Search

Silicide materials used as contacts in CMOS devices have evolved over many technology nodes. This article traces the often forgotten defectivity related reasons that were the primary drivers for a change in materials or process flow - evolving from Ti to Co and Ni silicides, and the more recent return to Ti-based liner silicides. The criteria used for the selection of these metal silicides have undergone a dramatic change with the advent of 3-D transistors and trench silicide contacts, and is now primarily guided by the value of interfacial contact resistivity (rc). Furthermore, using results from synchrotron X-ray diffraction and pole-figure analysis, we present how phase formation and microstructure of contacts vary with Ti thickness, alternative annealing treatments, and substrate composition and orientation. We show that microstructure in very thin films can change from amorphous to epitaxial, a factor likely to become important for the contacts in upcoming generations of devices.

Reduction of specific contact resistivity is one of the key items for downscaling of device feature size. This paper discusses an accurate method of measuring specific contact resistivity by using our simple four-point probe test structure having a small area, the interface resistance of which is to be measured, fabricated on a substrate that has low sheet resistance. We also show its application for direct specific contact resistivity measurements of several systems that have dopant dependence, i.e., the NiPtSi/Si, NiPtSi/SiGe, and Pt silicide systems, as well as the metal-metal system between silicide and contact plug metal.

The texture of thin films, originating from a solid state reaction between a deposited film and a single crystal substrate is investigated. The relation between the phase formation and texture is analyzed for a number of these systems, such as Co/Si, Ni/Si or Co/Ge, where a metal film is allowed to react with a semiconductor single crystal substrate during heating and a summary of these results in presented in this article. It was found that the texture of the resulting films can be very complex, consisting of a variety of simultaneously occurring texture components such as epitaxy, fiber and axiotaxy texture. The close connection between the phase formation and texture is demonstrated by the fact that even a small intervention in either one, can have a huge effect on the resulting phase and/or its texture. From this, we show that the effect of the addition of ternary elements (e.g. Pt, W, C) to the thin films can only be understood if one considers its effects on both the kinetics and the thermodynamics of the reactions, as well as on the texture of the phases. We show how this can be used to influence technologically important properties of the films, such of formation temperature or stability.

Smooth, continuous, and highly conformal nickel nitride films were deposited by direct liquid injection (DLI)-chemical vapor deposition (CVD) using a solution of bis(-di--butylacetamidinato)nickel(II) in tetrahydronaphthalene as the nickel (Ni) source and ammonia as the coreactant gas. The DLI-CVD films grown on HF-last (100) silicon and on highly doped polysilicon substrates served as the intermediate for subsequent conversion into nickel silicide (NiSi), which is a key material for source, drain, and gate contacts in microelectronic devices. Rapid thermal annealing in the forming gas of DLI-CVD films formed continuous NiSi films at temperatures above . The resistivity of the NiSi films was , close to the value for bulk crystals. The NiSi films have remarkably smooth and sharp interfaces with underlying Si substrates, thereby producing contacts for transistors with a higher drive current and a lower junction leakage. Resistivity and synchrotron X-ray diffraction in real-time during annealing of films showed the formation of a NiSi film at about , which is morphologically stable up to about . These NiSi films could find applications in future nanoscale complementary metal oxide semiconductor devices or three-dimensional metal-oxide-semiconductor devices such as Fin-type field effect transistors for the 22 nm technology node and beyond.

ABSTRACT Invasive mammals are premier drivers of extinction and ecosystem change, particularly on islands. In the 1960s, conservation practitioners started developing techniques to eradicate invasive mammal populations from islands. Larger and more biologically complex islands are being targeted for restoration worldwide. We conducted a feral goat (Capra hircus) eradication campaign on Santiago Island in the Galápagos archipelago, which was an unprecedented advance in the ability to reverse biodiversity impacts by invasive species. We removed >79,000 goats from Santiago Island (58,465 ha) in <4.5 years, at an approximate cost of US$6.1 million. An eradication ethic combined with a suite of techniques and technologies made eradication possible. A field-based Geographic Information System facilitated an adaptive management strategy, including adjustment and integration of hunting methods. Specialized ground hunting techniques with dogs removed most of the goat population. Aerial hunting by helicopter and Judas goat techniques were also critical. Mata Hari goats, sterilized female Judas goats induced into a long-term estrus, removed males from the remnant feral population at an elevated rate, which likely decreased the length and cost of the eradication campaign. The last 1,000 goats cost US$2.0 million to remove; we spent an additional US$467,064 on monitoring to confirm eradication. Aerial hunting is cost-effective even in countries where labor is inexpensive. Local sociopolitical environments and best practices emerging from large-scale, fast-paced eradications should drive future strategies. For nonnative ungulate eradications, island size is arguably no longer the limiting factor. Future challenges will involve removing invasive mammals from large inhabited islands while increasing cost-effectiveness of removing low-density populations and confirming eradication. Those challenges will require leveraging technology and applying theory from other disciplines, along with conservation practitioners working alongside sociologists and educators.

Silicide engineering for high-performance CMOS logic devices is challenged by aggressive scaling of critical dimensions, new high-performance elements, and requirements of morphological stability. While NiSi can satisfy many of the integration challenges, incorporating Pt forms a more robust [NixPt(1-x)]Si and improves morphological stability. In light of the challenges created by performance enablers, we review our latest results indicating whether a replacement for NiPt(1-x)Si is needed and highlight our investigations into alternative silicide materials such as Er, Yb, and Ir. We also discuss the architecture and performance needs beyond 32nm technology.

In an integration scheme where the nFETs and pFETs of CMOS devices are silicided with different materials (Dual Silicides), rare-earth erbium (Er) and ytterbium (Yb) silicides are potential candidates for contacts to n-Si because of their lower Schottky barrier heights, as compared to more conventional nickel and cobalt silicides. [1-3] Although the lower Schottky barrier across the silicide/n-silicon interface results in reduced contact resistivity, the microstructure can exhibit defects and morphology issues [3-7] which affect device integrity and may contribute to contact resistance degradation. [8] In this study, we compared the material and integration properties of Er and Yb silicides with those of Ni (Pt-alloyed) silicide. Using four point probe, AFM, optical inspection and SEM, we compared the silicides using sheet resistance, surface morphology, defects density and ease of formation in narrow lines. We found that the silicide morphology is affected by several process parameters such as the type of metal deposition process (sputtering vs. evaporated) and the anneal formation temperature. The silicides were also tested for their ability to withstand aggressive processing after their formation. The processes tested included exposure to PECVD plasma, contact hole reactive ion etching and forming-gas annealing. The rare-earth and Ni(Pt) silicides showed similar stability upon processing. Lastly, we quantified the residual metal remaining on dielectric surfaces after silicide processing. Overall, Er silicide showed better performance than Yb silicide. By optimizing various elements of the silicidation process, higher quality silicide films can be achieved for evaluation as suitable nFet contacts.

We demonstrate that the addition of alloying elements can substantially affect the formation and morphological stability of nickel silicide. A comprehensive study of phase formation was performed on 24 Ni alloys with varying alloying element concentrations. Using time resolved x-ray diffraction (XRD), we follow the formation of the silicide phases in real time during rapid thermal annealing. Simultaneous in situ resistance and light scattering measurements complement the XRD technique and show the onset of thermal degradation for the monosilicide films.

This letter presents a proof-of-concept process for tunable, self-limiting growth of ultrathin epitaxial NiSi2 films on Si (100). The process starts with metal sputter-deposition, followed by wet etching and then silicidation. By ionizing a fraction of the sputtered Ni atoms and biasing the Si substrate, the amount of Ni atoms incorporated in the substrate after wet etching can be controlled. As a result, the thickness of the NiSi2 films is increased from 4.7 to 7.2 nm by changing the nominal substrate bias from 0 to 600 V. The NiSi2 films are characterized by a specific resistivity around 50 uO cm.

Silicide materials used as contacts in CMOS devices have evolved over many technology nodes. This article traces the often forgotten defectivity related reasons that were the primary drivers for a change in materials or process flow – evolving from Ti to Co and Ni silicides, and the more recent return to Ti-based liner silicides. The criteria used for the selection of these metal silicides have undergone a dramatic change with the advent of 3-D transistors and trench silicide contacts, and is now primarily guided by the value of interfacial contact resistivity (ρc). Furthermore, using results from synchrotron X-ray diffraction and pole-figure analysis, we present how phase formation and microstructure of contacts vary with Ti thickness, alternative annealing treatments, and substrate composition and orientation. We show that microstructure in very thin films can change from amorphous to epitaxial, a factor likely to become important for the contacts in upcoming generations of devices.

We study the impact of improved dopant activation in a BiCMOS SiGe technology, using laser annealing to improve activation, and a low temperature contact module to avoid de-activation. We present the results of simple DC test structure measurements and high frequency bipolar transistor measurements. Improved activation significantly reduces sheet resistance, particularly for polysilicon layers. Likewise, reductions in the bipolar device resistance parasitics cause an improvement in maximum power gain frequency (fMAX).

This letter presents a proof-of-concept process for tunable, self-limiting growth of ultrathin epitaxial NiSi2films on Si (100). The process starts with metal sputter-deposition, followed by wet etching and then silicidation. By ionizing a fraction of the sputtered Ni atoms and biasing the Si substrate, the amount of Ni atoms incorporated in the substrate after wet etching can be controlled. As a result, the thickness of the NiSi2films is increased from 4.7 to 7.2 nm by changing the nominal substrate bias from 0 to 600 V. The NiSi2films are characterized by a specific resistivity around 50 μΩ cm.

Smooth, continuous, and highly conformal nickel nitride (NiNx)films were deposited by direct liquid injection (DLI)–chemical vapor deposition (CVD) using a solution of bis(N,N′-di-tert-butylacetamidinato)nickel(II) in tetrahydronaphthalene as the nickel (Ni) source and ammonia (NH3)as the coreactant gas. The DLI-CVD NiNxfilms grown on HF-last (100) silicon and on highly doped polysilicon substrates served as the intermediate for subsequent conversion into nickel silicide (NiSi), which is a key material for source, drain, and gate contacts in microelectronic devices. Rapid thermal annealing in the forming gas of DLI-CVD NiNxfilms formed continuous NiSi films at temperatures above 400°C. The resistivity of the NiSi films was 15μΩcm, close to the value for bulk crystals. The NiSi films have remarkably smooth and sharp interfaces with underlying Si substrates, thereby producing contacts for transistors with a higher drive current and a lower junction leakage. Resistivity and synchrotron X-ray diffraction in real-time during annealing of NiNxfilms showed the formation of a NiSi film at about 440°C, which is morphologically stable up to about 650°C. These NiSi films could find applications in future nanoscale complementary metal oxide semiconductor devices or three-dimensional metal-oxide-semiconductor devices such as Fin-type field effect transistors for the 22 nm technology node and beyond.

Silicide engineering for high-performance CMOS logic devices is challenged by aggressive scaling of critical dimensions, new high-performance elements, and requirements of morphological stability. While NiSi can satisfy many of the integration challenges, incorporating Pt forms a more robust [NixPt(1-x)]Si and improves morphological stability. In light of the challenges created by performance enablers, we review our latest results indicating whether a replacement for NiPt(1-x)Si is needed and highlight our investigations into alternative silicide materials such as Er, Yb, and Ir. We also discuss the architecture and performance needs beyond 32nm technology.

In an integration scheme where the nFETs and pFETs of CMOS devices are silicided with different materials (Dual Silicides), rare-earth erbium (Er) and ytterbium (Yb) silicides are potential candidates for contacts to n-Si because of their lower Schottky barrier heights, as compared to more conventional nickel and cobalt silicides. [1-3] Although the lower Schottky barrier across the silicide/n-silicon interface results in reduced contact resistivity, the microstructure can exhibit defects and morphology issues [3-7] which affect device integrity and may contribute to contact resistance degradation. [8] In this study, we compared the material and integration properties of Er and Yb silicides with those of Ni (Pt-alloyed) silicide. Using four point probe, AFM, optical inspection and SEM, we compared the silicides using sheet resistance, surface morphology, defects density and ease of formation in narrow lines. We found that the silicide morphology is affected by several process parameters such as the type of metal deposition process (sputtering vs. evaporated) and the anneal formation temperature. The silicides were also tested for their ability to withstand aggressive processing after their formation. The processes tested included exposure to PECVD plasma, contact hole reactive ion etching and forming-gas annealing. The rare-earth and Ni(Pt) silicides showed similar stability upon processing. Lastly, we quantified the residual metal remaining on dielectric surfaces after silicide processing. Overall, Er silicide showed better performance than Yb silicide. By optimizing various elements of the silicidation process, higher quality silicide films can be achieved for evaluation as suitable nFet contacts.

We demonstrate that the addition of alloying elements can substantially affect the formation and morphological stability of nickel silicide. A comprehensive study of phase formation was performed on 24 Ni alloys with varying alloying element concentrations. Using time resolved x-ray diffraction (XRD), we follow the formation of the silicide phases in real time during rapid thermal annealing. Simultaneous in situ resistance and light scattering measurements complement the XRD technique and show the onset of thermal degradation for the monosilicide films.