Fabrication of liquid emission device with asymmetrical electrostatic mandrel

Abstract

A liquid emission device includes a chamber having a nozzle orifice. Separately addressable dual electrodes are positioned on opposite sides of a central electrode. The three electrodes are aligned with the nozzle orifice. A rigid electrically insulating coupler connects the two addressable electrodes. To eject a drop, an electrostatic charge is applied to the addressable electrode nearest to the nozzle orifice, which pulls that electrode away from the orifice, drawing liquid into the expanding chamber. The other addressable electrode moves in conjunction, storing potential energy in the system. Subsequently the addressable electrode nearest to the nozzle is de-energized and the other addressable electrode is energized, causing the other electrode to be pulled toward the central electrode in conjunction with the release of the stored elastic potential energy. This action pressurizes the liquid in the chamber behind the nozzle orifice, causing a drop to be ejected from the nozzle orifice.

Description

FIELD OF THE INVENTION

[0002] The present invention relates generally to micro-electromechanical (MEM) drop-on-demand liquid emission devices such as, for example, ink jet printers, and more particularly such devices which employ an electrostatic actuator for driving liquid from the device.

BACKGROUND OF THE INVENTION

[0003] Drop-on-demand liquid emission devices with electrostatic actuators are known for ink printing systems. U.S. Pat. Nos. 5,644,341 and 5,668,579, which issued to Fujii et al. on Jul. 1, 1997 and Sep. 16, 1997, respectively, disclose such devices having electrostatic actuators composed of a single diaphragm and opposed electrode. The diaphragm is distorted by application of a first voltage to the electrode. Relaxation of the diaphragm expels an ink droplet from the device. Other devices that operate on the principle of electrostatic attraction are disclosed in U.S. Pat. Nos. 5,739,831, 6,127,198, and 6,318,841; and in U.S. Pub. No. 2001/0023523.

[0004] U.S. Pat. No. 6,345,884, teaches a device having an electrostatically deformable membrane with an ink refill hole in the membrane. An electric field applied across the ink deflects the membrane and expels an ink drop.

[0005] IEEE Conference Proceeding “MEMS 1998,” held Jan. 25-29, 2002 in Heidelberg, Germany, entitled “A Low Power, Small, Electrostatically-Driven Commercial Inkjet Head” by S. Darmisuki, et al., discloses a head made by anodically bonding three substrates, two of glass and one of silicon, to form an ink ejector. Drops from an ink cavity are expelled through an orifice in the top glass plate when a membrane formed in the silicon substrate is first pulled down to contact a conductor on the lower glass plate and subsequently released. There is no electric field in the ink. The device occupies a large area and is expensive to manufacture.

[0006] U.S. Pat. No. 6,357,865 by J. Kubby et al. teaches a surface micro-machined drop ejector made with deposited polysilicon layers. Drops from an ink cavity are expelled through an orifice in an upper polysilicon layer when a lower polysilicon layer is first pulled down to contact a conductor and is subsequently released.

[0007] One such device is disclosed in co-pending U.S. patent application Ser. No. 10/153,990 filed in the names of Gilbert A. Hawkins, et al on May 23, 2002. That device includes an electrostatic drop ejection mechanism that employs an electric field for driving liquid from a chamber in the device. Structurally coupled, separately addressable first and second dual electrodes are movable in a first direction to draw liquid into the chamber and in a second direction to emit a liquid drop from the chamber. A third electrode between the dual electrodes has opposed surfaces respectively facing each of said first and second electrodes at an angle of contact whereby movement of the dual electrodes in one of the first and second directions progressively increases contact between the first and third electrodes, and movement of the dual electrodes in the direction progressively increases contact between the second and third electrodes.

SUMMARY OF THE INVENTION

[0008] The device described in the Hawkins et al. patent application, and other multi-layer microelectromechanical electrostatic actuators for liquid emission devices, can be manufactured by chemical mechanical polishing in combination with a sacrificial layer to produce a member, having planar surface and a non-planar surface, that can move within a trench left when the sacrificial layer is removed to provide a separation from stationary parts.

[0009] According to a feature of the present invention, a drop-on-demand liquid emission device, such as for example an ink jet printer, includes an electrostatic drop ejection mechanism that employs an electric field for driving liquid from a chamber in the device. Structurally coupled, separately addressable first and second dual electrodes are movable in a first direction to draw liquid into the chamber and in a second direction to emit a liquid drop from the chamber. A third electrode between the dual electrodes has opposed surfaces respectively facing each of said first and second electrodes at an angle of contact whereby movement of the dual electrodes in one of the first and second directions progressively increases contact between the first and third electrodes, and movement of the dual electrodes in the direction progressively increases contact between the second and third electrodes.

BRIEF DESCRIPTION OF THE DRAWINGS

[0010] FIG. 1 is a schematic illustration of a drop-on-demand liquid emission device according to the present invention;

[0011] FIG. 2 is a cross-sectional view of a portion of drop-on-demand liquid emission device of FIG. 1;

[0012] FIGS. 3-5 are top plan views of alternative embodiments of a nozzle plate of the drop-on-demand liquid emission device of FIGS. 1 and 2;

[0013] FIG. 6 is a cross-sectional view of the drop-on-demand liquid emission device of FIG. 2 shown in a first actuation stage;

[0014] FIG. 7 is a cross-sectional view of the drop-on-demand liquid emission device of FIG. 2 shown in a second actuation stage;

[0015] FIG. 8 is a top view of a portion of another embodiment of the liquid emission device of FIG. 1;

[0016] FIGS. 9-22 are cross-sectional views taken along line A-A′ of FIG. 8 and showing the sequence of fabrication of a drop ejector;

[0017] FIG. 23 shows a cross-section through B-B′ of FIG. 8;

[0018] FIG. 24 shows a cross-section through C-C′ of FIG. 8;

[0019] FIG. 25 shows a cross-section through D-D′ of FIG. 8; and

[0020] FIGS. 26-40 are cross-sectional views of a second preferred embodiment of the present invention, taken along line A-A′ of FIG. 8 and showing the sequence of fabrication of a drop ejector;

[0021] FIG. 41 shows a cross-section through B-B′ of FIG. 8;

[0022] FIG. 42 shows a cross-section through C-C′ of FIG. 8; and

[0023] FIG. 43 shows a cross-section through D-D′ of FIG. 8.

DETAILED DESCRIPTION OF THE INVENTION

[0024] As described in detail herein below, the present invention provides a process for fabricating drop-on-demand liquid emission devices. The most familiar of such devices are used as printheads in ink jet printing systems. Many other applications are emerging which make use of devices similar to ink jet printheads, but which emit liquids (other than inks) that need to be finely metered and deposited with high spatial precision.

[0025] FIG. 1 shows a schematic representation of a drop-on-demand liquid emission device 10, such as an ink jet printer, which may be operated according to the present invention. The system includes a source 12 of data (say, image data) which provides signals that are interpreted by a controller 14 as being commands to emit drops. Controller 14 outputs signals to a source 16 of electrical energy pulses which are inputted to a drop-on-demand liquid emission device such as an ink jet printer 18.

[0026] Drop-on-demand liquid emission device 10 includes a plurality of electrostatic drop ejection mechanisms 20. FIG. 2 is a cross-sectional view of one of the plurality of electrostatically actuated drop ejection mechanisms 20. A nozzle orifice 22 is formed in a nozzle plate 24 for each mechanism 20. A wall or walls 26 that carry an electrically addressable electrode 28 bound each drop ejection mechanism 20.

[0027] The outer periphery of electrode 28 is sealingly attached to wall 26 to define a liquid chamber 30 adapted to receive the liquid, such as for example ink, to be ejected from nozzle orifice 22. The liquid is drawn into chamber 30 through one or more refill ports 32 from a supply, not shown, typically forming a meniscus in the nozzle orifice. Ports 32 are sized as discussed below. Dielectric fluid fills the region 34 on the side of electrode 28 opposed to chamber 30. The dielectric fluid is preferably air or other dielectric gas, although a dielectric liquid may be used.

[0028] Typically, electrode 28 is made of a somewhat flexible conductive material such as polysilicon, or, in the preferred embodiment, a combination of layers having a central conductive layer surrounded by an upper and lower insulative layer. For example a preferred electrode 28 comprises a thin film of polysilicon stacked between two thin films of silicon nitride, each film for example, being one micron thick. In the latter case, the nitride acts to stiffen the polysilicon film and to insulate it from liquid in the chamber 30. However, due to a coupler, described below, it is not necessary that the polysilicon film be made stiffer, since the electrode may be moved in either direction solely by electrostatic attractive forces.

[0029] A second electrode 36 between chamber 30 and a lower cavity 37 is preferably identical in composition to electrode 28, and is electrically addressable separately from electrode 28. Addressable electrodes 28 and 36 are preferably at least partially flexible and are positioned on opposite sides of a single central electrode 38 such that the three electrodes are generally axially aligned with nozzle orifice 22. Since there is no need for addressable electrode 36 to completely seal with wall 26, its peripheral region may by mere tabs tethering the central region of electrode 36 to wall 26.

[0030] Central electrode 38 is preferably made from a conductive central body surrounded by a thin insulator of uniform thickness, for example silicon oxide or silicon nitride, and is rigidly attached to walls 26. In a preferred embodiment, the central electrode is curved on one side, shown as the top side in FIG. 2, and is flat on the opposing side, shown as the bottom side in FIG. 2, and is in contact with addressable electrode 36 along its lower surface at walls 26. That is, the upper surface of central electrode 38 is concave away from addressable electrode 28, but the lower surface of central electrode 38 is planar and may be in contact with addressable electrode 36 along its entirety. The lower side of central electrode 38 is flat and addressable electrode 36 contacts the central electrode at its periphery along sidewall 26 in order to insure that the shape of addressable electrode 36, when in a position away from central electrode 36 (FIG. 6), is determined entirely by the materials properties of addressable electrode 36 and the length that rigid coupler 40 extends below the lower surface of central electrode 38. In this way, the position of addressable electrode 36, when extended downward, as in FIG. 6, will be very nearly identical for all ejectors on a single print head and for ejectors from print head to print head. The force exerted by addressable electrode 36 to expel drops during the drop expulsion portion of operation, as described later, will be nearly identical for all ejectors, irrespective of the exact shape of the curved portion of central electrode 38. As is well known in the art of semiconductor manufacture, a flat surface is more precisely and reliably obtained than a curved surface and films, such as the thin films forming addressable electrode 36, are deposited more consistently and are better understood when deposited on a flat substrate. Thereby the drops from all ejectors will be expelled with nearly identical velocities.

[0031] Additionally, due to the flat bottom surface of central electrode 38, addressable electrode 36 has a surface area that is a minimum when the addressable electrode contacts the lower surface of central electrode (FIG. 7). The surface area increases when addressable electrode 36 is pushed away from the central electrode (FIG. 6). Thereby, addressable electrode 36 is assured to contact completely the central electrode during operation, since the portion of addressable electrode 36 last to contact the central electrode will be in a state of lesser tension than if the central electrode were concave, as can be appreciated by one skilled in the theory of elastic deformation. This is opposite to addressable electrode 28 in FIG. 6, which is under its greatest tensile stress while contacting (or attempting to contact) the entire upper side of the central electrode since the surface area of addressable electrode 28 is a maximum when it contacts central electrode 38. Addressable electrode 28 may not fully contact central electrode 38 unless the voltage differential between them is very large, as shown in FIG. 6, whereas addressable electrode 36 will always contact central electrode 38, even for small voltage differentials between them. Thus, during the drop expulsion portion of operation, as described later, both addressable electrodes will be exerting a force to increase the pressure in ink cavity 30 because of their elastic properties as well as the voltage differential between the addressable electrode 36 and central electrode 38.

[0032] The two addressable electrodes are structurally connected via a rigid coupler 40. This coupler is electrically insulating, which term is intended to include a coupler of conductive material but having a non-conductive break therein. Coupler 40 ties the two addressable electrodes structurally together and insolates the electrodes so as to make possible distinct voltages on the two. The coupler may be made from conformally deposited silicon dioxide.

[0033] FIGS. 3-5 are top plan views of nozzle plate 24, showing several alternative embodiments of layout patterns for the several nozzle orifices 22 of a print head. Note that in FIGS. 3 and 4, the interior surface of walls 26 are annular, while in FIG. 5, walls 26 form rectangular chambers.

[0034] Referring to FIG. 6, to eject a drop, an electrostatic charge is applied to the polysilicon portion of addressable electrode 28 nearest to nozzle orifice 22 and the conductive portion of central electrode 38. The voltage of the conductive body of central electrode 38 and of the polysilicon portion of addressable electrode 36 are kept at the same. As shown in FIG. 6, addressable electrode 28 is attracted to central electrode 38 until it is deformed to substantially the surface shape of the central electrode, except in the region very near the central opening in the central electrode. In so conforming its shape, addressable electrode 28 presses down on addressable electrode 36 through rigid coupler 40, thereby deforming addressable electrode 36 downward, as shown in FIG. 6, and storing elastic potential energy in the system. Since addressable electrode 28 forms a wall portion of liquid chamber 30 behind the nozzle orifice, movement of electrode 28 away from nozzle plate 24 expands the chamber, drawing liquid into the expanding chamber through ports 32. Addressable electrode 36 does not receive an electrostatic charge, that is, its voltage is the same as electrode 38, and moves in conjunction with addressable electrode 28.

[0035] The angle of contact between the lower surface of addressable electrode 28 and the upper surface of central electrode 38 is preferably less than 10 degrees. In a preferred embodiment, this angle tends to 0 degrees at the point of contact between the lower surface of addressable electrode 28 and the upper surface of central electrode 38. This ensures the voltage difference required to pull addressable electrode 28 down into contact with central electrode 38 is small compared with the value that would be required if the angle were larger than 10 degrees. For example, for the shape of central electrode 38 shown in FIG. 6, the voltage required is typically less than half that required for the case in which the angle of contact between the lower surface of addressable electrode 28 and the upper surface of central electrode 38 is 90 degrees, as can be appreciated by one skilled in the art of electrostatic actuators.

[0036] Subsequently (say, several microseconds later) addressable electrode 28 is de-energized, that is, its voltage is the same as electrode 38, and addressable electrode 36 is energized, causing addressable electrode 36 to be pulled toward central electrode 38 in conjunction with the release of the stored elastic potential energy. The timing of the de-energization of electrode 28 and the energization of electrode 36 may be simultaneous, or there may be a short dwell period therebetween so that the structure begins to move from the position illustrated in FIG. 6 toward the position illustrated in FIG. 7 under the sole force of stored elastic potential energy in the system. Still referring to FIG. 7, this action pressurizes the liquid in chamber 30 behind the nozzle orifice, causing a drop to be ejected from the nozzle orifice. To optimize both refill and drop ejection, ports 32 should be properly sized to present sufficiently low flow resistance so that filling of chamber 30 is not significantly impeded when electrode 28 is energized, and yet present sufficiently high resistance to the back flow of liquid through the port during drop ejection.

[0037] The lower surface of central electrode 38 is planar, reducing the dependence of the displaced liquid volume during the ejection stroke on fabrication parameters, and allowing addressable electrode 28 to be planar at the peak of ejection height. In comparison with a symmetric central electrode having two concave surfaces, fabrication is simpler and less subject to process variations. Further, the onset of the ejection stroke is more precisely controlled.

[0038] FIG. 8 is a top view of a portion of drop ejection mechanism 20 of FIG. 2 formed according to a preferred embodiment of the present invention. In this and the following figures, the structure continues to be illustrated in schematic form, but in somewhat more detail than in the previous figures.

[0039] Still referring to FIG. 8, during operation, electrical signals are sent via electrical leads 42 to the three electrodes 28, 36 and 38 of FIG. 2. The three-layer electrode structure is anchored to outer wall 26 by structural supports 44. Rigid coupler 40 connects electrodes 28 and 36 of the three-layer electrode structure. A flow restrictor 46 (see also FIG. 22) prevents fluid from returning from liquid chamber 30 to the fluid reservoir (not visible here) via a fluid conduit 48 during drop ejection. A second fluid path 50 shown in FIG. 21 allows the dielectric fluid in region 37 to flow into and out of a dielectric fluid reservoir (not shown). In the preferred embodiment, the dielectric fluid is air, and the ambient atmosphere performs the function of a dielectric fluid reservoir.

[0040] A line A-A′ in FIG. 8 indicates the plane of the cross-sections depicted in FIGS. 9-22, which illustrate a single drop ejector of many which would normally be batch fabricated simultaneously.

[0041] FIG. 9 shows a substrate 52 of, say, a 550 μm thick, single crystal silicon wafer, for example. The substrate will be used to support the electrode structure and to form fluid conduits 48 that bring the fluid to nozzle orifice 22, and the second fluid paths 50 that bring the dielectric fluid to region 37.

[0042] FIG. 10 shows the preferred embodiment after deposition, patterning, and etching of a first structural layer 54 (e.g. 0.75 μm thick doped polysilicon) and a first passivation layer 56 formed for example of 0.1 μm low pressure chemical vapor deposition (LPCVD) silicon nitride. These two layers are patterned using photolithography and etched away to form a depression that will allow addressable electrode 36 to deform toward substrate 52 during pullback. First passivation layer 56 insulates addressable electrode 36 from first structural layer 54 and substrate 52, which may both be formed of conductive materials.

[0043] In FIG. 11, conformal deposition and planarization by chemical mechanical polishing (CMP) of an initial sacrificial layer 58 has occurred. The sacrificial layer may be, for example, 0.85 μm plasma enhanced chemical vapor deposition (PECVD) silicon dioxide, filling in the depression formed during the previous etch and providing a planar surface for the deposition of addressable electrode 36 as shown in FIG. 12. Addressable electrode 36 may be 3 μm to 5 μm doped polysilicon, and is relatively thick for a microdevice because it is advantageous to have a mechanically stiff electrode that will not deform, so that energy transfer from addressable electrode 36 to addressable electrode 28 through rigid coupler 40 is maximized when the addressable electrode 36 is energized to eject a drop. Although not shown in this figure, there are numerous perforations around the perimeter of the moving portion of addressable electrode 36 allowing it to move more easily. This reduces the energy required to pull the piston back to its “loaded” position.

[0044] FIG. 13 shows the preferred embodiment after deposition, patterning, and etching of a subsequent sacrificial layer 60 (e.g. 0.1 μm silicon dioxide). This thin layer provides mechanical separation between addressable electrode 36 and central electrode 38 shown in FIG. 15. Where subsequent sacrificial layer 60 is eliminated, the layers above will be attached to the layers below. The hole etched in the center will allow addressable electrode 36 and addressable electrode 28 can be mechanically coupled. The hole is preferably etched in the center, but could be etched elsewhere.

[0045] FIG. 14 shows the preferred embodiment after deposition, patterning, and etching of a second passivation layer 62 (e.g. 0.1 μm LPCVD silicon nitride). This layer provides electrical separation between addressable electrode 36 and central electrode 38, FIG. 15. LPCVD nitride is preferable to PECVD nitride in this layer, since the breakdown voltage of LPCVD nitride is higher, allowing a larger voltage to be supported without current leakage for the same layer thickness.

[0046] FIG. 15 shows the preferred embodiment after deposition, patterning, and etching of second electrode layer 38 (e.g. 5 μm doped polysilicon). This layer is non-uniform, increasing in thickness radially from the center of the device. This may be accomplished by one of the following well-known manufacturing techniques:

[0047] 1. Laser ablation (high cost, no advantage of batch processing).

[0048] 2. Making a 3-D mold with a release layer and perform a pattern transfer (high one-time expense but high accuracy). Re-usable if a proper release layer is used.

[0049] 3. Metal sputtering with a reusable shadow mask.

[0050] 4. Partial exposure of resist followed by an etch.

[0051] 5. Multiple exposures for differing lengths of time all aligned to the same point, causing resist to be underexposed at some points and properly exposed at others.

[0052] 6. Dithering of features on the mask to allow undercutting to occur during a subsequent isotropic etch.

[0053] 7. Blowing jets of air to form depressions at stagnation points in flow (works for a drying liquid or a curing polymer).

[0054] 8. Selective spatial exposure (shadow mask) of photoresist to an acetone vapor to cause variable degree of exposure based on the same light intensity.

[0055] 9. Using chemical mechanical polishing (CMP) to cause dishing by patterning a protective coating layer at high points and leaving low points exposed. Subsequent removal of the protective layer by etching.

[0056] 10. Reflowing a conductive conformal coating.

[0057] 11. Curing a conductive liquid drop.

[0058] FIG. 16 shows the preferred embodiment after deposition, planarization (e.g. CMP), patterning, and etching of a third sacrificial layer 66 (e.g. 0.55 μm silicon dioxide). This layer provides mechanical separation between second electrode layer 38 and third electrode layer 28. This step is provided for re-planarizing the system for deposition of third electrode layer 28.

[0059] FIG. 17 shows the preferred embodiment after deposition, planarization (e.g. CMP), patterning, and etching of a third passivation layer 64 (e.g. 0.12 μm silicon nitride). This layer mechanically couples first electrode layer 36 and third electrode layer 28, while insulating them from one another. This can be done in several ways. The method pictured is a thin insulating layer with its thickness determined by the breakdown voltage of the dielectric, followed by deposition of some other filler material as a second structural layer 40 (conductive or non-conductive) that is less expensive to deposit and planarize (e.g. spin-on polymer). Alternatively, a solid block of third passivation layer 64 can be employed. This would avoid the second deposition, but it requires a thick deposition and planarization down to a thin layer with some accuracy. Another alternative is to leave the center hollow, and allow the third electrode layer to partially fill it. This has the advantage of a less costly process, as well as a structurally weaker spacer, since third passivation layer 64 must be kept thin to minimize the voltage required to operate the device. In addition, the third electrode layer 28 would become non-planar due to the dip at the center of third passivation layer 64.

[0060] In FIG. 18, addressable electrode 28 (e.g. 2.5 μm doped polysilicon) has been deposited, patterned and etched. FIG. 19 shows the preferred embodiment after deposition, patterning, and etching of a third sacrificial layer 70 (e.g. 5 μm polyimide or silicon dioxide). This layer provides separation between addressable electrode 28 and nozzle plate 24 (FIG. 20) through which a drop will be ejected. The third sacrificial layer 70 will be eliminated later to form the liquid chamber 30. This layer is etched twice; once to provide a dimple that will create flow restrictor 46 (FIG. 8), and once to expose addressable electrode 28 for mechanical attachment.

[0061] In FIG. 20, nozzle plate 24 of, for example, 4 μm nitride or polyimide (if not used for the third sacrificial layer) has been deposited, patterned and etched. The hole in this layer forms nozzle orifice 22 through which the drop is ejected. FIG. 21 shows the preferred embodiment after substrate 52 is etched from the back side (the side not previously patterned), opening holes to first passivation layer 56 and first sacrificial layer 58, which act as etch stops during this process.

[0062] FIG. 22 shows the preferred embodiment after all sacrificial layers are removed (e.g. by immersion in HF to remove silicon dioxide sacrificial layers and/or by oxygen plasma to eliminate polyimide sacrificial layers). This is the completed device. Central electrode 38 is provided with external power through the lead 42 in this cross-section. FIG. 23 shows a cross-section through B-B′ of the preferred embodiment in its finished state. The difference between this and the previous figure is the electrode structure on the left side, where addressable electrode 36 is provided with external power through lead 42 in this cross-section. FIG. 24 shows a cross-section through C-C′ of the preferred embodiment in its finished state. The difference between this and the previous figure is the electrode structure on the left side, where addressable electrode 28 is provided with external power through lead 42 in this cross-section. FIG. 25 shows a cross-section through D-D′ of the preferred embodiment in its finished state. The difference between this and the previous figure is that the region shown does not intersect any of the lead structure. This represents the region through which the fluid flows freely from the fluid conduit to the ejection chamber.

[0063] FIGS. 26-39 are all cross sections through the line A-A′ in FIG. 8. FIG. 26 shows a substrate 52 such as a 550 μm thick single crystal silicon wafer for example. The substrate in this case will be used to support the electrode structure 28, 38, 36 and to form the fluid conduits 48 that bring the fluid to the nozzle.

[0064] FIG. 27 shows the preferred embodiment after deposition, patterning, and etching of the first structural layer 54 (e.g. 0.75 micron thick doped polysilicon) and the first passivation layer 56 (e.g. 0.1 μm LPCVD (low pressure chemical vapor deposition) silicon nitride). These two layers are patterned using photolithography and etched away to form a depression that will allow first electrode layer 36 to deform toward substrate 52 during pullback. First passivation layer 56 insulates first structural layer 54 and substrate 52, which may both be conductive materials, from first electrode layer 36.

[0065] FIG. 28 shows the preferred embodiment after conformal deposition and planarization (chemical mechanical polishing (CMP)) of first sacrificial layer 58 (e.g. 0.85 μm PECVD (plasma enhanced chemical vapor deposition) silicon dioxide), filling in the depression formed during the previous etch and providing a planar surface for the deposition of first electrode layer 36.

[0066] FIG. 29 shows the preferred embodiment after deposition, patterning, and etching of first electrode layer 36 (e.g. 3-5 μm doped polysilicon). First electrode layer 36 is relatively thick for a microdevice because it is advantageous to have a mechanically stiff electrode that will provide an elastic force in addition to the electrostatic attractive force that will eject a drop. Although not shown in this figure, there are numerous perforations around the perimeter of the moving portion of first electrode layer 36 allowing it to move more easily (see FIG. 18). This reduces the energy required to pull the piston back to its “loaded” position.

[0067] FIG. 30 shows the preferred embodiment after deposition, patterning, and etching of a second sacrificial layer 60 (e.g. 0.1 μm silicon dioxide). This thin layer provides mechanical separation between first electrode layer 36 and second electrode layer 38. Where second sacrificial layer 60 is eliminated, the layers above will be attached to the layers below.

[0068] FIG. 31 shows the preferred embodiment after deposition, patterning, and etching of a third sacrificial layer 72. This layer is non-uniform, decreasing in thickness radially from the center of the device. This is accomplished by one of the following methods:

[0069] 1. Curing a liquid drop. This is easier to process if a photopatternable polymer such as SU8 is used.

[0070] 2. Reflowing a conformal coating.

[0071] 3. Sputtering with a reusable shadow mask.

[0072] 4. Laser ablation.

[0073] 5. Making a 3-D Mold with a release layer and perform a pattern transfer.

[0074] 6. Partial exposure of resist followed by an etch.

[0075] 7. Multiple exposures for differing lengths of time all aligned to the same point, causing resist to be underexposed at some points and properly exposed at others.

[0076] 8. Dithering of features on the mask to allow undercutting to occur during a subsequent isotropic etch.

[0077] 9. Blowing jets of air to form depressions at stagnation points in flow.

[0078] 10. Pushing on an elastomer and locking it into place (by heating, for example).

[0079] 11. Selective spatial exposure (shadow mask) of photoresist to an acetone vapor to cause variable degree of exposure based on the same light intensity.

[0080] 12. Using chemical mechanical polishing (CMP) to cause dishing by patterning a protective coating layer at high points and leaving low points exposed. Followed by subsequent removal of the protective layer by etching.

[0081] FIG. 32 shows the preferred embodiment after deposition, patterning, and etching of a second passivation layer 62 (e.g. 0.1 μm LPCVD silicon nitride). This layer provides electrical separation between first electrode layer 36 and second electrode layer 38. LPCVD nitride is preferable to PECVD nitride in this layer, since the breakdown voltage of LPCVD nitride is higher, allowing a larger voltage to be supported for the same layer thickness. The hole etched in the center (preferred embodiment, but the hole could be etched elsewhere) will allow second sacrificial layer 60 below to be etched in subsequent steps, so that first electrode layer 36 and third electrode layer 28 can be mechanically coupled.

[0082] FIG. 33 shows the preferred embodiment after deposition, planarization, patterning, and etching of second electrode layer 38 (e.g. 5 μm doped polysilicon) and a third passivation layer 64 (e.g. 0.1 μm LPCVD silicon nitride).

[0083] FIG. 34 shows the preferred embodiment after deposition, patterning, and etching of a third sacrificial layer 66 (e.g. 0.55 μm silicon dioxide). This layer provides mechanical separation between second electrode layer 38 and third electrode layer 28. The patterning of the third sacrificial layer also removes part of the second sacrificial layer and exposes part of the first electrode.

[0084] FIG. 35 shows the preferred embodiment after deposition, planarization (e.g. CMP), patterning, and etching of a fourth passivation layer (e.g. 5 μm silicon nitride). This layer mechanically couples first electrode layer 36 and third electrode layer 28, while insulating them from one another. This can be done in several ways. The method pictured a solid block of the fourth passivation layer 40. This requires a deposition, planarization, patterning, and etch. Another method is a thin insulating layer with its thickness determined by the breakdown voltage of the dielectric, followed by deposition of some other filler material second structural layer (conductive or non-conductive) that is less expensive to deposit and planarize (e.g. spin-on polymer).

[0085] FIG. 36 shows the preferred embodiment after deposition, patterning, and etching of third electrode layer 28 (e.g. 2.5 μm doped polysilicon).

[0086] FIG. 37 shows the preferred embodiment after deposition, patterning, and etching of a fourth sacrificial layer 70 (e.g. 5 μm polyimide or silicon dioxide). This layer provides the separation between third electrode layer 28 and membrane layer 24 through which a drop will be ejected. This layer is etched twice; once to provide a dimple that will create flow restrictor 46, and once to expose third electrode layer 28 for mechanical attachment. For certain layer thickness combinations, it may be necessary to planarize before this step using deposition and CMP of a sacrificial material. Otherwise, the fluid conduit may be occluded where there is no lead structure or structural support.

[0087] FIG. 38 shows the preferred embodiment after deposition, patterning, and etching of membrane layer 24 (e.g. 4 μm nitride or polyimide if not used for the fourth sacrificial layer). The hole in this layer is nozzle 22 through which the drop is ejected.

[0088] FIG. 39 shows the preferred embodiment after the substrate 52 is etched from the back side (the side not previously patterned), opening holes to the first passivation layer 56 and first sacrificial layer 58, which act as etch stops during this process.

[0089] FIG. 40 shows the preferred embodiment after all sacrificial layers 58, 60, 66, 70 are removed (e.g. by immersion in HF to remove silicon dioxide sacrificial layers and/or by oxygen plasma to eliminate polyimide sacrificial layers). This is the completed device. The second electrode layer 38 is provided with external power through the lead 42 in this cross-section.

[0090] FIG. 41 shows a cross-section through B-B′ of the preferred embodiment in its finished state. The difference between this and the previous figure is the electrode structure on the left side, where the first electrode layer 36 is provided with external power through the lead 42 in this cross-section.

[0091] FIG. 42 shows a cross-section through C-C′ of the preferred embodiment in its finished state. The difference between this and the previous figure is the electrode structure on the left side, where the third electrode layer 28 is provided with external power through the lead 42 in this cross-section.

[0092] FIG. 43 shows a cross-section through D-D′ of the preferred embodiment in its finished state. The difference between this and the previous figure is that the region shown does not intersect any of the lead structure. This represents the region through which the fluid flows freely from the fluid conduit to the ejection chamber.

Claims

1. A method of making a multi-layer micro-electromechanical electrostatic actuator for producing drop-on-demand liquid emission devices, said method comprising: forming an initial patterned layer of sacrificial material on a substrate; depositing and patterning, at a position opposed to the substrate, a first electrode layer on the initial layer of sacrificial material; forming a subsequent patterned layer of sacrificial material on the first electrode layer such that a region of the first electrode layer is exposed through the subsequent layer of sacrificial material; depositing and patterning, at a position opposed to the first electrode layer, a second patterned electrode layer on subsequent layer of sacrificial material, said second electrode layer gradually varying in thickness; forming a third patterned layer of sacrificial material on the second electrode layer, said third patterned layer of sacrificial material having an opening there through to the exposed region of the first electrode layer; depositing and patterning a structure on the third layer of sacrificial material to a depth so as to at least fill the opening through the third layer of sacrificial material; planarizing structure to expose a surface of the third layer of sacrificial material; depositing and patterning a third electrode layer on planarized structure and the exposed surface of the third layer of sacrificial material, whereby the first electrode layer and the third electrode layer are attached by the structure; and removing sacrificial material from the initial layer, the subsequent layer, and the third layer, whereby the first electrode layer, the structure, and the third electrode layer are free to move together relative to the second electrode layer.

2. A method as set forth in claim 1, wherein the region of the first electrode layer is exposed through the subsequent layer of sacrificial material by etching through the subsequent layer of sacrificial material.

3. A method as set forth in claim 1, wherein the initial sacrificial layer is formed by conformal deposition and planarization by chemical mechanical polishing of a sacrificial material.

4. A method as set forth in claim 1, wherein the opening through the third layer of sacrificial material to the exposed region of the first electrode layer is formed by etching.