Patent Application
20110168317
This disclosure relates to silicon substrate processing.
A microelectromechanical system (MEMS) typically has mechanical structures formed in a semiconductor substrate using conventional semiconductor processing techniques. A MEMS can include a single structure or multiple structures. The electromechanical aspect of MEMS is that an electrical signal activates each or is produced by actuation of each structure in the MEMS.
Various processing techniques are used to form MEMS. These processing techniques can include layer formation, such as deposition and bonding, and layer modification, such as laser ablation, etching, punching and cutting. The techniques that are used are selected based on a desired pathway, recess and hole geometry to be formed in a body along with the material of the body.
One implementation of a MEMS includes a body having chambers formed therein and a piezoelectric actuator formed on an exterior surface of the body. The piezoelectric actuator includes a layer of piezoelectric material, such as a ceramic, and conductive elements, such as electrodes, on opposite sides of the piezoelectric material. The electrodes of the piezoelectric actuator can either apply a voltage across the piezoelectric material to cause it to deform, or deformation of the piezoelectric material can generate a voltage difference between the electrodes.
One type of MEMS with piezoelectric actuators is micro-fluidic ejection devices. An actuator can include piezoelectric material that can be actuated by electrodes, causing the piezoelectric material to deform. This deformed actuator pressurizes a chamber, causing fluid in the chamber to exit, for example, through a nozzle. The structure components, including the actuator, the chamber and the nozzle, can affect how much fluid is ejected. In a MEMS with multiple structures, forming uniformly sized components for each structure across the MEMS can improve the uniformity of performance of the MEMS, such as the uniformity of fluid quantities that are ejected. Forming structures with uniformity of size of a few microns can be challenging.
In general, in one aspect, a method of bonding two substrates includes placing a separating member between a first substrate and a second substrate, applying pressure to the first substrate to initiate a bond wave between the first substrate and the second substrates with the separating member between the first substrate and the second substrate, and controlling movement of the bond wave by translating the separating member away from a center of the first substrate or the second substrate.
This and other embodiments can optionally include one or more of the following features. The method can further include monitoring the bond wave as the bond wave moves between the first substrate and the second substrate. The method can further include removing the separating member from between the first substrate and the second substrate after translating the separating member. The separating member can include a tapered portion and a non-tapered portion, and removing the separating member can include removing the tapered portion after the non-tapered portion. The method can further include determining a stopping point of the bond wave, and controlling movement of the bond wave can begin after the stopping point has been determined.
The separating member can be translated at a rate that is less than a maximum rate above which voids and bubbles can be trapped between the first and second substrates. The separating member can be translated at a rate of between about 50 mm/s to 70 mm/s. Pressure can be applied at between about 0.5 psi and 5 psi, such as about 1 psi.
The first substrate or the second substrate can include a patterned region including at least one die. The method can further include positioning the substrate having the patterned region such that a length of the at least one die is positioned along an axis that is at an angle of less than 30° from an axis extending along a length of the separating member. The angle can be about 17°.
Placing the separating member between the first substrate and the second substrate can cause there to be a gap of between about 0.5 mm and 5 mm at least one point between the first substrate and the second substrate. The gap can be about 1 mm.
The separating member can be placed approximately along a radial axis of the first substrate or the second substrate, and the separating member can extend along the radial axis by an amount that is less than a radial distance of the first substrate or the second substrate. The separating member can extend about 0.5 mm to 50 mm along the radial axis. The separating member can extend about 3 mm along the radial axis.
The pressure can be applied with a manual mechanism. The pressure can be applied by air from an automated air cylinder. The bond can be further initiated by sliding a pressure mechanism across a surface of the first substrate or the second substrate. The pressure mechanism can include a compliant material. The compliant material can be rubber. The pressure can be applied at a single pressure point on the first or second substrate.
The separating member can be the only separating member between the first and second substrates.
In general, in one aspect, an apparatus for bonding two substrates includes a substrate holding member configured to hold a first substrate, a separating member configured to separate the first substrate and a second substrate, a pressure inducer configured to apply pressure to the first or second substrate and initiate a bond wave between the first substrate and the second substrate, a monitoring device configured to generate images of a bond wave between the first and second substrates, and a mechanism connected to the separating member. The mechanism is configured to translate the separating member away from a center of the first or second substrate to control movement of the bond wave.
This and other embodiments can optionally include one or more of the following features. The monitoring device can be an infrared camera. The separating member can include a tapered portion. The separating member can have a length that is less than a radial distance of the first substrate or the second substrate. The separating member can be configured to align about along a line that bisects a center of the first or second substrate and a point where pressure is applied to the first substrate or the second substrate. The apparatus can further include a handle configured to move the separating member away from the substrate holding member when not in use. The mechanism can include a pocket configured to hold the separating member when not in use. The pressure inducer can be capable of exerting a pressure on the first substrate or the second substrate at an angle other than parallel to a main surface of the first substrate. The pressure inducer can be configured to apply a pressure at an angle between 90 degrees and 45 degrees to the main surface. The pressure inducer can have a tip that is less than 5 mm in diameter. The pressure inducer can be actuatable.
By placing a separating member between two substrates and translating the separating member away from the center of the substrates, the bond wave between two substrates can be precisely controlled. Controlling the bond wave can avoid the formation of voids and bubbles between substrates. Avoiding bubbles and voids when bonding substrates can result in fewer defects in substrates, which can increase product yield. Moreover, controlling the bond wave to ensure that the bond is not defective can reduce the number of defects that need to be tested for in the completed device.
The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.
Like reference symbols in the various drawings indicate like elements.
When two substrates are bonded together, e.g., with room temperature fusion bonding, the bond typically begins at an initial bonding region and propagates outward in a bond wave. If at least one of the substrates includes patterned or etched features, the movement of the resulting bond wave is affected by the patterned regions of the substrate. As a result, the bond wave will move faster over some areas of the substrate than other areas. Such uneven movement of the bond wave can cause voids and air bubbles to be trapped between the substrates, reducing the strength of the bond and creating defects in unbonded areas. By placing a separating member between the substrates, monitoring the bond wave as it moves between the substrates, and translating the separating member away from the center of the substrates, the bond wave can be controlled to move uniformly across the substrates and avoid the formation of voids and bubbles between the substrates. In some devices, there are features, such as recesses or apertures that are formed in one or both of the substrates. The voids and bubbles that are avoided using the techniques and devices described herein are other than desired recesses and/or apertures that are purposely formed in a substrate and required for proper device construction. In some implementations, the voids and bubbles that are created by improper bonding of two substrates are greater than 2 millimeters in diameter.
Referring to
A separator unit 630 can be used to prevent portions of the substrates 200, 240 from contacting. The separator unit 630 can include a separating member 620. The separating member 620 can project from the separator unit 630 and can be positioned to project between the main surfaces of the upper substrate 200 and the lower substrate 240, e.g., generally parallel to the surface of the lower substrate.
As shown in
Each separator unit 630 can include a holding member 632, e.g., a clamp, for securing the separating member 620. A motor 650 (see
In some implementations, the separating member 620 can be mounted on the clamp 632 such that it can pivot freely in the vertical direction, i.e. rather than being mounted rigidly in the clamp 632. Mounting the separating member 620 to pivot freely in the vertical direction can both facilitate alignment of the separating member 620 and facilitate loading of substrates 200, 240. For example, if the separating member is mounted to pivot freely in the vertical direction, then the separating member can be pivoted such that both substrates 200, 240 follow the taper as the separating member is removed to ensure that the substrates 200, 240 will come together smoothly.
The separator unit 630 can further include a handle 634 to move the separating member 620 from an extended state as shown in
As shown in
In operation, a lower substrate 240 is placed on the substrate holders 612 of the substrate support 610, the separating member 620 is lowered, and then the upper substrate 200 is placed on top of the supported lower substrate 240 at one edge and on the separating member 620 on the opposite edge. The substrates can be, for example, silicon or piezoelectric (e.g. PZT) substrates. The interface between the two substrates 200, 240 can be, for example, silicon-to-silicon, silicon-to-oxide, oxide-to-oxide, or BCB-to-silicon. One substrate can be, for example, a sacrificial substrate.
At least one of the substrates can have an etched or patterned portion 202, as shown in
The substrates and separating member 620 can be positioned so that an axis through a center of the length of the separating member 620 can be at an angle to an axis that runs along a length of ones or more of the dies 204. The angle can be less than 30°, such as about 17° or about 0° (i.e., be parallel). Further, the separating member can be aligned approximately along an axis that intersects the center of the substrates 200, 240, i.e. is aligned along a radial axis of the substrates 200, 240.
Referring back to
Referring to
After the separating member 620 has been placed between the substrates 200, 240, a pressure can be applied to the substrates 200, 240, such as by pressing on upper substrate 200. The pressure can be applied at a point 414 that is about 180° from the separating member 620, i.e., the pressure point can be applied on the opposite side of the substrates 200, 240 as the separating member 620. In some implementations, the pressure point is close to the substrates' edge. The pressure can be applied with a pressure inducer 412, which can be manually actuatable. Alternatively, the pressure inducer 412 can be an automated pressure inducer that actuates on a signal from the controller 660. The pressure inducer 412 can be made, for example, of a resin, such as polypropylene, for example, if it is a manual pressure inducer. The pressure inducer 412 can also be made, for example, of a compliant material, such as rubber, for example if it is an automated pressure inducer, so that when the inducer contacts the surface, it can flex and slide slightly across the surface of the substrate to initiate the bond between the two substrates 200, 240. The pressure inducer can have a tip that is less than 5 mm in diameter. Alternatively, the pressure inducer 412 can be an air cylinder, which ejects air onto the substrates to put pressure between the two substrates. The pressure inducer 412 is capable of exerting a pressure on the upper substrate 200 that is at an angle other than parallel to the main surface 680 of the lower substrate 240, for example at an angle of between 45° and 90° with the surface 680. A pressure of between about 0.5 psi and 5 psi, such as about 1 psi can be applied with the pressure inducer 412 at the pressure point 414.
Referring to
As the substrates 200, 240 are bonded together, the bond wave can be monitored using the monitoring device 400. The monitoring device can reveal the position of the bond front 502 between the substrates 200, 240. At a certain point, for example when the monitoring device 400 shows that the bond wave has stopped due to the separating member 420 pulling the substrates 200, 240 too far apart to bond, or when a sensor detects a particular position of the bond wave, the separating member 620 can be translated radially away from a center of the substrates 200, 240 along the axis 422, for example using the motor 650. As shown in
The rate at which the separating member 620 is translated can relate to the rate at which the bond front 502 propagates across the substrates. Further, the rate at which the bond front 502 propagates can relate to the activation level of the substrates 200, 240. For example, a silicon-to-silicon bond is considered highly active and bonds quickly, causing the bond front to move quickly across the substrates. As a result, the rate of translation of the separating member can be faster. If, however, one of the substrate surfaces is contaminated, then the substrates will be less active, and the bond front will move slower. As a result, the rate of translation of the separating member may be slower. Similarly, a silicon-to-oxide bond or oxide-to-oxide is less active than a silicon-to-silicon bond, so the bond front moves slower across the substrates, and the rate of the separating member may therefore also be slower than with the silicon-to-silicon bond.
Referring to
When fusion bonding is used to bond two substrates together without using a separating member as described herein, the movement of the bond front can be uneven. For example, the bond wave can move slower across patterned areas than nonpatterned areas. Likewise, the bond wave can move slower across patterned areas with deep etchings than patterned areas with shallow etchings. In some cases, the bond wave can move around a circular area between the two substrates, creating an area of trapped air that prevents the substrates on either side of the air bubble from coming in close enough contact to form the requisite Van der Waals bonding. Thus, uneven movement can cause voids and air bubbles to be trapped between the substrates, which can reduce the effectiveness of the bond or even form defectively bonded dies or devices. By translating the separating member away from the center of the substrates, the bond wave between the substrates 200, 240 can be precisely controlled so that the bond front moves straight across the substrates. That is, the bond front does not move such that two portions of the front move faster across the substrates than a third portion between the two portions and meet one another, trapping the third portion of the front as an edge of an air bubble. As a result of using the separator described herein, the bond wave can be forced to move across all portions of the substrates, e.g. portions that are deeply etched, shallowly etched, or not etched, at about the same rate, thereby significantly reducing or avoiding the generation of voids or air bubbles between the substrates.
A number of embodiments of the invention have been described. Other embodiments are within the scope of the following claims.
