Bhushan B. Nanotribology and Nanomechanics: An Introduction

Подождите немного. Документ загружается.

1214 Bharat Bhushan

Millipede concept Single tip

Approaching/ thermal sensors

32× 32 array

Chip heater

2 mm

200μm

20μm

2μm

Fig. 22.6. Example of a NEMS device – AFM-probe-based nanoscale data storage system –

that experience tribological problems

tional multigrain magnetic rigid disk. The superparamagnetic eﬀect poses a serious

challenge for ever increasing areal density of disk drives. One of the promising

methods to circumvent the density limitations imposed by this eﬀect is the use of

nanopatterneddisks Fig. 22.7b.In conventionaldisks, the thin magnetic layer forms

a random mosaic of nanometer-scale grains and each recorded bit consists of many

tens of these random grains. In patterned disks, the magnetic layer is created as an

ordered array of highly uniform islands, each island capable of storing an individual

bit. These islands may be one or a few grains, rather than a collection of random

decoupled grains. This increases the density by a couple of orders of magnitude.

Figure 22.7c shows a schematic of an inductive-write GMR read head structure.

These are constructed from a variety of materials: magnetic alloys, metal conduc-

tors, ceramic, and polymer insulators in a complex three-dimensionalstructure. The

multilayered thin-ﬁlm structures used to construct the sensor and individual ﬁlms

are only a few nm thick. The head slider surface, which ﬂies over the disk surface,

is coated with diamond-like carbon coatings that are about 3 nm thick to protect the

thin-ﬁlm structure from electrostatic discharge. Any isolated contacts between the

disk and sensor and lubricant pickup pose tribological concerns [34].

22 Characterization of MEMS/NEMS and BioMEMS/BioNEMS 1215

Fig. 22.7. Schematic of

(a) sectional view of a con-

ventional multigrain mag-

netic rigid disk, (b) nanopat-

terned magnetic rigid disk,

and (c) an inductive-write

GMR read magnetic head

structure for magnetic data

storage (Hitachi Global Stor-

age Technologies)

BioMEMS

An example of a wristwatch-type biosensor based on microﬂuidics, referred to as

a lab-on-a-chip system, is shown in Fig. 22.8a [60,69]. These systems are designed

to either detect a single or class of (bio)chemicals, or system-level analytical cap-

abilities for a broad range of (bio)chemical species known as a micro total-analysis

system (µTas), andhavethe advantageof incorporatingsamplehandling,separation,

detection, and data analysis onto one platform. The chip relies on microﬂuidics and

1216 Bharat Bhushan

70–80μm

Magnetic beads

with dendrimer

Biofilter with immunosensor Microvalve

Control system and circuits

Flow sensor

Embedded

biofluidic

chips

Lab-on-a-chip

(Tang and Lee, [69])

75 mm

Cassette type

(Taylor et al., [70])

Wash Valve Filter region Elution chamber

Elution pump

PCR reaction tube

Post-PCR outlet

Post-PCR-pumpWasteFilter pump

Sample

Fig. 22.8. (a) MEMS-based bioﬂuidic chip, commonly known as a lab-on-a-chip, that can

be worn like a wristwatch [69]. (b) Cassette-type biosensor used for human genomic DNA

analysis [70]

involvesmanipulationof tinyamounts ofﬂuids in microchannelsusing microvalves.

The test ﬂuid is injected into the chip, generally using an external pump or syringe,

for analysis. Some chips have been designed with an integrated electrostatically ac-

tuated diaphragm-typemicropump.The sample,which can havea volumemeasured

in nanoliters, ﬂows through microﬂuidicchannels via an electric potential and capil-

lary action using microvalves(havingvariousdesigns includingmembrane type)for

variousanalyses. The ﬂuid is preprocessed and then analyzed usinga biosensor. An-

other exampleof a biosensoris the cassette-type biosensor used for human genomic

DNA analysis; integrated biological sample preparationis shown in Fig. 22.8b [70].

The implementation of micropumps and microvalves allows for ﬂuid manipulation

and multiple sample-processing steps in a single cassette. Blood or other aqueous

solutions can be pumped into the system, where various processes are performed.

Microvalves, which are found in most microﬂuidic components of BioMEMS,

can be classiﬁed into two categories: active microvalves (with an actuator) for ﬂow

regulation in microchannels and passive microvalves integrated with micropumps.

22 Characterization of MEMS/NEMS and BioMEMS/BioNEMS 1217

Active electrostatic microvalve

Electrode Plates

Dielectric

Base Plate

Silicon

Closure Plate

Exhaust

Inlet Orifice

Passive mechanical check valves

Valveless micr

opump

Valve 1 Valve 2Diaphragm Piezo

disk

Diaphragm micropump

Inlet Outlet

Piezo discPiezo bimorph cantilever

Membrane valveFlap valve

Inlet

SpacerMembrane

Outlet

Pressure chamber

Upper

electrode

Lower

electrode

Electrostatic micropump

Rotary micropumps

(Ohnstein et al., 1990)

(Woias, [111])

(Bustgens et al., 2001)

(Stehr et al., [110]) (Nguyen et al., [112])

(Ohnstein et al., 1990)

(Ahn and Allen, 1995)

(Doepper et al., 1997)

Fig. 22.8. (continued) (c) Multiple examples of valves and pumps found in BioMEMS de-

vices. Mechanical check valves, diaphragm micropump, valveless micropump, and rotary

micropump

Active microvalves consist of a valve seat and a diaphragm actuated by an external

actuator [61,109]. Diﬀerent types of actuators are based on piezoelectric, electro-

static, thermopneumatic, electromagnetic, and bimetallic materials, shape-memory

alloys and solenoid plungers. An example of an electrostatic cantilever-type active

microvalveis shown in Fig. 22.8c. Passive microvalvesused in micropumpsinclude

mechanical check valves and a diﬀuser/nozzle [61, 109–112]. Check valves con-

sist of a ﬂap or membrane that is capable of opening and closing with changes in

pressure; see Fig. 22.8c for schematics. A diﬀuser/nozzle uses an entirely diﬀerent

principle and only works with the presence of a reciprocating diaphragm.When one

convergentchannel works simultaneously with another convergentchannel oriented

in a speciﬁc direction, a change in pressure is possible.

There are four main types of mechanical micropumps,which include diaphragm

micropumps that involves mechanical check valves, valveless rectiﬁcation pumps

that use diﬀuser/nozzle-type valves, valveless pumps without a diﬀuser/nozzle,

1218 Bharat Bhushan

electrostatic micropumps, and rotary micropumps [61, 109–112]. Diaphragm mi-

cropumps consist of a reciprocating diaphragm, which can be piezoelectrically

driven, working in synchronization with two check valves Fig. 22.8c. Electrostatic

micropumps have a diaphragm as well, but it is driven using two electrodes. Valve-

less micropumps also consist of a diaphragm that is piezoelectrically driven, but

do not incorporate passive mechanical valves. Instead, these pumps use an elastic

buﬀermechanismor variablegap mechanisms.Finally, a rotarymicropumphas aro-

tating rotor that simply adds momentum to the ﬂuid by the fast-moving action of the

bladesFig. 22.8c. Rotarymicropumpscan be drivenusing an integratedelectromag-

netic motor or by the presence of an external electric ﬁeld. All of these micropumps

can be made of silicon or a polymer material.

During the operation of the microvalves and micropumps discussed above, ad-

hesion and friction properties become important in which contacts occur due to rel-

ative motion. During operation, active mechanical microvalves have an externally

actuated diaphragm which comes into contact with a valve seat to restrict the ﬂuid

ﬂow. Adhesion betweenthe diaphragm and valveseat will aﬀect the operation of the

microvalve.In diaphragm micropumps,two passive mechanicalcheck valves are in-

corporated into the design. Passive mechanical check valves also exhibit adhesion

when the ﬂap or membranecomes into contact with the valve seat when ﬂuid ﬂow is

removed.Adhesion also occurs during the operation of valveless micropumpswhen

the diaphragm, which is piezoelectrically driven, comes into contact with the rigid

outlet. Finally, adhesion and friction can also be seen during the operation of rotary

micropumps when the gears rotate, come into contact and rub against one another.

If the adhesion between the microchannel surface and the bioﬂuid is high, the

biomolecules will stick to the microchannel surface and restrict ﬂow. In order to

facilitate ﬂow, microchannel surfaces with low bioadhesion are required. Fluid ﬂow

in polymer channels can produce a triboelectric surface potential which may aﬀect

the ﬂow. Polymersare knownto generatesurfacepotentialsand themagnitudeof the

potential varies from one polymer to another [113–115]. Conductive surface layers

on the polymer channels can be deposited to reduce triboelectric eﬀects.

As mentioned, the microﬂuidic biosensor shown in Fig. 22.8a required the use

of micropumps and microvalves. For example, a microdevice with 1000 channels

requires 1000 micropumps and 2000 microvalves, which makes it bulky and poses

reliability concerns. Two methods can be used to drive the ﬂow of ﬂuids in mi-

crochannels: pressure and electrokinetic drive. The electrokinetic ﬂow is based on

the movement of molecules in an electric ﬁeld due to their charges. There are two

components to electrokinetic ﬂow: electrophoresis, which results from the accel-

erating force due to the charge of a molecule in an electric ﬁeld, and electroos-

mosis, which uses electrically controlled surface tension to drive uniform liquid

ﬂow. Biosensors based on electrokinetic ﬂow have also been developed.In so-called

digital-based microﬂuidics, based on the electroosmosis process, electrically con-

trolled surface tension is used to drive liquid droplets, thus eliminating the need

for valves and pumps [116,117]. These microdevices consist of a rectangular grid

of gold nanoelectrodes instead of micro/nanochannels. An externally applied elec-

22 Characterization of MEMS/NEMS and BioMEMS/BioNEMS 1219

tric ﬁeld enables manipulation of samples of a few nanoliters through the capillary

circuitry.

An exampleof a microarray-typebiosensorunder developmentin our laboratory

is that based on a ﬁeld-eﬀect transistor (FET), which is shown in Fig. 22.9 [72,118].

FETs are sensitive to the electrical ﬁeld produced due to the charge at the surface

of the gate insulator. In this sensor, the gate metal of a metal–oxide–semiconductor

ﬁeld-eﬀect transistor (MOSFET) is removed and replaced with a protein (recep-

tor layer) whose cognate is the analyte (e.g., virus or bacteria) that is meant to be

sensed. The binding of the receptor layer with the analyte produces a change in

the eﬀective charge, which creates a change in the electrical ﬁeld. This electrical

ﬁeld change may produce a measurable change in the current ﬂow through the de-

vice. Adhesion between the protein and silica substrate aﬀect the reliability of the

biosensor. In the case of implanted biosensors, the biosensors come in contact with

the exterior environment such as tissues and ﬂuids, and any relative motion of the

sensor surface with respect to exterior environment such as tissues or ﬂuids may

result in surface damage. A schematic of friction and points of wear generation

when an implanted biosensor surface comes into contact with living tissue is shown

in Fig. 22.9b. The, friction, wear, and adhesion of the biosensor surface may be

critical in these applications [72,119,120].

Field

insulator

(SiO

)

Source

metal

(Al)

Receptor

biological

molecules

Drain

metal

(Al)

–

Substrate

source Gate

insulator (SiO

)

drain

Point of friction

Silicon dioxide

Biomolecule layer

Cell cytoplasm

Cell membrane Integral

protein

Hydrophobic

helix

Lipid-linked protein

Fig. 22.9. (a) Schematic of

MOSFET-based bioFET sen-

sor [72], and (b) schematic

showing the generation of

friction and wear points

due to interaction of im-

planted biomolecule layer on

a biosensor with living tissue

1220 Bharat Bhushan

Anchor region

Cell active area

Combination beam array

Anchor regions

Combination beam array

Cell active

area

Combination beam array

Anchor regions

100 μm

Fig. 22.10. Schematic of

two designs for polymer

bioMEMS structures to meas-

ure cellular forces [121]

Polymer BioMEMS are designed to measure cellular surfaces. For two exam-

ples, see Fig. 22.10 [121]. The device on the left shows cantilevers anchored at the

periphery of the circular structure, while the device on the right has cantilevers an-

chored at the two corners on the top and the bottom.The cell adheres to the center of

the structure, and the contractile forces generated in the cells cytoskeleton cause the

cantilever to deﬂect. The deﬂection of the compliant polymer cantilevers is meas-

ured optically and related to the magnitude of the forces generated by the cell. Ad-

hesion between cells and polymer beam is desirable. In order to design the sensors,

micro- and nanoscale mechanical properties of polymer structures are needed.

BioNEMS

Micro/nanoﬂuidic devices provide a powerful platform for electrophoretic separa-

tions for a variety of biochemical and chemical analysis. Electrophoresis is a versa-

tile analytical method which is used for separation of small ions, neutral molecules,

and large biomolecules. Figure 22.11 shows an interdigitated micro/nanoﬂuidic

silicon array with nanochannels for the separation process. Figure 22.12a shows

a schematicof an implantable,immuno-isolationsubmicroscopicbiocapsules,aimed

at drug deliveryto treat signiﬁcant medical condition such as type I diabetes [76,77].

The purpose of the immuno-isolationbiocapsules is to create an implantable device

capable of supporting foreign living cells that can be transplanted into humans. It

is a silicon capsule consisting of two nanofabricated membranes bonded together

with the drug (e.g., encapsulated insulin-producingislet cells) contained within the

cavities for long-term delivery. Pores or nanochannels in a semipermeable mem-

brane as small as 6nm are used as ﬂux regulatorsfor the long-term release of drugs.

The nanomembrane also protects therapeutic substances from attack by the body’s

immune system. The pores are large enough to provide the ﬂow of nutrients (e.g.,

glucose molecules) and drug (e.g., insulin), but small enough to block natural anti-

bodies. Antibodies have the capability to pass through any oriﬁce larger than 18 nm.

The 50-nm pores in silicon were etched by using the sacriﬁcial-layer lithography

described in Sect. 22.A [77].

The main reliability concerns in the micro/nanoﬂuidic silicon array and im-

plantable biocapsules are biocompatibility and potential biofouling(undesirable ac-

22 Characterization of MEMS/NEMS and BioMEMS/BioNEMS 1221

Nanochannels for separation process

Interdigitated micro/nanofluidic array

Micro-

channels

50 μm

Nano-

channel

Exploded view of silicon nanochannels

1 μm

Fig. 22.11. Interdigitated

micro/nanoﬂuidic silicon

array for the separation pro-

cess [118]

cumulation of microorganisms) of the channels/membraneby means of protein and

cells adsorption from biological ﬂuids. Biofouling can also result in the clogging

of the nanochannels/nanopores, which potentially could render the device inef-

fective. The adhesion of proteins and cells to an implanted device can also cause

detrimental results such as inﬂammation and excessive ﬁbrosis. Deposition of the

self-assembled monolayers of selected organic molecules on the channels implants,

which makes them hydrophobic, presents an innovative solution to combat the ad-

verse eﬀects of the biological ﬂuids [122–125].

Figure 22.12b shows a conceptual model of an intravascular drug-delivery de-

vice – nanoparticles used to search for and destroy disease cells [78]. With lateral

dimensions of 1µm or less, the particles are smaller than any blood cells. These

particles can be injected into the blood stream and travel freely through the circu-

latory system. In order to direct these drug-delivery nanoparticles to cancer sites,

their external surfaces are chemically modiﬁed to carry molecules that have lock-

and-key binding speciﬁcity with molecules that support a growing cancer mass. As

soon as the particles dock onto the cells, a compound is released that forms a pore

on the membrane of the cells, which leads to cell death and ultimately to that of

the cancer mass that was being nourished by the blood vessel. Adhesion between

nanoparticles and disease cells is required. Furthermore, the particles should travel

close to the endothelium lining of vascular arteries. Decuzzi et al. [126] recently

1222 Bharat Bhushan

Immune molecules retained outside;

glucose, nutrients and insulin pass freely

Size-selective

nanoporous

membrane

Encapsulated

islet cells

One half

of biocapsule

1. Binding

(0 – 8 h

after

injection)

2. Plug

rupture,

drug

release

(12 – 48h)

3. Pore

formation –

cell lysis

and death

(12 – 48h)

Fig. 22.12. Schematics of (a) implantable, immuno-isolation submicroscopic biocapsules

(drug-delivery device) [77], and (b) intravascular nanoparticles to search for and destroy dis-

eased blood cells [78]

analyzed the margination of a particle circulating in the blood stream and calculated

the speed and time for margination (motion of the particles towards the walls of the

vessel) as a function of the density and diameter of particle, based on various forces

present between thecirculating particle and theendothelium lining. Humancapillar-

ies can have radii as small as 4–5µm. They reported that the particles used for drug

delivery should have a radius smaller than a critical value (of the order 100nm).

In summary, adhesion, stiction/friction and wear clearly limit the lifetime and

compromise the performance and reliability of MEMS/NEMS and BioMEMS/

BioNEMS. Figure 22.13a summarizes tribological problems encountered in some

of the MEMS, MOEMS, RF-MEMS, and BioMEMS devices discussed. In add-

ition to in-use stiction, stiction issues are also present in some processes used for

fabrication of MEMS/NEMS. For example, the last step in surface micromachin-

ing involves the removal of sacriﬁcial layer(s), known as the release process, since

the microstructures are released from the surrounding sacriﬁcial layer(s). The re-

22 Characterization of MEMS/NEMS and BioMEMS/BioNEMS 1223

Suspended

mass

Stationary

plates

Hinge Yoke

Tip

Landing site

Impact/wear

Stiction

(meniscus

effects)

Stiction

and wear

Field

insulator

(SiO

)

Source

metal

(Al)

Receptor

biological

molecules

Drain

metal

(Al)

–

Substrate

source Gate

insulator

(SiO

)

drain

Micro-

switch

Side

view

Highly resistive substrate

Off-state

Stiction and wear

Top view

GND

RF out

RF in

Flexible

metal bridge

signal

line

Dielectric

GND

Unreleased beam

Rinse liquid

Released beam before drying

Released beam collapsed to substrate

due to meniscus forces during drying

Silicon

Polysilicon

Sacrificial

material

Fig. 22.13. (a) Summary of tribological issues in MEMS, MOEMS and RF-MEMS device

operation, and (b) in microfabrication via surface micromachining

lease is accomplished by an aqueous chemical etch, rinsing and drying processes.

Due to meniscus eﬀects as a result of wet processes, the suspended structures can

sometimes collapse and permanently adhere to the underlying substrate, as shown

in Fig. 22.13b [127]. Adhesion is caused by water molecules adsorbed on the adher-

ing surfaces and/or because of the formation of adhesive bonds by silica residues

that remain on the surfaces after the water has evaporated. This so-called release

stiction is overcome by using dry release methods (e.g., CO

critical-point drying

or sublimation drying [128]).

Tribological Needs

MEMS/NEMS need to be designed to perform expected functions typically in

the millisecond to picosecond range. Expected life of the devices for high-speed

contacts can vary from a few hundred thousand to many billions of cycles, e.g.,

over a hundred billion cycles for DMDs, which puts serious requirements on

materials [13, 91, 129–132]. Adhesion between a biological molecular layer and

the substrate, referred to as bioadhesion, reduction of friction and wear of bio-