Observation of Crystallization, Precipitation, and Phase Transformation Phenomena in Si Rich Titanium Silicide Thin Film

  • PDF / 4,414,328 Bytes
  • 6 Pages / 414.72 x 648 pts Page_size
  • 10 Downloads / 211 Views

DOWNLOAD

REPORT


463 Mat. Res. Soc. Symp. Proc. Vol. 398 ©1996 Materials Research Society

were done in a JEOL 4000FX operating at 400 kV using a Gatan Model 628 single tilt heating holder. Other microstructural analyses were done in either a Philips 420 at 120 kV or a Philips CM20 at 200 kV. RESULTS Structure of the As-l)eopsited Film The film stoichiometry was determined from RBS data using the RUMP analysis program. A best fit of the modeled RBS spectra to the experimental data gave a film composition of Ti30 Si'7 . Figure 1 gives a plan view TEM micrograph of the as deposited film. The film was partially crystallized with a crystallite density of 5x10 8 cm2 . Electron diffraction analysis gave the C49 phase as the only one present. The was no evidence of Si precipitation in the electron diffraction patterns. The thickness of the film was determined from a cross sectional TEM to be 110( nm. A thin surface oxide(l.5nm) was present between the as deposited layer and the Si substrate(Fig. 2).

FiLgure 1. Plan view TEM micrograph of the as deposited Ti30Si7( film.

Figure 2. Cross sectional TEM image of the interfacial oxide.

Crystallization to a Continuous Film/Onset of Observable Si Precipitation The growth to a continuous film and its microstructure were studied first. A sample of the as-deposited film on Si was heated to 260°C in the TEM. A sequence of micrographs was taken as time was recorded with the electron beam continuously striking the sample. The temperature reported was the heating stage thermocouple temperature corrected for electron beam heating. A temperature correction of +30'C was determined by monitoring film growth rates with the electron beam on and off thin film samples. A plot of nuclei radii vs. time gave the linear plot shown in Fig. 3 from which a growth rate of 5.4x 10-1im/sec was extracted. This front velocity compared

464

0.30 Co

20.25 0 E w0.20 a:0.15 i

0

I

I

100 200 300 Time(sec)

Figure 3. Growth rate for Ti30Si'7 crystals at 260'C.

Figure 4. Microstructure for fully crystallized Ti'3Si70 film.

favorably with that of published by Raaijmaker for TiSi25 films[3]. The film growth was followed until a continuous thin film was formed. A grain size of 550 nm was estimated for this thin film by averaging the long diameters of 100 grains(Fig. 4). The continuous film was identified as the C49 phase of TiSi, using standard electron diffraction techniques. No evidence of Si precipitation could be seen in the transmission electron diffraction(TED) patterns taken. Observable precipitation did not occur until temperatures in the range of 650-750°C were reached. Fig. 5 gives a bright field TEM micrograph with a corresponding TED pattern after in-situ heating to a temperature of 700°C. A dense array of precipitates is seen. Grain size measurements indicated that a measureable grain growth had not occurred up until this point. Cross sectional TEM analysis(Fig. 6) showed that the precipitates, which were approximately oval in shape, were located within the volume of the TiSi 2 grains and did not nucleate at g