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

Подождите немного. Документ загружается.

964 P. Lin e t al.

Then a cross-linker was used to immobilize the bioactive proteins on a pre-adsorbed

BSA layer, with the chemistry as shown in Example 14. The employed strategy

(Fig. 13.27b) results in less blocking of protein A on the surface, and higher reac-

tivity with the antibody (Y). This new method signiﬁcantly improves the ELISA

detection efﬁciency using the same PDMS microﬂuidic system.

13.5.4.2 Case Study 2: Effective Enhancement of Fluorescence Detection

Efﬁciency Using Alternative Blocking Process in Protein Microarray

Assays

Aminosilanized silica substrates are used to prepare protein microarrays because of

their strong binding afﬁnity for proteins. In a typical microarray preparation proce-

dure, after the probe proteins are spotted on an aminosilanized silica substrate, the

background surface of the spotted silica substrate needs to be blocked to prevent

nonspeciﬁc adsorption of target proteins. Conventionally, protein-based blocking

solutions, such as albumin solutions (e.g. BSA solution), non-fat milk solutions,

casein solutions or serum solutions are used. Although aminosilanized silica sur-

faces can be passivated by protein-based blocking solutions, a wet blocking process

often causes smearing of individual protein spots and cross-contamination among

probe spots (Fig. 13.28, steps c1 to f1). To eliminate the smearing and cross-

contamination issues caused by conventional blocking processes using protein-

based blocking solutions, Hsieh et al. employed a vapor-phase blocking process

[93]. This method employs a highly ﬂuorinated organosilane, FOTS (see No. 10

in Fig. 13.7), as the blocking agent to passivate the background surfaces of pro-

tein arrays prepared on aminosilanized glass slides (Fig. 13.28,stepc2tof2).As

shown in Fig. 13.28 h and i, upon hydrolysis, FOTS can be made reactive, and

bonds to the APTS ( also called APTES, in Fig. 13.7) surface. During the vapor-

phase FOTS blocking treatment, the exterior protein molecules extensively exposed

to the FOTS vapor can serve as sacriﬁcial materials to protect the interior pro-

tein molecules. Moreover, these loosely bound exterior protein molecules can be

subsequently washed off from the spotted areas without causing undesired contami-

nation on the background surface that has been passivated by the vapor-phase FOTS

treatment.

Compared to several conventional blocking processes that use protein-based

solutions, it was demonstrated that the vapor-phase blocking process could signif-

icantly reduce the smearing of protein spots and effectively enhance ﬂuorescence

detection efﬁciency by up to ~64 times [93].

13.5.4.3 Case Study 3: Control of Speciﬁc Reaction Kinetics Involving

Bifunctional Cross-Linkers

AFM probes can be made into sensitive, chemically selective biosensors by attach-

ing ligand molecules to the tips of the probes [94]. Many model systems that

utilize ligands on AFM probes and complementary receptors on substrates have

13 Surface Treatment and Planarization 965

(a) cleaned slide

(b) array-spotting on an APTS-modified slide

(c1) wet blocking

(HS, MILK, or BSA)

(d1) completion of wet blocking

(e1) PBST washing

(f1) completion of conjugation

& final washing

(c2) vp-FOTS blocking

(d2) completion of

blocking

vp-FOTS

(e2) PBST washing

Conventional

Wet Blocking

Vapor-Phase

Blocking

Slide

APTS

Protein-Based

Blocking Agent

FOTS

2nd Ab (Cy5)

1st Ab (Cy3)

(f2) completion of conjugation

& final washing

APTS

treatment

(g)

(h)

(i)

Pressure meter

Temperature

controller

Valve

To t h e

mechanical

pump

Slides

(upside down)

Plastic box

FOTS vapor

FOTS

liquid

Fig. 13.28 Background blocking of protein arrays by a conventional protein-based wet blocking

process vs. a vapor-phase blocking process using a highly ﬂuorinated organosilane, 1H,1H,2H,2H-

Chemical Society

been developed to study the dynamics of molecular interactions. The tips of AFM

probes are made of polysilicon or silicon nitride. These surfaces are typically cov-

ered with a natural silicon oxide layer, and contain many r eactive -SiOH groups.

As described in Section 13.5.2.3, to start a multi-step modiﬁcation on such type

of surfaces, organosilanes with amine (such as APTES) or other functional groups

966 P. Lin e t al.

are r equired. After surface silanization of the AFM tip, a bifunctional cross-linker is

needed to bond the ligand (or antibody) to the AFM tip. The cross-linkers must have

a very ﬂexible chain, and PEG (polyethylene glycol) is typically used to provide

sufﬁcient chain ﬂexibility. As several surface modiﬁcation schemes with different

combinations of organosilanes and PEG-based bifunctional cross-linkers are pos-

sible, the reaction proﬁles and kinetics can differ from one scheme to another. To

address how to handle these issues, ﬁve different schemes for AFM tip coating are

provided as examples, and a discussion follows.

First, the linkage to the antibody is through the disulﬁde exchange reaction.

Ebner et al. has been working on the synthesis and application of PEG-based

heterobifunctional cross-linkers for the immobilization of antibodies onto AFM

tips [95]. In their earlier study, PDP-PEG-NHS was used as the cross-linker to

react with aminosilanized AFM tips [ PDP is 3-(2-pyridyl)-dithiopropionyl, which

contains a disulﬁde bond, i.e. PDP can undergo the exchange reaction described

in reaction (13.10) with a thiol group; NHS is N-hydroxysuccinimide group, see

reaction (13.6)]. Because there are no free thiol groups on antibodies, the thiol

groups need to be introduced to the antibody ﬁrst. The thiolated antibody needs

to be puriﬁed before usage because the residual reagents and byproducts from

the thiolation process can also react with the PDP group on the cross-linkers.

To avoid the extra purifying process, another strategy is to treat the aminosi-

lanized AFM tips with NHS-PEG-NHS to immobilize the antibodies. Compared

to PDP-PEG-NHS, NHS-PEG-NHS is preferred because there is no need to add

thiol groups to the antibody. However, the NHS-PEG-NHS has two amine-reactive

NHS groups, and it can lose them both if they each react with the surface-bound

amino groups. To prevent this, Ratto et al. developed another strategy, in which

a mixture of APTES (see Fig. 13.7) and MTES [CH

-Si-(OC

)

] was used

in the ratio of 1:250 to prepare an aminosilanized surface diluted by the methyl

groups of MTES [96]. This would reduce the reaction kinetics of the amino groups

on APTES towards the NHS group of the cross-linker, NHS-PEG-NHS. Using

this strategy, one NHS group in the cross-linker could r emain to react with the

antibody.

Alternatively, the AFM tip can be treated with 3-mercaptopropyl-

trimethoxysilane to produce reactive thiol groups, and a maleimide-PEG-NHS

cross-linker can be used to react with the AFM tip via a reaction between

the maleimide and thiol groups. Although a thiol functional group reacts with

maleimide relatively quickly, it still can react with NHS to some extent, though

less favorably. Thus this technique cannot completely prevent the cross-linker from

forming a loop on the AFM tip. Another strategy that was recently developed by

Ebner employed NHS-PEG-CHO as a cross-linker to react with aminosilanized

AFM tips [95]. Although both NHS and CHO can react with the amino group, the

kinetics of amine with NHS overwhelms the kinetics of amine with CHO; as such,

NHS-PEG-CHO preferentially reacts with the surface-bound amino group via the

NHS group. Following the treatment of the AFM tips with NHS-PEG-CHO, the

free CHO group can be exploited to react with the amino group of the antibody for

the preparation of biofunctional AFM tips.

13 Surface Treatment and Planarization 967

13.5.4.4 Case Study 4: Surface Modiﬁcation Using Elaborately Derivatized

Functional Groups

To exploit alternative chemical reactions for surface modiﬁcation in BioMEMS,

organic synthesis can be performed to prepare chemical species derivatized with

special functional groups for desired chemical reactions to occur. For example,

Houseman et al. exploited the Diels-Alder reaction (13.15) to perform mediated con-

jugation between the cyclopentadiene group derivatized on a peptide analogue and

the benzoquinone group immobilized on a gold surface (Fig. 13.29)[97]. Although

substantial synthetic work is required to synthesize a cyclopentadiene-derivatized

peptide analogue and a benzoquinone-derivatized thiol, this study shows that chem-

ical conjugation via the Diels-Alder reaction is highly efﬁcient and selective for the

(13.15)

Peptide Analogue

AuAu

(CH

)

(OC

)

R’

Fig. 13.29 Mediated conjugation of a peptide analogue to a Au-based substrate via Diels-Alder

reaction, as demonstrated by Houseman et al. [97]

968 P. Lin e t al.

preparation of peptide chips. As mentioned in Case study 3, the cross-linking reac-

tions are sometimes not very speciﬁc and can cause different side reactions. If a

very speciﬁc and efﬁcient reaction is required, preparation of certain compounds or

reagents with desired functional groups by elaborate synthesis may be necessary.

13.5.4.5 Case Study 5: Surface Patterning by Microcontact Printing

Microcontact printing is a surface patterning technique that can be conveniently

employed for many applications [98–105]. The basic working principle of micro-

contact printing is illustrated in Fig. 13.30. To perform surface modiﬁcation by

microcontact printing, a patterned elastomer stamp (usually a PDMS stamp) must

be fabricated. Then an “ink” solution containing a chemical (or a mixture of chem-

icals) that can react with the target substrate surface (e.g. organosilanes for silica

surfaces, thiols for noble metal surfaces) is applied to the patterned stamp surface.

Depending on the nature of the “ink” solution, (e.g. viscosity, reactivity, stability

etc.), some precautions must be considered during the preparation, handling and/or

Target substrate

PDMS stamp

PDMS stamp “treated”

with the “ink” solution

Microcontact printing

Target substrate surface

patterned by microcontact printing

Fig. 13.30 Schematic illustration of microcontact printing

13 Surface Treatment and Planarization 969

application of the “ink” solution (e.g. concentration of the “ink” solution, extent to

which the solvent of the “ink” solution should evaporate after the “ink” has been

applied to the patterned stamp surface, etc). Ideally, only a thin physically adsorbed

layer of the s urface-reactive chemical on the patterned stamp surface is desired after

the elastomer stamp is “treated” with the “ink” solution. Then the stamp is brought

into contact with the target substrate surface. As shown in Fig. 13.30, the surface-

reactive chemical reacts with the target substrate surface in the areas that contact the

patterned stamp surface. After the stamp is removed from the target substrate sur-

face, a geometrically and chemically deﬁned pattern can be obtained on the target

substrate surface for further use.

Because of its ease and convenience, microcontact printing is widely used for the

preparation of patterned bioactive surfaces [106]. For example, Chen et al. prepared

gold surfaces with patterned bioactive proteins (ﬁbronectin, vitronection, collagen

etc.) to study the growth and apoptosis of bovine capillary endothelial cells [107].

The gold surface was ﬁrst patterned with hexadecanethiol by microcontact printing.

The patterned gold surface was then treated with a tri(ethylene glycol)-terminated

hexadecanethiol solution to passivate the rest of the unpatterned gold surface.

When the above patterned/passivated gold surface was immersed in a protein solu-

tion, preferential adsorption of the protein occurred on the area with hydrophobic

hexadecyl SAM coating. Finally, cell cultures grown on the gold surfaces with var-

ious patterned bioactive proteins were investigated and compared. From this study,

Chen et al. found that it was the extent of cell spreading, not the area of adhesive

contact, that controlled the life and death of bovine capillary endothelial cells.

13.6 Surface Coating for Optical Applications

MEMS devices have been applied to optical sensing applications and light-

wave communications including optical switches, waveguides, tunable ﬁlters, and

reconﬁgurable wavelength add and drop multiplexers. Components in these MEMS

devices interact, receive, and transport light. As light interacts with MEMS devices,

three optical functions are relevant to performance: reﬂection, transmission, and

absorption. The surfaces on the MEMS devices have strong effects on these optical

functions.

Many optical materials can be processed in MEMS fabrication, including non-

crystalline materials (such as glasses and metals), crystalline materials (such as sil-

icon, quartz, and germanium wafers), and polymers (plastics, photoresist, and elas-

tomers). The fabrication techniques used to obtain the desired optical properties –

such as physical vapor deposition, chemical vapor deposition, sol-gel, spin coatings,

self assembly, phase separation, and imprinting – are discussed in the other chapters

of this book.

This section begins with theory and design concepts related to the optical

phenomena of surface ﬁlms, including refraction, dispersion, absorption, internal/

970 P. Lin e t al.

external reﬂection, polarization by reﬂection, and antireﬂection. The material prop-

erties and fabrication processes of these coatings and structures are discussed in

the second part of this section, along with their corresponding advantages and

challenges. Several useful case studies related to antireﬂection, absorption, and

refraction of designated light waves will be discussed as examples of the design and

fabrication of optical MEMS devices. Lastly, surface roughness and control will be

summarized.

13.6.1 Fundamentals of Optical Phenomena on Surface Coatings

The performance of optical MEMS devices usually relies on functional optical prop-

erties such as antireﬂection, high absorption, refraction, and polarization of device

surfaces. The desired properties are usually achieved by applying special coatings or

material structures to the surface. Most optical property transformations through sur-

face coatings and structures are based on the scattering or interference of light waves

at the interface [108, 109]. The interface is very sensitive to ﬁlm thickness relative

to the wavelength of the incident light, which can range from tens of nanometers

to several microns. As a result, surface planarization properties [109] have strong

effects on the phenomena occurring at the incident interfaces. Since the design and

selection of coatings and structures are typically the ﬁrst things to consider, we will

start with optical theory and design concepts in this section.

13.6.1.1 Index Variation of Materials Versus Wavelength [108]

Dielectric Materials

When light transmits in the medium of bulk matter, for example in a homogeneous,

isotropic dielectric material, the phase speed of the light in the medium can be

expressed by

v =

√

εμ

(13.16)

where ε is the permittivity and μ is the permeability of the medium. As a result, the

ratio of the speed of an electromagnetic wave in a vacuum to its speed in matter is

deﬁned as the absolute index of refraction, n:

n ≡



εμ

(13.17)

where c is the light speed, ε

is the permittivity, and μ

is the permeability in free

space.

The index n can also be expressed in t erms of the relative permittivity and relative

permeability of the medium:

13 Surface Treatment and Planarization 971

n =



≈



(13.18)

where K

= ε/ε

is the dielectric constant or relative permittivity, and K

= μ/μ

is the relative permeability. Since K

does not usually deviate much in the visible

regions of the spectrum, it can be neglected.

This relation is only valid for a few simple gases such as air, helium, and hydro-

gen. The approximate refractive indices of some commonly used materials are listed

in Table 13.3, and in other sections (Tables 13.4, 13.5, 13.6, and 13.7). However, the

indices of dielectric materials are often frequency-dependent. This is called “dis-

persion”, and can be expressed by the dispersion equation f or a dielectric material:

(ω) = 1 +



− ω



(13.19)

where q

is the electron charge, m

is the electron mass, N is the number of con-

tributing electrons per unit volume, and ω

is the resonance frequency. The index n

increases as ω approaches ω

from below, but drops to less than 1 when ω is larger

than ω

. The index of an optical material varies with wavelength (index dispersion),

as indicated in Fig. 13.31. The region with a large index jump corresponds to the

absorption bands of the material. After the jump, the index starts to decrease as

wavelength increases, until the wavelength reaches the next absorption band. Jumps

may be located in different regions of the spectrum, for example one in the infrared

(ω

) region and the other in the ultraviolet (ω

) region. The targeted visible spectra

lie between these regions.

Metals

When light interacts with bulk metal, the conduction electrons undergo oscillation

and collisions with the lattice, and thereby generate absorption. The electrical ﬁeld

Table 13.3 Approximate indices of refraction of various substances [108]

Air 1.00029 Light ﬂint glass 1.58

Ice 1.31 Polystyrene 1.59

Water 1.333 Carbon disulﬁde (CS

) 1.628

Ethyl alcohol (C

OH) 1.36 Dense ﬂint glass 1.66

Fused silica (SiO

) 1.4584 Lanthanum ﬂint glass 1.80

Carbon tetrachloride (CCl

)1.46 Zircon(ZrO

SiO

) 1.923

Turpentine 1.472 Fabulite (SrTiO

) 2.409

Benzene (C

) 1.501 Diamond (C) 2.417

Plexiglass 1.51 Rutile (TiO

) 2.907

Crown glass 1.52 Gallium phosphide 3.50

Sodium chloride (NaCl) 1.544 Silicon nitride (Si

)2.04

972 P. Lin e t al.

Table 13.4 Properties of antireﬂection coating materials [109, 110]

Material

Refractive index at

λ = 546 nm

Evaporation

temperature (

◦

C) Comments

Aluminum oxide (Al

) 1.66 2045 E-beam evaporation with substrate at 200

◦

C, hard and chemical

resistant ﬁlm

Cadmium sulphide (CdS) 2.42 800 E-beam evaporation in high vacuum, requires care

Cadmium telluride (CdTe) 2.69 E-beam evaporation

Ceric oxide (CeO

) 1.95 1600 Thermal evaporation, soft ﬁlm

Cerous ﬂuoride (CeF

) 1.63 1342 E-beam evaporation with substrate at 200

◦

Hafnium oxide (HfO

)2.05

Cryolite (Sodium aluminum

ﬂuoride, Na

AlF

)

1.22 1000 Thermal evaporation, soft ﬁlm

Lanthanum ﬂuoride (LaF

) 1.55 1490 Thermal evaporation

Lead ﬂuoride (PbF

) 1.75 855 Thermal evaporation, soft ﬁlm

Lithium ﬂuoride (LiF) 1.303 870 Thermal evaporation in high vacuum, soft transparent ﬁlm

Magnesium ﬂuoride (MgF

) 1.38 1266 Thermal evaporation in high vacuum

Magnesium oxide (MgO) 1.68 2800 E-beam evaporation, hard ﬁlm

Neodymium ﬂuoride (NdF

) 1.55 1410 Thermal evaporation, very hard and t ransparent ﬁlm

Neodymium oxide (Nd

) 1.79 1900 Thermal evaporation, very hard and transparent ﬁlm

Silicon monoxide (SiO) 1.49–1.80 1250 Depends on the rate of deposition, thermal evaporation at O

environment, hard ﬁlm

Silicon-silica mix (Si-SiO

) 1.49 Thermal evaporation, hard and transparent ﬁlm

Sodium ﬂuoride (NaF) 1.29 988 Thermal evaporation, soft and transparent ﬁlm, easily to be attacked

by moisture

13 Surface Treatment and Planarization 973

Table 13.4 (continued)

Material

Refractive index at

λ = 546 nm

Evaporation

temperature (

◦

C) Comments

Thorium oxide (ThO

) 1.86 3050 E-beam evaporation with substrate at 200

◦

Thorium oxyﬂuoride (ThOF

) 1.50 1100 Thermal evaporation, good adhesion to substrate

Titanium monoxide (TiO) 2.08–2.32 1750 Depends on the rate of deposition, slow thermal evaporation at O

environment, hard and transparent ﬁlm

Yttrium oxide (Y

) 1.79 2410 E-beam evaporation, very hard ﬁlm

Zinc selenide (ZnSe) 2.57 900 Thermal evaporation, soft and transparent ﬁlm

Zinc sulﬁde (ZnS) 2.35 1000 Thermal evaporation, soft and transparent ﬁlm

Zirconium oxide (ZrO

) 2.05 2714 E-beam evaporation with substrate at 200

◦

Germanium (Ge) 4.0

E-beam evaporation

Silicon 3.50

E-beam evaporation

not transparent in visible spectrum