Silicon-on-Silicon Differential Input Sensors

The present invention relates to diaphragm pressure sensors, and more particularly, to silicon-on-silicon differential input sensors with integrated signal conditioning circuitry and the process for producing them.

Sensors having a reduced thickness silicon diaphragm affixed to a glass wafer are known in the art. U.S. Pat. No. 4,543,457 issued to Petersen et al. on Sep. 24, 1985 ("the Petersen patent") discloses such a sensor. Sensors of this type may be used to measure a variety of parameters, such as pressure, temperature, acceleration or humidity. Silicon diaphragm sensors are reliable and may be fabricated at low cost. The output of such a sensor changes as the diaphragm is deflected by the condition sought to be measured.

A particular type of sensor uses variable capacitance. In variable capacitance sensors, the diaphragm is spaced apart from a confronting electrical plate. The capacitance between the diaphragm and the confronting plate changes in response to the deflection of the diaphragm. The usefulness of a given sensor is determined in part by the range of conditions over which the change in capacitance remains linear with respect to the change in input condition. Superior performance has been obtained with diaphragms of corrugated and bossed construction. These diaphragms deflect while remaining substantially parallel with the confronting contact, minimizing the undesirable effects of diaphragm curvature on sensor accuracy. Corrugated diaphragms have much larger regions of linearity than diaphragms without corrugations.

A major difficulty with conventional silicon-on-glass sensors of the type shown in the prior art is the difference between the coefficients of thermal expansion of silicon and glass. The silicon diaphragms are typically bonded to glass wafers by electrostatic bonding methods. Sensor inaccuracies result due to temperature changes because the silicon diaphragm and the glass wafer have different coefficients of expansion. This affects the geometric configuration of the device, resulting in degraded measurement accuracy. This problem could be avoided by mounting the silicon diaphragm on a wafer of similar silicon material; however, stray capacitances are introduced into the system when the diaphragm is bonded to silicon material.

Such stray capacitances present serious problems because they obscure accurate measurement of the capacitance between the diaphragm and the confronting electrode on the silicon wafer, effectively rendering measurements from the sensor useless. Historically, the thermal expansion problem, which arises from mounting the silicon diaphragm on glass, has been perceived by those skilled in the art as the least troublesome of the two. On the other hand, if the stray capacitance problem could be solved, silicon-on-silicon configurations free from thermal expansion problems could be produced.

Another problem with the silicon-on-glass construction is that it precludes the use of on board, integrated electronics for converting the output of the sensor into meaningful form and transmitting it to external electrodes. Instead, the circuitry must be produced separately and affixed to the glass wafer at a later time. These "hybrid" sensors are costly and inefficient to manufacture. Prior attempts to create one-piece silicon-on-silicon sensors have failed because the micro-electronics operations necessary to deposit integrated circuitry on the silicon are incompatible with the process of micro-machining operations necessary to create the diaphragm. No suitable solution to these problems has heretofore been found.

Accordingly, it is an object of the invention to provide a silicon-on-silicon capacitive sensor.

It is another object of the invention to provide such a sensor having monolithic integrated circuitry.

It is a further object of the invention to provide such a sensor having durable construction and capable of being produced at low cost.

These and other objects of the invention will become apparent to those skilled in the art when the following detailed description of the invention is read in conjunction with the accompanying drawings.

The present invention is a silicon-on-silicon capacitive sensor with monolithic integrated signal conditioning circuitry. It overcomes the thermal mismatch problem that occurs when a reduced thickness silicon diaphragm is mounted on a glass back plate. The support circuitry is designed to minimize the undesirable effects of stray capacitance historically encountered in prior attempts to utilize silicon-on-silicon technology. The support circuitry ma be configured to provide either analog or digital outputs. The sensor may be used to measure a variety of parameters, such as pressure, acceleration, temperature or humidity. It has a wide band of linearity and is particularly useful for accurately measuring pressures less than 0.5 PSI.

The sensor of the present invention is constructed by bonding together three silicon components: (1) a top plate having a mechanical pressure stop, (2) a reduced thickness diaphragm and (3) back plate containing CMOS circuitry. The diaphragm and back plate have two sets of confronting electrodes. One set provides a reference capacitance and the other a variable capacitance dependent on the deflection of the diaphragm. The three components are produced separately, avoiding the historic problem of incompatibility between micromachining and CMOS processing, which has previously made monolithic integrated support circuitry impractical. The components are then bonded together by eutectic soldering.

FIG. 1 is a cross-sectional side view of the capacitive sensor of the present invention.

FIG. 2 is a plan view in partial section of the capacitive sensor of the present invention.

FIG. 3 is a side view in partial section of a fully assembled sensor constructed according to the teachings of the present invention.

FIG. 4 is a schematic drawing of a first embodiment of integrated analog signal conditioning circuitry which may be deposited on the CMOS back plate of the present invention.

FIG. 4a is a timing diagram showing various signals associated with the circuit of FIG. 4.

FIG. 5 is a schematic drawing of a second embodiment of integrated analog signal conditioning circuitry which may be deposited on the CMOS back plate of the present invention.

FIG. 6 is a schematic drawing of an embodiment of integrated digital signal conditioning circuitry which may be deposited on the CMOS back plate of the present invention.

FIG. 7 is a schematic drawing of a relaxation oscillator suitable for use in the circuit of FIG. 6.

FIG. 7a is a schematic useful in understanding the effect of the guard region.

Referring to the drawings, FIG. 1 shows a cross-sectional side view of the capacitive sensor of the present invention. A silicon top plate 10 provides a mechanical pressure stop that limits the travel of the diaphragm when the sensor is exposed to pressures beyond its measurement range. The top plate 10 also acts as a buffer against package-induced stress. A first fluid access port 12 is formed in the top plate 10 by anisotropic etching, a process well-known in the art.

A diaphragm ring 14 has a central portion defining a reduced-thickness silicon diaphragm 16. The diaphragm 16 is typically 800-1200 microns square. The diaphragm 16 includes a boss structure 18, which contributes to the linearity of the sensor by reducing curvature of the diaphragm as it travels responsive to the force being measured. The boss 18 is created by anisotropic etching of the diaphragm ring 14 and is typically 500-800 microns square at the bottom tapering to 50-80 microns square at the top. The boss 18 also prevents destruction of the diaphragm under transient conditions by abutting the top plate 10. Thus, movement of the diaphragm 16 is limited to the useful range of the sensor. It should be noted that the location of the port 12 is offset from the position where the boss 18 would strike the top plate 10 to prevent it from becoming lodged therein.

Additionally, the diaphragm 16 is constructed to include corrugations 20. As previously noted, corrugations contribute to the linearity of the sensor. They reduce stress on the diaphragm 16 as it is deflected. The corrugations 20 are created by isotropic etching and are typically 10 microns deep and 25 microns wide. The diaphragm ring 14 also has a uniform recess 22 etched therein. The upper surface of recess 22 forms one of the electrode surfaces for a reference capacitance, as well as contributing to the flexibility of the diaphragm 16. An etch stop is diffused (typically 2 microns deep) into the bottom of the diaphragm to set the thickness of the reduced-thickness portion of the diaphragm. A P+ etch stop may be used; however, a lightly doped N-type electromechanical etch stop is preferred to preserve the well-behaved, low-stress nature of the bulk silicon.

The diaphragm 16 has a second fluid access port etched therein. The second port permits pressure equalization on both sides of the diaphragm. Thus, the measurements made by the sensor reflect the difference between the pressure source presented to the first port 12 and the second port. The top plate 10 is eutectically soldered to the diaphragm 14.

The diaphragm ring 14 is eutectically soldered to a back plate 26. The back plate 26 has a conductive pad 28 deposited thereon directly beneath the boss 18 to form a measurement capacitor in combination with the underside of diaphragm 16. The capacitance of the measurement capacitor changes as the diaphragm flexes responsive to the pressure differential provided to the first and second ports.

A second conductive pad 30 is deposited on the upper surface of the back plate 26 to create a reference capacitor when brought into confrontation with the underside of the diaphragm ring 14 exposed by the formation of the recess 22. The capacitance of the reference capacitor remains constant during operation of the sensor because this portion of the diaphragm ring remains at a fixed distance from the conductive pad 30.

A third conductive pad 32 (see FIG. 2) is deposited on the back plate 26 to act as a common terminal for the reference capacitor and the measurement capacitor. As will be apparent to those skilled in the art, the entire diaphragm ring 14 is conductive, thereby providing a common conductive surface for both the reference and measurement capacitors. The third conductive pad 32 electrically connects the ring 14 to the integrated signal conditioning circuitry, as will be fully described hereinafter.

FIG. 2 shows a plan view of the capacitive sensor of the present invention in partial section. The layout of the conductive pads 28, 30 and 32 may be clearly seen. The back plate 26 has an area 34 onto which integrated signal conditioning circuitry may be deposited. Inputs and outputs from this circuitry are electrically connected to a plurality of bond pads 36. The bond pads 36 may be connected to external leads by conventional wire bonding methods.

FIG. 3 shows a side view in partial section of a fully assembled sensor. A plastic housing 38 encloses the sensor assembly, providing openings for the first port 12 and the second port 24. The sensor assembly is secured to a base 40 having a plurality of external leads 42 mounted thereto. Wire bonds 44 connect the external leads 42 to the bond pads 36, effectively providing external access to the output from the signal conditioning circuitry.

A significant aspect of the present invention is the design of the integrated signal conditioning circuitry. The diaphragm has been designed such that its deflection is linear with the sensor input. This results in the measured capacitance C.sub.meas varying inversely with the sensor input. The object of the integrated circuitry is to generate an output directly proportional to the ratio of the reference capacitance to the measured capacitance and thus is directly proportional to the sensor input. Including the reference capacitance in the ratio cancels the effects of process variations on the gap between the capacitor plates and the variation of the dielectric constant of the fluid between the capacitor plates. A second object of the circuitry is to minimize the effects of stray capacitance introduced between each of the capacitor plates and the substrate which are introduced by the silicon on silicon technology. A constraint that the circuits must meet is that the diaphragm serves as a common terminal to both capacitors.

FIG. 4 shows an example of a switched capacitor analog signal conditioning circuit suitable for use with the sensor of the present invention. This circuit uses a switched capacitor gain cell to amplify a known dc input voltage V.sub.ref so that the output of the circuit V.sub.out, after sampling at the second stage, varies proportionally with the ratio of the reference capacitance C.sub.ref to the measured capacitance C.sub.meas. C.sub.ref is the fixed capacitance between the second conductive pad 30 and the portion of the diaphragm ring 14 exposed by the recess 22 (see FIG. 1). C.sub.meas is the capacitance between the first conductive pad 28 and the diaphragm 16. As previously described, this capacitance varies as the diaphragm 16 deforms responsive to the condition being measured.

The circuit of FIG. 4 is preferably implemented in CMOS technology where the switches will be simple complementary transistor pairs. The switches are turned on and off periodically as indicated in FIG. 4a. A logical 1 in FIG. 4a indicates that the associated switches are closed. .phi..sub.1 and .phi..sub.2 are non-overlapping waveforms of a two phase clock. Operation of the gain cell with this clock effects multiplication of V.sub.ref by the ratio of C.sub.ref to C.sub.meas. Clock signal .phi..sub.3 samples the output from operational amplifier 46. The sample is held by capacitor 49 and V.sub.out is buffered by operational amplifier 48.

Stray or parasitic capacitances 45 and 47 are introduced into the circuit because of the use of silicon in the construction of the diaphragm ring 14 and the CMOS back plate 26. The exact values of the stray capacitances 45 and 47 are difficult to determine with any precision, but are large compared to C.sub.ref and C.sub.meas. The topology of the circuit of FIG. 4 forces the stray capacitances 45 to low impedance nodes where their effect on the output is minimized. The stray capacitance 47 is connected to a virtual ground and its effective value is reduced by a factor equal to the open loop gain of the operational amplifier 46, thus minimizing its effect on the output.

FIG. 5 shows a second embodiment of analog signal conditioning circuitry employing an integrator/differentiator technique. This circuit employs the well-known fact that an op-amp integrator multiplies its input by 1/RC while an op-amp differentiator multiplies its input by RC. V.sub.out is forced to be proportional to C.sub.ref /C.sub.meas by employing an integrator, the output of which changes proportionally to 1/C.sub.meas, followed by a differentiator having an output proportional to C.sub.ref. An input voltage signal is generated by incorporating an integrating op-amp circuit 50 into a simple relaxation oscillator circuit 52. Two current sources 54 are employed to set the charging current. V.sub.int, the output of the integrator stage of the circuit, is presented to a differentiating stage 58 through C.sub.ref. Thus, V.sub.out is a square wave having an amplitude proportional to C.sub.ref /C.sub.meas. The effects of stray capacitance 60 is minimized because it is driven by the low impedance output of the integrating op-amp 50. The stray capacitances 56 are minimized by connection to virtual grounds. Their values are reduced by a factor equal to the open loop gain of the amplifiers. It will be apparent to those skilled in the art that V.sub.out may be input to a peak detection circuit on a sample and hold to obtain a dc voltage proportional to the parameter being measured by the sensor.

FIG. 6 shows an embodiment of digital signal conditioning circuitry suitable for use with the present invention. In this embodiment, the diaphragm, as the common terminal of both the measured and reference capacitors, is grounded. This results in stray capacitances in parallel with the reference and measured capacitances. In order to minimize the effect of the strays, electrically conductive guard regions 33 are introduced directly below electrodes 28 and 30. These regions are of opposite polarity to the polarity of the base and therefor form a PN junction as illustrated.

As shown in FIG. 7, these guard regions 33 are driven by a unity gain buffer amplifier 63 effectively removing C.sub.stray from the circuit. A relaxation oscillator incorporating this structure then is only responsive to the desired capacitance C.sub.meas.

Specifically, with reference to FIGS. 6 and 7, a first oscillator 62 is constructed so that the output frequency is proportional to the reciprocal of C.sub.meas. The output of the first oscillator 62 is connected to an input of counter 64 of the type commonly known in the art. A second oscillator 66 is constructed so that the output frequency is proportional to the reciprocal of C.sub.ref. The output of the second oscillator 66 is connected to a counter 68, the purpose of which is to control the gating or integration time of the first counter

The operation proceeds as follows. At the beginning of a cycle, counter 68 is preloaded with a value A and counter 64 is preloaded with a value B. Both counters accumulate counts from their respective oscillators for a period of time until counter 68 reaches zero. At this time, the value in counter 64 is transferred to external circuitry and the cycle is repeated. The value at the output is equal to Ka*C.sub.ref /C.sub.meas +Kb, where Ka and Kb are directly proportional to the preload values A and B, thus permitting simple adjustment of both the span and offset of the sensor. The circuitry of FIG. 6 allows a digital sensor output without costly and complex analog to digital conversion circuitry.

The present invention has been described with respect to certain embodiments and conditions, which are not meant to limit the invention. Those skilled in the art will understand that variations from the embodiments and conditions described herein may be made without departing from the invention as set forth in the appended claims.