Variable capacitor and method for manufacturing same

Details Variable capacitor and method for manufacturing same Variable capacitor and method for manufacturing same

2003-07-25

2004-11-02

Dinkins, Anthony (Department: 2831)

Electricity: electrical systems and devices

Electrostatic capacitors

Variable

C361S278000, C029S025420

Reexamination Certificate

active

06813135

ABSTRACT:

FIELD OF THE INVENTION
This invention relates to a mechanical variable capacitor and, more particularly, to a variable capacitor small in size and high in performance and quality, for use on a densely integrated electric circuit, and to a method for manufacturing the same.
BACKGROUND OF THE INVENTION
A variable capacitor is utilized as a microfabricated mechanical switch using a variable capacitance structure.
FIGS. 1A-1C
show a microwave switch introduced in IEEE MTT-S Digest 1999, pp. 1923-1926.
A gold contact
26
is provided, through an insulation layer
25
, at a tip lower part of a silicon cantilever
21
. On a surface opposed to the contact
26
, there are provided a circuit terminal
27
for forming a close circuit upon contacted with the contact
26
, and a driving electrode
28
for providing an electrostatic force to the contact
26
to thereby deflect the silicon cantilever
21
. The silicon cantilever
21
has a length of approximately 200 &mgr;m, a width of approximately 20 &mgr;m and a thickness of approximately 2.5 &mgr;m. The contact
26
and the circuit terminal
27
has a gap set at 10 &mgr;m or less. By applying a voltage of 50 V or higher to the driving electrode
28
, the beam
21
deflects to place the contact
26
into contact with the circuit terminal
27
, thereby closing the contact.
However, because the voltage required for closing the contact is as high as 50 V or greater, there is a need to mount an exclusive booster circuit, posing hindrance against miniaturizing the switch element. Also, as the area is broader in the pad formed at the tip of the silicon cantilever
21
, the more the viscosity resistance of ambient air is undergone during vertical driving, to decrease the operation speed. High speed switching on a several is order is difficult to attain.
FIG. 2
is a conventional beam structure that enables a low voltage driving and switching speed on an order of several &mgr;s. A beam
31
has a size of a width W=2 &mgr;m, a thickness t=2 &mgr;m and a length L=500 &mgr;m. On the substrate
34
, an electrode
32
formed, on its surface, an insulation layer
35
having a film thickness of 0.01 &mgr;m is arranged through an air gap of 0.6 &mgr;m to the beam
31
. In case a voltage V is applied to between the beam
31
and the electrode
32
, the beam
31
deflects in −z direction due to an electrostatic force. At a pull-in voltage or higher, the electrostatic force becomes greater than a restoration force of the beam
31
thereby increasing the force. Accordingly, the beam
31
is immediately attracted onto the insulation layer
35
. If the voltage is raised furthermore, the beam
31
gradually increases the capacitance with the electrode
32
while increasing the contact area with the insulation layer
35
.
In this manner, the beam
31
can be weakened in springiness by increasing the length of the beam
31
. Also, by narrowing the width of the beam
31
and thereby reducing the viscosity resistance of air, it is possible to attain low-voltage driving and switching speed on an order of several &mgr;s. When the beam
31
uses, as material, aluminum having a Young's modulus 77 GPa, the pull-in voltage is 0.25 V in the case the beam
31
is provided as a cantilever and 1.72 V when it is supported at both ends.
However, such an elongate beam geometry involves conspicuous problems of (1) residual stress, (2) thermal expansion, and (3) stiction.
The first problem of residual stress is mentioned. In fabricating a fine beam, used is a thin-film structure using a semiconductor process, a thin-rolled material junction structure, or the like. In any case, the residual stress within the beam is problematic. Such residual stress includes two kinds, i.e. one is compressive/tensile stress to act in a beam lengthwise direction, and the other is stress gradient along a beam thickness direction.
For example, in case the beam of
FIG. 2
is assumably a beam supported at both ends, when an excessive compressive stress remains in x and y directions in the figure, the stress release in the y direction does not cause a substantial change in the beam geometry. However, concerning the x direction in which the beam end surface is bound, buckling is caused in order to release stress. Thus, the beam deflects irrespectively of applying an electrostatic force.
Conversely, where tensile stress remains, the beam
31
apparently has no change. However, as the residual tensile stress increases, pull-in voltage increases to conspicuously change the beam driving characteristic. Namely, it is ideal to manufacture a beam with a residual stress of zero. However, unless internal stress is accurately controlled to a predetermined value in the beam manufacture process, variation is incurred in buckling or pull-in voltage, deteriorating element quality.
On the other hand, because this kind of stress on a cantilever is to be released, there is no occurrence of buckling or pull-in voltage variation. However, when the beam
31
is a cantilever, a stress gradient if exists in the z direction or in the beam thickness direction results in upward warps in the beam due to stress release. For example, in case a plus stress gradient 2 MPa/&mgr;m exists along the z direction within the beam, the beam at its tip warp up 2 &mgr;m. Unless the stress gradient value can be accurately controlled to a predetermined value in the beam manufacture process, there occurs variation in warp degree whereby it becomes impossible to suppress the capacitance-decrease variation and pull-in voltage increase variation due to increase in the distance between the beam
31
and the electrode
32
. For example, pull-in voltage is 0.25 V in the case the stress gradient is zero in the absence of a warp, whereas pull-in voltage increases up to 1.2 V in the state of warping up 2 &mgr;m at the tip.
It is quite difficult to control, in the manufacture process of the beam, the compressive/tensile strength in the lengthwise direction and the stress gradient in the thickness direction. Although there is “anneal” as a stress relaxing method in the manufacture process, this process is to expose a device to an elevated temperature which temperature has an effect upon the device structuring members other than the beam. For example, in case the sacrificial material or the like, temporarily provided beneath the beam and finally etched away in order to make an electrode metal or beam in a bridge structure, is exposed to an elevated temperature, the material characteristic thereof changes. For this reason, because the element cannot be exposed to a high temperature, it is impossible to completely remove stress.
In the second thermal expansion problem, the beam causes therein thermal expansion in a lengthwise direction due to temperature rise around the element. In the both-ends-supported beam structure bound at both ends, the beam causes buckling to deflect irrespectively of electrostatic force application.
Next, the third stiction problem is mentioned.
FIG. 3
represents a relationship between a voltage and a capacitance in the case residual stress is suppressed nearly zero on a structure that the beam
31
of
FIG. 2
is made in a both-end-supported type. In case voltage is applied, pull-in takes place at 1.72 V. In case a voltage equal to or greater than that is applied, the beam
31
and the electrode
32
go into contact through the insulation layer
35
, increasing the contact area and hence the capacitance. Conversely, in case voltage is lowered, when the voltage is lowered down to 0.64 V, the contact between the beam
31
and the electrode
32
is gradually released. This is because of weak springiness, or spring restoration force, of the beam
31
. This means that, when the voltage is returned to 0 V, in case there exists an adsorbing force through the ambient water molecules, an adsorbing force due to residual charge or a van der Waals force in the contact area, the beam
31
cannot return to the initial state with high possibility. In order to avoid this, there requires a complicated structure,

Affiliated with

Also associated with

Rating

Variable capacitor and method for manufacturing same does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Variable capacitor and method for manufacturing same, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Variable capacitor and method for manufacturing same will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3363757