Semiconductor device manufacturing: process – With measuring or testing – Electrical characteristic sensed
Reexamination Certificate
1998-07-01
2001-03-06
Fourson, George (Department: 2823)
Semiconductor device manufacturing: process
With measuring or testing
Electrical characteristic sensed
Reexamination Certificate
active
06197606
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to fabrication processes of integrated circuits, and, more particularly, to methods for controlling the quality of monocrystalline silicon substrates.
BACKGROUND OF THE INVENTION
The silicon generally used for fabricating integrated circuits is grown using a technique referred to as the Czochralski (CZ) technique, and, therefore, contains oxygen in a concentration comprised generally between 5 and 10×10
17
atoms·cm
−3
(ASTM 83 units) as disclosed, for example, in A. Borghesi, B. Pivac, A. Sassella and A. Stella, Appl. Phys. Rev. 77, 4169 (1995) and W. Zuhlener, J. of Crystal Growth 65, 189 (1983). A substantially oxygen-free silicon monocrystal may be grown by other techniques such as the so-called “flow-zone” technique, but an oxygen-free silicon is unsuitable for fabricating integrated circuits because the presence of oxygen improves its mechanical properties, as disclosed, for example, in K. Sumino, Proc. of Semiconductor Silicon, 1981, The Electrochem. Soc., 1981.
On the other hand, the presence of oxygen may generate defects in the crystalline structure of silicon unless great care is exercised in exposing the substrate to particularly critical conditions during the fabricating process of integrated circuits. Indeed, in CZ silicon, oxygen may be over-saturated at particularly elevated process temperatures and precipitate during thermal treatments, as disclosed, for example, in J. G. Wikes, J. of Crystal Growth 65, 189 (1983).
Oxygen precipitation induces the formation of extensive defects (dislocations, stacking faults) and may degrade the performance of integrated circuits if these defects occur in the active areas of the devices. In contrast, if the defects induced by oxygen precipitation grow sufficiently far from the active areas of the devices, for instance deep into the bulk of the monocrystalline wafer, they tend to have a positive effect as they act as gathering centers of metal impurities (atoms of transition metals of the periodic table), as disclosed, for example, in W. K. Tice and T. Y. Tan, Appl. Phys. Lett. 28, 564 (1976). For these reasons, treatment techniques have been developed for obtaining silicon wafers where defects induced by oxygen precipitation exist only in an innermost region (bulk) with respect to the wafer thickness, and have an oxygen defect-free superficial layer (denuded zone) as disclosed, for example, in K. Nagasawa, Y. Matsushita and S. Kishino, Appl. Phys. Lett. 37, 622 (1980). Wafers so processed are then used for fabricating integrated devices essentially in the defect-free surface region also referred to as the “denuded-zone” of one of the two sides of the wafer of monocrystalline silicon.
The thickness of the denuded zone depends on the properties of the starting material, on the denuding pre-treatment and/or, in many cases, on the thermal treatments that are contemplated in the specific fabrication process of the integrated devices. Accordingly, it becomes necessary to monitor the actual thickness of the denuded zone to ensure that the active regions are reliably defined in defect-free regions on the silicon wafer.
Moreover, it has been recognized that oxygen precipitation in the bulk region should be moderate to be beneficial, as disclosed, for example, in R. Falster, 6th International Symposium on the Structure and Properties of Disclocations in Semiconductors, Oxford, 1989 and J. Vanhellemont and C. Claeys, Proc. of ESSDERC 87, G. Soncini and P. U. Calzolari, eds., North Holland, Amsterdam, 1988, p. 451. Thus, the density and size of oxygen precipitates in the bulk of a wafer should be relatively small, and the current techniques used to measure the thickness of denuded zones often fail under conditions of insufficient oxygen precipitate density in the bulk region.
Denuded zone thicknesses are commonly measured by means of microscopy techniques. The monocrystal cleavage, the selective etching and inspections of the samples by a suitable microscopy technique (optical microscopy, scanning electron microscopy (SEM), atomic force microscopy or others) are limited by an insufficient selectivity of the etching of the sample surface.
Transmission Electron Microscopy (TEM) does not require a selective etching of the samples and possesses a very high sensitivity, though it is statistically limited. Indeed, TEM may provide for a relatively correct estimate of the depth of the denuded zone only when the defect density in the bulk region is sufficiently large. This is so because of the restricted region of a sample that may be explored with this technique.
As mentioned above, the density and size of defects in the bulk of the wafer of monocrystalline silicon tend, for other reasons, to be particularly small in present wafers. As a consequence, the monitoring of the depth of the denuded zone by the known techniques becomes problematic. Moreover, these methods are destructive and require proper sample preparation which is a burdensome operation, especially for TEM exploration. For these reasons, reliably producing depth maps of denuded zones is at present difficult and costly.
In theory, the monitoring of the denuded zone thickness could be performed, by electrical methods, under favorable conditions. Probably, the most reliable of these techniques is the so-called Electron Beam Induced Current (EBIC) method. This method is very sensitive, but time consuming. In addition, it requires laborious sample preparation. In any case, it remains a destructive technique. Therefore, the EBIC technique remains unsuitable as an effective monitoring technique and for producing maps of the depth of a denuded zone in terms of compatibility of the time required and of the costs to implement a reliable process quality control.
On the other hand, it is well known that lifetime measurements of current carriers in a semiconducting silicon monocrystal are very sensitive to the presence of defects in the crystal lattice of the semiconductor, and, hence, to the presence of oxygen precipitates in the silicon. This is so especially if the lifetime measurements are carried out under conditions of low injection of electric charges.
According to a technique and relative measuring equipment, known in the trade with the name of Elymat, as described in V. Lehmann, H. Föll, J. Electrochem Soc. 135, 2831 (1988), excess minority carriers are generated by a laser beam that illuminates the front of a wafer. These carriers are collected in the space charge region of a Schottky contact which may be established either on the wafer back-side (backside photocurrent or BPC) or on the same illuminated or frontside of the wafer (frontside photocurrent of FPC).
The sample is dipped in a diluted HF solution to establish a Schottky contact on one side of the wafer and passivate the surface layer states on the other side. The injection level &dgr;n/p
°
(where &dgr;n is the concentration of excess minority carriers and p
°
is the equilibrium concentration of minority carriers) can be varied by varying the power of the illuminating laser. Carrier lifetime (&tgr;) or diffusion length
(
L
diff
={square root over (D&tgr;)}
, where D is the minority carriers diffusivity)
data are extracted from photocurrent measurements. FPC measurements are used when the lifetime in the sample under consideration is so low that practically no current could be gathered at a contact established on the backside of the wafer. This is often the case in the presence of oxygen precipitates, and thus FPC measurements are more suited than BPC measurements in view of the objective of the present invention.
A particularly sensitive method for measuring the lifetime under low injection conditions is the so-called method of Surface Photovoltage Measurement (SPV). Surface photovoltage measurements are carried out by illuminating the sample with lights of different wavelengths. The minority carriers generated are collected in a depletion region on the same wafer surface (at a certain distance from the illuminated area) and produce a variation of the
Brambilla Marzio
Cazzaniga Francesco
Pavia Giuseppe
Polignano Maria Luisa
Zanderigo Federica
Abbott Barbara Elizabeth
Allen Dyer Doppelt Milbrath & Gilchrist, P.A.
Fourson George
Galanthay Theodore E.
STMicroelectronics S.r.l.
LandOfFree
Determination of the thickness of a denuded zone in a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Determination of the thickness of a denuded zone in a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Determination of the thickness of a denuded zone in a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2456001