O. Thomas

Reaction of titanium with germanium and silicon-germanium alloys

The reaction of Ti with pure Ge and several Ge‐Si alloys has been investigated with the double aim of understanding the reaction with Ge and of throwing some light on the still vexing problem of the Ti‐Si reaction. With pure Ge one observes first of all the formation of Ti6Ge5 until complete consumption of the Ti is present. This is followed by the clearly identifiable nucleation of TiGe2, initially forming islands that grow laterally. With a 50‐50 (at. %) alloy of Si and Ge, one still observes distinct growth steps, but there is overlap between the growth of the initial phase, and the nucleation and growth of Ti(Ge,Si)2.

Nucleation and growth in the reaction of titanium with germanium and some silicon-germanium alloys

Abstract Titanium reacts with pure Ge in two different ways. At low temperatures one observes the formation of Ti6Ge5 with some characteristics typical of diffusion-controlled reaction. Upon completion of this first stage Ti6Ge5 reacts with remaining Ge to form TiGe2, isomorphous with C54 TiSi2, in a process which is clearly controlled by nucleation. The same observations apply to reactions with an alloy containing 25 at% Si. With an alloy containing 50 at% Si the two stages become merged, so that while remaining identifiable, they are much less distinct than with the previous conditions. In the case of the 50-50 alloy, careful comparison of data obtained by backscattering, X-ray diffraction and resistance measurements reveals that in the formation of the mixed disilicide solid solution the formation of the C54 structure is preceded by that of the C49 structure as with pure Si. The reaction behavior observed with an alloy containing 80 at% Si resembles that generally obtained with pure Si: there are no easily identifiable steps between the initial SiTi sample and the final one, SiTiSi2. The gradual merging of the diffusion-controlled reaction and that controlled by nucleation as the concentration of Si in the substrate increases implies that nucleation plays a significant role in the formation of TiSi2, even if that role cannot be easily isolated. Values of the resistivity for both Ti6Ge5 and TiGe2 were obtained.

Comparison of the diffusion barrier properties of tungsten films prepared by hydrogen and silicon reduction of tungsten hexafluoride

Abstract We have made a comparative study of the thermal stability with respect to aluminum and silicon of tungsten films prepared by two different chemical vapor deposition processes. The first set of films were prepared via the direct reduction of WF 6 by the silicon substrate whereas the second set were deposited by hydrogen reduction of WF 6 . The latter layers have a metallurgical behavior comparable with that of tungsten films deposited by other techniques while the former exhibit a very high thermal stability with no reactions between tungsten and aluminum or silicon detected up to temperatures as high as 625°C. This results from the high oxygen content of these films which allows the formation of interfacial oxides and thereby inhibits further interreactions. In the absence of oxygen, the barrier begins to fail at 550°C, leading eventually to the formation of a hexagonal silicide structure of the form W(Si,Al) 2 . The role of aluminum in enhancing the formation of the silicide and the nature of the hexagonal aluminum-containing silicide are discussed in some detail.

The reaction of scandium thin films with silicon: diffusion, nucleation, resistivities

Abstract The reaction of scandium films with their (100) silicon substrates leads to the formation of two different silicides. In a temperature range from 500 to 600°C one observes the formation of the orthorhombic ScSi. The modalities of this reaction imply that it is diffusion-controlled, although that has not been explicitly verified. Above 900°C one observes within a narrow range of temperatures the formation of hexagonal ScSi 1.7 , which is clearly controlled by nucleation. The room-temperature resistivity of ScSi was measured at 21 μω· cm. Comparison with the resistivity at liquid helium temperature, 1.5 μω· cm provides a measure of the minimal room-temperature resistivity. The mechanical weakness of the ScSi 1.7 made it difficult to obtain reliable values for the resistivity of that compound; room temperature values of about 38 μω· cm were obtained. (The resistivity values reported here correspond to polycrystalline films with various degrees of random orientation. Of necessity they were derived without taking into account the effects of crystalline anisotropy.) A systematic analysis of the results obtained with scandium throws a welcomed light on the still poorly understood modes of formation of TiSi 2 and of the silicon-rich rare-earth compounds.

Respective mobilities of metal and silicon in disilicides: Bilayers of chromium with molybdenum or tungsten

The behaviors of metallic bilayers, chromium‐molybdenum and chromium‐tungsten, on (100) silicon during isochronal annealing have been studied by Rutherford backscattering of 2.3‐MeV 4He+ ions and x‐ray diffraction. These experiments were conducted with the aim of obtaining information about the respective mobilities of silicon and metal atoms in the different silicides through a comparison of the temperatures at which the silicides form and those at which they mix (through metal‐atom diffusion). The results confirm that the respective silicides form via silicon‐atom motion and that the mobilities of the metal atoms are markedly smaller than that of the silicon atoms.

Diffusion of dopants in tungsten disilicide: effects of diffusion paths

Abstract Dopants, mostly boron, but also arsenic and antimony, implanted in unannealed WSi2 films prepared by chemical vapor deposition diffuse quite readily at temperatures as low as 600°C. Complete homogenization in layers 200 nm thick is obtained at that temperature in periods of time less than 24 h. These results can be attributed to the large density of grain boundaries in layers that are initially amorphous and crystalline during the course of the diffusion annealings. They are in sharp contrast to what is observed when the layers are annealed at high temperature to cause grain growth and stabilization of the silicide structure prior to implantation. The dopants then are largely immobile not only at 600°C, but even during prolonged oxidation procedures at 800°C. Similar low temperature diffusion of boron in WSi2 had been observed earlier in bilayers of polysilicon and WSi2, but it is only presently that one is aware of the huge difference in diffusion behaviors between annealed and unannealed silicide layers.

Diffusion of elements implanted in amorphous titanium disilicide

Abstract The usual Si dopants, B, P, As, and Sb, plus Ge were implanted into thick (400 nm) TiSi 2 layers deposited in an amorphous state by sputtering from a compound target. Samples with the various implants were annealed at temperatures from 300 to 700°C and analyzed both by transmission electron microscopy and secondary ion spectroscopy. The annealed samples display a very large grain size, which complicates the interpretation of the concentration profiles obtained by SIMS. In the case of a high dose of B (1 × 10 16 atoms/cm 2 at least), there is an indication of grain boundary transport occurring mostly in the initial stage of the heat treatments, followed by phenomena dominated by a reduced solubility in the terminal large-grained matrix. With all elements, except Sb, evidence of diffusion is obtained at 400°C. Germanium diffusion is even observed at 300°C.

Ion-implantation-induced fluorine agglomeration in tungsten disilicide prepared by low-pressure chemical vapour deposition

Abstract Boron and antimony were implanted in tungsten disilicide/silicon structures. The silicide films with a Si/ ratio of 2.3 were prepared by low pressure chemical vapour deposition (LPCVD) from a mixture of WF6 and SiH4. The films were subsequently annealed at 1000° C prior to implantation. The effect of implantation and subsequent heat treatments on the distribution of fluorine have been investigated by means of transmission electron microscopy (TEM), secondary ion mass spectroscopy (SIMS) as well as the nuclear resonance broadening technique using the reaction 19F(p, αγ)16O. After implantation with 5 × 1014 150 keV 121Sb+ ions cm[su−2] the fluorine depth distribution was bimodal with one peak located about the silicide/silicon interface and a second peak at a depth corresponding approximately to the limit of the implant range distribution. This is also the position in the film where TEM revealed the presence of a large number of voids. The same observations were made in samples implanted with 11B. Some of the fluorine is lost from the implantation-induced peak after heat treatment at 300° C for 30 min, yet most of the fluorine remains even after annealing at 900° C. The overall picture is not substantially modified by raising the substrate temperature during implantation to 400° C.