GB2295441A

GB2295441A - Microminature fluid flow device

Info

Publication number: GB2295441A
Application number: GB9520719A
Authority: GB
Inventors: Philip W Barth; Christopher C Beatty
Original assignee: Hewlett Packard Co
Current assignee: HP Inc
Priority date: 1994-10-20
Filing date: 1995-10-10
Publication date: 1996-05-29
Anticipated expiration: 2015-10-10
Also published as: GB9520719D0; JPH08210519A; GB2295441B; DE19530843A1

Description

-- 1 1- N1ICRON1INIAMRE FLUID FLOW DEVICE 2295441 The present invention

relates to a rnicrominiature device for controlling the flow of a fluid.

The development of microminiature mechanical devices has advanced generally by use of a technique known as micromachining or microfabrication. See for instance, the discussion of microfabrication of mechanical devices by Angell et al. in "Silicon Micromechanical Devices," Scientific American, (April 1983). pp. 44-55.

A fundamental requirement of a micromechanical actuator (hereinafter, microactuator) is that some mechanical actuation means must be provided. A further requirement is that the actuation means must provide sufficient force for reliable actuation. For example, a microactuator may comprise part of a microvalve used to control the flow of a carrier gas through a capillary column in a gas chromatograph. The microactuator may be required to open or cAose a fluid passage by displacing a moveable member (typically a moveable membrane, diaphragm, or boss) against a pressure of 1375 kilopascals (200 pounds 4:5 per square inch) through a distance of as much as 100 micrometers.

Typically, electrical power from an external source is provided to the microactuator, which employs one of various techniques to convert the applied power to an actuating force. Often the applied electrical power is converted in part or whole to thermal power, and such microactuators can be considered as being thermally-driven.

An array of micromachined bi-metallic legs has been employed to provide a thermallydriven actuating force in a microminiature valve. As the bimetallic legs are heated, stresses are generated in the structure to deflect a protruding boss adjacent to an orifice, increasing or decreasing the flow of fluid to an attached fluid-bearing system. For example, and with reference to Figure t A which is reproduced from commonly-assigned US Pat- No. 5,333,831.

issued to Barth et al.. a microminiature valve 10 is shown as including a seat substrate 12, which acts as a base, and an upper substrate 13. A central flow via 14 is formed through the seat substrate 12. Supported atop the seat substrate 12 in the upper substrate 13 are a fixed periphery 16, a central boss 18 formed preferably in silicon, and legs 20, 22. The length and the width of the upper substrate 13 match the dimensions of the seat substrate 12.

The structure and the operation of the upper substrate 13 are described in commonlyassigned U.S. Pat. No. 5,058,856 to Gordon et al. Briefly, a layer of nickel is deposited and patterned on the upper substrate 13 using the techniques of sputtering, photolithography, and electroplating. The array of legs 20 and 22 join the fixed periphery 16 to the boss 18. A resisfive layer of nickel is present on the upper surface of the boss, and can be used to heat the boss and legs. When the upper substrate 13 is heated, the difference in coefficients of thermal expansion of the silicon and the nickel causes the legs 20, 22 to arch, lifting the boss 18 away from a valve seat 28 extending upwardly from an upper major surface 30 in the seat substrate 12. When the boss 18 is spaced apart from the seat substrate 12, the flow via 14 is in fluid communicabon with a surrounding volume 24. In turn, this volume 24 is in fluid communication with an apparatus to or from which flow is to be regulated by the microminiature valve 10. (Alternatively, there may be actuation by means other than arching legs.) The valve seat 28 includes a bearing surface 32 against which the boss IS is seated when the boss is in the cJosed position.

The performance of the microminiature valve 10 is determined in large part by its thermal resistance characterisfics. For example, if the microminiature valve 10 is normally closed when no power is applied, and if the thermal resistance from the actuator to its surroundings is low, the microminiature valve 10 will require a relatively large amount of power to open, but will cool rapidly when power is removed and so will close rapidly. If the thermal resistance from the microminiature valve 10 to its surroundings is high, the microminiature valve 10 will require less power for to open, but will cool more slowly, and so will be slower to close.

Thermal power flows from the upper substrate 13 through several paths and in several modes. When the microminiature valve 10 is closed, the boss 18 is in contact with the valve seat 28. Thermal power then flows by solid phase conduction from the boss 18 through the valve seat 28 and into the bulk of the seat substrate 12. Thermal power also flows by solid phase conduction from the actuator legs 20, 22 to the fixed periphery 16 and hence to the seat substrate 12. Gas phase conduction of thermal power occurs from the lower surface of the upper substrate 13 to the seat substrate 12 and also from the upper surface of the upper substrate 13 to any surrounding package (not shown). The thermal power flowing into the seat substrate 12 flows further into any thermally-conductive package structure that is contiguous with the seat substrate 12 such as a structure that is in contact with the bottom surface 38.

The present invention seeks to provide an improved microminiature device.

According to an aspect of the present invention, there is provided a microminiature device for controlling the flow of a fluid, comprising a seat substrate including opposed first and second major surfaces and a flow via extending from said first major surface to said second major surface, an integral annular wall structure including a valve seat and extending from said first major surface and surrounding the flow via, and a recess located adjacent said integral annular wall structure in a surround in said first major surface; and an upper substrate positioned adjacent the seat substrate and including a valve face positionable in a closed position with respect to the valve seat to obstruct fluid flow through said flow via and in an open position with respect to the valve seat to permit fluid flow through said flow via; said recess including a sufficient combination of surface area and depth. as compared respectively to the overall upstream surface area and overall thickness of the seat substrate, such that gas phase conduction between the upper substrate and the seat substrate is reduced.

It is possible to minimize the power consumed by a microminiature device, especially in a microminiature valve where fast actuation is not critical, by attending to the power lost by way of solid phase conduction or gas phase conduction from the upper substrate to the lower substrate.

A preferred device for controlling the flow of a fluid includes a seat substrate having opposed first and second major 1 surfaces and a flow via extending from said first major surface to said second major surface, an integral annular wall structure extending from said first major surface, the annular wall structure surrounding the flow via and including a valve seat, and a recess located in a surround in said first major surface, the recess being adjacent to said integral annular wall structure. An upper substrate is positioned adjacent the seat substrate and includes a valve face positionable in a closed position with respect to the valve seat to obstruct fluid now through said flow via and in an open position with respect to the valve seat to permit fluid flow through said flow via. The recess includes a sufficient combination of surface area and depth, as compared respectively to the overall upward-facing surface area and overall thickness of the seat substrate, such that the gas phase conduction between the upper substrate and the seat substrate is reduced. The annular wall structure may also be constructed in order to minimize solid phase conduction from the upper substrate to the seat substrate.

A preferred method of forming a microminiature valve is one in which an orifice in the central flow via and the valve seat are self-aligned so that the fabrication process is not susceptible to overetching and underetching. The anisotropic etching provides a constant valve seat geometry regardless of etching time. The duration of the etching affects the depth of the valve seat, but after formation of parallel (111) oriented walls, the cross sectional configuration of the valve seat is generally fixed.

The preferred fabrication of a microminiature valve includes steps for masking selected portions of both the upper and lower major surfaces of the seat substrate.

A first region is left exposed at the lower major surface. Etching the exposed first region forms a via either partially or entirely through the substrate. Preferably, the etching process employs an orientation dependent etchant that forms (111) oriented walls in the etched via. The front surface is patterned to define masked areas which are subsequently etched to form the recess and the annular wall structure. Orien ta tionde pendent etching from the front surface also further shapes the via, creating {1 11) walls that define a central flow via which is smallest in area at the orifice and whose cross-sectional area increases with increasing distance from the orifice.

An embodiment of the present invention is described below. by way of example only, with reference to the accompanying drawings, in which:

Figure I A is a side sectional view of a prior art microminiature valve;

Figure I B is a side sectional view of an embodiment of microminiature valve having a flow orifice and a valve seat, Figure I C is a detailed side sectional view of a portion of the microminiature valve of Figure IB; Figures 2A to 5B illustrate steps of fabricating one embodiment of the valve seat and flow orifice of the microminiature valve of Figure 113; Figures 6A and 6B are detailed side perspective, sectional views of a portion of the valve seat and flow orifice of the microminiature valve fabricated using the steps of Figures 2A-5B; Figure 7A is a side sectional view of a second embodiment of microminiature valve; Figure 7B is a detailed side sectional view of a portion of the valve of Figure 7A; Figure 8A is a side sectional view of a third embodiment of microminiature valve, Figure 8A; Figure 8B is a detailed side sectional view of a portion of the valve of Figure 9A is a side sectional view of a fourth embodiment of microminiature valve; Figure 9B is a detailed side sectional view of a portion of the valve of Figure 9A; Figure 10 is a plan view of the orifice area of the seat substrate fabricated using the steps of Figures 2A-5B, with a preferred seat etching mask superimposed for the purposes of illustration; and Figure I I is a plan view of a comer compensation mask preferred for use controlling the etch of the exterior portion of the throat wall fabricated using the steps of Figures 2A-5B.

It is to be understood that ranges and other values given herein may be extended or otherwise modified without losina the effects sou2ht, as will be apparent to the skilled person.

Whereas the following description is directed to a microactuator in the form of a microminiature valve, it is contemplated that the teachings herein may find application in other types of thermally-driven microdevices that operate at an elevated temperature with minimal power consumption. This characterization of microdevices as being 'thermallydriven' is meant to include those that operate on the conversion of an applied power into an actuation force for moving a movable member, wherein the conversion benefits from conservation or isolation of the thermal energy that may arise in the course of the conversion. Examples are microactuators that are driven by forces developed in a process of gas or liquid expansion/contraction, gas or liquid phase change, or according to changes in bi-metallic or

1 6 - shape-memory materials. Accordingly, the teachings will find use in a variety of microactuators that may be employed to operate upon a mechanical device or system, or upon a physical phenomena. such as the flow of fluids (including gases and liquids), electtrical and electronic parameters (such as capacitance, current flow, and voltage potential), acoustical and optical parameters (such as reflection, absorpt,ion, or diffraction) and simple dimensional parameters (such as acceleration, pressure, length, depth, and so on).

With reference to Figure 18, a first preferred embodiment of microactuator in the form of a microminiature valve 110 includes seat substrate 112 which acts as a base. The seat substrate 112 is preferably a silicon orifice chip which has been fabricated from a wafer using batch processing steps as will be described below with respect to Figures 2A to 5B. A central flow via 114 is formed through the seat substrate 112. (The term 'via' is used herein to describe a fine through-hole in a fabricated layer.) Supported atop the seat substrate 112 is an upper substrate 113 also formed from silicon that includes a fixed periphery 116 and a thermally- actuated member in the form of a central boss 118. The length and the width of the upper substrate 113 roughly match the respective dimensions of the seat substrate 112.

A thermally-driven actuator is preferably provided in the upper substrate 113, preferably in the form of an array of bi-metallic elements wherein one portion of the bi-metallic element has a higher coefficient of thermal expansion than the remaining portion, so that a temperature change in the bi-metallic element causes motion. The terms bi-metala and bi-metallico are not limited to their conventional sense; for example, one or both portion within the bi-metallic element may actually be non- metallic. Preferably, in the illustrated embodiment, one portion within the bi-metal element is nickel, and the other portion within the bi-metal element is silicon, which is a semiconductor.

Bi-metallic elements in the form of legs 120, 122 include a layer of nickel on a layer of silicon, wherein the thickness of the nickel and silicon layers may each be, for example, 30 micrometers. Legs 120 and 122 join the fixed periphery 116 to the central boss 118. A resistive heater (not shown) is located adjacent the upper surface of the upper substrate 113 in close proximity to legs 120 and 122. When the upper substrate 113 is heated by inducing an electrical current in the resistive heater, the difference in coefficients of thermal expansion of the silicon and the nickel causes the legs 120, 122 to arch, lifting the boss 118 away from the valve seat 128. When the boss 118 is spaced apart from the seat substrate 112, the flow via 114 provides fluid communication between the throat 132 and surrounding volume 124. In turn, the surrounding volume 124 is in fluid communication with an apparatus (not shown) to or from which flow is to be regulated by the microminiature valve 110. The preferred mode of thermally-driven actuation and other aspects of the structure and operation of the upper substrate 113 are disclosed in commonly-a s signed U.S. Pat. No. 5,058, 856 to Gordon et al. and commonly-assigned U.S. Patent 5,333,831 to Barth et aL, the disclosures of which are incorporated herein by reference.

While the microminiature valve 110 will be described as including an array of legs 120 and 122, it - is not limited to actuation by means of arching legs. For example, a structure that connects the central boss 118 to the fixed periphery 116 may instead be provided as a solid circular diaphragm which is selectively deflected to regulate fluid flow between the flow via 114 and the surrounding volume 124.

The seat substrate 112 includes an annular wall structure 119 preferably in the form of a hollow, truncated pyramid. For the purposes of this description, the term gannularo is meant to include polygonal as well as circular or conical formations. The annular wall structure 119 includes an orifice 127 circumscribed by a valve seat 128. The valve face 118A is seated against the valve seat 128 when the central boss 118 is in the closed position, The valve seat 128 is formed atop a seat wall 129 which extends from a throat wall 131. The annular wall structure 119 is surrounded by a surround 112C that defines a recess 130A formed in an upper major surface 130 of the seat substrate 112. As will be described more fully below, the annular wall structure 119 and recess 130A are preferably formed by orientation-dependent etching of the seat substrate 112 at the upper major surface 130 and lower major surface 138. The width of the valve seat 128 may be varied, but is chosen to be sufficiently great that the valve seat is not susceptible to fracturing upon repeated contact between the valve seat 128 and the valve face 11 8A.

The particular configuration of the recess 130A, throat wall 131, valve seat 128, and the orifice 127 in the seat substrate 112 offer improved pneumatic and thermal characteristics over the prior art. As previously noted, a goal in the design of a thermally actuated valve 110 is to minimize wasted thermal power. Typically, power flows from the upper substrate 113 in several ways: by thermal conductance through the valve seat 128; by thermal conductance paths through legs 120, 122; by convecfive heafing of the gas in the surrounding volume 124 and of the gas above the upper substrate 113; by gas phase conduction from the upper surface of upper substrate 113; by gas phase conduction from the upper surface of upper substrate 113 to any overlying package members; and by gas phase conductive flow through the gas in the surrounding volume 124.

The seat substrate 112 is constructed to include recess 130A so as to decrease the gas phase conduction between the upper substrate 113 and the seat substrate 112. Preferably, the seat substrate 112 is constructed to so as to provide a sufficient combination of surface area and depth of the recess 130A, as compared respectively to the overall upward-facing surface area and overall thickness 1 12D of the seat substrate 112, such that the gas phase conduction between the upper substrate 113 and the seat substrate 112 is reduced in comparison to the prior art. It is also preferred that the annular wall structure 119 be constructed to exhibit a vertical height sufficient to decrease the gas phase conduction between the upper substrate 113 and the seat substrate 112. It is further preferred that the.surround 112C be made sufficiently thick to provide adequate overall mechanical strength in seat substrate 112 while maximizing the depth of'recess 130A.

The teachings herein concerning the annular wall structure 119 are also directed to minimizing the overall thermal conductance between the upper substrate 113 and the seat substrate 112. For example, the seat wall 129 is made sufficiently thin without compromising robust mechanical operation, and the valve seat 128 is made sufficiently narrow, in order to minimize solid phase thermal conductance from the valve face 118A to the seat substrate 112. Further, the throat wall 131 is made sufficiently thin to minimize solid phase thermal conduction from the seat wall 129 to the seat substrate 112. As a result, the power consumption to achieve a given displacement of the boss 118 at a given temperature of the legs 120, 122 is greatly reduced.

In a first preferred embodiment, the recess 130A encompasses a surface area of at least ten times the cross-sectional area circumscribed by the valve seat 128; the depth of the recess I 30A is made substantially equal to the vertical height of the annular wall structure 119. In a second preferred embodiment of the seat substrate 112, the annular wall structure 119 has a vertical height that is greater than a range of two to five times the minimum lateral thickness of the throat wall 131. In a third preferred embodiment of the seat substrate 112, the surface area of the recess 130A is preferably between 25 to 95 percent of the overall upwardfacing surface area of the seat substrate 113; the depth of the recess 130A is preferably between 25 to 95 percent of the overall thickness 112 D of the seat substrate 112; or the surround 1 12C has a thickness selected from a range of approximately 5 to 75 per cent of the overall thickness 112D of the seat substrate 112.

In a fourth preferred embodiment of the seat substrate 112, the overall thickness 112D is less than 500 micrometers, the surround 112C has a thickness less than 300 micrometers, the recess 130A has a depth greater than 200 micrometers, or the flow orifice 114 has a cross sectional area of less than 360,000 square micrometers.

The fabrication of the seat substrate 112 in an odfice chip will now be described in general, and the specific steps for fabrication of the seat substrate 112 will be described With respect to Figures 2A-513, below. Further details on the fabrication of the actuator substrate 113 are disclosed in commonly-as signed U.S. Pat. No. 5,058,856 to Gordon and Barth.

Figures 2A-513 illustrate a first preferred method of fabdcating the seat substrate 112 of Figure 1 B. The fabrication process is preferably carried out in a batch mode wherein plural seat substrates 112 are fabricated simultaneously in one silicon wafer, and many silicon wafers are processed simultaneously in a cassette. The fabrication process comprises several operations to mask and etch thin films on the wafer surfaces plus three silicon etching operations.

The fabrication process begins at Figure 2A. A silicon wafer 260 is coated on both top and bottom with respective first and second layers of silicon nitride 262, 264 by low-pressure chemical vapor deposition (LPCVD). Preferenfially, the first and second silicon nitride layers - to- 262, 264 are low-stress layers achieved by silicon-rich deposition methods. Subsequently, a chrome layer 266 is deposited by sputter deposition on the first silicon nitride layer 262.

Using a double-sided photomask aligner such as a Karl Suss MA-25 aligner, photoresist patterns are simultaneously defined on the chrome layer 266 and the second silicon nitride layer 264. A central region 269 is defined in photoresist and is then etched through the second silicon nitride layer 264 using conventional plasma etching techniques to expose a portion of the silicon wafer 260 which will later be etched to form cavity 279A. The photoresist pattern on the chrome layer 266 is etched using wet etching techniques known for fabricating chrome photomasking patterns on glass. An inner annulus 272 defines an area that will later be the valve seat 128, and an outer annulus 274 defines the area which will later be the upper face of a shoulder region at the periphery of the seat substrate 112. For the purposes of this description, the terms "annularg and mannulus" are meant to include polygonal as well as circular formations. In particular, the preferred embodiment of the valve seat 128 (when viewed from above) defines a rectilinear form; however, other forms are contemplated

In Figure 2B, aqueous potassium hydroxide (KOH) is used in a first silicon etching operation to etch into the bottom of the silicon wafer 260 at the central region 269. Preferably, silicon wafer 260 is odented with respect to the silicon crystal lattice such that a <I 00> crystalline direction is perpendicular to the major surface of the silicon wafer 260. In such an orientation of the silicon wafer 260, etching by aqueous KOH proceeds rapidly in the < 1 00> direction perpendicular to the wafer surface, and etching proceeds much more slowly along the <111> planes that are oriented at 54.7 degrees away from the <100> direction. Asaresult, the etching substantially ends at the (11 1} crystal planes, thereby providing a cavity 279A with a flat roof and with four sloped throat walls comprising {1 1 1} crystal planes. In Figure 213, first and second throat walls 278, 280 of the four sloped walls are visible. Each of the first and second throat walls 278, 280 are oriented at an angle 54.7 degrees from the horizontal plane of the silicon wafer 260.

Preferably, the depth of cavity 279A obtained at this step should be greater than the thickness of the desired surround region 112C, so that the result of later etching step(s) will form a via through the silicon wafer 260. Also preferably, the extent of the central region 269 should be small enough that the projection of the first and second throat walls 278, 280 falls completely within the confines of inner annulus 272. Although the cavity 279A may at this point in the fabrication process be etched completely through silicon wafer 260 to stop at silicon nitride layer 262, it is preferred that in order to save fabrication time, the etching process is stopped when the depth of the cavity 279A is greater than the desired thickness of the surround region 112C. For example, if surround region 112C is desired to be 250 micrometers thick, the depth of cavity 279A at this point in the fabrication process may be made 300 micrometers, providing a comfortable margin of 50 micrometers. Also, at this point in the fabrication process, the cavity 279A should not be etched completely to the first silicon nitride layer 262 because a subsequent photomasking operation must be performed, and the first silicon nitride layer 262 may not possess sufficient strength to survive mechanical handling during the subsequent photomasking process.

Next, a second photomasking operation is carried out. Plasma etching is performed on the first silicon nitride layer 262 to leave an outer annular region 270 of silicon nitride and an inner annular region 276 of silicon nitride. A second silicon etching operation is conducted in aqueous KOH, resulting in the features shown in Figures 3A and 3B. The second silicon etching operation etches the exposed portion of silicon wafer 260 in a region situated between annular region 270 and annular region 276, creating an annular cavity with recess face 282. The second silicon etching operation also deepens cavity 279A. The second silicon etching operation produces throat wall 281. In a preferred embodiment, the second silicon operation etches 725 micrometers deep into the top surface of a silicon wafer 260, and etches the roof of the cavity 279A to create a central flow via 279B extending completely through the silicon wafer 260.

Preferably, as shown in Figures 3A and 38, the inner annular region 276 allows silicon etching in a central via from the top side of the silicon wafer 260. However, an annulus is not strictly necessary and the central via in the inner annular region 276 serves as a fabricabon convenience to ease tolerance in fabrication. Tolerances are eased in the second silicon etching operabon because, as the downward-moving (100)-oriented etching plane meets the - II - upward-moving (100)-oriented etching plane, the original extent of the width of central region 269 may be reduced in a manner calculated according to the crystal directions and known principles of orientation- dependent etching. Further, it is not strictly necessary to form a complete via through the silicon wafer 260 during the second silicon etching operation. However, the fabrication of a complete via allows one to visually ensure that the top side and the bottom side masking patterns are correctly aligned to one another. Accordingly, the second silicon etching operation preferably provides a central flow via 279B to be formed entirely through the silicon wafer 260.

Next, the inner annulus 272 and outer annulus 274 are employed as masking regions in a plasma etching operation which etches the exposed silicon nitride in the inner annular region 276 and outer annular region 270. A inner annular subregion 276A of silicon nitride, which replicates the lateral shape of annulus 272, and an outer annular subregion 270A which replicates the lateral shape of annular region 274, are formed. The annular subregion 276A when viewed from above defines a particular shape illustrated in Figure 10 as feature 702 and which is discussed below. Silicon regions now exposed are etched in aqueous KOH in a third silicon etching operation, such as the recess face 282 and first, second, third, and fourth opposed recess walls 294A, 294B, 295A, 295B as illustrated in Figures 4A and 4B. Preferably, the etching of flat (100) surfaces in the <100> direction that occurs in the third silicon etching operation is lessthan the etching of similar surfaces in the first and second etching operations; for example, the recess face 282 will typically be etched an additional 25 micrometers during the third silicon etching operation.

Referring to Figures 4A - 4B, the third silicon etching operation occurring at the upper major surface results in four adjoined walls with (111) oriented inner and outer lateral wall surfaces and whose combined upper surface is the same shape as that of the valve seat. First and second sloped seat inner walls 296, 298 are illustrated in cross section in Figures 4A and 48. First and second substantially vertical walls 290 and 292 connect the inner surfaces of first and second sloped seat inner walls 296, 298 to the previously-formed first and second throat inner walls 278, 280. First and second vertical walls 290, 292 are substantially composed of (11 0} oriented crystal planes and the etch rate in the < 11 0> crystal direction in aqueous KOH is typically approximately 1.9 times as great as the vertical etch rate of (100) planes in the <100> direction. Accordingly, the first and second vertical walls 290 and 292 are extended downward and outward, thus maintaining substantially the same vertical dimension while their lateral dimensions are increased, the horizontal upper surfaces of the throat wall 281 is also etched downward in the <100> crystal direction. However, the location and slope of the first and second sloped seat inner walls 296, 298 is established at an early time in the third silicon etching operation and remains substantially unchanged during lateral etching of the first and second vertical walls 290, 292 and the vertical etching of the horizontal upper surfaces of the throat wall 281.

As shown in Figure 5A, and after the third silicon etching operation, the second silicon nitride layer 264 is removed from the bottom of the silicon wafer 260 by plasma etching. Next, the upper surface of the silicon wafer 260 is exposed to a plasma etch in order to etch any portion of the inner annular subregion 276A beneath the inner annulus 272 which was exposed by undercutting of the inner annular subregion 276A during the third silicon etching operation. The lateral extent of the remaining portion of inner annular subregion 276A becomes substantially coextensive with the second and fourth opposed recess walls 294B, 295B. Thus, a self-alignment characteristic is obtained between the etched silicon features at the top of the first and second sloped seat inner walls 296, 298 and the edges of inner annular subregion 276A, whereby the edges of inner annular subregion 276A are laterally registered with sub- micrometer accuracy to the edges of the upper surfaces of the first and second sloped seat inner walls 296, 298.

As a final step in the batch fabrication process, wet chemical etching is used to remove any remaining &rome regions from the silicon wafer 260. The result is a silicon wafer 260 containing a plurality of individual valve orifices and seats. The silicon wafer 260 can then be separated into individual orifice chips in a dicing step using a conventional high- speed dicing saw. The final dicing step can generate dust which may later impair the operation of the orifice chips. Thus, in cases where certain considerations warrant a modified fabrication process, the dicing step may be performed before the third silicon etching operation, and the individual silicon chips obtained after dicing may then be exposed to the third silicon etch, the two plasma etches, and the chrome etch as described hereinabove, with the exception that the individual chips are placed in fixtures which facilitate such individualized handling. The modified fabrication process can in some cases be beneficial because the third silicon etching operation tends to dissolve any microscopic silicon dust that remains on the orifice chips after the dicing step.

Figure 6A - 6B are cutaway perspective views of the orifice area of the completed seat substrate 112 in an individual orifice chip.

Figures 7A and 78, 8A and 8B, 9A and 9B illustrate respective second, third, fourth preferred embodiments of the seat substrate, formed in accordance with the foregoing steps, but with minor modifications, as set forth below.

In Figures 7A-7B, a second seat substrate 412 includes a second central flow via 414, second valve seat 428, a second seat wall 429, and a second throat wall 431. This embodiment is accomplished by controlling the abovedescribed first silicon etching operation to etch a depth from the bottom side equal to the thickness desired for the portion of the seat substrate 412 that surrounds the second throat wall 431. Then, during the third silicon etching operation, etching is stopped when the second central flow via 414 extends completely through the wafer. Over-etching would result in lateral undercutting or destruction of the second valve seat 428.

In Figures 8A-8B, a third seat substrate 512 includes a third central flow via 514, a third valve seat 528, a third seat wall 529, and third throat wall 531. This embodiment is accomplished using the first and second silicon etching operations, and may be simpler to perform.

In Figures 9A-9B, a fourth seat substrate 612 includes a fourth central flow via 614, a fourth valve seat 628, a fourth seat wall 629, and a fourth throat wall 631. This embodiment is conceptually similar to the embodiments illustrated in Figures 8A and 8B, except that extended contact portions 61 8A, 61 8B have been added to lower face of a central boss 618.

Figure 10 shows a seat etching mask 702 that is preferred for controlling the etch of the various valve seats described herein. The seat etching mask 702 is superimposed over the orifice area of the seat substrate 112 for the purposes of illustrating their relative alignment. The mask consists of a rectilinear frame with features oriented along the < 11 0> directions on the (100) planar surface of the silicon wafer 260, and with diagonal tabs at its comers oriented in <100> directions on the (1001 planar surface. The purpose of the diagonal tabs is to ensure that the (111) planes that will subsequently be formed at the exterior edges of the frame will meet at sharp comers without undercutting beneath the rectilinear features, so as to achieve improved comer shapes.

Figure 11 illustrates a photolithographic comer compensation mask 802 preferred for controlling the etch of the valve seat supporting structure described herein. The purpose of this comer compensation mask is to ensure that the very deeply etched (111) planes which form the outer surface of the annular wall structure 119 will form appropriate comers. Further details of etching convex comers in (100)-silicon in aqueous KOH are described by Mayer et al., in "Fabrication of Non-Underetched Convex Comers in Anisotropic Etching of (100)-Silicon in Aqueous KOH with Respect To Novel Micromechanic Elements", J Electrochem. Soc., Vol. 137, No. 12, p. 3947-3951.

Modifications in the structure of the disclosed embodiments may be effected by use of differing patterns in the etch-resistant coatings. In addition, alternative coatings, such as silicon dioxide or polyimide, are contemplated as being deposited or grown on the surface of the completed structure. Furthermore, while the disclosed embodiments have been described as being fabricated from a silicon substrate, other materials such as metal, glass, ceramic, or polymers, and other semiconductor or crystalline substrates such as gallium arsenide, may also be used. For example, the structures described herein may be fabricated according to one or more of the following alternatives: borosilicate glass may be fabricated using ultrasonic machining; photosensitive glass may be formed by lithography; a ceramic material may be ultrasonically machined or may be cast and fired; a metal or machinable ceramic may be formed by conventional machining; or a polymer may be machined, cast, or injection molded.

]be disclosures in United States patent application no. 08/326,454, from which this application claims priority, and in the abstract accompanying this application are incorporated herein by reference.

16

Claims

1. A microminiature device for controlling the flow of a fluid, comprising a seat substrate including opposed first and second major surfaces and a flow via extending from said first major surface to said second major surface, an integral annular wall structure including a valve seat and extending from said first major surface and surrounding the flow via, and a recess located adjacent said integral annular wall structure in a surround in said first major surface; and an upper substrate positioned adjacent the seat substrate and including a valve face positionable in a closed position with respect to the valve seat to obstruct fluid flow through said flow via and in an open position with respect to the valve seat to permit fluid flow through said flow via; said recess including a sufficient combination of surface area and depth, as compared respectively to the overall upstream surface area and overall thickness of the seat substrate, such that gas phase conduction between the upper substrate and the seat substrate is reduced.

2. A microminiature device as in claim 1, wherein the surround is sufficiently thick to provide adequate overall mechanical strength in seat substrate while maximizing the recess depth.

3. A microminiature device as in claim I or 2, wherein the annular wall structure exhibits a vertical height sufficient to decrease the gas phase conduction between the upper substrate and the seat substrate.

4. A microminiature device as in claim 1, 2 or 3, wherein the annular wall structure comprises a throat wall sufficiently thin to minimize solid phase thermal conductance from the valve face to the seat substrate.

5. A microminiature device as in any preceding claim. wherein the annular wall structure comprises a seat wall for supporting the valve seat and wherein the C, recess encompasses a surface area at least ten times the cross-sectional area 17 circumscribed by the valve seat.

6. A microminiature device as in any preceding claim, wherein the depth of the recess is substantially equal to the vertical height of the annular wall structure.

7. A microminiature device as in any preceding claim, wherein the surface area of the recess is between 25 to 95 percent of the overall upstream surface area of the seat substrate.

8. A microminiature device as in any preceding claim, wherein the depth of the recess is between 25 to 95 percent of the overall thickness of the seat substrate.

9. A microminiature device as in any preceding claim, wherein the surround has a thickness approximately 5 to 75 percent of the overall thickness of the seat substrate.

10. A microminiature device as in any preceding claim, wherein the overall thickness of the seat substrate is greater than 500 micrometers and the surround has a thickness of less than 300 micrometers.

11. A microminiature device as in any preceding claim, wherein the recess has a depth greater than 200 micrometers.

12. A microminiature device as in any preceding claim, wherein the annular wall structure comprises a throat wall, and the annular wall structure has a vertical height that is greater than two to five times the minimum lateral thickness of the C throat wall.

13. A microminiature device as in any preceding claim, wherein the flow via is substantially centrally located in said annular wall structure and said flow via increases in cross-sectional area with distance from said valve seat.

18

14. A microminiature device as in any preceding claim, wherein said seat substrate comprises a semiconductor material.

is. A microminiature device as in claim 14, wherein said annular wall structure is formed by (111) oriented crystal planes.

16. A microminiature device as in any preceding claim, wherein said seat substrate comprises a material selected from the following group of materials: metal, glass, ceramic, and polymers.

17. A microminiature device as in any preceding claim, wherein said annular wall structure is a truncated pyramidal structure.

18. A microminiature device as in claim 17, wherein said flow via is formed by a four-sided throat wall.

19. A microminiature device as in any preceding claim, wherein the upper substrate comprises a thermal actuator operable for positioning the valve face in a selected one of the open and closed positions.

20. A microminiature device as in claim 19, wherein the thermal actuator comprises first and second material layers having respectively different coefficients of thermal expansion.

21. A micron-liniature device substantially as hereinbefore described with reference to and as illustrated in Figures 1 B, 1 C, 2A to 6B. 10 and 11; Figures 7A and 7B; Figures 8A and 8B; or Figures 9A and 9B of the accompanying drawings.