Semiconductor device manufacturing: process – Making device or circuit responsive to nonelectrical signal – Physical stress responsive
Reexamination Certificate
2001-11-07
2003-10-21
Niebling, John F. (Department: 2812)
Semiconductor device manufacturing: process
Making device or circuit responsive to nonelectrical signal
Physical stress responsive
C200S181000, C200S600000, C073S504140, C073S514160, C073S862632
Reexamination Certificate
active
06635506
ABSTRACT:
FIELD OF THE INVENTION
This invention generally relates to the fabrication of micro-electromechanical switches (MEMS), and more particularly, to the manufacture of MEMS which can be integrated into current state of the art semiconductor fabrication processes.
BACKGROUND OF THE INVENTION
Switching operations are a fundamental part of many electrical, mechanical and electromechanical applications. MEM switches have drawn considerable interest over the last few years. Products using MEMS technology are widespread in biomedical, aerospace, and communications systems.
Conventional MEMS typically utilize cantilever switches, membrane switches, and tunable capacitor structures as described, e.g., in U.S. Pat. No. 6,160,230 to McMillan et al., U.S. Pat. No. 6,143,997 to Feng et al., U.S. Pat. No. 5,970,315 to Carley et al., and U.S. Pat. No. 5,880,921 to Tham et al. MEMS devices are manufactured using micro-electromechanical techniques and are used to control electrical, mechanical or optical signal flows. Such devices, however, present many problems because their structure and innate material properties require them to be manufactured in lines that are separate from conventional semiconductor processing. This is usually due to the different materials and processes which are not compatible and, therefore, which cannot be integrated in standard semiconductor fabrication processes.
The use of materials typically used in the manufacture of MEMS, such as gold, pose obvious integration problems for integrating devices directly to on-chip applications. Even the use of polysilicon, which is widely found in the literature, poses problems due to the temperature cycles and the usual segregation of front-end of the line (FEOL) tools where the actual semiconductor devices are fabricated and the back-end of the line (BEOL) where interconnect metals are processed. Typically, the two sets are not allowed to have process crossovers from one to the other in order to prevent metallic contamination of the active devices. It is therefore unlikely to see polysilicon deposition in the back-end of the line.
Most existing processes suffer from a serious drawback in that by using standard metalization, no encapsulation is provided to protect the metal. Moreover, more than one substrate is used, oftentimes bonded together, with corresponding inherent disadvantages.
Other existing techniques only provide switching capabilities at the top of the structure, making it unlikely that integration can be achieved at all levels, as will be described hereinafter in the present invention.
Accordingly, there is a need for a process that is capable of providing MEMS devices using established BEOL materials coupled to processing that can be fully integrated so that these devices can be manufactured either in conjunction with or as an add-on module to the conventional BEOL or interconnect levels.
In order to gain a better understanding of the present invention, a conventional MEM switch will now be described with reference to
FIG. 1
, which shows a cross-section view of a MEM switch having both ends of a deformable beam
1
anchored in dielectric
4
. The lowest level consists of a dielectric material
5
containing conductive elements
2
,
2
a
, and
3
which will be used subsequently to connect or form the various electrical portions of the device. The conductors referenced by numerals
2
and
2
a
are used to provide an operating potential that cause the beam to deform. Conductor
3
, which conducts a signal, is in turn connected to the beam when it is in operation.
FIG. 2
shows a planar view of the same prior art MEM device of FIG.
1
. In a typical implementation, deformable beam
1
is formed by polysilicon over dielectric
4
, e.g., SiO
2
, and the surrounding material is etched away leaving a raised structure, i.e., the beam suspended above the conductors that were previously formed or which, themselves, are made of polysilicon. Then the device is subjected to electroless plating, usually of gold, that adheres to the polysilicon forming the conductive elements
1
,
2
,
2
a
and
3
. The switch is operated by providing a potential difference between the beam and electrodes
2
and
2
a
. This voltage generates an electrostatic attraction which pulls beam
1
in contact with electrode
3
, thus closing the switch.
One should note that these are all typically raised structures having a large topography when compared to conventional semiconductor devices. This in itself makes them virtually impossible to integrate into the semiconductor chip fabrication process. These devices are a typically made using surface micro-machining techniques which include building on photoresist or building on a substrate, such as silicon, and then removing a portion of the substrate under the device from the backside of the substrate, again precluding integration with standard semiconductor processing.
FIG. 3
illustrates a cross-section view of another version of a conventional MEM switch, wherein only one end of the beam
1
is anchored within the dielectric
4
. All the other parts perform as described in FIG.
1
. The same applies with regard to
FIG. 4
, illustrating a top-down view of the corresponding device illustrated in FIG.
3
. In the latter case, the switch is operated by applying a voltage between beam
1
and control electrode
2
. This causes the beam to be pulled down into contact with the signal electrode
3
. When the voltage is dropped, beam
1
returns to its original position.
Typically, the gap between the beam and the control electrode substantially determines the voltage required to pull down the beam. Most literature describes devices having gaps ranging from 1 to several micrometers. These gaps are large and the voltage required is therefore higher than would be desired for most consumer applications. Reported activation voltages range from around 30 to 75 volts. This is far too high for applications like cell phones which typically operate between 3 to 5 volts. The structure of the present invention operates with gaps ranging from 200 angstroms to several thousand angstroms, producing switches having an activation voltage below 5 volts.
The aforementioned illustrative switch configurations are only some of many possible structures which are known in the art. It is worth noting that MEM switches may also be configured in an arrangement of multiple beams wired in a variety of combinations.
Stiction is of primary concern in MEMS devices. Stiction is defined as two or more surface making contact that will not release without causing some damage to the device. Impingement is a major cause of this phenomena. The present invention addresses this problem in at least one embodiment by providing an air gap
200
when the switch is closed, as will be shown in detail with reference to FIG.
19
A. Surface tension is also believed to be another major cause of stiction. That explains why the present invention utilizes dry etches and processes for the release of the moving parts and subsequent processing.
OBJECT OF THE INVENTION
Accordingly, it is an object of the invention to build MEM switches and other similar structures which are fully integrated within CMOS , bipolar or BiCMOS wafers.
It is another object to manufacture MEM switches and other similar structures with a modified damascene process.
It is a further object to build MEM switches and other similar structures utilizing copper encapsulated in a barrier material to protect the metal.
It is yet another object to ensure that the encapsulation can be integrated into BEOL copper at a temperature compatible with such a process.
SUMMARY OF THE INVENTION
These and other objects are addressed by the present invention by providing a method of fabricating MEMS switches integrated with conventional semiconductor interconnect levels, using compatible processes and materials.
The invention described herein provides a method of fabricating a capacitive switch adaptable to produce various configurations used for contact switching and/or metal-dielectric-metal switches.
In a p
Bisson John C.
Cote Donna R.
Dalton Timothy J.
Groves Robert A.
Petrarca Kevin S.
Niebling John F.
Schnurmann H. Daniel
Simkovic Viktor
LandOfFree
Method of fabricating micro-electromechanical switches on... 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 of fabricating micro-electromechanical switches on..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of fabricating micro-electromechanical switches on... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3131881