Controlled Synthesis of Ferromagnetic Semiconducting Silicon

Mar 27, 2012 - Recently, Mu et al. reported a multistep template replication ..... solution containing NiSO4 (0.4 M, Sigma Inc.), H3BO3 (0.6 M), and N...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Controlled Synthesis of Ferromagnetic Semiconducting Silicon Nanotubes Nava Shpaisman, Uri Givan, Moria Kwiat, Alexander Pevzner, Roey Elnathan, and Fernando Patolsky* School of Chemistry, The Raymond and Beverly Sackler Faculty of Exact Sciences, Tel-Aviv University, Tel Aviv 69978, Israel ABSTRACT: Recently, transition-metal-doped semiconductor nanostructures, so-called diluted magnetic semiconductors, such as dots, rods, wires, and films, have been the subject of intense research efforts due to their fascinating properties and potential applications in bioimaging, spintronics, and quantum interference information processing. Here, we present a method for synthesizing superdiluted Ni-doped ferromagnetic silicon nanotubes (SiNTs) (with room-temperature ferromagnetism), with minimal synthetic steps and with maximal control of the resultant SiNTs structure and composition. The unique advantage of our approach is the simplicity that provides us precise control of the ferromagnetic SiNT parameters, length, outer and inner diameter, wall thickness, Ni concentration, and crystallinity, by changing the template membrane (pore diameter), dipping time in the catalyst, growth time, and decomposition temperature. Numerous combinations of SiNT parameters can therefore be prepared that can influence their magnetic and electronic properties. This level of control can lead to novel future nanoelectronic and nanospintronic devices.

S

materials. Interest in ferromagnetic semiconductors (FMSs) was revived with the discovery of spontaneous ferromagnetic order in In1−xMnxAs and in Ga1−xMnxAs,20−23 when FM properties were realized in semiconductor hosts already widely recognized for semiconductor device applications. These new FMS materials exhibit Curie temperatures up to 35 and 173 K, respectively, for Mn concentrations of the order of 5−15% and sufficiently high hole densities, and have been closely studied for their potential applications in future spin-dependent semiconductor device technologies.24 Recently, nonmetallic diluted magnetic semiconductors (DMSs), based on group-IV semiconductors, attracted special attention due to their reliability for a straightforward integration into Si-based electronics, relatively high Curie temperature, and specific magnetic and transport properties compared to metallic GaMnAs DMS.25−36 Diluted magnetic semiconducting nanomaterials are considered favorable candidates to improve the efficiency of spin injection from ferromagnetic materials to semiconductors. These materials are important for developing spin-controlled electronic device and integrated magnetic devices applications.37 Diluted magnetic semiconductor materials promise the possibility of combining the advantage of semiconductor band gap engineering with controllable magnetic properties in a single semiconductor device. Thus, the emerging field of spintronics would be dramatically boosted if room-temperature ferromagnetism could be added to semiconductor nanostructures that are compatible with silicon technology. The achievement of silicon nanostructure-based spintronics38−42 is believed to be significant due to the long spin-coherence lifetime,43 well-developed materials pro-

ince the discovery of inorganic nanowires (NWs) and carbon nanotubes (CNTs), nanoscale one-dimensional (1D) materials have attracted great attention, both scientifically and technologically.1−6 The growth of reproducible and controllable nanotubes with uniform inner diameter and wall thickness, and controllable morphology, shape, and chemical composition, would be advantageous in potential nanoscale electronics, biochemical-sensing applications, and fluid-transport devices. In this regard, silicon nanotubes are of great importance. Silicon has been widely recognized as the most important material of the 20th century. This is largely due to its role as a fundamental component in integrated circuits and consequently in the microelectronics revolution. Recently, a few groups have reported on the synthesis of silicon nanotubes,7−15 each using a different growth process. Mesoporous materials have been used as templates for the synthesis of a variety of nanotubes.16−18 The template-modulated method has an advantage in obtaining well-aligned and ordered arrays of nanotubes with uniform tube dimensions. Silicon nanotubes (SiNTs) have been synthesized by Sha et al. using the Au catalyst-assisted vapor−liquid−solid (VLS) process, into the nanochannels Al2O3 substrate.16 This method is, however, uncontrollable since both Si NWs and SiNTs grow simultaneously. Recently, Mu et al. reported a multistep template replication approach for the fabrication of uniform SiNT arrays. Nevertheless, this method is time-consuming and employs a large number of synthetic steps.18 Ferromagnetic semiconductors are materials that simultaneously exhibit semiconducting properties and spontaneous long-range ferromagnetic ordering. Classic examples comprise the europium chalcogenides and the chalcogenide spinels.19 Nevertheless, device applications have languished because of low Curie temperatures and the inability to incorporate these materials in thin films with mainstream semiconductor device © 2012 American Chemical Society

Received: April 24, 2011 Revised: February 12, 2012 Published: March 27, 2012 8000

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

NiRed/Al2O3 clusters on the membrane surface or alternatively to form NixSiy (nickel silicide) crystalline clusters. The NiRed/ Al2O3 (or NiSi/Al2O3) clusters serve as catalysts for further SiH4 decomposition resulting in absorption of Si onto the membrane’s nanochannel inner surfaces, in a manner similar to the decomposition of methane during CNT growth.44−47 The absorbed Si serves in turn to catalyze the decomposition of SiH4 and encourages the continuous growth of the SiNTs. Nickel atoms can diffuse into the forming silicon layer, spreading evenly into the SiNT walls volume, at the growth temperatures between 450 and 650 °C. Such a growth mechanism is comparable to the catalyst-free SiNW growth. For instance, in the oxide-assisted growth mechanism, the oxide serves only as a nucleation center, and the continuous growth is achieved by the dangling bonds of the reactive Si atoms on the nucleus surface.49 Figure 1A shows representative top-view scanning electron microscopy (SEM) images of the SiNTs grown for 10 min, at 450 °C, after partially dissolving the AAO membrane template. The SiNT attachment to the membrane’s walls implies that the SiNT growth occurs with the assistance of the Al2O3 (Figure 1B). Once the AAO membrane is completely dissolved, the SiNTs remain well aligned and highly ordered, supported by a silicon thin film decomposed on top of the membrane (Figure 1E). Figure 1D shows the dispersed SiNTs after their release from the Si base film by sonication and dispersed onto a Si wafer. The SiNT outer diameter and length are ca. 250 nm and 60 μm, respectively. These values are comparable to the nominal thickness and pore size of the AAO membrane template. In addition, both parameters (length and outer diameter) depend only on the AAO characteristics (regardless of growth parameters such as growth time). Hence, those two parameters are readily controlled by choosing the membrane’s pore size and length. Furthermore, dipping the AAO template membrane in the NiSO4 solution for various duration times prior to the growth process leads to SiNTs containing different Ni concentrations. Please note that EDX chemical analysis performed by TEM and SEM did not reveal Ni residues inside the SiNTs, showing apparently pure silicon- or metal-free nanotubes. These analytical techniques are not sensitive enough to detect Ni concentrations below 1% in silicon; thus, based on these results it could be easily and erroneously assumed that the obtained SiNTs are Ni-free. Thus, ultrasensitive chemical composition analysis of the SiNTs was carried out by employing time-offlight secondary ion mass spectroscopy (TOF-SIMS) analysis. To eliminate the potential AAO membrane interference, the measurements were preformed on dispersed SiNTs, which were deposited onto a GaAs substrate (silicon-free substrate), after dissolving the AAO membrane in a solution containing both chromic and phosphoric acids. The Ni compositional analysis shown in Figure 2 clearly reveals that the SiNTs indeed contain Ni and that the Ni concentration increases with the dipping duration in the NiSO4 solution. Importantly, depth analysis performed by TOF-SIMS reveals a homogeneous distribution of the Ni element over the whole thickness of the NT walls, as well as along the NT length. Thus, Ni atoms are well distributed into the whole volume of the silicon walls and not only on the outer surface of the nanotubes, as originally assumed. Clearly, these low concentrations of Ni doping cannot be detected by the lowsensitivity EDX measurements. Ni atoms, originally adsorbed on the alumina membrane surface, diffuse into the growing walls of the SiNTs during the deposition process.50 Thus, the

cessing, and the compatibility with the active electronic devices. Nevertheless, to the best of our knowledge, no work on semiconducting silicon-based room-temperature ferromagnetic nanotubes has been yet demonstrated. Herein, we present an easy, predictable, single-step, wellcontrolled template-based process for producing uniform roomtemperature ferromagnetic crystalline Ni-implanted SiNTs, employing Ni solution-dipped anodized aluminum oxide (AAO). The SiNTs have been produced by a conventional CVD technique (at temperatures below the Si−Ni eutectic point). The use of Ni/Al2O3 as catalyst for methane decomposition is a well-known route for the growth of CNTs.44−47 Also, metal nickel (and copper) substrates were shown to act as an effective catalyst for the decomposition of gaseous silane and germane precursors at a broad range of temperatures.48 We found that Ni/Al2O3, simply by dipping the AAO membrane in a Ni solution, served as an excellent catalyst for the chemical decomposition of silane onto the AAO pore surface, thus enabling the controllable formation of SiNTs. In our approach, the Ni-dipped AAO membrane holds a double task in the formation of the SiNTs, serving both as the template structure and as the catalyst that triggers the SiNT nucleation and growth. We would like to point out that neither SiNTs nor SiNWs were found to grow at a growth temperature of 450 °C and even 650 °C with metal Ni as catalyst (using membranes with an evaporated Ni film) since the eutectic temperature of Ni/Si is 963 °C which is higher than the growth temperatures used in this study. We therefore conclude that the growth process of the SiNTs via our approach is different from that of the conventional VLS mechanism.1 The obtained SiNTs have a uniform wall thickness that varies from a few nanometers to several dozens of nanometers (depending on growth parameters) along the entire tube’s length (up to 60 μm). These highly ordered and well-aligned vertical SiNT arrays can be used in many applications as is, or they can be further dispersed on planar substrates, while maintaining their high aspect ratio due to their rigidness. This synthetic approach enables precise control of diameter, wall thickness, length, crystallinity, and magnetic/ electrical characteristic of the resultant SiNTs. Furthermore, our template-assisted growth approach enables the further electrochemical filling of the SiNTs with various materials such as metals or polymers, leading to novel building blocks with intriguing future potential applications. All SiNTs produced in this work were grown in a home-built, computer-controlled, hot-wall CVD setup using commercial AAO membranes as a template, high-purity silane (99.999% SiH4) as precursor gas, and high-purity Argon (99.9999%) as carrier gas. AAO membranes with pores diameter ranging from 20 to 200 nm were dipped in a basic NiSO4 solution (0.4 M) for 1, 15, 180, and 960 min, washed, dried, and finally attached to the Si substrate with carbon tape and placed at the center of a horizontal quartz tube in a hot-wall furnace reactor. A pregrowth vacuuming and purging step was performed. The growth step of all SiNTs was carried out at either 450 or 650 °C, and growth duration varied between 10 and 120 min. For n- or p-doped SiNT growth, either PH3 or B2H6 was added, respectively. Finally, the AAO membrane was chemically dissolved revealing the free-standing SiNTs. A schematic illustration of our proposed growth approach is shown in Figure 1. First, Ni2+ ions are absorbed on the Al2O3 pore surface as a result of dipping the AAO membrane in a basic NiSO4 solution, a process in which Ni(OH)2/Al2O3 clusters are formed. During the growth process, once SiH4 starts to flow at a temperature of 450 °C, it reduces the absorbed Ni2+ to form 8001

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

Figure 1. Schematic illustration of the synthesis process of ferromagnetic SiNTs. (1) Ni ions are attached onto the walls of the AAO membrane pores. (2) Reduction of the Ni ions to form Ni/Al2O3 clusters. (3) Decomposition of SiH4 on the formed Ni/Al2O3 cluster to form the SiNT wall. Top- and side-view SEM images of the ferromagnetic SiNTs grown for 10 min at 450 °C, after partial (A−B) and complete (C and E) dissolution of the AAO membrane. (D) SEM picture of dispersed SiNTs.

(TEM) images of the obtained SiNTs grown for 10, 30, 60, and 120 min (Figure 3A−D), exhibiting uniform wall thickness of about 10, 17, 25, and 40 nm, respectively. Figure 3E shows the linear correlation between the wall thickness and the silicon deposition time. Notably, SiNTs could be grown directly on the AAO membrane without the Ni catalyst. Nevertheless, the Ni catalyst-free growth does not yield SiNTs for growth periods of less than 60 min and was ca. 10-fold slower than the Ni-catalyzed process. This result indicates that the presence of Ni indeed accelerates the growth process by catalyzing the silane decomposition as per our suggested growth mechanism. The crystalline structure of the SiNTs was investigated employing high-resolution TEM (HRTEM) and Fast Fourier Transform analysis (FFT). The SiNTs prepared at 450 °C for 10 min (15 min dipping in NiSO4 solution) were found to possess an amorphous structure, as shown in Figure 4A. The amorphous structure was further confirmed by FFT measurements as presented in the Figure 4B inset. Nevertheless, nanosized crystalline grains are found in these nanotubes, which might indicate a silicon growth catalyzed by the Ni catalyst via the vapor solid solid (VSS) mechanism (Figure 4B). Figure 4C represents SiNTs grown at 650 °C (identical dipping and growth time conditions as per Figure 4A). At this growth temperature, the SiNTs have a polycrystalline structure, as revealed in Figure 4D and by the FFT analysis in the Figure 4D inset. It should be noted that polycrystalline NTs were also achieved by annealing the amorphous SiNTs for 10 min at 650 °C. In all cases, HRTEM and FFT analysis of the nanocrystalline domains

Figure 2. Ni compositional profile of the Ni-doped SiNTs, at different periods of incubation of the membrane template in Ni2+ solution, probed by time-of-flight secondary-ion mass spectroscopy (TOF-SIMS) measurements. Insets: (top) Depth compositional analysis by TOF-SIMS (red curve) nickel counts and (black curve) silicon counts, and (bottom) depth-dependent Ni concentration derived from the top inset.

concentration of Ni obtained is homogeneous through the nanotubes walls (Figure 2 insets). Also, control over the wall thickness was achieved by adjusting the SiNTs’ deposition time. SiNTs have been grown for 10, 30, 60, and 120 min using an AAO membrane previously dipped 15 min in a NiSO4 solution. Figure 3 shows a series of high-resolution transmission electron microscopy 8002

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

In all HR-TEM experiments performed, we have found no evidence for the presence of Ni-rich nanodomains or NiSi phases, neither on the nanotube surfaces nor in their bulk. Additionally, the depth analysis from TOF-SIMS measurements shows that Ni impurities, whether single atoms or atomic clusters, are present and distributed all over the bulk of the silicon nanotubes, and not only, or mostly, on the surface of the nanostructures. Several cycles of TOF-SIMS measurements, with etching steps in between, clearly show that the percent of Ni in the first measurement cycle (nanotubes surface) is only 10−12% higher than the values obtained for the Ni percent in the following cycles of etching along the depth of nanotubes. This means that, regardless of the surface coverage of nanotubes on the GaAs surface and the distribution of nanowire dimensions in the sample (the diameter and wall thickness distribution of nanotubes is very narrow), surface-adsorbed nanotube walls are etched similarly during the TOF-SIMS analysis cycles, from their outer shell through the Si bulk material forming the wall of the tubes. Thus, the measured Ni/Si ratio by TOF-SIMS depth analysis is not influenced by the dimensions of the nanotubes on the sample and by their small size distribution. Finally, due to the small amounts of Ni present in our nanotubular structures, we cannot completely rule out the presence of a Ni-rich cluster. Comparing the SiNTs grown at different temperatures reveals major differences of the wall thickness due to growth temperature dependence. The wall thickness is two times thicker (about 20 nm) for SiNTs grown at 650 °C as compared to those grown at 450 °C under the same growth conditions. The prominent dependence of the SiNT wall thickness upon growth temperature implies the growth to be dominated by the pyrolitic decomposition of the SiH4 gas precursor. As pointed out before, the SiNTs obtained by this method contain Ni homogeneously distributed over the whole nanotube volume, for which the amount can be easily controlled by the dipping time of the AAO membrane in the NiSO4 solution. Figure 5A represents the isothermal magnetization curves versus applied field (M(H)) at 298 K for the SiNTs containing different doses of Ni (0, 2.15 × 1012, 3.11 × 1012 atoms/cm2). It is readily observed that the M(H) curves for all Ni-free SiNT samples (prepared without the use of Ni as catalyst for the NT growth) exhibit diamagnetic behavior owning to the bulk Si. Nevertheless, a hysteresis loop observed at low field (low inset in Figure 5A) for the samples containing 2.15 × 1012 and 3.11 × 1012 atoms/cm2 of Ni indicates a small signal of bulk ferromagnetism behavior belonging to the Ni doping. Figure 5B shows the curve of magnetization versus applied field at 5 K of all discussed samples. Despite the diamagnetic behavior observed in the Ni-free SiNTs, a well-defined hysteresis loop exhibiting a ferromagnetic behavior is observed for the samples with 2.15 × 1012 and 3.11 × 1012 atoms/cm2 of Ni. Furthermore, the sample with the highest Ni concentration (3.11 × 1012 atoms/cm2) exhibits a pure bulk ferromagnetic behavior, with no diamagnetic residue. Figure 5B clearly demonstrates the increase of the saturation magnetization with the increase in the Ni concentration. The saturation magnetizations of the samples which contain 2.15 × 1012 and 3.11 × 1012 atoms/ cm2 of Ni are 6.3 × 10−5 and 3.3 × 10−4 emu g−1, and the coercivity of those samples are 811 and 714 Oe, respectively. These results indicate that long-range ferromagnetic coupling of the magnetic moments is sensitive to the Ni concentration. The temperature-dependent remnant magnetization curves of ferromagnetic SiNTs with the highest Ni concentration, between 5 and

Figure 3. TEM images of the edge of representative ferromagnetic Nidoped SiNTs grown at 450 °C for 10, 30, 60, and 120 min (A−D); scale bars are 20 nm. (E) Wall thickness of resulting SiNTs as a function of the deposition time.

Figure 4. HRTEM images of the ferromagnetic SiNTs prepared at 450 °C (A−B) and at 650 °C (C−D); scale bar 50 nm for A and C, 2 nm for B, and 5 nm for D.

only reveal the presence of cubic phase silicon (from this analysis, it was seen that the evaluated d-spacing value, of ∼0.314, 0.191, and 0.121 nm, fit well to the expected values for the interplanar distances of {111}, {220}, and {422} planes of fcc silicon). 8003

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

Figure 5. M(H) curves of SiNTs containing different concentration of Ni: 0 (Blue), 2.15 × 1012 (Red), and 3.11 × 1012 (Black) atom/cm2, measured at 298 K (A) and 5 K (B). Close look of the low-field region (low inset). M(T) curves of the ferromagnetic SiNTs containing 3.11 × 1012 atom/cm2 of Ni measured by applying a field of 500 Oe (upper inset). Figure 6. (A) Electrical measurements of ferromagnetic undoped SiNTs and n- (phosphorus) and p-doped (boron) ferromagnetic SiNT vertical arrays (1:4000). (B) Representative Isd vs Vg for a (a) backgated p-type boron-doped ferromagnet with a nickel concentration of 0.15% (SiNT sample annealed at 650 °C for 2 h), (b) an undoped SiNT FET, and (c) undoped and Ni-free SiNTs. All curves were recorded at Vsd = −5 V.

330 K, were obtained under the field-cooled (FC) mode at an applied magnetic field of 500 Oe (top inset of Figure 5A). The Curie temperature for this sample was found to be near room temperature, ∼260−280 K. Interestingly, a clear ferromagnetic behavior is observed for the Ni-doped SiNTs with only ∼0.2% Ni content. It is important to note that no segregation of metal nickel, as well as no formation of nickel or nickel silicide nanoparticles on the surface of SiNTs, is observed by HRTEM measurements. Thus, this is the first report showing ferromagmetic behavior of silicon 1D nanostructures doped with trace concentrations (∼0.2% Ni) of transition metal atoms. In spintronics applications, magnetic properties (remnant states and spin-polarized carriers) may be conveyed by conventional metallic ferromagnets or by so-called ferromagnetic semiconductors (FMS). In addition, because ferromagnetism is driven by carriers in FMS, magnetic properties may be controlled by an electrical field through the application of a gate voltage. Among many possible applications, carrier-controlled ferromagnetism, spin injection into nanostructures, and spin collection could permit giant magnetoresistance-type memories, field sensors, spin transistors, and reconfigurable logics. The integration of ferromagnetic SiNTs as building blocks for the fabrication of electrical devices requires control over their electrical properties. Such control was achieved by doping with either phosphorus or boron via coflow of phosphine (PH3) or diborane (B2H6) with silane during their growth.12,13 Thus, vertical SiNT-based devices were made by evaporating a Ni bottom and upper contact on the growth membrane where each circular upper electrode covers about 4 × 104 NTs (see Experimental Section for details). The I/V curves of 1/4000 (2.5 × 10−4) p-type, n-type, and nominally undoped SiNTs present in Figure 6A verify the success of dopant insertion

into the ferromagnetic SiNTs. The phosphorus-doped and boron-doped I/V curves exhibit, as expected, reduction in the resistance of almost two orders of magnitude in comparison to the undoped SiNTs. Also, single SiNT field-effect transistors were fabricated on silicon wafers as previously described.13,51 Undoped SiNTs, Ni-doped SiNTs (3.11 atoms/cm2), and boron-doped NidopedSiNTs were used as building blocks for the fabrication of two-terminal nanowire electrical devices. Figure 6B, curve (a), shows the representative electrical characteristics of a p-type boron-doped ferromagnetic SiNTbased field-effect transitor device (SiNTs were annealed at 650 °C postsynthesis to crystallize the structures prior to the fabrication of devices). Clearly, these devices show the expected behavior for the p-type FETs, as previously observed for SiNW and SiNT electrical devices.13,52 The doping of NTs with Ni atoms does not negatively influence the electrical performance of the resulting devices. Ni-doped nanotubes show a slight p-type behavior (Figure 6B, curve (b)), presumably as Ni atoms behave as acceptors similar to boron dopants in boron-doped silicon devices. Intrinsic SiNTs (Figure 6B, curve (c)) are not electrically active, and no clear gating effect is observed. Furthermore, SiNTs fabricated by our approach benefit the advantage of being electrochemically filled by various materials. For example, Figure 7 shows SEM and TEM images of Au-filled SiNTs and demonstrates the potential for novel applications. 8004

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

metal electrode and (ii) the isolation of the silicon walls from the metal-filled core by a dielectrics thin layer for the future fabrication of gate-embedded nanotransistors. To summarize, we have developed a method for synthesizing ferromagnetic SiNTs with minimal synthetic steps and with maximal control of the resultant SiNT structure and composition. The unique advantage of our approach is the simplicity that provides us precise control of the ferromagnetic SiNT parameters, such as length, outer and inner diameter, wall thickness, Ni concentration, and crystallinity, by changing the membrane (pore diameter), dipping time in the catalyst, growth time, and decomposition temperature. Numerous combinations of SiNT parameters can therefore be prepared that can influence their magnetic and electronic properties. This level of control can lead to novel nanoelectronic and nanospintronic devices.

■

EXPERIMENTAL SECTION Ferromagnetic SiNTs were prepared using AAO membranes with a specified pore diameter of 20−200 nm and a thickness of 60 μm and were purchased from Whatman Inc. (Cat. No.: 6809−7003). The AAO membrane was dipped in a basic solution containing NiSO4 (0.4 M, Sigma Inc.), H3BO3 (0.6 M), and NaOH (3 mM) for 1, 15, 180, and 960 min. The AAO membrane was then washed with deionized water, dried, and finally attached to the Si substrate with carbon tape which was placed at the center of a horizontal quartz tube in a hot-wall furnace. A pregrowth step included vacuuming and purging with Ar. The growth step of all ferromagnetic SiNTs was carried out at either 450 or 650 °C, with the growth duration varied between 10 and 120 min, using high purity 99.999% SiH4 (5 sccm) as precursor and 99.9999% Argon (10 sccm) as carrier gas, at a growth pressure of 30−50 Torr. For n- or pdoped SiNT growth, either PH3 (1.25 sccm) or B2H6 (5 sccm) was added, respectively. Finally, the AAO membrane was removed using a mixture of 1.8 wt % chromic acid and 6 wt % phosphoric acid at rt for 24 h, followed by washing with water and ethanol. After the residue was washed, the tube was sonicated 10 s to release and suspend the ferromagnetic SiNTs. Apparatus. FEG environmental SEM (FEG-ESEM, Quanta 200) and FEG-HRTEM (Philips Tecnai F20) were used. Energy-dispersive X-ray analysis (EDS) of the nanowires was performed with an EDAX acquisition system (Ametek Inc., Mahwah) installed on the SEM imaging instrument. Chemical composition analysis was carried out employing time-of-flight secondary ion mass spectroscopy (TOF-SIMS) analysis. Electrical measurements were made employing a probe station (Janis, ST-500) connected to a DAQ card (National Instruments, PCIMIO-16XE) by a computer-controlled, rack-mounted breakout accessory (National Instruments, BNC 2090) via a preamplifier (DL 1211). Magnetic measurements were carried out by a superconducting quantum interference device (SQUID) magnetometer. Potential controlled experiments were performed using a CHI621 A potentiostat (CH Instruments, Austin, TX). TOF-SIMS analysis was performed by a time-of-flight secondary ion mass spectrometer (PHI TRIFT II, PHI Corp.), on a gallium arsenide wafer containing deposited silicon nanotubes. The Ga+ primary ion beam was operated at 25 keV and 20 nA. Positive secondary ion spectra were acquired from 50 × 50 μm2 areas of the surfaces. Sputtering was performed by Ga+ ions for 1 s followed by spectrum acquisition. The sputtering area is 150 × 150 μm2. Preparation of Vertical Devices and IV Measurements. Multiple SiNT measurements were carried out in the AAO

Figure 7. Backscattered SEM images (A; top view, B) and TEM images (C) of Au-filled SiNTs.

To an AAO membrane containing SiNTs with a 40 nm wall thickness, Au was electrochemically deposited. The contrast between the Au core and the Si shell can be clearly shown by the backscattered SEM images (Figures 7A−B). Furthermore, from the top-view image (Figure 7A), it can be seen that the Au fills the entire SiNTs. To achieve this goal, the interior walls of the SiNTs can be selectively oxidized, at 800 °C in a 100 Torr O2 atmosphere, before the nanotubes are electrochemically filled with metal cores. This oxidation step serves two important purposes: (i) the electrical isolation of silicon nanotube walls from the metal electrodeposition solution to prevent deposition of metal directly from the inner walls and only from the bottom 8005

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

(8) Seifert, G.; Kohler, T.; Urbassek, H. M.; Hernandez, E.; FrauenHeim, T. Tubular Structures of Silicon. Phys. Rev. B 2001, 63, 193409. (9) Zhang, M.; Kan, Y. H.; Zang, Q. J.; Su, Z. M.; Wang, R. S. Why Silicon Nanotubes Stably Exist in Armchair Structure? Chem. Phys. Lett. 2003, 379, 81−86. (10) Bai, J.; Zeng, X. C.; Tanaka, H.; Zeng, J. Y. Metallic SingleWalled Silicon Nanotubes. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 2664−2668. (11) Hu, J. Q.; Bando, Y.; Liu, Z. W.; Zhan, J. H.; Golberg, D.; Sekiguchi, T. Synthesis of Crystalline Silicon Tubular Nanostructures with ZnS Nanowires as Removable Templates. Angew. Chem., Int. Ed. 2004, 43, 63−66. (12) Ben Ishai, M.; Patolsky, F. Shape and Dimension-Controlled Single-Crystalline Silicon and SiGe Nanotubes: Towards NanoFluidic FET Devices. J .Am .Chem. Soc. 2009, 131, 3679−3689. (13) Ben Ishai, M.; Patolsky, F. Tube-in-Tube and Wire-in-Tube NanoBuilding Blocks: Towards the Realization of Multifunctional Nanoelectronic Devices. Angew. Chem., Int. Ed. 2009, 46, 8699−8702. (14) Ben Ishai, M.; Patolsky, F. Wall-Selective Chemical Alteration of Silicon Nanotube Molecular Carriers. J. Am. Chem. Soc. 2011, 133, 1545−1552. (15) Ben Ishai, M.; Patolsky, F. From Crystalline Germanium-Silicon Axial Heterostructures to Silicon Nanowire-Nanotubes. Nano Lett. 2012, DOI: 10.1021/nl204263k. (16) Sha, J.; Niu, J.; Ma, X.; Xu, J.; Zhang, X.; Yang, Q.; Yang, D. Silicon Nanotubes. Adv. Mater. 2002, 14, 1219−1221. (17) Mu, C.; Zhao, Q.; Xu, D.; Zhuang, Q.; Shao, Y. Silicon Nanotube Array/Gold Electrode for Direct Electrochemistry of Cytochrome c. J. Phys. Chem. B 2007, 111, 1491−1495. (18) Mu, C.; Yu, Y. X.; Liao, W.; Zhao, X. S.; Xu, D. S.; Chen, X. H.; Yu, D. P. Controlling growth and field emission properties of silicon nanotube arrays by multistep template replication and chemical vapor deposition. Appl. Phys. Lett. 2005, 87, 113104. (19) Haas, C. Spin Disorder Scattering and Transport Properties of Ferromagnetic Semiconductors of Chalcogenide-Spinel Type. Bull. Am. Phys. Soc. 1970, 15, 157−164. (20) Motsnyi, V. F.; De Boeck, J.; Das, J.; Van Roy, W.; Borghs, G.; Goovaerts, E.; Safarov, V. I. Electrical Spin Injection in a Ferromagnet/ Tunnel Barrier/Semiconductor Heterostructure. App. Phys. Lett. 2002, 81, 265−267. (21) Ohno, H.; Shen, A.; Matsukura, F.; Oiwa, A.; Endo, A.; Katsumoto, S.; Iye, Y. (Ga,Mn)As: A New Diluted Magnetic Semiconductor Based on GaAs. Appl. Phys. Lett. 1996, 363−365. (22) Ohno, H. Making Nonmagnetic Semiconductors Ferromagnetic. Science 1998, 281, 951−956. (23) Medvedkin, G. A.; Ishibashi, T.; Nishi, T.; Hayata, K.; Hasegawa, Y.; Sato, K. Room Temperature Ferromagnetism in Novel Diluted Magnetic Semiconductor Cd1-xMnxGeP2. Jpn. J. Appl. Phys. 2000, 39, L949−L951. (24) Ohno, K.; Ohya, S.; Tanaka, M. Properties of Heavily Mndoped GaMnAs with Curie Temperature of 172.5 K. J. Supercond. Novel Magn. 2007, 20, 417−420. (25) Park, Y. D.; Hanbicki, A. T.; Erwin, S. C.; Hellberg, C. S.; Sullivan, J. M.; Mattson, J. E.; Wilson, A.; Ambrose, T. F.; Spanos, G.; Jonker, B. T. A Group-IV Ferromagnetic Semiconductor: MnxGe1-x. Science 2002, 295, 651−654. (26) Tsui, F.; He, L.; Ma, L.; Tkachuk, A.; Chu, Y. S.; Nakajima, K.; Chikyow, T. Novel Germanium-Based Magnetic Semiconductors. Phys. Rev. Lett. 2003, 91, 177203. (27) Park, Y. D.; Wilson, A.; Hanbicki, A. T.; Mattson, J. E.; Ambrose, T.; Spanos, G.; Jonker, B. T. Magnetoresistance of Mn:Ge Ferromagnetic Nanoclusters in a Diluted Magnetic Semiconductor Matrix. Appl. Phys. Lett. 2001, 78, 2739−2741. (28) Cho, S.; Choi, S.; Hong, S. C.; Kim, Y.; Ketterson, J. B.; Kim, Y. C.; Jung, J.-H. Ferromagnetism in Mn-doped Ge. Phys. Rev. B 2002, 66, 033303. (29) Choi, S.; Hong, S. C.; Cho, S.; Kim, Y.; Ketterson, J. B.; Jung, C.-L.; Rhie, K.; Kim, B.-J.; Kim, Y. C. Ferromagnetic Properties in Cr, Fe-doped Ge Single Crystals. J. Appl. Phys. 2003, 93, 7670−7672.

membrane. The bottom electrode was thermally evaporated 200 nm Ni, and the upper electrodes were deposited employing shadow mask defining its geometry and hence the number of NTs included in a such device. For example, a 400 μm diameter circular electrode deposited on the AAO membrane will cover about 4 × 104 NTs (according to the membrane’s pore size and density). Once Ni contacts were deposited, the SiNT containing AAO was annealed at 600 °C for 1 min in Ar/H2 ambient in a rapid thermal processor (RTP), and its bottom electrode was attached to metal foil with silver paste for the IV measurement. Single SiNT FETs were fabricated using simple lithographic procedures as described previously.12 SiNT samples were annealed at 650 °C for a period of 2 h after their synthesis to crystallize the nanostructures and improve their electrical characteristics. Preparation of the Au-Core SiNT Shell. A 200 nm Ni contact was thermally evaporated on the bottom of the AAO membrane containing ferromagnetic SiNTs grown at 450 °C for 120 min (previous dipping time 15 min). Prior to the evaporation, selective inner-wall oxidation can be readily performed at 800 °C in an oxygen atmosphere, at a pressure of 100 Torr, to create a thin dielectric layer before metal electrodeposition is performed to fill the SiNT interior voids. The AAO membrane was then assembled in an electrochemical cell with the open pores facing a plating bath containing KAu(CN)2 (3.4 mM) pH = 5.8. The deposition of the Au was performed at ambient temperature using a galvanostatic technique, at 2 mA/cm2 for 3 h. Once the Au was plated, the electrolyte solution was removed, and the electrochemical cell was washed with water. Afterward, the AAO membrane was dissolved as mentioned previously.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. Alexander Tsukernik for his help with electrical device preparation using e-beam lithography and Dr. Alexander Gladkih for his help with TOF-SIMS measurements. We also thank the Legacy Foundation, ISF, for the financial support.

■

REFERENCES

(1) Morales, A. M.; Lieber, C. M. A Laser Ablation Method for the Synthesis of Crystalline Semiconductor Nanowires. Science 1998, 279, 208−211. (2) Wagner, R. S.; Ellis, W. C. Vapor-Liquid-Solid Mechanism of Single Crystal Growth. Appl. Phys. Lett. 1964, 4, 89−97. (3) Iijima, S. Helical Microtubules of Graphitic Carbon. Nature 1991, 354, 56−58. (4) Rapoport, L.; Fleischer, N.; Tenne, R. Applications of WS2 (MoS2) Inorganic Nanotubes and Fullerene-like Nanoparticles for Solid Lubrication and for Structural Nanocomposites. J. Mater. Chem. 2005, 15, 1782−1788. (5) Tenne, R. Inorganic Nanoclusters with Fullerene-like Structure and Nanotubes. Prog. Inorg. Chem. 2001, 50, 269−315. (6) Tenne, R.; Zettle, A. K. Nanotubes from Inorganic Materials. Carbon Nanotubes 2001, 80, 81−112. (7) Fagan, S. B.; Baierle, R. J.; Mota, R.; Silva, A. J. R.; Fazzio, A. Ab Initio Calculations for a Hypothetical Material: Silicon Nanotubes. Phys. Rev. B 2000, 61, 9994−9996. 8006

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007

The Journal of Physical Chemistry C

Article

(30) Braak, H.; Gareev, R. R.; Bürgler, D. E.; Grünberg, P.; Schreiber, R.; Schneider, C. M. Magnetic Characteristics of Epitaxial Ge(Mn,Fe) Diluted Films - A New Room Temperature Magnetic Semiconductor? J. Magn. Magn. Mater. 2005, 286, 46−50. (31) Li, A. P.; Shen, J.; Thompson, J. R.; Weitering, H. H. Ferromagnetic Percolation in MnxGe1-x Dilute Magnetic Semiconductor. Appl. Phys. Lett. 2005, 86, 152507. (32) Bougeard, D.; Ahlers, S.; Trampert, A.; Sircar, N.; Abstreiter, G. Clustering in a precipitate-free GeMn magnetic semiconductor. Phys. Rev. Lett. 2006, 97, 237202. (33) Li, A. P.; Zeng, C.; van Benthem, K.; Chisholm, M. F.; Shen, J.; Rao, S. V. S. N.; Dixit, S. K.; Feldman, L. C.; Petukhov, A. G.; Foygel, M.; Weitering, H. H. Dopant segregation and giant magnetoresistance in manganese-doped germanium. Phys. Rev. B 2007, 75, 201201. (34) Devillers, T.; Jamet, M.; Barski, A.; Poydenot, V.; BayleGuillemaud, P.; Bellet-Amalric, E.; Cherifi, S.; Cibert, J. Structure and magnetism of self-organized Ge(1-x)Mn(x) nanocolumns on Ge(001). Phys. Rev. B 2007, 76, 205306. (35) Luo, J.; Liang, J. K.; Liu, Q. L.; Liu, F. S.; Zhang, Y.; Sun, B. J.; Rao, G. H. Structure and magnetic properties of Mn-doped ZnO nanoparticles. J. Appl. Phys. 2005, 97, 086106. (36) Dietl, T. Self-organized growth controlled by charge states of magnetic impurities. Nat. Mater. 2006, 5, 673. (37) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. M.; von Molnár, S.; Roukes, M. L.; Chetchelkanova, A. Y.; Treger, D. M. Spintronics A Spin-Based Electronics Vision for the Future. Science 2001, 294, 1488−1495. (38) Seo, K.; Lee, S.; Younghun, J.; Jo, Y.; Jung, M. H.; Kim, J.; Churchill, D. G.; Kim, B. Room Temperature Ferromagnetism in Single-Crystalline Fe5Si3 Nanowires. J. Phys. Chem. C 2009, 113, 6902−6905. (39) Schmitt, A. L.; Bierman, M. J.; Schmeisser, D.; Himpsel, F. J.; Jin, S. Synthesis and Properties of Single-Crystal FeSi Nanowires. Nano Lett. 2006, 6, 1617−1621. (40) Pearton, S. Magnetic Semiconductors - Silicon-Based Spintronics. Nat. Mater. 2004, 3, 203−204. (41) Singh, A. K.; Briere, T. M.; Kumar, V.; Kawazoe, Y. Magnetism in Transition-Metal-Doped Silicon Nanotubes. Phys. Rev. Lett. 2003, 91, 146802. (42) Kulkarni, J. S.; Kazakova, O.; Holmes, J. D. Dilute Magnetic Semiconductor Nanowires. Appl. Phys. A: Mater. Sci. Process. 2006, 85, 277−286. (43) Appelbaum, I.; Huang, B.; Monsma, D. J. Electronic Measurement and Control of Spin Transport in Silicon. Nature 2007, 447, 295−298. (44) Liu, J.; Harris, A. T. Synthesis of Multiwalled Carbon Nanotubes on Al2O3 Supported Ni Catalysts in a Fluidized-Bed. J. AIChE 2010, 56, 102−113. (45) Li, H.; Zhaoa, N.; Hea, C.; Shi, C.; Dua, X.; Li, J. Low Temperature Fabrication of Hollow Carbon Nanospheres over Ni/ Al2O3 by the Catalytic Method. J. Alloys Compd. 2008, 465, 387−390. (46) Zhu, X.; Cheng, D.; Kuai, P. Catalytic Decomposition of Methane over Ni/Al2O3 Catalysts: Effect of Plasma Treatment on Carbon Formation. Energy Fuels 2008, 22, 1480−1484. (47) Avdeeva, L. B.; Reshetenko, T. V.; Ismagilov, Z. R.; Likholobov, V. A. Iron-Containing Catalysts of Methane Decomposition: Accumulation of Filamentous Carbon. Appl. Catal., A 2002, 228, 53−63. (48) (a) Dubois, L. H.; Nuzzo, R. G. The Decomposition of Silane and Germane on Ni(III): Implications for Heterogeneous Catalysis. Surf. Sci. 1985, 149, 133−145. (b) Wiegand, B. C.; Lohokare, S. P.; Nuzzo, R. G. Si-H Bond Activation on Cu: Reaction of Silane on Cu(111). J. Phys. Chem. 1993, 97, 11553−11562. (49) Zhang, R. Q.; Lifshitz, Y.; Lee, S. T. Oxide-Assisted Growth of Semiconducting Nanowires. Adv. Mater. 2003, 15, 635−640. (50) Bonzel, H. P. Diffusion of Nickel in Silicon. Phys. Status Solidi 1967, 90, 493−499.

(51) Patolsky, F.; Zheng, G. F.; Lieber, C. M. Fabrication of Silicon Nanowire Devices for Ultrasensitive, Label-free, Real-time Detection of Biological and Chemical Species. Nat. Protoc. 2006, 1, 1711−1724. (52) Cui, Y.; Duan, X. F.; Hu, J. T.; Lieber, C. M. Doping and Electrical Transport in Silicon Nanowires. J. Phys. Chem. B 2000, 104, 5213−5216.

8007

dx.doi.org/10.1021/jp2037944 | J. Phys. Chem. C 2012, 116, 8000−8007