Chemical mechanical polishing composition and process

Semiconductor device manufacturing: process – Chemical etching – Combined with the removal of material by nonchemical means

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C438S692000, C106S003000, C051S307000, C051S308000, C051S309000, C216S089000, C252S079100

Reexamination Certificate

active

06313039

ABSTRACT:

INTRODUCTION
1. Technical Field
This invention relates to an improved composition and process for the chemical mechanical polishing or planarization of semiconductor wafers. More particularly, it relates to such a composition and process which are tailored to meet more stringent requirements of advanced integrated circuit fabrication.
2. Background
Chemical mechanical polishing (or planarization) (CMP) is a rapidly growing segment of the semiconductor industry. CMP provides global planarization on the wafer surface (millimeters in area instead of the usual nanometer dimensions). This planarity improves the coverage of the wafer with dielectric (insulators) and metal substrates and increases lithography, etching and deposition process latitudes. Numerous equipment companies and consumables producers (slurries, polishing pads, etc.) are entering the market.
CMP has been evolving for the last ten years and has been adapted for the planarization of inter-layer dielectrics (ILD) and for multilayered metal (MLM) structures. During the 80's, IBM developed the fundamentals for the CMP process. Previously (and still used in many fabs today) plasma etching or reactive ion etching (RIE), SOG (“spin on glass”), or reflow, e.g., with boron phosphorous spin on glass (BPSG), were the only methods for achieving some type of local planarization. Global planarization deals with the entire chip while “local” planarization normally only covers a ~50 micron
2
area.
At the 1991 VMIC Conference in Santa Clara, Calif., IBM presented the first data about CMP processes. In 1993 at the VMIC Conference, IBM showed that a copper damascene (laying metal lines in an insulator trench) process was feasible for the MLM requirements with CMP processing steps. In 1995 the first tungsten polishing slurry was commercialized.
The National Technology Roadmap for the Semiconductor Industries (1994) indicates that the current computer chips with 0.35 micron feature sizes will be reduced to 0.18 micron feature size in 2001. The DRAM chip will have a memory of 1 gigabit, and a typical CPU will have 13 million transistors/cm
2
(currently they only contain 4 million). The number of metal layers (the “wires”) will increase from the current 2-3 to 5-6 and the operating frequency, which is currently 200 MHz, will increase to 500 MHz. This to will increase the need for a three dimensional construction on the wafer chip to reduce delays of the electrical signals. Currently there are about 840 meters of “wires”/chip, but by 2001 (without any significant design changes) a typical chip would have 10,000 meters. This length of wire would severely compromise the chip's speed performance.
The global planarization required for today's wafer CDs (critical dimensions) improves the depth of focus, resulting in better thin metal film deposition and step coverage and subsequently increases wafer yields and lowers the cost/device. It is currently estimated (1996) that it costs $~114/layer/wafer with current limited planarization processes. As the geometries become smaller than 0.35 micron, the planarity requirements for better lithography become critical. CMP is becoming important, if not essential, for multiple metal levels and damascene processes.
The CMP process would appear to be the simple rotation of a wafer on a rotary platen in the presence of a polishing medium and a polishing pad that grinds (chips away) the surface material. The CMP process is actually considered to be a two part mechanism: step one consists of chemically modifying the surface of the material and then in the final step the altered material is removed by mechanical grinding. The challenge of the process is to control the chemical attack of the substrate and the rate of the grinding and yet maintain a high selectivity (preference) for removing the offending wafer features without significant damage to the desired features. The CMP process is very much like a controlled corrosion process.
An added complexity is that the wafer is actually a complex sandwich of materials with widely differing mechanical, electrical and chemical characteristics, all built on an extremely thin substrate that is flexible.
The CMP processes are very sensitive to structural pattern density which will affect metal structure “dishing” and oxide erosion. Large area features are planarized slower than small area features.
At the recent SEMICON/Southwest 95 Technical program on CMP, it was stated that “Metal CMP has an opportunity to become the principal process for conductor definition deep 5submicron integrated circuits.” Whether or not it does so depends on the relative success of CMP technologists in achieving the successful integrated process flow at competitive cost.
Slurries: CMP has been successfully applied to the planarization of interdielectric levels (IDL) of silicon oxides, BPSG, and silicon nitride and also metal films. The metal films currently being studied include tungsten (W), aluminum (Al) and copper (Cu).
The polishing slurries are a critical part of the CMP process. The polishing slurries consist of an abrasive suspension (silica, alumina, etc.) usually in a water solution. The type and size of the abrasive, the solution pH and presence of (or lack of) oxidizing chemistry are very important to the success of the CMP process.
Metal CMP slurries must have a high selectivity for removing the unwanted metal compared to the dielectric features on the wafers. The metal removal rate should be between 1700 to 3500 Å/min) without excessive “dishing” of the metal plugs or erosion of the oxide substrate.
The oxide CMP has similar requirements and polishing rates close to 1700 Å/minute.
Metal Polishing: This type of polishing relies on the oxidation of the metal surface and the subsequent abrasion of the oxide surface with an emulsion slurry. In this mechanism, the chemistry's pH is important. The general equations are (M=metal atom):
M
o
→M
n+
+ne
M
n+
+[Ox]
y
→MO
y
or [M(OH)
x
]
Under ideal conditions the rate of metal oxide (MO
y
) formation (V
f
) will equal the rate of oxide polishing (V
p
), (V
f
=V
p
). If the pH is too low (acidic) then the chemistry can rapidly penetrate the oxide and attack the metal (V
f
<V
P
), thus exposing the metal without any further oxide formation. This means that all metal surfaces, at high points and in valleys are removed at the same rate. Planarization of the surface is not achieved. This could cause metal plug connectors to be recessed below (“dishing”) the planarization surface which will lead eventually to poor step coverage and possible poor contact resistance.
When the pH is too high (caustic), then the oxide layer may become impenetrable to the chemistry and the metal becomes passive, (V
f
>V
p
) and the metal polishing rate becomes slow. Metal polishing selectivity to oxide generally ranges from 20 to 100:1, depending on the metal type. Tungsten metal should have selectivities >50:1 for the metal to oxide, and copper could have >140:1 metal to oxide selectivity. Etch rates can be up to 7000 Å/min. The chemical diffusion rate and the type of metal oxide surface are important to the successful planarization process. A detailed mechanism has been proposed by Kaufman.
In practice, the low pH and highly corrosive oxidants (ferric nitrate) being used with an example metal CMP process has created corrosion problems with the polishing equipment. Currently the oxidant used in the metal polishing step has ranged from nitric acid to hydrogen peroxide, cesium and ferric nitrate solutions and even ferric cyanide solutions. Because of chemical stability problems, many slurries are made up at the point of use which means that there is little or no shelf life.
Metal planarization needs an oxidizing reagent that is stable and is not going to contribute to mobile ion contamination, will not “stain” the equipment, will not affect the slurry composition.and slurry particle distribution and is generally environmentally friendly. The current hyd

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Chemical mechanical polishing composition and 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 Chemical mechanical polishing composition and process, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Chemical mechanical polishing composition and process will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-2615010

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.