Letter pubs.acs.org/JPCL
Epitaxial Heterostructures of Lead Selenide Quantum Dots on Hematite Nanowires Rachel S. Selinsky, Sanghun Shin, Mark A. Lukowski, and Song Jin* Department of Chemistry, University of Wisconsin−Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States S Supporting Information *
ABSTRACT: We present a novel method for synthesizing epitaxial quantum dotnanowire (QD-NW) heterostructures using the example of colloidal PbSe QDs decorated on furnace-grown hematite (α-Fe2O3) NWs. The direct heterogeneous nucleation of QDs on Fe2O3 NWs relies upon an aggressive surface dehydration of the as-synthesized Fe2O3 NWs at 350 °C under vacuum and subsequent introduction of colloidal reactants resulting in direct growth of PbSe QDs on Fe2O3. The synthesis is tunable: the QD diameter distribution and density of QDs on the NWs increase with increased dehydration time, and QD diameters and size distributions decrease with decreased injection temperature of the colloidal synthesis. Transmission electron microscopy (TEM) structural analysis reveals direct heteroepitaxial heterojunctions where the matching faces can be PbSe (002) and Fe2O3 (003) with their respective [11̅0] crystallographic directions aligned. This can be a general approach for integrating colloidal and furnace synthetic techniques, thus broadening possible material combinations for future high-quality, epitaxial nanoscale heterostructures for solar applications. SECTION: Physical Processes in Nanomaterials and Nanostructures
S
junctions through proximal contact of nanomaterials. Composite nanostructures were formed by separately synthesizing QD dispersions and nanostructure covered substrates. The QDs were then integrated by drop-casting or spin-coating the QD dispersions onto the decorated substrates. Examples of proximally contacted heterostructures include PbSe QDs infiltrated into a vertically oriented array of ZnO nanowires (NWs),14 PbSe QDs mixed with TiO2 nanobelts,15,16 PbSe QDs on TiO2 and SnO2 mesoporous films,17,18 and CuInS2 NCs, CdSe NCs, and InP QDs spin-coated onto TiO2 nanoparticle films.19−21 The lack of direct electrical connection between the two materials can result in charge transfer by an inefficient hopping mechanism. The second type of heterostructures is formed by purposefully linking nanoscale materials with an organic or inorganic molecule. Examples include PbS and PbSe, CdSe, CdTe, or Bi2 S3 QDs linked to TiO2 nanoparticles by 3-mercaptopropionic acid (MPA).17,22−27 These purposefully designed linkers can improve the electronic communication, but the lack of lattice connection could still pose a barrier to efficient charge transfer. Indeed the electron transfer rate depends on the nature and the length of the linker molecules.23,26 The third type is linker-free, epitaxially connected, nanocrystalline materials that are made through the direct growth of one crystalline material onto another. Examples include PbSe QDs epitaxially grown onto the surface of colloidal TiO2 nanorod (NR) suspensions or TiO2 NC paste
emiconductor heterostructures containing nanocrystals (NCs) or quantum dots (QDs) are of current technological interest for applications requiring charge transfer including solar energy conversion, electronic and optoelectronic devices, and catalysis.1−6 To effectively facilitate charge separation, suitably designed heterostructures should have both appropriate band gap alignment and a defect-free epitaxial interface.1,3,7,8 Both requirements can benefit from nanoscale specific effects, namely, band gap tunability and tolerance for high lattice mismatch. First, by reducing the dimensions of a material to the nanoscale, electrons and holes can be quantum confined resulting in an increase in band gap.9,10 As such, the band alignment of nanoscale heterostructures can be tuned by adjusting the sizes of the two component materials.2,11,12 Second, lattice mismatch can limit the combinations of materials possible for bulk semiconductor heterostructures, because one material can be grown on another with different lattice parameters, forming a coherent interface only to a critical thickness defined by the degree of mismatch. Beyond that thickness, dislocations will form to relieve elastic strain. In contrast, the size scale of nanomaterials is usually far below the critical thickness, and as such, nanoscale heterojunctions can tolerate large lattice mismatch before defects are observed.5,13 As a result, combinations of materials that could not be grown as bulk or thin film junctions can often be synthesized as nanoscale heterostructures with a single crystalline, defect-free interface. There have been several synthetic approaches that integrate
Received: May 24, 2012 Accepted: June 4, 2012 Published: June 4, 2012
colloidal QDs into semiconductor heterostructures, and some exhibit efficient charge transfer. The first approach forms © 2012 American Chemical Society
1649
dx.doi.org/10.1021/jz300672x | J. Phys. Chem. Lett. 2012, 3, 1649−1656
The Journal of Physical Chemistry Letters
Letter
thin films on conductive glass substrates.11,28,29 While the two components in the heterostructures have direct electronic communication, these colloidally suspended heterostructures or mesoporous composites are not directly connected to conductive substrates, which could limit their utility. There are also other noncolloidal synthesis methods for growing heterostructures, such as chemical vapor transport (CVT) growth of CdS NRs on TiO2 NWs30 and chemical vapor deposition (CVD) growth of PbS on TiO2, muscovite mica, and NaCl;31 successive ionic layer adsorption and reaction method (SILAR) growth of NCs,32−34 including CuInS2 and CdS on TiO2 nanotube arrays35 and SnS on CdS; 36 electrodeposition of Cu(In,Ga)Se2 on CdS NWs;37 electrodeposition of CdSe NPs on hydrothermal TiO2 NW arrays;38 chemical bath deposition of CdS and CdSe QDs and sonication assisted sequential chemical bath deposition (S-CBD) of CdS QDs on TiO2 NT arrays;39−43 microwave assisted chemical bath deposition (MACBD) of CdS on TiO2;37 and close space sublimation of CdS NPs on TiO2 NT arrays.44 However, vapor phase growth, SILAR, electrodeposition, and MACBD methods lack control over the NC sizes, dispersity, and morphologies. Here we present a method for the direct epitaxial growth of colloidal PbSe QDs onto furnace-grown α-Fe2O3 NWs affixed to a steel substrate. PbSe−Fe2O3 QD-NW heterostructures are potentially useful for solar energy conversion. PbSe is a small band gap (0.27 eV) semiconductor that is highly quantum confineable with a Bohr radius of 46 nm; however, the band gap is only tunable across the near-infrared region giving it limited coverage of the solar spectrum.45−49 Fe2O3 has the advantage of being nontoxic, stable, inexpensive, and abundant. The difficulty with Fe2O3 is its short exciton lifetime, small diffusion length, and poor conductivity. In addition, its band gap at 2.1 eV (590 nm) only covers part of the visible region of the solar spectrum.50−56 By combining PbSe QDs with vapor phase grown Fe2O3 NWs in heterostructures, we could take advantage of the benefits of each, while ameliorating some of their limitations. Nanocrystalline PbSe and Fe2O3 have a staggered band alignment (Scheme 1); by reducing the size of PbSe, its band gap width and relative band position can be tuned11,47 with respect to Fe2O350 to obtain an appropriate type II heterostructure for charge separation. The combination of the materials covers the solar spectrum more completely than either alone. In addition to band gap considerations, epitaxially grown PbSe QDs may improve Fe2O3 properties by passivating Fe2O3 surface defects and improving Fe2O3 NW conductivity.57−61 In this Letter, we developed a unique and facile method for direct nucleation and epitaxial growth of PbSe QDs on the surface of Fe2O3 NWs. This is achieved by first dehydrating the surface of the Fe2O3 NWs in a Schlenk line compatible flask and then immediately introducing colloidal PbSe precursors at temperatures sufficient to initiate nucleation. This method for integrating colloidal QDs into heterostructures has several advantages. First, the direct lattice connection leads to direct electrical connection from the QDs to the NWs. Second, unlike colloidal suspensions of heterostructures,11,29 the Fe2O3 NWs herein are anchored directly to a substrate, which may enhance their utility in device applications. Lastly, this method integrates a furnace NW growth method with a colloidal QD growth. Given the vast literature of both furnace-grown NWs62−67 and colloidal QDs,9,68 this method opens the door to many new possible materials combinations for nanoscale heterostructures.
Scheme 1. Band Position Schematic for Fe2O3 and PbSe QDs. Small Diameter, Quantum-Confined PbSe QDs, Display a Staggered Band Gap Alignment (Type II)a
a
The band positions and alignment of the PbSe and Fe2O3 are taken from the indicated references.11,47,50.
Our goal is to integrate tunable, high-yield colloidal QD and furnace NW growth methods to synthesize epitaxial heterostructures. Because heterogeneous nucleation on a surface should have a lower energy barrier than that required for homogeneous QD nucleation in solution, if NWs with clean surfaces are in the presence of high temperature colloidal reactants, the surfaces of these wires could act as preferential heterogeneous nucleation sites for QD growth. Our general strategy is to utilize this property to nucleate PbSe QDs directly on the surface of Fe2O3 NWs. The Fe2O3 NWs are prepared by following an improved thermal oxidation synthesis using steel substrates69 (a representative image is shown in Figure 1B). A steel substrate covered with as-synthesized Fe2O3 NWs is placed into a Schlenk flask compatible glass holder that consists of a glass tube with an angled slit at the bottom into which the substrate is inserted (Figure 1A). This allows the NW substrate to be suspended within a vigorously stirring mixture of colloidal PbSe reactants without significantly damaging the substrate. This holder is then placed in a three-necked cylindrical flask on a Schlenk line and a cooled mixture of lead oleate, trioctylphosphine, and diphenyl ether is added under inert atmosphere. Next, these reactants are heated to a temperature sufficient for PbSe nucleation. The Se precursor, trioctylphosphine selenide (TOP-Se), is injected, resulting in a darkening of the solution surrounding the NW substrate indicating homogeneous PbSe nucleation.48 However, initial attempts failed to grow PbSe−Fe2O3 heterostructures by simply adding the PbSe precursors directly to as-synthesized Fe2O3 NWs. No PbSe NCs were observed on the NW surface through scanning electron microscopy (SEM) (Figure 2A) or through transmission electron microscopy (TEM). In addition, no lead signature was observed by electron dispersive spectroscopy (EDS) measurement of individual NWs in the TEM. Since PbSe NCs were indeed homogeneously nucleated in solution during the course of the reaction, this implied that there was a barrier preventing the direct heterogeneous nucleation of PbSe on the Fe2O3 surface. Due 1650
dx.doi.org/10.1021/jz300672x | J. Phys. Chem. Lett. 2012, 3, 1649−1656
The Journal of Physical Chemistry Letters
Letter
despite a water vapor pressure of 1 Torr. As we utilize both a higher temperature and lower pressure than the Nilsson report, our dehydration procedure should remove surface water effectively.70,71 The dehydrated NW substrate was kept in the same Schlenk flask, and the QD growth was then immediately carried out by transferring the pretreated reactants from another flask. Such procedures successfully led to the nucleation and growth of PbSe QDs on the surface of Fe2O3 NWs. Once we determined a method for synthesizing these heterostructures, our next goal was to determine the parameters for structural tunability. We tested the dehydration time dependence on the resulting heterostructures. The dehydration was carried out for 0, 1, 2, and 5 h, while the rest of the synthetic conditions were kept constant (125 °C injection temperature and 2 min reaction time). The representative SEM images of these reaction products (Figure 2A−2D) clearly show that longer dehydration times result in a decrease in the size distribution of the PbSe QDs and an increase in the density of PbSe QDs per unit NW surface area (Figure 2). The average diameters and QD densities from statistical analysis of the SEM images are displayed in Figure 2E. (The QD diameter is defined as the cross sectional size projected onto the heterointerface as some QDs may not be perfectly spherical.) The ability to tune the diameter and thus the band gap of PbSe QDs is important for achieving the desired type-II PbSe−Fe2O3 band alignment as seen in Scheme 1. Since longer dehydration times were correlated with narrower diameter distributions, longer dehydration times are preferable as they allow for tighter band gap control. In addition to the tunability of the surface coverage and the distribution of PbSe QD diameters, proper control of the band alignment of Fe2O3 and PbSe requires PbSe diameter control. As found in traditional PbSe QD synthesis, injection temperature strongly influences QD diameters. As shown in Figure 3, QD diameter and diameter distributions increase with the nucleation temperature (NW reaction conditions are listed in Figure 3). An injection temperature of 110 and 195 °C resulted in QD diameters of 8.7 ± 1.3 nm and 27.5 ± 9.9 nm, respectively. The ability to tune across such a wide range of sizes gives us access to a broad range of PbSe band gaps and, as a result, allows tuning of the PbSe−Fe2O3 band alignment (Scheme 1). Although the PbSe nucleation temperature-diameter trend mirrors the trend observed for homogeneously nucleated PbSe QDs, the diameters of the heterogeneously nucleated PbSe
Figure 1. (A) The illustration shows the glassware designed and utilized for this heterostructure synthesis with (B) a representative SEM image of as-prepared Fe2O3 NWs on the substrate.
to the hydrophilic nature of the Fe2O3 surface, as reported in the literature,70−74 we expect this barrier to be surface adsorbed water. Given this hypothesis, the Fe2O3 surface must be dehydrated before the QD growth. Therefore, the key to the successful growth of PbSe−Fe2O3 heterostructures is an aggressive in situ surface dehydration. This was accomplished by placing the Fe2O3 NW substrate in the glassware as described above (Figure 1A) and heating it in a sand bath at 350 °C on the Schlenk line under dynamic vacuum (
