Ghodssi R., Lin P., MEMS Materials and Processes Handbook

Подождите немного. Документ загружается.

204 E. Meng et al.

Fig. 4.6 Dimensionless form of force-displacement relationships for PDMS microcantilever

shown by: (1) small deﬂection beam theory, (2) the elastic FEA model for short beams, and (3)

the viscoelastic FEA model with different loading rates (Reprinted with permission from [153],

adoption of PDMS in biological research laboratories. An example of the PDMS

structure is given below [144]. First, the base PDMS prepolymer and the curing

agent are mixed with a ratio of 10:1 and then the mixture is stirred thoroughly for

uniformity. The PDMS mixture is placed in a vacuum for 5 min to get rid of the air

bubbles introduced during stirring. The PDMS mixture is then poured on top of the

substrate, and spun at a low rate (about 500 rpm) to create a ﬂat surface. Afterwards,

the stack is placed on top of a hotplate for initial thermal curing (or “precuring”) at

◦

C (Fig. 4.7a). The curing time is carefully controlled so that the PDMS mixture

does not fully cross-link. At the completion of the precuring, the molded pattern is

slightly dipped onto the PDMS mixture and a light weight (e.g., 15 g) is applied on

top (Fig. 4.7b). The substrate is put back into the vacuum for 5 min (Fig. 4.7c). Upon

returning to atmospheric pressure (Fig. 4.7d) the PDMS mixture is thermally cured

at 65

◦

C for a second time ( Fig. 4.7e). Following removal of the master template,

the PDMS specimens, with various aspect ratios, are formed on the base substrate

(Fig. 4.7f).

An alternative approach to PDMS fabrication is to make the PDMS pre-

polymer sensitive to ultraviolet (UV) wavelengths, and then it can be directly

photopatternable. Several companies have commercialized photosensitive silicon

containing polymers: WL-5000 silicones by Dow Corning for electronic packaging

4 Additive Processes for Polymeric Materials 205

Secondary curing

(f)

Vacuum pump

V < V

atm

(c)

(g)

Precuring

(a)

(b)

Silicon mold

Partially cured PDMS

Glass slide

(d)

V = V

vac

V = V

atm

(e)

Fig. 4.7 Illustration of pressure-assisted micromolding: (a) PDMS prepolymer is precured on

a glass slide. (b) The mold is slightly dipped into partially cured prepolymer. (c) Trapped air

expands and some escapes with a certain vacuum level. (d) Trapped air reaches the equilibrium.

(e) The polymer is raised when increasing the chamber pressure. (f) Secondary curing is conducted.

(g) Cantilevers with various aspect ratios are formed after the removal of the master mold [144]

applications [158], cyclotene photosensitive resins from Dow Chemical (derived

from benzocyclobutene (BCB) monomers) as dielectrics [159], and inorganic–

organic hybrid polymers (ORMOCER) [160] for optical applications developed

at Fraunhofer Institute and commercialized by Micro Resist. However, these new

materials have different chemical nature and softness compared to the vinyl copoly-

mers or pure PDMS polymers used in lithography thus far. Recently there have

been some reports on photosensitive polymers by introducing photoinitiators, s uch

as 2,2-dimethoxy-2-phenyl acetophenone (DMAP) [161] and benzophenone [162].

The application to PDMS has also been reported [163, 164]. The photodeﬁnable

mixture of PDMS and a photoinitiator, which eliminates the need of a master mold

and the compatibility issues related to molding, provides a more efﬁcient way of

rapid prototyping polymer-based MEMS devices.

4.2.3 Biological Application Guide

Generally speaking, the biocompatible nature, rheological property, surface chem-

istry, and insulative property made PDMS a versatile material choice in biological

applications as a functional, structural, and transfer material.

206 E. Meng et al.

4.2.3.1 Stamp Material for Protein Transfer: Microcontact Printing

The microcontact printing (μCP) technique was ﬁrst introduced by Kumar and

Whitesides and coworkers at Harvard University in 1993, for patterning self-

assembled monolayers (SAMs) of alkanethiols onto gold substrates [165]. The

concept of μCP is straightforward: This technique is similar to ﬂexography [166]

which was adapted from the paper printing industry: alkanethiol “ink” was trans-

ferred to a gold “paper” by a prepatterned PDMS “stamp”. Since then, the concept

has been widely used for patterning substrates with biomolecules that enable future

cell attachment and growth [167]. This includes, for example, folding of proteins

and t-RNAs [168], formation of the DNA double-helix [169], and formation of the

cell membranes from phospholipids [170]. First, the PDMS stamp is fabricated by

the replica process from the PDMS prepolymer and mold (Section 5.2.2). Certain

types of biomolecules are attached to a PDMS stamp pattern (Fig. 4.8a) and then

transferred to a chemically activated substrate by direct contact of the substrate with

the stamp (Fig. 4.8b). Unstamped regions are chemically treated with cell adhesion

inhibitor (Fig. 4.8c), such as albumin, polyethylene glycol, or pluronic [171, 172].

However, biomolecules will remain in the stamped regions and allow cells to attach

(Fig. 4.8d).

Substrate

PDMS Stamp

Biomolecule

(a)

(b)

(c)

(d)

Substrate

PDMS Stamp

Biomolecule

(a)

(b)

(c)

(d)

Fig. 4.8 The PDMS stamp

transfers the biomolecule

“ink” to a substrate upon brief

contact: (a) Biomolecules are

attached to a PDMS stamp.

(b) Transfer to a chemically

activated substrate by direct

contact of the substrate with

the stamp. (c)Unstamped

regions are chemically treated

with cell adhesion inhibitor.

(d) Cells to attach to the

biomolecules

4.2.3.2 Microﬂuidic Devices

Microﬂuidic devices have provided extraordinary advantages for biochemical anal-

ysis or synthesis. The basic concept of the microﬂuidic device is the integration

of various chemical operations involved in conventional analytical processes done

in a laboratory, such as mixing, reaction, and separation, into a miniaturized ﬂow

system. Ever-increasing attention has been garnered by microﬂuidic devices due to

the extraordinary advantages of parallelization, automation, integration of sample

preparation, low cost, short analysis time, and high sensitivity on separation and

detection [173, 174].

4 Additive Processes for Polymeric Materials 207

PDMS has become a prime candidate as the building material of microﬂuidic

devices for a variety of reasons. PDMS is much less expensive, easy to fabricate and

bond with external components (e.g., glass, silicon, and other PDMS), has a vari-

able thickness, and is less fragile compared to silicon and glass. Soft lithography

techniques in PDMS have made the fabrication of microﬂuidic prototype devices

with much shorter time and convenience than using silicon fabrication technol-

ogy. Valves, mixers, and pumps can be easily fabricated and molded based on soft

lithography procedures described in Section 4.2.2. The sealing of PDMS with other

materials can be achieved either reversibly (e.g., van der Waals contact of PDMS-

PDMS or PDMS-glass) or irreversibly (with the assistance of plasma oxidation or

bonding of precured PDMS to a fully cured PDMS).

The physical properties of PDMS also make it favorable candidate as the

functioning material for microﬂuidic devices. With its optical transparency for

wavelengths ranging from 235 nm to near-infrared, PDMS enables optical detection

over the entire visible spectrum, which is particularly attractive for the fabrication

of microﬂuidic devices with an integrated optical detection or actuation system as

shown in Fig. 4.9 [175].

Fig. 4.9 The experimental setup for ﬂuorescence detection. Green light is emitted across a

microﬂuidic channel ﬁlled with phosphate-buffered saline (PBS) (Reprinted from [175], copyright

2006, IEEE)

PDMS is attractive for microﬂuidic applications because it is waterproof and has

high gas permeability. These properties enable the inexpensive PDMS membrane

to function without the integration of an electrostatic or electromagnetic ﬂow con-

trol system. Many researchers have turned to PDMS microvalves and micropumps,

along with pumps that rely on the gas permeability of PDMS, to produce ﬂuid move-

ment. For example, a thin membrane can be placed between a control channel layer

and a ﬂow channel layer. Fluid ﬂow is triggered from the application of pressure

across the thin PDMS membrane and the ﬂuid ﬂow can be controlled bidirectionally

(Fig. 4.10)[176–178].

208 E. Meng et al.

Fig. 4.10 Diagram of PDMS-based permeation pump mechanism within a microﬂuidic device

(Reprinted with permission from [178], IOP)

The elastic property of PDMS also allows integration of microﬂudic devices

with PDMS-based biomicroactuators that provide energy to drive both the solid

microstructures and the ﬂuid in the microchip. The low elastic modulus of PDMS

enables bending of a PDMS micropost by the force exerted from a single cardiac

myocyte. Unlike conventional microactuators, for example, PDMS-based cardiomy-

ocyte biomicroactuators do not need an outside electrical power supply or stimulus.

Because of their intrinsic cellular behavior, these biomicroactuators are gaining pop-

ularity as an energy source for biochemical reactors and biosensors [179]. Tanaka

et al. presented the working principles of a PDMS-based cardiomyocyte biomicroac-

tuator that uses PDMS micropillars driven repetitively by spontaneously pulsating

cardiomyocytes as shown in Fig. 4.11. Both ﬂuid and solid microstructures were

successfully actuated utilizing the chemical energy of glucose which is subsequently

converted into mechanical energy by cardiomyocytes. This prototype demonstrated

the future potential of self-actuated and energy efﬁcient microtransducers for in vivo

applications [180].

4.2.4 Case Study

In this section, we give an example of PDMS being employed as a ﬂexible mate-

rial choice for a mechanical sensor for single-cell biomechano analysis. In recent

efforts to measure the cellular traction forces of living cells, researchers developed

4 Additive Processes for Polymeric Materials 209

Micropillar

PDMS

Cardiomyocyte

Fig. 4.11 Actuating

principles of PDMS

micropillars coupled to and

poweredbycultured

cardiomyocyte

a technique that uses polymer micropillar arrays as extracellular matrices that cells

can grow on, as shown in Fig. 4.12. Instead of measuring the cell reaction forces

directly, this approach measures the deﬂection ﬁrst and converts the deﬂections into

reaction forces using appropriate mechanics of the mechanical sensor [144, 153,

181–185].

Fig. 4.12 (a) A SEM image of micropillar array of cellular traction force measurement system;

(b) SEM image showing a smooth muscle cell cultured by on a PDMS micropillar array; and

The low elastic modulus and ﬂexible molding properties render PDMS an ideal

material choice as mechanical sensors f or cellular mechanics study [144, 153, 181–

188]. Some of its unique advantages over other materials include (1) low stiffness

allowing for the measurement of mechanical forces of cells in the piconewtons to

micronewtons level; (2) ability to operate in biological environments, such as culti-

vated liquid or normal saline; and (3) a good biological compatibility, so that cells

can survive on PDMS for a long period of time. The highly compliant and biocom-

patible PDMS meets all the required material characteristics as biomechanosensors.

This is because the mechanical forces studied in cellular mechanics have consid-

erable variation. For example, the forces generated by ﬁbroblasts are on the order

of hundreds of nN [181]; and the contractile forces of cardiac myocytes are mea-

sured in μN[182], whereas the forces generated by single ﬁlament, myosin, and

210 E. Meng et al.

Table 4.2 Spring constant and probing range for microstructures with various geometries

Geometry

D (μm) H (μm)

Spring constant

(nN/μm) Sensitivity (nN)

Probing forces on

the order of

2 4 62.64 12.56 Over a hundred nN

2 6 10.20 2.04 Tens of nN

2 8 4.00 0.80 A few nN

2 12 1.36 0.27 Sub nN

8 24 58.15 11.63 Over a hundred nN

kinesin are only 5–7 pN [189]. Taking advantage of the rheological properties

of PDMS, cantilevers made of PDMS with varying mechanical stiffness are con-

veniently fabricated and tuned with desired aspect ratios [144]. Table 4.2 shows

PDMS cantilevers with various aspect ratios replicated from a single silicon mold.

Figure 4.13 illustrates the probing range with different geometry parameters [144].

Based on the adjusted spring constants, cantilevers with appropriate geometries for

the measurements were selected.

Aspect Ratio

2 4 6 8

4:1

5:1

6:1

1:1

2:1

3:1

10 12 14

Precuring Duration (min)

Fig. 4.13 PDMS cantilevers with various aspect ratios were replicated from a single silicon mold.

Note that the cantilevers with aspect ratios 6:1 or higher collapse due to the low elastic modu-

Institute of Physics)

Here, the force measurement in isolated cardiomyocytes was demonstrated.

Considering that the contraction force of cardiomyocytes is about a few nN, and

the lateral dimension is about 20 μm by 100 μm, cantilevers with aspect ratios of

2:1 were manufactured and employed in cardiomyocyte force measurements. The

resulting cantilever was 4 μm tall. The diameter of the root surface was 2.5 μm, and

the diameter of the top surface was 2 μm with lateral penetration less than 0.1 mm.

The spring constant in the presence of small deﬂection was derived as 62.64 nN/μm.

In order to perform monitoring of cell morphology and mechanical forces, the

cells need to be viable on the stage of an optical microscope for a period of time

4 Additive Processes for Polymeric Materials 211

Vacuum

pump

Inverted microscope

CCD camera

Computer system

for imaging analysis

Buffer solution

Liquid pump

Feedback control

Heating rod

Waste solution

Inlet

Electrical

contact pair

Thermometer

Outlet

Perfusion chamber

Fig. 4.14 Schematic of the in-situ force probing system for living cells (Reprinted with permission

(preferably for a couple of hours). The experimental setup includes a perfusion

chamber, temperature control components, ﬂuidic connections, electrical connec-

tions, and image acquisition and analysis systems as shown in Fig. 4.14.The

deﬂection of the pillars was measured by an inverted phase-contrast microscope

and background noise was removed using MATLAB. A displacement map was cal-

culated using the differences in pillar top positions (Fig. 4.15)[182]. Given that

the magnitude of force exerted by the myocyte is dependent on the number and

the alignment of the sarcomeres (the contractile unit of the myocyte), the vector

component along the length of the sarcomeres is expected to be greater than the

vector component transverse to the length of the sarcomeres. This conforms to the

Fig. 4.15 The deformation vector maps and corresponding force vector maps obtained at systole

212 E. Meng et al.

physiological behavior of cardiomyocytes [182]. The amplitude of the force varia-

tion is a direct index of myocardial performance, which directly affects the blood

pressure and pumping volume between the diastole and systole stages.

4.3 Polyimide

Polyimides have long been used in the microelectronics industry as both an insulator

and packaging material. In particular, polyimides have been employed extensively

as a planarization layer in forming multilevel interconnects [190–192] and for form-

ing multichip modules [193]. Since the synthesis of the ﬁrst aromatic forms of

polyimide in 1908, advances in polyimide chemistry have led to its availability in

a variety of forms including bulk (ﬁlm or tapes with pressure-sensitive adhesive) or

spin-on versions (photodeﬁnable and nonphotodeﬁnable) [194]. These are commer-

cially available from HD MicroSystems (Parlin, NJ) and DuPont (Wilmington, DE).

Polyimides were ﬁrst introduced to MEMS as a ﬂexible substrate material for sen-

sor arrays [195] and multielectrode arrays [196, 197], both biomedical applications.

Earlier reviews of polyimide in MEMS include [198, 199].

Polyimides may possess either an aliphatic (linear) or aromatic (cyclic) chemical

structure and exhibit either thermoset or thermoplastic behavior (Fig. 4.16)[194].

To obtain the ﬁnal chemical structure of polyimide, a polyamic acid precursor is

imidized by baking at elevated temperatures (∼300–500

◦

C) [192, 195]. Polyamic

acid is soluble in polar inorganic solvents (NMP, dimethyl formamide (DMF) and

dimethylsulfoxide (DMSO)). The imidization process removes this solvent and in

aromatic versions, closes the ring structure [200].

Fig. 4.16 Chemical structure

of polyamic acid (left)and

polyimide (right)inwhich

the R



and R



may be

aliphatic or aromatic

(Adapted from [192])

4.3.1 Material Properties

Early applications of polyimide as a MEMS material capitalized on its low Young’s

modulus and ability to serve as a ﬂexible substrate with potential biocompatibil-

ity and biostability for in vivo biomedical applications [195–197]. Its inertness

[201] and low cytotoxicity [202] have been demonstrated in preliminary stud-

ies. Biocompatibility of polyimide-supported implantable microelectrodes has been

demonstrated [203], however, many manufacturers expressly prohibit the use of

their polyimides in implantable devices. Furthermore, polyimide is stiffer in com-

parison to other thin-ﬁlm polymers (such as Parylene and PDMS) and has led to

damage of delicate neural tissues [204, 205].

4 Additive Processes for Polymeric Materials 213

Other notable features of polyimides include high glass transition tempera-

ture, high thermal and chemical stability, low dielectric constant, high mechanical

strength, low moisture absorption, and high solvent resistance [194, 199, 206].

This combination of features has led to the use of polyimides as replacements for

ceramics [194 ], as chemically resistant electroplating masks [207], and as sacri-

ﬁcial layers [208]. Further review of the electrical and mechanical properties are

available in [200, 209] and modiﬁcation of electrical properties through the addi-

tion of graphite particles is described in [210]. Chemical modiﬁcation of polyimide

chemistry yields photosensitive versions having negative properties [207], however,

these suffer from signiﬁcant shrinkage during the imidization process. Thus, etching

of standard nonphotodeﬁnable polyimides yields higher resolutions [200, 207].

4.3.2 Processing Variations

4.3.2.1 Removal of Polyimide

Chemical removal of both cured and uncured polyimides is possible although wet

removal of cured polyimides is difﬁcult.

• Cured polyimides: hot bases and strong acids [200], H

(selective over

silicon nitride and oxide) [211].

• Uncured polyimides: bulk removal potassium hydroxide (KOH) (5–30%) [199,

207, 212, 213] and selective removal in KOH with photoresist mask [200].

• Dry etching is more effective for removing polyimide, including when used as a

sacriﬁcial layer [208, 214].

• Chemistries: O

plasma [215]or O

+ CF

,CHF

,orSF

plasmas [199, 216–

221].

• Etch masks: Al [199, 201, 208, 217, 219, 221–223], Cr/Au [218], PECVD silicon

nitride [224], oxide [208], SiC [208].

Process optimization (gas concentration, plasma power, and ambient pressure)

enables control of etched sidewall angles [220, 225–227] and high aspect ratios are

possible with an electron–cyclotron resonance source [228]. Uncured polyimides

may also be removed by dry etching [199].

Alternative dry removal processes include focused ion beam (FIB) [217] and

excimer laser machining [213, 229–232].

4.3.2.2 Release of Polyimide

The simplest release technique is to peel polyimide away from Si wafers [201, 223,

233, 234]. Flexible electrode arrays are released from substrate carriers in this man-

ner (Fig. 4.17). A number of sacriﬁcial materials have also been used to promote

release: