Material selection for Micro-Electro-Mechanical-Systems (MEMS) using Ashby's approach

Accepted Manuscript Material selection for Micro-Electro-Mechanical-Systems (MEMS) using Ashby's approach

Zahid Mehmood, Ibraheem Haneef, Florin Udrea PII: DOI: Reference:

S0264-1275(18)30598-7 doi:10.1016/j.matdes.2018.07.058 JMADE 7291

To appear in:

Materials & Design

Received date: Revised date: Accepted date:

17 March 2018 27 July 2018 28 July 2018

Please cite this article as: Zahid Mehmood, Ibraheem Haneef, Florin Udrea , Material selection for Micro-Electro-Mechanical-Systems (MEMS) using Ashby's approach. Jmade (2018), doi:10.1016/j.matdes.2018.07.058

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT MATERIAL SELECTION FOR MICRO-ELECTRO-MECHANICAL-SYSTEMS (MEMS) USING ASHBY’S APPROACH Zahid Mehmooda, b, Ibraheem Haneefa, Florin Udreab a Institute of Avionics & Aeronautics, Air University, E-9, Islamabad 44000, Pakistan bDepartment of Engineering, University of Cambridge, Cambridge CB3 0FA, UK Abstract

US

CR

IP

T

A key aspect in design optimization of a product or a system is the selection of materials that best meet the design needs, ensuring maximum performance and minimum cost. Ashby’s approach, originally introduced for macro-systems and products, has been very successfully employed for Micro-ElectroMechanical-Systems (MEMS) / micromachined sensors, actuators and devices. This paper presents a comprehensive review and critical analysis of MEMS material selection studies using Ashby’s approach reported in the literature during the last two decades. Performance and Material Indices derived for various microsystems and MEMS devices have been summarized. Moreover, all MEMS materials reported in the literature and the most suitable materials proposed for a variety of MEMS systems and devices have also been consolidated. A material selection case study utilizing micro-scale properties of 51 MEMS compatible materials has been presented to demonstrate that the use of different materials’ bulk properties is not the best choice for MEMS materials selection. This paper will serve as a reference guide and useful resource for researchers and engineers engaged in the design and fabrication of various microsystems and MEMS sensors, actuators and devices. Key Words:

AN

Material Selection, Ashby’s Methodology, Microsystems, Micro-Electro-Mechanical-Systems (MEMS), Performance Index, Material Index, Design Optimization, Product Design

[Pa]

𝐸

[Pa]

𝑎

[m]

𝐶1 , 𝐶2

[1]

ED

𝜎𝑓

𝑝

Nomenclature

[Units] :

Fracture strength

:

Young’s modulus

:

Diaphragm length

:

Constants

:

Pressure change

:

Length

PT

Symbols

M

List of Symbols

CE

[Pa] [m]

𝜌

[Kg/m3]

:

Density

𝐾

[N/m]

:

Stiffness

[Kg]

:

Effective mass

:

Width

ℎ

[m] [m]

:

Thickness

𝐵

[J / oK]

:

Boltzmann constant

𝑇

oK

:

Absolute temperature

𝛽

[Hz]

:

Band width

𝜒

[1]

:

Loss coefficient

𝑝

[Pa]

:

Static Pressure

𝛿

[m]

:

Deflection

𝑚∗ 𝑏

AC

𝐿

1

ACCEPTED MANUSCRIPT [N]

:

Force

𝜎𝑚𝑎𝑥

[Pa]

:

Maximum stress

𝜀𝑚𝑎𝑥

[m/m]

:

Maximum strain

𝐼

[m4]

:

Moment of inertia

𝑉

[V]

:

Voltage

𝑓

[Hz]

:

Frequency

𝐴

[m2]

:

Area

𝑔𝑜

[m]

:

Un-actuated gap

𝑔

[m]

:

Actuated gap

𝜀

[F/m]

:

Permittivity

𝜆

[1]

:

Modulus ratio

𝜉0

[1]

:

Optimal thickness ratio

𝑑31

[m/V]

:

Material constant of piezoelectric layer

𝜂

[1]

:

Efficiency

𝐾′

[1]

:

Electromechanical coupling

𝑑

[m]

:

Distance between capacitor plates

𝑑1

[m]

:

Final position achieved by capacitor plate

𝑚

[Kg]

:

𝑎′

[m/s2]

:

𝜈

[1]

:

[1/

𝛾

[1]

𝑘

[W/m-oK]

𝑁

[1]

𝜙

[ions/m2-s]

𝑀

[Kg/mol]

𝐷

[Kg/m3]

𝑅

[Ω]

IP

CR

US

AN M

Acceleration Poisson’s ratio

:

Coefficient of Thermal Expansion (CTE)

:

Reducing factor

:

Thermal conductivity

PT

𝛼

Mass

ED

oK]

T

𝐹

Numerical factor

:

Particle flux

:

Molar mass

:

Mass density

:

Resistance

[F]

:

Capacitance

[H]

:

Inductance

𝑊

[m]

:

Length of electrode

𝐽

[A]

:

Current

𝜌𝑒

[Ω-m]

:

Resistivity

𝑈

[J]

:

Kinetic energy

𝐴𝑑

[m]

:

Vibration amplitude in driving mode

Ω

[Radian/s]

:

Rate of rotation

[1]

:

Quality factor in transverse direction

𝐿′

𝑄𝑦

AC

𝐶

CE

:

2

ACCEPTED MANUSCRIPT 𝜔𝑦

[Radian/s]

:

Frequency in transverse direction

𝛾𝐺

[1]

:

Gruneisen’s constant

:

Specific heat at constant pressure

𝐶𝑝

[J/

ℎ𝐶

[1]

:

Contact coefficient

[W/m-oK]

:

Thermal conductivity of bar A and B respectively

[m]

:

Thickness of bar A and B respectively

𝑘𝐴 , 𝑘𝐵 Δ𝑥𝐴 , Δ𝑥𝐵

Introduction

T

1.

oK]

AN

US

CR

IP

Designers, engineers and manufacturers are always in search of new and better materials for performance improvements and cost reduction of their products to remain competitive in the market. This requirement has resulted in the availability of thousands of material choices to the designers. It has been estimated that over 160,000 materials have been invented in the world [1]. In the same context, the initial limited material choices available to the microsystems / Micro-Electro-Mechanical-Systems (MEMS) designers have also expanded due to improvement and introduction of new microfabrication techniques and processes. It is claimed that virtually all materials that can be electroplated from solution are useable as MEMS materials [2]. Therefore, choosing the best suited material using the traditional design approach, such as those based on Design of Experiments (DOE), is extremely challenging, or even impossible. It is also likely that many materials with superior performance are missed out during the design cycle.

CE

PT

ED

M

To overcome this difficulty and to make the design process more effective and efficient, many systematic material selection methods have been proposed over the years. Jahan et al. [3] have listed 22 methods used by researchers for materials’ screening, comparing, choosing and optimization. Some notable methods include ‘Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS)’ [4], ‘Analytic Hierarchy Process (AHP)’ [5], ‘Elimination and Choice Expressing Reality (ELECTRE)’ [6], ‘Vise Kriterijumska Optimizacija Kompromisno Resenje’ i.e. multi-criteria optimization and compromise solution (VIKOR) [7], Fuzzy [8] and Ashby’s Method [9]. While these material selection techniques have been frequently applied to traditional macro-sized products and systems, microsystems / MEMS have so far received little attention. For comparison, out of all the material screening, comparing, selection and optimization methods reported in the literature in recent years, only four have been demonstrated for selection of materials for MEMS devices. Interestingly, out of these four techniques, Ashby’s method of material selection, originally developed for materials selection at macro-scale, has been applied most frequently for microsystems / MEMS devices. The other examples are the work of Chauhan et al. [10], Yazdani et al. [11] and Zha et al. [12].

AC

Chauhan et al. [10] compared Ashby’s method with TOPSIS and VIKOR for material selection for MEMS gyroscope and Capacitive Micromachined Ultrasonic Transducer (CMUT). Yazdani et al. [11] compared Ashby’s method with TOPSIS and VIKOR for material selection of MEMS electrostatic actuators, whereas Zha et al. [12] developed a Fuzzy decision process for MEMS materials and process selection. In the TOPSIS approach, a positive and negative ideal solution is identified. The attribute closest to the positive ideal and farthest from the negative ideal is the best compromised solution. This process is easy to implement and program, and the total number of steps involved in the process does not vary with the addition of any other attribute. However, the major drawback with this approach is that it does not consider the correlation between the attributes [13]. In a similar fashion to TOPSIS, the VIKOR method also relies on the closeness to an ideal solution but unlike TOPSIS, this method introduces a ranking index depending upon the measure of closeness from the ideal solution. Moreover, this method eliminates the unit of criterion functions by using linear normalization [14]. This method is very helpful in multi-criteria decision making if the designer is not clear about the preferences at the start of the design cycle. However, the major disadvantage of this method is that its accuracy is compromised during the normalization process [15]. The Fuzzy decision-making method takes imprecise and incomplete data to

3

ACCEPTED MANUSCRIPT reach a solution, which is helpful if the available data is limited. However, this approach is sometimes difficult to implement and requires a large amount of computational modelling before it can be implemented on a real world problem [16]. TOPSIS, VIKOR and Fuzzy decisions are primarily statistical methods, therefore, the inherent issues of statistical methods are embedded in them. These methods are applied before knowing any exact relationship between the material properties of the system under study, thus neglecting the effect of one property on the other [10]. However, in Ashby’s method, exact relationships between different materials’ properties involved in the design are established. This is achieved using Performance and Material Indices, which are derived specific to the part, device or component’s functional requirements, thus efficiently integrating the design needs in the material selection process. This makes the Ashby’s method simple, highly accurate, efficient and reliable.

2.

US

CR

IP

T

Having discussed the usefulness of Ashby’s methodology in relation to the other three techniques and the frequency with which Ashby’s methodology for materials selection for microsystems / MEMS is used, this paper will now present a review of all the MEMS materials selection studies published over last two decades that are based on Ashby’s method and summarize their key findings. The remaining paper is organized as follows: Section 2 of the article gives a brief review of the Ashby’s material selection methodology. Section 3 reviews applications of the Ashby’s method for material selection of microsystems / MEMS. In section 4, MEMS materials proposed for different applications have been summarized. A discussion and concluding remarks are presented in section 5. Ashby’s Material Selection Methodology

 Function

ED

M

AN

Ashby’s material selection approach is a well-established, tried and tested methodology, which falls in the category of Multi Objective Decision Making (MODM) techniques. In addition to its wide-spread and successful application for material selection in macro-scale systems such as trusses [17], bipolar plates [18], natural materials [19], wind turbine blades [20, 21], precision instruments [22] and hard coatings [23] etc., it is probably the only approach that has also been successfully applied for material selection in numerous microsystems / MEMS sensors, actuators and devices. In Ashby’s material selection approach [9], the performance of the product / system under consideration is calculated by the equation:-

  Material    Geometric ,  ,  requirements, F   requirements, G   requirements, M  

PT

P  

(1)

AC

CE

The function, geometry and material requirements in equation (1) are independent and separable, thus offering great simplification. For given ‘𝐹’ and ‘𝐺’ requirements, the performance can be optimized by just optimizing the material requirements. Based upon the functional and geometric requirements of the product, performance indices are derived in the form of an equation (1). The material part of the equation (1) i.e. material index, is then optimized using graphs / charts. The axes of these graphs are the material properties of the derived material index. Different stages of the Ashby’s material selection approach are also shown graphically in Figure 1.

4

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

3.

M

Figure 1. Materials Selection Stages of Ashby’s Methodology, taken from Ref [24]. Ashby’s Material Selection Method Applied to Microsystems / MEMS Devices

CE

PT

ED

Material selection case-studies, using Ashby’s approach, have been reported in the literature for a total of 20 key categories of MEMS devices. The performance indices have been developed for these devices and using the developed performance indices, material selection has been performed through comparison of material properties of different materials in the form of tables or charts. For the ease of comprehension, the 20 categories of MEMS devices having similar functional and structural requirements have been grouped into five classes of MEMS devices. These are MEMS sensors / transducers, MEMS actuators, micro-beams / flexures, RF-MEMS and others (micro-motors, micro-pumps, micro-filters, micro-turbines, micro- heat sinks, micro-switches, micro-springs, microphones and micro-gyroscopes).

AC

In each class, first the types of MEMS devices and their functional /structural requirements have been discussed. The reported performance indices to achieve the functional requirements of these MEMS devices have then been presented. From the reported performance indices, the corresponding material indices for each category of MEMS devices have been derived. Based upon the material indices, the governing material properties responsible for achieving optimum devices performance have been identified and analyzed. After having discussed the performance and their corresponding material indices for five different classes of MEMS devices, the proposed materials along with different methods used to select these materials have been discussed. The strengths and weaknesses of material selection methods and the proposed materials have been highlighted. Solutions to overcome the identified shortcomings of these methods have also been proposed and demonstrated using case studies. Tables 1-5 summarize the performance and material indices reported for five different classes of MEMS devices during last two decades, each of which will be discussed separately in the subsequent paragraphs.

5

ACCEPTED MANUSCRIPT 3.1

MEMS Sensors / Transducers

US

CR

Applied Pressure

IP

T

In this category, the MEMS sensors and transducers used for measuring pressure and force [2, 25-27], detecting ultrasonic signals (ultrasonic transducers) [28] and mercury vapor, moisture or volatile mercaptans [25] are combined. The reason for grouping these sensors together is the fact that all these sensors / transducers use diaphragm deflection for transduction. A schematic of a membrane / diaphragm-based MEMS pressure sensor is shown in Figure 2.There are only two exceptions [25, 27] which use cantilever deflection for sensing and transduction, therefore the main mechanical part involved in these MEMS sensors’ / transducers’ design is their diaphragm or cantilever beam. Maximizing the diaphragm / cantileverbeam deflection and its vibrating frequency are the two design goals to be achieved for all these sensors and thus the Performance and Material Indices for these two parameters were formulated from this. By using an appropriate material index involving three material properties, 𝝈𝒇 , 𝑬 and 𝝆 , as suggested by the formulated performance indices, optimized diaphragm design for maximum deflection and highest frequency can be achieved. The Performance and Material Indices reported for these sensors / transducers are summarized in Table 1.

Cavity

AN

Diaphragm

M

Reference Pressure

ED

Figure 2. Schematic diagram showing the cross section of a typical membrane / diaphragm based pressure sensor

MEMS Actuators

AC

3.2

CE

PT

It is worth mentioning that mechanical structures also dissipate energy during their vibrations and these dissipations are required to be minimized for an efficient design. This can be achieved by minimizing the intrinsic loss coefficient, as reported in [27]. Moreover, ultrasonic transducers may also be required to operate at high temperatures resulting in a changed frequency response due to thermal strain. Therefore, an additional performance index (thermal stability) has also been studied for such devices. The corresponding material properties that define the thermal stability of the device are thermal conductivity and coefficient of thermal expansion. The material index for thermal stability has been reported by [28] for MEMS CMUTs (Capacitive Micromachined Ultrasonic Transducers).

Micro actuators are used in a variety of applications such as micro-grippers [29], relays [30], switches [31] and precision positioning [31, 32]. Micro actuation can be achieved using different actuation principles. Material selection has been reported for micro actuators working on three different principles: electrostatic actuation [33, 34], thermal actuation [28] and Bi-material Piezo-Electric (BPE) actuation [35]. A typical MEMS electrostatic actuator consists of two parallel plates, as illustrated in Figure 3, with an actuating electrode and a fixed electrode. The movement of the actuating electrode is controlled by the electrostatic force developed between the two electrodes. The important design parameters for electrostatic actuation are actuation voltage, actuation speed (directly related to natural frequency of vibration) and actuation displacement / stroke. However, the requirement to maximize or minimize these parameters depends upon the target application. For example, applications such as displays and switches require low actuation voltage, whereas for applications requiring large forces (e.g. stepper motors [36]), a high actuation voltage is essential [33]. By adjusting the contributions of the three material properties, 𝝈𝒇 , 𝑬 and 𝝆 , performance parameters can be either maximized or minimized. The exact relationships for maximizing or minimizing these material properties are given in Table 2.

6

CR

IP

T

ACCEPTED MANUSCRIPT

US

Figure 3. Electrostatic actuator schematic: (a) lumped mass representation, (b) cross section representation of beam electrostatic actuator. Figure has been taken from Ref [33].

AC

CE

PT

ED

M

AN

The simple structure of a thermal actuator consists of a single beam anchored with the substrate at both ends. The applied voltage between the two anchor points, as shown in Figure 4, causes ohmic (or Joule) heating and an expansion of the beam. This results in beam buckling and in-plane movement at its mid-point. For higher reliability of such structures, thermal distortion governed by the material’s thermal conductivity and coefficient of thermal expansion [28, 37, 38] is required to be minimized. Accordingly, the index for achieving minimum thermal distortion, listed in Table 2, requires a material with higher thermal conductivity and lower coefficient of thermal expansion.

Figure 4. A single beam thermal actuator. Beam is anchored at its two ends. The applied voltage across the anchor point causes joule heating and in-plane movement of the beam. Figure has been taken from Ref [39]. A typical BimaterialPiezo-Electric (BPE) micro-actuator, as shown in Figure 5, consists of two layers; a piezoelectric material layer and a supporting substrate material layer. The piezoelectric effect in the piezoelectric material layer causes the generation of mechanical stress, which results in the actuation. The important performance parameters for BPE micro-actuators are actuation frequency, actuation efficiency, required voltage for actuation (voltage index), and the loss coefficient. Performance and Materials Index for these parameters are given in Table 2. The effect of actuation frequency on the material selection is negligible as the density of the two materials considered for BPE are very close to

7

ACCEPTED MANUSCRIPT

IP

T

each other. Required actuation voltage can be optimized by optimizing the thickness ratio of the two materials, while the loss coefficient can be controlled through material stiffness.

3.3

CR

Figure 5. Schematic of a cantilever type Bimaterial Piezo-Electric (BPE) actuator. Figure has been taken from Ref [35]. Micro-Beams / Flexures

AC

CE

PT

ED

M

AN

US

Micro-beams are integral parts of many MEMS sensors and actuators [40]. They can also be used as independent sensors; one such example is the inertial sensor shown in Figure 6. Failure by shock-induced stiction is a common issue for micro-beams but surface coatings and enhanced surface roughness have provided a practical solution to this problem. These structures are also required to resist the inertial loads resulting from sudden shocks and therefore must have high stiffness to reduce the deformation and low density to lower the inertial loads. In order to fulfill all these requirements for microbeams, Performance and Material Indices developed and reported in literature [25, 27] have been summarized in Table 3. The ratio of the material properties 𝑬 and 𝝆 can be adjusted to achieve the desired design for specific applications.

Beams contacting the ground plane

Figure 6. Inertial sensor failure due to stiction. Figure has been taken from Ref [41]. Micromachined flexures are also prevalent in many MEMS sensors and actuators [27]. The performance requirement for these flexures (used in place of hinges or other bearings) is to have a maximum displacement without failure, with minimum force application. This material requirement can be achieved by maximizing the ratio of material properties 𝝈𝒇 and 𝑬. 3.4

RF-MEMS

8

ACCEPTED MANUSCRIPT RF MEMS have numerous components (e.g. bridges, switches, switched capacitors, varactors, resonators, and oscillators) and actuation methods (e.g. electro-thermal, electrostatic, piezoelectric and magneto-static). Four cases [42-45] of RF MEMS materials selection using Ashby’s methodology have been reported in literature. All four are based on an electrostatic actuation mechanism. Performance and Material Indices have been reported for RF bridge, RF switch and RF varicap (variable capacitor). The performance parameters governing these components are intrinsic residual stresses induced during deposition process of RF thin films, pull-in voltage (for bridge and switches), available capacitance and quality factor.

CE

PT

ED

M

AN

US

CR

IP

T

In RF-MEMS, while minimizing intrinsic residual stresses, the quality factor and available capacitance are required to be maximized. However, the value of the required pull-in voltage is application dependent. This can be explained by considering a normal design of a MEMS RF switch / varicap, which consists of two parallel plates that form a capacitor. The top plate is mobile, while the bottom plate is fixed. When a DC voltage is applied between parallel plates, an electrostatic force is generated. When this applied voltage reaches a specific value, called pull-in voltage, the mobile plate collapses on the fixed plate. This is not required for an RF varicap, as illustrated in Figure 7; however, this is the normal function of an RF switch (Figure 8) [43]. Material properties,𝝈𝒇 , 𝑬, 𝝆,𝝂 and 𝒌 can be used to achieve the desired maximum or minimum of different RF MEMS devices, as summarized in Table 4.

AC

Figure7. RF varicap in three different states : (a) initial state, (b) up state, (c) downstate. In down state, mechanical stoppers has been added to avoid collision of top plate with the bottom plate. Figure has been taken from Ref [46].

9

ACCEPTED MANUSCRIPT Figure 8. RF switch in two different states (on state and off state). Figure has been taken from Ref [47].

3.5

Other MEMS / Micro Devices

IP

T

In this category, material selection for MEMS devices such as gyroscopes [2, 28], micro-motors / micro-pumps / micro-turbine [2, 25, 28], micro-filters [27], micro-electro heat sinks [48] and micro-phones, resonators, accelerometers, switches [28] has been reported. A gyroscope (as shown in Figure 9) consists of one or two proof masses, which are driven with in-plane vibration. When a rotational rate is experienced by the gyroscope, it produces an out of plane vibration, which is used to detect the angular rate. The liquid and gas management at micro level is achieved by using micro-motors and micropumps. The active structural element for these is a rotating disc as shown in Fig 10. The micro-turbines, on the other hand, are used for energy generation at micro-level. Micro-filters are potential candidates for use in communication circuits for filtering out selective frequencies. Heat sinks are used for the most cost effective thermal management of MEMS circuits and devices.

CE

PT

ED

M

AN

US

CR

The performance parameters for each of these micro-devices / microsystems are different. Natural frequency is important for gyroscopes, resonators and filters, energy is an important parameter for rotating disc of micro-motors / micro-pumps / micro-turbines, whereas deflection is the critical parameter for microphones and switches. Performance and Material Indices for all these devices are summarized in Table 5.

AC

Figure 9. Schematic diagram of a MEMS tuning fork gyroscope. Figure has been taken from Ref [28]

(a)

(b)

Figure 10. Rotating Disc for (a) mirco motor, (b) micro pump. Figure has been taken from Ref [25].

10

ACCEPTED MANUSCRIPT

4.

Review of Materials Proposed for MEMS Applications

Materials for MEMS Sensors / Transducers

US

4.1

CR

IP

T

Ashby’s material selection methodology has four steps;(1) formulation of performance indices of the element under consideration, (2) extracting material indices from the performance indices, (3) plotting material indices to generate material selection charts and finally, (4) material selection. The real advantage of Ashby’s methodology is its integration with the CES software, which has an updated and accurate database of almost 3000 available materials. While reviewing the application of Ashby’s material selection methodology for MEMS / microsystems (Tables 1- 5) it became evident that steps 1 and 2 of Ashby’s methodology have been applied fully for most of the cases. However, steps 3 and 4 have not been implementedfully in many cases. It is worth highlighting that the material selection for different MEMS devices / applications reported so far is generally based on the performance comparison of very few materials (less than 30 materials in most cases). There are however, two exceptions. First is the material selection for RF-MEMS [42, 43], where a material database of 167 materials has been used. The second exception is material selection for MEMS heat sinks [48], where CES-generated material selection charts have been used. Keeping this fact and background in mind, the materials proposed for different MEMS devices and systems will now be discussed in the subsequent paragraphs.

Materials for two types of MEMS sensors / transducers (i.e. diaphragm and cantilever based) have been reported (S.No 1, Table 6). The performance parameters for diaphragm based pressure sensors are diaphragm deflection and its vibration frequency. To achieve large diaphragm deflection,

AN

𝟑⁄

M

materials with higher values of 𝝈𝒇 𝟐 and a lower value of 𝑬 are desirable. The material selection for diaphragm based pressure sensors, reported in literature is primarily based upon the material properties comparison of very few materials (maximum 09) in a tabular form. The dominant material proposed, based upon this comparison, is silicon. It is pertinent to highlight that materials better than silicon could become prevalent if the MEMS material data base is expanded.

4.2

AC

CE

PT

ED

The key performance parameter for a cantilever based force sensor is its sensitivity (minimum 𝟏 𝟏 detectable force), which is function of (𝑬𝝆) ⁄𝟒 and 𝝌 (loss coefficient). For higher sensitivity, both (𝑬𝝆) ⁄𝟒 and 𝝌 are required to be minimized. The initial choices of materials for a cantilever based force sensor from Figure 11 are silicon oxide, quartz, silicon, gallium arsenide, and silicon nitride. However, under ambient condition, the loss coefficient is largely determined by the extrinsic loss, which makes it independent of material properties. In such cases, the effective material requirement is reduced to just 𝟏 having a minimum value of index (𝑬𝝆) ⁄𝟒 . Figure 11 then suggests that polymers are also an attractive choice [27]. Fabrication of polymer probe for scanning force microscopy by Genolet et al. [49] is consistent with these findings.

Materials for MEMS Actuators

Candidate materials for two types of micro-actuators; electrostatic and BPE actuators have been reported. Since the material requirements for the micro-actuators are application dependent, materials for three different applications of electrostatic micro-actuators and two applications of BPE actuators have been discussed below. 4.2.1

Materials for high speed and high force electrostatic actuation

The speed of actuation is related to frequency of vibration of the actuator and is governed by the material index √𝑬⁄𝝆, whereas actuation force is related to the material Young’s modulus 𝑬. In order to achieve high speed and high force actuation, both the material indices √𝑬⁄𝝆 and 𝑬 are required to be maximized. Figure 12 is the plot of these two material indices and it is evident that diamond is the best

11

ACCEPTED MANUSCRIPT

AN

US

CR

𝛘

IP

𝟏⁄ 𝟐

T

choice in this case. After diamond comes silicon carbide, silicon nitride, alumina, titanium carbide and silicon as possible materials.

𝟏⁄ 𝟒 (𝐏𝐚

(𝐄𝛒)

− 𝐊𝐠 − 𝐦−𝟑 )

𝟏⁄ 𝟒

PT CE AC

𝐄 (𝐆𝐩𝐚)

ED

M

Figure11. Material selection chart for cantilever based force sensor. Figure has been taken fromRef [27].

√𝐄⁄𝛒 √𝐆𝐏𝐚⁄ 𝐠𝐜𝐦𝟑 Figure 12. Material selection chart for electrostatic actuators requiring high speed and high force actuation. Figure has been taken from Ref [11]. 4.2.2

Materials for high speed and low voltage electrostatic actuation

12

ACCEPTED MANUSCRIPT

ED

M

AN

US

CR

IP

T

A high speed of actuation requires a high value of √𝑬⁄𝝆, whereas low actuation voltage requires a low value of √𝑬. This is a case of conflicting requirements of material indices because a high actuation speed will also result in a high actuation voltage. Such conflicting requirements can be handled by drawing a trade-off line as shown in Figure 13 [33]. The materials falling on this line or close to this line are the best-suited materials. Silicon oxide, quartz and silicon are thus the candidate materials for moderately high speed and low voltage applications.

4.2.3

PT

Figure 13. Material selection chart for electrostatic actuators requiring high speed and low actuation voltage. Figure has been taken from Ref [33]. Materials for large displacement and low voltage electrostatic actuation

4.2.4

AC

CE

Large displacement actuation requires higher values of material index 𝝈𝒇 ⁄𝑬 (maximum value of 𝝈𝒇 and minimum value of 𝑬), whereas low voltage actuation requires a low value of material index √𝑬. Figure 14 [11] is the plot of these material indices, where 𝝈𝒇 ⁄𝑬 has been plotted on x-axis and √𝑬 has been plotted on the y-axis. It is evident from Figure 14 that polymers are an attractive choice for high displacement and low voltage actuation. Amongst the polymers, PMMA is optimal followed by PVDF and Polyimide. However, linearity and hysteresis may be an issue in the mechanical response of polymers. Materials for Bimaterial Piezo-Electric (BPE) actuators

Piezoelectric and substrate materials for two different applications, high force / work and high frequency, within BPE has been proposed in the literature [35]. Figure 15 is a plot of a range of piezoelectric materials on two different substrate materials; Si (Silicon) and DLC (Diamond Like Carbon). It is evident from Figure 15 that when used on Si and DLC substrate, ferroelectric piezo-ceramics (Rochelle salt (RSAL) and PZT) are superior when compared with the other materials. Due to the unstable nature and low transition temperature, Rochella salt is unlikely to be suitable for micro-actuators. Quartz, which is traditionally used for sensors and macro-actuators is not suitable for MEMS actuators due to its low piezoelectric constant [35]. From Figure 15, the best material choices for high force / work actuation are PZT (Lead Zirconate Titanate), PMNT (Lead Magnesium Niobate-Lead Titanate), PZNPT

13

ACCEPTED MANUSCRIPT

US

𝛔𝐟 ⁄𝐄 (𝟏𝟎−𝟑 )

CR

IP

T

√𝐄(𝐆𝐏𝐚)𝟎.𝟓

(Lead Zinc Niobate-Lead Titanate), PMNPT (Lead Magnesium Niobate - Lead Titanate) and BaTiO3 (Barium Titanate) in combination of both Si and DLC substrate.

AC

CE

PT

ED

M

AN

Figure 14. Material selection chart for electrostatic actuators requiring large displacement and low actuation voltage. Figure has been taken from Ref [11].

14

US

CR

IP

T

ACCEPTED MANUSCRIPT

AN

Figure 15. Contours of equal work per volume ( 𝑙𝑜𝑔10 (𝑊𝑛𝑜 )), blocked moment (𝑙𝑜𝑔10 (𝑀𝑛𝑜 )) and tip slope (𝑙𝑜𝑔10 (Θ𝑛𝑜 ) for a range of piezoelectric materials on (a) Si, (b) DLC. The Figure has been taken from Ref [35].

AC

CE

PT

ED

M

The material selection chart for selecting candidate materials for high frequency applications is given in Figure 16 [35]. AIN on Si or DLC substrate are superior choices compared with the rest of the piezoelectric materials for this type of application. A combination of piezoelectric materials on polymer substrates however has lower performance.

Figure 16. Contours of equal actuation frequency(𝑙𝑜𝑔10 (𝑓)) on Si, DLC and PMMA substrate for different piezoelectric materials. Figure has been taken from Ref [35].

15

ACCEPTED MANUSCRIPT 4.3

Materials for Micro-Beams / Flexures

PT

ED

M

AN

US

CR

IP

T

Micro-beams are prone to shock induced stiction during the loading event, therefore they should have high stiffness and low inertia. The material property governing the stiffness is the Young’s modulus, whereas inertia is controlled by density. In order to resist shock induced stiction, materials with a higher value of Young’s modulus and a lower value of density are the candidate materials. Maximizing material index 𝑬⁄𝝆 satisfies this requirement. This material index has been plotted in Figure 17 [27], where 𝝆 is plotted on the x-axis and 𝑬 on y-axis. The candidate materials from Figure 17 are diamond, silicon carbide, alumina, silicon nitride and silicon.

CE

Figure 17. Plot of material Young’s modulus versus density for selecting optimum material for microbeams. Figure has been taken from Ref [27].

AC

The requirement of the flexures is to produce a large displacement with minimum application of force. They are also required to produce large displacement without fracturing. The material requirement of flexures is to have minimum 𝑬 and maximum 𝝈𝒇 ⁄𝑬. These two material indices have been plotted in Figure 18 [27]. It is evident from Figure 18 that polymers are the most attractive choice for applications involving maximum flexure.

16

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Materials for RF-MEMS

ED

4.4

M

Figure 18. Plot of 𝜎𝑓 ⁄𝐸 versus 𝐸for selecting optimum material for micro-flexures. Figure has been taken from [27].

4.4.1

PT

Materials for three different requirements (S. No. 4, Table 6); (a) decreasing the intrinsic stresses during deposition process of RF thin films, (b) high pull in voltage and high quality factor, and (c) low pull in voltage and high quality factor, are proposed in the literature. The proposed materials for each requirement are discussed below. Materials for low intrinsic stresses

AC

CE

Generally, the residual stresses are developed during the deposition process. These stresses come in the three forms: intrinsic, thermal and external. Thermal residual stresses are developed during heating / cooling cycles, whereas the sources of external residual stresses are oxidation and impurities. The intrinsic residual stresses are the only one which deals with the material properties of the deposited layer. The material index to minimize the intrinsic stresses is given in Table 4. According to the reported material index, materials with minimum value of 𝜶𝑬𝑻𝒎𝒆𝒍𝒕𝒊𝒏𝒈 ⁄(𝟏 − 𝝊) and maximum value of thermal conductivity 𝒌 result in low intrinsic stress during the deposition process of RF thin films. The candidate materials, from the database containing 167 materials, qualifying this criterion are rubidium, boron nitride, potassium, silver, magnesium, gold, copper and aluminum. Rubidium, potassium and magnesium are highly reactive materials, which makes them unsuitable for RF thin film deposition [43]. Therefore, the final choice of materials for decreasing intrinsic stresses during thin film deposition is reduced to boron nitride, silver, copper, aluminum and gold. 4.4.2

Materials for high pull-in voltage and high quality factor

A high pull-in voltage and high quality factor are the requirements for RF-varicap [43]. By maximizing material index √𝑬𝝆−𝟏 , suitable materials for high pull-in voltage and high quality factor can be chosen. The materials qualifying this criterion are copper, silver, iridium, gold, rhodium, tungsten and

17

ACCEPTED MANUSCRIPT molybdenum. However, if low intrinsic stress in the deposited film is also the requirement, in addition to the high pull-in voltage and high quality factor, then the choice of material is reduced to Copper, Silver and Gold. 4.4.3

Materials for low pull-in voltage and high quality factor

Materials for other MEMS devices

CR

4.5

IP

T

A low pull in voltage and high quality factor are the requirements for RF-switches. By maximizing material index √𝑬/𝝆−𝟏 , suitable materials for low pull-in voltage and high quality factor can be chosen. The materials qualifying this criterion are gold, copper, magnesium, calcium, sodium, silver, potassium and rubidium [43-45]. Again, if low intrinsic stress in the deposited film is required in addition to a low pullin voltage and high quality factor, then the choice of material is reduced to Copper, Silver and Gold. It is pertinent to highlight that Copper, Silver and Gold qualify as candidate materials for all three considered material requirements of RF-MEMS. However, Copper and Silver are sensitive to oxygen and humidity, which leaves Gold as the most suitable bridge material for RF MEMS application. This has been theoretically and experimentally demonstrated by Guisbiers et al. [44] by depositing a Gold bridge for RF MEMS applications.

Materials for pumps / turbomachinery

AN

4.5.1

US

In this category, material selection for pumps / turbomachinery, micromachined filters, gyroscopes and micro-electronic heat sinks from literature are reported. Proposed materials for each device category are discussed below.

Materials for micromachined filters

PT

4.5.2

ED

M

The main structure for pumps / turbomachinery is the rotating disc as shown in Figure 10 [2, 25]. Maximizing the total kinetic energy stored / mass in this rotating disc is the functional requirement for optimum performance. This is achieved by maximizing material index 𝝈𝒇 ⁄𝝆. Two authors [2, 25] have reported suitable materials for pumps / turbomachinery. They compared the material properties of qty 08 and 09 materials in a tabular form, respectively. The material proposed for the optimum performance of pump / turbomachinery by both studies [2, 25] is silicon. However, since the materials database considered for optimum material selection in their studies is very small, the option of exploring other / new materials for these applications remains open.

Materials for gyroscopes

AC

4.5.3

CE

The main requirement for the micromachined filter is to operate at a higher frequency to ensure the selected signal of interest can be matched. This is achieved by maximizing the material index √𝑬⁄𝝆. Figure 17 above [27] can be used to make the initial selection of the materials for the micromachined filters. It is evident that diamond is the optimal material followed by silicon carbide, alumina and silicon nitride. Diamond has also been proposed by Kohn et al. [50] and Wang et al. [51] for such applications.

The functional requirement of gyroscopes is to have maximum sensitivity, which is related to the amount of out of plane deflection for a given rotational rate and the maximum achievable deflection. The amount of out of plane deflection for a given rotational rate is governed by 𝒌⁄(𝑬𝜶)𝟐 , while maximum achievable deflection is given by material index 𝝈𝒇 ⁄𝑬. These two material indices have been plotted in Figure 19 [28]. The materials falling in the upper right corner of the plot are the candidate materials and include diamond, silicon nitride, silicon carbide, poly-silicon and silicon oxide.

18

US

CR

IP

T

𝐤⁄(𝐄𝛂)𝟐

ACCEPTED MANUSCRIPT

𝝈𝒇 ⁄𝑬

4.5.4

AN

Figure 19. Plot of 𝜎𝑓 ⁄𝐸 versus 𝑘⁄(𝐸𝛼)2 for selecting optimum materials for MEMS gyroscopes. Figure has been taken from [28]. Materials for micro-electronic heat sinks

PT

ED

M

Micro-electronic heat sinks have various functional requirements. It should be a good electrical insulator (i.e. having maximum electrical resistivity) to prevent stray capacitance and electrical coupling between the microchip and heat sink. At the same time, it should be a good thermal conductor (i.e. having maximum thermal conductivity) so that heat from microchip is carried away efficiently. The thermal stresses during heating cycles and the contact resistance between heat sink and the microchip also need to be minimized. Material needs for the first two requirements are fulfilled by the CES-generated material selection chart shown in Figure 20 [48]. The materials falling at the top right corner of the figure have the highest values of electrical resistivity and thermal conductivity. It is evident from the figure that aluminum nitride and aluminum oxide are the candidate materials fulfilling the chosen requirements.

AC

CE

In order to reduce the thermal stresses, materials that have a larger value of Young’s modulus and whose convective heat transfer coefficient (h) increases with increasing temperature are the candidate materials. The proposed materials satisfying these requirements are aluminum nitride, aluminum oxide and silicon nitride. Aluminum and its alloys are also good for lowering the contact resistance [48]. Therefore, aluminum and its alloys are best overall candidate materials for MEMS based heat sink applications. However, it should be noted that this selection is based upon the bulk material properties within the CES software due to the database containing only bulk material properties.

19

US

CR

IP

T

ACCEPTED MANUSCRIPT

Discussion

M

5.

AN

Figure 20. Plot of electrical resistivity versus thermal conductivity for selecting optimum material for microelectronic heat sinks. Figure has been taken from Ref [48].

PT

ED

The review of MEMS material selection studies based on Ashby’s methodology presented in preceding sections reveals that so far different researchers have reported material selection studies for a total of 20 key categories of MEMS devices. Based upon the Performance and Material Indices derived for these MEMS devices (Tables1 – 5) many authors have suggested the most suitable materials (Table 6) for these MEMS devices.

AC

CE

It is worth noting, however, that the materials selection performed in these studies is based primarily on the performance parameters of these MEMS devices. However, other factors such as reliability, cost (and its relationship associated with manufacturing volume) and yield are also critical for MEMS design and must be taken into account. Moreover, the material selected in most of these studies is based upon the manual comparison of material properties of a limited set of materials, and in this process the complete material universe available to MEMS designers / engineers is not explored. These manual methods based upon the data of a limited set of materials become inadequate for large material data sets and for customized material selection targeting, for example if the designer wants only eco- friendly materials or materials with a cost or weight below a certain threshold. These limitations can be overcome by developing (or using) a suitable material selection software for MEMS materials selection, which should be accompanied by a reasonably comprehensive MEMScompatible or MEMS-specific materials database. Among a number of general databases and software made available for materials selection by different organizations, the Cambridge Engineering Selector (CES) stands out as the most powerful and versatile option [1]. It is an industrial standard tool, which provides graphical analysis of material properties. A number of materials databases are embedded within CES, which can be used with its interactive graphing and comparison tool, enabling smart material selection decisions. Other important requirements such as cost minimization, eco-friendliness, medical compatibility, etc., can also be fulfilled whilst selecting the materials that best-match the design / performance requirements for the considered applications.

20

ACCEPTED MANUSCRIPT It is equally important to point out that despite having all the above strengths and versatilities, CES in its current state has one limitation in the context of microsystems / MEMS devices; it lacks availability of data on MEMS-compatible or MEMS-specific materials with micro-scale properties. Since the CES software was originally developed for material selection and optimization for macro-systems [1722], its database mainly contains bulk material properties, which are often different from the micro-scale properties. As an example, the reported micro-scale tensile strength of silicon is 4000 MPa [2, 25], whereas the tensile strength for silicon used in CES is 165 – 180 MPA. Similarly, the reported micro-scale tensile strength of silicon oxide is 1000 MPA [11, 27], whereas that used in CES is 45.7–50.4MPa. Moreover, many extensively used and reported MEMS materials, e.g. silicon (110), silicon (111), polysilicon, SU-8, parylene, gallium arsenide and gallium nitride etc. are not included in the material database of CES software.

CR

IP

T

This limitation of CES software can be addressed by adding micro-scale material properties for the MEMS compatible materials in its main database and / or by developing a separate MEMS material database to be used as an extension module with the software. One good example in this context is that of COMSOL multi-physics [52], which has been upgraded by incorporating a MEMS module to keep pace with this emerging technology.

US

To address this limitation of CES, we have developed a MEMS Materials Database, which consolidates micro-scale material properties for most of the MEMS materials reported in the literature. To elucidate the disparity between bulk material properties (typically used in CES software) and the microscale material properties (reported for MEMS materials), both have been put side-by-side and presented in Table 7. Based upon these micro-scale material properties, a dedicated MEMS Materials Database has also been developed as an Add-on Module for CES software and integrated with it.

ED

M

AN

To illustrate the usefulness of MEMS-compatible materials data versus the bulk materials properties provided by the CES, material selection charts using bulk and MEMS (micro-scale) material properties have been developed and compared for selection of materials with conflicting requirements (maximizing diaphragm deflection andnatural frequency of vibration simultaneously) of a MEMS pressure sensor. The material selection chart generated using CES default database (bulk material properties) is shown in Figure 21 while that using MEMS-compatible materials database is shown in Figure 22. These charts have been divided into four sectors, taking quartz as the reference material at point ‘O’. The materials in sector A are better than quartz in terms of maximizing diaphragm deflection and natural frequency of vibration simultaneously, whereas those in section B, C and D are inferior to quartz.

AC

CE

PT

While considering the bulk material properties (Figure 21), only two materials (diamond and chromium) emerge as better material than the quartz, whereas while considering micro-scale material properties (Figure 22), there is a much larger set of candidate materials that are better than quartz. At the same time it is evident that a number of MEMS materials are missed out while considering the bulk properties of the materials included in the CES database. An example to this effect is silicon (110), which is one of the key candidate materials for this application but has been missed out altogether from the material selection chart based upon the bulk material properties (Figure 21). Furthermore, several materials are placed at incorrect locations on the materials comparison chart, leading to the selection of inappropriate materials due to use of currently available bulk material properties database for selection of materials for MEMS devices. Silicon nitride (SiN), which is erroneously shown to be inferior to quartz when using CES database with bulk material properties (Figure 21) rightly appears superior to the quartz when considering the newly developed MEMS-compatible / specific materials database (Figure 22). As demonstrated by the above case study, at the start of any MEMS device design, the developed micro-scale material properties database used with the developed performance indices can give a reasonably good estimate of candidate materials. This database can be used for the material optimization of 20 different categories of MEMS devices for which the performance indices have already been established. Moreover, using advanced features of CES software, material selection for customized requirements such as materials for specific temperature range or any other specific operating environmental condition can also be selected very easily. Although, performance indices for a variety of MEMS devices have been developed but these do not cover the whole range of MEMS devices and need to be expanded in future. For example, development of the performance indices for the MEMS flow / shear stress sensors and MEMS temperature sensors are such two cases. Similarly, the MEMS material

21

ACCEPTED MANUSCRIPT database developed in this paper contains the material properties of 51 MEMS materials, which can also be expanded on availability of new MEMS materials in future. Most of the developed performance indices are of preliminary nature and inclusion of more specific design parameters in the performance indices can provide even better material choices. For example, in case of MEMS pressure sensor, the derived performance indices focused on maximizing the diaphragm deflection and its natural frequency. Instead of focusing on diaphragm deflection, going for a more specific performance attribute such as device sensitivity, burst pressure, package stresses and signal to noise ratio will provide more meaningful material choices.

AlO

SiC

Sector C

SiO2

Ceramics

Metals and alloys

Al

100

T

Sector D

SiN

O

Ag

IP

Si

Cu PVDF

M1,0/M1

PMMA

Plastics

Ti Diamond

W Polyimide

AN

0.1

Silicone Rubber)

US

Quartz

1

Elastomers

CR

Cr

10

Sector A 1

Sector B 10

100

M2,0/M2

Metals and alloys

Sector D

W

Cr SiCN

Cu

SiO

Poly-Si 1

DLC Diamond 3H-SiC Si

Polymers

Parylene PVDF

PMMA

O

Ni-Fe Alloy PDMS

Ge

Quartz Silicone Rubber

AC

SiN

Au

Ti

CE

M1,0/M1

UNCD

0.1

Si(100)

0.01

PT

AlO

Polyimide

SU-8

Al

SiC

Sector C

Ag

Ceramics

10

ED

M

Figure 21. CES generated material selection chart based upon CES default (bulk material properties) materials database for material selection against conflicting requirements of MEMS pressure sensor

Si(110)

Sector A 0.1

Sector B 1

10

100

1000

M2,0/M2

. Figure 22. CES generated material selection chart based upon MEMS-compatible materials database for material selection against conflicting requirement of MEMS pressure sensor 6.

Conclusion

In this paper, different material selection approaches for microsystems / MEMS devices have been reviewed. It has been established that with few exceptions, Ashby’s material selection methodology,

22

ACCEPTED MANUSCRIPT due to its simplicity, reliability, versatility and accuracy, remains the most widely and extensively used technique for microsystems / MEMS materials selection to-date. As a reference for the global community of MEMS designers, engineers, and scientists, all the performance and material indices reported in literature for 20 different categories of microsystems / MEMS during last two decades have been collated and presented (Tables 1-5). Moreover, the most suitable materials suggested in the literature for different types / categories of MEMS devices and applications have also been presented (Table 6).

IP

T

This paper also highlights both the unique strengths as well as a limitation of Ashby’s materials selection methodology, when applied by using the CES software for selecting the best suited materials for MEMS. Since micro-scale properties of materials commonly used for micro-fabricationdo not match those of the bulk materials, a separate database having micro-scale properties of materials used for MEMS sensors, actuators and devices was essential. To address this limitation, a comprehensive database containing micro-scale properties of MEMS materials has been developed (Table 7).

US

CR

To demonstrate the advantage of using the developed MEMS specific database versus the previously available database in CES software (with bulk material properties), a case study on the selection of materials for conflicting performance requirements of MEMS pressure sensors has also been presented. Our selection showed that certain MEMS material (e.g. SiN and Si (110)) were incorrectly assessed and therefore misplaced on the materials selection charts while using the general CES database. These application-specific superior materials correctly emerged as the best candidate materials using the MEMS database extension.

ED

M

AN

It is therefore concluded that Ashby’s method has remained and will continue to remain the preferred technique for selection of materials for different microsystems / MEMS devices. This work allows Ashby’s material selection methodology to be fully exploited for MEMS design by using the MEMS-compatible materials’ properties (Table 7) along with the relevant Performance Indices and Material Indices for different categories of microsystems / MEMS devices (Tables 1-5).Using the Performance Indices, Material Indices and materials suggested by different researchers for different types of MEMS devices (Table 6) alongside the micro-scale properties of MEMS compatible materials presented in this paper, it will serve as a reference and a useful resource for researchers, engineers and scientists engaged in materials based design optimization of various microsystems and MEMS devices.

PT

Acknowledgement

CE

This work was supported by British Council (BC) and Higher Education Commission (HEC), Pakistan grant No KEP-031 awarded to Dr Ibraheem Haneef and Prof Florin Udrea under BC-HEC Knowledge Economy Partnership (KEP) Programme.

AC

The authors also very gratefully acknowledge the extremely useful comments and suggestions given by Prof Mark Spearing [Vice President (Research and Enterprise) and Professor of Engineering Materials, Faculty of Engineering & Environment, University of Southampton, UK] and Prof Michael F. Ashby [Emeritus Professor of Materials, Department of Engineering, University of Cambridge, Trumpington Street, Cambridge CB2 1PZ, UK], which greatly helped in improving the research work presented in this paper. The authors are very grateful to Mr Ethan Gardner, Graduate Student, University of Cambridge, UK for his help with a diligent review of the final manuscript for improving its grammar, syntax and overall style.

23

ACCEPTED MANUSCRIPT Authors’ Biographies Florin Udrea is a professor in semiconductor engineering and head of the High Voltage Microelectronics and Sensors Laboratory at University of Cambridge, UK. He received his BSc from University of Bucharest, Romania in 1991, an MSc in smart sensors from the University of Warwick, UK, in 1992 and the PhD degree in power devices from the University of Cambridge, UK, in 1995. Since October 1998, Prof. Florin Udrea has been an academic withthe Department of Engineering, University of Cambridge, UK. Between August 1998 and July 2003 he was an advanced EPSRC Research Fellow and prior to this, a College Fellow in Girton College, University of Cambridge, UK.

CR

IP

T

He is currently leading a research group in power semiconductor devices and solid-state sensors that has won an international reputation during the last 20 years. Prof. Udrea has published over 350 papers in refereed journals and international conferences. He holds 70 patents with 20 more patent applications in power semiconductor devices and sensors.

US

Prof. Florin Udrea co-founded four companies; Cambridge Semiconductor (Camsemi) in power ICs, Cambridge CMOS Sensors (CCS) in the field of smart sensors, Cambridge Microelectronics in power devices and Flusso in flow sensors. Prof. Florin Udrea is a Board Member of Cambridge Enterprise Ltd. He has received several awards from The Institute of Electrical and Electronics Engineers (IEEE) and The Liverhulme Trust. For his ‘outstanding personal contribution to British Engineering’ he has been awarded the Silver Medal from the Royal Academy of Engineering.

AC

CE

PT

ED

M

AN

Ibraheem Haneef is an Associate Professor and Director International Cooperation Office at Air University, Islamabad, Pakistan. He received his B.Engg. degree from NED University of Engineering & Technology, Karachi, Pakistan, in 1991. After gaining 11 years of field experience in aviation industry, he did his MS from the National University of Sciences andTechnology (NUST), Islamabad, Pakistan with research focus on high strain rate deformation modelling of composite materials in 2004. He then won a prestigious PhD fellowship from Higher Education Commission of Pakistan and completed his PhD focusing on development of novel SOI CMOS MEMS flow sensors at Department of Engineering, University of Cambridge, U.K. in 2009. Since 2009, he has worked at Air University, Islamabad, Pakistan and National University of Sciences & Technology (NUST), Islamabad as an academic, and held different positions including Head of Research, Head of Department and Dean at College of Aeronautical Engineering, NUST. He has contributed to 27 international refereed journal and conference papers as main or co-author and is a co-inventor on one granted US patent and two Pakistani patent applications. He is also recipient of a Best Student Paper Award at IEEE International Conference on Sensors (IEEE Sensors 2008) and co-author of a paper that won Best Paper Award at 36th International Semiconductor Conference (CAS 2013). His research interests include high strain behavior of composite materials, and design, material optimization and packaging for CMOS MEMS (flow, pressure, temperature and gas) sensors. He is also interested in laser micromachining for MEMS and devices based on advanced materials including Carbon Nano Tubes (CNTs) and Graphene.

24

ACCEPTED MANUSCRIPT

IP

T

Zahid Mehmood received his B.Engg. from National University of Sciences and Technology (NUST), Islamabad, Pakistan in 2003. He then worked as a Technical Manager in Aviation industry for six years before undertaking MS (Solid Mechanics) studies at Air University, Islamabad, Pakistan from 2009 to 2011. He was as a lecturer at National University of Sciences and Technology, Islamabad (NUST), Pakistan from 2011-2014. Currently he is pursuing a PhD focusing at “Design and Materials Optimization for MEMS sensors” at Air University, Islamabad with a Visiting Researcher position at University of Cambridge, UK under the Knowledge Economy Partnership grant provided jointly by British Council and Higher Education Commission of Pakistan. He has authored / co-authored 10 international refereed journal papers and conference papers. His research interests include materials and design optimization of CMOS MEMS pressure sensors, flow sensors and temperature sensors.

7. 8. 9. 10. 11. 12. 13. 14.

15.

16. 17. 18. 19.

US

AN

6.

M

5.

ED

4.

PT

3.

CE

2.

Ramalhete, P., A. Senos, and C. Aguiar, Digital tools for material selection in product design. Materials & Design (1980-2015), 2010. 31(5): p. 2275-2287. Spearing, S.M., Materials issues in microelectromechanical systems (MEMS). Acta Materialia, 2000. 48(1): p. 179-196. Jahan, A., M.Y. Ismail, S.M. Sapuan, and F. Mustapha, Material screening and choosing methods – A review. Materials & Design, 2010. 31(2): p. 696-705. Yoon, K.P. and C.-L. Hwang, Multiple attribute decision making: an introduction. Vol. 104. 1995: Sage Publications. Saaty, T.L., The Analytical Hierarchy Process, Planning, Priority. Resource Allocation. RWS Publications, USA, 1980. Roy, B., The outranking approach and the foundations of ELECTRE methods. Theory and Decision, 1991. 31(1): p. 49-73. Opricovic, S., Multicriteria optimization of civil engineering systems. Faculty of Civil Engineering, Belgrade, 1998. 2(1): p. 5-21. Zadeh, L.A., Fuzzy sets. Information and Control, 1965. 8(3): p. 338-353. Ashby, M.F. and D. Cebon, Materials selection in mechanical design. Le Journal de Physique IV, 1993. 3(C7): p. C7-1-C7-9. Chauhan, A. and R. Vaish, A comparative study on material selection for micro-electromechanical systems. Materials & Design, 2012. 41: p. 177-181. Yazdani, M. and A.F. Payam, A comparative study on material selection of microelectromechanical systems electrostatic actuators using Ashby, VIKOR and TOPSIS. Materials & Design, 2015. 65: p. 328-334. Zha, X.F. and H. Du, Manufacturing process and material selection in concurrent collaborative design of MEMS devices. Journal of Micromechanics and Microengineering, 2003. 13(5): p. 509. Kolios, A., V. Mytilinou, E. Lozano-Minguez, and K. Salonitis, A comparative study of Multiple-Criteria Decision-Making (MCDM) methods under stochastic inputs. Energies, 2016. 9(7): p. 566. Shemshadi, A., H. Shirazi, M. Toreihi, and M.J. Tarokh, A fuzzy VIKOR method for supplier selection based on entropy measure for objective weighting. Expert Systems with Applications, 2011. 38(10): p. 1216012167. Huang, J.-J., G.-H. Tzeng, and H.-H. Liu, A revised VIKOR model for multiple criteria decision making-The perspective of regret theory, in Cutting-Edge Research Topics on Multiple Criteria Decision Making. 2009, Springer. p. 761-768. Velasquez, M. and P.T. Hester, An analysis of multi-criteria decision making methods. International Journal of Operations Research, 2013. 10(2): p. 56-66. Ananthasuresh, G. and M. Ashby, Concurrent design and material selection for trusses. 2003. de Oliveira, M.C.L., G. Ett, and R.A. Antunes, Materials selection for bipolar plates for polymer electrolyte membrane fuel cells using the Ashby approach. Journal of Power Sources, 2012. 206: p. 3-13. Wegst, U. and M. Ashby, Material selection charts for natural materials. 1997.

AC

1.

CR

References

25

ACCEPTED MANUSCRIPT

29.

30.

31.

32. 33. 34. 35. 36. 37. 38. 39.

40. 41. 42. 43. 44.

T

IP

CR

28.

US

27.

AN

26.

M

24. 25.

ED

23.

PT

22.

CE

21.

Rashedi, A., I. Sridhar, and K. Tseng, Multi-objective material selection for wind turbine blade and tower: Ashby’s approach. Materials & Design, 2012. 37: p. 521-532. Grujicic, M., G. Arakere, E. Subramanian, V. Sellappan, A. Vallejo, and M. Ozen, Structural-response analysis, fatigue-life prediction, and material selection for 1 MW horizontal-axis wind-turbine blades. Journal of materials engineering and performance, 2010. 19(6): p. 790-801. Cebon, D. and N. Ashby, Materials selection for precision instruments. Measurement Science and Technology, 1994. 5(3): p. 296. Chauhan, A. and R. Vaish, Hard coating material selection using multi-criteria decision making. Materials & Design, 2013. 44: p. 240-245. Ashby, M.F., Materials selection in mechanical design. Pergamon Press Oxford. 2005. Qian, J. and Y.-P. Zhao, Materials selection in mechanical design for microsensors and microactuators. Materials & Design, 2002. 23(7): p. 619-625. Madhavi, K., S. KA, M. Krishna, and A. Dharani. Diaphragm Design for MEMS Pressure Sensors using a Data Mining Tool. in Proceedings of the World Congress on Engineering. 2011. Srikar, V.T. and S.M. Spearing, Materials selection in micromechanical design: an application of the Ashby approach. Journal of Microelectromechanical Systems 2003. 12(1): p. 3-10. Pratap, R. and A. Arunkumar, Material selection for MEMS devices. Indian Journal of Pure and Applied Physics, 2007. 45(4): p. 358-367. Millet, O., P. Bernardoni, S. Régnier, P. Bidaud, E. Tsitsiris, D. Collard, and L. Buchaillot, Electrostatic actuated micro gripper using an amplification mechanism. Sensors and Actuators A: physical, 2004. 114(23): p. 371-378. Seki, T., M. Sakata, T. Nakajima, and M. Matsumoto. Thermal buckling actuator for micro relays. in Solid State Sensors and Actuators, 1997. TRANSDUCERS'97 Chicago., 1997 International Conference on. 1997: IEEE. Chu, C.-H., W.-P. Shih, S.-Y. Chung, H.-C. Tsai, T.-K. Shing, and P.-Z. Chang, A low actuation voltage electrostatic actuator for RF MEMS switch applications. Journal of Micromechanics and Microengineering, 2007. 17(8): p. 1649. Horsley, D.A., N. Wongkomet, R. Horowitz, and A.P. Pisano, Precision positioning using a microfabricated electrostatic actuator. IEEE Transactions on Magnetics, 1999. 35(2): p. 993-999. Srikar, V.T. and S.M. Spearing, Materials selection for microfabricated electrostatic actuators. Sensors and Actuators A: physical, 2003. 102(3): p. 279-285. Parate, O. and N. Gupta, Material selection for electrostatic microactuators using Ashby approach. Materials & Design, 2011. 32(3): p. 1577-1581. Srinivasan, P. and S.M. Spearing, Optimal materials selection for bimaterial piezoelectric microactuators. Journal of Microelectromechanical Systems 2008. 17(2): p. 462-472. Comtois, J.H. and V.M. Bright. Surface micromachined polysilicon thermal actuator arrays and applications. in Proc. Solid-State Sensor and Actuator Workshop. 1996. Phinney, L.M., M.S. Baker, and J.R. Serrano, Thermal Microactuators, in Microelectromechanical Systems and Devices. 2012, InTech. Baker, M.S., J.A. Walraven, T.J. Headley, and R.A. Plass, Compliant Thermo-mechanical MEMS Actuators, LDRD# 52553. 2004: United States. Department of Energy. Sinclair, M.J. A high force low area MEMS thermal actuator. in Thermal and Thermomechanical Phenomena in Electronic Systems, 2000. ITHERM 2000. The Seventh Intersociety Conference on. 2000: IEEE. Kong, S., S. Zhou, Z. Nie, and K. Wang, Static and dynamic analysis of micro beams based on strain gradient elasticity theory. International Journal of Engineering Science, 2009. 47(4): p. 487-498. Walraven, J.A. Failure mechanisms in MEMS. in International test conference. 2003. Guisbiers, G. and M. Wautelet, Materials selection for micro-electromechanical systems. Materials & Design, 2007. 28(1): p. 246-248. Guisbiers, G., O. Van Overschelde, and M. Wautelet, Materials selection for thin films for radio frequency microelectromechanical systems. Materials & Design, 2007. 28(6): p. 1994-1997. Guisbiers, G., E. Herth, B. Legrand, N. Rolland, T. Lasri, and L. Buchaillot, Materials selection procedure for RF-MEMS. Microelectronic Engineering, 2010. 87(9): p. 1792-1795.

AC

20.

26

ACCEPTED MANUSCRIPT

54. 55. 56. 57. 58. 59.

60. 61. 62. 63. 64. 65. 66. 67. 68. 69.

T

IP

CR

52. 53.

US

51.

AN

50.

M

49.

ED

48.

PT

47.

CE

46.

Sharma, A.K. and N. Gupta, Material Selection of RF-MEMS switch used for reconfigurable antenna using Ashby's Methodology. Progress In Electromagnetics Research Letters, 2012. 31. Cazzorla, A., P. Farinelli, R. Sorrentino, and B. Margesin, Reliability test of a RF MEMS varactor based on a double actuation mechanism. Microelectronics Reliability, 2015. 55(9-10): p. 1906-1910. Bansal, D., A. Kumar, A. Sharma, P. Kumar, and K. Rangra, Design of novel compact anti-stiction and low insertion loss RF MEMS switch. Microsystem Technologies, 2014. 20(2): p. 337-340. P. Reddy, G. and N. Gupta, Material selection for microelectronic heat sinks: An application of the Ashby approach. Materials & Design, 2010. 31(1): p. 113-117. Genolet, G., J. Brugger, M. Despont, U. Drechsler, P. Vettiger, N. De Rooij, and D. Anselmetti, Soft, entirely photoplastic probes for scanning force microscopy. Review of scientific instruments, 1999. 70(5): p. 23982401. Kohn, E., M. Adamschik, P. Schmid, S. Ertl, and A. Flöter, Diamond electro-mechanical micro devices— technology and performance. Diamond and Related Materials, 2001. 10(9-10): p. 1684-1691. Wang, J., J.E. Butler, D. Hsu, and C.T.-C. Nguyen. High-Q micromechanical resonators in CH4-reactantoptimized high acoustic velocity CVD polydiamond. in Solid State Sensor, Actuator, and Microsystems Workshop Proceedings, Hilton Head Island, SC. 2002. Multiphysics, C., Comsol multiphysics user guide (version 4.3 a). COMSOL, AB, 2012: p. 39-40. Phan, H.-P., D.V. Dao, K. Nakamura, S. Dimitrijev, and N.-T. Nguyen, The piezoresistive effect of SiC for MEMS sensors at high temperatures: A review. Journal of Microelectromechanical Systems, 2015. 24(6): p. 1663-1677. Jiang, L. and R. Cheung, A review of silicon carbide development in MEMS applications. International Journal of Computational Materials Science and Surface Engineering, 2009. 2(3-4): p. 227-242. Huang, X. and T.K. Chuan, MEMS-micropumps: a review. Journal of Fluids Engineering-Transactions of the ASME, , 2002. 124(2): p. 384-392. Prasanna, S. and S.M. Spearing, Materials selection and design of microelectrothermal bimaterial actuators. Journal of Microelectromechanical Systems 2007. 16(2): p. 248-259. Wang, X., B. Li, O.L. Russo, H.T. Roman, K.K. Chin, and K.R. Farmer, Diaphragm design guidelines and an optical pressure sensor based on MEMS technique. Microelectronics Journal, 2006. 37(1): p. 50-56. Bogue, R., MEMS sensors: past, present and future. Sensor Review, 2007. 27(1): p. 7-13. Rajavelu, M., D. Sivakumar, R. Joseph Daniel, and K. Sumangala, Perforated diaphragms employed piezoresistive MEMS pressure sensor for sensitivity enhancement in gas flow measurement. Flow Measurement and Instrumentation, 2014. 35: p. 63-75. Manikam, V.R. and K.Y. Cheong, Die attach materials for high temperature applications: A review. IEEE Transactions on Components, Packaging and Manufacturing Technology, 2011. 1(4): p. 457-478. Ando, T., M. Shikida, and K. Sato, Tensile-mode fatigue testing of silicon films as structural materials for MEMS. Sensors and Actuators A: Physical, 2001. 93(1): p. 70-75. Yi, T. and C.-J. Kim, Measurement of mechanical properties for MEMS materials. Measurement Science and Technology, 1999. 10(8): p. 706. Eaton, W.P. and J.H. Smith. Micromachined pressure sensors: review and recent developments. in Smart Structures and Materials. 1997: International Society for Optics and Photonics. Sharpe, W., J. Bagdahn, K. Jackson, and G. Coles, Tensile testing of MEMS materials—recent progress. Journal of Materials Science, 2003. 38(20): p. 4075-4079. Haque, M. and M. Saif, A review of MEMS-based microscale and nanoscale tensile and bending testing. Experimental Mechanics, 2003. 43(3): p. 248-255. Fu, Y., W. Huang, H. Du, X. Huang, J. Tan, and X. Gao, Characterization of TiNi shape-memory alloy thin films for MEMS applications. Surface and Coatings Technology, 2001. 145(1): p. 107-112. Lorenz, H., M. Despont, N. Fahrni, N. LaBianca, P. Renaud, and P. Vettiger, SU-8: a low-cost negative resist for MEMS. Journal of Micromechanics and Microengineering, 1997. 7(3): p. 121. Von Metzen, R.P. and T. Stieglitz, The effects of annealing on mechanical, chemical, and physical properties and structural stability of Parylene C. Biomedical Microdevices, 2013. 15(5): p. 727-735. Cho, S., I. Chasiotis, T.A. Friedmann, and J.P. Sullivan, Young's modulus, Poisson's ratio and failure properties of tetrahedral amorphous diamond-like carbon for MEMS devices. Journal of Micromechanics and Microengineering, 2005. 15(4): p. 728.

AC

45.

27

ACCEPTED MANUSCRIPT

T IP CR US AN M ED PT CE

71.

Auciello, O., J. Birrell, J.A. Carlisle, J.E. Gerbi, X. Xiao, B. Peng, and H.D. Espinosa, Materials science and fabrication processes for a new MEMS technology based on ultrananocrystalline diamond thin films. Journal of Physics: Condensed Matter, 2004. 16(16): p. R539. Schulz, M., Polymer derived ceramics in MEMS/NEMS–a review on production processes and application. Advances in Applied Ceramics, 2009. 108(8): p. 454-460.

AC

70.

28

ACCEPTED MANUSCRIPT Table 1. Performance and Material Indices for MEMS Sensors / Transducers Ref

Year

Devices

Variable Studied

Performance Indices Formulated

[2]

2000

Pressure sensor

Diaphragm deflection

-

Diaphragm deflection [25]

2002

𝛿=

𝐶1 3⁄ 2

𝐶2

Pressure sensor

[27]

[28]

2002

Microcantileversen sor

Vibration frequency

2003

Microfabricated cantilever basedsensors

Minimum detectable force (Sensitivity)

2007

Sensitivity / Deflection

𝑀= 3⁄ 2

𝑎 Δ𝑝

1⁄ 2

𝑓=

Vibration frequency

Capacitive Micromachined Ultrasonic Transducer (CMUT)

3⁄ 2

)(

𝜎𝑓

𝐸

(1 − 𝜈 2 ))

D E

U N

𝐹𝑚𝑖𝑛

A

M

T P

Thermal stability

𝑏ℎ2 =( ) 𝐿

C A

29

1⁄ 2

(𝐵𝑇𝛽)

1 1 [𝜒 ⁄2 (𝐸𝜌) ⁄4 ]

1⁄ 2

𝜎𝑓

3⁄ 2

𝑀1 =

Material Charts Developed No

𝐸 𝜎𝑓

No

𝐸

𝑀2 = √ 𝑀=

𝐸 𝜌

1 𝜌

No

No

1

𝑀1 = 𝜒 ⁄2 1 𝑀2 = (𝐸𝜌) ⁄4

Yes

𝐸 𝜌

Yes

-

𝑀1 = √

-

𝜎𝑓

-

E C

C S

1 𝐾 √ ∗ 2𝜋 𝑚

T P

I R

1 𝐸 𝑓∝ √ 𝐿 𝜌

Vibration frequency [25]

(

Material Indices Formulated

3⁄ 2

𝑀2 =

𝐸 𝑘 𝑀3 = 𝛼

Yes Yes

ACCEPTED MANUSCRIPT Table 2. Performance and Material Indices for MEMS Actuators Ref

Year

Devices

Variable Studied

𝐾=

Beam stiffness

𝐶1 𝐸𝐼 𝐿3 𝑉=√

Actuation voltage [33]

2003 Vibration frequency

Electrostatic actuators

[34]

𝑓=

Thermal actuators

Actuation voltage

2𝐾(𝑑 − 𝑑1 ) 𝑉 = 𝑑√ 𝜖𝐴

Thermal distortion

Moment

D E

2008

BimaterialPiezoEl ectric (BPE) micro-actuators

T P

Work

C A

E C

Actuation frequency

A

U N

M

𝑓=

1 𝐾 √ 2𝜋 𝑚

-

Θno = 6𝑑31 /4 𝐸𝑑 𝑀𝑛𝑜 = 1 31 2 ⁄ 𝜉0 + 1 2( ⁄𝜉 ) 0 3𝐸1 (𝑑31 )2 𝑊𝑛𝑜 = 2 ⁄ 𝜉0 + 1 32 ( ⁄𝜉 ) 0 1 1.8751 2 𝑓= ( ) 2𝜋 𝐿 𝐸1 ℎ2 √ 3𝜆(𝜌1 𝜉0 + 𝜌2 )(𝜉0 + 1)

30

T P

I R

C S

Optimum slope

[35]

𝐶2 𝐸ℎ2 √ 2𝜋 12𝜌𝐿4

2𝐿2 𝜎𝑓 𝛿= 3ℎ 𝐸

2011

2007

2𝐾𝑔2 (𝑔𝑜 − 𝑔) 𝜖𝐴

Actuation stroke

Actuation speed [28]

Material Indices Formulated

Performance Indices Formulated

Material Charts Developed

𝑀1 = 𝐸

Yes

𝑀2 = √𝐸

Yes

𝐸 𝑀3 = √ 𝜌 𝜎𝑓 𝑀4 = 𝐸

Yes Yes

𝑀1 = √𝐸

No

𝐸 𝑀2 = √ 𝜌

No

𝑀=

𝑘 𝛼

No

-

Yes

-

Yes

-

Yes

-

Yes

ACCEPTED MANUSCRIPT Table 3. Performance and Material Indices for Micro-beams / Flexures Ref

[27]

Year

2003

Devices

Shock Resistant Micro-beams

Variable Studied

𝑓=

Inertial load

𝑃 = (𝐿𝑏ℎ)(𝜌𝑎′ )

Deflection

𝛿=

𝑏ℎ3 (𝐸) 4𝐿3

𝛿=(

2𝐿2 𝜎𝑓 )( ) 3ℎ 𝐸

2003

Micro-machined Flexures

Displacement

[25]

2002

Micro-springs

Elastic Energy

D E

M

T P

E C

C A

31

1 𝜎𝑓2 𝑈= 2𝐶1 𝐸

T P

𝑀=

𝐸 𝜌

Yes

𝑀=

𝜎𝑓 𝐸

Yes

𝑀=

𝜎𝑓2 𝐸

No

I R

C S

U N

A

Material Charts Developed

0.16 ℎ 𝐸 √ 𝐿2 𝜌

Resonating frequency

[27]

Material Indices Formulated

Performance Indices Formulated

ACCEPTED MANUSCRIPT Table 4. Performance and Material Indices for RF-MEMS Ref

[42]

[43]

[44]

Year

2007

2007

2010

Devices

RF-MEMS

RF-MEMS

RF-MEMS (RF Bridge)

Variable Studied Intrinsic residual stresses in evaporation process deposition High pull-in voltage (RF – Bridge) Intrinsic residual stresses in evaporation process deposition

𝐸 𝜎𝑖𝑛𝑡 = 𝛾 ( ) 𝛼(𝑇𝑚𝑒𝑙𝑡𝑖𝑛𝑔 − 𝑇𝑠𝑢𝑏𝑡𝑟𝑎𝑡𝑒 1−𝜈 𝑉𝑝𝑢𝑙𝑙−𝑖𝑛 = √

Available Capacitance (RF-Varicap) Intrinsic residual stresses in evaporation process deposition

A

D E

M

𝜎𝑖𝑛𝑡 =

T P

[45]

2012

𝐸 𝑀 𝑁𝜙√𝐸𝑝 1−𝜈 𝐷

𝑉𝑝𝑢𝑙𝑙−𝑖𝑛 = √

E C

C A

-

𝑄=

𝑉𝑝 = √

Pull in voltage

8𝐾𝑑 3 27𝐴𝜀

√𝐿′ 𝑅√𝐶 8𝐾𝑑 3 27𝜖𝑊𝑏

𝐶1 𝜌𝑒 𝐿 4ℎ𝑏 𝜀𝐿 Δ𝜎 = 𝐸Δ𝛼𝑃𝑙𝑜𝑠𝑠 4𝑘ℎ𝑏 𝑃𝑙𝑜𝑠𝑠 = 𝐽2

RF loss Thermal residual stress

32

T P

I R

C S

8𝐸𝑑 3 =√ 27𝐴𝜀

U N

𝑉𝑝𝑢𝑙𝑙−𝑖𝑛

Maximum Quality factor

RF-MEMS bridge switch for Antenna

8𝐾𝑑 3 27𝐴𝜀

𝐸 𝜎𝑖𝑛𝑡 = 𝛾 ( ) 𝛼(𝑇𝑚𝑒𝑙𝑡𝑖𝑛𝑔 − 𝑇𝑠𝑢𝑏𝑡𝑟𝑎𝑡𝑒 1−𝜈

Low pull-in voltage ((RF-Switch)

Minimum pull-in voltage

Material Indices Formulated

Performance Indices Formulated

Material Charts Developed

𝑀1 = 𝛼𝐸𝑇𝑚𝑒𝑙𝑡𝑖𝑛𝑔 /(𝑘(1 − 𝜈))

No (Table prepared)

𝑀2 = √𝐸𝜌−1

Yes

𝑀1 = 𝛼𝐸𝑇𝑚𝑒𝑙𝑡𝑖𝑛𝑔 /(𝑘(1 − 𝜈))

Yes

𝑀2 = √𝐸/𝜌−1

Yes

𝑀3 = √𝐸𝜌−1

Yes

𝑀1 =

𝐸𝑀 (1 − 𝜈)𝐷

Yes

𝑀2 = √𝐸

Yes

𝑀3 = 𝜌

Yes

𝑀1 = √𝐸 𝑀2 = 𝜈 𝑀3 = 𝛼

Yes

𝑀4 = 𝜌

Yes

𝜌 𝐾

Yes

𝑀5 =

ACCEPTED MANUSCRIPT Table 5. Performance and Material Indices for Miscellaneous MEMS Devices Ref

Year

Devices

Variable Studied

Performance Indices Formulated

[2]

2000

Gyroscopes

Vibrating frequency

-

[2]

2000

-

-

[25]

2002

[27]

2003

Rotating Disc Pumps Micromotors/Pumps

Kinetic energy of disc

Micro-Mechanical Filters

Resonating frequency

Microphones

Gyroscope

D E

Natural frequency [28]

2007 Resonators Accelerometers and Switches Micro-motors, Micro-pumps and Micro-turbines

[48]

2010

Micro-electronic Heat Sinks

U N 𝐴=

A

M

T P

Natural frequency

E C

C A

Microheaters

C S

-

2𝐴𝑑 Ω𝑄𝑦 𝜔𝑦

𝐸 𝜌

3⁄ 2

𝑀=

𝜎𝑓

𝐸 𝑘 𝑀1 = (𝐸𝛼)2

No No No Yes No Yes

𝑀2 =

𝜎𝑓 𝐸

Yes

-

𝐸 𝑀3 = √ 𝜌

No

-

𝑀=√

-

Inertial force

-

Thermal distortion

-

Resistivity and thermal conductivity

𝛼 = 𝛾𝐺 𝜌𝐶𝑝 /3𝐸

33

𝑀=√

Material Charts Developed

-

Deflection

Thermal stresses

T P

I R

-

Out of plane deflection Maximum deflection without fracture

𝐸 𝜌 𝜎𝑓 𝑀= 𝜌 𝜎𝑓 𝑀= 𝜌 𝑀=

𝜎𝑓 𝑈 2 =( )( ) (3 + 𝜈) 𝜌 𝑚

Deflection

Material Indices Formulated

𝐸 𝜌 𝜎𝑓 𝑀= 𝐸 𝜎𝑓 𝑀= 𝜌 𝑀=

No No No

𝑘 𝛼

𝜌𝑒 > 1019 𝜇Ω𝑐𝑚 Minimum 𝛼 Maximum E

Yes Yes

ACCEPTED MANUSCRIPT Thermal contact resistance

𝑞 = (𝑇1 − 𝑇3 )/(

Δ𝑥𝐴 ⁄𝐴𝑘 + (𝐴ℎ𝑐 )−1 𝐴 Δ𝑥𝐵 + ⁄𝐴𝑘 𝐵

Highest ℎ𝑐 Highest 𝑘𝐵

T P

I R

C S

A

U N

D E

M

T P

E C

C A

34

No

ACCEPTED MANUSCRIPT SNo

MEMS Systems

1

MEMS Sensors / Transducers

Table 6. Candidate Materials forDifferentCategories of Microsystems / MEMS Year Devices Method Used Materials Proposed By comparing material Diaphragm Based [2] 2000 properties of 08 materials in a Silicon Pressure Sensor table Diaphragm Based Silicon By comparing material Pressure Sensor [25] 2002 properties of 09 materials in a Micro-cantilever Based table Silicon oxide Sensor By comparing material Micro-cantilever Based Silicon oxide, quartz, silicon, gallium [27] 2003 properties of 14 materials in a Sensor arsenide, and silicon nitride plot By comparing material Membrane Based Diamond, Silicon carbide, Silicon nitride, [28] 2007 properties of 16 materials in a Ultrasonic Transducer Polysilicon plot For high speed and high force actuators - Diamond, Silicon Carbide, Alumina,Silicon Nitride, Silicon By comparing material For large displacement and low [33] 2003 properties of 30 materials in a voltage actuators - Polymers plot For high speed and low voltage actuators - Silicon dioxide, quartz, Aluminum For high voltage and high force Electrostatic Actuators actuation- Diamond, Silicon Carbide, By comparing material Silicon nitride [34] 2011 properties of 11 materials in a plot For high speed actuation - Aluminum, Nickel, Copper For high speed high force actuators Diamond, Silicon Carbide By comparing material [11] 2015 properties of 15 materials in a For large displacement and low plot voltage actuators - Polymers (PMMA, PVDF) Ref

T P

I R

C S

U N

A

2

MEMS Actuators

D E

M

T P

E C

C A [35]

2008

BimaterialPiezo-Electric (BPE) Micro-actuator

35

By comparing material properties of 15 materials in a plot

For large force/work actuators – PZNPT, PZT, PMNPT and BaTiO3 on Silicon or DLC For high frequency application – AIN on silicon or DLC

ACCEPTED MANUSCRIPT

3

Micro-beams / Flexures

[27]

2003

Shock Resistant Microbeams

[27]

2003

Micro-machined Flexures

[25]

2002

Micro-springs

[42]

2007

RF-MEMS (Material for RF-Switch and Varicap)

RF - MEMS

D E

[44]

Others

I R

By comparing material properties of 167 materials in a self-developed material data base

C S

T P

U N

A

M

By comparing material properties of 167 materials in a self-developed material data base

RF-MEMS

By comparing material properties of 07 materials in a plot By comparing material properties of 08 materials in a plot

[45]

E C 2012

RF-MEMS

[2]

2000

Pumps/ Turbomacniery

[25]

2002

Pumps

[27]

2003

Micro-mechanical Filters

C A 5

2010

36

Diamond, Silicon Carbide, Alumina, Silicon Nitride,and Silicon Polymers, Silicon Nitride, Ni-Fe alloys, and Silicon Silicon

T P

RF-MEMS (Material for RF-Bridge)

2007

[43] 4

By comparing material properties of 25 materials in a plot By comparing material properties of 12 materials in a plot By comparing material properties of 09 materials in a table

By comparing material properties of 08 materials in a table By comparing material properties of 09 materials in a table By comparing material properties of 25 materials in a

Intrinsic Stress - Rubidium, Boron Nitride, Potassium, Silver, Arsenic, Magnesium, Carbon, Gold, Copper, Aluminum High pull in voltage and high quality factor- Copper, Silver, Iridium, Gold, Rhodium, Tungsten, Molybdenum Intrinsic Stress - Rubidium, Boron Nitride, Potassium, Silver, Arsenic, Magnesium, Carbon, Gold, Copper, Aluminum Low pull in voltage and high quality factor- Rubidium, Potassium, Silver, Magnesium, Gold, Copper, Aluminum High pull in voltage and high quality factor – Copper, Silver, Gold For intrinsic stress : Aluminum, Copper and Gold For low pull in voltage and high quality factor : Aluminum, Copper and Gold Low pull in voltage - Aluminum, Gold RF-Loss - Aluminum, Gold Thermal residual stress – Gold, Copper, Aluminum Silicon

Silicon Diamond, Silicon Carbide, Alumina, Silicon Nitride, and Silicon

ACCEPTED MANUSCRIPT [28]

[48]

2007

plot By comparing material properties of 25 materials in a plot

Vibratory Gyroscope

Using CES generated material selection charts

Micro-electronic Heat Sink

2010

C S

U N

A

M

T P

E C

C A

37

For high resistivity and thermal conductivity – Aluminum Nitride and Alumina Thermal stresses, Aluminum, Copper, Zinc alloy, Aluminum Nitride, Aluminum oxide Thermal contact resistance Aluminum

T P

I R

D E

Diamond, Silicon Carbide, Silicon Nitride, Silicon oxide, Polysilicon

ACCEPTED MANUSCRIPT First Author, Year [Ref] Spearing, 2000 [2] Qian, 2002 [25] Reddy, 2010[48], Parate, 2011 [34] Phan, 2015 [53] Jiang, 2009[54] Nguyen, 2002 [55] Srikar, 2003 [27] Prasann, 2007 [56] Wang, 2006[57] Bogue, 2007 [58] Rajavelu, 2014 [59] Yazdani, 2015[11] Manikam, 2011 [60] Pratap, 2007[28] Ando, 2001 [61] Chauhan, 2012 [10] Pratap, 2007 [28] Yi, 1999 [62] Ando, 2001 [61] Chauhan, 2012[10] Ando, 2001 [61] Spearing, 2000[2] Qian, 2002 [25] Prasann, 2007[56] Srikar, 2003[27] Yazdani, 2015 [11] Pratap, 2007 [28] Reddy,2010[48] Parate, 2011[34] Chauhan, 2012 [10] Spearing, 2000 [2] Eaton, 1997 [63] Bogue, 2007 [58] Qian, 2002 [25] Prasann, 2007 [56]

Table 7. MEMS / Microsystems’ Material Properties Reported in the Literature Density, 𝝆 (kg/m3) Modulus, 𝑬 (GPa) Material Published CES Published CES 2330 129-187 2330 129-187

Silicon

Silicon (100)

Silicon (110)

E C

C A

Silicon Oxide

Silicon Nitride

NR*

125-180

>1000

NR* 2330 NR* NR* NR* NR* NR* NR* 2000 2330 2300 NR* 2300 2300 NR* NR* 2300

130-180 130-185 200 125-180 165 NR* NR* 106.8 160 NR* 115-142 122 130 147-188 164-172 140 168

NR* NR* NR* >1000 NR* NR* NR* NR* 4000 NR* 2000-4300 NR* 3400 6000-8000 1200 NR* 7000

D E

T P

Silicon (111)

Fracture Strength, 𝝈𝒇 (MPa) Published CES 4000 4000

NR*

2200 2200 NR* NR* 2000 2500

2280-2380

T P

140-180

I R

C S

Material not included in CES

A

U N

M

Material not included in CES Material not included in CES

2170-2220

111 73 73 75 70 73 57-92

Material not included in CES Material not included in CES Material not included in CES

72-74

NR* 1000 1000 NR* 1000 1000 800-1100

NR*

70

1000

2500 3300 NR* NR* 3300 NR*

70 304 NR* NR* 304 260

1000 1000 1000-2000 1000-2000 1000 NR*

3150-3210

38

288-302

165-180

Material not included in CES Material not included in CES Material not included in CES

45.7-50.4

240-270

ACCEPTED MANUSCRIPT Srikar, 2003 [27] Pratap, 2007 [28] Reddy, 2010[48], Parate, 2011[34] Sharma, 2012 [45] Yazdani, 2015 [11] Sharpe, 2003[64] Chauhan, 2012[10] Nguyen, 2002 [55] Spearing, 2000[2] Bogue, 2007[58] Qian, 2002[25] Prasann, 2007 [56] Srikar, 2003 [27] Reddy, 2010[48], Parate, 2011 [34] Yazdani, 2015 [11] Phan 2015[53] Sharpe, 2003 [64] Manikam, 2011 [60] Jiang 2009[54] Chauhan, 2012 [10] Pratap, 2007 [28] Pratap, 2007 [28] Srikar, 2003 [27] Chauhan, 2012 [10] Reddy, 2010[48], Parate, 2011 [34] Yi, 1999 [62] Sharpe, 2003 [64] Spearing, 2000[2] Qian, 2002 [25] Prasann, 2007 [56] Srikar, 2003 [27] Pratap, 2007 [28] Guisbiers, 2007, 2010 [42-44] Sharma, 2012 [45] Reddy, 2010[48], Parate, 2011 [34]

Silicon Carbide

NR* 3100

250 230-290

6000 5000-8000

NR*

250

6000

NR* 3000 NR* 3100 NR* 3300 3300 3300 NR* NR*

304 323 252-262 250 300 430 430 430 460 400

NR* 1000 5830 ± 250 6400 NR* 2000 2000 2000 NR* NR*

NR* 3000 NR* NR* 3170-3210 3210 3200

D E

3H-SiC

3200

Polysilicon

T P

E C

C A Nickel

2300 NR* 230 NR*

NR* NR* 8900 8900 NR* NR* 8910

3140-3210

M

Material not included in CES

8830-8930

390-410

U N

A

Material not included in CES

I R

C S

400

T P

450 300-500 417 NR* 392-694 400 331-470 140-169 160 159 160 130-174 NR* 207 207 207 180 168-214

Material not included in CES

Material not included in CES

190-220

NR* 2000 NR* 810 NR* NR* 7000 4000-9000 1210-2800 1200-3000 1650

Material not included in CES

1200-3000

Material not included in CES

1250-2500 3000 500 500 NR* 500 320-780

360-445

8910

221

NR*

NR*

204

NR*

NR*

180

500

39

476-525

ACCEPTED MANUSCRIPT Yazdani, 2015 [11] Yi, 1999 [62] Chauhan, 2012 [10] Spearing, 2000 [2] Qian, 2002 [25] Prasann, 2007 [56] Srikar, 2003 [27] Pratap, 2007 [28] Guisbiers, 2007, 2010 [42-44] Chauhan, 2012 [10] Haque, 2003 [65] Yazdani, 2015 [11] Reddy, 2010[48], Parate, 2011 [34] Sharma, 2012 [45] Pratap, 2007 [28] Srikar, 2003 [27] Prasann, 2007 [56] Guisbiers, 2007, 2010 [42-44] Reddy, 2010[48], Parate, 2011[34] Sharma, 2012 [45] Yazdani, 2015 [11] Chauhan, 2012 [10] Qian, 2002 [25] Spearing, 2000 [2] Manikam, 2011 [60] Chauhan, 2012 [10] Phan 2015 [53] Yazdani, 2015 [11] Pratap, 2007 [28] Srikar, 2003 [27] Prasann, 2007 [56] Guisbiers, 2007, 2010 [42-44] Reddy, 2010 [48], Parate, 2011 [34] Sharma, 2012 [45] Yazdani, 2015 [11]

8902 NR* 8910 2710 2710 NR* NR* 2700 Aluminum

193 176 185 69 69 68 69 47-85

2710

2670-2730

2700 NR* 2700

70 69.6-74.6 70

NR*

69

D E

T P

E C

Diamond

C A Gold

M

3440-3580

19300

NR* 170 NR* 300

NR* 120-260 350 NR*

115

NR* 120-125

45-55

124

350

115 117 120 1035 1035 NR* 800 1000 1200 600-1100 70 77

NR* 250 1000 1000 NR* 8500 NR* 1000 8000-10000 300 NR*

1050-1210

75 19300-19400

2800-2930

NR* 76-81

180-220

NR*

70

300

NR* 19300

77 70

NR* 300

40

76-84

150

69 86-137 124 110

8940-8950

8960 8960 3510 3510 3520 3500 NR* 4000 3500 NR* NR*

I R

U N

A

8890 NR*

69-72

C S

NR* 8960 NR* NR* Copper

T P

68

500 560 400 300 300 NR* 150 150-300

ACCEPTED MANUSCRIPT Guisbiers, 2007, 2010 [42-44] Sharma, 2012 [45] Yazdani, 2015 [11] Yazdani, 2015 [11] Guisbiers, 2007, 2010 [42-44] Pratap, 2007 [28] Chauhan, 2012 [10] Yi, 1999 [62] Pratap, 2007 [28] Yazdani, 2015 [11] Chauhan, 2012 [10] Yazdani, 2015 [11] Guisbiers, 2007, 2010 [42-44] Yazdani, 2015 [11]

Srikar, 2003 [27] Reddy, 2010 [48], Parate, 2011[34] Yazdani, 2015 [11] Yazdani, 2015 [11] Nguyen, 2002 [55] Fu, 2001 [66] Pratap, 2007 [28] Prasann, 2007 [56] Yazdani, 2015 [11] Chauhan, 2012 [10] Nguyen, 2002 [55] Lorenz, 1997 [67] Pratap, 2007 [28]

Nguyen, 2002 [55]

147 21500

NR* 21450 4506

Tungsten

4510 4510 NR* 19300 19250 19300 7190

Chromium

Ni-Fe alloy

Titanium Carbide Stainless Steel TiNi

C A SU-8

Silicone Rubber

7100-7200

NR*

8100-8200

D E

8000 5000 NR* NR*

1420 NR* 1000 1420 NR* NR* 1164

NR* NR*

A

M

4810-5010 7710-7790 Material not Included In CES

1330-1430

Material not included in CES Material not included in CES 1050-1070

41

96-115 110 96 410 411 410 279

NR* NR* 500

120-165

NR* 100-105

T P

I R

340-350

440-790 500 950 700 700 700 NR*

240-360

1670-2050

C S

245-285

83 121 209 211 120

70-74 109-115 199-215 204-212

NR* NR* NR* NR* 1600

110-175 320-420 1020-1250 230-295

120

137-145

1600

655-810

1600 NR* NR*

260-330 1110-1220

289

U N

10500 12000-12100 8800-8900 7860

NR* Parylene

19300-19400

10490 12023 8900 7874 NR*

T P

E C

Polyimide

116 4510-4520

7190

Silver Palladium Cobalt Iron

NR* 168-172

171 168 116

4510

Titanium

Nguyen, 2002 [55] Metzen, 2013 [68]

21440

Platinum

120 439 240 60-80 4-15 4 8 8 10 4.05 1.8-4.2 3 2.9 0.0005

420-450 200-210 Material not Included In CES

2.07-2.76

Material not included in CES Material not included in CES .0002-.00021

NR*

NR* 23-70 NR* 40 40 NR* NR* 30-50 NR*

370-460

Material Not Included in CES

72.4-118

Material not included in CES

68.9

Material not included in CES

NR*

1.69-2.13

ACCEPTED MANUSCRIPT Yazdani, 2015 [11] Chauhan, 2012 [10] Pratap, 2007 [28] Pratap, 2007 [28] Yazdani, 2015 [11] Qian, 2002 [25] Spearing, 2000[2] Sharma, 2012 [45] Yazdani, 2015 [11] Manikam, 2011 [60]

2000 1780 1780 1200

Poly-Vinylidene-DiFluoride (PVDF) Poly-Mmethyl-MethAcrylate (PMMA)

3970 3970 NR* 4000 5320

Gallium Arsenide

Jiang 2009 [54]

Prasann, 2007 [56] Srikar, 2003 [27] Reddy, 2010 [48], Parate, 2011 [34] Cho, 2005 [69] Auciello, 2004 [70]

NR*

Diamond Like Carbon (DLC)

NR* NR*

Ultra-Nano-Crystalline Diamond

T P

Yazdani, 2015 [11]

3000

Qian, 2002 [25]

Schulz, 2009 [71]

E C

Tin Lead Molybdenum

C A

7365 11340 NR*

Carbon Single Walled Nano-Tubes (SWNT)

1330

Polymer Derived Ceramic SiCN

NR*

U N

M

Material not included in CES Material not included in CES 7280-7310 11300-11400 10100-10300 Material not included in CES Material not included in CES

D E

NR*

NR*

(*) Legend: NR-Not Reported

42

200-300 700 800 800

T P

361-381

I R

C S

85.5

A

NR*

Quartz

Sharma, 2012 [45]

Material not included in CES

2.21-2.41

393 393 380 275 NR*

Material not included in CES Material not included in CES

6100 Gallium Nitride

2-2.5

2

3820-3900

5320

Manikam, 2011 [60] Phan 2015 [53]

1140-1160

1000

Aluminum Oxide

2 2.3 1.1-4 1.8-3.1

1770-1780

Material not included in CES Material not included in CES Material not included in CES

759 ± 22 980

107 50 16 320 >1000

150

50 50 48-60 48-80 80 2000 2000 NR* 2000 NR* NR* NR* NR* NR* 8000 8000

24.1-50

55.2-62.1

252-265

Material not included in CES Material not included in CES

Material not included in CES

7300 ± 1200 Material not included in CES Material not included in CES 41-45 13-14 315-335 Material not included in CES Material not included in CES

4000-5000

Material not included in CES

1700

Material not included in CES

NR* NR* NR*

11-18 15-20 450-795

NR*

Material not included in CES

NR*

Material not included in CES

ACCEPTED MANUSCRIPT

AUTHORS CONTRIBUTION STATEMENT

T

Zahid Mehmood : Data Curation, Formal analysis, Writing - original draft

IP

Ibraheem Haneef: Conceptualization, Funding acquisition, Formal analysis, Investigation,

CR

Supervision, Writing - review & editing

AC

CE

PT

ED

M

AN

US

Florin Udrea : Supervision, Funding acquisition, Writing - review & editing

43

ACCEPTED MANUSCRIPT MATERIAL SELECTION FOR MICRO-ELCTRO-MECHANICAL-SYSTEMS (MEMS) USING ASHBY’S APPROACH : GRAPHICAL ABSTRACT

T P

I R

C S

A

U N

D E

M

T P

E C

C A

44

ACCEPTED MANUSCRIPT

HIGHLIGHTS A review of material selection methodologies applied to Micro-Electro-MechanicalSystems during the last two decades is presented.



Ashby’s material selection methodology was found to be the most frequently used method for MEMS material selection



The performance and material indices for 20 types of MEMS devices have been analyzed and a database of 51 MEMS materials has been developed.



A case study demonstrates that Ashby’s methodology gives most accurate materials choices for MEMS devices only with micro-scale material properties.

AC

CE

PT

ED

M

AN

US

CR

IP

T



45