View the related article
J. Phys. Soc. Jpn. 75 (2006) 044602
Intrinsic Point Defects in Crystalline Silicon [April 10, 2006]
|
| Fig. 1: Diamond structure of Si crystal. ● is Si atom and ○ is vacancy (Courtesy: T. Goto). |
Intrinsic gettering[2] was effectively demonstrated in various mass-produced ULSI devices in the late 1980s by understanding and controlling oxygen precipitation. In 1990, a new type of grown-in defect, which was termed as a crystal originated particle (COP)[3] was found to degrade device performance and yield due to the nearly same size as the design rule; and this defect has attracted considerable attention.
It is well known that vacancies that are incorporated from the solid-melt interface during crystal growth agglomerate during the cooling process and form octahedral-shape voids, which are delineated as COPs on the wafer surface. In order to control and suppress void formation, it is important to understand the incorporation and diffusion of intrinsic point defects. A theory proposed by Voronkov[4] based on the recombination and diffusion of point defects predicted the behavior and control of defect formation in a crystal. At present, "pure silicon wafers" which are free of COPs and large dislocations[5] are being mass-produced with diameters as large as 300mm by improving the hot zone structure and precisely controlling the growth parameters.
In the case of silicon crystals with such large diameters, it becomes difficult to experimentally investigate the growth parameters that must be controlled in order to suppress the grown-in defects because of the large amounts of time and expenses involved in the investigation. Therefore, the prediction of point defects and COPs using computer simulation is very important; however, the scattered values of point defects such as thermal equilibrium concentration and diffusion constant have been reported and used in the calculation. The precise determination of these point defect parameter values would be very useful for the control of COPs, the progress of the diffusion model, and further advancements in gettering technology; moreover, it would contribute to the development of silicon crystals and processes.
In order to measure the thermal equilibrium concentration, positron lifetime measurement was performed at high temperatures[6]. The vacancy concentration was estimated to be 1012 ∼ 1014atoms/cm3 at 1300K; however, it is pointed out that the analysis of these data is very difficult. On the other hand, a difference between changes in the macroscopic length, and also in the lattice parameter induced by vacancies and self-interstitials are theoretically discussed. The concentration of the intrinsic point defects in a FZ Si single crystal at high temperatures was directly determined from the difference between the macroscopic length changes[7] and the lattice parameter thermal expansion[8]. It was found that vacancies are dominant and the total concentration of intrinsic point defects in thermal equilibrium at 1300K is (1.8 ± 1.2)x1016atoms/cm3[8]. Nevertheless, it is difficult to measure these values at high temperatures, and reliable values have not been obtained thus far.
Recently, Goto et al.[9] succeeded in the direct observation of the isolated vacancy in crystalline silicon using the low-temperature ultrasonic measurement with a high ultrasonic velocity resolution ∼ 10-6. They observed a low-temperature softening of the elastic constants in non-doped and B-doped FZ silicon crystals, which indicates that the vacancy with the triply degenerate states gives rise to the elastic softening through the Jahn-Teller effect. The widely spread orbital around the vacancy and the strong electron-strain interaction enabled the direct observation of vacancies with extremely low concentrations.
|
| Fig. 2: The low-temperature ultrasonic measurement system using He-refrigerator (Courtesy: T. Goto). |
The vacancy concentration was estimated to be ∼ 1015atoms/cm3 at 1620K. In the future, it is expected that the detailed observation of the quenched samples and COP-controlled CZ wafers will be carried out. It is beyond doubt that this fundamental and innovative result will contribute to the further development of silicon crystals.
References
[1]W. G. Allen and K. V. Anand:Solid-State Electronics 14 (1971) 397.
[2]T. Y. Tan et al.:Appl. Phys. Lett. 30 (1977) 175.
[3]J. Ryuta et al.:Jpn. J. Appl. Phys. 29 (1990) L1947.
[4]V. V. Voronkov: J. Cryst. Growth 59 (1982) 625.
[5]J. G. Park et al.: Silicon Wafer Sympo., Portland, 1998, p.E-1.
[6]S. Dannefear et al.: Phys. Rev. Lett. 56 (1986) 2195.
[7]M. Okaji: Int. J. Thermophys. 9 (1988) 1101.
[8]Y. Okada: Phys. Rev B. 41 (1990) 10741.
[9]T. Goto et al.: J. Phys. Soc. Jpn. 75 (2006) 044602.
The above article should be referred as "H. Tsuya: JPSJ Online-News and Comments [Apr. 10, 2006]" when citing.
Copyright © the Physical Society of Japan.