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

Подождите немного. Документ загружается.

224 E. Meng et al.

Fig. 4.24 Chemical structure of poly (p-xylylene)s (Original ﬁgure from [346], used with

permission of Institute of Physics Publishing Ltd. 2008)

Table 4.3 Comparison of parylene types and their key features

Parylene type Key features

N High crevice penetration, lubricity,

excellent electrical properties

C Low permeability barrier, excellent

combination of electrical and

mechanical properties

D High thermal stability

HT Superior thermal, barrier, and electrical

properties

A and AM Amino group convenient for bonding to

biomolecules

Fig. 4.25 Summary of deposition process (After [350])

4.5.1 Material Properties

Parylene is chemically inert, coats surfaces conformally, and is an excellent poly-

mer barrier [339]. Historically, it has been used as a medical coating, an insulator

4 Additive Processes for Polymeric Materials 225

for implantable neural wire probes [351–354], an interlayer dielectric [355], and

was later introduced to the MEMS community in 1997 as a coating material for

ﬂuidic interconnects [356]. Parylene has gained popularity as a standard MEMS

material for its deposition method (vacuum-based), pinhole-free deposition, low

process temperature, transparency, and compatibility with microfabrication pro-

cesses [342]. Since its introduction to MEMS, Parylene was explored as a structural

material [357] in a wide variety of MEMS applications [123, 204, 205, 329,

358–370]. Parylene in carbonized/pyrolyzed [371–373] and ion-implanted [374]

forms have also found applications in MEMS as a sacriﬁcial and sensing mate-

rial. Further modiﬁcation of Parylene surfaces has been demonstrated, especially

for biological/biomedical applications [375–377].

4.5.2 Processing Techniques

Parylene is resistant to removal by solvents below its melting temperature. Only

chloronaphthalene or benzoyl benzoate above 150

◦

C are effective [378]. Liftoff

is not possible due to the conformal structure obtained during the deposition

process.

Physical processes are best for removing Parylene. Plasma processes effectively

remove both Parylene N and C; the removal mechanism is discussed in [379, 380].

Plasma etching [335, 346, 358, 377], reactive ion etching [346, 355, 381], reac-

tive ion beam etching [382], high-density plasma etching [383], and Bosch-like

switched chemistry etching [346, 384, 385] have all been demonstrated. Oxide and

metal masks may redeposit during the etch process and result in rough surfaces or

the formation of micrograss [346, 355, 381]. Spin-on glass, nitride, and sputtered

a-Si are also mediocre masking materials [346, 383]. Photoresist masks are pre-

ferred, however, the etch rate is comparable to that of Parylene and thus exhibits

low selectivity [346]. SU-8 masks exhibit higher selectivity although its removal

after etching was not required in the single study to date [386]. Anisotropic sidewall

proﬁles are possible by using high-density plasma and switched chemistry etching

[346, 383–385].

Other methods to remove Parylene include ultraviolet laser ablation [353, 354]

and manual removal [352]. Release agents (such as 2% Micro



lab cleaning solu-

tion, International Product Corp.) applied to surfaces allows Parylene to be peeled

away after deposition. Peeling of Parylene without damage to the ﬁlm was also

demonstrated on Si surfaces having a native oxide layer; immersion in water can

facilitate the release [204]. Photoresist sacriﬁcial layers are practical for the release

of smaller Parylene structures [361]. As removal can pose fabrication challenges

in some applications, alternative patterning methods were investigated. Selective

deposition of Parylene is achieved by controlling substrate temperature [387].

This technique exploits the phenomenon that deposition thickness is a function

of substrate temperature; heating the substrate (70

◦

C) selectively limits Parylene

deposition to cooler regions [388, 389].

226 E. Meng et al.

Parylene’s thermoplastic nature also enables thermal imprint patterning (Ni

molds at 250

◦

C) [390]. Micromolding combined with thermocompression bonding

has been used to form channels [370, 391]. Other modes of Parylene-to-Parylene

and Parylene-to-substrate bonding were also developed [392, 393].

4.5.3 Lessons Learned

Adhesion to most substrates (such as silicon) requires pretreatment with adhesion

promotion A-174 (gamma-methacryloxypropyltrimethoxysilane) prior to deposi-

tion. However, A-174 is incompatible with common positive photoresists causing

dissolution of photoresist patterns during the adhesion promotion process. It is

suggested that a short oxygen plasma treatment can promote Parylene-to-Parylene

adhesion by roughening the surface [358]. To further improve Parylene-to-Parylene

adhesion, especially in wet biomedical applications, thermal annealing was per-

formed (2 days at 200

◦

C in vacuum) [204]. Robust Parylene-to-substrate adhesion

was promoted by surface roughening with BrF

or XeF

gas phase etching [358,

394, 395] of Si or by anchoring Parylene to the substrate by deposition into etched

trenches [366, 396–398].

4.5.4 Case Study

Implantable microelectrode arrays and telemetry coils were fabricated using single-

and dual-layer Parylene/metal processes to realize an intraocular retinal prosthe-

sis (Fig. 4.26)[204, 205, 399]. Parylene was selected over other polymers for its

ﬂexibility, process compatibility, and biocompatibility. Fabricated electrode arrays

and coils, however, cannot tolerate accelerated lifetime testing due to water inﬁl-

tration between the Parylene-to-Parylene interface. Thermal annealing was found to

Fig. 4.26 Parylene multielectrode array with biomimetic arrangement of electrodes. The close-

up shows an array that has been annealed and heat formed to match the curvature of the retina

4 Additive Processes for Polymeric Materials 227

improve Parylene-to-Parylene adhesion (200

◦

C, 2 days, vacuum with nitrogen back-

ﬁll) and provide an extrapolated mean time to failure exceeding 20 years following

accelerated lifetime testing [399]. As the interface between the Parylene array and

the tissue is curved, thermoforming of the thin-ﬁlm devices was performed using

aluminum molds to impart an anatomically matched curvature to the array portion.

In electronic neuroprostheses, it is desirable to fabricate electrodes using materials

with high charger delivery capacity. Ir is thus preferred over Pt. However, the high

melting temperature of Ir results in cracking of ﬁlms deposited in Parylene C. In

this case, Parylene HT, which has greater thermal stability, was used to successfully

fabricate Parylene HT–Ir electrode arrays [204].

4.6 Conductive Polymers

Polymers are usually considered as solely having plastic properties. They are excel-

lent insulators and have been extensively used as coatings to protect electrical wires

from short-circuits. However, Alan J. Heeger, Alan G. MacDiarmid, and Hideki

Shirakawa changed this view with their discovery of a new class of conducting

polymers in 1974 for which they were awarded the Nobel Prize in Chemistry in

2000. Although this class of polymer is in its infancy, their potential applications

have signiﬁcant implications.

Polyacetylene was the ﬁrst conducting polymer. The key feature of the polymer

is that it consists of alternating single and double bonds, called conjugated double

bonds along the backbone of the polymer chain [231] (Fig. 4.27). The π-electron in

a double bond has less energy and can be delocalized.

Fig. 4.27 Polyacetylene

structure

However, it is not enough to have conjugated double bonds. To become electri-

cally conducting, the polymer has to be disturbed: either by removing electrons from

(oxidation), or inserting them into (reduction), the material. The process is known

as doping. In the doped state, the polymer can achieve conductivity even compa-

rable to metals, which is an increase of about 13 orders of magnitude compared to

common nonconducting polymers [400] (Fig. 4.28).

Although featuring the highest conductivity among other conducting polymers,

polyacetylene is not stable, which limits its use. In the next three decades following

the development of polyacetylene, a wide range of derivatives were developed, such

as polypyrrole, polythiophene, and polyaniline [402] (Fig. 4.29).

228 E. Meng et al.

Fig. 4.28 Conductivities of conducting polymers relative to other common polymers and inor-

ganic materials. PE is polyethylene, PVC is polyvinylchloride and PTFE is polytetraﬂuoroethylene,

also known as Teﬂon

. The upper limits in conductivity for the conducting polymers polyaniline,

polypyrrole, and polyacetylene are shown (Reprinted from [401] with permission of, the MIT

Technology Licensing Ofﬁce)

PolythiophenePolyaniline

Polyfuran

Polyphenylene

sulfide

PolypyrrolePolyacetylene

Fig. 4.29 Types of conducting polymers (Reprinted from [402] with permission from ASME)

4.6.1 Material Properties

Some of the properties of conducting polymers can be controlled by changing the

redox state. By an electrochemical process, electrons can be added to or removed

from the polymer via an electrode, thereby changing the oxidation state. More

important, t his change is reversible through chemical or electrochemical means. For

example, the conductivity can be switched by 13 orders of magnitude, an effect

used in organic transistors. Optical absorption, permeability, hydrophobicity, stored

charge, and volume all change in a controllable manner. This enables devices such

as chemical sensors, ﬁlters, capacitors, and batteries [403].

Here, we focus on the actuator applications of conducting polymer, where the

volume change relies on the ion and solvent migrations [404–407]. By applying

a positive potential, electrons are removed from the polymer via an electrode and

the polymer is oxidized. To compensate, oppositely charged ions are incorporated

4 Additive Processes for Polymeric Materials 229

into the polymer from the supporting electrolyte and thus the volume expands. The

polymer can be reversibly reduced to neutral state by applying a more negative

potential and allowing the polymer to shrink to the original volume.

Although electrostatic and piezoelectric materials have major technological

importance for the direct conversion of electrical energy to mechanical energy in

actuators, conducting polymers, as an alternative, could provide a similar function

in a manner analogous to natural muscle. Key features include:

1. Large strain (2∼20%)

2. High stress (100 times greater than muscle)

3. Low actuation voltage (∼1 V or less)

4. Can be positioned continuously between minimum and maximum values

5. Can be kept in a ﬁxed position without consuming additional power; consumes

current only when switching states

6. Can operate in liquid electrolytes, including body ﬂuids

However, conducting polymer-based actuators have a low actuation rate and

low conversion efﬁciency of electrical energy to mechanical energy. These can be

improved through miniaturization (fabrication of conducting polymer-based devices

on a smaller scale). Miniaturization could lead to improvements in conductivity and

electroactivity [408]. Higher conductivity is important, but reducing resistive loss is

the key requirement. One attractive feature of miniaturization is the possibility of

having highly ordered and less defective material to contribute to high conductivity.

In addition, the efﬁciency of the electroactive process through a redox cycle can

be improved because the surface area to volume ratio is higher, providing a large

interface between the electrode and electrolyte. This can increase the rate-limiting

ion diffusion process, and reduce the double-layer capacitance and the RC time

constant.

4.6.2 Actuation Mechanism and Theories

During the last few decades, researchers have proposed many models to describe

the actuation mechanism of conducting polymer actuators. The complex nature

of the actuation mechanism involves a large number of electrical, chemical, and

mechanical variables, some of which are interrelated. Madden, Madden, and Hunter

described the actuation of a free-standing polypyrrole ﬁlm as a simple elastic

mechanical model superimposed with the electrochemically generated strain that

is assumed to be directly proportional to the charge density [409]. Della Santa, De

Rossi, and Mazzoldi investigated the modeling and characterization of a muscle-

like conducting polymer axial/linear actuator operating in an electrolytic cell [410].

Their model utilized simple lumped parameters identiﬁed using force and length

change data. Based on the beam-bending model, Pei and Inganaes developed a

mathematical model for a bilayer strip made of a gold-coated polyethylene layer

230 E. Meng et al.

and a polypyrrole layer [411]. It was concluded that the actuation consisted of

a fast cation insertion for volume expansion and a subsequent slow salt draining

process for phase relaxation. Alici et al. developed a mathematical model of a strip-

type polypyrrole/PVDF/polypyrrole trilayer actuator operated in air, using a similar

beam-bending method [412]. The static characteristics of the actuator were ana-

lyzed and veriﬁed by experiments, such as the relationship among the input voltage,

output deﬂection, and force.

4.6.3 Applications

4.6.3.1 Actuators

Conducting polymer actuators are commonly fabricated in two forms: linear and

bilayer bending actuators. For linear actuators, Hara et al. built a polypyrrole–zigzag

metal wire composite linear actuator and showed an enhanced electrochemical

strain up to 21.4% [413]. Spinks et al. developed a helix tube polypyrrole lin-

ear actuator for a Braille display screen [414]. The bilayer actuator is a simple

mechanical ampliﬁcation structure that can convert a small linear strain into a

large bending angle. One of the earliest demonstrations was carried out by Smela

et al. with a polypyrrole/gold bilayer actuator to produce an electrically con-

trolled micro ﬁnger [415]. This actuator, however, could only be used i n a liquid

electrolyte.

A further improvement was made by Kaneto et al. to enable the actuator to oper-

ate in air [416]. They used an insulating paper soaked with the electrolyte or a solid

polymer electrolyte sandwiched between the two conducting polymer layers. This

type of actuator has some limitations: either the ionic conductivity in solid polymer

electrolyte is very low at room temperature, or the solvent in the solvent-swollen

gel, typically water, will evaporate in the long run.

A promising approach to dry actuators has become feasible with the synthesis of

new molecules called ionic liquids. These materials are salts that are liquid at room

temperature; they typically consist of nitrogen-containing organic cations, such as

1-butyl-3-methyl imidazolium, and inorganic anions (Fig. 4.30)[417]. Furthermore,

some ionic liquid relatives are plastic solids at room temperature, yet still maintain

a reasonable conductivity, whereas others can be transformed into soft elastomeric

solids at room temperature by the addition of small amounts (∼5%) of a suitable

polymer. These materials have only recently been applied to conducting polymer

actuators, but the results are impressive: they show signiﬁcantly enhanced lifetimes

Fig. 4.30 Structure of the ionic liquid BF

1-butyl-3-methyl imidazolium (Reprinted from [417]

with permission from Wiley-VCH)

4 Additive Processes for Polymeric Materials 231

compared to other electrolytes, as well as fast switching speeds [418]. The com-

bination of gels with ionic liquids is therefore the next logical step in the effort to

develop dry actuators.

4.6.3.2 Conducting Polymer as a Strain Gauge Material

Conducting polymers have also been reported to have a strain-sensitive effect,

where the resistance changes as the gauge is stretched. The resistance change is

given by:

R

= Gε,

where R

is the unstrained resistance of the gauge and  is the strain. The factor

G is known as the gauge factor of the sensor. A higher gauge factor improves the

sensitivity of the strain gauge. Pure polypyrrole has been used to make strain gauges,

however, the best conducting polymer strain gauges have been made by coating a

ﬂexible fabric with a l ayer of polypyrrole; the reported gauge factor was around 13

[419].

4.6.4 Processing Techniques

4.6.4.1 Deposition

Depositing thin, uniform polymer ﬁlms by spin-coating is a widely used tech-

nique in microfabrication. This technique can be applied to photoresist, such as

Shipley 1818. For metal layers, such as Cr, Au, Ti, an E-beam evaporation or ther-

mal evaporation can be used. E-beam evaporation has a small effect on samples

thus making it easier to pattern by etching or a lift-off process; thermal evapora-

tion will raise the sample temperature thus it is not suitable for a lift-off process

due to overheating of the photoresist and difﬁculty stripping the photoresist after

evaporation.

There are many ways to deposit conducting polymers. Polyaniline can be made

into powder and dissolved or dispersed in an acid solution [420]. Polythiophene can

be dissolved in certain types of organic solvent by special treatment such as alkyl

substitution on carbon rings [421]. Then they can be spin-cast into ﬁlms onto many

types of substrates. If the polymer is insoluble, such as polypyrrole, the electro-

chemical deposition technique is a common alternative. This process is conducted

in a typical three-electrode electrochemical cell setup, which consists of a working

electrode (where the conducting polymer deposits), a counterelectrode, and a ref-

erence electrode (providing precise potential control of the synthesis). The electric

power can be applied by means of either potentiostatic (constant potential) [405,

422] or galvanostatic (constant current) mode [423].

232 E. Meng et al.

4.6.4.2 Patterning

There are several ways to pattern conducting polymers. The main methods are listed

as follows (Fig. 4.31)[424].

1. Patterning the electrode onto which the conducting polymer is deposited.

2. Using photoresist as a mold to template deposition.

3. Depositing the conducting polymer in the whole layer then etching it.

4. Printing the polymer (this only works with soluble conducting polymers).

Fig. 4.31 Patterning methods: (a) patterned electrodes; (b) template deposition; and (c) etching

(Reprinted from [424] with permission from IOP)

Patterning electrodes is the simplest way because conducting polymer will

directly deposit on a substrate with no need of further patterning. The main draw-

back is the nonuniform ﬁlm thickness; higher electric ﬁelds and reactant availability

at the edges of electrodes produce thicker deposits. Also the polymer has ﬁnite lat-

eral growth, which may bridge adjacent electrodes or make it harder to release the

structure after deposition.

The second method is template deposition, using photoresist as a mold. The con-

ducting polymer will only deposit in the opening windows, and photoresist can be

subsequently removed with acetone or ethanol. This method can eliminate the lateral

growth in patterned electrodes to beneﬁt the structure release.

Conducting polymers can also be patterned by removal of unwanted material

by reactive ion etching (RIE). First, the polymer is deposited on the entire surface

and then the photoresist is patterned to form an etch mask. Finally, RIE is applied

(usually oxygen plasma). Because oxygen will etch both the photoresist and con-

ducting polymer, the etching process should be carefully monitored. Christophersen

et al. reported that the oxygen plasma etch selectivity (ratio of etch rates) between

polypyrrole and the masking 2 μm thick photoresist was 2.5:1 [425]. However, our

experiments (40 sccm O

ﬂow rate, 100 mtorr pressure, 300 W power) demonstrated

an S1818 etch rate of 0.18 μm/min, whereas that of polypyrrole was 0.20 μm/min.

Thus, this method limits the conducting polymer thickness to be less than the

photoresist thickness.

4.6.4.3 Release

Microactuators require a method to partially release them from the substrate. As

shown in Fig. 4.32 [424], a sacriﬁcial layer is commonly used. But for this method,

4 Additive Processes for Polymeric Materials 233

Fig. 4.32 Release methods: (a) sacriﬁcial layer and (b) differential adhesion (Reprinted from

[424] with permission from IOP)

the overhang should not be very large to avoid anchor failure during actuation. Also,

the undercut release of large areas requires long etch times.

The second method uses differential adhesion. It is well known that gold adheres

weakly to bare silicon, silicon dioxide, and glass. Therefore a thin layer of metal

such as chromium is normally deposited ﬁrst as an adhesion-promoting layer. By

patterning the adhesion layer only at the anchor parts, the structure on the bare

silicon can be released with the stress generated by the activated conducting polymer

layer.

4.6.4.4 Process Considerations

Although there is great ﬂexibility in fabrication techniques, conducting polymers are

nevertheless s ensitive to some chemicals. Therefore care must be taken to devise

an appropriate process. Table 4.4 summarizes some chemical compatibilities and

processes [424].

Table 4.4 Processing hazards

Damaging Use care, test ﬁrst Harmless

Cr etchant Acids Au etchant

High temperatures Hot plate at 100

◦

Developer (KOH),

bases

Solvents Photoresist

Resist

stripper/remover

UV light in mask

aligner

Reprinted from [424] with permission from IOP

4.6.5 Case Study

In this case study, we design two structures with detailed processes based on the

aforementioned fabrication techniques (Fig. 4.33). The conducting polymer we

use is dodecylbenzene sulfonate (DBS

–

)-doped polypyrrole. The synthesis is con-

ducted in 0.1 M pyrrole monomer (Sigma Aldrich) and 0.1 M NaDBS (Sigma

Aldrich) aqueous solution, by applying a constant potential of 0.5 V versus Ag/AgCl

reference electrode.