1. Introduction
Many thin films are capable of exhibiting large elastic deflection under transverse loading [
1,
2,
3,
4,
5,
6], which provides the possibility of the design and development of elastic-deflection-based devices [
7,
8,
9,
10,
11,
12,
13,
14]. Pressure sensors based on thin film elastic deflection have widespread applications in many areas, such as bio-medical applications, robotics, automobiles, and environmental monitoring, and capacitive structures are widely used. These capacitive types of pressure sensors convert the elastic deformation of thin films, corresponding to the pressure applied on the thin films, into a change in capacitance. They usually use single-crystal silicon and polysilicon [
15], polymer/ceramic [
16], low-temperature co-fired ceramic [
17], silicon carbide [
18,
19], or graphene–polymer heterostructure [
20,
21,
22,
23] thin films in microelectromechanical systems (MEMS), and have the advantages of low cost, small volume, high stability, high sensitivity, low temperature drift, and lower sensitivity to environment effects.
Figure 1 illustrates the typical structure and modes of operation of a traditional capacitive pressure sensor. On application of a pressure
q, the conductive membrane, as the upper electrode plate of the capacitor, elastically deflects in response to the applied pressure
q. This elastic deflection is a measure of the applied pressure
q and also changes the capacitance of the capacitor. Therefore, the applied pressure
q can be determined by measuring the change in capacitance. In the so-called normal mode of operation, the deflected conductive membrane, as the upper electrode plate of the capacitor, is always kept at a distance away from the isolation layer coating the lower electrode plate, as shown in
Figure 1b, and thus a device operating in this state may also be called a non-touch-mode capacitive pressure sensor. When in the so-called touch mode of operation, the deflected conductive membrane is always kept in contact with the isolation layer, as shown in
Figure 1c, and therefore a device operating in this state is called a touch-mode capacitive pressure sensor. Usually, the output capacitance of a non-touch-mode capacitive pressure sensor is nonlinear with respect to the input pressure changes, and the sensitivity in the near-linear region is not high enough to ignore the many stray capacitance effects. Touch-mode capacitive pressure sensors are known to have robust structures to withstand harsh industrial environments and higher sensitivity by one or two orders of magnitude than in the normal mode of operation in the near-linear operation range, so that some of the stray capacitance effects can be neglected. Moreover, the output capacitance of touch-mode devices is mainly the isolation-gap capacitance of the touched area due to the very thin isolation layer, and its capacitance per unit area is much larger than the air-gap capacitance in the untouched area. This is the main reason why touch-mode capacitive pressure sensors are often called linear sensors: the change in the touched area is usually designed to be almost proportional to the applied pressure
q, and thus the output capacitance–input pressure characteristic is nearly linear.
However, there are two types of difficulties or problems in the design of traditional capacitive pressure sensors. Firstly, there are difficulties in the preparation or selection of conductive elastic membranes. The conductive membrane is used as both the movable upper electrode plate of the capacitor and the deformation element that elastically responds to the applied pressure. Therefore, the preparation or selection of the conductive membrane depends not only on its high electrical conductivity but also on its good elastic deformation ability, which obviously places high requirements on the preparation or selection of materials. Secondly, there are difficulties in balancing the linear input–output characteristic and the wide operational pressure range. It is in fact very difficult to design a touch-mode capacitive pressure sensor with a wide operational pressure range and a nearly linear characteristic between the output capacitance and the input pressure. In general, when the operational pressure range is too wide, there often is a strong nonlinear relationship between the change in touched area and the applied pressure. As a result, the designer must often choose between a wide operational pressure range and a nearly linear input–output characteristic.
In our earlier work [
24], we proposed an improved capacitive pressure sensor using non-conductive elastic membranes and conductive rigid thin plates instead of traditional conductive elastic membranes. The topside structure of the proposed capacitive pressure sensor uses a non-conductive elastic annular membrane (as an elastic deformation element to respond to the applied pressure), whose inner edge is rigidly connected to the outer edge of a conductive, rigid, flat, concentric-circular thin plate (as a movable upper electrode plate of the capacitor), to substitute for the dual-function topside structure of traditional capacitive pressure sensors, i.e., the traditional conductive elastic membrane used as both an elastic deformation element and an upper electrode plate. The proposed capacitive pressure sensor uses independent elastic deformation elements and upper electrode plates, overcoming the shortcomings of traditional capacitive pressure sensors. Non-conductive membranes with very good elasticity are abundant, and rigid thin plates with high electrical conductivity are easier to find, making the preparation or selection of materials very easy. Furthermore, the convenience of material preparation or selection allows a wider range of material parameters to be selected, such as the Poisson’s ratio, Young’s modulus of elasticity, and the thickness of the membrane, as well as the radius of the conductive, rigid, flat, concentric-circular thin plate. In particular, the conductive, rigid, flat, concentric-circular thin plate, as a movable upper electrode plate, forms a parallel plate capacitor with the flat lower electrode plate, and the calculation for a parallel plate capacitor is well known to be easier than that for a non-parallel plate capacitor. All of these advantages provide great convenience for balancing a wide operational pressure range and a nearly linear input–output characteristic.
The improved capacitive pressure sensors proposed in [
24] have many advantages over traditional capacitive pressure sensors: they are more suitable for large-volume (size) sensors such as those used for building-facade wind pressure measurements, etc. However, our earlier work [
24] failed to accurately solve the behavior of the elastic deformation of the annular membrane analytically, due to the complexity of the problem. An accurate analytical solution is usually very important for sensor design, and the closed-form solution presented in [
24] could not meet the design requirements of the proposed capacitive pressure sensor, due to the adopted assumption condition that the rotation angle
θ of the annular membrane is so small that “sin
θ = tan
θ” can be used to replace “sin
θ = 1/(1 + 1/tan
2θ)
1/2”. Obviously, such an assumption inevitably introduces computational errors and affects the accuracy of the closed-form solution presented in [
24] when the rotation angle of the annular membrane is relatively large, i.e., when the applied pressure is relatively large. As is well known, the sine function sin
θ can be approximated by the tangent function tan
θ only when
θ is relatively small, and a large rotation angle
θ will give rise to a significant approximation error. For instance, the error caused by approximating sin
θ to tan
θ is about 1.54% when
θ = 10°, 6.42% when
θ = 20°, 15.47% when
θ = 30°, and 30.54% when
θ = 40°. In fact, the rotation angle
θ of the annular membrane in the proposed parallel-plate-capacitor-based pressure sensor may exceed 40°. Therefore, it is necessary to reject the approximation of replacing “sin
θ = 1/(1 + 1/tan
2θ)
1/2” with “sin
θ = tan
θ” in the derivation of the closed-form solution. Hence, the behavior of the elastic deformation of the annular membrane under pressure must be analytically solved again.
Our earlier work [
24] also failed to give an illustration of how to use the closed-form solution to achieve the numerical calibration of the relationship between the output pressure and the input capacitance of the proposed capacitive pressure sensor. As can be seen below, changes in material parameters such as the initial gap between the upper and lower electrode plates of the parallel plate capacitor, the Young’s modulus of elasticity, and the thickness of the membranes, have an important effect on the relationship between the output pressure and the input capacitance of the proposed capacitive pressure sensor. Therefore, the numerical calibration plays a very important role in the design phase of the proposed capacitive pressure sensor as it determines the required operational pressure ranges and input–output characteristics. In other words, it is impossible to achieve the required operational pressure ranges and input–output characteristics by just changing the actual materials used (i.e., by experimental calibration). Thus, closed-form solutions are of unparalleled and irreplaceable value in sensor design. Therefore, as a purely theoretical study, since the closed-form solution is given, it is necessary to address how to use the closed-form solution to achieve the numerical calibration of the proposed capacitive pressure sensor. However, our previous work [
24] failed to do this.
This paper renames the improved capacitive pressure sensor proposed in [
24] as a capacitive pressure sensor based on thin film elastic deflection and a parallel plate capacitor (or an elastic-deflection-and-parallel-plate-capacitor-based pressure sensor for short) and presents a further theoretical study of the pressure sensor. In this paper, in order to improve the accuracy of analytical solutions, the assumption adopted in [
24] is rejected, resulting in a new and more refined closed-form solution for the behavior of the elastic deformation of the annular membrane. For the first time, examples are given to illustrate how to use the resulting closed-form solution to achieve the numerical calibration of the relationship between the output pressure and input capacitance of the elastic-deflection-and-parallel-plate-capacitor-based pressure sensor. In addition, the effect of important parametric variations on the input–output characteristics is also discussed numerically. The novelty or innovation of this paper mainly lies in the following three aspects. A new and more refined closed-form solution is presented, where the newly presented closed-form solution can greatly reduce the pressure measurement error for the same input capacitance (i.e., under the same maximum elastic deflection, in comparison with the previously presented closed-form solution. A numerical example of how to use the resulting closed-form solution to numerically calibrate the input–output characteristics is given for the first time. The effect of important parametric variations on the input–output characteristics is addressed, which has important theoretical significance for guiding the design of capacitive pressure sensors based on thin film elastic deflection and a parallel plate capacitor. By changing some important parameters and carrying out a series of numerical calibrations, the variation trend of the operational pressure ranges and input–output characteristics with important parametric variations can be found. This can clarify how to appropriately prepare or select materials to achieve the desired operational pressure ranges and input–output characteristics.
The paper is organized as follows. In the following section, the structure, the mode of operation, and the working principle of the elastic-deflection-and-parallel-plate-capacitor-based pressure sensor are briefly described. In
Section 3, the behavior of the elastic deformation of the annular membrane of the elastic-deflection-and-parallel-plate-capacitor-based pressure sensor under pressure is analytically solved again, the assumption condition adopted in [
24] is rejected, and a new and more refined closed-form solution is given. In
Section 4, some important issues are addressed. The validity of the closed-form solution obtained in
Section 3 is first addressed. Secondly, the new closed-form solution given in this paper is numerically compared with the one given in [
24], in terms of the pressure measurement error under the same maximum elastic deflection (i.e., under the same input capacitance). Next, an example is given to illustrate how to use the closed-form solution obtained in
Section 3 to achieve the numerical calibration of the relationship between the output pressure and input capacitance of the elastic-deflection-and parallel-plate-capacitor-based pressure sensor. Finally, the effect of important parametric variations on input–output relationships is numerically discussed, showing how the required operational pressure ranges and input–output characteristics can be achieved based on a series of numerical calibrations. Concluding remarks are given in
Section 5.
2. Materials and Methods
The structure or geometry of the proposed elastic-deflection-and-parallel-plate-capacitor-based pressure sensor is shown in
Figure 2a, where
a denotes the outer radius of the initially flat non-conductive elastic annular membrane,
b denotes the inner radius of the annular membrane, as well as the outer radius of the conductive, rigid, flat, concentric-circular thin plate, and
g denotes the initial gap between the initially flat non-conductive elastic annular membrane and the flat lower electrode plate. The inner edge of the initially flat non-conductive elastic annular membrane is rigidly connected to the outer edge of the conductive, rigid, flat, concentric-circular thin plate, forming the topside structure of the proposed pressure sensor. The conductive, rigid, flat, concentric-circular thin plate, as a movable upper electrode plate, forms a parallel plate capacitor with the flat lower electrode plate. On application of the pressure
q, as shown in
Figure 2b, the initially flat non-conductive elastic annular membrane will deflect towards the lower electrode plate and work as an elastic deformation element in response to the applied pressure
q, resulting in the upper electrode plate moving a distance
wm (the maximum elastic deflection) from its initial position (that of the initially flat annular membrane) towards the lower electrode plate. Clearly, the movement of the upper electrode plate will result in a capacitance change in the parallel plate capacitor. Therefore, the applied pressure
q can be determined by measuring the capacitance change caused in the parallel plate capacitor.
The topside structure of the proposed elastic-deflection-and-parallel-plate-capacitor-based pressure sensor can also be formed by a non-conductive elastic circular membrane whose central region firmly adheres to the conductive, rigid, flat, concentric-circular thin plate (such that the central region membrane will not produce elastic deformation when the upper electrode plate moves). In addition, the initial gap
g between the upper and lower electrode plates should be far less than the diameter 2
b of the upper and lower electrode plates, such that the fringe effect in the capacitance calculation of the parallel plate capacitor can be ignored. Therefore, the capacitance between the two parallel conductive, flat, circular thin plates with radius
b, dielectric constant
ε, and air gap
g-wm (see
Figure 2b), after neglecting the fringe effect, may be written as [
25]
As mentioned above, the application of a pressure q will result in the maximum elastic deflection wm. In other words, there is a one-to-one correspondence. Hence, wm is a continuous function of q, i.e., wm(q). Therefore, once wm(q) is obtained, the relationship between the pressure q and the capacitance C can be determined. We must solve for the behavior of the elastic deformation with large deflection of the annular membrane analytically, to obtain an accurate continuous function wm(q).
Analytical solutions for the large-deflection phenomenon of elastic membranes are available in only a few cases, due to the difficulties of analysis. However, for the design and development of elastic-deflection-based devices, accurate analytical solutions are often found to be necessary [
26,
27]. Our earlier work [
24] failed to accurately solve the behavior of the elastic deformation with large deflection of the annular membrane (see
Figure 2b) analytically. The closed-form solution presented in [
24] could not meet the accuracy requirements for designing the proposed elastic-deflection-and-parallel-plate-capacitor-based pressure sensor, as it introduced too many pressure-measurement errors for the same input capacitance (i.e., under the same maximum elastic deflection
wm). Therefore, the next section is devoted to the new and more refined closed-form solution for the behavior of the elastic deformation with large deflection of the annular membrane.
3. Refined Closed-Form Solution
An initially flat, linearly elastic annular membrane with thickness
h, outer radius
a, inner radius
b, Poisson’s ratio
v, and Young’s modulus of elasticity
E is tightly fixed at its outer edge and connected at its inner edge to a movable, weightless, rigid, concentric-circular thin plate of radius
b, resulting in an immovable and non-deformable outer edge and a movable but non-deformable inner edge. At the same time, a uniformly distributed transverse load
q is quasi-statically applied to the annular membrane and movable, weightless, rigid, concentric-circular thin plate, resulting in an out-of-plane displacement (deflection) of the annular membrane, as shown in
Figure 3. In the figure, a cylindrical coordinate system (
r, φ, w) is introduced, with the polar coordinate plane (
r, φ) located in the plane in which the geometric middle plane of the initially flat annular membrane is located, and where
o denotes the origin of the introduced cylindrical coordinate system (
r, φ, w) (which is placed in the centroid of the geometric middle plane),
r denotes the radial coordinate,
φ denotes the angle coordinate (not represented in
Figure 3), and
w denotes the axial coordinate as well as the transverse displacement of the deflected membrane. A free body, a piece of annular membrane with radius
r (
b ≤ r ≤ a), is taken from the central portion of the deflected annular membrane, to study the static problem of equilibrium of this free body under the joint action of the external active force
πr2q produced by the uniformly distributed transverse loads
q and the reactive force 2
πrσrh produced by the membrane force
σrh acting on the boundary
r, as shown in
Figure 4, where
σr denotes the radial stress and
θ denotes the rotation angle of the deflected annular membrane.
The so-called out-of-plane equilibrium equation can be obtained from the equilibrium condition that the resultant force in the transverse (vertical) direction is equal to zero, and is given by
If the transverse displacement of the deflected annular membrane at
r is denoted by
w(
r), then
Substituting Equation (3) into Equation (2), the out-of-plane equilibrium equation can be written as
In the horizontal direction parallel to the initially flat annular membrane, there are the actions of the radial membrane force
σrh and the circumferential membrane force
σth, where
σt denotes the circumferential stress. Therefore, the so-called in-plane equilibrium equation may be written as
If the radial strain, circumferential strain, and radial displacement are denoted by
er,
et, and
u(
r), respectively, then the relations between the strain and displacement, the so-called geometric equations, may be written as
and
Moreover, the relations between the stress and strain, the so-called physical equations, are still assumed to satisfy linear elasticity
and
Substituting Equations (6) and (7) into Equations (8) and (9) yields
and
Eliminating d
u/d
r from Equations (10) and (11) and further using Equation (5), it is found that
After substituting the
u of Equation (12) into Equation (10), the so-called consistency equation may be written as
Equations (4), (5) and (13) are three equations for the solutions of
σr,
σt, and
w. The boundary conditions for solving Equations (4), (5) and (13) are
and
Let us introduce the nondimensionalization
and transform Equations (4), (5) and (12)–(16) into
and
From Equation (18), it is found that
Eliminating d
W/d
x from Equations (21) and (25), we can obtain an equation which contains only
Sr:
In view of the physical phenomenon that the values of stress, strain, and displacement are all finite within the range
α ≤
x ≤ 1, we can expand
Sr and
W into a power series of (
x−
β), i.e.,
and
where
β = (1 +
α)/2. For convenience we introduce
X =
x −
β, then Equations (25)–(28) can be transformed into
and
Substituting Equation (31) into Equation (30) and letting the sums of all coefficients of the same powers of
X be equal to zero yields a system of equations for determining the recursion formulas for the coefficients
ci. The solution to this system of equations shows that the coefficients
ci (
i = 2, 3, 4, …) can be expressed as polynomial functions with regard to the first two coefficient
c0 and
c1 (see
Appendix A). Further, by substituting Equations (31) and (32) into Equation (29), the coefficients
di (
i = 1, 2, 3, …) can also be expressed in terms of
c0 and
c1 (see
Appendix B).
The remaining three coefficients
c0,
c1, and
d0 are three undetermined constants, which depend on the specific problem addressed and can be determined by using the boundary conditions of Equations (22) and (23), as follows. Substituting Equation (31) into Equations (22) and (23) yields
and
If all the recursion formulas for the coefficients
ci in
Appendix A are repeatedly substituted into Equations (33) and (34), Equations (33) and (34) will contain only
c0 and
c1. Therefore, the values of
c0 and
c1 can be determined by simultaneously solving Equations (33) and (34), and the expression for
Sr can be determined. Further, substituting Equation (32) into the boundary condition of Equation (24) yields
Therefore, with the known c0 and c1, the value of d0 can finally be determined using Equation (35), and the expression for W can also be determined. The expression for St, with the known expression of Sr, can easily be determined by Equation (19).
The problem addressed here is thus solved analytically, and its closed-form solutions for stress and deflection are given. The closed-form solution for stress will be used to check whether the thin film used meets the mechanical strength required, while the closed-form solution for deflection will be used to determine the important analytical relationship between the applied pressure q and the maximum deflection wm.
The maximum deflection of the deflected annular membrane
wm is at
x =
α, and from Equations (17) and (28) it may finally be written as
The maximum stress of the deflected annular membrane
σm is also at
x =
α, and from Equations (17) and (27) it may be written as