CA1334798C - Micro-machined accelerometer - Google Patents

Abstract

A spring-mass-support structure and associated electrical circuitry for measuring acceleration applied to the support of the structure is disclosed. According to one aspect of the invention, the spring-mass-support structure of the transducer is fabricated using micro-machining techniques of a silicon wafer. The structure includes E-shaped leaf springs of silicon dioxide which suspend a mass from a support structure.
The springs as well as the support structure and the suspended mass, are formed by chemical etching through openings of opposite faces of a silicon wafer on which etch stop layer patterns are diffused. Sense and force conductive patterns are diffused onto the opposite faces of the suspended mass. The spring-mass-support structure is then sandwiched between opposite plates having corresponding sense and force conductive patterns which face such patterns on the suspended mass. According to another aspect of the invention, electrical circuitry is provided by which a sense voltage and a force voltage are applied between opposite sense and force conductive patterns of the opposite plates creating a sense electric field and a force electric field across the sense and force patterns of the mass. A feedback circuit arrangement is provided to substantially maintain the mass at a predetermined reference position between the support plates. Forward circuitry of the feedback arrangement produces a displacement signal in response to a voltage induced on the sense area of the mass as it moves in response to acceleration applied to the support structure.
Feedback circuitry of the feedback arrangement produces an amount of charge in response to the displacement signal which is applied to the force conducting area of the mass.
The force on said mass is proportional to the product of that charge and the force electric field intensity which causes the mass to remain close to a reference position between the plates. Alternative circuitry for providing analog or binary bit stream representations of acceleration is disclosed.

Description

BACKGROUND OF THE L~V~N 1 ION
MICRO-MACHINED ACCELEROMETER
Field of the Invention This invention relates generally to transducers which may be fabricated from semi-conductor material such as silicon by batch photolithographic techniques from a monocrystalline wafer. More particularly, the invention relates to a spring-mass-support structure for detecting a signal proportional to acceleration applied to the support structure. The invention also relates to electronic circuitr~ in combination with the structure for detecting displacement of the spring supported mass to produce analog or binary representations of acceleration applied to the support structure.
Description of the Prior Art A publication by Kurt E. Petersen entitled "Silicon as a Mechanical Material" published in Vol. 70, No. 5 (May, 1982) of the Proceedings of the IEEE, surve~s the art of micro-machining of silicon to fabricate a mechanical device. The Petersen publication describes etching techniques commonly used in fabricating micromechanical devices and describes several devices which have been made using such techniques. The publication also describes a cantilever beam light modulator arrav the beams of which are fahricated of silicon dioxide. Another survey article appeared in the April, 1983 publication of Scientific American entitled, ~' Silicon Micromechanical Devices by Angell, et al.
An accelerometer having a micromachined cantilever beam intended for in vivo biomedial studies and which includes a piezoresistive sensing element is described in a publication entitled, "A Batch-Fabricated Silicon Accelerometer" by Roylance and Angell in IEEE Transactions on Electron Devices, Vol. ED-26, No. 12, December, 1979.
rJ.s. patent Re.31,459 (reissue of U.S. patent No.
4,071,838) reissued December 6, 1983 in the name of Barrv n Block discloses a silicon folded leaf spring force transducer fabricated by batch photolitho~raphic and etchinq techniques. One embodiment of such transducer consists of four E-springs located at 90 angles about the central axis of sensitivity. Piezoresistors disposed on legs of the E-springs provide a measure of the displacement of the springs in response to a force applied to the transducer.
U.S. patent 4,155,516 issued ~arch 13, 1979, in the name of Harry Aine discloses first and second leaf spring ~o structures, as supported from support structures, which are coupled together in axiall~ spaced mutually opposed relation so that the irst and second leaf spring structures are coupled together for deflection in unison relative to the surrounding support structures in response to deflection of either of the first and second sprinq structures. A capacitive detector structure is provided for detecting displacement of the leaf spring.

~ T.S. patent 4,597,003 issued June 24, 1986, in the names of Aine and Rlock discloses a method of chemical etching a silicon wafer by undercutting an etch stopped layer to fabricate a cantilever beam supported at one end by a frame structure with an integral mass formed by the etch. Etching from opposite etch stopped layers of a silicon wafer is described.
While the art described above has advanced the art of micro-machining of silicon wafers to fabricate sensors in general and accelerometers in particular, there remains a need for a micro-machined structure of silicon and other materials and associated electronics meeting extremely precise specifications. For example, extremely low distortion for an accelerometer is required for certain applications. Prior art spring configurations and materials are believed to produce spring constants which change unacceptably over a deflection range of interest.
Very low sensitivity to accelerations orthogonal to the sensitivity axis is desired. Prior art spring mass ~n configurations do not provide sufficient insensitivity to orthogonal acceleration and angular motion applied to such configurations. Very high dynamic range (requiring extremelv low noise) is desired in an accelerometer which has not been provided by prior art acceleromet~rs and associated detecting electronics systems. Shock impact insensitivity is required in an accelerometer intended for use in certain applications such as in the seismic, automotive, or aeronautics fields.
Identification of Objects of the Invention Accordingly, the invention seeks to provide a micro-machined spring-mass sensor which has springs having spring constants which can be precisely controlled during the fabrication process.
Further the invention seeks to provide a geometry of a spring-mass-support transducing element which results in an accelerometer which has very low distortion and has a very low sensitivity to motion orthogonal to its sensitivity axis and to angular motion.

Further still the invention seeks to provlde an accelerometer sensor which when subjected to extremely high shock impacts will survive undamaged.
The invention in another aspect seeks to provide a micro-machined sensing device and associated electronics resulting in an accelerometer with an extremely broad dynamic range.
The invention in a further aspect seeks to provide a batch fabricated micro-machined sensing element with associated electronics for producing a signal representative of acceleration applied along the sensitivity axis of the element.

In a still further aspect, the invention seeks to provide a batch fabricated micro-machined sensing element with associated electronics to produce a binary bit stream representation of acceleration applied along the sensitivity axis of the element.
The invention in one broad aspect provides an improved spring structure in a micro-machined transducer having a support frame and a sensing mass, where the sensing mass is at least partially supported from the support frame via at least one E-shaped leaf spring structure. The spring structure is defined by a base and three legs including two outer legs and an inner leg, each of the three legs being connected to the base at one end, the two outer legs '`';`' being connected to the frame at their other ends and the inner leg being connected to the mass at its other end. The improvement comprises the base having a coefficient of stiffness substantially greater than that of the legs.
A further broad aspect of the invention provides a transducer comprising, sensing means for generating a displacement signal representative of displacement of a mass from a reference position with respect to a support structure, forward circuit means responsive to the displacement signal for converting the displacement signal to an output binary bit stream and feedback circuit means responsive to the output binary bit stream for generating a binary force on the mass in a direction tending to restore the mass to the reference position, whereby the output binary bit stream is representative of a characteristic of motion of the support structure.
Still further the invention pertains to a method for making a transducer including a spring mass system by chemical etching of a semiconductor wafer, including the step of diffusing etch stop layer patterns on opposite major faces of a semiconductor wafer forming a first intermediate wafer where the patterned etch stopped layers define a support frame structure, a sensing mass structure and an E-shaped leaf spring structure, the spring structure defined by a base and three legs including two outer legs and an inner leg with each of the three legs being connected to the base at one end and the two outer legs being connected to the frame at their other ends, the inner leg being connected to the mass structure at its other end.
Additional etch stop material is diffused into the support frame structures and the sensing mass structures and the bases of the E-shaped leaf spring structure of the first intermediate wafer, to form a second intermediate wafer. Additional silicon dioxide is formed on each of the three legs of the spring structures of the second intermediate wafer to form a third intermediate wafer and a semiconductor etchant is introduced via openings on opposite faces of the wafer where no etch stop layer is diffused to undercut substantial portions of the etch stop layers including the E-shaped leaf spring structure. The etch of semiconductor material is terminated so as to leave undercut structure in both of the etch stopped layers interconnected to each other by interconnecting substrate semi-conductor material and etch stop material is removed from beneath the silicon dioxide layers of the three legs of the undercut E-shaped leaf spring structure.
More particularly, the sensing structure of the invention includes a micro-machined transducer having a support frame and a sensing mass supported from the frame by at least one, preferably eight, E-shaped leaf springs. Each of the eight E-springs have a base with two outer legs and an inner leg. Each leg is connected to the base at one of its ends. The two outer legs connect to the frame at their other ends. The inner leg is connected to the mass at its other end. The base is characterized by a coefficient of stiffness substantially greater than that of the legs. Each of the legs are preferably the same length. In order to precisely define the spring constant of the springs, the legs are fabricated of silicon dioxide, because the thickness of the silicon dioxide springs may be precisely controlled during the fabrication process.
The mass-spring-support sensing structure is fabricated by orienting four E-shaped leaf structure patterns on top and bottom (100) faces of a silicon wafer such that they are oriented at forty-five degree angles with respect to intersection lines of (111) planes of the wafer and the opposite faces of the wafer. The opposite faces are masked and silicon dioxide grown on surfaces which will ultimately be etched away by an etching chemical. A thin laver which is a stop layer is formed. This layer can be formed by diffusing into this layer an impurity, such as arsenic, that will change the semiconductor type from P to N in this layer for an impurity such as arsenic which will form a P/N junction between the laver and the substrate. A voltage bias can then be applied to this laver to form an electromechanical etch stop layer. The arsenic is diffused into opposite - surfaces of the wafer everywhere except the "opening"
surfaces and the surfaces which will ultimately serve as the legs of the sprinq.
A thick oxide layer is then grown bY diffusing a precise amount of oxygen into the silicon to form a precise thickness of silicon dioxide on the silicon wafer.
Next, the opposite surfaces of the wafer are masked exposinq only the regions which will not be the springs of the structure as well as some other selected areas.
These exposed areas of silicon dioxide are then etched ~0 down to the silicon surfaces. A thin silicon dioxide la~er is then grown, masked, and the exposed silicon dioxide etched down to the silicon surface in the regions that will be exposed to the silicon etchant. Chemical etch of the silicon material then is applied which acts through the openings of the silicon dioxide to etch silicon material awav while undercutting the springs, leaving a central mass supported bY silicon dioxide spring legs. Stiff etched-stop laYers remain for the bases of the springs and the opposite faces of the suspended~mass and support structures.
Force and sense conducting areas are then deposited on opposite faces of the suspended mass. Conductors are deposited which rtln from the conducting areas via spring legs to the support structure. First and second plates having corresponding force and sense plates with conductors running to external connecting pads are fixed to opposite support surfaces of the mass-spring-support structure such that each of the conducting surfaces of the mass precisely face a corresponding conducting surface of the first and second plates.
An electrostatic electronic feedback system is provided to detect displacement of the mass of the mass-spring-support structure in response to acceleration or force applied to the support. For frequencies of the applied acceleration below the resonant frequency of the mass-spring-support structure, displacement of the mass is ~n proportional to acceleration. Displacement of the mass (and consequently the acceleration) applied to the support structure is measured bv applying a first voltage between the sense conducting areas of the opposite plates and applying a second voltage between the force conducting areas of the oppo.site plates. The first voltage creates a sense e]ectric field between the sense conductinq areas of the suspended mass and induces a sense voltage on it - -lo- 1 334798 proportional to the displacement of the mass between the opposite plates. The second voltage creates a ~force electric field between the force conducting areas of the mass.
A forward circuit responsive to the sense voltage generates a displacement signal which is representative of the displacement that the mass has moved from a reference position between the plates. A feedback circuit responsive to the displacement signal applies an amount of electric charge on the force conducting area of the mass surface such that a force is created on the mass proportional to the product of the amount of charge and - the force electric field so as to substantially restore the mass to the reference position. The displacement signal is an analoq representation of acceleration. The first and second voltages may be d.c. voltages, or may be square waves operating at a modulation frequency higher than expected frequencies of the acceleration to be measured. A demodulator is required to produce an ~n unmodulated analog output signal representative of acceleration.
Another embodiment of the invention, like the ones described above, includes a sensor for generating a displacement signal representative of displacement of a mass from a reference position with respect to a support structure. A forward circuit is provided which, in response to the displacement signal, converts the displacement signal to an output binary stream. A
feedback circuit, in response to such output binary stream, generates a hinary force on the mass in a direction tending to restore the mass to the reference position. Consequentlv, the output binary stream is representative of a characteristic of motion, preferably acceleration, of the support structure.

BRIEF DESCRIPTI~N OF THE DRAWINGS
The objects, advantages and features of the invention will become more apparent bv reference to the drawings which are appended hereto and wherein like numerals indicate like parts and wherein an illustrative embodiment of the invent,ion is shown, of which Figure 1 is a transverse sectional view of a wafer of monocrYstalline material from which a sprinq mass system is to be fabricated according to micro-machining techniques of the invention;
Figure 2 is a view similar to that of Figure 1 showing an oxide layer formed on opposite sides of the wafer;
Fiqure 3 is a plan view of the wafer of Figure 2 showing areas for which spring-mass-structures will be fabricated;
Figure 4A is an enlarged view of an area of Figure 3 as indicated hy line 4A of Figure 3 which indicates the orientation of the crvstal top and bottom surfaces and its -12- l 334798 relation to (111) planes of the crystal;
Figure 4B is a sectional view of the wafer showing a part of the layout taken along the lines 4B-4B of Figure 4A with photo-resist material having been applied to the opposite sides of the wafer;
Figures 5A and 5B are similar to Figures 4A and 4B
but show a wafer portion after it has been masked with a pattern of a spring-mass-support structure, and after portions of the oxide layers have been removed, leaving a pattern of oxide partially defining ultimate etching openings on the oppos.ite sides of the wafer;
Figures 6A and 6B are similar to Figures 5A and 5B
but ref]ect further processing after etch stop material has been d.ifused into the crvstal of the wafer on opposite surfaces in areas other than the etching openings and spring legs;
Figures 7A and 7B are si.milar to Figures 6A and 6B
but reflect further processing after opposite wafer surfaces have been masked to allow further etch stop layer ~0 material onto wafer surfaces where surrounding support structure, central mass and bases of E-springs are patterned;
Figures 8A and 8B are similar to Figures 7A and 7B
but reflect further processing after the etching openings on both sides of the wafer have had oxide previously covering them removed and show a relativelv thick layer of silicon dioxide material having heen qrown over areas of ~ -13- l 334798 the sides of the wafer defining E-spring legs;
Figure 9A shows a plan view of the one-half o,f the wafer after chemical etching of the wafer through etch openings has etched to (111) planes of the crystal of the wafer thereby undercutting the E-spring structures and severing the mass portion of the structure from the support portion while suspending the mass from the support and after further processing where the thin layer of etch stop material beneath oxide siAes of the springs have been n removed;
Figure 9~ is a sectional view of the wafer illustrated in Figure 9A taken along section lines 9~-9~;
Fiaure 1~ is a perspective view of the sen.sing element structure including support, E-springs and sensing mass, the view being partiallv cut-away to illustrate the central mass severed from the support structure to show the oxide springs and undercut etch stop layers of the mass, springs and support structure;
Figure 11 illustrates a cross-section through the sensing element showing the sensing mass supported by springs to ~undercut portions of the support member where gravity forces the mass to be disposed beneath the top plane of the support member, the view further showing a composite structure including top and bottom plates having force and sense conducting areas deposited on their faces correspondinq to similar force and sense conducting areas on opposite surfaces of the sensing mass;

Figure 1~ is a plan view of the sensinq structure illustrating deposited conducting leads to the forc~ and sense conducting areas of the mass surface from the support structure via the E-sprinq leads;
Figure 13 is a schematic illustration of an electrostatic feedback control system .for detecting acceleration applied to the support system by substantially maintaining the sensing mass at a predetermined position between the opposite plates of the composite structure, the output of the system being an analog signal proportional to applied acceleration;
Figure 14 shows a schematic illustration of a system similar to that of Figure 13 but having an alternating voltage applied to the covering plate force and sense conducting areas, with a demodulator provided to develop an analog signal proportional to applied acceleration;
Figure 15 is a schematic illustration of a charge generator circuit used directly or with modification in the transducer systems of Figures 13, 14 and 16, 17, 18 ~0 and 19;
Figure 16 is a schematic illustration of a transducer having a binary bit stream output representative of the displacement or acceleration of a mass in a mass-spring-support structure with binary bit stream voltages applied to t.op and bottom force plates and with a constant charge applied to force plates of the mass with the result that a binary force is applied to the mass -15- l 334798 tending to restore it to its reference position;
Figure l6A is an illustration of a binarv sampler used in the embodiments of Figures 16, 17, 18 and 19;
Figure 17 is similar to the transducer of Figure 16 except that a binary hit stream of charge is applied to the force plates of the mass and a constant electric field is applied~across the force plates of the top and bottom ~ plates;
- Figure 18 is similar to the transducer of Figure 17 lnexcept that an electric field applied to the force plates of the top and bottom plates is modulated with an a.c.
siqnal, and consequentlv the binarv bit stream of charge applied to the force plates of the mass is also modulated;
and Fiqure 19 is similar to the transducer of Fiqure 18 except that the charge applied to the force plates of the mass is modulated and modulated bit streams of the output signal are fed back to the force plates of the top and bottom plates.

?n DESCRIPTION OF THE INVENTION
Two important motivating factors, among others, led to the accelerometer of this invention. The first is low manufacturing cost so that the resulting accelerometer may find wide application where manv force sensing transducers are required. The second is an accelerometer having extremelv hiqh performance characteristics: wide dynamic ranqe; low distortion; low sensitivity to motion orthogonal its sensitivity axis; and resistance to high shock impact. These requirements lead to the fabrication of the accelerometer by micro-machining techniques from a monocrystalline material such as silicon.
Referring now to Figure 1, a typical wafer 10 is shown from which a batch of accelerometers are to be fabricated according to the process of the invention. In a typical example, the wafer 10 is made of a nonmetallic ln monocrystalline material such as silicon, germanium, quartz, gallium arsenide, gallium phosphide, etc. In a preferred embodiment, the wafer 10 is made of a diamond cubic material, such as ~silicon. The wafer 10 has a thickness of about 500 microns, has polished top and hottom surfaces 12, 14 and has a convenient diameter, such as 3 to 8 inches. In the case of diamond cubic material, the crystallographic plane (100) is preferably formed at the upper 12 and lower 14 major surfaces of the wafer 10.
Furthermore the wafer 10, in the case of silicon, is preferably doped with a P type dopant to a conductivity of between 0.1 - 10.0 ohm-cm.
In the next step of the process, the wafer 10 is oxidized on opposite sides to form oxidized layers 16, 18 of 2500 angstroms in thickness as illustrated in Figure 2.
This is conveniently achieved by putting the wafers in a furnace at 1100 C, in the presence of oxvqen. Next, the oxide lavers 16, 18 are each coated with a photoresist material such as KMER.
Figuxe 3 is a plan view of the wafer with mul~tiple areas each of which are to be processed simultaneously to batch produce the mass-sprinq-support structures of the invention. One~particular area is illustrated in Figure 3 with line 4A-4A, shown much enlarged in Figure 4A. The boundary edges 24 of the structure to be fabricated are oriented parallel to the (111~ planes of the crystal as illustrated in Figure 4A. Rach structure of the array of structures of Figure 3 will ultimatelv be cut along the boundary edges 24 parallel to the (111) planes of the crystal after batch processing. Figure 4~ shows a cross-section through a portion of the wafer showing the KMER photoresist layers 20, 22 deposited over silicon dioxide laYers 16, 18.
Fiqure 5A shows one side of the wafer (for convenience called here the top side) after masking and exposure of the pattern as depicted thereon. The unexposed photoresist portion covers the etch opening patterns 30' and the spring legs 30'', but the remainder of the photoresist portions are removed by exposure and development of the photoresist. The wafer is then placed in a silicon oxide etch solution such as buffered hydrofluoric acid, which attacks the oxide but not the photoresist or the underlying silicon. The photoresist protects the pattern 30 as illustrated in Figure 5A. The pattern 30 includes two parts: the regions in~icated by -18- l 334798 30' will ultimately be etch openings; the regions marked 30'' will ultimately be spring legs. Fiqure 5B shQws a cross-section through lines 5B-5B of Fiqure 5A and - illustrates the oxide portions 30i and 30'' on top and bottom surfaces 12, 14 which respectively ultimately form a slot opening behind the base of the E-spring and the middle leg of the E-spring. ~xide portions 30' and 30'' are greatly enlarged in Figure 5B for ease of illustration.
ln The oxide pattern 30' of Figures 5A and 5B will ultimatelv serve as the etch openings to fabricate the three dimensional mass-spring-support structure of the accelerometer. The oxide patterns of 30'' will ultimately be the location of the sprinq legs of the E-sPrings. A
perspective view of the structure may be seen in cut-awav sectional view of Figure 10 showing the structure at a certain stage in the processing.
Figure 5A illustrates particular features of the grown oxide patterns 30' and 30'' which ultimately will ~0 result in the structure of Figure 10. The lines 32-3~ of Figure 5B show a region in which an "E"-shaped spring 40 will be fabricated to partially support a mass 36 from a surrounding suDport structure 38. The E-spring 40 will include a base 42 and three legs 44, 46, 48. Slots or openings 43, 45, 47 will ultimately outline the shape of the E-spring. The inner leg 46 is twice the width of outer legs 44, 48. The lengths of the legs 44, 46, 48 are precisely the same.
Rach E-shaped spring 40 is placed at the cornçr of intersecting (111) planes, that is, at 45 angles to intersecting lines of (111~ planes which intersect the surfaces 12, 14 of the mono-crystalline wafer lO. The base 42 of each spring will ultimatelv be fabricated of stiff etch stop material while each leg 44, 46, 48 will ultimately be fabricated of silicon dioxide, as described below in more detail.
n Two characteristics of the shape of the pattern of lines 32-32 are important. The first relates to the area identified by arrow 50, which connects the inner leg 46 to the sensing mass 36. Solely for purposes of illustration and description, this region is called the "cat head with ears" region, and is pointed to by reference arrow 50.
The top of the cat head is attached to inner leg 46 with its "neck" or "body" attached to mass reqion 36. The "ears" of the cat head and ears region extend into oxide areas which ultimatelv will be openings between sensing ?n mass region 36 and support region 38. The cat head and ears qeometry provides a stiff region of mass 36, ultimatelv to be fabricated of undercut etch-stop material, to which inner leg 46 is connected.
The second shape relates to an area pointed to bv reference arrow 52 of Figure 5A. Solely for purposes of illustration and designation, this region is called the "hockey sti~ck" region because of its general resemblance -20- l 334798 to the shape of a hockey stick. This hockey stick region, ultimately to be undercut etch stopped material between regions 43 and 56 (which ultimately will be openings formed by means of silicon etching), provides a relatively wide reqion at the top of the foot of the hockey stick by which outer leq 44 (or 48~ attaches to etch stopped material. Such wide region pro~Jides a strong connection between the ends of the outer leqs of the E-springs and the etch stopped material of the surrounding support structure.
The oxide areas S6 (which ultimately will be openings) have edges 58, 60 which are coexistent with intersecting lines of (111) planes of the crystal with the (100) surface 12 of the wafer. These areas 56 serve as "etch-time and undercut control" areas and are given the short-hand notation, "ETUC" areas. They ultimately serve as large open areas by which etchant chemical will undercut E-springs 40. Their area must be relatively large to allow etchant to etch to intersect (111) planes 2n havinq intersecting lines 62, 64 with the surface 12.
Areas 56 must be large enouqh to speed the etching process to completion to the intersection of the planes of the intersecting lines 62, 64, but must be shaped as shown to provide the "hockey stick" etch stopped areas 52 on its sides as discussed above. The inner edge 60 defines the intersecting line of a (111) plane of the crystal which will define the outer surfaces of the mass region 36.

Another factor dictating the shape of ~TUC areas 56 is the requirement that during chemical etching, the mass region 36 must be severed from the surrounding support frame 38. Chemical etching of the silicon wafer occurs in one direction along (111~ planes which intersect with lines such as 64 and 6~ of the top and bottom surfaces which are extensions of edqes 58 of ETUC areas. Chemical etchinq of the silicon wafer occurs in the opposite direction alonq (111) planes which intersect lines which ln extend from inner edqes 60 of the ETUC areas of the top and bottom surfaces. When the etching from the opposite directions along (111~ planes meet, the mass 36 is separated from support frame 38 as can be seen in Figure 9B. Consequently, the distance between edges 58 and 60 must be great enough to allow the mass region 36 to be separated from support frame 36 during chemical etching.
Turning now to Figure 6A, the silicon dioxide regions 30' defining openings of Figure 5A and the spring leg regions 30'' are masked, and the uncovered silicon ~n portions are implanted with a thin laver of arsenic. The arsenic is diffused a short distance, (6 to 7 microns) in the silicon to form N layers on top and bottom surfaces that are used as electrochemical etch stop la~ers in the silicon crystal. This stage of the processing is illustrated in the cross-section of Figure 6B. Next, the oxide of spring leg region 30'' (of legs 44, 46, 48~ is removed and an additional two micron thickness of arsenic -22- l 33~ 7q8 is diffused into the silicon. Figure 7B shows this staqe of the processing. The thickness of the etch stop layer 70' is now about 8-9 microns thick.
In preferred processing of the mass-spring-support structure of this invention, the thick laver of arsenic is especially important to impart stiffness to certain regions of the top and bottom etch stop layers. The base of the springs 42, cat head region 50 and the hockev stick region 52 require stiffness and strength to connect with the spring legs 44, 46, 48 so that the mass region 36 mav be adequately supported from support region 38.
Figures 8A and 8B show the condition of the wafer after the silicon dioxide regions 30' have been stripped from the wafer 10, a thick layer (1.2 microns) of silicon dioxide has been regrown everywhere over top and bottom surfaces, and finally such silicon dioxide has been stripped awav evervwhere except over spring leg portions 44, 46 and 48. Silicon dioxide layer 30''' remains for such legs as is shown in Figure 8B. At this point in the ~n processing, conductive areas are provided on the opposite faces of the mass region 36 and conductive leads running from the support region via the legs 48 or 44, 46 and base 42 are deposited. For simplification of illustration, the results of providing such conductive region and leads are illustrated in Figure 12 and described below.
Next, the wafer is sub~ected to electro-chemical etching such that the wafer is etched away from the ETUC

areas 56 and other open areas, such as slots 43. The etch chemical etches beneath etch stop layer 70' preferentially along (111) planes of the crystal of the wafer. The preferred etching process is similar to that described in U.S. patent 4,597,003, described above and which may be referred to for its electro-chemical etching methods.
Next, the thin layers 70''' (Figures 7B, 8B) of etch stop material are removed by subjecting the wafer to a silicon etch (either dip or plasma) everywhere. Such etching causes etch stop layer 70' to be reduced in thickness by about 2 microns, that is, to an ultimate thickness of about 6-7 microns.

~ he result of the processing described above is a three dimensional support-spring-mass structure illustrated in plan and cross-section views of Figures 9A
and 9~ and the perspective view partially broken away of Figure 10. The etching process has etched the silicon material awav, along the (111) planes defined by the intersection of such planes with lines parallel to edges 58, 60 of each of the ETUC areas 56. Undercut etch stop layers 70' are visible in Figures 9B and 10 as well as the suspended mass 36, surrounding support structure 38 and silicon dioxide spring layers 30'''. The bases 4~ of the E-springs 40 are formed of undercut etch stop material 70' as seen in Figure 9B and 10.
An important feature of the invention is apparent from Figures 9A and 9B. That feature is the act ~ha~

~~4~ 1 334798 bases 42 of the E-springs are fabricated of etch stop material 70' while the legs 44, 46, 48 of the E-sp~ings are substantially entirely of silicon dioxide. This difference in materials allows the stiffness of the base of each E-spring to be substantially greater than that of the legs.
The coefficient of stiffness (Young!s modlus) is approximatel~ three times greater for silicon than it is for silicon dioxide. The thickness of the base 42 at each 1n E-spring is controlled during processing to be greater than seven times as thick as that of the silicon dioxide leqs. The resulting stiffness of the base, consequently, is seven cubed times three, or more than 1000 times stiffer than the legs for similar lengths and widths of material.
Each spring leg 42, 46, 48 of each E-spring 40 is precisely the same length from its connection at one end of its respective base 42 to its respective connection to either the top of the foot of the hockey stick area or to ~0 the head of the "cat head with ears" area. It is fabricated of silicon dioxide, as described above, because the thickness of the silicon dioxide legs mav be verv accurately controlled. The length of each leg and its thickness must be accuratel~r controlled so as to control the spring constant of the sensing memher. It can be shown that the spring constant of the E-spring is proportional to the cube of the thickness of its legs.

-25- 1 3347~8 Another reason for fabricating the legs of each E-spring of silicon dioxide is that silicon dioxide is a well-behaved material in that it has no measurable hysteresis and it does not fatigue over many cycles of flexing.
Legs 44 and 48 are precisely the same width. Leg 46 is made up of one or more side by side legs whose composite width is precisely twice the width of legs 44 or 48. Although not illustrated, leg 46 is preferably a ~ composite of three side by side legs, each leg of course havinq a smaller width than the composite width. Dividing the total width of leg 46 into three parts has been found to prevent buckling or crinkling of leg 46 after processing.
Figure 11 shows a cross-section of sensing element 80 sandwiched between opposite supporting members 82, 84.
Sensing element 80 shows central mass 36 supported via spring legs 46, 48 and legs 46, 44 via bases 42 from surrounding support member 38. The mass 36 is shown 2n displaced vertically by the force of gravity. The sense conducting areas 90 formed on opposite sides of the faces of mass 36 surround force conducting areas 92. Figure 12 shows a plan view of the support-spring-mass structure 80 with the force and sense conducting areas 92, 90 formed on the top face of the mass 36. Such areas 90, 92 are conductively lsolated from each other. Although not shown in Figure 12 for simplification of illustration, guard regions are provided between areas 90,92 to reduce the capacitive couplinq between such areas. Correspo~ding force and sense conductinq areas 92', 90' are provided on plate surfaces formed on opposite supporting members 82, 84. Such plate surfaces and force and sense conducting areas 92', 90' correspond in size and pattern to the opposite surfaces of mass 36 and its force and sense conducting areas 92, 90. Opposite supporting members are bonded to the support member 38 bv conventional means.
ln Figure 12 shows conductors running from external connection pads to the force and sense plates. Such conductors are illustrated for one surface of the support-sprinq-mass structure 80 as an illustration of such conductors provided to the sense and force conducting axeas of the top and bottom plates 82, 84 and the opposite surfaces of mass 36. Conductors 94 and 96, preferably gold, are deposited on opposite surfaces of the wafer and structure 80 such that they run respectively from connecting pads 98 and 100 via separate E-springs 40A and ~0 40B. As illustrated, lead 94 connects to force conducting area 92 via legs 48 and 46 and base 42 of E-spring 40A, while lead 96 connects to sense conducting area 90 via legs 44 and 46 and base 42 of E-spring 40B.
The leads 94 and 96 respectively pass through tunnels 102, 104 provided for a seal (not shown) between sandwiched surfaces of suppoxt member 38 and opposite supporting members 82, 84. Conductor 94' is deposited on -27- l 33479B

opposite leg 44 and a portion of base 42 of spring 40A to balance the weight and stiffness of conductor 94 on l~eg 48 and base 42. Likewise, conductor 96' is deposited on opposite leg 48 and a portion of base 42 of spring 40~ to balance the~weiqht and stiffness of conductor 96 or leg 44 and base 42. Conductor 94' is not electricallY connected to conductor 94. Likewise, conductor 96' is not electrically connected to conductor 96.
Conductors 106 and 108 are deposited on E-springs 40C
l~ and 40D, but have no external electrical connection. Such conductors 106 and 108 provide weight and stiffness to springs 40D and 40C to balance the weiqht and stiffness of conductors deposited on springs 40A and 40B. Each of the four surfaces, two opposite faces of mass 36 and opposing surfaces of opposite support members 8~ and 84, have similar conduct.ing leads attached to their respective force and sense conducting areas. Such leads are connected to appropriate pads illustrated adjacent pads 98,100 of Figure 1~ which commun.icate with electronic ~n circuitry described below.
Turning now to the electronic circuitr~ associated with the spring mass transducer described above, three alternative embodiments of transducer systems, or accelerometers, are illustrated respectivel~ in Figures 13, 14 and 15. Figure 13 shows a mechanical schematic of the spring mass transducer of Figure 11.
Mass 36 is supported from a support 38 via springs 40 having substantially equal spring constants of 2ks. A
d.c. voltage of magnitude Vcc is applied between sense conducting areas 90' on plates of opposite support members 82, 84, which of course are fixed with support 38. A d.c.
voltage of magnitude Vqq (which may be the same as or different from magnitude Vcc) is applied between force conducting areas 92'. Leads 96 connected to top and bottom sense conducting areas 90 of mass 36 are connected to lead 110. Leads 94 connected to top and bottom force conducting areas 92 of mass 36 are connected to lead 112.
A differential amplifier 120 has one of its two inputs connected to lead 110 with another of its leads connected to a reference voltage which is provided with some fraction of magnitude Vcc, preferably, one half Vcc.
The voltage appearing on lead llO is proportional to the position of mass element 36 between the plates of support members 82 and 84, because a sense electric field is created between sense conducting plates 90' by voltage Vcc, and because sense conducting plates 90 of mass 36 are disposed in the path of that sense electric ield. When the mass element 36 moves toward the plate of top support memher 82, the voltage on sense plates 90 and lead 110 approach the magnitude Vcc; converselv, when the mass element 36 moves toward the plate of bottom support member 84, the voltage on sense plate 90 and lead llO approaches zero or ground magnitude.
Consequentl~,r, with the reference voltage Vcc/2 applied to differential amplifier 20, its output on lead 122 is a sense displacement signal proportional to the distance that mass element 36 has moved from a reference position, halfway between the reference plate of supports 82, 84. The displacement signal produced on lead 122 is smoothed and filtered by filter circuitry 124, resulting in an output siqnal on lead 126.
The voltage Vqq applied between force conducting areas of pl~ates of top and bottom support members 82, 84, creates a force electric field across the force conducting areas 92 of mass 36. A neqative feedback circuit, labelled as charge generator 130 (described below and illustrated in Figure 15) produces an output on its output lead 112 in response to the output displacement signal from lead 128 to apply an amount and magnitude of electric charge on force conducting areas 92 to move the mass toward its reference position, preferably halfwav between support members 82, 84. The orce on mass 36 is proportional to the numerical product of the amount of ?.0 charge deposited on plates 92 times the force electric field, provided the mass is maintained close to or at its reference position.
In other words, force or acceleration appl-ed to support 38 causes displacement of the mass 36 via the springs 40. Displacement of the mass from a reference position is sensed as a displacement signal which causes a charge generator 130 to generate a charge which is applied ~30- 1 334798 to conducting areas 92 so as to drive the mass back to its reference position. Consequently, the difference be,tween the sense displacement signal on lead 110 and the reference potential VCc/2 applied to lead 120 is extremely small, yet the output displacement signal on lead 126 after amplification, and filtering of the output of differential amplifier 120 is representative of the displacement of mass 36 in response to acceleration applied to support 38.
As long as the frequency of applied acceleration is below the effective resonant frequency of the spring-mass-support transducer, displacement of the mass of the tran~ducer is directly proportional to acceleration, and the system of Figure 13 can be described as an accelerometer. Consequently, the output voltage signal of lead 126 is an analog signal proportional to acceleration for acceleration frequencies below the effective resonant frequency of the transducer.
Figurei 14 is a block diagram of an accelerometer similar to that of Figure 13 but has an alternating voltage, preferably a square wave, applied between sense conducting areas 90' and force conducting areas 92' of the plates of opposite support members 82, 84. Different amplitudes of the square waves may be applied to the force and conducting areas. A square wave~train, ]abelled VO1, is illustrated~ in Figure 14 representative of the alternating voltage applied to the force and sense ~ 334798 - conducting areas 92',90' of the opposite plates. The frequenc~ of the square wave-train is high comparçd to that of an illustration of the displacement of mass element 36 in response to acceleration applied to support 38. The voltage appearing on lead 110 is an amplitude modulated signal (suppressed carrier) representative of the displacement of mass 36 from a reference position.
The difference signal of differential amplifier 120 is applied to band pass filter 124 via lead 122 which ln results in a modulated displacement signal on lead 126.
The charqe generator circuit 130 produces modulated voltages to apply charge to force plates 92 of mass 36 in svnchronism with the square wave voltage applied to force conducting areas 92' of opposite support members 82 and 84 to dri~Te the mass to its reference position.
The voltage on lead 126 is applied to demodulator 134 alonq with VO1 and VO1 inputs to produce an analog voltage on lead 136 representative of acceleration applied to supports 38 of the transducer. The advantage of the carrier system of Figure 14, over the analog system of Figure 13, is that the mass displacement signal is shifted out of the base band of the acceleration signal to a higher carrier signal of the square wave train thereby minimizing 1/f noise and d.c. offset problems.
The charge generator 130 of Figure 13 is illustrated in detail in Figure 15. The voltage from the output of the system of Figure 13 is applied via lead 128 to charge -~- 1 334798 aenerator 1~0. Two control si~nal voltages, ~1 and ~2 as illustrated in Figllre 15, are used to control condu~ction of MOS transistor pairs Q1' ~? and Q3, ~4 respectivelv.
Control siqnal voltages d1 and ~2 provide non-coincident turn on signals of a period which is small compared to the period T of the highest frequency acceleration input to the system.
Capacitor CREF is disposed in the conduction path of transistors Ql and Q2. Consequently when Ql and Q2 are ln turned on by control voltage ~1' a charge is p3aced on CREF proportional to the output voltage VOut and lead 128.
A positive charge is placed on capacitor CREF if the output voltage VOut is positive, and vice versa. After control voltage ~1 is turned off, transistor Q3 and Q4 are turned on bv control voltaqe ~2 therebv placinq the charqe stored on capacitor CREF across the inputs of very high gain operational amplifier 137. In response, the output of amplifier 137 changes in a direction that causes the charge stored on capacitor CREF to discharge. ~,ead 112 is ~0 connected between force plates 9~ of the mass 36 of the sensor. Consequently, the discharge path for the charge stored on capacitor CREF is throuqh the capacitance of the force electro~es 92 via lead 11~. The charge on the reference capacltance CREF is totallv transferred to force electrodes 92.
As long as control voltaqe ~ is in the off condition, or the ~I siqnal is applied to transistor pair -33~ 1 334798 Q7' Q8' voltages Vg~ and ground are applied respectivelY
to corresponding force plates 92' of the top and bottom support plates 82, 84. During the time when control signal ~1 is positive and ~1 is not present, transistor pair Q5, Q6 conduct and cause top and hottom pairs of plates 92' and 92 to be connected together, therebv equalizing charge on the respective plates.
The circllit 130 could he modified such that the voltage Vq~ is constantlv applied to Ql on lead 12~ and the output siqnal from lead 126 is applied to transistor p7. This causes a fixed value of charqe to be applied to plates 92 while the voltage across plates 92' is equal to the output voltaqe on lead 126.
The circuit could also be modified such that positive and negative voltages are applied respectively to plates 92' (or 92 as indicated above) such that the electric field across plates 92' is the difference between the positive voltage and the negative voltage. Likewise, the voltage applied across capacitor CREF may be a negative ~n voltage rather than ground as illustrated. Such modif,ication will be apparent to one of skill in the electronic design art.
The charge qenerator 130 o~ Figure 14 .is similar to that illustrated in Figure 15 except that the control siqnal voltages ~1 and ~z are in synchronism with the modulation siqnal ~7~1-Figures 16 through 1~ illustrate different arrangements of the embodiment of the inventio,n where a binary bit stream representation of the displacement of the mass 36 is fed back to the mass-spring-supPort structure to generate a binary force on the mass in a dlrection tending to restore the mass to its reference pos.ition.
Fiqure 16 ;llustrates a support 38-sprinq-mass 36 svstem which is similar to the d.c. svstem of Fiqure 13 in ~,hat a displacement siqnal on lead 110 is applied to a n - differential amplifier 120 havinq a reference potential Vol/~ applied to it corresponding to a reference position of mass 36 between support plates 82 and 84. A difference signal on lead 112 is applied to a filter 124. The output of filter 1~ on lead 126 is then applied to a binarv sampler 150 which is illustrated in Fiqure 16A. The filtered differential displacement signal on lead 126 is applied to a threshold circuit 160 which produces a "1"
("hiqh"~ or "0" ~"low") signal on lead 161 dependinq on whether the displacement signal on lead 1~.6 is qreater ~n than or less than a voltaqe threshold siqnal applied to lead 159. Consequentlv, the signal on lead 16l is e;.ther a "1" or a "0".
The binarv siqnal on lead 161 is app]ied to flip flop circuit 162. to which is applied a clock siqnal labeled TcLocK and which is illustrated in Figure 16. The flip flop circuit produces outputs on its Q and Q leads 164, 166 which is a binarY stream of pulses, as illustrated in _35_ l 334798 Figure 16.
When the siqnal on lead 161 is a "1", the ~inarY
stream on the Q output (lead 164) is a stream of "1"
pulses each of the same period as the pulses of the TCLocK
signal. Simultaneouslv, the Q output llead 166~ is a binary stream of "0" pulses each of the same period. The bit streams are reversed when the signal on lead 161 is a "0". ~he illustration of "l"'s and "0"'s shown in Figures 16, 17, 18 and 19 is an accurate representation of how nthis signal will appear for a sigma delta modulation encoder.
It is preferred that the accelerometer of Figure 16, as well as those of Figures 17-19, have very high noise reduction characteristics. In order to achieve such noise reduction the preferred embodiment of Figure 16 and those of Figures 17-19 includes one or more integrators in the forward loop. With such an integrator, for example in coniunction with filter 124, the filter-inteqrator 1~4 and binarv sampler 15n, with feedback o the output siqnal to ?~the mass 36 to be summed with the input signal applied to support 38, is defined as a sigma-delta-modulation system.
It is preferred to have a single integrator in the forward loop, thus achieving a first order sigma-delta-modulation s~stem. I.ower or higher order systems may find application with the invention of Figures 16-19. Sigma delta modulation systems are described in an IEEE
publication, "Stability Analysis of ~igh-Order Sigma-Delta -36- l 334798 Modulators", bv Ardalan and Paulos, CH 2255 8/86/nO00-0715, 1986.
The output binarv bit stream of a sigma delta moAulation svstem is mostl~ alternating bit patterns with the densitv of " 1 " ' S versus "O"'s being the representation of the acceleration input. In the embodiment of Figure 16, the voltage bit streams on leads 164 and 166 are applied to force conducting plates 92' of the top and bottom support members 82, 84.
ln A constant charge is applied to force conducting plates 92 of mass 36 via charge generator 130 and leads 112, 94. Conse~uently, the force electric field across the mass ~6 is applied in a series of binary voltage pulses to the support member conducting plates to drive mass 36 to its reference position as determined by the d.c. reference potential Vol/2. The binarv bit stream applied to plates 92' reverses in polarity of its bit density as the displacement signal on lead 126 crosses the threshold voltage level applieA to threshold circuit 160.
Such reversal of polarity of the bit streams on leads 164, 166 causes the mass 36 to be driven in the opposite direction.
The charge generator circuit 130 is similar in construction to that illustrated in Figure 15 except that a constant voltage V is applied to lead 128 which causes a constant amount of charge to ~e applied to conducting plates 9~ of mass 36.

-37~ l 334 798 The output of the transducer of Figure 16 on lead 168 is a serial binary bit stream of O's or l's representative of the displacement of the mass 36 with respect to a reference position as specified bv the reference potential Vol/2 applied to amplifier 120. As discussed previousl~, flisplacement of the mass 36 is proportional to acce]erati~n applied to support 38 of the mass 36 spring support 38 svstem.
Fiqure 17 is an alternative binary bit stream nembodiment of the transducer, similar to the transducer of Figure 16, except that a d.c. potential difference V0l -V0l is applied across force conducting plates 92' of top and bottom plates 82, 84. The binary hit stream output from the Q output of binary sampler 150 is applied to lead 128 to charge generator 130, which in response, applies "l"s or "O"s of charge to force conducting plates 92 via leads 112 and 94. In other words, the emhodiment of Figure 17 is the inverse arrangement of that of Figure 16 in that a constant voltage is applied to force plates 92' 20of the top and bottom plates, but binary charge proportional to displacement of the mass 36 (in response to acceleration applied to support 38) is applied to the force plates 9~. As before, mass 36 is driven in a direction toward its reference position as ~etermined by ge (V0l V0l)/~ applied to amplifier 120. The output on lead 168, like that of Figure 16, is a serial binary bit stream proportional to acceleration applied to the support 38 of the transducer.
Fiqure 18 is another alternative arrangement of the embodiment of the invention having a binarv bit stream output and feedback which is representative of displacement or acceleration applied to the transducer.
The arrangement of ~igure 18 is similar to the analog embodiment of Figure 14 in that it applies an a.c. carrier voltage (square wave) to modulate the force electric field on force plates 92' and to sense plates 90' of top and bottom supports 82, 84. The arrangement o Figure 18 is also similar to the binary output arrangement of Figure 17 in that a binarv stream of charge pulses, (but in this arrangement modulated bv the carrier signal Vol! is fefl back to force plates 92 of mass 36 in response to the operation of digital sampler 150. Like in Figure 17, the digital sampler 150 produces a hinary bit stream on lead 168, but modulated by signal V0l, representative of displacement signal 126. Charge generator 130 produces a modulated binarv bit stream on leads 112 and 94 in response to voltage bits applied on lead 128 from lead 168. The modulated bit stream on lead 168 is flemodulatefl bv demodulator 153 rin a similar fashion as the analog signal on lead 126 of Figure 14 is flemodulated) to produce a binary bit stream on lead 170 representative of displacement or acceleration applied to the transducer.
Figure 19 is similar to Figure 18, but feeds a modulated binarv bit stream of voltage pulses back to plates 92' of top and bottom support plates 82, 84 and _39_ l 334798 applies modulated charge, but of constant magnitude, to force plates 92 of mass 36. The result, ~after demodulation of the modulated binarv bit stream signal on lead 152 by demodulator 1~3 i9 a serial binary bit stream representation of the displacement or acceleration applied to the transducer system.
Various modifications and alternatives in the described structures will be apparent to those skilled in the art of the foregoing description which does not depart n from the spirit of the invention. For example, although the preferred method of fabricating the force transducer is by chemical etching of a semi-conductor wafer, plasma etching may also be used effectively. Although the preferred N doping of the spring bases and other regions is arsenic, boron or phosphorous may also be used dependinq upon the original crystal type. Although the preferred method of sensing the displacement of the mass of the mass-spring-support svstem is bv electric field means, a displacement signal representative of the nosition of the mass between the ~op and bottom support plates mav be generated by piezoresistors disposed on or in the springs which change their resistance depending on the flex of the springs. A current through such piezoresistors would produce a voltage signal used in the same way as the voltage signal of the source conducting plates as described above with the preferred embodiments of the invention. For this reason, these changes are _40_ 1 334798 desired to be included in the appended claims. The appended claims recite the only limitations to the pr~esent invention and the descriptive manner which is emploved for settin~ forth the emhodiments and is to be incorporated as illustrative and not limitative.

Claims (39)

1. An improved spring structure in a micro-machined transducer having a support frame and a sensing mass, where said sensing mass is at least partially supported from said support frame via at least one E-shaped leaf spring structure, said spring structure defined by a base and three legs including two outer legs and an inner leg, each of said three legs being connected to said base at one end, said two outer legs connected to said frame at their other ends, said inner leg connected to said mass at its other end, the improvement comprising, said base characterized by a coefficient of stiffness substantially greater than that of said legs.

2. The improved spring structure of claim 1 wherein said improvement further includes said legs each having substantially the same length.

3. The improved spring structure of claim 2 wherein said force transducer is fabricated by etching of a semi-conductor wafer which has been doped with a dopant of one type and wherein said improved spring structure further includes said base of said spring structure being fabricated of semi-conductor material and having an impurity diffused therein of an opposite type from that of which said semi-conductor wafer was doped and wherein said legs of said spring structure are each fabricated of a layer of insulating material, said legs each characterized by a thickness which is small compared to the thickness of said base.

4. The improved spring structure of claim 1 wherein said insulating material of which said legs are fabricated is silicon dioxide.

5. The improved spring structure of claim 3 wherein said semiconductor material is silicon and said impurity diffused therein is of an N type impurity.

6. The improved spring structure of claim 3 wherein said semi-conductor material is silicon and said impurity diffused therein is of a P type impurity.

7. The improved spring structure of claim 1 wherein said improved spring structure further includes said two outer legs having substantially the same width and having substantially equal coefficients of stiffness, and said inner leg having a width about twice the width of said outer legs.

8. The improved spring structure of claim 7 wherein said inner leg has at least one groove disposed along its length effectively creating side-by-side inner legs each having substantially the same combined width and coefficient of stiffness as said combined outer legs.

9. In a method for making a transducer including a spring mass system by chemical etching of a semiconductor wafer, the steps of diffusing etch stop layer patterns on opposite major faces of a semi-conductor wafer forming a first intermediate wafer where said patterned etch stopped layers define a support frame structure, a sensing mass structure and an E-shaped leaf spring structure, said spring structure defined by a base and three legs including two outer legs and an inner leg, each of said three legs being connected to said base at one end, said two outer legs connected to said frame at their other ends, said inner leg connected to said mass structure at its other end;
diffusing additional etch stop material into said support frame structures and said sensing mass structures and said bases of said E-shaped leaf spring structure of said first intermediate wafer, to form a second intermediate wafer;
forming additional silicon dioxide on each of said three legs of said spring structures of said second intermediate wafer to form a third intermediate wafer;
introducing a semiconductor etchant via openings on opposite faces of said wafer where no etch stop layer is diffused to undercut substantial portions of said etch stop layers including said E-shaped leaf spring structure;
terminating the etch of semiconductor material so as to leave undercut structure in both of said etch stopped layers interconnected to each other by interconnecting substrate semiconductor material, and removing etch stop material beneath said silicon dioxide layers of said three legs of said undercut E-shaped leaf spring structure.

10. In the steps of claim 9 wherein said major faces of said semiconductor wafer are (100) faces, and wherein said E-shaped leaf spring structures are oriented substantially at forty-five degree angles with respect to the intersection lines of (111) planes of the wafer.

11. In the steps of claim 9 wherein diagonally opposed E-shaped undercut spring structures are formed on opposite sides of said wafer for supporting said sensing mass structure from said support frame structure.

12. In the steps of claim 11 wherein two pair of diagonally opposed E-shaped undercut spring structures are formed on opposite sides of said wafer for supporting said sensing mass structure from said support frame structure, each of said two pair of spring structures being orthogonal to the other.

13. In the steps of claim 12 wherein each of said opposite faces of said wafer have open areas where no etch stop layer is diffused including etch-time and undercut control (ETUC) areas between each of said E-shaped undercut spring structures, each of said ETUC areas having an outer edge coinciding with an intersection line of a (111) plane of the wafer and a major face of said wafer, each of said ETUC areas having an inner edge coinciding with an intersection line of a (111) plane of the wafer and major face of said wafer which defines an outer edge of said sensing mass structure, each of said outer edges of said ETUC coinciding with (111), planes which extend outwardly beyond said E-shaped undercut spring structure on either side of said opening, whereby when etchant is introduced into said ETUC area openings, said E-shaped spring structures are undercut to said outer edge coinciding (111) planes.

14. In the steps of claim 13 wherein said etch stop layer pattern defining said support frame structure includes a hockey stick shaped area between one of said outer legs of said E-shaped spring structure and an adjacent etch time and undercut control area, said hockey stick shaped area including a foot and a leg, where said outer leg of said E-shaped spring structure is connected to the top of the foot of said hockey stick shaped area, said foot of said hockey stick shaped area providing a strength region for connecting said outer leg to said support frame structure.

15. In the steps of claim 13, wherein adjacent outer edges coinciding with a (111) plane of the wafer and a major face of said wafer which define outer edges of said sensing mass structure are connected by a cat head and ears shaped area of etch stop material, whereby, said inner leg of an E-shaped leaf spring structure is connected to the head of said cat head and ears shaped area, and where each of said ears of said sensing mass structures extend into a respective space between an outer leg and said inner leg of said E-shaped leaf spring structure.

16. In the method of claim 9, the additional steps of forming force conductive areas on said opposite sides of said sensing mass structure; and forming sense conductive areas on said opposite sides of said sensing mass structure, while conductively isolating said sense area from said force areas.

17. In the method of claim 16, the additional steps of forming force conductors from a location exterior of said spring structures and said sensing mass structures to both of said force conductive areas on opposite sides of said sensing mass structure via one set of oppositely faced E-springs, and forming a sense conductor from a location exterior of said spring structures and said sensing mass structures to both of said sense conductive areas on opposite sides of said sensing mass structures via a different set of oppositely faced E-springs.

18. In the method of claim 17, the additional steps of forming first and second plates of semi-conductive material, each of said plates having a peripheral outside area disposed about force and sense conductive areas which are electrically isolated from each other and which are mirror images of said force and sense conductive areas disposed on said opposite sides of said sensing mass structure, forming first and second force conductors respectively to said force conductive areas of said first and second plates, forming first and second sense conductors respectively to said sense conductive areas of said first and second plates, and assembling a sandwich structure with said sensing mass structure disposed between said first and second plates with said force and sense conductive areas of said first and second plates facing said force and sense conductive areas of said opposite sides of said sensing mass structure, said sensing mass structure being supported via said E-spring structure from said support frame structure, said support frame structure secured to peripheral outside areas of said first and second plates.

19. A transducer comprising:
a movable mass supported by springs from a support structure;
said mass having a surface with a sense conducting area and a force conducting area, said mass being free to move between corresponding first and second sense conducting areas and first and second force conducting areas on plates which face said surface of said mass and which are secured to said support structure;
means for applying a first voltage between said first and second sense conducting areas of said plates thereby creating a sense electric field between said sense conducting areas and inducing a sense voltage on said mass sense conducting area of said mass surface proportional to the displacement of said mass surface between said plates;
means for applying a second voltage between said first and second force conducting areas of said plates thereby creating a force electric field between said force conducting areas;
forward circuit means responsive to said sense voltage for generating a displacement signal representative of the distance said mass has moved from a reference position between said plates;
and feedback circuit means responsive to said displacement signal for applying an amount of electric charge on said force conducting area of said mass surface such that a force is created on said mass proportional to the numerical product of said amount of charge and said force electric field in a direction to restore said mass to said reference position.

20. The transducer of claim 19 wherein said mass has opposing surfaces, each surface having a sense conducting area and a force conducting area, with said sense conducting areas of said opposing surfaces being conductively connected and with said force conducting areas of said opposing surfaces being conductively connected.

21. The transducer of claim 19 wherein said first and second voltages applied respectively between said first and second sense and force conducting areas of said plates which face said mass are d.c. voltages.

22. The transducer of claim 19 wherein said springs and said mass are characterized by a resonant frequency, and wherein said displacement signal is substantially proportional to acceleration of said support structure for acceleration of frequency less than said resonant frequency.

23. The transducer of claim 21 wherein said forward circuit means includes a differential amplifier having one of its input leads connected to said sense conducting area of said mass and having another of its input leads connected to a reference voltage source, the level of said reference voltage source being a percentage of said sense voltage corresponding to said reference position of said mass between said first and second conducting areas between said plates.

24. The transducer of claim 23 wherein said reference voltage is about one-half said first voltage between said first and second sense conducting areas and said reference position of said plate is about one-half the distance between said plates which face said mass.

25. The transducer of claim 19 wherein said first and second voltages applied respectively between said first and second sense and force conducting areas of said plates which face said surface of said mass are a.c. voltages whereby said sense electric field and said force electric field are a.c. electric fields.

26. The transducer of claim 25 wherein said first and second voltages are square wave voltage signals.

27. The transducer of claim 26 wherein said forward circuit means includes a differential amplifier having one of its input leads connected to said sense conducting area of said mass and having another of its input leads connected to a reference voltage source, the level of said reference voltage source being a percentage of the voltage level difference between upper and lower amplitude levels of said square wave voltage signals.

28. The transducer of claim 25 wherein said displacement signal is an a.c. signal, the peak to peak amplitude of which is proportional to the distance said mass has moved from said reference position between said plates;
said springs and said mass are characterized by a resonant frequency, and wherein said peak to peak amplitude of said a.c. displacement signal is substantially proportional to acceleration of said support structure for accelerations of frequency less than said resonant frequency.

29. The transducer of claim 28 further comprising, demodulator circuit means responsive to said a.c.
displacement signal and to said a.c. voltage applied between said first and second sense conducting areas for generating an analog signal representative of acceleration applied to said support structure.

30. The transducer of claim 28 further comprising:
circuit means responsive to said displacement signal for generating a binary output signal representative of said displacement signal which is a binary representation of acceleration of said support structure.

31. The transducer of claim 19 further comprising:
circuit means responsive to said displacement signal for generating a binary output signal representative of said displacement signal which is a binary representation of acceleration of said support structure.

32. A transducer comprising:

sensing means for generating a displacement signal representative of displacement of a mass from a reference position with respect to a support structure;
forward circuit means responsive to said displacement signal for converting said displacement signal to an output binary bit stream;
feedback circuit means responsive to said output binary bit stream for generating a binary force on said mass in a direction tending to restore said mass to said reference position;
whereby said output binary bit stream is representative of a characteristic of motion of said support structure.

33. The transducer of claim 32 wherein:
said mass is supported by springs from said support structure; and said mass has a surface with a sense conducting area and a force conducting area, said mass being free to move between corresponding first and second sense conducting areas and first and second force conducting areas on plates which face said surface of said mass and which are secured to said support structure.

34. The transducer of claim 33 wherein:
said displacement signal is generated on said sense conducting area by applying a d.c. potential across said first and second sense conducting areas of said plates secured to said support structure; and said binary force on said mass is generated by applying complementary binary bit streams to said first and second force conducting areas of said plates secured to said support structure and by applying a constant charge to said force conducting area of said mass.

35. The transducer of claim 33 wherein:
said displacement signal is generated on said sense conducting area by applying a d.c. potential across said first and second sense conducting areas of said plates secured to said support structure; and said binary force on said mass is generated by applying a d.c. potential across said first and second force conducting areas of said plates secured to said support structure and by applying a binary bit stream of charge, proportional to said output binary bit stream, to said force conducting area of said mass.

36. The transducer of claim 33 wherein:
said displacement signal is generated on said sense conducting area by applying an a.c. potential across said first and second sense conducting areas of said plates secured to said support structure; and said binary force on said mass is generated by applying an a.c. potential across said first and second force conducting areas of said plates secured to said support structure and by applying a modulated binary bit stream of charge, proportional to said output binary bit stream which is modulated by said a.c.
potential, to said force conducting area of said mass.

37. The transducer of claim 33 wherein said displacement signal is generated on said sense conducting area by applying an a.c. potential across said first and second sense conducting areas of said plates secured to said support structure; and said binary force on said mass is generated by applying an a.c. modulated charge to said force conducting area of said mass and by applying complementary modulated bit streams of voltage, proportional to said output binary bit stream which is modulated by said a.c. potential to said first and second force conducting areas of said plates secured to said support structure.

38. The transducer of claim 32 whereby said forward circuit means includes integration means; and wherein said sensing means, said forward circuit means and said feedback circuit means define a sigma delta modulation system.

39. A transducer comprising:
a movable mass supported by springs from a support structure;
sensing means for generating a sensing signal representative of displacement of said mass from a reference position with respect to said support structure;
said mass having a surface with a force conducting area, said mass being free to move between first and second force conducting areas on plates which face said surface of said mass and which are secured to said support structure;
means for applying a voltage between said first and second force conducting areas of said plates thereby creating a force electric field between said force conducting areas;
forward circuit means responsive to said sensing signal for generating a displacement signal representative of the distance said mass has moved from a reference position between said plates;
and feedback circuit means responsive to said displacement signal for applying an amount of electric charge on said force conducting area of said mass surface such that a force is created on said mass proportional to the numerical product of said amount of charge and said force electric field in a direction to restore said mass to said reference position.