Single-crystal – oriented-crystal – and epitaxy growth processes; – Processes of growth from liquid or supercritical state – Having pulling during growth
Reexamination Certificate
1999-06-10
2001-03-06
Hiteshew, Felisa (Department: 1765)
Single-crystal, oriented-crystal, and epitaxy growth processes;
Processes of growth from liquid or supercritical state
Having pulling during growth
C117S013000, C117S020000
Reexamination Certificate
active
06197109
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a silicon single crystal wafer with few crystal defects and a method for producing the same.
2. Description of Related Art
In recent years, with the use of finer semiconductor devices required for higher integration degree of semiconductor devices such as DRAM, increasingly higher quality of silicon single crystals produced by the Czochralski method (abbreviated as CZ method hereinafter) and used for substrates of the devices has been desired. In particular, such crystals are likely to contain defects introduced during the crystal growth, which are called grown-in defects such as FPD, LSTD and COP, and degrade the oxide dielectric breakdown voltage and other device characteristics, and it is considered important to reduce the density and the size of the defects.
For the reference of the explanation of those defects, there will be given first general knowledge of factors determining densities of defects introduced in silicon single crystals, a void type point defect called vacancy (occasionally abbreviated as V hereinafter), and an interstitial type silicon point defect called interstitial silicon (occasionally abbreviated as I hereinafter).
A V-region in a silicon single crystal means a region containing many vacancies, i.e., depressions, pits, voids and the like generated due to missing of silicon atoms, and an I-region means a region containing many dislocations and aggregations of excessive silicon atoms generated due to excessive amount of silicon atoms. Between the V-region and the I-region, there should be a neutral (occasionally abbreviated as N hereinafter) region with no (or little) shortage or no (or little) surplus of the atoms. It has become clear that the aforementioned grown-in defects (FPD, LSTD, COP etc.) should be generated strictly only with supersaturated V or I, and they would not be present as defects even though there is little ununiformity of atoms so long as V or I is not saturated.
The densities of these two kinds of defects are determined by the relationship between the crystal pulling rate (growing rate) V, and the temperature gradient G in the vicinity of the solid-liquid interface in the crystal in the CZ method. It has been confirmed that defects distributed in a ring shape called OSF (Oxidation Induced Stacking Fault) are present around the boundary between the V-region and the I-region.
Those defects generated during the crystal growth include those mentioned below. For example, when the growth rate is relatively high, i.e., around 0.6 mm/min or higher, grown-in defects considered to be originated from voids, i.e., aggregations of void-type defects, such as FPD, LSTD, and COP are distributed over the entire cross-section of the crystal along the radial direction at a high density, and a region containing such defects is called V-rich region. When the growth rate is 0.6 mm/min or lower, the aforementioned OSF ring is initially generated at the circumferential part of the crystal with the decrease of the growth rate, and L/D (large dislocations, which include LSEPD, LFPD and the like, and are also called interstitial dislocation loops or dislocation clusters), which are considered to be originated from dislocation loops, are present outside the ring at a low density, and a region containing such defects is called. I-rich region. When the growth rate is further lowered to around 0.4 mm/min or lower, the OSF ring shrinks at the center of wafer and disappears, and thus the entire plane becomes the I-rich region.
Recently, there has been discovered a region called N-region containing neither the void-originated FPD, LSTD, and COP, nor the dislocation loop-originated LSEPD and LFPD, which region is present between the V-rich region and the I-rich region, and outside the OSF ring. This region exists outside the OSF ring, and shows substantially no oxygen precipitation occurred when subjected to a heat treatment for oxygen precipitation and examined by X-ray analysis or the like as for the precipitation contrast. This region is present at rather I-rich side, and the defects are not so rich as to form dislocation clusters such as LSEPD and LFPD.
Because this N-region is obliquely formed with respect to the growing axis when the growth rate is lowered in a conventional growing method, it exists as only a part of the wafer plane.
With respect to the defects mentioned above, according to the Voronkov's theory (V. V. Voronkov, Journal of Crystal Growth, 59 (1982) 625-643), it was proposed that a parameter of V/G, which is a ratio of the pulling rate (V) and the crystal solid-liquid interface temperature gradient (G) along the growing axis, determined the type and the total density of the point defect. In view of this, because the pulling rate should be constant in a plane, for example, a crystal having a V-rich region at the center, I-rich region at the periphery, and N-region therebetween is inevitably obtained at a certain pulling rate due to the ununiformity of the gradient G in the plane.
Therefore, improvement of such ununiformity of the gradient G has recently been attempted, and it has become possible to produce a crystal-having the N-region spreading over an entire transverse cross-section of the crystal (this region can previously exist only obliquely), for example, when the crystal is pulled with a gradually decreasing pulling rate V. Further, such an N-region spreading over an entire transverse cross-section can be made larger to some extent along the longitudinal direction of the crystal by pulling the crystal at a pulling rate maintained at a value at which the N-region transversely spreads. Furthermore, it has also become possible to make the N-region spreading over the entire transverse cross-section somewhat larger along the growing direction by controlling the pulling rate together with correction of G with reference to its variation with the crystal growth, so that the V/G should strictly be maintained constant. This N-region spreading over the entire transverse cross-section does not contain grown-in defects at all, and exhibits good oxide dielectric breakdown voltage.
Other than the aforementioned techniques, the gradual cooling method can be mentioned as a currently used method for reducing the defects. In this method, a crystal having a region containing excessive voids, called V-rich region, over the entire cross-section is pulled at a relatively high pulling rate, and the time for passing the crystal through a temperature range of 1150°-1080° C. during the pulling of crystal is prolonged to decrease the defect density. This method can improve the oxide dielectric breakdown voltage.
There has also been a method for pulling a crystal wherein the crystal is pulled at a low pulling rate so that the crystal should have a region containing excessive interstitial silicon, a region called I-rich region. This method substantially eliminates COP and the like, and affords good oxide dielectric breakdown voltage.
Further, V-rich crystals has been conventionally doped with nitrogen to afford crystals with extremely few FPD and COP.
When it is desired to use a higher pulling rate in the production of a crystal having an extremely low defect region such as one having the N-region over the entire transverse cross-section, the solid-liquid interface temperature gradient G of the crystal along the growing axis direction can be made larger based on the Voronkov's theory. However, the gradient G also must be made uniform with respect to the transverse direction of the crystal, and therefore such increase of the pulling rate is restricted by the limitation concerning the internal structure of furnace (hot zone, HZ) provided in an apparatus for crystal growing. In addition, in order to obtain the N-region, the pulling rate must be controlled within a narrow range, and therefore it is difficult to make the N-region larger along the crystal growth direction. Thus, such attempts do not suit the large scale production.
Further, the aforementioned gradual cooling me
Iida Makoto
Kimura Masanori
Kusaki Wataru
Muraoka Shozo
Tamatsuka Masaro
Hiteshew Felisa
Hogan & Hartson LLP.
Shin-Etsu Handotai & Co., Ltd.
LandOfFree
Method for producing low defect silicon single crystal doped... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for producing low defect silicon single crystal doped..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for producing low defect silicon single crystal doped... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2545187