First Phase Nucleation and Growth of Titanium Disilicide with an Emphasis on the Influence of Oxygen
- PDF / 344,691 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 9 Downloads / 228 Views
sample
contamination
oxygen content*
integrated 2
Ttr(°C)
at%**
(I/cm )
1
reference
2.1.1016
1.8
675+25
2
4 nm native oxide
3.1.1016
2.6
675*25
3
9 nm native oxide
7.0.1016
5.8
675+25
4
sputtered at
7.4.1016
6.2
675*25
27.1016
23
675*25
increased back-
ground pressure (10-5 Torr) low temperature
5
treatment (350C 24h)
in
HV fur-
nace at PH2 0=
10-6 Torr *
measured with 2 MeV He+ backscatterinq.
**
assuming a flat 0 profile at the start of silicidation
1103.
39
In a separate experiment the thickness of the Ti layer was varied. These samples were deposited and annealed without additional contamination. Elemental depth profiles and phase identification were obtained with the common techniques (Rutherford backscattering Auger electron spectroscopy (AES), spectroscopy (RBS), Transmission electron microscopy (TEM) and X-ray diffraction (XRD)). In addition resistivity measurements were performed using a standard four point probe. RESULTS AND DISCUSSION (i) Phase identification Support for the occurrence of a phase with the stoichiometry of TiSi2 in diffusion couples after anneal at lower temperaobtained from RBS showing that ture (500°C to 6500C) is first Ti/Si=0.50 + 0.02. XRD patterns of reacted couples generally show only two reflections which can be assigned to C49 TiSi2. Phase identification from only two reflections however is ambiguous. Fortunately cosputtered samples annealed at low temperatures show some additional reflections because of the lack of texture in such layers. XRD analyses of numerous cosputtered samples and diffusion couples never showed any indication that the phase which concerns us here is different in the cosputtered samples. For this reason the final phase identification is performed on a 200 nm thick TiSi 2 .2 cosputterad sample annealed at 650°C for 24 h. The d-values obtained from the XRD patterns are listed in Table 2 together with the calculated values, which are obtained with the lattice constants a=0.3545, b= 1.3502, c=0.3550 nm for C49 TiSi 2 . No fit can be obtained for all these reflections with other Ti-Si compounds as listed in the JCPDS diffraction standards. In addition, we measured with TEM diffraction the dvalues of the reflections originating from a 30 nm thick Ti(100) Si substrate diffusion couple annealed at 650'C for 60 min. All 10 visible diffraction rings could be identified with the C49 TiSi2 phase. No other phases were detected. Since the measurement inaccuracy from TEM diffraction patterns is much larger than that of XRD we did not include these dates in table 2.
Table 2.
Measured
(XRD) and calculated
d-values for C49
TiSi2.
observed d-values
calculated (nm)
d-values
(nm)
Miller indices
0.337
0.3375
0.247
0.2467
040 111
0.225
0.2250
060
0.219
0.2191
131
0.1773
0.1773
200
0.1687
0.1688
080
40
Threshold temperature for the C49-*C54 transition Titanium layers of varying thickness were deposited on 0 (100) Si substrates and annealed for 120 min. at 600 C.The resistivity of the resulting layers is 80 4
