This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: ARTICLES High-resolution detection of Au catalyst atoms in Si nanowires
JONATHAN E. ALLEN1†, ERIC R. HEMESATH1†, DANIEL E. PEREA1, JESSICA L. LENSCH-FALK1, Z.Y. LI2, FENG YIN2, MHAIRI H. GASS3, PENG WANG3, ANDREW L. BLELOCH3, RICHARD E. PALMER2 AND LINCOLN J. LAUHON1*
1 2 Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, USA Nanoscale Physics Research Laboratory, School of Physics and Astronomy, The University of Birmingham, Birmingham B15 2TT, UK 3 SuperSTEM Laboratory, STFC Daresbury, Daresbury WA4 4AD, UK † These authors contributed equally to this work. *e-mail: [email protected] Published online: 10 February 2008; doi:10.1038/nnano.2008.5 The potential for the metal nanocatalyst to contaminate vapour–liquid –solid grown semiconductor nanowires has been a long-standing concern, because the most common catalyst material, Au, is highly detrimental to the performance of minority carrier electronic devices. We have detected single Au atoms in Si nanowires grown using Au nanocatalyst particles in a vapour– liquid –solid process. Using high-angle annular dark-ﬁeld scanning transmission electron microscopy, Au atoms were observed in higher numbers than expected from a simple extrapolation of the bulk solubility to the low growth temperature. Direct measurements of the minority carrier diffusion length versus nanowire diameter, however, demonstrate that surface recombination controls minority carrier transport in as-grown n-type nanowires; the inﬂuence of Au is negligible. These results advance the quantitative correlation of atomic-scale structure with the properties of nanomaterials and can provide essential guidance to the development of nanowire-based device technologies. Si nanowires have been used as the active components in a wide variety of proof-of-principle electronic devices including transistors, photodetectors and photoemitters, and chemical sensors1–5. In recent years, signiﬁcant progress has been made in nanowire alignment and integration into devices over large areas6–8. Although these developments represent important steps towards the commercialization of nanowire-based devices, reliable control over the size, composition and properties of nanowires is also essential. The vapour–liquid–solid (VLS) mechanism of nanowire growth9 is widely used to realize precise control of morphology while providing high-quality single-crystal material10. The VLS approach also enables in situ impurity doping through the metal catalyst1 and the formation of complex axial and radial heterostructures11,12. However, the use of a metal nanocatalyst introduces the potential for unintentional incorporation of impurities. Au, the most commonly used catalyst material, is notoriously deleterious to the electronic properties of Si13,14, and this has motivated work on alternative catalysts and growth techniques15. Despite widespread concern over this issue, there have been no direct measurements of Au concentrations in Si nanowires and therefore no means to correlate the presence or absence of Au impurities with the electronic properties of nanowires. Interestingly, early electrical measurements on Au- and Cucatalysed whiskers with diameters of 20–60 mm suggested that incorporated catalyst atoms were responsible for a decreasing resistivity with increasing growth temperature16. Here we show that Au is incorporated through the catalyst in VLS-grown Si nanowires by identifying and spatially localizing
168 single Au atoms in three dimensions. Using complementary analysis techniques, we have placed an upper bound on the Au concentration and ruled out post-growth diffusion as the mode of incorporation. The observed Au levels exceed expectations based on an extrapolation of the equilibrium bulk solubility to low temperatures. Despite contamination from the catalyst, direct measurements of the minority carrier diffusion length in asgrown n-type Si nanowires reveal that surface recombination controls minority carrier transport. The results reported herein quantify the impact of surface chemistry on the electronic properties of semiconductor nanowires and provide essential insights into the relative roles of surface and bulk recombination processes in highly scaled semiconductor nanomaterials. IMAGING Au ATOMS
High-angle annular dark-ﬁeld (HAADF) scanning transmission electron microscopy (STEM) images resolving single Au atoms in Si nanowires are shown in Fig. 1. The Si nanowires used in this study were synthesized by low-pressure chemical vapour deposition using solution-deposited Au colloids to catalyse the growth of nanowires of selected diameters10. STEM measurements were carried out in a 100-kV VG HB501 STEM retroﬁtted with a Nion Co. aberration corrector. Figure 1a,b shows ﬁltered and unprocessed HAADF images, respectively, of a defect-free nanowire on the  zone axis. Bright spots associated with impurities are visible, two of which are highlighted. Atom 2 moved during imaging, creating a double
nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology © 2008 Nature Publishing Group ARTICLES
1 b2 d1 2 d2
y z x b1 10 Focal depth (nm) 5 0 d2 b1 d1 d1 b1 Intensity 100 0 –5 z –10
0 d2 x Intensity 800 400 0 0 1 2 50 100 Excess intensity 150 5 10 15 Lateral distance (nm) 20 Figure 1 Aberration-corrected STEM imaging of Au atoms in a phosphorus-doped Si nanowire. a,b, Filtered (a) and unprocessed (b) HAADF-STEM images of a Si nanowire with a k110l growth direction; the nanowire is oriented on a  zone axis. Scale bars are 5 nm. c, Intensity proﬁles across the unprocessed HAADF-STEM image taken through two bright spots associated with impurities 1 (blue plot) and 2 (green plot) indicated in a. The plot for atom 1 is offset by 200 counts for ease of comparison. The excess counts associated with each impurity are shown in the upper axis in c. Periodic intensity ﬂuctuations from the Si lattice are visible in each plot. Figure 2 Three-dimensional localization of Au atoms. a, Sum of seven aligned HAADF-STEM images of an intrinsic Si nanowire showing impurities trapped at a twin defect (d1 ,d2) and bulk impurities (b1 ,b2). Scale bar is 5 nm. (See Supplementary Information, Movie M1, for aligned but otherwise unprocessed individual images.) b, Excess intensity of Au atoms indicated in a plotted as a function of focal depth. The peak associated with atom b1 lies between the defect atom peaks and is therefore located within the nanowire as indicated in the schematic diagram. The dashed line extending across the nanowire represents the (111) twin defect bisecting the nanowire (as discussed in the Supplementary Information, Fig. S2). spot. Mobile impurities, which were observed in several images and for several wires, may be associated with the surface, although additional studies are needed to conﬁrm this assignment. The HAADF scattering intensity (Fig. 1c) can be used to place a lower bound on the impurity atomic number based on a comparison of the scattering from a single impurity atom with that of a column of Si atoms17. Assuming screened Rutherford scattering with a cross-section depending on Z1.7 (ref. 17) and estimating that the thickness of the approximately circular nanowire is 68 atoms, we ﬁnd that Zimpurity . 50. Single impurities imaged near the Au catalyst tip produce similar excess intensities, providing circumstantial evidence that the bright spots are due to single Au atoms. One cannot deduce, however, that atom 1 of Fig. 1a is in fact within the bulk and not on the upper or lower surface; the interaction of the coherent STEM beam with the crystalline Si nanowire can lead to variations in the beam intensity within the nanowire and hence to variations in excess counts depending sensitively on the location of the impurity (see Supplementary Information, Fig. S1). To conﬁrm the presence of single Au atoms and to localize their positions in all three dimensions, through-focal series HAADF images of a twinned nanowire were taken 8 mm from the nanowire tip (Fig. 2a). The summed HAADF image of Fig. 2a reveals a
nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology strikingly periodic array of bright spots associated with a twin defect. These features extend to the Au catalyst tip, providing strong circumstantial evidence that single Au atoms are being detected. As the focal depth is varied, two distinct lines of impurity atoms come into and out of focus (see Supplementary Information, Movie M1). The intriguing nature of the crystalline defect that leads to this impurity ordering is not the focus of this report. Rather, the well-localized vertical positions of these two lines of Au atoms at defect sites are used to establish that single Au atoms are present in defect-free regions of the nanowire. A series of HAADF-STEM images at varying focal depths was analysed to determine the relative positions of selected atoms along the imaging axis. The intensity versus defocus for three atoms (Fig. 2b) shows that the intensity of the bulk impurity, b1, is highest at a focal position that is bounded above and below by the peaks from the defect impurities d1 and d2. Because the impurities at the defect cannot lie outside the nanowire, we conclude that the bulk impurity lies within the wire. (See Supplementary Information, Fig. S2, for tilted images of the nanowire in Fig. 2a that support our interpretation of the intensity versus defocus proﬁles.) The equilibrium solubility of Au in bulk Si at the growth temperature of 450 8C is not known, but can be conservatively estimated to be less than 1 Â 1014 cm23 based on extrapolation from bulk measurements at higher temperatures18. At this concentration, one would expect to ﬁnd one Au atom for every 5 Â 108 Si atoms. The imaging volume sampled by the entire through-focal series, which included the full depth of the nanowire and the lateral boundaries of Fig. 2a, contains $ 5 Â 105
169 © 2008 Nature Publishing Group ARTICLES
wherein the solubility of Au in Si decreases drastically with temperature, but a supersaturated Au–Si liquid catalyst could shift the solidus towards higher Au concentrations. It is also important to note, however, that the Au impurities may not represent an equilibrium concentration. The growth velocity ( $ 20 nm s21) is much lower than the velocities of $ 1 m s21 associated with kinetic impurity trapping in Si (ref. 21), but the concentration of Au in the catalyst is also much higher than that of the dilute impurities typically considered in Si solidiﬁcation studies21. Kinetic trapping in our low-temperature growth regime could be explored by replacing the silane (SiH4) used in the present study with disilane (Si2H6) to achieve much higher growth rates, and perhaps higher Au concentrations. Finally, we note that the twin defect itself cannot be strictly excluded as a source of the two Au atoms observed in the crystalline regions of Fig. 2a, because bulk diffusion rates of Au at the growth temperature are not known. Although additional studies are needed to explain the incorporation rate, we have conducted local electrode atom probe (LEAP) tomographic analysis to rule out the possibility that the Au diffuses into the nanowire after growth. Recently, we demonstrated the ﬁrst LEAP analysis of semiconductor nanowires to produce three-dimensional maps of nanowire composition with 0.1 –0.3 nm spatial resolution and single-atom sensitivity22. In the present study, we analysed as-grown Si nanowires to look for evidence of bulk Au diffusion and to establish an upper bound on the Au concentration over large regions of the nanowire (Fig. 3). A one-dimensional composition proﬁle through the Au tip into the Si nanowire reveals a chemically abrupt Au –Si interface with a width of , 0.5 nm, and single Au ion counts quickly fall to the background level within 1 nm of the interface (Fig. 3b). There is no detectable concentration gradient within the nanowire, suggesting that the Au catalyst does not act as a post-growth source of diffusing Au atoms. Analyses of complete mass-spectra (see Supplementary Information, Fig. S3) provide an upper bound on the total Au concentration of $ 5 Â 1017 –1.5 Â 1018 cm23 depending on the signal-to-noise characteristics of the spectrum in question, and the background count rate in Au regions of the spectrum does not vary with position. The absence of a detectable gradient in Au concentration along the nanowire, in both STEM and LEAP analysis, is consistent with incorporation during growth. 100 Composition (%) 80 60 40 20 0 –2.0 –1.0 0.0 1.0 Distance (nm) 2.0 4 0 12 Au counts 8 Figure 3 Atom probe analysis of the catalyst – nanowire interface. a, Three-dimensional tomographic reconstruction of the Au catalyst – Si nanowire interface region from LEAP analysis. The interface between the Au (yellow spheres) and Si (red spheres) is abrupt. The dimensions are 4 Â 4 Â 4 nm3. b, One-dimensional composition proﬁle of the interface in a. The right-hand axis shows individual counts detected in the Au mass window beyond the Au – Si interface at 0 nm. Si atoms and at least two bulk Au atoms. If the concentration were established by the bulk phase diagram, the probability of observing two Au atoms in the imaged volume would be approximately 1 in 1,000,000. This suggests that Au atoms are present in excess of the bulk solubility. NATURE OF Au INCORPORATION SCALING OF MINORITY CARRIER DIFFUSION LENGTH
The identiﬁcation of Au atoms in Si nanowires has important implications for our understanding of both the growth process and electronic properties, which we address in that order. There are three types of Au atoms to consider: surface, defect-related and bulk. The origin and effects of surface Au atoms will be discussed in a later section. The Au atoms at the twin defect are intriguing, and the periodicity of their incorporation is suggestive of a stable defect site, but they will not be the focus of this article because they have little statistical impact on the electrical measurements described below. k111l-oriented nanowires predominate for the range of diameters studied in our electrical measurements19, P whereas f111g ¼ 3 twin boundaries parallel to the growth axis occur only in k112l-oriented nanowires. One explanation for the presence of the bulk Au atoms is that they are incorporated directly from a supersaturated, that is, non-equilibrium, Au–Si liquid catalyst. Au-mediated VLS growth of Ge nanowires was recently shown to be a non-equilibrium process involving a supersaturated catalyst particle20, indicating that the bulk equilibrium phase diagram cannot be used to predict the composition of the catalyst and nanowire components during and after growth. The Au–Si system, which has a very similar phase diagram to the Au–Ge system, exhibits a retrograde solidus
170 We next turn to the consequences of Au impurities for the electrical properties of Si nanowires and show that the Au impurities do not signiﬁcantly inﬂuence the electrical properties of as-grown phosphorus-doped nanowires. Au introduces mid-gap traps in bulk Si that act as recombination centres for electrons and holes, signiﬁcantly decreasing the diffusion lengths of excess carriers and thereby compromising the performance of devices such as photovoltaics14. Au can also act as a generation centre, increasing off-currents in transistors and dark currents in photodiodes. Because the minority carrier diffusion length in Si is extremely sensitive to Au contamination, minute concentrations (, 1 Â 1012 cm23) of Au in bulk Si have been quantiﬁed through well-established electrical and optical measurements13. Although many bulk analysis techniques cannot be applied to nanowires because of the very small sample volumes, we have applied electronbeam-induced current (EBIC) microscopy23 to quantify minority carrier diffusion lengths in semiconductor nanowires (Fig. 4). EBIC measurements were conducted on two-terminal n-type Si nanowire Schottky diodes fabricated using electron-beam lithography (Fig. 4a). Devices were loaded into a scanning electron microscope (SEM) in which a focused electron beam
nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology © 2008 Nature Publishing Group ARTICLES
e– Schottky Ohmic Normalized current Vg V
+ 1 0.1 0.01 I 0 50 100 Position (nm) 150 200 100 EF L p (nm) 80 60 40 20 0 20 40 60 80 Diameter (nm) 100 2.0 1.5 (nA ) 1.0 0.5 0.0 –100 0 100 200 Distance (nm) 300 Figure 5 Dependence on diameter of minority carrier diffusion length. a, Semi-logarithmic plot of EBIC proﬁles for nanowires having diameters of 93 nm (ﬁlled blue squares), 55 nm (ﬁlled green circles) and 35 nm (ﬁlled red triangles) with minority carrier diffusion lengths of 78 nm, 57 nm and 28 nm respectively. The dashed lines are ﬁts to the current along the wire, which decays as a eÀx =Lp , where L p is the minority carrier diffusion length. b, Plot of L p versus diameter for varying gas-phase Si:P doping ratios: 500:1 (ﬁlled red triangles), 1,000:1 (ﬁlled green squares), 1,500:1 (ﬁlled blue diamonds), 2,000:1 (ﬁlled magenta circles). Grey lines are model calculations (see Supplementary Information, S5) of L p versus nanowire diameter for various values of the surface recombination velocity S . Top, bottom and dashed lines correspond to S ¼ 1 Â 105 cm s21, 1 Â 106 cm s21 and 3 Â 105 cm s21, respectively. Figure 4 Electron beam induced current measurement on Si nanowire devices. a, Schematic of device geometry. b, Band diagram near the Schottky contact. Excess minority carriers (holes) diffuse to the space-charge region where they are swept to the contact by the internal electric ﬁeld. The gradient shading represents the relative probability of carrier collection. c,d, False-colour SEM image of a nanowire Schottky contact (c) and the induced current image (d) taken simultaneously under a reverse bias of 0.3 V. Scale bars are 100 nm. e, Line proﬁles of the data in d taken along (solid green line) and perpendicular to (dashed line) the nanowire axis showing the signal decay. The abrupt drop-off of the signal away from the wire conﬁrms that backscattered electrons make a negligible contribution to the signal. was scanned over the nanowire device while changes in device current were measured as a function of the beam position (Fig. 4c,d). The energetic electron beam locally generates excess minority carriers (holes), which diffuse in the nanowire channel and are collected at the Schottky contact (Fig. 4b) if they reach the contact before recombining. The minority carrier diffusion length was determined by analysing the rate of decay of the current along the nanowire axis (Fig. 4e). (See Supplementary Information, Figs S4 and S5, for details regarding the diffusion length extraction and control experiments.) The minority carrier diffusion length displays a strong dependence on nanowire diameter (Fig. 5). Diffusion lengths
nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology increase monotonically, from 25 nm up to 80 nm for nanowires 30 nm to 100 nm in diameter, respectively, and are independent of the doping level. In contrast, the hole diffusion lengths in bulk n-type Si at comparable doping levels ($1 Â 1019 –1 Â 1020 cm23) are 1–10 mm, indicating a 100- to 1,000-fold decrease in nanowires compared to bulk. If recombination at bulk Au impurities were limiting the hole diffusion length, we would not observe a diameter dependence. Because the diffusion length shows no sign of saturating with increases in diameter (Fig. 5b), we conclude that the nanowire surface, not bulk Au impurities, controls minority carrier diffusion. This does not completely rule out the inﬂuence of surface Au atoms; Au was recently found to diffuse on the surface of Si nanowires in ultrahigh-vacuum conditions24, and several mobile Au atoms were observed in our HAADF-STEM measurements (see Supplementary Information, Movie M1). It is important to note, however, that we can rule out the presence of a surface monolayer of Au because we do not observe strong high-angle scattering at the edges of the nanowires. It has been proposed that minute partial pressures of oxygen in our low-pressure CVD reactor may be sufﬁcient to suppress diffusion of Au on the Si surface25. Furthermore, surface Au atoms could be removed through straightforward chemical etching following growth. IMPLICATIONS FOR NANOWIRE DEVICES
To account quantitatively for the diameter-dependent diffusion length, and to gain physical insight into the relative inﬂuence of
171 © 2008 Nature Publishing Group ARTICLES
surface and bulk recombination processes, a simple model of minority carrier diffusion and recombination was applied based on previous work on quantum dots26. The basic physical picture and arguments are given below (see Supplementary Information, S5, for details of the model). The diffusion length Lh is proportional to (Dhth)1/2, where Dh is the hole diffusion constant and th the hole lifetime. The diffusion constant is directly proportional to the mobility mh, with the proportionality constant given by the Einstein relation Dh ¼ mhkBT. At the doping levels used, the minority carrier mobility will be limited by ionized impurity scattering27 suggesting that Dh will be comparable to the bulk value and that the diffusion length will be limited by the lifetime th due to recombination at the surface. If the nanowire diameter is much less than the bulk diffusion length, we can deﬁne a one-dimensional effective diffusion length LÃ ¼ (DhtÃ )1/2, where the effective lifetime tÃ is determined by h h h the recombination rate at the surface. This relation was used to generate the curves bounding the data plotted in Fig. 5b, which indicate that the scaling of effective diffusion length with diameter is consistent with a constant surface recombination velocity S % 3 Â 105 cm s21. Using this value for S and the known capture cross-sections for holes14, one can estimate a Si– SiOx interface state density of $ 1.7 Â 1013 cm22, which is consistent with values reported for native oxide on Si (ref. 28). Based on the analysis above, one can make two important conclusions regarding the relative inﬂuence of Au impurities and the nanowire surface. First, given the known relationship between minority carrier lifetime (and diffusion length) and the Au concentration in bulk Si, the longest measured diffusion length of $ 100 nm can be used to place a maximum on the concentration of Au of $ 5 Â 1017 cm23, which is consistent with the LEAP and HAADF-STEM data. Second, even 5 Â 1016 cm23 (1 p.p.m.) of Au will not have a signiﬁcant inﬂuence on minority carrier diffusion in 20 nm nanowires until the surface state density is reduced by at least two orders of magnitude (to $ 2 Â 1011 cm22), which represents a high level of control even for planar structures. This comparison serves as an important reminder that for semiconductor nanomaterials, control over the surface chemistry is at least as important as control over the bulk chemistry. Various Si nanowire surface passivation strategies have been pursued recently29,30, but the robust passivation of nonplanar interfaces remains an important challenge. We emphasize that Si nanowires are being synthesized by a large number of groups under varying conditions, and that factors such as growth temperature, reactant partial pressure and intentional impurity doping may inﬂuence the level of Au incorporation. Clearly, a quantitative comparison of materials from different laboratories is highly desirable. Although additional HAADFSTEM studies are necessary to establish bounds on the Au concentration for different synthesis conditions, we have demonstrated the feasibility of such measurements. By advancing the quantitative correlation of atomic-scale structure with the properties of nanomaterials, these scientiﬁc ﬁndings and the capabilities we have demonstrated can provide essential guidance to the development of nanowire-based device technologies.
isopropyl alcohol and drop-cast onto degenerately doped Si substrates with 400 nm of thermal oxide. Contact regions were deﬁned by electron beam lithography using a PMMA/MMA bilayer resist. Thirty-second oxygen plasma cleaning of the electrode regions was followed by a 7-s etch in buffered hydroﬂuoric acid, after which the substrates were immediately loaded into an electron beam evaporator for contact evaporation. Samples were lifted off in acetone and rinsed with isopropyl alcohol. The contact metal for ohmic contacts was Ni, and Au was used to form Schottky contacts.
SCANNING TRANSMISSION ELECTRON MICROSCOPY (STEM) Nanowires were suspended in solution by sonication in isopropyl alcohol, and the solution was dropped directly onto holey carbon grids for STEM imaging. High-resolution electron microscopy was performed at the UK SuperSTEM laboratory on a 100-kV VG HB501 dedicated STEM ﬁtted with a Nion secondgeneration spherical aberration corrector. The convergence semi-angle of the electron probe was 24 mrad and the HAADF semi-angle was 70 – 200 mrad. STEM enables both bright-ﬁeld and HAADF images to be acquired simultaneously. Through-focal HAADF series were acquired at nanometre intervals, with the ﬁrst image under-focused (beyond the beam exit surface) and the ﬁnal image over-focused (before the beam entrance surface). The images were then manually aligned to remove the effects of sample drift. STEM images were acquired at a resolution of 1,024 Â 1,024 and ﬁltered using a bandpass ﬁlter (3 – 40 pixels) in ImageJ freeware.
LOCAL-ELECTRODE ATOM PROBE TOMOGRAPHY Laser-assisted local electrode atom probe analysis was conducted in the NUCAPT facility at Northwestern University on an Imago Scientiﬁc Instruments LEAP 3000X Si system. Si nanowires were grown epitaxially on Si(111) substrates as described previously31. The specimen temperature was held below 100 K during analysis. Typical laser pulse energies and frequencies were 0.1 nJ and 100 kHz, respectively.
ELECTRON BEAM INDUCED CURRENT Electron beam induced current measurements were made on an FEI Quanta FESEM using the Nabity NPGS system to control data acquisition. The induced current was recorded as a function of the position of a focused electron beam using a lock-in ampliﬁer and an electron beam modulated at 1,919 Hz by the beam blanker. The electron beam current and voltage were 20 pA and 5 kV, respectively. (See Supplementary Information, Fig. S4, for a discussion of other scan conditions used in the control experiments.) The scan speed was 600 nm s21, with a dwell time of 10 ms per point. Received 13 September 2007; accepted 20 December 2007; published 10 February 2008. References
1. Cui, Y. & Lieber, C. M. Functional nanoscale electronic devices assembled using silicon nanowire building blocks. Science 291, 851 –853 (2001). 2. Beckman, R., Johnston-Halperin, E., Luo, Y., Green, J. E. & Heath, J. R. Bridging dimensions: Demultiplexing ultrahigh-density nanowire circuits. Science 310, 465 –468 (2005). 3. Huang, Y. et al. Logic gates and computation from assembled nanowire building blocks. Science 294, 1313–1317 (2001). 4. Patolsky, F. et al. Detection, stimulation, and inhibition of neuronal signals with high-density nanowire transistor arrays. Science 313, 1100– 1104 (2006). 5. Xiang, J. et al. Ge/Si nanowire heterostructures as high-performance ﬁeld-effect transistors. Nature 441, 489 –493 (2006). 6. Huang, Y., Duan, X. F., Wei, Q. Q. & Lieber, C. M. Directed assembly of one-dimensional nanostructures into functional networks. Science 291, 630– 633 (2001). 7. Jin, S. et al. Scalable interconnection and integration of nanowire devices without registration. Nano Lett. 4, 915 –919 (2004). 8. Javey, A., Nam, S., Friedman, R. S., Yan, H. & Lieber, C. M. Layer-by-layer assembly of nanowires for three-dimensional, multifunctional electronics. Nano Lett. 7, 773– 777 (2007). 9. Wagner, R. S. & Ellis, W. C. Vapor–liquid –solid mechanism of single crystal growth. Appl. Phys. Lett. 4, 89– 90 (1964). 10. Cui, Y., Lauhon, L. J., Gudiksen, M. S., Wang, J. F. & Lieber, C. M. Diameter-controlled synthesis of single-crystal silicon nanowires. Appl. Phys. Lett. 78, 2214 –2216 (2001). 11. Gudiksen, M. S., Lauhon, L. J., Wang, J., Smith, D. C. & Lieber, C. M. Growth of nanowire superlattice structures for nanoscale photonics and electronics. Nature 415, 617 –620 (2002). 12. Lauhon, L. J., Gudiksen, M. S., Wang, C. L. & Lieber, C. M. Epitaxial core –shell and core –multishell nanowire heterostructures. Nature 420, 57–61 (2002). 13. Sprokel, G. J. & Fairﬁeld, J. M. Diffusion of gold into silicon crystals. J. Electrochem. Soc. 112, 200– 203 (1965). 14. Bullis, W. M. Properties of gold in silicon. Solid State Electron. 9, 143 –168 (1966). 15. Wang, Y., Schmidt, V., Senz, S. & Gosele, U. Epitaxial growth of silicon nanowires using an aluminium catalyst. Nature Nanotech. 1, 186 –189 (2006). 16. Shchetinin, A. A. et al. Distribution coefﬁcient of the initiating impurity and its inﬂuence on the electrical-resistivity of ﬁlamentary silicon-crystals. Inorgan. Mater. 27, 1137–1139 (1991). 17. Voyles, P. M., Muller, D. A., Grazul, J. L., Citrin, P. H. & Gossmann, H. J. L. Atomic-scale imaging of individual dopant atoms and clusters in highly n-type bulk Si. Nature 416, 826– 829 (2002). METHODS
NANOWIRE SYNTHESIS AND DEVICE FABRICATION Nanowires were grown at 450 8C (460 8C for n-type) in a 1-inch quartz tube furnace from colloidal Au catalyst particles dispersed on Si substrates. The Si source was silane at a partial pressure of 2 torr and ﬂow rates of 0.2 and 2 s.c.c.m. for the intrinsic and n-type wires, respectively. The phosphorus source for the n-type wires was phosphine (2.5– 20 s.c.c.m.) with N2 used as a dilutant gas at 30 – 47.5 s.c.c.m. to maintain a constant face velocity at a total pressure of 40 torr. For device fabrication, nanowires were suspended in solution by sonication in
172 nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology © 2008 Nature Publishing Group ARTICLES
18. Stolwijk, N. A., Schuster, B. & Holzl, J. Diffusion of gold in silicon studied by means of neutron-activation analysis and spreading-resistance measurements. Appl. Phys. A Mater. Sci. Process. 33, 133– 140 (1984). 19. Wu, Y. et al. Controlled growth and structures of molecular-scale silicon nanowires. Nano Lett. 4, 433– 436 (2004). 20. Kodambaka, S., Tersoff, J., Reuter, M. C. & Ross, F. M. Germanium nanowire growth below the eutectic temperature. Science 316, 729 –732 (2007). 21. Reitano, R., Smith, P. M. & Aziz, M. J. Solute trapping of group-III, IV, and V elements in silicon by an aperiodic stepwise growth-mechanism. J. Appl. Phys. 76, 1518– 1529 (1994). 22. Perea, D. E. et al. Three-dimensional nanoscale composition mapping of semiconductor nanowires. Nano Lett. 6, 181 –185 (2006). 23. Ioannou, D. E. & Dimitriadis, C. A. A SEM-EBIC minority-carrier diffusion-length measurement technique. IEEE Trans Electron. Dev. 29, 445– 450 (1982). 24. Hannon, J. B., Kodambaka, S., Ross, F. M. & Tromp, R. M. The inﬂuence of the surface migration of gold on the growth of silicon nanowires. Nature 440, 69–71 (2006). 25. Kodambaka, S., Hannon, J. B., Tromp, R. M. & Ross, F. M. Control of Si nanowire growth by oxygen. Nano Lett. 6, 1292 –1296 (2006). 26. Daiminger, F., Schmidt, A., Faller, F. & Forchel, A. Picosecond time resolved investigations of carrier lifetime and carrier capture in InGaAs/GaAs quantum dots. Proc. SPIE 2139, 213 –221 (1994). 27. Klaassen, D. B. M. A uniﬁed mobility model for device simulation .2. Temperature-dependence of carrier mobility and lifetime. Solid State Electron. 35, 961– 967 (1992). 28. Gray, P. V. & Brown, D. M. Density of SiO2 – Si interface states. Appl. Phys. Lett. 8, 31– 33 (1966). 29. Seo, K. I., Sharma, S., Yasseri, A. A., Stewart, D. R. & Kamins, T. I. Surface charge density of unpassivated and passivated metal-catalyzed silicon nanowires. Electrochem. Solid State Lett. 9, G69 –G72 (2006). 30. Haick, H., Hurley, P. T., Hochbaum, A. I., Yang, P. D. & Lewis, N. S. Electrical characteristics and chemical stability of non-oxidized, methyl-terminated silicon nanowires. J. Am. Chem. Soc. 128, 8990– 8991 (2006). 31. Perea, D. E., Lensch, J. L., May, S. J., Wessels, B. W. & Lauhon, L. J. Composition analysis of single semiconductor nanowires using pulsed-laser atom probe tomography. Appl. Phys. A Mater. Sci. Process. 85, 271 –275 (2006). Acknowledgements
The work at Northwestern was supported by the National Science Foundation (NSF) through the Materials Research Science and Engineering Center (MRSEC) (J.A.), CAREER (L.L.), NIRT (E.H.) and Graduate Research Fellowship (J.L.) programmes; the Ofﬁce of Naval Research and a Ford Foundation Fellowship (D.P.); and an Alfred P. Sloan Research Fellowship (L.L.). E.H. acknowledges a travel grant from the Northwestern University Nanoscale Science and Engineering Center (NSEC). We acknowledge the Northwestern University Center for Atom – Probe Tomography facility and the Northwestern University Atomic- and Nanoscale Characterization Experimental Center (NUANCE). The NUANCE Center is supported by NSF-NSEC, NSF-MRSEC, the Keck Foundation, the State of Illinois and Northwestern University. The work at Birmingham and Daresbury was supported by the Engineering and Physical Sciences Research Council. Correspondence and requests for materials should be addressed to L.L. Supplementary information accompanies this paper on www.nature.com/naturenanotechnology. Author contributions
J.A. and E.H. contributed equally to this work. J.A. performed the device experiments and analysed the data with L.L. D.P. performed the atom probe experiments and analysed the data with L.L. E.H. and J.L. synthesized the materials. The STEM work and analysis were conducted by the Birmingham-SuperSTEM collaboration (Z.Y.L, F.Y, R.E.P., M.H.G., P.W. and A.L.B.) in association with the Northwestern group (E.H. and L.L.). L.L. coordinated the design and execution of the experiments. J.A. and L.L. co-wrote the paper. All authors discussed the results and commented on the manuscript. Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/ nature nanotechnology | VOL 3 | MARCH 2008 | www.nature.com/naturenanotechnology 173 © 2008 Nature Publishing Group ...
View Full Document
This document was uploaded on 08/16/2010.
- Spring '09