TMS Home Page Presenting a Web-Enhanced
Feature Article from JOM

View Current Issue




Overview: Phase Transformations Vol. 57, No.9, pp. 24-31

Metal Silicides: An Integral Part of Microelectronics


About the September 2005 Issue





View the Print Version
The print and/or PDF versions of the article can be acquired via the TMS Document Center.




Figure 1
Figure 1. A cross-section scanning electron microscope image of a six-level metal backend structure. (Courtesy UMC.)




Figure 2
Figure 2. A TEM image of amorphous interlayer at the Ti/(001)Si interface in an as-deposited sample.




Figure 3
Figure 3. A schematic diagram showing the formation of multiphases in a Ti/Si sample.




Figure 4
Figure 4. An atomic resolution TEM image of Si/TbSi2/Si heterostructure with simulated images pasted for direct comparison.




Figure 5
Figure 5. A planview TEM image of an Ni(2 nm) /a-Si(2 nm)/Si0.7Ge0.3 sample annealed at 600°C for 1 h.




Figure 6
Figure 6. A planview TEM image of an Ni(7 nm)/a-Si(13 nm)/Si0.7Ge0.3 sample annealed at 500°C for 1 h.




Figure 7
Figure 7. A scanning tunneling microscope image (100 nm × 100 nm) of submonolayer titanium deposited on Si(111) 7×7 surface at 700°C.




Figure 8
Figure 8. An SEM image of NiSi2 nanowires on blank (001)Si substrate by reactive deposition epitaxy at 730°C.




Figure 9
Figure 9. An SEM image of NiSi2 nanowires on nitride-capped (001)Si substrate by reactive deposition epitaxy at 730°C.




Figure 10
Figure 10. Plots of length and width of NiSi2 islands versus island area.




Figure D
Figure A. A cross-section transmission electron microscope image of a 0.1 μm TiSi2 salicide structure.


Questions? Contact
2005 The Minerals, Metals & Materials Society

This article presents an overview of the recent developments in the fundamental understandings and microelectronics applications of metal silicides. The synthesis and characterization of nanoscale silicides with potential applications in nanotechnology are reviewed.


Metal silicide thin films are integral parts of all microelectronics devices. They have been used as ohmic contacts, Schottky barrier contacts, gate electrodes, local interconnects, and diffusion barriers. With advances in semiconductor device fabrication technology, the shrinkage in line width continues at a fast pace. The International Technology Roadmap for Semiconductors (ITRS) predicted that in 2005, in the 90 nm generation devices, the gate length and thickness of silicide at the contact window would be 32 nm and 20 nm, respectively. In the year 2007, for the 65 nm generation devices, these numbers are predicted to further decrease to 25 nm and 17 nm, respectively.1

"Interconnectors provide flexibility in circuit design and substantial reduction in die size, and, thus, chip cost. "

In addition, more transistors will be incorporated in one chip. However, owing to the demand for increased integration level, the surface area will not be adequate to meet the interconnect demand. Multi-level interconnections provide flexibility in circuit design and a substantial reduction in die size and, thus, chip cost. Figure 1 shows a scanning electron microscope (SEM) cross section of a six-level metal backend structure. Electrical connection between the various metal layers is provided by vertical interconnects commonly referred to as vias. See the sidebar for device application details.


For metallization of integrated circuit (IC) devices, transition metal silicides, including near-noble and refractory metal silicides, are used. The general requirements are: low resistivity; good adhesion to silicon; low contact resistance to silicon; appropriate Schottky barrier height or Ohmic with heavily doped silicon (n+ or p+); thermal stability; appropriate morphology for subsequent lithography and etching; high corrosion resistance; oxidation resistance; good adhesion to and minimal reaction with SiO2; low interface stress, compatible with other processing steps such as lithography and etching, minimizing metal penetration; high electromigration resistance; and formability at low temperature. The requirements are rather stringent and at present, only three silicides, TiSi2, CoSi2, and NiSi, are being considered for metal contacts for advanced devices.2

PtSi and Pd2Si were used early on for metal contacts to lower the contact resistance of aluminum alloys as well as to serve as a diffusion barrier layer between aluminum alloy film and silicon. In the early 1980s, as the linewidth decreased to about 1 μm, many refractory metal silicide films, such as MoSi2, WSi2, TiSi2, and TaSi2 were used by different manufacturers. For the 0.25 μm technology, TiSi2 was almost used exclusively.3 For devices with linewidth of 0.18 µm or smaller, TiSi2, CoSi2, and NiSi are possible candidate contact materials.4,5

Many different deposition techniques can be used to deposit metal thin films. Currently, sputtering is used almost exclusively to deposit metal layers for contacts or in the self-aligned silicidation (salicide) process. Figure A shows a self-aligned TiSi2, which was formed on source, drain, and gate simultaneously. On the other hand, chemical vapor deposition of WSix and tungsten films is the dominant method to form gate electrodes or local interconnects and metal plugs, respectively.

The usual steps to form a silicide begin with the cleaning of the wafers consecutively by organic solution, dilute hydrochloric acid (HF), and deionized water. The wafers are blown dry with a nitrogen gun or in a “spin-rinse-dry” process. An alternative is to dip the wafer in dilute HF then blow dry with a nitrogen gun or “spin dry.” The wafers are immediately placed in the metal deposition chamber and the surface is sputter-cleaned by argon ions if necessary (argon sputtering may cause particle issue). Next, metal thin films are deposited on silicon at room temperature or at a higher temperature, and finally, the wafers are heat treated either by traditional furnace annealing or by rapid thermal annealing to form silicides.

Prior to the deposition of metal thin films, a 1.5-nm to 2-nm-thick SiO2 layer was usually present at the silicon substrate surface following the etching of the thermal oxide. It is necessary for the contact metal layers to penetrate the thin oxide layer to react with the silicon to form silicides. Titanium and nickel atoms are capable of penetrating through the thin oxide. On the other hand, cobalt atoms have difficulty forming silicide with silicon if a thin oxide layer is present at the interface. An argon ion sputter-cleaning step is usually required. Since CoSi2 is widely used in devices with linewidths of 0.18 μm or smaller, the formation of CoSi2 is used as an example to illustrate the steps to form silicides on silicon. The deposition of cobalt thin films by sputtering is kept at room temperature. A mixture of Co2Si and CoSi is formed at 300°C. CoSi2 forms at 550°C.4 For rapid thermal annealing, the first-step and second-step annealings are conducted at 500–550°C for 30–60 s and 700–850°C for 30–60 s, respectively.


The impetus for the study of silicide formation on silicon was stimulated by the expectation of device applications of silicides in the late 1970s and early 1980s. Two review chapters have succinctly summarized the knowledge accumulated up to the early 1980s.6,7 This article focuses on the most important developments in recent years.

Solid-State Amorphization
In device applications, interfacial reactions of metal thin films with silicon are rather peculiar in that polycrystalline metal film reacts with single-crystal silicon. The substrate is covalently bonded and the thin film is metallic. As a result, the microstructure of the silicide film and orientation of the substrate may play an important role in influencing the reaction. Some silicides can form at a temperature as low as 100°C. The mechanism for the break up of silicon bonds at such a low temperature is rather intriguing.7 Furthermore, the silicide phases formed at relatively low temperature are apparently related more to the growth kinetics than they are dictated by the thermodynamic consideration.

The formation of an amorphous interlayer (a–interlayer) by solid-state diffusion in diffusion couples has been one of the most challenging problems in condensed matter physics in recent years. The a-interlayer has been found to occur in all refractory metal/silicon and a number of rare-earth (RE) metal and platinum-group metal and crystalline silicon systems. A systematic survey and review of extensive studies on the subject in the past years showed:

  • A negative heat of mixing provides the driving force for the reaction and fast diffusion of one component in the other preempts the formation of crystalline compounds
  • The growth follows a linear law at the initial stage with activation energy around 1–1.5 eV for refractory metal/silicon systems and 0.5 eV for RE metal/silicon systems
  • The dominant diffusing species is silicon
  • The stability of the amorphous interlayer depends on the composition
  • Multiphases are present simultaneously in the initial stage of metal/ silicon interaction
  • Good correlations exist between physical parameters and kinetic data

From the investigation of amorphous interlayers, mechanisms of roughing of epitaxial RE silicide/(001)silicon interface, formation of stacking faults, and pinholes in RE silicides have gained in basic understanding. The insight led to successful growth of a pinhole-free epitaxial RE silicide layer on (111)Si. Furthermore, the enhanced formation of technologically important C54-TiSi2 by high-temperature sputtering, a thin interposing molybdenum layer, and tensile stress can all be explained involving some aspects of the amorphous interlayers.8

The First Nucleated Phase and Simultaneous Occurrence of Multiphases
In the transition metal-silicon binary phase diagrams, three or more silicide phases usually can be found. However, only selective phases are detected after thermal annealing of metal thin films on silicon. From x-ray diffraction and Rutherford backscattering spectrometry data, it was concluded initially that only one phase grows at a time for a clean system. This is consistent with the assertion that the formation of silicides is determined more by the growth kinetics than by energetics. However, more refined analysis by high-resolution transmission-electron microscopy (HRTEM) in conjunction with the fast Fourier transform analysis as well as auto-correlation function analysis indicated that formation of multiphases occurred in a number of refractory metal/silicon systems.8–11

In the Ti/Si system, Ti5Si3, located at the Ti/a–interlayer interface was identified to be the first nucleated phase.9 Ti5Si3, Ti5Si5, TiSi, and C49-TiSi2, along with an amorphous interlayer, were observed to be present simultaneously in samples annealed at higher temperatures.10 Examples are shown in Figures 2 and 3. Similar results were obtained for many refractory metal-silicon systems.8 For the near-noble silicides, a complex formation sequence was also found recently. The complex sequence of nickel silicide formation has been observed with the sheet resistance measurements combined with in-situ x-ray and light-scattering measurements in a synchrotron radiation facility.5

Growth Kinetics of Silicides
Kinetic data are crucial for a basic understanding of interfacial reactions between metal thin films and silicon. Most silicides are formed at a temperature far lower than the eutectic temperature. The growth is often diffusion controlled or interface-reaction controlled. The thickness of the silicide is proportional to the square root of time t and t, respectively. The presence of contaminating or doping impurities was found to influence the growth rate. For platinum films deposited in ultrahigh vacuum, the growth rate of PtSi was found to increase significantly. However, the growth law remained the same.12

Cross-section transmission electron microscopy (XTEM) has been demonstrated to provide direct and accurate kinetic data, such as the sequence of phase formation, the dependence of the phase growth, and morphology of phase and interface structure in the growth of silicides on silicon.13

In TiSi2, CoSi2, NiSi2, and a number of RE silicides, the silicide formation took place within a narrow temperature range and nucleation was suggested as a controlling mechanism.14 The nucleation effects are eliminated when these phases are formed on an amorphous layer.15 The importance of nucleation effects in silicide formation has been discussed extensively by d’Heurle.14 The films produced from nucleation-limited reactions are often rather rough.

Dominant Diffusing Species
In the silicide formation, metal atoms diffuse across the metal/silicide interface, silicon atoms diffuse across the silicide/silicon interface, or both. In order to determine the dominant diffusing species, it is common to introduce an inert marker. In thin film reactions, the markers are usually tens of nanometers in size and should not influence the growth kinetics of silicide formation. Ideally, the markers should be inert and remain immobile as the diffusing species streams by. An additional constraint is that the marker should be located in the silicide layer to avoid possible influence due to the presence of the interface.7

From the marker experiments, it was revealed for metal-rich silicides such as M2Si, the dominant diffusing species are mostly metal atoms. On the other hand, in the formation of monosilicide and disilicide, silicon atoms are generally the dominant diffusing species. However, there are exceptions. Important silicides in ultralarge-scale integrated-circuit technology, the dominant diffusing species in the growth of TiSi2, CoSi2, WSi2, and NiSi are Si, Co, Si, and Ni, respectively.6,7,14 For the TiSi2 salicide process, if the temperature, time, and ambient for the rapid thermal annealing were not optimized, C49-TiSi2 and/or C54-TiSi2, which are not easily removed by ammonia and peroxide solution, are prone to form on the dielectric sidewall between the poly-gate and source/drai. This results in the so-called bridging problem, which may lead to device failure. Since cobalt is the dominant diffusing species in the formation of CoSi2, the bridging problem is less troublesome in the CoSi2 salicide technology.

Epitaxial Growth of Silicides
Epitaxial silicides belong to a special class of silicides that exhibit a definite orientation relationship with respect to the silicon substrate. A silicide is expected to grow epitaxially on silicon if the crystal structures are similar and the lattice mismatch between them is small. The impetus for the study of epitaxial silicides mainly stemmed from several favorable characteristics of epitaxial silicides in comparison with their polycrystalline counterparts, including greater stability and a lower stress at the interface, alleviation of grain boundary effects, as well as conductivity enhancement.16

NiSi2 and CoSi2 can be grown in single-crystal form on silicon.17 Many hexagonal RE silicides have been grown on Si(111) for the almost perfect lattice matches between RE silicide (0001) and Si(111) planes. Furthermore, on top of the silicide layer, a single-crystal silicon layer can be grown. An example of the Si/TbSi2/Si heterostructure is shown in Figure 4.18 On the other hand, almost all transition metal silicides can be grown epitaxially on silicon to a certain extent. In particular, FeSi and TiSi2 can be grown to tens of micrometers in grain size.16

Initial studies on the epitaxial growth of silicides on silicon were mostly on the growth of silicides on a large area. However, in device applications, silicides were grown on laterally confined silicon. Lateral confinement was found to exert significant influence on the epitaxial growth of NiSi2 and CoSi2 on silicon.19–21 The epitaxial silicides were relevant to the device applications as the contact size shrank to sub-100 nm.

In an Ni/(001)Si system, low-resistivity NiSi is at the center of attention in device applications. In nickel on blank (001)Si, NiSi is formed and stable at 350–700°C.6 It has been reported that dopants do not affect NiSi formation.22 However, striking effects of B+ and BF2+ implantation on the growth of epitaxial NiSi2 on silicon were observed. As a result of ion implantation into (001)Si, epitaxial NiSi2 was found to grow at 200–280°C instead of the usual formation temperature of about 800°C on blank (001)Si. Both boron and fluorine atoms introduced by ion implant into silicon were found to promote the epitaxial growth of NiSi2 on silicon at low temperatures. Good correlation was found between the atomic size factor and the resulting stress and NiSi2 epitaxy at low temperatures. The final structure of the silicide layer was found to depend critically on the thickness of the starting nickel overlayer and the annealing temperature. The amorphicity of the substrate apparently played an important role in promoting the formation of polycrystalline NiSi2 at low temperatures.23–25


Nanoscale silicides are named nanosilicides. As the integrated circuit industry moves into the nano-era, metal silicide contacts are naturally falling into this category. On the other hand, many efforts have been made to fabricate nanosilicides employing the bottom-up approach without elaborate microlithography.

Quantum dots are envisioned to be useful in devices such as single-electron transistors, high-density memories, light emission, semiconductor lasers, and tunnel diodes.26 In principle, any ultrathin (~ 1 nm) silicide forming metal film may react with silicon substrate to form silicide nanodots under appropriate annealing conditions. Other means, such as ion implantation of metal ions into silicon nanowires followed by annealing, may also produce silicide nanoparticles.27 To meet the requirements of microelectronics and optoelectronics, it is imperative to control the size, density, and ordering of the dots.

Self-assembly is an attractive nanofabrication technique because it provides the means to precisely engineer structures on the nanometer scale over large sample areas. Self-organizing nanocrystal assemblies have already shown the degree of control necessary to address the challenges of building nanometer-scale technologies.28

Self-Assembled Low-Resistivity Metal Silicide Quantum Dot Arrays on Epitaxial Si0.7Ge0.3 on (001)Si
Si1–xGex/Si heterostructures are used to fabricate high-speed transistors that extend the range of applications of silicon technology.29 Self-assembled NiSi quantum-dot arrays have been grown on relaxed epitaxial Si0.7Ge0.3 on (001)Si. The formation of the one-dimensional (1-D) ordered structure is attributed to the nucleation of NiSi nanodots on the surface undulations induced by step bunching on the surface of SiGe film. This results from the miscut of the wafers from normal to the (001)Si direction. The two-dimensional (2-D), pseudohexagonal structure was achieved under the influence of repulsive stress between nanodots. Since the periodicity of surface bunching can be tuned with appropriate vicinality and misfit, the undulated templates promise to facilitate the growth of ordered silicide quantum dots with selected periodicity and size.30

Figure 5 shows a planview TEM micrograph of an Ni(2 nm)/a-Si(2 nm)/ Si0.7Ge0.3 sample revealing the ordered, equally spaced NiSi dot arrays, oriented along the [110] surface direction. The apparent 1-D alignment and less ordered 2-D arrangement features rule out the direct influence of the misfit-dislocation strain. The average size of nanodots and spacings between adjacent arrays are about 15 nm and 20–40 nm, respectively. In contrast, NiSi nanodots in Ni/a-Si/ Si(001) samples were found to be randomly distributed. It indicated that the use of an Si0.7Gex/Si heterostructure template induces the highly ordered alignment of NiSi dots.

A close look at the Si0.7Ge0.3/(001)Si and Ni(2 nm)/a-Si(2 nm)/Si0.7Ge0.3 surfaces with HRTEM indeed revealed the presence of the atomic steps, about 5–20 nm in spacing and 10 nm in average spacing. The HRTEM images further showed that the irregularity in step spacing indicating the presence of step bunching. In a particular instance, the nanodot arrays, about 100–800 nm apart, were found to align with the cross-hatch pattern in a 500°C annealed Ni(7 nm)/a- Si(13 nm)/Si0.7Ge0.3 sample, as shown in Figure 6. The nanodots tended to be connected along individual arrays. The alignment of nanodots is apparently under the influence of the strain fields associated with the cross-hatch patterns. It is conjectured that the alignment with the cross-hatch pattern is most prominent in places where step bunching is of low density and exerts weak influence on the formation of nanodot patterns. Similarly, CoSi2 and TiSi2 nanodot arrays were formed.31

Formation of Epitaxial β-FeSi2 Nanodot Arrays on Strained Si/ Si0.8Ge0.2 (001) Substrate
Epitaxial β-FeSi2 nanodots were grown on strained Si/Si0.8Ge0.3 (001) substrates by the solid-phase epitaxy method. High-quality β-FeSi2 nanodots were grown at 800°C by employing strained Si/Si0.8Ge0.2 substrates, owing to a decrease of the in-plane lattice mismatch between the lattice spacing of the β-FeSi2 [001] and [010] directions and that of a silicon substrate. Ordered β-FeSi2 arrays along <110> direction were observed to form on surfaces of strained Si/Si0.8Ge0.2 substrate. It is shown that dislocation slip originating from compositionally graded Si1–xGex layers can produce local surface-strain and local thickness variation. The surface features are used for the fabrication of epitaxial β-FeSi2 nanostructures on strained Si/Si0.8Ge0.2 substrate.32

First Nucleated Phase and the Dominant Diffusing Species
Atomic resolution techniques have been successful in studying nanoscale silicides. A particularly pertinent example is seen in the identification of Ti5Si4 as the first nucleated phase in submonolayer titanium deposited on the Si(111)-7×7 surface by ultrahigh vacuum scanning tunneling microscopy in conjunction with atomic-resolution TEM. The direct observation of the formation of clusters surrounded by the heavily damaged silicon lattice strongly suggested that silicon is the dominant diffusing species in forming the silicide. An example is shown in Figure 7.33

One-dimensional building blocks, such as nanowires and nanotubes, are especially attractive candidates around which to develop a bottom-up paradigm for nanotechnology-enabled architectures. As opposed to zero-dimensional nanocrystals, which have been the subject of intense study but are challenging to contact electrically, nanowires and nanotubes can act both as interconnects for the transport of charge carriers as well as active device elements.34,35 Nanowires are intrinsically suitable as highly sensitive sensor elements, due to their high surface/volume ratio and the extreme sensitivity of 1-D transport to gating fields or adsorbates.

Self-Assembled Nanowires
Self-assembled silicide nanowires are envisioned to possess advantages of perfect single crystallinity, metallic resistivity, compatibility with silicon device processing, and high thermal stability. A large number of self-assembled epitaxial silicide nanowires were investigated in the past.36–44 Many RE silicide nanowires were grown on silicon substrates. These RE nanowires are commensurate with nearly perfect lattice match in their long direction and are limited in their ability to grow coherently with the substrate in the lateral direction). For PtSi on Si(001), the long direction is aligned with [001]PtSi direction in parallel with the Si(220) plane with smaller lattice mismatch.45 On the other hand, for the growth of C49-TiSi2, the range of structural variants argues against a simple interface-energy explanation.40 It is, however, interesting that TiSi2 nanowires are incommensurate (8%) in their long direction.46 However, the interface structure for the nanowires may not be the same as that inferred from the bulk lattices.

"The self-assembly of nanowires usually requires that the substrate be crystalline, precluding their use for many potential applications."

For systems of isotropic lattice mismatch, such as Ni/Si and Co/Si systems, the aspect ratio of nanowires in these systems was generally small and unsatisfactory for practical applications.42–44 Strained epitaxial layers may form while the interface between the overlayer and the substrate is commensurate. These layers are inherently unstable and have interesting properties, which are of importance in semiconductor devices. Two kinds of strained relief mechanism were recognized: one is the formation of dislocations and the other is shape transition.

In recent years, it has been recognized that shape changes such as island formation constitute a major mechanism for strain relief.42,47,48 Tersoff and Tromp reported that a strain-induced shape transition may occur. Below a critical size, islands have a compact symmetric shape. For larger sizes, they adopt a long thin shape that allows better elastic relaxation of the island’s stress.47 Experimental data on silicide island formation [e.g., Au4Si/Si(111)48 and CoSi2/Si(100)]48 also exhibited the elongated island growth. For the Ti/Si system, a series of phase transformations was reported in thin-film reactions.49 Titanium silicide islands of various shapes were observed.50 The shapes were found to depend on the thickness of titanium deposition and the thermal treatment process. A previous work showed that the formation of CoSi2 nanowires involved the mechanism of “endotaxy.”44 The twinning relationship with the substrate breaks the symmetry of the surface and leads to the asymmetric growth of islands. By combining the methods of reactive deposition epitaxy and nitridemediated epitaxy, the formation of high aspect ratio NiSi2 nanowires can be achieved. Examples are shown in Figures 8 to 10.51 The nanowires were successfully grown with high aspect ratios despite the four-fold symmetric epitaxial relationship between NiSi2 (of cubic CaF2 structure) and silicon (of diamond cubic structure). Nitride-mediated epitaxy was presented by Chong et al. to complement the use of oxide mediated epitaxy in promoting epitaxial growth of CoSi2 on (001)Si.52,53 The thin amorphous interlayer acts as a physical barrier to control the flux of metal atoms on the silicon substrate. Such a concept was used in the growth of self-assembled silicide nanowires to control the kinetic process during the growth. A similar effect is expected to be applicable to other strained epitaxial layer systems.

The challenges for self-assembled silicide nanowires are the control of aspect ratio and location. In addition, the self-assembly of nanowires usually requires that the substrate be crystalline, precluding their use for many potential applications.

Alternative Growth of Silicide Nanowires
Alternative approaches have been adopted to grow nanowires without relying on the mismatch between the nanowires and the substrate.

Wu et al. prepared single-crystal metallic NiSi nanowires using free-standing silicon nanowires as the template. NiSi nanowires were produced by annealing the nickel-metal-coated silicon nanowires at 550°C. They also prepared NiSi/Si nanowire heterostructures with NiSi formed using crossed Si/SiO2 core-shell nanowires as masks to define the lengths of the unreacted silicon regions. Electrical measurements show that the single-crystal nickel silicide nanowires have ideal resistivities of about 10μΩcm and remarkably high failure current densities. In addition, the nickel silicide/silicon (NiSi/Si) nanowire heterostructures have been used to produce field-effect transistors in which the source–drain contacts are defined by the metallic NiSi nanowire regions.33 On the other hand, carbon-coated NiSi nanowires were prepared in a radio-frequency-induction heating chemical vapor deposition reactor. The growth of the NiSi nanowires and the coating of the nanowires with carbon layers simultaneously took place in the reaction. The nanowires were more than 10 μm long and with an average diameter of 20–40 nm. The resistivity of individual NiSi nanowire was about 370μΩcm at room temperature, indicating the presence of considerable impurities and/or defects.54 Nickel silicide nanowires were also grown on nickel surfaces by decomposition of silane at 320–420°C. Depending on the growth conditions, single-phase Ni2Si, Ni3Si2, and NiSi nanowires were formed. It has been demonstrated that directed growth of silicide nanowires can be achieved with the aid of applied electric field.55

Xiang et al. used a vapor-phase deposition method to grow TiSi2 nanowires on silicon wafers. Field emission and cathodoluminescence measurements reveal the potential applications in vacuum microelectronics.56

TaSi2 nanowires have been synthesized by annealing FeSi2 thin film and nanodots grown on silicon substrate in an ambient containing tantalum vapor. The TaSi2 nanowires are formed in three steps: segregation of silicon atoms from the FeSi2 underlayer to form a silicon base, epitaxial growth of TaSi2 nanodots on a silicon base, and elongation of the TaSi2 nanowire along the growth direction. Strong field emission properties promise future electronics and optoelectronics applications.57


The research was supported by the Republic of China National Science Council through grant No. NSC 93-2215- E-007-011 and Ministry of Education grant No. 91-E-FA04-1-4.


1. The International Technology Roadmap for Semiconductors, 2004 Update, Semiconductor Industry Association (San Jose, CA: Semiconductor Industry Association, 2004),
2. L.J. Chen, editor, Silicide Technology for Integrated Circuits (London: IEE, 2004).
3. Z. Ma and L.H. Allen, “Titanium Silicide Technology,” in Ref. 2, pp. 49–76.
4. T. Kikkawa, K. Inoue, and K. Imai, “Cobalt Silicide Technology,” in Ref. 2, pp. 77–94.
5. C. Lavoie, C. Detavernier, and P. Besser, “Nickel Silicide Technology,” in Ref. 2, pp. 95–152.
6. K.N. Tu and J.W. Mayer, Thin Films-Interdiffusion and Reactions, ed. J.M. Poate, K.N. Tu, and J.W. Mayer (New York: Wiley, 1978), pp. 359–405.
7. M.A. Nicolet and S.S. Lau, Materials Process and Characterization, ed. N.G. Einspruch and G.R. Larrabee (New York: Academic, 1983), pp. 329–464.
8. L.J. Chen, “Solid-State Amorphization in Metal-Si Systems,” Mater. Sci. Engineering R, 29 (2000), pp. 115–152.
9. M.H. Wang and L.J. Chen, “Identification of the First Nucleated Phase in the Interfacial Reactions of Ultrahigh Vacuum Deposited Titanium Thin Films on Silicon,” Appl. Phys. Lett., 58 (1991), pp. 463–465.
10. M.H. Wang and L.J. Chen, “Simultaneous Occurrence of Multiphases in the Interfacial Reactions of Ultrahigh Vacuum Deposited Titanium Thin Films on Silicon,” Appl. Phys. Lett., 59 (1991), pp. 2460–2462.
11. J.M. Liang and L.J. Chen, “Auto-Correlation Analysis for the Determination of the Structure of Amorphous Interlayers in Ultrahigh Vacuum Deposited Molybdenum Thin Films on Silicon,” Appl. Phys. Lett., 64 (1994), pp. 1224–1226.
12. C.A. Crider and J.M. Poate, “Growth-Rates for Pt2Si and PtSi Formation Under UHV and Controlled Impurity Atmospheres,” Appl. Phys. Lett., 36 (1980), pp. 417–419.
13. J.Y. Cheng, H.C. Cheng, and L.J. Chen, “Cross-Sectional Transmission Electron Microscope Study of Growth Kinetics of MoSi2 on (001)Si,” J. Appl. Phys., 61 (1987), pp. 2218–2223.
14. F.M. d’Heurle, “Nucleation of a New Phase from the Interaction of Two Adjacent Phases: Some Silicides,” J. Mater. Res., 3 (1988), pp. 167–195.
15. L.S. Hung et al., “Kinetics of TiSi2 Formation by Thin Ti Films on Si,” J. Appl. Phys., 54 (1983), pp. 5076–5080.
16. L.J. Chen and K.N. Tu, “Epitaxial Growth of Metal Silicides on Silicon,” Mater. Sci. Reports, 6 (1991), pp. 53–140.
17. R.T. Tung, “Epitaxial CoSi2 and NiSi2 Thin-Films,” Mater. Chem. Phys., 32 (1992), pp. 107–133.
18. C.H. Luo, F.R. Chen, and L.J. Chen, “Atomic Structure of Si/TbSi2-x/(111)Si Double Heterostructure Interfaces,” J. Appl. Phys., 76 (1994), pp. 5744–5747.
19. C.S. Chang, C.W. Nieh, and L.J. Chen, “Formation of Epitaxial NiSi2 of Single Orientation on (111)Si inside Miniature Size Oxide Openings,” Appl. Phys. Lett., 50 (1987), pp. 259–261.
20. J.Y. Yew, L.J. Chen, and K. Nakamura, “Epitaxial Growth of NiSi2 on (111)Si inside 0.1-0.6 μm Oxide Openings Prepared by Electron Beam Lithography,” Appl. Phys. Lett., 69 (1996), pp. 999–1001.
21. J.Y. Yew et al., “Formation of CoSi2 on Selective Epitaxial Growth Silicon inside 0.1–0.6 μm Oxide Openings,” Appl. Phys. Lett., 69 (1996), pp. 3692– 3694.
22. T. Morimoto et al., “Self-Aligned Nickel-Mono-Silicide Technology for High-Speed Deep-Submicrometer Logic CMOS ULSI,” IEEE Trans. Electron Dev., 42 (1995), pp. 915–922.
23. S.W. Lu, C.W. Nieh, and L.J. Chen, “Epitaxial Growth of NiSi2 on Ion-Implanted Silicon at 250–280°C,” Appl. Phys. Lett., 49 (1986), pp. 1770–1772.
24. L.J. Chen et al., “The Effects of Implantation Impurities and Crystallinity on the Formation of Epitaxial NiSi2 on Silicon at 200–280°C,” J. Appl. Phys., 62 (1987), pp. 2789–2792.
25. W.J. Chen and L.J. Chen, “Interfacial Reactions in Nickel Thin Films on BF2+-Implanted (001)Si,” J. Appl. Phys., 70 (1991), pp. 2628–2633.
26. C.B. Murray et al., “Monodisperse 3D Transition- Metal (Co, Ni, Fe) Nanoparticles and Their Assembly into Nanoparticle Superlattices,” MRS Bull., 26 (2001), pp. 985–991.
27. C.P. Li et al., “Metal Silicide/Silicon Nanowires from Metal Vapor Vacuum Arc Implantation,” Adv, Mater., 14 (2002), pp. 218–221.
28. G.M. Whitesides and B. Grzybowski, “Self-Assembly at All Scales,” Science, 295 (2002), pp. 2418–2421.
29. C.W. Liu and L.J. Chen, “SiGe Heterostructures,” Encyclopedia of Nanoscience and Nanotechnology, Vol. 9, ed. H.S. Nalwa (Stevenson Ranch, CA: American Scientific Publishers, 2004), pp. 775–792.
30. W.W. Wu et al., “Self-Assembled NiSi Quantum-dot Arrays on Epitaxial Si0.7Ge0.3 on (001)Si,” Appl. Phys. Lett., 83 (2003), pp. 1836–1838.
31. L.J. Chen et al., “Nanostructures on Epitaxial SiGe Films on Silicon,” Electrochem. Soc. PV, 2004-02 (2004), pp. 241–252.
32. H.C. Chen et al., “Growth of Beta-FeSi2 Nanodots on Strained Si on Si-Ge,” Thin Solid Films, 461 (2004), pp. 44–47.
33. H.F. Hsu et al., “Identification of the First Nucleated Phase in Submonolayer Ti Deposited on Si(111)-7×7 by Atomic Resolution Techniques,” Ultramicroscopy, 100 (2004), pp. 347–351.
34. Y. Wu et al., “Single-Crystal Metallic Nanowires and Metal/Semiconductor Nanowire Heterostructures,” Nature, 430 (2004), pp. 61–65.
35. J.F. Lin et al., “Signatures of Quantum Transport In Self-Assembled Epitaxial Nickel Silicide Nanowires,” Appl. Phys. Lett., 85 (2004), pp. 281–283.
36. C. Preinesberger et al., “Formation of Dysprosium Silicide Wires on Si(001),” J. Phys. D: Appl. Phys., 31 (1998), pp. L43–L45.
37. Y. Chen et al., “Self-Assembled Growth of Epitaxial Erbium Disilicide Nanowires on Silicon (001),” Appl. Phys. Lett., 76 (2000), pp. 4004–4006.
38. J. Nogami et al., “Self-Assembled Rare-Earth Silicide Nanowires on Si(001),” Phys. Rev. B, 63 (2001), p. 233305.
39. Y. Chen, D.A.A. Ohlberg, and R.S. Williams,“Nanowires of Four Epitaxial Hexagonal Silicides Grown on Si(001),” J. Appl. Phys., 91 (2002), pp. 3213–3218.
40. M. Stevens et al., “Structure and Orientation of Epitaxial Titanium Silicide Nanowires Determined by Electron Microdiffraction,” J. Appl. Phys., 93 (2003), pp. 5670–5674.
41. W.C. Yang, H. Ade, and R.J. Nemanich, “Shape Stability of TiSi2 Islands on Si (111),” J. Appl. Phys., 95 (2004), pp. 1572–1576.
42. S.H. Brongersma et al., “Stress-Induced Shape Transition of CoSi2 Clusters on Si(100),” Phys. Rev. Lett., 80 (1998), pp. 3795–3798.
43. J.D. Carter, G. Cheng, and T. Guo, “Growth of Self-Aligned Crystalline Cobalt Silicide Nanostructures from Co Nanoparticles,” J. Phys. Chem. B, 108 (2004),pp. 6901–6904.
44. Z. He, D.J. Smith, and P.A. Bennett, “Endotaxial Silicide Nanowires,” Phys. Rev. Lett., 93 (2004), p. 256102.
45. K.L. Kavanagh, M.C. Reuter, and R.M. Tromp, “High Temperature Epitaxy of PtSi/Si(001),” J. Cryst. Growth, 173 (1997), pp. 393–401.
46. H.F. Hsu et al., “Shape Transition in the Initial Growth of Titanium Silicide Clusters on Si(111),” Jpn. J. Appl. Phys., 43 (2004), pp. 4541–4544.
47. J. Tersoff and R.M. Tromp, “Shape Transition in Growth of Strained Islands—Spontaneous Formation of Quantum Wires,” Phys. Rev. Lett., 70 (1993), pp. 2782–2785.
48. K. Sekar et al., “Shape Transition in the Epitaxial-Growth of Gold Silicide in Au Thin-Films on Si(111),” Phys. Rev. B, 51 (1995), pp. 14330–14336.
49. M.H. Wang and L.J. Chen, “Phase Formation in Ultrahigh Vacuum Deposited Titanium Thin Films on (001)Si,” J. Appl. Phys., 71 (1992), pp. 5918–5925.
50. K. Ezoe et al., “Scanning Tunnelling Microscopy Study of Initial Growth of Titanium Silicide on Si(111),” Appl. Surf. Sci., 130-132 (1998), pp. 13–17.
51. S.Y. Chen and L.J. Chen, unpublished work (2005).
52. R.K.K. Chong et al., “Nitride-Mediated Epitaxy of CoSi2 on Si(001),” Appl. Phys. Lett., 82 (2003), pp. 1833–1835.
53. R.T. Tung, “Oxide-Mediated Epitaxy of CoSi2 on Si(001),” Appl. Phys. Lett., 68 (1996), pp. 3461–3463.
54. K.S. Lee et al., “Anomalous Growth and Characterization of Carbon-Coated Nickel Silicide Nanowires,” Chem. Phys. Lett., 384 (2004), pp. 215–218.
55. C.A. Decker et al., “Directed Growth of Nickel Silicide Naowires,” Appl. Phys. Lett., 84 (2004), pp. 1389–1391.
56. B. Xiang et al., “Synthesis and Field Emission Properties of TiSi2 Nanowires,” Appl. Phys. Lett., 86 (2005), pp. 243101–243103.
57. Y.L. Chueh et al., “Synthesis and Characterization of Metallic TaSi2 Nanowires,” unpublished work (2005).

L.J. Chen is Ministry of Education National Chair Professor of the Department of Materials Science and Engineering at National Tsing Hua University in Hsinchu, Taiwan.

For more information, contact L.J. Chen, National Tsing Hua University, Department of Materials Science and Engineering, Hsinchu, Taiwan, +886-3-573-1166; fax +886-3-571-8328; e-mail