mader, “Metal-seeded growth mechanism of ZnO nanowires,” Cryst. Growth Des. 13, 572–580 (2013).
