Table of Contents
Fetching ...

Topology-Directed Silicide Formation: An Explanation for the Growth of C49-TiSi$_2$ on the Si(100) Surface

Lukas Hückmann, Jonathon Cottom, Jörg Meyer, Emilia Olsson

TL;DR

This study explains why metastable $C49$-TiSi$_2$ forms first at Ti/Si(100) interfaces by showing that surface topology of the reconstructed Si(100) drives Ti adsorption into a TiSi bilayer that templates the $C49$ phase. Through density-functional theory, it reveals that initial Ti adsorption on $c(4\times2)$ Si(100) is site-selective, with strong Ti–Ti pairing in the first subsurface layer lifting Si dimers and enabling bilayer formation. The resulting TiSi bilayer reproduces characteristic motifs of $C49$-TiSi$_2$ and provides a nucleation seed that explains the Stranski–Krastanov growth mode and the preferential formation of $C49$ over $C54$. The work also suggests surface-disruption or refractory overlayers as practical routes to suppress $C49$ formation, guiding more controllable TiSi$_2$ integration in future devices."

Abstract

Designing metal-semiconductor junctions is essential for optimizing the performance of modern nanoelectronic devices. A widely used material is TiSi$_2$, which combines low electronic resistivity with good endurance. However, its multitude of polymorphs continues to pose a challenge for device fabrication. In particular, the naturally occurring formation of the metastable C49-TiSi$_2$ modification remains poorly understood and is problematic due to its unfavorable electronic properties. Based on extensive DFT calculations, we present a comprehensive model of Ti adsorption on Si(100) that highlights the pivotal role of surface topology for the initial stages of the interfacial TiSi$_2$ formation process. We show that the interplay between Si surface dimers, the symmetry of the Si(100) surface, and the incorporation of Ti adsorbates below the surface drives an adsorption pattern that yields a nucleation template for the C49-TiSi$_2$ phase. Our atomistic model rationalizes experimental observations like the Stranski-Krastanov growth mode, the preferential formation of C49-TiSi$_2$ despite it being less favorable than the competing C54 phase, and why disruption of the surface structure restores thermodynamically driven growth of the latter. Ultimately, this novel perspective on the unique growth of TiSi$_2$ will help to pave the way for next-generation electronic devices.

Topology-Directed Silicide Formation: An Explanation for the Growth of C49-TiSi$_2$ on the Si(100) Surface

TL;DR

This study explains why metastable -TiSi forms first at Ti/Si(100) interfaces by showing that surface topology of the reconstructed Si(100) drives Ti adsorption into a TiSi bilayer that templates the phase. Through density-functional theory, it reveals that initial Ti adsorption on Si(100) is site-selective, with strong Ti–Ti pairing in the first subsurface layer lifting Si dimers and enabling bilayer formation. The resulting TiSi bilayer reproduces characteristic motifs of -TiSi and provides a nucleation seed that explains the Stranski–Krastanov growth mode and the preferential formation of over . The work also suggests surface-disruption or refractory overlayers as practical routes to suppress formation, guiding more controllable TiSi integration in future devices."

Abstract

Designing metal-semiconductor junctions is essential for optimizing the performance of modern nanoelectronic devices. A widely used material is TiSi, which combines low electronic resistivity with good endurance. However, its multitude of polymorphs continues to pose a challenge for device fabrication. In particular, the naturally occurring formation of the metastable C49-TiSi modification remains poorly understood and is problematic due to its unfavorable electronic properties. Based on extensive DFT calculations, we present a comprehensive model of Ti adsorption on Si(100) that highlights the pivotal role of surface topology for the initial stages of the interfacial TiSi formation process. We show that the interplay between Si surface dimers, the symmetry of the Si(100) surface, and the incorporation of Ti adsorbates below the surface drives an adsorption pattern that yields a nucleation template for the C49-TiSi phase. Our atomistic model rationalizes experimental observations like the Stranski-Krastanov growth mode, the preferential formation of C49-TiSi despite it being less favorable than the competing C54 phase, and why disruption of the surface structure restores thermodynamically driven growth of the latter. Ultimately, this novel perspective on the unique growth of TiSi will help to pave the way for next-generation electronic devices.
Paper Structure (9 sections, 2 equations, 5 figures, 4 tables)

This paper contains 9 sections, 2 equations, 5 figures, 4 tables.

Figures (5)

  • Figure 1: a) Schematic of the Ti adsorption sites on the $c(4\times2)~\ce{Si}(100)$ surface (top view). The numbered blue circles in the inset represent Ti adatoms, the orange and yellow ones represent buckled Si-dimers with the orange ones being elevated, and the gray, the dashed-hollow, and the dotted hollow are the Si atoms of first, second, and third subsurface layer, respectively. Depiction of a Ti atom adsorbed at position b) 1 and c) 6, respectively, where Si is colored yellow, Ti blue.
  • Figure 2: a) Sketch of the $c(4\times2)~\ce{Si}(100)$ surface to illustrate the direction-dependent pairwise arrangement of Ti adsorbates with the sites on the Si dimer ridge being colored blue, those in the trench being red, and those across along the [1$\bar{1}$0]-direction being purple. The colored stripes are added to guide the eye. Subsurface Si atoms are omitted for visibility. b)/c) Depiction of two Ti atoms at position b) $\ce{Ti}^{(\mathbf{1})}\xspace$-$\ce{Ti}^{(\mathbf{1}')}$ and c) $\ce{Ti}^{(\mathbf{1})}\xspace$-$\ce{Ti}^{(\mathbf{6})}\xspace$, respectively, where Si is colored yellow, and Ti is blue.
  • Figure 3: a) Adsorption energy ($E_\text{ads}$) and b) the work function $\Phi$ as a function on the Ti coverage. The adsorbate coverage is given in monolayers (ML) defined via the number of adsorbates per surface unit cell and in atoms per surface area ($N_\mathrm{Ti}A^{-1})$. The error bars indicate the highest/lowest $E_\text{ads}$ and the boxes are the standard deviation of $E_\text{ads}$ of the ensemble of configurations. For comparison, the $\Phi$ of pristine Si(100) and $\text{C49-}\ce{TiSi2}$(010) is added as dashed and dotted line, respectively. The hollow diamonds are UPS data measured by Taubenblatt and Helms.taubenblatt_silicide_1982 c)-e) Three representative examples of the slab in c) the dilute limit, d) at low concentration, and e) at medium coverage (from left to right). All slabs are displayed along the [110]-axis.
  • Figure 4: Depiction of a slab covered by two monolayers of Ti, with close-up views of the bilayer along the $[110]$ and $[1\bar{1}0]$ directions, together with a $3\times1\times3$$\text{C49-}\ce{TiSi2}$ cell for comparison. The blue lines outline the $1\times1$ Si(100) surface unit cell.
  • Figure 5: a) The $c(4\times2)~\ce{Si}(100)$ slab viewed from the [110]-direction b) Schematic top view of the $c(4\times2)~\ce{Si}(100)$ slab with the adsorption sites and the reduced cell used for the exhaustive sampling with icet code.angqvist2019