Heterogeneous and Homogeneous Nucleation of ... - ACS Publications


Heterogeneous and Homogeneous Nucleation of...

0 downloads 114 Views 2MB Size

J. Phys. Chem. C 2011, 115, 397–401

397

Heterogeneous and Homogeneous Nucleation of Epitaxial NiSi2 in [110] Si Nanowires Yi-Chia Chou,*,† Wen-Wei Wu,‡ Chung-Yang Lee,| Chun-Yi Liu,| Lih-Juann Chen,| and King-Ning Tu† Department of Materials Science and Engineering, UniVersity of California at Los Angeles, Los Angeles, California 90095, United States, Department of Materials Science and Engineering, National Chiao Tung UniVersity, Hsinchu 300, Taiwan, and Department of Materials Science and Engineering, National Tsing Hua UniVersity, Hsinchu 300, Taiwan ReceiVed: September 12, 2010; ReVised Manuscript ReceiVed: NoVember 16, 2010

We report here both heterogeneous and homogeneous nucleation of epitaxial silicide of NiSi2 in Si nanowires grown in [110] direction by in situ observation in high-resolution transmission electron microscopy (TEM). Owing to the excellent lattice match between Si and NiSi2, a giant epitaxial step of 2-8 nm wide forms at the interface between Si and NiSi2 during the growth of the latter. The step formation results in two silicide/ Si interfaces parallel to each other in TEM observations. Heterogeneous nucleation of NiSi2 occurs at the intersection where the step meets the interfaces. However, we have also observed the epitaxial growth of NiSi2 having a single interface with the Si, i.e., without a giant step. Homogeneous nucleation of NiSi2 occurs on the single interface. Incubation time of heterogeneous nucleation of NiSi2 has been measured by highresolution video to be much shorter than that of homogeneous nucleation. The overall growth rate of NiSi2 for the case of heterogeneous nucleation is faster than that for the case of homogeneous nucleation. Kinetic analysis of both types of nucleation is presented for a direct comparison in order to have a better understanding of the nucleation events. Introduction

Experimental Section

As we approach the end of Moore’s law of scaling of MOSFET devices, much effort has been made recently to synthesize nanoscale building blocks based on Si nanowires for integrated circuits in nanoelectronics.1-5 The morphology, size, and electrical properties of Si nanowires are favorable for assembling nanoelectronic devices as well as for biosensors.6-10 The formation of nanosilicide contacts to nanowires of Si is crucial in enabling device performance.9,10 Thus, it is essential to have a detailed understanding of the mechanism of nucleation and growth of silicide contacts in Si nanowires and the silicide/ Si interfacial structure and properties.11,12 Ledge nucleation and flow were also observed at the silicide (catalyst)/Si interface during Si nanowire growth by a vapor-solid-solid mechanism in Cu-Si and Pt-Si systems.13,14 The primary epitaxial growth direction of Si nanowires is 〈110〉 but depends on the relative orientation of the catalyst(silicide) and substrate. Both NiSi and NiSi2 silicide nanowires have been synthesized as low resistivity contacts in low-power memory devices and NiSi2 as optical absorbers in optical communication applications.15,16 At the NiSi/ Si interface, a large misfit strain was found because of the large mismatch in their lattice parameters and lattice structures; however, stain is nearly negligible in the interface between NiSi2 and Si due to the same cubic lattice structure and close lattice parameters. Moreover, recent works on the transport measurements of silicide nanowires show excellent electrical transport and field emission properties.4,8

Point Contact Sample Preparation. The Si nanowires were prepared on Si wafers by the vapor-liquid-solid (VLS) method using gold nanodots as catalyst for the growth of single crystal Si nanowires with [110] growth direction.18-20 The Si nanowires having a diameter about 70 nm and length of a few micrometers were detached from the wafer in ethanol solution by ultrasonic vibration. Samples for silicide formation were prepared by depositing 5 nm thick Ni thin film on a Cu grid coated with a 50 nm thick amorphous SiO2 film and followed by a hightemperature annealing to form Ni nanodots. Then a drop of solution containing Si nanowires was dripped on top of Ni nanodots to make point contact samples. The samples were dried at room temperature. Characterization Techniques. In situ annealing at ∼800 °C under ultrahigh vacuum for point contact reactions was performed and images were taken with a JEOL 2000 V ultrahigh vacuum transmission electron microscope. High-resolution lattice imaging and line scan were performed with a JEOL 3000F high-resolution transmission electron microscope.

* To whom correspondence should be addressed, [email protected]. † Department of Materials Science and Engineering, University of California at Los Angeles. ‡ Department of Materials Science and Engineering, National Chiao Tung University. | Department of Materials Science and Engineering, National Tsing Hua University.

Results and Discussion Using Si nanowires grown in the [110] instead of [111] direction, we have observed in situ epitaxial growth of NiSi2 with a sharp NiSi2/Si interface with high-resolution transmission electron microscopy (TEM). Both events of heterogeneous and homogeneous nucleation of the NiSi2 were observed, analyzed and are reported here. The heterogeneous nucleation does not occur from the side wall of the Si nanowire, rather a giant step was found in the middle of the silicide/Si interface and the edges of the step served as the heterogeneity for heterogeneous nucleation of NiSi2. Because of the step, we observed in TEM two sharp interfaces between the NiSi2 and

10.1021/jp108686y  2011 American Chemical Society Published on Web 12/13/2010

398

J. Phys. Chem. C, Vol. 115, No. 2, 2011

Chou et al.

Figure 1. High-resolution TEM images and line scan analyses of interfaces of NiSi2 and Si nanowires. (a) Low magnification in situ TEM capture of NiSi2 growth in a Si nanowire. The darker region is NiSi2 and the brighter region is Si. Two flat interfaces were shown in the image. (b) High-resolution TEM image of the interface between NiSi2 and Si. The insets in panel b are the fast Fourier transform (FFT) pattern of Si and NiSi2 zone axis. (c) An enlarged high-resolution TEM image of the two interfaces, which divided into three regions. (d) Line scan analysis through the two interfaces showing Ni and Si concentration profiles along the line.

Si. The epitaxial growth rates of these two interfaces were measured to be almost the same. The growth requires repeating heterogeneous nucleation of each atomic layer of NiSi2, and the incubation time of heterogeneous nucleation has been measured. On the other hand, the homogeneous nucleation of NiSi2 in Si nanowires grown in the [110] direction was also observed. It occurs when the NiSi2 forms only one epitaxial interface with the Si, i.e., without the giant step. The incubation time of homogeneous nucleation was measured to be about 5 times longer than that of heterogeneous nucleation; in turn the rate of epitaxial growth with homogeneous nucleation was measured to be slower by a factor of 3 than that with heterogeneous nucleation. At ∼800 °C under ultrahigh vacuum, the silicide phase formed is NiSi2. Two flat interfaces with ∼8 nm distance were observed in situ between the NiSi2 and the Si, as shown in Figure 1a. The high-resolution TEM images (Figure 1b,c) show the lattice imaging of three regions across the two interfaces. The epitaxial relations across the two vertical (to the nanowire orientation) interfaces between NiSi2 and Si (region A and B in Figure 1c, respectively) are Si(-220)//NiSi2(-220) and Si[111]//NiSi2[111]. Since NiSi2 and Si have cubic CaF2 and diamond crystal structure, respectively, and the lattice parameters are 0.542 and 0.543 nm, respectively, a good match results for forming a sharp epitaxial interface. The two atomically sharp interfaces indicate the formation of a giant step with good epitaxial relations between the phases, as depicted in Figure 2a. The lattice in the middle region C is essentially the same as that in NiSi2 in region A and that in Si in region B (Figure 1b,c). In other words, region C is a composite of NiSi2 and Si. The width of the step, the width in region C, is about 2-8 nm. Because of the good lattice match between Si and NiSi2, we also found the epitaxial relations across the horizontal interface or across the giant step, which are Si(111)//NiSi2(111) and Si[211]//NiSi2[211]. The angle between the vertical (0-22) and the horizontal (111) interfaces is 90°. Their intersections or the edges of the step are along the [211] direction. To verify the composition of region C, line scan EDX analysis was performed by moving an electron beam with a diameter of

Figure 2. Schematic diagram of interfaces of NiSi2 and Si and the stepwise growth curve of NiSi2 formation by heterogeneous nucleation. (a) Schematic diagram of interfaces of NiSi2 and Si and heterogeneity sites in heterogeneous nucleation mode. (b) Stepwise growth curve of two interfaces of NiSi2 and Si by heterogeneous nucleation.

25 nm slowly across the two interfaces between NiSi2 and Si to measure the concentrations of Ni and Si along the line. The resolution of the line scan is better than 5 nm, so we can infer two concentration drops within ∼8 nm width at the interface. The concentration profiles are shown in Figure 1d. The variation of concentration of Si is small because of the interference of the SiO2 coating on the substrate, but the concentration profile of Ni varies much more clearly from a high concentration of about 33% in the NiSi2 to a low concentration of 0% in the Si region. The circle in Figure 1d highlights the step-interface between the NiSi2 and the Si.

Nucleation of NiSi2 in [110] Si Nanowires

Figure 3. In situ TEM images of interfaces between the two phases and its schematic diagram. (a, b) In situ TEM images with different magnification taken at different times show the displacement of the two interfaces due to silicide growth. The first two numbers in the upper-right corner are in units of minutes and the following two numbers are in units of seconds. (c, d) In situ TEM images with different magnification show a third interface existing and moving between the two interfaces of Si and NiSi2, while the silicide growth is shown by the displacement of the two interfaces. (e) A schematic diagram of NiSi2 nuclei on the horizontal (111) interface of the giant step.

During the epitaxial growth, we have examined the motion of the two (0-22) vertical interfaces in high-resolution video, and we found that their motion is not continuous; rather it is jerky, consisting of sequential advancements of very small steps. Between two sequential steps, there is an incubation time. In Figure 2b, the horizontal scale is the incubation time and it is about 1 s. The vertical scale is the step height which is about 0.19 nm. Specifically, the horizontal length of each step stands for the incubation period of nucleation of a NiSi2 atomic layer and the width of the vertical line is the growth period of a NiSi2 atomic layer. In Figure 2a, we have depicted the heterogeneous nucleus by the half circular disk at the edges of the two (0-22) vertical interfaces. We shall discuss the kinetics of heterogeneous nucleation and its incubation time later. Furthermore, when we examined the epitaxial growth in situ using low magnitude video, we observed often the image of an additional or a third interface which fluctuates between the two migrating (0-22) vertical interfaces as indicated by an arrow in the TEM images in panels c and d of Figure 3. In panels a and b of Figure 3, a pair of TEM images taken at two different times is shown. The interval between them is about 5.5 min. Using the image of particle at the lower left corner as a marker, we can recognize the motion of the two interfaces. In Figure 3b, they have reached the particle. In the video recording, we found the pair of images shown in panels c and d of Figure 3, taken close to that in panels a and b of Figure 3, respectively. The recording showed that as the third interface moved and reached the front interface (the one on the left-hand side in Figure 3a and b), it disappeared and was followed by a new cycle of the growth of the two (0-22)

J. Phys. Chem. C, Vol. 115, No. 2, 2011 399

Figure 4. Single interface of NiSi2 and Si, and the stepwise growth curve of NiSi2 formation by homogeneous nucleation. (a) Highresolution image of single interface of NiSi2 and Si. (b) Stepwise growth curve of NiSi2 epitaxial growth with Si by homogeneous nucleation. The upper and lower curves are from two different sets of data.

vertical interfaces. The appearance of the third interface is infrequent. In Figure 3e, we depict the heterogeneous nucleation of half disks on the horizontal interface or the step. The probability of a heterogeneous nucleation should be nearly the same as those depicted in Figure 2a. The growth of a layer of NiSi2 on the horizontal interface will form the image of the third interface. However, it is plausible that the third interface could mean a medium size step in the giant step. During the epitaxial growth, if the front interface moves faster than the rear interface, it will increase the width of the step, which is energetically unfavorable. On the other hand, if the rear interface moves faster than the front interface, they will merge together to become a single epitaxial interface. Without the step, it means without the heterogeneity, the nucleation event will change from heterogeneous nucleation to homogeneous nucleation, in turn the reaction rate will slow down. Figure 4a shows the TEM image of a single epitaxial interface between Si and NiSi2. Figure 4b shows the steps of advancement of the single epitaxial interface and the incubation time of each step. Significantly, we found that the average incubation time is about 5 s, which is much longer than that shown in Figure 2b. We interpret the difference in incubation time as a change from heterogeneous to homogeneous nucleation. Affected by the incubation time, the overall growth rate of NiSi2 with heterogeneous and homogeneous nucleation was measured to be about 0.133 and 0.0415 nm/s, respectively. Our previous work on homogeneous nucleation is the epitaxial growth of NiSi on Si nanowires grown in the [111] direction. The repeating event of homogeneous nucleation was revealed by a jerky motion of epitaxial growth with atomic layer by atomic layer of NiSi on the Si. A nucleus forms at the middle region of the Si nanowires by homogeneous nucleation since no giant step was found.11 This is because NiSi has a poor lattice match with Si, so NiSi cannot have a horizontal step which is

400

J. Phys. Chem. C, Vol. 115, No. 2, 2011

Chou et al.

epitaxial with Si. Thus, the giant step formation is unique in the epitaxial growth of NiSi2 on Si, and we have found both heterogeneous nucleation and homogeneous nucleation in the same system. It is of interest to ask if we grow NiSi2 on Si nanowires grown in the [111] direction shall we find the giant step? Since CoSi2 has the same crystal structure and nearly the same lattice parameter as NiSi2, it is again of interest to know if giant steps can form or not. To estimate the activation energy of heterogeneous nucleation, nuclei are assumed to form at the heterogeneity sites as two half circular disks, p and q, shown in Figure 2a, rather than disks which have a wetting angle of θ. This is because we can compare p and q nuclei more easily. The total free energy for the heterogeneous nucleation of the p nucleus is

∆Gn,p

On the other hand, total free energy of homogeneous nucleation is ∆Gn ) -πR2a(∆Gv - ∆Gs) + 2πRaγ1 ) -n(∆Gv - ∆Gs) + 4πa 1/2 1/2 n γ1 ) -n(∆Gv - ∆Gs) + bn1/2γ1 Ω

( )

where b ) (4πa/Ω)1/2. Likewise, the critical energy of it is

∆G*homo )

∆G*homo ) πγ12

and that for the heterogeneous nucleation of q nucleus is

1 ∆Gn,q ) - πR2a(∆Gv - ∆Gs) + (πRaγ1 + 2Raγ2) 2

1/2

γ1n1/2 )

-n(∆Gv - ∆Gs) + b1γ1n1/2 and

∆Gn,q ) -n(∆Gv - ∆Gs) + 8a ( πΩ )

1/2

( 2πa Ω )

1/2

γ1n1/2 +

γ2n1/2 ) -n(∆Gv - ∆Gs) + b1γ1n1/2 + b2γ2n1/2

where b1 ) (2πa/Ω)1/2 and b2 ) (8a/πΩ)1/2. The critical energy of formation of a heterogeneous nucleus can be obtained

∆G*hetero )

(b1γ1)2 4(∆Gv - ∆Gs)

∆G*hetero )

(b1γ1 + b2γ1)2 4(∆Gv - ∆Gs)

for p and

for q.

1 a Ω ∆Gv - ∆Gs

1 a 1 ∆G*hetero,p ) πγ12 2 Ω ∆Gv - ∆Gs

where ∆Gv and ∆Gs are Gibbs free energy of formation of NiSi2 and misfit strain energy per unit volume, respectively, R is radius of the disk, a is atomic layer thickness, γ is interfacial energy per unit area, and γ1 is for the curved interface and γ2 is for the interface of (111) NiSi2 and (111) Si. We must consider strain energy in the nucleation of an atomic thick epitaxial disk. This is because the edge of the disk is similar to the edge of an edge dislocation. Then γ1 and γ2 are actually line energies rather than interfacial energies. Dislocation has a long-range strain field. By assuming nΩ ) 1/2 (πR2a), where n is number of atoms and Ω is the atomic volume, we can change the energies from per unit volume (or per unit area) to per atom, we have the following equations for p and q, respectively.

( 2πa Ω )

(2)

By substituting back b1 for p, and b1 and b2 for q into eq 1, and b in eq 2, we have

1 ) - πR2a(∆Gv - ∆Gs) + πRaγ1 2

∆Gn,p ) -n(∆Gv - ∆Gs) +

(bγ1)2 4(∆Gv - ∆Gs)

(1)

and

∆G*hetero,q )

[ 21 πγ

2

1

+

1 a 2 2 γ + 2γ1γ2 π 2 Ω ∆Gv - ∆Gs

]

Thus, we can compare the critical energy of formation of homogeneous and heterogeneous nucleation of p and q nuclei. For the p nucleus, we obtain

∆G*homo =2 ∆G*hetero,p

(3)

For the q nucleus, the case is much more complicated because γ1 can be higher than γ2. While both are NiSi2/Si interfaces, we note that γ1 is for a curved interface and γ2 is for the step interface, which should be a low energy interface, otherwise the step cannot be formed. Because of uncertainly, we take a range of values of the ratio of γ2 and γ1 as γ2/γ1 ) 0.5 ( 0.1. So we obtain

∆G*homo = 1.15 ( 0.11 ∆G*hetero,q

(4)

Comparing eqs 3 and 4, we might expect the rate of frequency of nucleation of q nucleus would be much slower than that of p nucleus. However, the two experimental curves as shown in Figure 2b indicate that the rate of heterogeneous nucleation of p and q nucleus is nearly the same. In the derivation of eqs 3 and 4, we have assumed that the strain energy is the same. This may not be true and it is possible that the strain energy of the q nucleus is less than that of the p nucleus. If so, it will enhance the nucleation of the q nucleus. Furthermore, if we consider the edges of the p and q nuclei as dislocation edges, we should take line energy instead of surface energy in forming the nucleus. Yet the calculation of the line energy and the strain energy in this case is beyond the scope of our work.

Nucleation of NiSi2 in [110] Si Nanowires

J. Phys. Chem. C, Vol. 115, No. 2, 2011 401

TABLE 1: The Required Number of Molecules to Form a Stable Silicide Nucleusa

Z ) 0.1 Z ) 0.05

n* homogeneous nucleation

n* heterogeneous nucleation (q nucleus)

n* heterogeneous nucleation (p nucleus)

16-18 32-36

16 32

11-13 23-25

a n* is the number of molecules required to form a stable NiSi2 nucleus. Z is Zeldovich factor.

Activation energy of NiSi2 formation in thin film reactions has been reported to be 2.95 eV/atom, which is by heterogeneous nucleation.17 To make a conservative estimation, we take this value to be the activation energy for the q nucleus because we note that the p nucleus is a rather special type of heterogeneous nucleation. Then we use eq 4 to estimate the activation energy of homogeneous nucleation of NiSi2; it will be 3.4 ( 0.33 eV/ atom. Equation 4 shows that the number of molecules required to form a critical nucleus by homogeneous nucleation is about 36-11% higher than that by heterogeneous nucleation. In Table 1, the number of NiSi2 molecules in a stable nucleus is given, assuming the Zeldovich factor equals 0.1 and 0.05, and the activation energies used are 2.95 and 3.4 ( 0.33 eV/atom for heterogeneous and homogeneous nucleation, respectively. The Zeldovich factor is a kinetic factor that stands for the percentage of critical nuclei which become stable nuclei by the assumption of a thermally activated process of incremental growth of subcritical nuclei and a steady-state distribution of subcritical clusters. The nuclei formation in a homogeneous nucleation occurs randomly on the silicide/Si interface (a flat interface) and in a heterogeneous nucleation occurs randomly along the edge line of the step as sketched by yellow lines in Figure 2a. The Zeldovich factor which applies to both cases is

Z)

[

1 ∆G* 4πkT (n*)2

]

1/2

for a disk nucleus. The major findings reported here on epitaxial growth of NiSi2 in nanowire of Si are that both homogeneous and heterogeneous nucleation can occur and the latter is facilitated by a giant step. The formation of a giant step is energetically unfavorable; why it forms can tentatively be explained by a kinetic reason because it leads to a faster phase change or a higher rate of free energy change. Yet, what determines the width of the step is unclear and whether there is a critical width remains to be studied. We note that no misfit dislocations have been found at the epitaxial interfaces. Conclusions Epitaxial growth of nano NiSi2 in Si nanowires grown in [110] direction has been studied by in situ high-resolution TEM. The phase change can take place by either homogeneous or heterogeneous nucleation with a large difference in incubation time. This is because of the excellent lattice match between Si and NiSi2, a giant epitaxial step can be formed in the interface. The edges of the step serve as sites for heterogeneous nucleation of NiSi2. Without the step, homogeneous nucleation occurs. Incubation time of heterogeneous and homogeneous nucleation has been measured

to be about 1 and 5 s, respectively. The overall growth rate of NiSi2 with heterogeneous and homogeneous nucleation was measured to be about 0.133 and 0.0415 nm/s, respectively. Acknowledgment. The authors at UCLA acknowledge support by NSF/NIRT project CMS-0506841. The authors at NTHU acknowledge support by National Science Council project 982221-E-007-104-MY3. The author at NCTU acknowledges support by National Science Council project 97-2218-E-009027-MY3. We thank Professor A. M. Gusak at Cherkasy National University, Ukraine for helpful comments. References and Notes (1) Cui, Y.; Lieber, C. M. Functional Nanoscale Electronic Devices Assembled Using Silicon Nanowire Building Blocks. Science 2001, 291, 851–853. (2) Huang, Y.; Duan, X. F.; Cui, Y.; Lauhon, L. J.; Kim, K. H.; Lieber, C. M. Logic Gates and Computation from Assembled Nanowire Building Blocks. Science 2001, 294, 1313–1317. (3) Wu, Y.; Xiang, J.; Yang, C.; Lu, W.; Lieber, C. M. Single-Crystal Metallic Nanowires and Metal/Semiconductor Nanowire Heterostructures. Nature 2004, 430, 61–65. (4) Lu, W.; Lieber, C. M. Nanoelectronics from the Bottom up. Nat. Mater. 2007, 6, 841–850. (5) Chen, L. J. Silicon Nanowires: the key Building Block for Future Electronic Devices. J. Mater. Chem. 2007, 17, 4639–4643. (6) Xia, Y.; Yang, P.; Sun, Y.; Wu, Y.; Mayers, B.; Gates, B.; Yin, Y.; Kim, F.; Yan, H. One-Dimensional Nanostructures: Synthesis, Characterization, and Applications. AdV. Mater. 2003, 15, 353–388. (7) Javey, A.; Nam, S.; Friedman, R. S.; Yan, H.; Lieber, C. M. Layerby-Layer Assembly of Nanowires for Three-Dimensional, Multifunctional Electronics. Nano Lett. 2007, 7, 773–777. (8) Lee, C. Y.; Lu, M. P.; Liao, K. F.; Lee, W. F.; Huang, C. T.; Chen, S. Y.; Chen, L. J. Free-Standing Single Crystal NiSi Nanowires with Excellent Electrical Transport and Field Emission Properties. J. Phys. Chem. C 2009, 113, 2286–2289. (9) Lu, K. C.; Wu, W. W.; Wu, H. W.; Tanner, G. M.; Chang, J. P.; Chen, L. J.; Tu, K. N. In situ Control of Atomic-Scale Si Layer with Huge Strain in the Nanoheterostructure NiSi/Si/NiSi through Point Contact Reaction. Nano Lett. 2007, 7, 2389–2394. (10) Weber, W. M.; Geelhaar, L.; Graham, A. P.; Unger, E.; Duesberg, G. S.; Liebau, M.; Pamler, W.; Cheze, C.; Riechert, H.; Lugli, P.; Kreupl, F. Silicon-Nanowire Transistors with Intruded Nickel-Silicide Contacts. Nano Lett. 2006, 6, 2660–2666. (11) Chou, Y. C.; Wu, W. W.; Chen, L. J.; Tu, K. N. Homogeneous Nucleation of Epitaxial CoSi2 and NiSi in Si Nanowires. Nano Lett. 2009, 9, 2337–2342. (12) Chou, Y. C.; Wu, W. W.; Cheng, S. L.; Yoo, B. Y.; Myung, N.; Chen, L. J.; Tu, K. N. In-situ TEM Observation of Repeating Events of Nucleation in Epitaxial Growth of Nano CoSi2 in Nanowires of Si. Nano Lett. 2008, 8, 2194–2199. (13) Wen, C.-Y.; Reuter, M. C.; Tersoff, J.; Stach, E. A.; Ross, F. M. Structure, Growth Kinetics, and Ledge Flow during Vapor-Solid-Solid Growth of Copper-Catalyzed Silicon Nanowires. Nano Lett. 2010, 10, 514–519. (14) Hofmann, S.; Sharma, R.; Wirth, C. T.; Cervantes-Sodi, F.; Ducati, C.; Kasama, T.; Dunn-Borkowsli, R. E.; Drucker, J.; Bennett, P.; Robertson, J. Ledge-Flow-Controlled Catalyst Interface Dynamics during Si Nanowire Growth. Nat. Mater. 2008, 7, 372–375. (15) Yeh, P. H.; Yu, C. H.; Chen, L. J.; Wu, H. H.; Liu, P. T.; Chang, T. C. Low-Power Memory Device with NiSi2 Nanocrystals Embedded in Silicon Dioxide Layer. Appl. Phys. Lett. 2005, 87, 193504. (16) Zhu, S. Y.; Yu, M. B.; Lo, G. Q.; Kwong, D. L. Near-Infrared Waveguide-Based Nickel Silicide Schottky-Barrier Photodetector for Optical Communications. Appl. Phys. Lett. 2008, 92, 081103-1-3. (17) Ma, D.; Chi, D. Z.; Loomans, M. E.; Wang, W. D.; Wong, A. S. W.; Chua, S. J. Kinetics of NiSi-to-NiSi2 Transformation and Morphological Evolution in Nickel Silicide Thin Films on Si(001). Acta Mater. 2006, 54, 4905–4911. (18) Schmidt, V.; Senz, S.; Gosele, U. Diameter-Dependent Growth Direction of Epitaxial Silicon Nanowires. Nano Lett. 2005, 5, 931–935. (19) Cui, Y.; Lauhon, L. J.; Gudiksen, M. S.; Wang, J.; Lieber, C. M. Diameter-Controlled Synthesis of Single-Srystal Silicon Nanowires. Appl. Phys. Lett. 2001, 78, 2214–2216. (20) Kodambaka, S.; Hannon, J. B.; Tromp, R. M.; Ross, F. M. Control of Si Nanowire Growth by Oxygen. Nano Lett. 2006, 6, 1292–1296.

JP108686Y