Digital micro-machined ultrasonic transducer

Abstract

A pressure sensor device includes at least one membrane deflectable by an incident pressure or rarefaction wave, a cavity under the at least one membrane, a lower electrode placed at a lower portion of the cavity, an upper electrode placed on the at least one membrane, and at least one multi-state sensor placed at one or several locations inside the cavity and configured, upon reaching a threshold value, to change output from one state to a different state and vice versa when the at least one membrane comes within a specific distance from the at least one multi-state sensor.

Description

DIGITAL MICRO-MACHINED ULTRASONIC TRANSDUCERFIELD OF THE INVENTION

[0001] The present invention generally relates to ultrasonic transducers, and particularly to a digital micro-machined ultrasonic transceiver.BACKGROUND OF THE INVENTION

[0002] Standard, modern, silicon-based pressure sensors (like CMUT and PMUT) use thin silicon membranes as a sensing element for pressure waves. The thin membranes vibrate, or move, with incident pressure / rarefaction and external analog and digital circuitry is used to sense the instantaneous position of a sensing element (membrane) and present it in digital format.

[0003] Pressure sensors are devices designed to convert pressure into electrical signals. A pressure wave sensor, in its basic form, has a surface that interacts with an incident pressure (wave) and electrical or optical circuitry that converts the state (or movement) induced in this surface to measurable electrical signals. This conversion can be done directly, e.g., using piezoelectric materials to convert pressure to voltage, or using some intermediate phenomenon that is easier to measure and convert to electric current. The shape, size and material of the sensing surface are designed to maximize the performance (i.e., mechanical movement) of the sensor over the desired operating range.

[0004] In recent years, the size of pressure sensors has been reduced and their sensitivity increased using micro-machined electromechanical system (MEMS) manufacturing processes. The increased sensitivity comes at a price - mechanical movements in the order of a few nanometers need to be sensed electrically before conversion to digital format.

[0005] In the prior art, pressure sensors convert incident pressure to an analog electric signal whose strength is related to the deflection of a sensing membrane. The deflection is proportional to the incident pressure. The low signal generated by the membrane deflection is amplified and then passed through an analog-to-digital converter (ADC) to obtain a digital signal consumable by computers.

[0006] In the prior art, the main designs can be divided into 4 main families based on their conversion mechanisms:

[0007] PMUT - Piezoelectric micro-machined ultrasound transducers. Converting mechanical stress induced by the pressure wave to strain in a piezoelectric material, resulting in electrical voltage. Modern designs place a thin piezoelectric film on a thin flexible membrane. As the membrane is deflected by a pressure wave, voltage is induced on the piezoelectric film.

[0008] CMUT - Capacitive micro-machined ultrasound transducers. These sensors use the varying capacitance between a fixed substrate and a movable membrane to sense pressure. As the membrane is deflected by a pressure wave, the capacitance changes. The capacitance can be measured using current monitoring.

[0009] OMUT - Optical-based sensors. These types of sensors use optical sensors to measure the deflection of a mechanical part affected by the pressure wave.

[0010] FET-UT - Field effect transistor (FET) based sensors. The FET transistor is manufactured to be flexible, so the channel formed (and hence the current) depends on the pressure. One existing design uses a biased membrane above a FET acting as the gate. Another design places the FET on the membrane itself and uses the bending to control the channel.

[0011] Deficiencies of analog sensors

[0012] In all cases, these devices produce a low electric signal which is amplified and then passed through an analog-to-digital converter (ADC) to obtain a digital signal consumable by computers.

[0013] The first problem with this approach is that small signals are susceptible to thermal and electric noise which limits their resolution. To reduce the effect one usually increases the area of the moving part susceptible to the pressure (e.g., membrane) or uses multiple CMUTs / PMUTs / FET-UT elements connected in parallel to increase the signal strength.

[0014] The second problem arises from the amplifier. Such amplifiers are usually complex, power-hungry, and expensive as they need to minimize added noise and also be relatively linear to preserve the details of the input.

[0015] The third problem arises from the ADC. An ADC with high resolution and range is complex and expensive in both cost and die area.

[0016] Many MEMS sensors try to circumvent some of the problems by changing some bias voltage in the system that changes the operating / sensitivity range, i.e., the sensor can be operated in different ranges but not simultaneously.

[0017] Others attempt to improve accuracy by optical means, in which a laser beam is projected on the pressure sensor membrane and the reflected signal is analyzed to determine the amount of deflection caused by pressure. Optical sensing provides better signal / noise and more accurate reading.

[0018] The emergence of MEMS-based ultrasonic imaging created new areas of application for ultrasound imaging. Those new applications have power, size and cost constraints that are difficult to meet with standard xMUT technologies: the low levels of electrical signals require power-hungry, expensive, external circuitry and need multiple elements (membranes) to achieve the required sensitivity. There is a clear need to reduce the complexity and physical size of ultrasonic sensors.

[0019] The invention described hereinbelow is a device and method that converts pressure directly into a digital signal without using an ADC and sophisticated amplifiers. This design can allow for greater pressure sensitivity at high frequencies and simpler design and integration with silicon IC processes.

[0020] It is noted that US Patents 7589456 and 8014231 describe capacitive membranes used as digital sensors. The membranes act as binary devices, which are in a collapsed or non-collapsed state. As the membranes collapse and touch the electrodes on the bottom a digital signal is formed. By providing drum heads (membranes and associated gaps) with different response characteristics, the drum heads of an element digitally indicate the amplitude of the acoustic force by which of the drum heads are triggered or change states. This approach places several severe restrictions on the proposed device: First, for a high resolution sensor with, say, 14 bit resolution, the device will have 4096 drumheads in the case of simple binary detection, one for each resolution step, resulting in a very large physical device that is impractical for any wearable or flexible implementation. Even in the case of using multiple electrodes per drum head, say, 2 or 4 thresholds per drumhead, -2,000-1,000 drumheads per device - the device will still be too large. The number of adjacent electrodes under a single membrane is severely limitedas entire rings are touched by the membrane at a given pressure and the electrodes need to be wide enough to discern the pressure and act without an amplifier.

[0021] Second, the design increases the measurement error because collapse is a highly non-linear effect (for multiple states after collapse) and is far more sensitive to fabrication error.

[0022] Lastly, the design is far more complicated than designing many membranes when taking into account all other operating requirements: Each different membrane has a different frequency band, the collapse has hysteresis (the releasing pressure is different from the collapse and takes time), the collapse also causes charging in the insulator, etc. The only way to circumvent all of those problems without another technological leap is to add more membranes, thus increasing the area further.

[0023] As will be described below, the current invention decouples the membrane design from the binary / multi-stage sensor design configuration, enabling improved design flexibility, as well as allowing for sensing mechanisms that can integrate a significantly higher number of binary or multi-stage sensors under a single membrane. This reduces the overall DMUT area, a critical factor for achieving high resolution imaging.

[0024] The following explanation is given to understand a difference between US Patents 7589456 and 8014231.

[0025] CMUTs have three operating zones:

[0026] a) Traditional: the membrane does not touch the cavity bottom, Vbias<Vcollapse. That is, the voltage that biases the membrane from its neutral, unbiased state is less than the voltage required to collapse the membrane to touch the cavity bottom.

[0027] b) Collapse: Vbias>=Vcollapse, the membrane collapses on the cavity bottom.

[0028] c) Snapback: Vbias<Vsnapback, the membrane disengages from the cavity bottom.

[0029] However, due to inertia and hysteresis, when the bias voltage Vbias is reduced below Vcollapse, the membrane does not immediately completely lift off the cavity bottom and snapback to the “no contact” state. Rather, it starts to lift gradually off the cavity bottom, starting from the membrane edge inwards. Thus, when Vbias starts to decrease and is reduced below Vcollapse, the membrane initially still contacts some of the cavity bottom. As Vbias continues to decrease, the membrane continues lifting untilVsnapback is reached. At Vsnapback, the membrane does not touch the cavity bottom. Just before reaching Vsnapback, a very small amount of the membrane touches the cavity bottom.

[0030] The most effective zone for the sensor is the area between collapse and snapback voltages. In this area the membrane is closer to the bottom, capacitance is higher and traditional CMUT sensors operate best. In this zone the membrane displacement due to incoming pressure is the largest.

[0031] US Patents 7589456 and 8014231 operate the sensors in the collapsed state. The sensors do not sense pressure in the region between collapse and snapback.

[0032] Background of MOSFET transistors

[0033] This section includes a brief background of MOSFET transistor characteristics. The discussion and example used in the patent description are for n-channel enhancement mode MOSFET, but the invention can equally work with other types of MOSFET with the relevant modifications.

[0034] In n-channel enhancement-mode MOSFET, there is no conduction channel between the source and drain when the voltage between the gate and substrate is below a certain threshold value Vtfl. The threshold value depends on the geometry of the transistor and the voltage between the source and substrate. More specifically:

[0036] Where:

[0037] VTh0the threshold voltage for 0 substrate bias.

[0038] VSB- source to body (substrate) voltage.

[0039] (f>F- Surface potential.

[0040] tox- thickness of oxide between the gate and substrate (or other insulating material).

[0041] eox " Permittivity of silicon and oxide respectively.

[0042] qe, NA- elementary charge of an electron and doping concentration, respectively.

[0043] Y - body effect coefficient: y = —j2q eeesiNAox

[0044] For the case of a FET transistor with a gate separated by air (tair, eaLr) and oxide (tox, eox) as is the case of in the FET-based pressure sensor, the body constant changes to.

[0045] The threshold voltage depends on the ambient temperature through the surface potential and Ef / l0parameters. For a small enough temperature range the dependency can be linearized to:

[0046] T) = Vth(T0) + ath-T(T - To)

[0047] The current from the source to the drain is given by: IDSK(^GS ~ Vrn)2-SUMMARY

[0048] The present invention seeks to provide a digital micro-machined ultrasonic transceiver, which uses an element or elements located internally to the pressure sensor (intra-cavity) to sense the instantaneous position of a membrane and output a digital representation of the position, thereby eliminating expensive external circuitry as well as reducing the area of the sensor. This provides significant technical and financial benefits.

[0049] A membrane is clamped to the edge of a cavity and is used as a sensor to detect static pressure levels and pressure wave fields. The membrane operates as a continuous, linear and non-linear device that converts the instantaneous pressure to mechanical movement and location. One or more sensors, each designed to sense the mechanical distance between the multi-state sensor and the membrane, is placed within the cavity, below the membrane. Each multi-state sensor is designed to switch state when the membrane comes within a specific distance from the specific sensor. A circuitry aggregates the outputs from the multi-state sensors to a multi-state representation of the instantaneous location and movement of the membrane.

[0050] The present invention takes a new approach that avoids the above mentioned issues found in the prior art. The sensing mechanism operates independently of membrane collapse and physical contact, and the operating design of the membrane is decoupled from the design of the sensing mechanism resulting in a much smaller device. Moreover, the multi-state sensor, which is located at the bottom of the cavity, electronically measures the mechanical distance between the multi-state sensor and the membrane using technology from the mainstream silicon industry that enables significantreduction in sensor size leading to significantly smaller device size. The multi-state sensor is thus significantly greater in accuracy and cost-efficiency compared to the prior art sensors.

[0051] There is provided in accordance with a non-limiting embodiment of the invention a pressure sensor device including at least one membrane deflectable by an incident pressure or rarefaction wave, a cavity under the at least one membrane, a lower electrode placed at a lower portion of the cavity, an upper electrode placed on the at least one membrane, and at least one multi-state sensor placed at one or several locations inside the cavity and configured, upon reaching a threshold value, to change output from one state to a different state and vice versa when the at least one membrane comes within a specific distance from the at least one multi-state sensor. In one embodiment, the membrane does not touch the cavity bottom.BRIEF DESCRIPTION OF DRAWINGS

[0052] The present invention will be understood and appreciated more fully from the following detailed description, taken in conjunction with the drawings in which:

[0053] Fig. 1 is a simplified illustration of a digital micro-machined ultrasonic transceiver (DMUT), in accordance with a non-limiting embodiment of the invention, showing a gap with a membrane and multi-state element arrays at the bottom of the device.

[0054] Fig. 2 is a simplified illustration of the digital micro-machined ultrasonic transceiver, showing the full sensor with several differently biased membranes and a temperature calibration.

[0055] Figs. 3A and 3B are simplified illustrations of the device with FET-transistor arrays at the bottom of the gap, DMUT based on FET transistor array with different gate thickness, respectively without applied pressure (Fig. 3A), and with applied pressure (Fig. 3B).

[0056] Figs. 4A and 4B are simplified illustrations of the DMUT based on a cantilever array with different shapes, in which Fig. 4A shows a side view of a gap, membrane and a single element, and Fig. 4B shows multiple cantilevers of different shapes forming an array.

[0057] Fig. 5 is a simplified illustration of encoding the information from the element arrays to a digital signal.

[0058] Fig. 6 is a simplified illustration of encoding the information from multiple element arrays in multiple separate gaps.

[0059] Fig. 7 is a simplified illustration of placing element arrays at different locations under the membrane.

[0060] Fig. 8 is a simplified illustration of cantilever operation of the DMUT.

[0061] Fig. 9 A is a simplified schematic illustration of the DMUT operation, in accordance with a non-limiting embodiment of the invention.

[0062] Fig. 9B is a simplified schematic illustration of bias dependent multipoint triggering using an equal-strength electric field curve.

[0063] Fig. 9C is a simplified schematic illustration of a comparison of pressure sensing schemes using binary sensor devices.

[0064] Fig. 9D is a simplified schematic illustration of binary / multi-state sensor states when applying different incident pressure scenarios at different Vbias.DETAILED DESCRIPTION

[0065] A preferred embodiment described in this application includes a device, a system and methods for a digital micro-machined ultrasound transceiver. A membrane is used as an analog sensor to convert pressure to mechanical deflection and a multi-state sensor, having two or more distance / proximity thresholds, is used to convert the deflection into a digital signal representing the said deflection.

[0066] A digital sensing element (which could be binary having two states, or which could be multi-state, having more than two states) is integrated into the cavity of the pressure sensor. Each of the multi-state elements emits a signal when the pressure on the membrane is above a certain pressure level:

[0067] The nth multi-state sensor emits a signal snwhen the pressure p on the membrane exceeds a certain threshold:

[0069] By design, each two-state digital element triggers at a different pressure on the membrane. For simplicity, it is assumed that the elements are ordered so that pn< pn+1.

[0070] A device comprises a membrane, a cavity and a plurality of binary sensors, each tuned to a different threshold value, said threshold representing a specific distance between the membrane and the binary sensor. The device is able to represent the instantaneous position and movement of the membrane.

[0071] A system may be constructed when the mechanical size of the device is not able to include the total amount of binary sensors required to achieve a certain level of accuracy. A system is built of two or more devices. The close proximity of the two or more devices guarantee practically identical properties and dividing the number of binary sensors amongst several devices also allows for better accuracy of each binary sensor.

[0072] A method is used to aggregate the outputs of all binary sensors into an nBit digital word that may be used by computer code. Other methods control the sensitivity of binary sensors and may be used for calibration, both during manufacturing and standard operation, of a device or a system.

[0073] In the description, there are two distinct implementations using two different types of two state elements:

[0074] 1) A device built by fabricating multiple FET transistors inside the gap without a gate (i.e. only the doped source, drain and a channel with oxide above). The transistors are fabricated with different insulator (such as oxide) thicknesses and at different locations below the membrane to give different threshold voltages.

[0075] 2) A device built by fabricating microscale cantilevers inside the gap.

[0076] The membrane described in this preferred implementation may be designed to sense physical or chemical properties other than pressure while maintaining the methods of converting the deflection and motion of the membrane.

[0077] Fig. 1 depicts the basic structure of the device and invention.

[0078] The device includes a cavity (20) fabricated between a substrate (30) and a membrane (40), a bottom electrode and a top electrode (10), voltage may be applied between the top and bottom electrode to cause the membrane to bend (60). The cavity may hold a vacuum or may be at least partially filled with a gas (such as an inert gas), a liquid, or a gel.

[0079] The invention includes sensors (5xx), placed at the bottom of the cavity and designed to continuously sense the state and motion of the membrane.

[0080] The device is fabricated using available MEMS processes. In one embodiment the fabrication involves starting with a plain silicon wafer, etching cavities, oxidizing surfaces and bonding the patterned silicon wafer to a second wafer that serves as the membrane. Other variations use glass wafer as the substrate, with or without through- silicon or through-glass vias, Sol wafer as the temporary carrier for the membrane layer or use non-silicon material as a membrane.

[0081] Figure 9 depicts the principle of operation of a DMUT:

[0082] The sensing membrane 205 is brought to an idle, concave, position 207. This may be done either by fabricating the cavity 201 with vacuum inside or by applying a bias voltage 210 or both. Incident pressure, static or dynamic, causes the membrane to move towards the bottom electrode 202 (when pressure is present) or away from the bottom electrode (in the case of rarefaction). The maximum membrane deflection 211 and the membrane motion range are set during the design phase of the device but may be finetuned at a later stage. They may be designed never to reach a collapsed state, but the invention is not limited to this. In one embodiment of the invention voltage is applied between top electrode 204 and bottom electrode 202 and the resulting electric field present between the membrane 205 and the bottom electrode 202 is sensed by the binary sensors 203 and the output of the sensors is sampled and converted to a digital word. The electric field varies as the membrane 205 deflection changes. Each binary sensor 203 is designed to switch state when the electric field experienced by the specific binary sensor exceeds a set threshold. To reach a resolution of N bits 2Nbinary sensors 203 are required, each tuned to switch state at the appropriate electric field level that corresponds to a specific membrane deflection.

[0083] In one implementation, each binary sensor 203 is designed to trigger at a specific electric field level corresponding to a specific membrane-electrode distance and bias voltage pair. The binary sensor is designed so a specific incident pressure will deflect the membrane to the specific membrane-electrode distance and cause the binary switch to change state. The voltage applied between top electrode 204 and bottom electrode 202 may be used to allow the same binary sensor 203 to trigger at different membrane to bottom electrode distances, hence at different incident pressure levels. This is depicted in Fig 9B. The solid graph represents all the combinations of Voltage / Distance operatingpoints that result in the same effective electric field at the binary sensor. By applying voltages Vbiasl to VbiasN a given binary sensor 203, designed to trigger at specific membrane-electrode distance and bias voltage pair, can be adjusted to trigger at several different membrane to bottom electrode distances (4 are depicted in Fig 9B. in this example, four levels are depicted in Fig. 9B, though this number can be greater depending on the number of distinct bias voltages applied and corresponding sensor settings, hence the number of different binary switch designs (namely the electric field trigger level) can be reduced significantly resulting in fewer unique sensor designs, and simpler, more reliable manufacturing process while maintaining digital resolution.

[0084] In a fixed single bias implementation, achieving an N-bit resolution may require up to 2AN distinct binary sensors, which can be impractical. In an alternative bias tuned implementation, the bias voltage between electrodes 204 and 202 is varied so that the same sensor 203, with the same electric field trigger level, is made to trigger at different membrane-to-electrode distances (and thus different incident pressures). This reduces the number of distinct sensor types while preserving resolution.

[0085] Pre-Calibration Step

[0086] Controller operation using discrete biases. During each measurement, the controller changes the bias voltage in a series of steps (Vbiasl to VbiasN) in sequence and samples the output of sensor 203 at each set point, all while measuring the physical membrane deflection.

[0087] The controller detects a transition of sensor 203 at point V and registers the membrane deflection (d) 211, the bias voltage and incident pressure and organizes the data in a calibration table Vbias-d-Incident Pressure.

[0088] The calibration data is then processed to generate correction factors that will bring the c-Incident Pressure graph to a required linearity level.

[0089] In one embodiment, a binary switch may be calibrated in the following process:

[0090] A. Applying initial Vbias 1 , increasing the incident pressure wave magnitude until the binary switch changes state and registering the pressure level Pl causing this change of state.

[0091] B. Adjusting the incident pressure wave magnitude to Pl -A, A being the smallest observable pressure change or other higher value, and increasing Vbias until the binary switch is triggered and changes state.

[0092] C. Repeating the above steps two or more times.

[0093] D. Constructing a calibration table with Vbias-IncidentP levels.

[0094] In another embodiment of this invention magnetic sensing may be used.

[0095] The digital outputs of the 2Nbinary sensors 203 are connected to a converter where an N bit word is generated.

[0096] Non-limiting Reduction to Practice #1: Binary Sensors (5xx in Fig. 1) using FET transistors

[0097] The device is built by fabricating multiple FET transistors on a substrate without a gate (i.e. only the doped source, drain and a channel with oxide above. The transistors are fabricated with different insulator (such as an oxide) thicknesses and at different locations below the membrane to give different threshold voltages. It is noted that oxide is just one example of an insulator that can be used. Non-limiting examples of insulators include silicon dioxide (SiO2), silicon nitride (Si3N4), hafnium oxide (HfO2), tantalum pentoxide (Ta2Os), aluminum oxide (A12O3), zirconium dioxide (ZrO2), barium strontium titanate (BST), and strontium titanate (SrTiO3).

[0098] By design, each FET allows current to flow at a different level of pressure on the membrane. All transistors are linked to a digital unicode-to-binary encoder to provide a binary readout of the pressure. For example, a 10-bit resolution requires 1024 transistors, each enabling conduction at a different pressure.

[0099] In terms of fabrication, it is possible to fabricate noxdifferent thicknesses using log2nox) masks. A more practical approach for the fabrication of the device:

[0100] Use nano-imprinting lithography (NIL) to create a multi-level mask followed by ion-milling. The ion-milling will remove the NIL resist first, resulting in a multi-level insulator.

[0101] Note that fabricating the NIL template itself requires multiple masks, but this is done only once.

[0102] Using reflow to reshape the resist and form a convex, lens-like shape with different heights.

[0103] Variations of the FET array design

[0104] It is possible to further reduce the number of heights by placing the FET transistors at different locations below the membrane. Because the membrane is attached at the edges, areas closer to the center move further than those near the edges, so that FETs with the same thickness of insulation become active at different pressures. Due to the differences in the electric field between the bottom electrode(s) and the different areas below the membrane, fewer oxide thicknesses are needed. Using the same 10-bit resolution example, the 1024 levels can be realized by 8 locations and only 128 heights. The sensing at each region is linear but does not share the same scale as the other regions, and the digital part combines them into one signal.

[0105] Another variation of the device is to use an array of multiple membranes, each with its own copy of straight-to-digital FET transistors. Each membrane is supplied a different bias voltage (essentially a different baseline gate-to-substrate voltage) so that the channel becomes active at different pressure ranges.

[0106] Yet another version is a hybrid between a digital and analog version. The analog is responsible for either the low-bits or the high-bits and the FET transistors are responsible for the rest of the bits. This allows for simpler and faster amplifiers and ADC. For example, a full 12-bit sensor can be realized with the high 8-bit straight-to-digital part, which is used to drive a digital-to-analog (DAC) to produce a reference voltage to an ADC to obtain the rest of the 4 bits.

[0107] Variations of the unicode-to-binary encoder

[0108] Due to fabrication imperfections, the pressure thresholds of all transistors to emit a signal are not distributed evenly (usually they still form an increasing series). To correctly determine the original pressure the result of the unicode-to-binary encoder goes into a lookup table to retrieve the original pressure.

[0109] Correcting for temperature variations

[0110] Because the threshold voltage of the FET transistors is temperature-dependent, a separate set of “standard” FET transistors is placed near the DMUT membranes. “Standard” in the sense stands for having an electrode on the gate and no membrane ontop. From this transistor set it is possible to obtain the temperature and calibrate the rest of the system.

[0111] Non-limiting Reduction to Practice #2 - Binary Sensors (5xx in Fig 1) are Cantilever Switches

[0112] MEMS based cantilevers are used extensively in applications like accelerometers and are widely known and well understood. In a typical application, the cantilever is used to measure acceleration - acceleration causes the cantilever to bend, and this bending is sensed and converted to a measurement of acceleration. In accelerometer application the force applied to the cantilever is generated by acceleration and the displacement can be measured using several methods, e.g. a piezo sensor.

[0113] In this invention, the cantilever is used as a binary switch. Fig. 8 depicts the operation of a single cantilever. The cantilever (104) has a mechanical constant K that is proportional to its length, width, height and type of material and is connected between a post (101) and contact (100). A voltage source Vth (103) is connected to the cantilever and the resulting force Fl pulls the edge of the cantilever towards the contact (100). The force Fl is proportional to distance d* (105) and Vth (103).

[0114] At the same time the membrane, connected to Vbias, generates a force Fm that opposes Fl.

[0115] As long as Fl>Fm the cantilever is in contact with the contact (101) and detector (106) indicates “state 0”. Once the membrane deflects to a distance that causes Fm to be higher than Fl the balance of forces releases the cantilever from the contact (101) detector (106) indicates “state 1”.

[0116] Several parameters can be used to set the trigger point for a cantilever: Vth, d* and K (the shape of the cantilever). This means that trigger points can be controlled during design (d* and K) and during operation (Vth).

[0117] Fig. 4 A depicts the operation of a complete device.

[0118] The dotted line depicts the membrane location at rest (i.e. only Vbias is applied, without any external pressure), the lower line depicts the membrane location at the maximal designed pressure and the upper line depicts the membrane location at the maximal designed rarefaction.

[0119] At any membrane location / deflection within the motion range, the force between the membrane and each specific cantilever changes. The mechanical properties of the cantilever, combined with its mechanical gap d* and Vth, are designed to cause an upward movement once Fm is greater than Fl. A complete device is built by fabricating multiple cantilevers, each designed to trigger at different pressure / rarefaction levels, covering the entire motion range. The digital output of the cantilevers is then fed to a unicode-to-binary encoder to generate an equivalent digital sample. Fig. 4B depicts a set of cantilevers with different trigger points.

[0120] The actual number of cantilevers depends on the required resolution. At low- resolution, a single membrane may be enough to implement a complete sensor. At high resolutions, where many cantilevers are required, the size of the cavity and membrane may be too small to accommodate them in a single membrane device. In this case, several membranes / cells will be used and the total amount of cantilevers will be divided between them. This mode adds an additional control parameter to the design as the design of each membrane cell may be different in order to have a better membrane / bias / cantilever design.

[0121] Fig. 9 A is a simplified schematic illustration of the DMUT operation, in accordance with a non-limiting embodiment of the invention.

[0122] Fig. 9B is a simplified schematic illustration of bias dependent multipoint triggering using an equal-strength electric field curve. Plot of inter electrode bias voltage (Vbias) versus membrane to sensor distance (<7) showing an equal-strength electric field curve. Discrete bias levels Vbiasl-VbiasN intersect the curve at respective distances dl- dN, indicating that the same binary sensor 203 can be caused to trigger at multiple membrane positions by changing Vbias. The respective distances dl-dN correspond to the motion range Ad of Fig. 9, that is, the different positions corresponding to different bias levels.

[0123] Fig. 9C is a simplified schematic illustration of a comparison of pressure sensing schemes using binary sensor devices. The left panel illustrates a simple approach (Fig. 2 or Fig. 3) where each pressure level is detected by a distinct binary sensor, requiring multiple sensor designs for multiple pressure levels. The right panel illustrates the proposed method in which a single binary sensor is used for N pressure sensing levels byapplying different bias voltages, allowing detection of corresponding pressure levels and significantly reducing the number of unique sensor devices required.

[0124] Fig. 9D is a simplified schematic illustration of binary / multi-state sensor states when applying different incident pressure scenarios at different Vbias. The left panel illustrates the incident pressure scenarios while the four columns on the right indicate the status of binary / multi-state sensor (on / off) at each Vbias. Pressure levels (L3> L2> L2> LI) and Vbias levels ( Vbias 1> Vbias2> Vbias3> Vbias4).

[0125] It is noted that the invention also provides a method for reducing the number of different binary or multi-state sensor device designs in a DMUT while maintaining fixed resolution, the method comprising: determining, such as but not limited to, through simulations or lab experiments, two or more Vbias-distance combinations that result in the same electric field on the binary or multi-state sensor; determining, such as but not limited to, through simulation or lab experiments, the level of incident pressure that will cause the membrane to deflect to each of the distances in the Vbias-distance list; and creating, for each binary or multi-state sensor design, a calibration table that lists the incident pressure level that causes the binary or multi-state sensor to switch state Vs the bias voltage. Using different bias voltage values for different trigger points may reduce a number of distinct binary sensors otherwise required at a fixed bias to realize an N-bit equivalent resolution.

Claims

CLAIMSWhat is claimed is:

1. A pressure sensor device comprising: at least one membrane deflectable by an incident pressure or rarefaction wave; a cavity under said at least one membrane; a lower electrode placed at a lower portion of said cavity; an upper electrode placed on said at least one membrane; and at least one multi-state sensor placed at one or several locations inside said cavity and configured, upon reaching a threshold value, to change output from one state to a different state and vice versa when said at least one membrane comes within a specific distance from said at least one multi-state sensor, the membrane not being in collapsed state.

2. The pressure sensor device according to claim 1 , wherein said cavity has a cavity bottom and said membrane does not touch said cavity bottom.

3. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor comprises at least one field-effect (FET) transistor that comprises a source, a drain, a substrate and a layer of an insulator on said substrate, but said FET transistor is without a gate, and said threshold value depends on a geometry of said FET transistor and a voltage between said source and said substrate.

4. The pressure sensor device according to claim 3, wherein said at least one multi-state sensor is located at a bottom of said cavity and said threshold value depends on a thickness of said insulator.

5. The pressure sensor device according to claim 3, wherein said at least one multi-state sensor is located above a bottom of said cavity and said threshold value depends on a height of said at least one multi-state sensor above the bottom of said cavity.

6. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor comprises one or more relatively less accurate sensors that control a threshold value of a relatively more accurate sensor.

7. The pressure sensor device according to claim 1, wherein a bias voltage is applied between said lower and upper electrodes.

8. The pressure sensor device according to claim 7, wherein the device is configured to apply a plurality of discrete bias voltage levels between the lower and upper electrodes so as tocause the same multi-state sensor to switch state at a plurality of different membrane to sensor distances corresponding to a plurality of incident pressure levels.

9. The pressure sensor device according to claim 1, wherein said cavity contains a vacuum.

10. The pressure sensor device according to claim 1, wherein said cavity is at least partially filled with a liquid, gas or gel.

11. The pressure sensor device according to claim 1, wherein said cavity is held at atmospheric pressure.

12. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor comprises a cantilever.

13. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor comprises a magnetic sensor.

14. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor is fabricated as part of a fabrication process of the entire pressure sensor device.

15. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor is fabricated separately from a fabrication process of the entire pressure sensor device.

16. The pressure sensor device according to claim 1, wherein said at least one multi-state sensor comprises at least one binary sensor.

17. The pressure sensor device according to claim 16, further comprising a unicode-to- parallel encoder configured to convert 2Nbinary inputs from said at least one binary sensor to a N bit word representing a deflection of said at least one membrane.

18. The pressure sensor device according to claim 1, further comprising a lookup table or conversion function for converting a deflection of said at least one membrane to instantaneous incident pressure or rarefaction.

19. The pressure sensor device according to claim 1, further comprising a processor configured to determine elevation of said pressure sensor device above or below sea level by comparing a sensed deflection of said at least one membrane with a known deflection of said at least one membrane at sea level.

20. The pressure sensor device according to claim 19, wherein said processor is configured to calibrate said pressure sensor device by obtaining a difference between an actual deflection measurement of said at least one membrane at a local elevation above or below sea level and anexpected deflection at said local elevation, and using the difference to correct the deflection measurement.

21. A method for reducing the number of different binary or multi-state sensor device designs in a DMUT, the method comprising: determining two or more Vbias-distance combinations that result in the same electric field on the binary or multi-state sensor; determining the level of incident pressure that will cause the membrane to deflect to each of the distances in the Vbias-distance list; and creating, for each binary or multi-state sensor design, a calibration table that lists the incident pressure level that causes the binary or multi-state sensor to switch state Vs the bias voltage.

22. The method of claim 21, wherein using different bias voltage values for different trigger points reduces a number of distinct binary sensors otherwise required at a fixed bias to realize an N-bit equivalent resolution.

Images