Oriented Growth of Single-Crystalline Bi2s3 Nanowire Arrays

Ju Xu,Nikolay Petkov,Xueyan Wu,Daniela Iacopino,Aidan J. Quinn,Gareth Redmond,Thomas Bein,Michael A. Morris,Justin D. Holmes

DOI: https://doi.org/10.1002/cphc.200600681

IF: 3.52

2007-01-01

ChemPhysChem

Abstract:Highly oriented single crystal Bi2S3 nanowire arrays are formed inside anodised alumina membranes (AAMs) using a new solventless approach. A high yield of pore filling is achieved by injecting the melted single source precursor liquid into the channels followed by thermolysis. The absorption profile (bandgap) of these nanowires shifts to higher energies as the mean diameter of the nanowires decreases. Various approaches, including electrochemical and electroless deposition, sol–gel, paired-cell, chemical vapour deposition and supercritical fluid deposition have been utilised to produce a range of inorganic nanowires, including metals, metal oxides and semiconductors within the channels of anodic alumina membranes (AAMs).1–4 However, these methods typically produce amorphous or polycrystalline nanowires with poorly defined structures.5 Post-treatment of the nanowires, for example, by annealing at elevated temperatures, is often required to achieve the desired crystallinity and mechanical integrity necessary for many applications.6, 7 Electrodeposition techniques have been used to obtain ordered single crystal arrays of metal and II–VI nanowires within AAMs.8–16 However, a significant limitation of this synthetic approach is that it requires the deposited material to be both easily reduced and conductive.17 Also, the optimal deposition conditions depend on the metal or compound to be deposited, which may restrict the compositional variety of the single crystal nanowires that can be formed. Several compounds such as CdS, TiO2 and YBa2Cu3O7-δ, with a single crystal structure, have also been obtained as nanowire arrays in AAMs using direct sol–gel or electric field assisted sol–gel methods.5, 18, 19 However, low pore filling efficiencies and shrinkage/cracking of the nanowires upon calcination are often observed with sol–gel-based methods. Additionally, rough surfaces and nonconsistent growth directions are typically obtained with single crystal nanowires grown by both electrodeposition and sol–gel approaches.5, 8–16, 18–20 A pressure injection process for growing arrays of single crystal nanowires of low melting point metal, for example, Bi, Sn, In, Al,21, 22 has been reported by Zhang et al.21 However, this method is limited to low melting point metals due to the deformation of the template at high temperatures.21 Thus, it is still a significant challenge to fabricate high density arrays of oriented single crystal nanowires, with high aspect ratios, within the channels of templates such as AAMs.5, 14, 16, 18, 23 Bi2S3, an ionic n-type semiconductor with a direct bandgap of 1.3 eV, has numerous potential applications as a material for photovoltaic24, 25 and thermoelectric26 devices. Several techniques have previously been reported for synthesizing randomly oriented Bi2S3 nanowires including solvothermal and hydrothermal decomposition,27–29 microwave irradiation,30 crystallization of amorphous colloids,31 chemical vapour deposition,32 sonochemical,33 solventless,[34]and ionic liquid-assisted methods.35 Polycrystalline Bi2S3 nanowires embedded inside the channels of AAMs have also been reported by electrochemical deposition20 and photochemical synthesis.36 To date, there have been no reports in the literature of the templated growth of arrays of single crystal Bi2S3 nanowires. Herein, we demonstrate a new and simple template-based solventless approach, without the use of a catalyst or complexing agent, to synthesize single crystal Bi2S3 nanowires inside the channels of AAMs by the thermolysis of a single source precursor. A high yield of pore filling, up to 80 %, of Bi2S3 nanowires inside the channels of the template can be achieved by injection of the melted precursor liquid into the channels followed by thermolysis. X-ray diffraction (XRD) and high-resolution transmission electron microscopy (HRTEM) studies suggest that highly oriented growth of the Bi2S3 nanowires, with respect to the channel direction of the AAM, can be achieved. Additionally, the mean diameter of the nanowires can readily be controlled by using AAMs with different pore diameters. The absorption profile (bandgap) of these nanowires was found to be blue-shifted to higher energies as the mean diameter of the nanowires decreased. Our approach utilises bismuth bis(diethyldithiocarbamate) [Bi(S2CNEt2)3], an organometallic molecule which has been used in the preparation of Bi2S3 thin films by metal organic chemical vapour deposition (MOCVD),37 as a single source precursor to deposit Bi2S3 nanowire arrays within the pores of AAM templates. The precursor has a melting point of approximately 200 °C, and decomposition temperature around 300 °C, which makes it an ideal liquid precursor, after melting, to be injected within the pores of AAMs. AAMs with mean pore diameters of approximately 80 and 20 nm were used as hard templates for the synthesis of arrays of Bi2S3 nanowires. An XRD pattern of the as-decomposed material on the membrane surface shows a number of reflections that can be indexed to the orthorhombic bismuthinite structure (reference pattern, JCPDS No. 170320; Figure 1 a)1. The XRD patterns from the polished membranes, in which the surface material has been removed leaving the material decomposed within the pores only, show one very high intensity reflection emanating from the (002) set of planes, at 48.2 2θ degrees, indicating that the Bi2S3 nanowires have a highly preferential orientation along the [002] direction (Figures 1 b 1and 1 c). This remarkable difference in the XRD characteristics of the Bi2S3 crystals on the surface vs the polished samples can be explained by the fact that in the reflection XRD experiment (θ–2θ) only the crystallographic planes parallel to the sample surface can be detected. Thus randomly oriented Bi2S3 crystals on the membrane surface give rise to powder-like patterns whereas the preferentially oriented nanowires inside the channels of the AAMs results in a reduced number of reflections, corresponding to planes parallel to the membrane surface, that is, the (200) reflection. XRD patterns of a) Bi2S3 crystals formed outside an AAM template and Bi2S3 nanowire arrays formed within AAMs with mean diameters of b) 80 nm and c) 20 nm. All samples exhibit an orthorhombic Bi2S3 crystal structure (bismuthinite, JCPDS file 17-0320) and the Bi2S3 nanowires within the templates have a preferred orientation along the [002] direction. Side-view scanning electron microscopy (SEM) images of the samples subjected to thermolysis at 300 °C reveal that continuous Bi2S3 nanowires protrude out of the channels of the AAMs upon breakage of the membranes (Figured 2 a 2and 2 b). The nanowires are embedded and restricted by the dimensions of the membrane channels, and at the same time isolated from each other inside the porous alumina matrix. The TEM images show that the liberated nanowires have lengths of up to tens of micrometres and mean diameters of approximately 80 nm and 20 nm, depending on the AAM used as the template (Figures 3 a 3and 3 b). Energy dispersive X-ray analysis (EDAX) shown in the Supporting Information reveals that the nanowires are composed of Bi and S (no Al was detected suggesting the template was completely removed). Quantitative analysis of the EDXA data indicates that the atomic composition of Bi:S in the nanowires is close to a 3:2 stoichiometry. The nanowires also display very smooth surfaces with very few defects. Small gaps between the nanowires and the walls of the columnar channels of the AAMs are observed in the plane-view TEM images (Figures 4 4a and 4 b). Nanowire growth inside the channels of the AAMs probably occurs through an anisotropic growth due to the highly anisotropic crystalline structure of Bi2S3. The role of the porous template is to separate and isolate the nanowires in an ordered array. The diameters of the nanowires formed were however often smaller than the dimensions of the pores in which they were grown. This observation is probably due to 1) a contraction in the volume of the material inside the pores in going from the bulky precursor to nanowires during the crystallisation process and 2) nonuniform filling of the pores in the templates with the melted precursor. Thus the nanowires show a little broader distribution of diameters than the channel size distribution of the templating membranes, such as 10–80 nm for the 80 nm membranes and 8–20 nm for the 20 nm membranes, respectively (the pore size and nanowire diameter distributions were calculated from at least five different plane-view TEM images). The total pore filling ratio, calculated by the plane view TEM images, is up to 80 %, with densities of 1.5×109 and 8.0×109 nanowires per cm2 for the 80 and 20 nm membranes respectively. Side-view SEM images of the Bi2S3 nanowire arrays within AAMs with mean pore diameters of a) 80 nm and b) 20 nm. TEM images of Bi2S3 nanowires liberated from AAMs with mean pore diameters of a) 80 nm and b) 20 nm. Plane-view TEM images of the Bi2S3 nanowire arrays within the pores of a) 80 nm and b) 20 nm AAMs. The crystal structure of the Bi2S3 nanowires was examined by selected area electron diffraction (SAED) and HRTEM combined with fast fourier transforms (FFTs). To help interpret the SAED and FFTs patterns, electron diffraction patterns were simulated with a crystal studio software package using appropriate orthorhombic unit cell dimensions with a Pbmn space group. Figure 5 5shows TEM images of isolated 60 and 15 nm Bi2S3 nanowires after removal from the membranes. The corresponding SAED patterns verify their single-crystalline nature, indicated by the sharp diffraction dots. The SAED patterns were taken with the electron beam along the [010] direction and confirmed that the nanowires were grown along the [002] direction (in line with XRD data). The HRTEM images shown in Figures 55 b and 5 d also provide evidence of the single-crystalline nature of the nanowires with clearly visible lattice fringes. For example in Figure 55 b, lattice fringes with a d-spacing of about 5.6 Å, attributed to the (200) lattice planes, are observed parallel to the growth direction of the nanowires. The FFTs of the HRTEM shown in Figures 55 b and 5 d further validate the [002] growth direction of the nanowires. All of the nanowires (about 10–15 nanowires) that were studied by SAED and FFTs of the HRTEM images, isolated from both types of membranes (80 and 20 nm pores), showed growth directions along the [002] zone axis. These data are in good agreement with the XRD results reported in the literature for single-crystalline Bi2S3 nanowires and nanorods grown by various methods.29, 34, 40 HRTEM images with the corresponding SAED and FFTs for isolated Bi2S3 nanowires with diameters of approximately 60 nm and 15 nm (insets: SAED and FFTs are indexed in the the Pbmn space group with the electron beam along the [010] direction, the arrows show the growth direction of the nanowires that is along the [002] zone axis). To further confirm the orientation of the nanowires inside the anodic alumina channels, Bi2S3 nanowires were imaged inside the AAMs (Figure 6) 6with the electron beam along the direction of the channels (parallel to the growth direction). The results show that both the FFTs of the HRTEM images and the SAED patterns can be indexed assuming the beam is incident along the [002] direction (several tens of SAED and FFTs were recorded and compared to simulated patterns). This result is consistent with XRD, SAED and HRTEM analysis presented herein, and further confirms the alignment of the single-crystalline Bi2S3 nanowires with a [002] growth direction parallel to the long axis of the channels of the AAMs. Plan-view HRTEM images of Bi2S3 nanowires inside the channels of the AAMs with mean pore diameters of a) 80 nm and b) 20 nm (inset is the corresponding SAED, indexed in the Pbmn space group with the electron beam along the [002] direction). Figure 7 a 7shows the optical properties of Bi2S3 powder, formed by the decomposition of the precursor in the absence of a template, Bi2S3 nanowires within the channels of AAMs and an empty membrane. No significant absorption arises from the unfilled AAM template. There is an obvious blue shift of the absorption edge for the Bi2S3 nanowires grown within small pore AAMs (mean diameter of 20 nm) compared with the Bi2S3 powders and Bi2S3 nanowires grown within the large pore AAMs (mean diameter of 80 nm). For a direct bandgap semiconductor such as Bi2S3, the bandgap energy is usually defined by the x-axis intercept of (αhν)2 versus hν, where α is the absorption coefficient.41, 42 Applied to the near-edge absorption for Bi2S3 nanowires, Figure 77 b shows the (αhν)2 versus hν plot obtained from curves 1 to 3 in Figure 77 a. Extrapolation of the plot to the energy axis gives a value of approximately 1.35 eV for the bandgap of the Bi2S3 powder. The bandgap energy of the Bi2S3 nanowires within the pores of the 80 and 20 nm AAMs was 1.36 and 1.51 eV respectively. The bandgap energies reported here lie between the energy range of 1.25–2.0 eV, which is similar to the energies previously reported for Bi2S3 thin films and nanowires.20, 43–45 The bigger optical bandgap (Eg) for the Bi2S3 nanowires grown within small pore AAMs (below 20 nm) may be related to quantum confinement effects within the nanowires at these dimensions. a) Absorption spectra of Bi2S3 powder, Bi2S3 nanowires inside AAM templates and a blank AAM membrane, b) (αhν)2 versus hν obtained from curves 1,2,3 shown in (a). In summary, a solventless, template-based approach has been developed for growing high density arrays of oriented single crystal Bi2S3 nanowires. The role of the porous structure is to separate and isolate these nanowires and guide their growth in an ordered array. A high loading of the melted precursor can be obtained in the channels of the AAMs, which can subsequently be crystallised to form continuous single crystal nanowires. XRD, SAED and HRTEM imaging provided evidence that the nanowires produced are preferentially oriented inside the channels of the AAMs along the [200] growth direction. By varying the pore size of the templates, the mean diameter of the nanowires could also be controlled. The absorption bandgap energies of the synthesised nanowires were found to increase as the mean diameter of the nanowires decreased. By choosing suitable precursors, single crystal nanowire arrays of other materials may also be prepared using this approach. Compared with other synthetic approaches, our solventless technique is a simple method for obtaining arrays of single crystal compound nanowires with a preferential growth direction. Sample Preparation: The single source precursor, bismuth bis(diethyldithiocarbamate) [Bi(S2CNEt2)3] was prepared by the reaction of Bi2O3 with carbon disulfide and diethylamine in methanol and was thoroughly purified by repeated recrystallisation as described by Monteiro et al.46 The AAMs were made in-house using a well known two-step anodisation procedure described elsewhere.4, 47 The synthesis of the single crystal Bi2S3 nanowire arrays was carried out in a flat bottom glass tube (inner diameter 1.4 cm). In a typical experiment, the organometallic precursor (≈300 mg) was put on the surface of the AAM in the glass tube. The tube was then heated to 200 °C under vacuum. After the precursor had melted, the temperature was maintained for another 30 min. The vacuum was then disconnected and the glass tube containing the sample was placed on a hot plate at 300 °C for 30 min until the Bi2S3 precursor had completely decomposed. The layer of Bi2S3 formed on the surface of the template was removed by polishing with 5 μm diamond paste and washing with acetone whilst sonicating. The back aluminium metal of the AAM was removed using a saturated copper (II) chloride solution. The templates were dissolved completely with a 6 M aqueous solution of sodium hydroxide at 60 °C. The resulting suspension was washed with deionised water several times. X-ray Diffraction Measurements: XRD measurements were performed with a Philips X′Pert diffractometer using Cu Kα1 radiation with an anode current of 35 mA and an accelerating current of 40 kV. The measurements were done in reflection (θ–2θ) geometry. Electron Microscopy: SEM images of the broken Bi2S3 filled AAM templates, put on conductive carbon tape, were obtained using a field-emission scanning electron microscope JEOL JSM 6700F at an accelerating voltage of 5 kV. The TEM and electron diffraction images were investigated in a JEOL JEM-2000FX microscope operated at 200 kV. The sample was dispersed in absolute ethanol, sonicated and dropped onto copper grids coated with a holey carbon film. The plane-view TEM samples were made after mechanical polishing both sides of the template, followed by Ar-ion milling until the sample was transparent for TEM imaging. High-resolution TEM images were recorded with a Joel JEM 2010 microscope operating at 200 kV. Optical Measurements: Absorption measurements on the samples, attached to a quartz plate with wax, were made using an Agilent 8453E UV/Vis spectrophotometer with a Labsphere RSA-HP-53 diffuse reflectance and transmittance accessory. The authors acknowledge financial support for their work from the Higher Education Authority (HEA) in Ireland under PRTLI 3 funding and Science Foundation Ireland (Grant 03/IN3/I375). We are grateful to Joseph Tobin for some TEM analysis and Brian Daly for useful discussions. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2267/2007/z600681_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.