Field of the Invention
This invention is generally directed to the field of micromachined devices, and specifically for capacitive sensor devices associated with this field.
Background of the Invention
Contemporary capacitive pressure sensors can be grouped into at least two types of distinct architectures, each having their associated fabrication processes. Each of these architectures use a silicon substrate as a first capacitive electrode, separated by another substrate, of either silicon or a metal coated glass substrate, forming a complementary capacitor. Typically, the first silicon electrode forms a diaphragm whose deflection depends on an applied pressure, the flexibility determining a distance between the electrodes--thus a capacitance. To measure absolute pressure, the electrodes are encapsulated within a cavity. This cavity is loaded with a reference pressure, typically a vacuum.
In the case of the glass-silicon structure, an electrical connection for the complementary capacitive electrode, associated with the metal coated glass substrate, must be brought outside of the glass structure to connect it to a measurement circuit. This is typically done by tunneling a conductive feedthrough via through the glass structure. The conductive feedthrough is typically constructed of a metal material. Because the metal is confined within the glass tunnel, and the thermal coefficients of expansion are widely different between the metal and the glass, these type of structures suffer from poor field reliability over large temperature swings due to cracks forming either in the glass or in the metal due to stress formed between the relatively rigid and fragile glass and the relatively deformable metal material. When the cracks form, the reference vacuum pressure can be released thereby adversely effecting the accuracy of the device. This is a well characterized performance of the glass-silicon type of architecture. Further, since the conductive feedthrough adds dimension to the glass-silicon structure, these type of devices are necessarily larger.
On the other hand, those devices based on a silicon-silicon architecture suffer from non-linearity associated with non-uniform geometry of the diaphragm in relation to the complementary capacitive electrode as the diaphragm is deflected during pressure changes. Additionally, it is difficult at best to fabricate a consistent thickness for the flexible diaphragm because of time etch techniques often used to define the flexible diaphragm thickness. In those cases where an etch stop is used to ease processing requirements, the diaphragm is under significant stress because of heavy p-type doping of the silicon required to form the etch stop. This compressive stress makes the diaphragm buckle, thereby prohibiting a manufacturable thin diaphragm. Other etch-stop techniques include a electrochemical one by biasing a p-n junction during etch, which is cumbersome to fabricate. Both etch-stop techniques pose limits to the diaphragm thickness that only a relatively thin diaphragm is obtainable. Both the heavy doping and the p-n junction methods limit the thickness of the diaphragm to a few micrometers, thereby restricting applicable pressure range. Furthermore, silicon becomes fragile with doping-limiting the robustness of the doped diaphragm.
What is needed is an improved capacitive pressure sensor and method for fabricating same, that is more compact, reliable, has a more linear response to pressure changes and is more manufacturable.
Brief Description of the Drawings
FIG. 1 shows a cross-sectional view of a capacitive pressure sensor in accordance with a preferred embodiment of the invention;
FIG. 2 shows a cross-sectional view of various method steps for fabricating the capacitive pressure sensor shown in FIG. 1; and
FIG. 3 shows a top view of the capacitive pressure sensor shown in FIG. 1 highlighting an orientation of a rigidizing member.
Detailed Description of a Preferred Embodiment
A capacitive pressure sensor and method of fabricating the sensor includes bonding three silicon wafers at approximately 1200.degree. C. by using silicon oxide as a bonding agent. At 1200.degree. C. the oxide is sufficiently soft to fuse the silicon layers together. A top silicon layer is used as a rigid center. A thinner middle silicon layer serves as a pressure-sensitive diaphragm. This middle layer is ground and polished mechanically to achieve its thickness. A bottom silicon layer is used as a substrate supporting the whole structure. As mentioned above, silicon oxide layers are disposed between the silicon layers as a bonding agent. Furthermore, the silicon oxide between the top and middle silicon layers is used as an etch stop when the rigid center is formed using anisotropic etch. The oxide between the middle and bottom silicon layers serves as a spacer, providing space for the diaphragm to deflect when pressure is applied.
FIG. 1 shows a cross-sectional view of a capacitive pressure sensor constructed using a method in accordance with a preferred embodiment of the invention. FIG. 2 illustrates the various method steps in cross-sectional view for fabricating the capacitive pressure sensor shown in FIG. 1. The details of the method steps will be described first.
In FIG. 2, steps 1 and 2 include providing first and second silicon layers 123, 115, preferably in wafer form. Each of the first and second silicon layers 123, 115 have a layer of insulator disposed thereon. Preferably silicon oxide, shown at reference numbers 201, 203, 205, 207, is grown on a first major surface 202, 204 and on a surface 206, 208 opposing the first major surface 202, 204. The oxide thickness is typically in the order of one micrometer. For robust adhesion properties, required later in the process, the silicon oxide on the major surfaces 203, 205 of the first and second silicon layers 123, 115 are typically polished.
In STEP 3, a bonded interface 209 is formed between the first and second silicon layers 123, 115 by bonding together each of the silicon oxide layers 203, 205 at approximately 1200.degree. C. Preferably, the bonding is effectuated in steam or oxygen ambient.
A thickness dimension of the second silicon layer 115 is then reduced to a predetermined thickness 116, as shown in STEP 4, by removing the oxide layer 207, oriented on the surface 208 opposing the first major surface 204 of the second silicon layer 115, and a portion of the second silicon layer 115. This step of reducing exposes a second major surface 220 on the second silicon layer 115. The method of reducing a thickness 210 of these layers 207, 115 is mechanical grinding and polishing. Typically, the thickness 116 is in a range of about 10 to 200 micrometers and forms the basis of the above-mentioned pressure-sensitive diaphragm structure. STEP 4 represents a significant departure from prior art approaches. These prior art approaches used various etching process to define the thickness of the pressure-sensitive diaphragm. As mentioned in the background section it is difficult at best to fabricate a consistent thickness for the flexible diaphragm using time etch techniques. Also, etch stop techniques using heavy doping induce significant stress on the diaphragm causing it to buckle making relatively thin diaphragm structures impractical. Furthermore, electrochemical methods are cumbersome and render making relatively thicker diaphragm structures required in higher pressure sensory applications impractical. By mechanically reducing these layers 207, 115 first by grinding and then by polishing several benefits result. For instance this technique uses less process time than etch based methods. Also, the grinding and polishing method yields better dimensional definition of the diaphragm. Furthermore, a wide range of pressure applications is possible based on wide range of thicknesses possible by the mechanical reduction method for determination of the diaphragm thickness.
Next, in STEP 5, a thin oxide layer 211, typically ranging from 0.05 to 0.5 micrometers, is then grown on the surface 220 of the second silicon layer 115, wherein the second silicon layer 115 and the oxide layer 211 form a complete diaphragm structure 298. This oxide layer, or third insulating layer 211 is used to form an electrically insulating barrier between the surface 220 of the second silicon layer 115 and the above-mentioned, and yet to be detailed, bottom silicon layer.
In STEP 6, a third, or bottom, silicon layer 101 is provided. Two oxide layers 109, 217 are grown respectively on a third major surface 214 and on a surface 216 opposing the third major surface 214. Note that the oxide layer 109 forms a fourth insulating layer.
In STEP 7, in order to provide a space for the diaphragm 298 to deflect, a predefined volumetric area 114 of the oxide layer 109, oriented adjacent the third major surface 214 of the third silicon layer 101, is removed. The removal of the predefined volumetric area 114 is performed using a well known photolithographic process. This includes screening a photo mask onto a top surface 270 of the oxide layer 109, then etching out the predefined volumetric area 114 of the oxide layer 109 using buffered hydrofluoric acid. This process leaves an exposed surface of uninsulated silicon 218 bounded within a remaining perimeter structure 215 comprised of a portion of the oxide layer 109 in contact a portion 110 of the major surface 214 of the third silicon layer 101 surrounding the exposed surface of uninsulated silicon 218.
In STEP 8, a portion 240 of a fourth major surface 241 of the electrically insulating layer 211 is bonded to the top surface 270 of the perimeter structure 215 oriented on the third silicon layer 101. By doing this a cavity 113 is formed between an unbonded portion 242 of the fourth major surface 241 of the electrically insulating layer 211, the perimeter structure 215, and the exposed surface of uninsulated silicon 218.
All the above-mentioned bonds are formed by fusion at approximately 1200.degree. C. where the silicon oxide is sufficiently soft to bond the various silicon layers together. Alternatively, a low-temperature oxide such as Pyrex can be used. A layer of Pyrex film can be sputtered onto a silicon layer. Layers of silicon can then be bonded electrostatically at approximately 400.degree. C.
As mentioned above the oxide layer 211 was disposed to form an electrically insulating barrier between the surface 220 of the second silicon layer 115 and the exposed surface of uninsulated silicon 218 as the diaphragm 298 structure deflects under pressure. Alternatively, rather than growing the thin oxide layer 211 on the second major surface 220, the oxide layer may be grown on the exposed surface of uninsulated silicon 218 within the well area (the predefined volumetric area 114 of the oxide layer 109) of the third silicon layer 101.
In STEP 9, a masking layer 221, preferably silicon nitride, is then deposited onto a major surface 222 of the silicon oxide layer 201.
In STEP 10, using photolithography and preferably a dry, plasma etch method, the masking layer 221 is patterned to a predetermined area 224 so that it becomes a mask illustrated by reference number 223. This mask 223 is applied next to delimit removal of a portion of the silicon oxide layer 201.
The result of the removal of the portion of the silicon oxide layer 201 is shown in STEP 11. This removal is performed by the well known above-mentioned buffered hydrofluoric acid etch technique. The remaining portion of the silicon oxide layer 201 is shown by reference number 225. The remaining silicon oxide 225 is contained within an area 224 of substantially the same dimension as the predetermined area 224 of the mask 223 provided in the step (STEP 10) of patterning a mask 223.
Next, in STEP 12, a portion of the first silicon layer 123 is removed. Preferably the removal of this portion of the first silicon layer 123 is accomplished using an anisotropic etching method. This leaves a remaining silicon structure 227 bounded within substantially the same dimension as the predetermined area 224 proximate the remaining silicon oxide layer 225 and a larger area 228 adjacent the first major surface 202 of the silicon oxide layer 203. The relationship between the dimensions 224 and 228 bounding the remaining silicon structure 227 is well controlled because as a result of the anisotropic etching a resulting geometry including 4 facets, one facet shown by reference number 171, always form crystalline planes. During the anisotropic etch step, the silicon oxide layer 203 serves as an etch stop. By using the silicon oxide layer 203 as an etch stop process time is minimized. Also, due to both the use of the silicon oxide layer 203 as an etch stop and the well defined geometrical relationship of the crystalline planes accurate dimensional definition of the resulting structure is possible.
In STEP 13 the second silicon layer 115 is partially exposed so that it can be electrically connected to external signal conditioning circuits. This is accomplished by etching an additional portion of the silicon oxide layers 203 and 205. The remaining portions 229 and 231 are substantially bounded within the larger area 228 associated with the remaining silicon structure 227.
Then, in STEP 14, selective areas of each of the exposed surfaces of each of the second and third silicon layers are metalized to form electrically conductive contact pads 129 and 131. These electrically conductive contact contact, are bonding, pads 129 and 131 may be disposed for instance by sputtering a pad 129 onto a first surface 119 of the flexible silicone layer 115, and a pad 131 onto an opposing second surface 105 of the third silicone layer 101. The final structure is shown in FIG. 1.
In FIG. 1 the second and third silicon layers 115 and 101 act as opposing plates of a capacitor. The second silicon layer 115 is relatively flexible, and the third silicon layer 101 is relatively rigid. When an increase in pressure difference is applied between these layers 115 and 101, the opposing plates of the capacitor approach each other. This causes the capacitance of the structure to increase. This is typically measured across the electrically conductive contact pads 129 and 131.
A rigidizing structure comprised of elements 127, 125, 123, and 121 is used to improve the linearity of a change in capacitance of the sensor 100 due to a change in pressure. These elements 127, 125, 123, and 121 provide a stiffening action so that the diaphragm 298 structure does not bow as much when an increasing pressure is provided between the opposing plates of the capacitor. Preferably, the layered rigidizing structure is located proximate a center position with relationship to a perimeter of the first surface 103 of the rigid third silicon layer 101.
FIG. 3 shows a top view of the capacitive pressure sensor shown in FIG. 1 highlighting a preferred orientation of the rigidizing member. Since this is a top view only elements 127, 171, 173, 175, and 177 of the rigidizing structure are visible. Reference numbers 171, 173, 175, and 177 illustrate the above-described 4 facets of the first silicon layer 123. The chamber 113 is shown delimited by dashed lines and the second silicon layer 115 is also shown.
In conclusion, method steps for fabricating an improved capacitive pressure sensor have been taught above. The preferred embodiment uses mechanical grinding then polishing to define the diaphragm thickness. Furthermore, the usage of an oxide layer as an etch stop in combination with wafer bonding techniques allows a wider range of diaphragm thickness. These steps produce a sensor that is more compact, reliable, has a more linear response to pressure changes and is more manufacturable than prior art solutions.