Reduction in temperature of nearly 100 - 150°C to form low resistivity C54-TiSi2 phase from the high resistivity C49- TiSi2 phase is achieved by the ion implantation of little amount of Mo or W (slightly lower improvement) into the silicon substrate prior to the deposition of Titanium. Even the geometric size of C54 formation of the silicide structure is greatly reduced. This is mainly due to the formation of nucleation of C54-TiSi2 phase from the high resistivity C49- TiSi2 phase but not with the ion mixing of the Ti/Si interface or implant-induced damage.
Titanium disilicide, a polymorphic material, has become predominant in semiconductor devices. It is available in two forms - C49 Orthorhombic base-centered and C54 orthorhombic face-centered structures. Former has high resistivity of about 60-70µâ„¦cm while the latter has low resistivity of around 15-20µâ„¦cm. When the temperature is gradually increasing, Ti started to react with Si. At the temperature of about 550 to 700 °C, high resistance C49 phase is formed first (due to the lower barrier to nucleation) followed by the formation of low resistivity C54 phase at temperatures greater than 750 °C.
Get your grade
or your money back
using our Essay Writing Service!
Formation of C54 from C49 is limited by two factors- low driving force and high activation energy (>5eV). Because of these factors, on heating, formation of C54 nuclei density is very little, which results in the fineline effect. Lowering the linewidth of the structure to submicron regime, due to the lack of C54 nuclei, TiSi2 film would be either in C49 phase or contains mixture of both phases or have high resistances even in C54 phase, which in turn decreases the performance of CMOS devices. This method not only reduces the temperature of C54 phase formation, but also removes the fineline effect by triggering the C54 nuclei density of about 0.4 µm of TiSi2 features to entirely convert from C49 to C54 during thermal annealing.
Either by using lightly doped p-type silicon or both (doped/undoped) polysilicon as a substrate at 45keV, implantation of Tungsten and Molybdenum(concentration of 1-10^13 -1-10^14 ions/cm2) takes place. The former uses filament material of 184 W+ ion beam to be placed into Si substrate while the latter uses arc chamber(98 Mo+ ion beam). Thickness of Titanium around 25, 35, or 55 nm, was sputtered on Silicon, followed by annealation of Nitrogen to simulate the activation of junction, which was then followed by the buffered HF.
FIG 1 depicts for the given energy of 45keV, the highest concentration of about 3-10^18 at/cm3 is found below the silicon surface of about 35 nm, for a dosage of 5-10^13 at/cm3 Molybdenum has been implanted through SIMS profile. Even Tungsten follows the similar shape though no proper standards were available to approximate it. Due to the activation anneal of nitrogen prior to Titanium deposition, which effectively removes the damage triggered by ion implantation. Few damages like crystal defects, precipitation of metal or silicide, and amorphous regions are reduced significantly, which was shown clearly in the TEM cross sectional analysis.
From the FIG 2, we can analyze that, in Mo controlled sample resistivity had gone down to about 16µâ„¦cm which in turn forms C54 TiSi2 at a temperature of about 600 °C and 30 min anneal while Samples without Mo still in C49 phase of high resistivity greater than 60µâ„¦cm. Rather Tungsten implanted samples had the combination of both C49 and C54 TiSi2, due to the resistivity of about 40µâ„¦cm.
Consider two samples with and without Mo to analyze the phase transformation by in situ resistance method. The below graph(FIG 3.a) distinctly illustrates the formation of C49 standard sample(without Mo implant) at around 580°C and sudden transformation to C54 phase at about >780 °C. In Comparison with Mo sample, temperature range of around 700 - 800°C there exists no clear resistance plateau, which means it was already converted to the C54 phase at ~700°C. Thin film X-ray analysis method (Fig 3.b) also gives the same result, which was annealed at 15 °C/min to 700 °C. These methods were valid for both unpatterned silicide films and submicron silicide structures.
Always on Time
Marked to Standard
Taken samples of both with and without implantation of Molybdenum for TiSi2 /N+ doped polysilicon lines of sheet resistance of 0.4 µm shown in FIG 4. Initially both were given a 650 °C, 30 min anneal respectively to form either C49 or C54-TiSi2. The lines which had a Molybdenum implant have fully converted to C54 phase while the lines without Mo implant received an additional 825 °C to transform to C54 phase. Even though it received an additional anneal, still struggling to transform to C54 phase, which was clearly visible from the spread of resistance values. The same results had been obtained with submicron P+ doped Polysilicon lines.
From all our techniques we had been observed till now that the implantation of Mo brings down the temperature of C54 formation to 100 - 150°C which effectively boosts the C54 nucleation. C54 nucleation density of grain size had been raised by a factor greater than 100 with Mo implanted samples which was confirmed by TEM measurements. High percentage of enhancement in the nucleation density (C54) effectively raises the number of active triple junctions in C49 due to the addition of Mo. In contrast, the normal phase transformations showed increment of only 10-15%.
In conclusion, nearly 100 - 150°C of temperature had been reduced greatly with the help of Mo or W ion implantation, into Silicon substrate prior to the Titanium deposition to form the low resistivity C54 phase (TiSi2). Even the size of the silicide structure had been decreased dramatically. On heating, the barrier to form C54 nucleation had been reduced due to segregation of Mo-C49 triple points.