Single-crystal – oriented-crystal – and epitaxy growth processes; – Apparatus
Reexamination Certificate
2002-08-28
2003-11-25
Kunemund, Robert (Department: 1765)
Single-crystal, oriented-crystal, and epitaxy growth processes;
Apparatus
C117S088000, C117S089000, C117S900000, C118S715000, C118S725000
Reexamination Certificate
active
06652650
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a modified susceptor for use in a chemical vapor deposition process. More particularly, the present invention relates to a modified susceptor having a plurality of holes for use in an epitaxial deposition reactor and process such that autodoping of the front surface of a semiconductor wafer and discontinuous silicon growth on the back surface of the semiconductor wafer is significantly reduced or eliminated.
In the production of single silicon crystals grown by the Czochralski method, polycrystalline silicon is first melted within a quartz crucible with or without dopant. After the polycrystalline silicon has melted and the temperature equilibrated, a seed crystal is dipped into the melt and subsequently extracted to form a single crystal silicon ingot while the quartz crucible is rotated. The single crystal silicon ingot is subsequently sliced into individual semiconductor wafers which are subjected to several processing steps including lapping/grinding, etching, and polishing to produce a finished semiconductor wafer having a front surface with specular gloss. To prepare the finished wafer for device manufacturing, the wafer may be subjected to a chemical vapor deposition process such as an epitaxial deposition process to grow a thin layer of silicon generally between about 0.1 and about 200 micrometers thick on the front surface of the wafer such that devices can be fabricated directly on the epitaxial layer. Conventional epitaxial deposition processes are disclosed in U.S. Pat. Nos. 5,904,769 and 5,769,942.
The epitaxial deposition process is typically comprised of two steps. In the first step after the semiconductor wafer is loaded into a deposition chamber and lowered onto a susceptor, the front surface of the wafer is subjected to a cleaning gas such as hydrogen or a hydrogen/hydrochloric acid mixture at about 1150° C. to “pre-bake” and clean the front surface of the semiconductor wafer and remove any native oxide on that surface to allow the epitaxial silicon layer to grow continuously and evenly onto the front surface. In the second step of the epitaxial deposition process the front surface of the wafer is subjected to a vaporous silicon source such as silane or trichlorosilane at about 800° C. or higher to deposit and grow an epitaxial layer of silicon on the front surface. During both steps of the epitaxial deposition process the semiconductor wafer is supported in the epitaxial deposition chamber by the susceptor which is generally rotated during the process to ensure even growth of the epitaxial layer. The susceptor is generally comprised of high purity graphite and has a silicon carbide layer completely covering the graphite to reduce the amount of contaminants such as iron released from the graphite into the surrounding ambient during high temperature processes. Conventional susceptors used in epitaxial growth processes are well known in the art and described in U.S. Pat. Nos. 4,322,592, 4,496,609, 5,200,157, and 5,242,501.
When a conventional susceptor is utilized in the epitaxial deposition process, during the loading process as the wafer is lowered onto the susceptor, gas can be trapped between the susceptor and the wafer causing the wafer to “float” and slide onto the susceptor in a tilted position. This can result in uneven epitaxial growth. Furthermore, during the pre-bake step a small amount of cleaning gas such as hydrogen can effuse around the wafer edge between the wafer and the susceptor and into the space between the wafer and the susceptor. If the back surface of the wafer is sealed with an oxide layer, such as a low temperature oxide layer, the effused hydrogen will not react sufficiently with the oxide layer to create pinholes in the layer or completely remove the oxide layer. If the back surface is an etched or polished surface as desired by many device manufacturers and only contains a thin native oxide layer, the hydrogen or hydrogen/hydrochloric acid mixture will completely remove the native oxide layer near the outer edge of the back surface where the cleaning gas effuses around the wafer and will create pinhole openings in the native oxide layer exposing the silicon surface as etching moves away from the outer edge of the wafer. These pinhole openings typically form a ring or “halo” around the wafer.
During the epitaxial deposition process a small amount of silicon containing source gas can also effuse around the wafer edge between the wafer and the susceptor and into space between the wafer and the susceptor. If the back surface of the wafer is oxide sealed, nucleation and growth of a silicon film is substantially suppressed. In areas where the native oxide layer has been completely etched away by the cleaning gas a smooth continuous layer of silicon is grown. Where the cleaning gas has not completely removed the native oxide layer and pinholes have created exposed silicon, however, the silicon containing source gas can deposit silicon in the pinholes and create a non-uniform silicon film on the wafer backside during the epitaxial deposition. Thus, for wafers with etched or polished back surfaces having only a native oxide layer, pinholes created in the native oxide layer during the pre-bake step may lead to discontinuous silicon growth on the back surface which appears hazy under bright light illumination. This haziness or “halo” on the back surface of the wafer is comprised of small silicon growths or bumps having a diameter of about 0.5 micrometers and being about 10 nanometers high. These bumps of silicon scatter light and lead to the haziness and are undesirable as they can interfere with machine vision and optical pyrometry systems that view the back surface of the wafer during device processing.
Another problem encountered during the high temperature growth of the epitaxial silicon layer is the out-diffusion of dopant atoms such as boron or phosphorus through the back surface of the semiconductor wafer during the high temperature pre-bake and the epitaxial growth steps. With conventional susceptors, the dopant atoms that out-diffuse from the back surface are trapped between the susceptor and the wafer itself and can effuse between the wafer edge and the susceptor toward the front surface of the wafer. These dopant atoms can be incorporated into and contaminate the growing deposition layer and degrade the resistivity uniformity near the wafer edge. If the back surface of the semiconductor wafer is oxide sealed using for example, a low temperature oxide, the dopant atoms will not substantially out-diffuse from the back surface. Semiconductor wafers having etched or polished back surfaces, however, are subject to out-diffusion of dopant atoms from the back surface during the epitaxial deposition process which can lead to unwanted autodoping of the front surface.
Several methods have been suggested for attempting to eliminate back surface halos and autodoping. To eliminate back surface halos Nakamura (Japanese Unexamined Patent Application No. JP11-16844) disclosed performing a hydrogen fluoride strip and/or a high-temperature hydrogen annealing step of the back surface up to 10 days before the wafers are loaded into the epitaxial reactor. The process adds additional processing steps which can greatly increase complexity and cost of the deposition process. Deaton et al. (U.S. Pat. No. 5,960,555) disclosed a method of preventing the frontside reactive source gas from effusing to the wafer backside by utilizing a susceptor with built-in channels along the wafer edge for directing purge gas flows to the edge of the wafer. This process requires substantial modification of existing epitaxial deposition chambers and utilizes increased purge gas flows which can cause the purge gas to spill over to the front surface and mix with the source gas which can degrade the resulting epitaxial film.
To reduce autodoping, Hoshi (Japanese Unexamined Patent Application No. JP11-87250) disclosed using vacuum sucking on the edge of a susceptor to evacuate boron dopant on the edge of the susceptor a
Standley Robert W.
Yang Charles Chiun-Chieh
Kunemund Robert
MEMC Electronic Materials , Inc.
Senniger Powers Leavitt & Roedel
LandOfFree
Modified susceptor for use in chemical vapor deposition process does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Modified susceptor for use in chemical vapor deposition process, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Modified susceptor for use in chemical vapor deposition process will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3176552