MICROELECTROMECHANICAL SYSTEM HAVING MOVABLE ELEMENT INTEGRATED INTO SUBSTRATE-BASED PACKAGE
A method of fabricating a MEMS device is disclosed. A metal layer is provided over a first surface of a substrate including over an opening. The metal layer is patterned to define a membrane segment and a pad, with the membrane segment extending at least partially across the opening. An integrated circuit chip is attached over the opening to the membrane segment and pad, with the integrated circuit separated from an extending portion of the membrane segment by a gap. The integrated circuit chip includes a conductive member so that deflection of the extending portion relative to the conductive member can be measured as a change in capacitance.
Latest Texas Instruments Incorporated Patents:
This application is a divisional of application Ser. No. 13/670,242 filed Nov. 6, 2012, incorporated by reference herein, which is a continuation of application Ser. No. 12/701,683 filed Feb. 8, 2010 (now U.S. Pat. No. 8,304,274), incorporated by reference herein, which claims the benefit of Provisional Application Nos. 61/291,767 filed Dec. 31, 2009, and 61/152,607 filed Feb. 13, 2009.FIELD OF THE INVENTION
This relates in general to the field of semiconductor devices and processes, and more specifically to the structure and fabrication of microelectromechanical system (MEMS) devices having a movable element integrated into a substrate-based ball grid array package and a sensing element built on an integrated circuit.DESCRIPTION OF RELATED ART
The wide variety of products collectively called microelectromechanical system (MEMS) devices are small, low weight devices on the micrometer to millimeter scale produced on the basis of batch fabrication techniques similar to those used for semiconductor microelectronics devices. MEMS devices integrate mechanical elements, sensors, actuators, and electronics on a common carrier. MEMS devices have been developed to sense mechanical, thermal, chemical, radiant, magnetic, and biological quantities and inputs, and produce signals as outputs.
MEMS devices may have parts moving mechanically under the influence of an energy flow (acoustic, thermal, or optical), a temperature or voltage difference, or an external force or torque. Certain MEMS devices with a membrane, plate or beam can be used as a pressure sensor or actuator (for instance microphone and speaker), inertial sensor (for instance accelerometer), or capacitive sensor (for instance strain gauge and RF switch); other MEMS devices operate as movement sensors for displacement or tilt; bimetal membranes work as temperature sensors. Besides small size, the general requirements for the membrane- or plate-operated sensors include long term stability, small temperature sensitivity, low hysteresis for pressure and temperature, resistance to corrosive environments, and often hermeticity.
In a MEMS device, the mechanically moving parts are fabricated together with the sensors and actuators in the process flow of the electronic integrated circuit (IC) on a semiconductor chip. As an example, the mechanically moving parts may be produced by an undercutting etch at some step during the IC fabrication. Bulk micromachining processes employed in MEMS device sensor production for creating, in bulk semiconductor crystals, the movable elements and the cavities for their movements include anisotropic wet etching, reactive ion etching (RIE), and deep reactive ion etching (DRIE). These techniques employ photolithographic masking, are dependent on crystalline orientation, and need etch stops, all of which are expensive in terms of time and throughput. In addition, there are bulk and surface micromachining techniques for building up structures in thin films on the surface of semiconductor wafers, also expensive techniques. While many of the processes are expensive to implement, some processes, such as automatic wafer bonding, are inexpensive.
Because of the moving and sensitive parts, MEMS devices have a need for physical and atmospheric protection. Consequently, MEMS devices are surrounded by a housing or package, which has to shield the MEMS device against ambient and electrical disturbances, and against stress. For many devices, fully hermetic and even quasi-hermetic packages represent a significant cost adder, especially when ceramic packages or precision parts such as glass plates are required.
Among the basic operating principles of pressure sensors are piezoresistive, capacitive, and resonant operation. In the piezoresistive operation, the pressure is converted to an electronically detectable signal, wherein the conversion relies on the elastic deformation of a structure such as a membrane exposed to the pressure; pressure causes strain, and strain causes change of electrical resistivity. In MEMS device silicon technology, controlling the membrane thickness, size, and alignment involves precision process steps. In the resonant operation, the pressure causes mechanical stress in the vibrating microstructure; the resonance frequency is measured in dependence on the mechanical stress. Excitation and damping of the MEMS device silicon diaphragm and the nonlinear frequency-pressure relationship require sophisticated calibration. In the capacitive operation, the pressure causes a displacement-dependent output signal. The change in pressure causes a displacement, the displacement causes a capacitor change, and the capacitor change causes electrical signal—similar to the operation of a condenser microphone. Nonlinearity and parasitic capacitances and residual membrane stress represent challenges for MEMS device membrane fabrication of silicon and epitaxial silicon.
Taking the example of capacitive pressure sensors, several fabrication methods may be chosen. In one method, the sensors are bulk micro-machined as a glass-silicon-glass structure with vertical feed-throughs. In another method, a preferentially etched wafer receives deep and shallow boron diffusions and dielectric depositions, which are mounted on glass so that the wafer can finally be dissolved. In yet another method, a surface micro-machined capacitive pressure sensor is created by a polysilicon layer (1.5 μm thick) separated by a gap (0.8 μm wide) over the n-F doped silicon electrode; the sensor is monolithically integrated with the sensing circuitry. The sensors are small and span an operating range from about 1 bar to 350 bar, have high overpressure stability, low temperature dependence and low power consumption.
In the basic operating principle of accelerometers, the mechanical and electrical sensitivity are a function of the vertical displacement of the movable plate's center. In displacement sensing accelerometers, the applied acceleration as input is transformed into the displacement of the movable mass (plate) as output; a suspension beam serves as the elastic spring. Force sensing accelerometers detect directly the force applied on a proof mass. The MEMS device fabrication in bulk single-crystal silicon of the movable plate, the suspension beam, and the proof mass requires sensitive semiconductor etching techniques.SUMMARY
Applicants believe manufacturing cost is the dominant factor preventing the widespread integration of pressure sensors, microphones, accelerometers and other applications in which a movable member is needed to convert an external analog input into an electrical output, into systems in the automotive, medical, and aerospace industries.
Applicants saw that building a MEMS device on the surface or within the wafer by standard wafer fab technology and standard wafer fab lithographic methods is not only a high cost approach, but also limits the choice of materials and configuration available for the MEMS device components, which have to be compatible with the standard wafer process. After the wafer fabrication, in standard technology the MEMS devices still have to be packaged using known packaging material and processes—another cost adder.
Applicants solved the problem of mass-producing low cost semiconductor-centered MEMS devices by integrating the movable MEMS device parts, such as mechanical elements and sensors, including their complete fabrication with low-cost device materials and packages, and by leaving only the electronics and signal-processing parts in the integrated circuitry. The package, into which the movable parts are integrated, may either be a leadframe-based or a substrate-based plastic molded housing. With this invention, the MEMS devices may use a standard CMOS chip without any movable structure and a packaging component with movable structures built therein.
Applicants further discovered that the separation of movable and electronics parts provides greater system level integration with other components such as package-on-package MEMS devices, thus increasing the electrical product efficiency.
In embodiments, which have the movable elements integrated into a substrate-based package, the substrate may be a stiff multi-layer substrate, such a multi-metal-layer FR-4 board, or a flexible film substrate, such as a metalized polyimide tape. The latter devices need a molded encapsulation for robustness. Packages can be stacked with solder bodies as connecting elements.
Embodiments of this invention include the usage of electrostatic force, acceleration, air pressure, etc., to deflect a beam or membrane for building microphones, pressure sensors, accelerometers, and other applications where a movable member is needed to convert an external input into an electrical output.
Example MEMS devices of the pressure sensor family, operating on capacitive changes caused by a movable membrane, may offer 80% lower fabrication cost, when the membrane is integrated into the plastic device package instead of being fabricated in conventional manner as a portion of the silicon chip.
One embodiment of the present invention provides a MEMS device comprising: a flat substrate having a thickness, a first surface and an opposite second surface; an opening through the thickness of the substrate, the opening extending from the first to the second surface; a metal foil attached onto the first surface of the substrate, the foil including a plurality of pads and a membrane extending at least partially across the opening; and an integrated circuit chip flip-assembled to the pads, the chip at least partially spanning across the opening, separated from the membrane by a gap.
Another embodiment of the present invention provides a method for fabricating a MEMS device comprising the steps of: forming an opening from a first surface to an opposite second surface of a flat substrate; laminating a metal foil onto the first substrate surface and at least partially across the opening so that the foil adheres to the substrate; patterning the metal layer into a plurality of pads and a segment; and flip-connecting a semiconductor chip having electronic circuitry onto the pads so that the chip spans across at least partially across the opening, separated from the segment by a gap.
In the embodiment of
MEMS device 100 has an opening 122 through the thickness 121 of substrate 120. In
Since pads 110 and 111, and membrane 112 are portions of the metal foil attached to substrate surface 120a, they have the same thickness 113; for many embodiments, thickness typically is between 10 and 25 μm. In this thickness range, membrane 112 is flexible in the direction normal to the first substrate surface and movable in the space of the opening 122 and of the gap 107.
The example of
In the example embodiment of
The MEMS device of
Since pads 210 and 211, and membrane 212 are portions of the metal foil attached to substrate surface 220a, they have the same thickness 213; for many embodiments, thickness 213 is between about 5 and 50 μm, typically between 10 and 25 μm, but may be thicker or thinner in other embodiments. In this thickness range, membrane 212 is flexible in the direction normal to the first substrate surface and movable in the space of the opening 222 and of the gap 207.
Another example embodiment of the invention, generally designated 300, is illustrated in
Substrate 320 further has an opening 322 through the thickness of substrate 320, extending from first substrate surface 320a to second substrate surface 320b. In
A metal foil 310 is attached onto first substrate surface 320a, typically using an adhesive layer 330 to support the attachment. The metal foil may be made of copper or a copper alloy; other options include nickel, or an alloy containing an iron-nickel alloy (such as Alloy 42 or Invar™), or aluminum. The preferred foil thickness is between 10 and 25 μm, but may be thicker or thinner in other embodiments. The foil is patterned to form a membrane 312 without the adhesive layer. In the foil thickness range indicated, membrane 312 is flexible in the direction normal to the first substrate surface 320a and movable in the space of the opening 322 and of the gap 307. Acceleration is transformed into the displacement of the movable membrane 312, and the position change is measured as a change of the capacitance relative to the fixed metal layer 108 on chip 101.
Since the stack of tape 320, adhesive layer 330, and metal foil 310 has a total thickness in the range from approximately 70 to 150 μm, it is flexible. When some applications require a mechanically more robust MEMS device, the stack may be strengthened by adding a body 350 of hardened plastic compound such as a polymerized epoxy-based molding compound, optionally filled with inorganic particles of silicon dioxide or silicon nitride. The strengthening body 350 has an opening 352 through the body thickness. Opening 352 feeds into opening 322 so that is allows the unobstructed operation of membrane 312 in moving in the z-direction. Opening 352 may be shaped as a cylinder, a truncated cone wider on the outside and narrowing towards width 322a, as depicted in
Flip-assembled chip 101 spans at least partially across the opening 322 so that metal plate 108 forms a capacitor with membrane 312. Plate 108 is separated from membrane 312 by a distance, which is composed of the sum of gap height 307a, the thickness of tape 320, and the thickness of adhesive layer 330. As stated above, the movable part, membrane 312, can move in this distance in the z-direction, normal to the plane of the membrane. In some embodiments, movable part 312 includes the suspension beam of length 315 and the movable plate of length 316. Movable plate 316 has an area equal to the area of the fixed plate 108 on the chip surface in order to form a capacitor. In addition, for some embodiments the mass of the movable plate 316 can be enlarged by adding the mass of a deformed gold sphere 314, as formed in the well-known wire ball bond process. Mass 314 represents the proof mass.
The invention allows the selection of the materials and dimensions for opening 322 (and 352), length of suspension beam 315, area of movable plate 316, mass 314, and capacitance between movable plate 316 and fixed plate 108. Consequently, the accelerometer of
The example embodiments shown in
Substrate 420 has an opening 422 through the thickness of the substrate, extending from first substrate surface 420a to second substrate surface 420b. In
Substrate 420 has a patterned metal foil attached to its first substrate 420a; an optional adhesive attachment layer is not shown in
For the example embodiment of
For the substrate pad sets 410 and 413, a plurality of chip terminals 402 allow the connections to solder bodies, gold bumps, or gold alloy. The gold bumps may be produced by a wire ball bonding technique, followed by a flattening process with a coining technique. The gold alloy may be a low melting gold/germanium eutectic with 12.5 weight % Ge and an eutectic temperature of 361° C. In addition, chip terminal 403 may be configured as a seal ring to allow the formation of a seal ring made of solder or gold alloy to seal the enclosed space at least quasi-hermetically, e.g. against environmental disturbances such as particles, but not completely against gaseous and moisture molecules.
Sensing plate 408 and membrane 412, typically having the same area and being separated by a gap, form a capacitor. As stated above, membrane 412 is made of a metal (for example, copper) thin enough (for example, 10 μm) to be flexible and sensitive to pressure changes. The assembled device 400, therefore, works as a MEMS device for pressure sensor and microphone. Responding to pressure arriving through opening 422 by bending inward and outward, membrane 412 modifies distance 407 relative to stationary plate 408. Let the area of membrane 412, and plate 408, as electrodes be A; the distance between the electrodes under original pressure be Do; and the dielectric constant of the space between the electrodes be ∈, then the capacitance C of the electrodes is given by
Pressure in z-direction deforms the flexible membrane so that the deformed area has to be calculated as an integral over small area elements dx·dy, while the distance Do is modified in both x-direction and y-direction by a deflection wx,y. The resulting change of capacitance is measured by the circuitry of chip 401, operating as a microphone or a pressure sensor. A miniature speaker can be built in a similar way by driving the membrane electrostatically.
It should be stressed that the embodiments shown in
In order to give a cost estimate for the example pressure sensor MEMS device, the side lengths of the molded material 450 in
The concept to integrate the movable part of a MEMS device with a substrate-based package rather than with the chip allows numerous variations with the goal to sensitize certain aspects of the measurement, or to include new aspects into the measurement.
The symmetrically balanced design of a single membrane is shown in
Another embodiment of the invention is a process for fabricating substrate based MEMS devices with the movable element integrated into the device packages. For the process flow shown in
The patterning of metal layer 910 into a plurality of pads and a segment-to-become-membrane begins with depositing a photoresist layer 960 across the metal foil on the flat first substrate surface and another photoresist layer 961 across the contoured second substrate surface. As
The result of the patterning steps for metal layer 910 is illustrated in
In the step depicted in
In the next process step, illustrated in
Alternatively, before the cuts along line 980 in
Next, substrate 920 is prepared for the cuts along line 980, for instance by a saw, to singulate the substrate into discrete MEMS devices. These separated units are depicted in
While this invention has been described in reference to illustrative embodiments, this description is not intended to be construed in a limiting sense. Various modifications and combinations of the illustrative embodiments, as well as other embodiments of the invention, will be apparent to persons skilled in the art upon reference to the description. As an example, the invention applies to any material for the MEMS device package, including plastics and ceramics, and the semiconductor device, integrated circuits as well as discrete devices, including silicon, silicon germanium, gallium arsenide, or any other semiconductor or compound material used in manufacturing.
As another example, the integration of the movable element into the leadframe-based package of a MEMS device can be applied to piezoresistive pressure sensors, where the conversion of pressure to an electronically detectable signal relies on the elastic deformation of a membrane, or generally of a structure, that is exposed to the pressure.
As another example, the integration of the movable element into the leadframe-based package of a MEMS device can be applied to resonant pressure sensors, where the resonance frequency depends on the mechanical stress in the vibrating microstructure.
As another example, the method of integrating the movable element into the MEMS device package allows an inexpensive fine-tuning of the mechanical transfer function by controlling the thickness of the membrane and by adding one or more mass units of squashed balls produced in wire bonding technique.
As another example, the integration of the movable element into the substrate-based package of a MEMS device can be applied to piezoresistive pressure sensors, where the conversion of pressure to an electronically detectable signal relies on the elastic deformation of a membrane, or generally of a structure, that is exposed to the pressure.
As another example, the integration of the movable element into the substrate-based package of a MEMS device can be applied to resonant pressure sensors, where the resonance frequency depends on the mechanical stress in the vibrating microstructure.
As another example, the method of integrating the movable element into the substrate-based MEMS device package allows an inexpensive fine-tuning of the mechanical transfer function by controlling the thickness of the membrane and by adding one or more mass units of squashed balls produced in wire bonding technique.
It is intended that the appended claims encompass any such modifications or embodiments.
1. A method of fabricating a MEMS device, comprising:
- providing a substrate having an opening extending between first and second surfaces;
- providing a metal layer over the first surface including over the opening;
- patterning the metal layer to define a membrane segment and a pad, with the membrane segment extending at least partially across the opening; and
- attaching an integrated circuit chip over the opening to the membrane segment and pad, with the integrated circuit separated from an extending portion of the membrane segment by a gap; the integrated circuit chip including a conductive member positioned and configured so that deflection of the extending portion relative to the conductive member can be measured as a change in capacitance.
2. The method of claim 1, wherein providing the metal layer includes providing an adhesive layer over the first surface including over the opening; and providing the metal layer over the adhesive layer.
3. The method of claim 2, wherein providing the metal layer includes providing a tape over the first surface including over the opening; and attaching the metal layer by adhesive to the tape.
4. The method of claim 2, further comprising patterning the adhesive layer to remove the adhesive layer from over the opening.
5. The method of claim 4, wherein patterning the adhesive layer includes patterning the adhesive layer by removing a portion of the adhesive layer through the opening.
6. The method of claim 5, wherein patterning the metal layer includes forming a first layer of photoresist over the metal layer over the first surface, forming a second layer of photoresist over the second surface including within the opening, and patterning the metal layer by patterning the first layer of photoresist without patterning the second layer of photoresist.
7. The method of claim 5, wherein patterning the metal layer defines further pads spaced laterally outwardly from the membrane segment and pad; the method further comprising attaching solder balls to the further pads.
8. The method of claim 7, further comprising forming an encapsulation material over the second surface, the encapsulation material having an opening aligned with the substrate opening.
9. The method of claim 8, wherein the encapsulation material opening is tapered outwardly in a direction away from the substrate.
10. The method of claim 9, wherein the metal layer is a metal foil.
11. The method of claim 10, wherein the metal foil comprises at least one of copper, copper alloy, nickel, an iron-nickel alloy, or aluminum.
12. The method of claim 1, wherein patterning the metal layer includes forming a first layer of photoresist over the metal layer over the first surface, forming a second layer of photoresist over the second surface including within the opening, and patterning the metal layer by patterning the first layer of photoresist without patterning the second layer of photoresist.
13. The method of claim 1, wherein the metal layer is a metal foil.
14. The method of claim 1, wherein patterning the metal layer defines further pads spaced laterally outwardly from the membrane segment and pad; the method further comprises attaching solder balls to the further pads.
15. The method of claim 1, further comprising forming an encapsulation material over the second surface, the encapsulation material having an opening aligned with the substrate opening.
16. The method of claim 15, wherein the encapsulation material opening is tapered outwardly in a direction away from the substrate.
17. The method of claim 1, wherein the integrated circuit is attached by solder material to the membrane segment and pad, and the gap is set by the solder material.
18. A method of fabricating a MEMS device, comprising:
- providing a substrate having openings extending between first and second surfaces;
- attaching a layer of adhesive material over the first surface including across the openings;
- providing a metal foil over the layer of adhesive material including over the openings;
- patterning the metal layer to define a membrane segment and pad respectively associated with each opening, the membrane segment extending at least partially across the associated opening;
- attaching an integrated circuit chip over each opening to the associated membrane segment and pad, with the integrated circuit separated from an extending portion of the membrane segment by a gap; the integrated circuit chip including a conductive member positioned and configured so that deflection of the extending portion relative to the conductive member can be measured as a change in capacitance.
19. The method of claim 18, further comprising patterning the adhesive layer by removing through the openings the portions of the adhesive layer covering the openings.
20. The method of claim 19, further comprising singulating the substrate into discrete MEMS devices.
International Classification: H01L 29/84 (20060101);