Nano Today (2011) 6, 131—154
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/nanotoday
REVIEW
Silicon nanowire field-effect transistor-based
biosensors for biomedical diagnosis and cellular
recording investigation
Kuan-I Chen a,b,1, Bor-Ran Li a,b,1, Yit-Tsong Chen a,b,∗
a
b
Department of Chemistry, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 106, Taiwan
Institute of Atomic and Molecular Sciences, Academia Sinica, P.O. Box 23-166, Taipei 106, Taiwan
Received 31 October 2010; received in revised form 18 December 2010; accepted 7 February 2011
Available online 8 March 2011
KEYWORDS
Silicon nanowire;
Field-effect
transistor;
Protein—protein
interaction;
DNA hybridization;
Peptide—small
molecule interaction;
Biomarker detection;
Three-dimensional
localized bioprobe
Summary Silicon nanowire field-effect transistors (SiNW-FETs) have recently drawn tremendous attention as a promising tool in biosensor design because of their ultrasensitivity,
selectivity, and label-free and real-time detection capabilities. Here, we review the recently
published literature that describes the device fabrication and biomedical applications of SiNWFET sensors. For practical uses, SiNW-FETs can be delicately designed to be a reusable device via
a reversible surface functionalization method. In the fields of biological research, SiNW-FETs are
employed in the detections of proteins, DNA sequences, small molecules, cancer biomarkers,
and viruses. The methods by which the SiNW-FET devices were integrated with these representative examples and advanced to virus infection diagnosis or early cancer detection will
be discussed. In addition, the utilization of SiNW-FETs in recording the physiological responses
from cells or tissues will also be reviewed. Finally, the novel design of a three dimensional (3D)
nano-FET probe with kinked SiNWs for recording intracellular signals will be highlighted in this
review.
© 2011 Elsevier Ltd. All rights reserved.
Abbreviations: AFM, atomic force microscopy; ATP, adenosine triphosphate; CA, carbohydrate antigen; CaM, calmodulin; CEA, carcinoembryonic antigen; CgA, chromogranin A; CNT, carbon nanotube; CVD, chemical vapor deposition; DNA, deoxyribonucleic acid; EDTA,
ethylenediaminetetraacetic acid; FET, field-effect transistor; GSH, glutathione; GST, glutathione S-transferase; His-tag, histidine-tag;
miRNA, microRNA; MPC, microfluidic purification chip; NTA, nitrilotriacetic acid; PBS, phosphate buffered saline; PDMS, polydimethylsiloxane; PNA, peptide nucleic acid; PS, phosphate solution; PSA, prostate specific antigen; RNA, ribonucleic acid; RT-PCR, reverse
transcription-polymerase chain reaction; SiNW, silicon nanowire; TnI, troponin I; VGCC, voltage-gated Ca2+ channel; VLS, vapor—liquid—solid.
∗ Corresponding author at: Department of Chemistry, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 106, Taiwan.
Tel.: +886 2 2366 8238; fax: +886 2 2362 0200.
E-mail address: ytcchem@ntu.edu.tw (Y.-T. Chen).
1 These authors contributed equally to this work.
1748-0132/$ — see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.nantod.2011.02.001
132
Introduction
Quantification and analysis of biological processes are of
utmost importance for biomedical applications and cellular programming investigation. However, it is challenging
to convert the biological information into an electronic
signal due to the difficulties of connecting an electronic
device into a biological environment. In recent years, there
has been dramatic development of electrochemical biosensors because of their applications in toxicity testing [1],
chemical analysis [2], medical diagnosis [3], food industry [4], environmental monitoring, and many other areas.
An electrochemical biosensor, as defined by IUPAC, is a
self-contained integrated device that allows for specific analytical detection by using a biological recognition element
(a biochemical receptor) in direct spatial contact with a
transduction element (Fig. 1(a)) [5,6]. Different from a bioanalytical system (e.g., immunoprecipitation usually used
for protein analysis) that requires a reagent addition to pro-
K.-I. Chen et al.
ceed the analysis, an electrochemical biosensor provides
an attractive platform to analyze the contents of biological samples because of the direct conversion of biological
events to electronic signals (that can be detected directly),
thus allowing more rapid and convenient sensing detection.
Investigations of the materials and methods to construct an electrochemical biosensor have been underway
for decades. Over the past 20 years, nanomaterials, such
as quantum dots, nanoparticles, nanowires, nanotubes,
nanogaps, and nanoscale films [7—13], have received enormous attention due to their suitable properties for designing
novel nanoscale biosensors. For example, the dimension of
nanomaterials of ∼1—100 nm provides a perfect feature to
study most biological entities, such as nucleic acids, proteins, viruses, and cells (as illustrated in Fig. 1(b)) [14].
In addition, the high surface-to-volume ratios for nanomaterials allow a huge proportion of the constituent atoms
in the material to be located at or close to the surface. This characteristic makes the surface atoms play
Figure 1 (a) The construction of typical biosensors with elements and selected components. The procedures are described as
follows: (i) receptors specifically bind the analyte; (ii) an interface architecture where a specific biological event takes place and
gives rise to a signal recorded by (iii) the transducer element; (iv) computer software to convert the signal into a meaningful physical
parameter; finally, the resulting quantity is displayed through (v) an interface to the human operator. (b) The sizes of nanomaterials
(NW and NT) in comparison to some biological entities, such as bacteria, viruses, proteins, and DNA.
Reprinted from [6,14].
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
133
an extremely important role in determining the physical,
chemical, or even electronic properties of nanomaterials.
Moreover, some particular nanomaterials with surfaces that
are easily chemically modified have made them significant
candidates for nanoscale sensing applications.
To date, a variety of nanoscale sensing techniques have
been used for biological research and applications. In particular, when monitoring living systems, requiring rapid and
precise detection, the demands of sensor architectures
become challenging. Several essential factors, such as ultrasensitivity, specificity, high-speed sample delivery, and low
cost must be considered when designing and fabricating
nanoscale biosensors. Some sensing devices selecting quantum dots as their sensing elements possess the merits of high
sensitivity, selectivity, and short response time. However,
this kind of sensing technique generally requires integration with optical instruments to translate the successful
binding phenomena into a readable signal [7], making the
sensing measurements costly. On the other hand, devices
like field-effect transistors (FETs) can be suitable candidates
for designated sensors, owing to their ability to directly
translate the interaction with target molecules taking place
on the FET surface into a readable signal [15]. In recent
years, one-dimensional semiconducting nanomaterials, such
as silicon nanowires and carbon nanotubes, configured with
FETs (referred to as SiNW-FET [16,17] and CNT-FET [18—20],
respectively) have attracted great attention because they
are an ideal biosensor with high selectivity and sensitivity,
real-time response, and label-free detection capabilities. In
this review article, we will mainly focus on the device fabrication of SiNW-FETs and their applications in biomedical
diagnosis and cellular research.
Field-effect transistor-based biosensors
From the electrochemical point of view, SiNW-FET-based
biosensors are a three-electrode system, including source,
drain, and gate electrodes. The function of the source
and drain electrodes is to bridge the semiconductor channel made of SiNWs and the gate electrode is responsible
for modulating the channel conductance. In a representative NW-FET example illustrated in Fig. 2(a), the biological
receptors were anchored to the surface of the semiconductor channel by chemical modification to recognize the target
analytes through their high specificity and strong binding
affinity in the buffer environment. The target—receptor
Figure 2 (a) The illustration of a nanoscale FET biosensor
with a cross-sectional view. The semiconductor channel (NW or
NT) is placed between the source and drain electrodes with a
gate electrode on the bottom to modulate the conductivity of
the semiconductor channel. Target molecules can be recognized
by the receptor modified on the channel surface through strong
binding affinity. (b) When positively charged target molecules
bind the receptor modified on a p-type NW, positive carriers
(holes) are depleted in the NW, resulting in a decrease in conductance. On the contrary, negatively charged target molecules
captured by the receptor would make an accumulation of hole
carriers, causing an increase in conductance. (c) Schematic
representation of a CNT-FET device including the surface modification and molecular recognition procedures: (1) modification
of linkers onto the single-walled CNT through ␲—␲ interaction; (2) immobilization of antibody; (3) detection of antigen by
antibody. (d) CgA was released from neurons stimulated by glutamate and was detected by CgA-Ab/CNT-FET. A coverslip with
grown neurons was positioned on the CgA-Ab/CNT-FET device
with neurons facing the FET circuits. Immediately after the glutamate (50 ␮M) stimulation, a dramatic increase in current was
detected due to the binding of the released CgA to CgA-Ab/CNTFET.
Reprinted from [14,38].
134
interaction then varied the surface potential of the semiconductor channel and modulated the channel conductance,
and the signal was eventually collected by a detection system.
A diversity of FET-based biosensors has been employed
for biological applications. Here, we tried to classify these
biosensors into enzyme-modified FETs, cell-based FETs, and
immunologically functionalized FETs. Enzyme-modified FETs
comprise a redox active enzyme integrated with an electronic circuitry to give a real-time quantitative analysis
of the enzyme substrate [21], e.g., sensing glucose from
a catalytic reaction in the presence of glucose oxidase.
Cell-based FETs were exploited to detect the released biochemical agents or real-time cellular responses from living
cells, such as action potentials from neuron cells [22] or
electrical recordings from chicken hearts [23].
In general, immunologically functionalized FETs are the
most frequently used biosensors. For example, an antibodymodified FET sensor can be used to detect the corresponding
antigen. Depending on the charge carriers in the semiconductive channel (holes for a p-type channel and electrons for
an n-type channel), the direction of the conductance change
represents the sign of the charges carried by the target antigen, and the magnitude of the conductance change reflects
the antigen—antibody interaction. In an example of a p-type
NW-FET illustrated in Fig. 2(b), when the positively charged
analytes bind the receptor-anchored NW-FET, a depletion of
charge carriers occurs in the conductance channel, causing
a decrease in the device conductivity. On the contrary, an
increase in the device conductivity would result from the
accumulation of charge carriers in the conductance channel
while negatively charged molecules, such as DNA or RNA,
bind the p-type NW-FET.
In recent years, many types of semiconducting materials,
such as carbon materials (e.g., CNT and graphene) [24—26]
and metal-oxide nanowires (e.g., In2 O3 -NW and ZnO-NW)
[27,28], have been selected as promising candidates for
the development of FET-biosensors. For instance, graphenebased FETs were constructed for electrically detecting
pH values, bovine serum albumin adsorption [25], and
cellular recording [26]. In2 O3 -NW [27] and ZnO-NW [28]
configured with FET-biosensors were also used to monitor
protein—protein interactions. Among them, CNTs, singledwalled CNTs in particular, were at the forefront of these
explorations. Several recent articles about CNT-FETs, as
represented in Fig. 2(c), have reviewed their biological
applications [29—31], such as antigen—antibody interactions
[27,32—34], DNA hybridization [35,36], and enzymatic glucose detection [37]. As depicted in Fig. 2(d), a CNT-FET
was specifically applied to the real-time detection of a
cancer marker for neuroendocrine tumors, namely chromogranin A (CgA), released from embryonic cortical neurons
[38,39]. The CNT-FET device modified with the antibody of
CgA (referred to as CgA-Ab/CNT-FET) was employed to monitor the in situ release of CgA from living neurons in response
to glutamate stimulation.
Despite these advances of CNT-FETs in biosensory applications, several shortcomings were encountered in the
fabrication and applications of CNT-FETs. First, in the fabrication of CNT-FETs, the mixtures of semiconducting and
metallic CNTs still hamper future developments in nanoelectronics. Secondly, the determining factors for the sensing
K.-I. Chen et al.
mechanisms of a CNT-FET are somewhat complex and were
reported to involve field-effects [18,40], electron transfer
[18], Schottky barriers [41,42], etc. In contrast, the sensing mechanism of a SiNW-FET sensor is straightforward and
simply dominated by the field-effect [16,43] due to the
interaction between the target analyte and the receptor
modified on the surface of the SiNW-FET.
Silicon nanowire field-effect transistors
Taking advantage of the well-developed silicon industry,
SiNW-FETs can benefit from existing and mature silicon
industry processing techniques and fabrications. In the synthetic reactions that prepare SiNWs, different sizes [44,45],
shapes [46], and dopants [47] of SiNWs could be precisely
tailored. Because SiNWs could be well-controlled during the
wire growth, the performance exhibits high reproducibility.
Therefore, the n-/p-type semiconducting property, doping
density, and charge mobility in a SiNW-FET can be designed
in advance. In the following sections, we selectively discuss the nanowire fabrication, assembly techniques, device
array design, and electrical measurement setup used in the
performance of SiNW-FETs.
Fabrication of SiNW-FETs
There are two major fabrication techniques in preparing
SiNW-FETs: ‘‘top-down’’ and ‘‘bottom-up’’. The ‘‘topdown’’ method is carried out through lithographic processes
combined with an electron-beam technique that defines
SiNW-FETs by physically etching a single-crystalline silicon
wafer [48]. On the other hand, the ‘‘bottom-up’’ processes
start with the growth of SiNWs, normally in a chemical vapor
deposition (CVD) reaction, followed by SiNW assembly and
electrode fabrication via the photolithographic or electronbeam lithographic procedures [43].
‘‘Top-down’’ SiNW-FETs
The ‘‘top-down’’ method for the SiNW-FET fabrication is
based on lithographic processing on a silicon-on-insulator
(SOI) wafer. As illustrated in Fig. 3(a, i), the structure
of an SOI wafer contains three layers: substrate Si wafer,
buried silicon dioxide (about 200—400 nm thick), and top Si
layer (about 50—100 nm thick). Through the standard procedures of photolithography, reactive ion etching (RIE), ion
implantation, electron-beam lithography, and thermal evaporation, SiNWs and the connecting electrodes can be defined
to form SiNW-FET devices, in which the width of SiNWs could
reach the scale of ∼100 nm.
A typical ‘‘top-down’’ process to fabricate SiNW-FETs is
illustrated schematically in Fig. 3(a) [49—56]. In Step 1, the
Si layer is doped with low-density boron or phosphorous
of ∼1015 /cm3 (about 10—20 cm). Therefore, the n-/ptype semiconducting property and doping ratio of SiNWs are
determined (Fig. 3(a, i)). Step 2 is to define the source and
drain leads with heavy doping (1019 /cm3 ), of which the patterns are drawn with a photomask design (Fig. 3(a, ii)). In
Step 3, the micrometer-sized source and drain electrodes
are finished by RIE etching (Fig. 3(a, iii)). The following
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
135
Figure 3 (a) Schematic illustration of a typical ‘‘top-down’’ process to fabricate SiNW-FETs. (i) In Step 1, the silicon layer is doped
with low-density boron or phosphorous of ∼1015 /cm3 . (ii) In Step 2, specific regions defined with a photomask pattern receive heavy
doping (1019 /cm3 ). (iii) In Step 3, the micrometer-sized source and drain electrodes are finished by RIE etching. (iii) The following
Step 4 is to fabricate the nanometer-sized SiNWs with an electric-resist pattern and RIE etching. (b) An illustration of a ‘‘bottomup’’ method to fabricate SiNW-FETs. (i) The growth of SiNWs in CVD reaction via the VLS mechanism. (ii) Deposition/alignment of
SiNWs on a silicon substrate. (iii) A photomask pattern to define source/drain electrodes. (iv) Thermal evaporation to deposit the
source/drain contacts. (v) Lift-off the remaining photoresist with Remover PG.
Step 4 is to fabricate the nanometer-sized SiNWs with an
electric-resist pattern and RIE etching (or tetramethylammonium hydroxide etching [49]) (Fig. 3(a, iv)). Subsequently,
a thermal evaporation is used to make the contact leads
and back-gate, and finally an insulator layer (e.g., Al2 O3
[49,53], SiO2 [51], or Si3 N4 [56]) is coated on the SiNW-FET
devices.
Compared with the ‘‘bottom-up’’ method, the ‘‘topdown’’ approach is more complex because the process
relies on high-resolution lithography. For this reason,
electron-beam lithography is necessary. Although the ‘‘topdown’’ approach needs many luxurious instruments, it has
advantages of using standard semiconductor techniques to
precisely design a desired device-array pattern without
problems of positioning SiNWs. Another challenge to the
‘‘top-down’’ method is that the minimum width of the produced SiNWs is around 100 nm. To overcome this barrier,
single SiNWs of triangular section were fabricated to reach
the transverse dimension of ≤20 nm with the length of several micrometers [55].
‘‘Bottom-up’’ SiNW-FETs
The ‘‘bottom-up’’ processes start with the growth of SiNWs,
normally in a chemical vapor deposition (CVD) reaction, followed by SiNW assembly (assisted by various techniques
that are discussed in the next section), and finally the
device fabrication via the photolithographic or electronbeam lithographic procedures [43]. With the ‘‘bottom-up’’
method, SiNWs can be synthesized catalytically in a CVD
reaction via the vapor—liquid—solid (VLS) growing mechanism (Fig. 3(b, i)) [57]. The synthesis is usually catalytically
assisted with metal nanoparticles [58,59] that not only catalyze the SiNW formation, but also control the size of
the as-synthesized SiNWs. Subsequently, the as-synthesized
SiNWs are suspended in ethanol solution and dispersed onto
a support silicon substrate (Fig. 3(b, ii)). In the following
photolithographic steps, a two-layer photoresist consisting
of LOR3A and S1805 was first deposited onto a silicon substrate by spin coating where the electrodes were defined
with a photomask design (Fig. 3(b, iii)). The next step is
136
to deposit metal for the source/drain contacts by thermal
evaporation (Fig. 3(b, iv)). Finally, the remaining photoresist
layer was lifted off by Remover PG (Fig. 3(b, v)) [43]. Compared with the ‘‘top-down’’ technique, the ‘‘bottom-up’’
method has the advantages of synthesizing SiNWs of high
crystallinity, designated dopant density, thin silicon oxide
sheaths, and easily controlled diameters in a cost-effective
preparation. However, without a deliberate alignment for
the randomly orientated SiNWs on the silicon substrate,
the device fabrication would suffer from inefficient fabrication yields, which could also limit their development in the
industrial applications. Therefore, the success of producing
high-quality SiNW-FETs calls for developing a uniform assembly of the ‘‘bottom-up’’ synthesized SiNWs on the support
substrates.
Nanowire assembly techniques
Significant efforts have been invested in developing generic
methods to align NWs for the device assembly to fabricate NW-FETs. Several techniques for the assembly of
NWs have been achieved, including flow-assisted alignment
[60], Langmuir—Blodgett technique [61—64], bubble-blown
technique [65], electric-field-directed assembly [66—69],
smearing-transfer method [70], roll-to-roll printing assembly [71], and polydimethylsiloxane (PDMS) transfer method
[72,73].
Flow-assisted alignment. In the flow-assisted alignment method (Table 1(a)), the suspended SiNWs were
passed through the microfluidic channel structures formed
between a PDMS mold and a flat SiO2 /Si substrate
[60]. The SiO2 /Si substrate was pre-modified with 3aminopropyltriethoxysilane (APTES), of which one end is
anchored to the SiO2 surface and the other end forms an
NH2 -terminated surface. This NH2 -terminated surface will
help the alignment of SiNWs via electrostatic interactions.
While the angular spread of the SiNWs in the flow direction
is flow-rate dependent, the density of the SiNWs assembled
on the SiO2 /Si substrate is time dependent.
Langmuir—Blodgett technique. The Langmuir—Blodgett
technique can be applied for the alignment of NWs/NTs. As
shown in Table 1(b), this solution-based method assembles
SiNWs in a monolayer of surfactant at the air—water interface and then compresses the SiNWs on a Langmuir—Blodgett
trough to a specified pitch. The aligned SiNWs are then
transferred to the surface of a substrate to make a uniform parallel array. Crossed SiNW structures could further
be formed by uniform transfer of a second layer of aligned
parallel SiNWs perpendicular to the first layer [61—64].
Compared with other methods, the Langmuir—Blodgett
technique can prepare an ultrahigh-density SiNW alignment.
Bubble-blown technique. The bubble-blown technique is
another physical assembly method (Table 1(c)), in which the
SiNWs were suspended in tetrahydrofuran solution and then
blown into a single bubble using a nitrogen flow to form SiNW
blown-bubble films [65]. The uniqueness of this method is
that blown-bubble films can be transferred to both rigid and
flexible substrates during the expansion process. It can also
be scaled to large wafers and non-rigid substrates [65].
Electric-field-directed
assembly. The
electric-fielddirected assembly of SiNWs is an intriguing and desirable
method. From the appropriate electrode design and adjust-
K.-I. Chen et al.
ment of the applied gate voltage (Vg ) vs. source-drain
voltage (Vsd ), Freer et al. reported that single SiNWs
could be assembled over 98.5% of 16,000 pre-patterned
electrode sites through controlling the balance of surface,
hydrodynamic, and dielectrophoretic forces (Table 1(d))
[66].
Smearing-transfer method. The smearing-transfer (or
contact-printing) method is one of a series of alignment
methods developed by Ali Javey’s group. This method is
based on a direct contact printing process that enables the
direct transfer and positioning of SiNWs from a donor substrate to a receiver chip. This simple method can efficiently
transfer a variety of NWs (such as SiNWs and Ge-NWs) to a
wide range of receiver substrates, including silicon and flexible plastics. The technique actually uses chemical ‘‘lawn’’
and ‘‘lubricant’’ to increase the density and alignment quality and can be regarded as a rapid, efficient, and economic
method (Table 1(e)) [70].
Roll-printing assembly. On the basis of the contactprinting process, Ali Javey’s group also developed an
approach for a scalable and large-area printing. Roll-to-roll
assembly has made it possible to produce highly ordered,
dense, aligned, and regular arrays of NWs with high uniformity and reproducibility using differential roll printing. The
schematic of the differential-roll-printing system setup and
the results of the wire alignment are shown in Table 1(f).
The optical and scanning electron microscopy (SEM) images
of the roller printed Ge-NWs on a Si/SiO2 substrate clearly
indicate the well-aligned and dense (about 6 NWs/␮m) NW
parallel arrays. The differential-roll-printing process is compatible with the smearing-transfer method and can also be
implemented in a wide range of rigid and flexible substrates
[71].
PDMS-transfer method. Chang et al. developed a NW alignment method using a PDMS stamp (Table 1(g)) [72]. In the
report, a high-speed roller (20—80 cm/min) was used to
assist the transfer of ZnO-NWs from the growth substrate
to a PDMS stamp; these NWs were then re-transferred from
the PDMS stamp to another receiver substrate. With this
method, NWs can be aligned with high density, providing
a convenient and efficient approach for the fabrication of
NW-FETs.
Array design and electrical measurement setup
The SiNW-FET devices could be fabricated following a
standard photolithographic procedure with a mask design
depicted in Fig. 4(a). The synthesized SiNWs were dispersed
on a SiO2 /Si substrate (typically 400 nm-thick SiO2 ). The
as-dispersed SiNWs in the central area (the reddish rectangles in Fig. 4(a) and (b)) were electrically connected by
metal leads (represented in yellow in Fig. 4(a)). The surfaces of the metal electrodes were further coated with an
insulating layer to prevent electric leakage during sensing experiments. The bottom inset graph enlarged from
the red region in Fig. 4(b) displays the array design, in
which the individual nanowire device was connected by
metal electrodes with a separation of several micrometers.
The individual SiNW situation can be seen with the image
scanned by atomic force microscope (AFM), as shown in
Fig. 4(c).
The experimental setup involved in electrical measurements includes a silicon chip (1.5 mm × 1.5 mm) containing
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
137
Table 1 Nanowire assembly techniques. (a) Schematic and results of a flow-assisted NW assembly method. The suspending NWs
were passed through a fluidic channel resulting in the alignment of NWs on a flat substrate, but often with low NW density. (b) A
flow-assisted NW Langmuir—Blodgett assembly method. This method involves packing aligned NWs in a monolayer of surfactant
at the air—water interface, followed by transferring to the surface of a substrate. The scalability and uniformity of alignment
for large-scale is still challenging; however, the assembled NWs are highly aligned and of high density. (c) A bubble-blown NW
assembly method. The NWs were suspended in tetrahydrofuran solution and then blown into a single bubble using a nitrogen
flow to form the SiNW blown-bubble films, followed by transferring to the surface of a substrate by direct contact. Although
this approach cannot align NWs to be of high density, it has advantage of fitting the receiver substrate in different shapes. (d)
An electric-field NW assembly method. The alignment is induced by polarizing NWs in an applied electric field. Under a delicate
control, the ratio for a successful alignment could be over 90%. (e) A smearing-transfer NW assembly method. This method is
based on a direct contact-printing process that enables direct transferring and positioning of NWs from a donor substrate to a
receiver chip. This method can be employed for large scale NWs transfer with high alignment and density. (f) A roll-printing
NW assembly method. This method applies a glass roller as the substrate for NWs growth, and then uses this roller as a NWs
donor to transfer NWs to a receiver substrate through shear force. (g) A PDMS-transfer NW assembly method. This method is an
amboceptor in the NWs transfer process, which adheres NWs by stamping the donor substrate and transferring NWs to another
receiver substrate. Although the directions of aligned NWs were not perfect, it is a convenient method with a potential to be
developed further in the future.
(a) Flow-assisted [60]
(b) Langmuir—Blodgett [61]
(c) Bubble-blown [66]
(d) Electric-field [65]
(e) Smearing-transfer [70]
(f) Roll-printing [71]
138
K.-I. Chen et al.
Table 1 (Continued )
(g) PDMS-transfer [72]
Reprinted from [60,61,65,66,70—72].
SiNW-FET device arrays, a PDMS microfluidic channel
(6.25 mm × 0.5 mm × 0.55 mm), and a detection system.
First, the silicon chip containing SiNW-FET device arrays
was mounted on a plastic circuit board and electrically connected with ∼30 ␮m-diameter aluminum wires (Fig. 4(d))
before electrical measurement. The PDMS microfluidic
channel was then placed in the middle of the chip to
allow sample solution delivery onto the SiNW-FET arrays
(Fig. 4(e)). The detection system including a lock-in
amplifier and a current pre-amplifier was to record the
electrical signals resulting from the binding events occurring on the SiNW-FET surface during sensing experiments
(Fig. 4(f)).
It is noted that the laminar flow in an ordinary microfluidic channel used in FET-based measurements may restrict
the detection sensitivity due to the diffusion-limited sample
delivery [74]. Comparatively, a specially designed microscale solution chamber with efficient sample mixing during
the fluid exchange has been demonstrated to improve the
detection sensitivity [49,75].
Reusable SiNW-FET system
In the application of SiNW-FETs for the biomedical diagnosis
of a particular target (e.g., an antigen), the corresponding receptor (e.g., the antibody) is usually modified on
the SiNW-FET surface prior to the detection. By virtue
of the strong and specific binding affinity between antigen and antibody under normal physiological conditions,
the receptor-modified SiNW-FET can serve as an extremely
sensitive sensor with high selectivity. By the same token,
because of this strong binding between the antigen and
antibody, it is difficult to remove the antigen—antibody
complex from the surface of SiNW-FET after detection,
meaning that a SiNW-FET could be used only for a single
measurement. With this limitation, consecutively quantitative analysis by a calibratable SiNW-FET is hard to
achieve.
To solve this problem, several reversible surface modification techniques have been developed recently, leading
to reusable SiNW-FET devices. Two well-known protein
trapping systems wildly used in protein purifications,
the GSH/GST-tag [51,76] and Ni2+ /His6 -tag [77—79], were
adapted to serve as a reversible surface modification method
on SiNWs and will be discussed in the following section. In
addition, the application of a cleavable disulfide bond that
served as a linker between the receptor and a SiNW-FET has
also been reported for use as a reusable SiNW-FET system
[49].
GSH/GST-tag
The reversible binding between glutathione S-transferase
(GST) and glutathione (GSH) has long been applied in protein purification. Through molecular cloning techniques,
the GST sequence can be incorporated into an expression
vector alongside the gene sequence encoding the protein of interest. Thus, various GST-fusion proteins can be
easily produced in a large scale via bacterial or mammalian expression systems. By using GSH-conjugated resins
to trap GST recombinant protein from whole cell extract
and then washing the resins with buffer to remove contaminating bacterial or mammalian proteins, the pure
GST-fused protein can be eluted easily by a high concentration of GSH. Taking advantage of the reversible GSH—GST
association—dissociation, Lin et al. adapted this method
to make SiNW-FET a reusable biosensor [51]. As illustrated
schematically in Fig. 5(a), a SiNW-FET was first modified with
GSH (referred to as GSH/SiNW-FET) and then anchored with
a particular GST-fused protein (referred to as protein-GST).
This protein-modified SiNW-FET could then be employed
to screen possible interacting proteins. After the sensing
measurements of protein—protein interactions, the used
protein-GST on the GSH/SiNW-FET could be easily removed
with a GSH (≥1 mM) washing solution. The reversible
GSH—GST association—dissociation has made the SiNW-FET
sensorial device reusable and calibratable, thus allowing
for quantitative analysis in sensing measurements. This
biologically modified SiNW-FET can be applied as an ultrasensitive biosensor for fast high-throughput screening of
biomolecular associations, such as protein—protein interactions, protein—DNA interactions, and protein—carbohydrate
interactions. Very recently, Lin et al. also applied this technique of using a reusable SiNW-FET to detect the interactions
of calmodulin with purified cardiac troponin I (∼7 nM) and
crude N-type Ca2+ channel extracts [76].
Ni2+ /His6 -tag
The polyhistidine-tag is an amino acid motif that consists
of multiple histidine (His) residues at the N- or C-terminus
of the protein. The total number of His residues may vary,
but there are normally six in the tag; therefore, it generally named a His6 -tag. Similar to the GST-tag system,
the His6 -tag is a popular and efficient system for purifying proteins via the reversible association—dissociation
between the His6 -tag and the affinity resins; the association is assisted with metal ions, either nickel or cobalt. The
reversible immobilization of His6 -tagged proteins to a sensor
surface was recently applied to CNT-FET [77] and demon-
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
139
Figure 4 (a) Mask design for the photolithographic fabrication of SiNW-FET device arrays. (b) Device arrays on a magnified scale.
Top: Optical image of the circuits in the area of the yellow square in (a); bottom: SEM image of a SiNW-FET array with a source-drain
separation of 2 ␮m. The scale bar is 50 ␮m. (c) The topograph of a SiNW-FET scanned by AFM. A SiNW of ∼50 nm in diameter is
connected by two Ni/Al (70 nm/100 nm in thickness) electrodes of ∼2 ␮m in separation. (d) The SiNW-FET device arrays on a silicon
chip (1.5 mm × 1.5 mm) are connected to a plastic circuit board with aluminum wires (∼30 ␮m in diameter). (e) A sample solution
was delivered onto the SiNW-FET arrays through a PDMS microfluidic channel (6.25 mm × 0.5 mm × 0.05 mm), which was designed to
couple with the SiNW-FET device arrays. (f) The variation of electrical signals was monitored by a detection system that combined
a lock-in amplifier and a current preamplifier.
Reprinted from [38].
strated on SiNW [78]. As demonstrated in Fig. 5(b), the
hexavalent Ni2+ ions held by the nitrilotriacetic acid (NTA)
chelator groups were chemically modified to the FET surface
and then bound to His6 -tagged protein through the coordination between His residues and the remaining two sites
of the hexavalent Ni2+ ions. The addition of imidazole or
ethylenediaminetetraacetic acid (EDTA) can compete with
the His—metal interaction to cause a reversed process of the
protein immobilization, resulting in the retrieval of the FET
surface. In comparison with the GSH/GST-tag system, the
smaller Ni2+ /His6 -tag has its advantages in the FET-based
sensing measurements. First, because of its smaller size,
more Ni2+ /His6 -tags could be anchored on the FET surface,
thus increasing the sensing sensitivity. Secondly, the binding
sites located on the smaller-sized Ni2+ /His6 -tag are closer
to the FET surface, resulting in a lesser screening effect
140
K.-I. Chen et al.
Figure 5 A schematic illustration for the reversible SiNW-FET system. (a) A SiNW-FET is first modified with 3-aminopropyltrimethoxysilane (APTMS) and 3-maleimidobenzoic acid N-hydroxysuccinimide ester (MBS) linkers, then functionalized with GSH
to form GSH/SiNW-FET. A particular GST-fusion protein (referred to as protein-GST) is anchored on the GSH/SiNW-FET via the
GST—GSH association. The protein-immobilized SiNW-FET is then employed for screening possible interacting proteins. At the end
of each measurement, the used protein-GSTs are removed with GSH (≥1 mM) washing solution, making the GSH/SiNW-FET a reusable
biosensor. (b) The Ni2+ /His6 -tag functionalized sensor surface was formed by (3-glycidyloxypropyl) trimethoxysilane (GPTMS) and
N-(5-amino-1-carboxypentyl) iminodiacetic acid (AB-NTA). His6 -tagged proteins can then be trapped on the sensor surface. After
the measurement, the His6 -tagged proteins can be removed by imidazole to retrieve the sensor surface.
Reprinted from [51,78,79].
on the FET detections, which also increases the sensing
sensitivity.
Sensing measurements
Size effect on sensing sensitivity
The wire size of a SiNW-FET can also affect the sensitivity
of the FET device. As illustrated in Fig. 6, the surface-tovolume ratio of thick wires is relatively small (Fig. 6(a))
compared to that of thin wires (Fig. 6(b)). Therefore, when
thick wires are approached by charged particles, the area
affected by the electric field exerted from the charged particles is only located at or close to the wire surface. Namely,
the interior areas of the wires could still be unaffected.
In sharp contrast, as the wire diameter decreased, say to
nanoscale, the surface-to-volume ratio increases drastically
and the influence of the external electric field could reach
the whole cross-section of the NW. As such, the induced
conductance change inside the NW-FET could overwhelmingly prevail over microwire-FETs [6,54]. Elfstrom et al.
have demonstrated the size-dependent sensitivity of SiNWFETs [54]. When SiNW-FETs of different wire widths are
immersed in an acidic buffer solution, the FET devices configured by smaller diameter SiNWs exhibit large conductance
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
141
Figure 6 An illustration for the concept of a size effect on the
conductance change in a wire. (a) For thick wires, the surfaceto-volume ratio is relatively small. When the wire surface is
approached by charge particles (red ball), only the conductance
near the wire surface is affected. There is still a large interior
area of the wire that might not be influenced (gray circle). (b)
As the wire diameter is reduced to nanoscale, the surface-tovolume ratio drastically increases. Therefore, the same external
electrical field (pink area) caused by the charge particles (red
ball) could influence most of the interior area of the NW, thus
drastically changing its conductance.
Reprinted from [6].
changes, whereas those of larger diameter SiNWs remain
unaffected.
Debye—Hückel screening
In the FET-based biosensing measurements, the solution
environment plays an important role in determining the
sensing performance. In order to create a surrounding similar to a normal physiological environment, like human
serum or urine, phosphate buffered saline (1× PBS, 137 mM
NaCl, 2.7 mM KCl, 10 mM Na2 HPO4 , 2 mM KH2 PO4 , pH 7.4
with NaOH) or phosphate solution (1× PS, 2.4 mM NaH2 PO4 ,
7.6 mM Na2 HPO4 , pH 7.4 with NaOH) was generally selected
as the matrix, in which analytes dissolved during the measurements. However, in solutions containing such high-salt
concentrations, the interaction potential (V(r)) between the
receptor and analytes that cause the conductance change
in the FET sensor could be partially screened by the strong
ionic strength of the electrolytic buffer solution, thus reducing the signals obtained from the electrical measurements.
The screening of V(r) in the FET measurements is enhanced
exponentially by the distance (rbs ) measured from the binding site of receptor—analyte complex to the FET surface and
can be represented as
V (r)e−r/D
at r = rbs
(1)
where D is the Debye—Hückel length [80,81] and is given by
D =
ε0 εr k B T
2NA e2 I
(2)
Figure 7 (a) A schematic showing the height of D from the
sensor surface for an electrolytic buffer solution. The horizontal
dashed lines mark the heights of D = 0.7, 2.4, and 7.4 nm for 1×
PBS (blue), 0.1× PBS (red), and 0.01× PBS (black), respectively.
(b and c) Real-time electrical measurements of the association
and dissociation of GST on a GSH/SiNW-FET in (b) 0.1× PBS
(black curve, D = 2.4 nm) and 1× PS (red curve, D = 1.9 nm)
and (c) 0.01× PBS (black curve, D = 7.4 nm) and 0.1× PS (red
curve, D = 6.1 nm).
Reprinted from [51].
where ε0 represents the vacuum permittivity, εr is the
relative permittivity of the medium, kB is the Boltzmann
constant, T represents the absolute temperature, NA is Avogadro’s number, e stands for the elementary charge, and I
represents the ionic strength of the electrolytic buffer solution. It is obvious that an electrolytic solution of higher ionic
strength (I) has a shorter D , thus creating a more severe
screening effect on the FET-based sensing measurements.
Calculations from Eq. (2) give D = 0.74 nm for the 1× PBS
solution and D = 1.94 nm for the 1× PS solution.
As represented schematically in Fig. 7(a), depending on
the rbs value in a FET measurement, the electrolytic buffer
solution needs to be properly selected to be of an appropriate D without jeopardizing the signal collection. In Fig. 7(b)
and (c), the screening effect due to the electrolytic buffer
solution was experimentally demonstrated from the binding of GST to GSH/SiNW-FET [51], where the conductance
142
K.-I. Chen et al.
Figure 8 (a) Real-time electrical measurements of the association of CaM-GST with a GSH/SiNW-FET in 0.1× PS solution
supplemented with 0.5 mM EDTA (pH 7.4). The arrow indicates the arrival of the CaM-GST solution. (b) Real-time detection of the
binding of K+ (red), Al3+ (green), and Ca2+ (blue) to CaM/SiNW-FET. The arrow indicates the arrival of the appropriate ion solution.
(c) Real-time detection of the binding of cardiac TnI to CaM/SiNW-FET in 0.1× PS solution supplemented with 100 ␮M Ca2+ . (d)
Specificity of CaM/SiNW-FET. The electrical conductance of CaM/SiNW-FET showed no response until binding of TnI. (e) Plot of
G vs. log[TnI]. The addition of various concentrations of TnI in 0.1× PS supplemented with Ca2+ (square) or 0.5 mM EDTA (circle). The red line represents a linear fit to the five concentration data points (correlation coefficient = 0.987). (f) Real-time electrical
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
change (G) of the GSH/SiNW-FET in measuring 15 nM GST
(Fig. 7(c)) in 0.01× PBS (black curve, D = 7.4 nm) or 0.1× PS
(red curve, D = 6.1 nm) is enhanced roughly fourfold compared to the measurements (Fig. 7(b)) in 0.1× PBS (black
curve, D = 2.4 nm) or 1× PS (red curve, D = 1.9 nm).
Gao et al. have recently reported that the optimal sensitivity of a SiNW-FET in biosensing measurement can be
achieved by judiciously selecting the subthreshold regime,
where the gating effect from target molecules is most effective due to the reduced screening of carriers inside the SiNW
[82]. The effectiveness of gating effect induced by target
molecules at the surface of a SiNW-FET sensor is determined
by the relative magnitude between carrier screening length
(Si ) and SiNW radius (R). In the high carrier concentration
regime (Si R), SiNW-FET works in a linear regime and the
conductance varies with gate voltage linearly. In the low
carrier concentration regime (Si R), the SiNW-FET works
in the depletion (subthreshold) regime and the conductance
varies with gate voltage exponentially. It is demonstrated
that the most sensitive SiNW-FET biosensor should utilize
the field gating effect of surface charges throughout the
whole cross-section of SiNW, which requires Si > R. In the
subthreshold regime of a SiNW-FET, carriers in the SiNW have
long screening length (Si > R) and the field effect of surface
charges can gate the whole SiNW, fully utilizing the high
surface-to-volume ratio of SiNW and effectively reaching the
optimal detection sensitivity of the FET sensor.
Applications of SiNW-FET sensors
Protein—protein interaction
A huge number of approaches have been developed
to understand molecular complex interactions, such as
protein—protein or protein—small molecule interactions.
For example, a fluorescence detection method combined
with a fiber-optic biosensor has been established to study
the binding kinetics of immunoglobulin G (IgG)/anti-mouse
IgG and human heart-type fatty acid-binding protein (its
antibody) [83]. However, this labeling detection method was
limited by some drawbacks. For instance, the surface characteristics of small proteins might be changed after chemical labeling, thus varying the labeling efficiency for different
proteins, which consequently makes accurate quantification detection difficult. Also, this labeling technique usually
requires a huge amount of time for the labeling procedures.
In recent years, some label-free detection techniques, such
as surface plasmon resonance imaging (SPRI) [84], AFM [85],
and SiNW-FET and CNT-FET [86,87] have been invented for
sensing protein—protein interactions. Among these sensing
approaches, SiNW-FET has attracted more and more atten-
143
tion lately for studying protein interaction mechanisms, not
only because of its real-time and label-free detection, but
also due to its high sensitivity and selectivity. An early measurement made by Cui et al. demonstrated the real-time
detection of streptavidin binding to biotin-modified SiNWFET [88]. They also explored the ability of biotin-modified
SiNW-FET to detect streptavidin at the concentration of
10 pM, which is much lower than the nanomolar-range detection level obtained from other techniques, such as the
stochastic sensing of single molecules [89].
In addition to the biotin—streptavidin investigation, the
concept of detecting protein—protein interactions with
SiNW-FET could be extended to broad applications. A
calmodulin (CaM)-modified SiNW-FET sensor has recently
been used to detect calcium ions (Ca2+ ) by Lin et al. [76];
their experiments also showed that Ca2+ -bound CaM is able
to activate various proteins involved in physiological activities, such as the binding between Ca2+ -bound CaM and
cardiac troponin I (TnI). CaM was anchored to a reusable
SiNW-FET (referred to as CaM/SiNW-FET) via the aforementioned GSH—GST association—dissociation. As shown in
Fig. 8(a), a dramatic increase in conductance verified the
successful binding of negatively charged GST-fused CaM
(referred to as CaM-GST, pI of CaM ∼4 and pI of GST ∼6.72)
to the p-type GSH/SiNW-FET. In order to examine how the
binding of various metal ions affects the conductance of
CaM/SiNW-FET, three different metal ions (Ca2+ , Al3+ , and
K+ ) were selected to be examined in this system. As demonstrated in Fig. 8(b), CaM/SiNW-FET showed a preference for
binding Ca2+ (blue curve), but not K+ (red curve) or Al3+
(green curve), according to the conductance change (G)
after the arrival of each metal ion solution. These results
reveal a high specificity of CaM/SiNW-FET for sensing Ca2+ .
Shown in Fig. 8(c) is the binding of the positively charged
protein troponin I (TnI, pI ∼9.3) onto CaM/SiNW-FET in 0.1×
PS (pH 7.4) containing 100 ␮M Ca2+ , which led to a sizable
decrease in the conductance of the FET sensor. The control experiments carried out in Fig. 8(d) reflected that the
association between CaM and TnI is specific and can only
be triggered in the presence of Ca2+ . It has been proven in
Fig. 8(e) that the G increased with a rising concentration of
TnI (i.e., log [CaM]) in the presence of Ca2+ . It is also demonstrated in Fig. 8(f) that the concentration of Ca2+ required
to activate the interaction between CaM and TnI was at the
micromolar level (i.e., 10−6 M Ca2+ ).
In addition, CaM/SiNW-FET was applied to detect Ca2+
channels in cell lysate. The N-type voltage-gated Ca2+
channels (VGCCs) located at the plasma membrane mediate the entry of Ca2+ into cells in response to membrane
depolarization. As illustrated in Fig. 8(g), transfected 293
T cells containing N-type VGCCs resuspended in 1× PBS
were sonicated and then centrifuged to isolate the mem-
measurement for determining the [Ca2+ ] required to activate the interaction between TnI and CaM, where the binding of TnI to
CaM/SiNW-FET was detected at various [Ca2+ ]. The result revealed that the minimal [Ca2+ ] ∼1 ␮M is needed to trigger the CaM
activation. (g) Schematic illustration for the detection of membrane fractions containing N-type VGCCs utilizing CaM/SiNW-FET.
Real-time electrical detections of the binding of N-type VGCCs to CaM/SiNW-FET in 0.1× PS supplemented with (h) 100 ␮M Ca2+ ; (i)
0.5 mM EDTA. (j) (Top graph) Electrical detection of the membrane fraction without the ␣1b subunit by CaM/SiNW-FET in 0.1× PS
supplemented with 100 ␮M Ca2+ and (bottom graph) electrical detection of N-type VGCCs by GST/SiNW-FET in 0.1× PS supplemented
with 100 ␮M Ca2+ .
Reprinted from [76].
144
K.-I. Chen et al.
Figure 9 (a) A structure of the associated PNA and DNA. (b) Schematic representation showing that the distance from the bound
DNA to the SiNW surface could be varied by controlling the location of the DNA—PNA hybridization. (c) Distinguishable resistance
changes in the PNA-modified SiNW-FET resulting from the varied hybridization sites measured for two different concentrations of
target DNAs. (d) Plot of the experimental ratio of resistance change vs. calculated distance (L) from DNA strands to the SiNW
surface.
Reprinted from [81].
brane fractions. The solution retreated in 0.1× PS was
subsequently introduced into CaM/SiNW-FET; the decreased
conductance shown in Fig. 8(h) indicates the binding of
VGCCs to CaM/SiNW-FET. On the other hand, an appreciable decrease in conductance of CaM/SiNW-FET was also
observed for the association of CaM with a peptide covering the IQ domain at the C-terminal of the N-type VGCC in
the absence of Ca2+ (Fig. 8(i)), which is consistent with a
previous study [90]. Finally in Fig. 8(j), two control experiments were performed to ensure that CaM/SiNW-FET was
Ca2+ channel-specific and that CaM was essential for the
detection of N-type VGCCs.
Lately, Zheng et al. has also demonstrated a new
methodology based on a frequency (f) domain electrical
measurement utilizing a SiNW-FET for protein detection
[91]. The power spectral density of voltage from a currentbiased SiNW-FET shows 1/f-dependence in frequency
domain for the measurements of antibody-functionalized
SiNW-FET devices in buffer solution or in the presence of
protein not specific to the antibody receptor. In the presence of the protein (antigen) which can be recognized
specifically by the antibody-functionalized SiNW-FET, the
frequency spectrum exhibits a Lorentzian shape with a characteristic frequency of several kHz. They observed the shape
of the frequency spectrum to monitor the binding events,
and further to determine the detection limit. With the help
of this new method in the frequency-domain measurement,
the detection sensitivity was claimed to increase by 10-fold.
All of the results outlined above suggest that proteinmodified SiNW-FET sensors exhibit high sensitivity and
excellent specificity and are able to detect target molecules
rapidly and precisely. This novel technique provides a
promising tool to study protein—protein or protein—small
molecule interactions and can be further applied to biomedical diagnosis.
DNA hybridization
In addition to the protein—protein interactions mentioned
above, SiNW-FETs were adapted for the detection of DNA
or RNA. Due to the large amount of negative charges in
the phosphate backbones of DNA or RNA, SiNW-FETs offer
a good candidate for monitoring DNA or RNA hybridizations, because the hybridizations cause the accumulation
or depletion of charge carriers in the SiNW-FET, leading to a
conductance change. Peptide nucleic acid (PNA), an artificially synthesized polymer similar to DNA, is commonly used
in biological research, especially in DNA or RNA hybridizations. As shown in Fig. 9(a), PNA hybridizes with DNA by
base pairing through hydrogen bonds. Because PNA has no
phosphate groups in its backbone, the binding of PNA/DNA
or PNA/RNA strands is stronger than that of DNA/DNA or
DNA/RNA duplexes due to the lacking of electrostatic repulsion. Hahm et al. have reported the real-time and label-free
detection of DNA with a PNA-modified SiNW-FET [92]. In
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
that study, PNA was anchored to the SiNW surface by the
strong interaction between avidin and biotin. The successful
PNA—DNA duplex formation was demonstrated by the observation of a sizable increase in conductance in the p-type
PNA-modified SiNW-FET, because of the negatively charged
phosphate backbones of DNA. Even the surface of the PNAmodified SiNW-FET was covered with a layer of avidin; this
ultrasensitive biosensor for sensing DNA is capable of detecting down to 10 fM.
The strategy utilizing PNA as a capture receptor has
also been applied to study the detection sensitivity of a
SiNW-FET by examining the distance from a binding site
of charged analytes to the SiNW surface [81]. The illustration in Fig. 9(b) shows that the distance from the bound
DNA to the SiNW surface can vary by controlling the location of the hybridization site. The neutral character of PNA
avoids background electric charges to interfere with the
binding phenomenon and allows for the hybridization to be
performed in a low ionic-strength environment with a high
signal-to-noise ratio. The species of target DNAs with different nucleotides (nt) in the study were designed to be
separated into a 22-nt fully complementary, a 19-nt complementary, a 16-nt complementary, a 13-nt complementary, a
10-nt complementary, and a 7-nt complementary DNA fragment. In addition, a non-complementary DNA was used as
a control. The hybridization of PNA—DNA was monitored by
the resistance change that results from the accumulation
of negative charges on the n-type PNA-modified SiNW-FET.
The resistance changes due to the hybridizations of the PNA
receptors with these seven different target DNAs at two
different concentrations have been recorded in Fig. 9(c).
It is noted that the resistance change of the 7-nt complementary DNA (∼11%) is much smaller than that of fully
complementary DNA (∼50%). The result reveals that when
the complementary segments become shorter, which means
that the distance between the bound DNA and the SiNW surface becomes longer, the ability of SiNW-FET to detect DNA
hybridization is reduced exponentially, as shown in Fig. 9(d).
This observation suggests that the corresponding detection
sensitivity is mostly dependent on the distance of the charge
layer to the SiNW surface.
Viral infection monitoring and early cancer
detection
Specific PNA-modified SiNW-FET sensors have recently been
established to diagnose Dengus virus infection [93]. As represented schematically in Fig. 10(a), the synthetic PNA
receptors were first anchored to the SiNW-FET surface.
A specific fragment (69 bp) derived from Dengus serotype
2 (DEN-2) virus genome sequences was selected as the
target DNA and amplified by the reverse transcriptionpolymerase chain reaction (RT-PCR). Distinctive resistance
changes between the two different PNA receptors (i.e.,
complementary and non-complementary to the target DNAs)
can be distinguished, as seen in Fig. 10(b). The detection
limit of this biosensor based on SiNW-FET was claimed to
be 10 fM (Fig. 10(c)). These investigations suggest that the
PNA-modified SiNW-FET sensor incorporated with RT-PCR has
145
been successfully developed for a rapid and ultrasensitive
diagnostic method of detecting Dengus virus.
In addition, this promising method allows the detection of
microRNAs (miRNAs) for early cancer diagnosis [94]. MiRNAs
have been characterized to play a significant role in the cell
development and to be related to a number of cancers and
neurological disorders. Therefore, the detection of miRNAs
becomes more and more important in the field of medical science. As illustrated in Fig. 10(d), a PNA-immobilized
SiNW-FET was used to probe miRNA by detecting PNA-miRNA
hybridization via base pairing. As shown in Fig. 10(e), from
the resultant resistance change in the PNA-immobilized
SiNW-FET, this approach exhibits an excellent detecting
specificity capable of discriminating a single base mismatch in miRNA. Moreover, as depicted in Fig. 10(f), the
application of a PNA-functionalized SiNW-FET to probe the
hybridization with complementary miRNAs is obviously preferential to a DNA-functionalized SiNW-FET, again indicating
that neutral PNA prefers to hybridize miRNAs. Also, the
PNA-functionalized SiNW-FET sensor is capable of sensing
a specific miRNA in total RNA extracted from HeLa cells.
This technique provides a promising tool for early cancer
detection in which the species and the amount of miRNAs
in the cancer cells were suggested to be different from
those of normal cells. The combination of PNA and SiNW-FET
provides a powerful technique to detect target molecules
rapidly and precisely. This PNA-functionalized SiNW-FET sensor could also be applied in medical diagnosis of cancer cell
growth or other diseases by simply varying the sequences of
the PNA capture receptors.
Peptide—small molecule interaction
The SiNW-FET system has also been applied to study
peptide—small molecule interactions, including ammonia
(NH3 ) and acetic acid (AcOH) [95]. As shown in Fig. 11(a),
the specific peptides were modified covalently to a SiNWFET (referred to as peptide/SiNW-FET). X-ray photoelectron
spectroscopy and water contact angle were utilized to
verify the successful attachment of peptides on the SiNWFET surface (Fig. 11(b)). To test the selectivity of a
peptide/SiNW-FET to AcOH, several experiments have been
conducted to detect the AcOH diluted in acetone, which
is a similar molecule to AcOH. In Fig. 11(c), an obvious increase (black curve) is obtained by subtracting the
response caused by the binding of AcOH to peptide/SiNWFET (referred to as AcOH-peptide/SiNW-FET, green curve)
from that of the addition of AcOH to peptide-free/SiNWFET (blue curve), indicating that the peptide/SiNW-FET
has an excellent specificity to AcOH in compound chemical backgrounds. Moreover, the performances of both
AcOH-peptide/SiNW-FET and NH3 -peptide/SiNW-FET were
investigated in simulated breath backgrounds as a closer
approximation toward medical applications. As demonstrated in Fig. 11(d), the electrical response from both FET
sensors can be observed after introducing the target analyte
(AcOH or NH3 ) in a background of 6% CO2 . These results show
that peptide/SiNW-FET can be used to monitor the exhaled
breath content at high sensitivities and is able to serve as
an electronic nose for further medical diagnosis.
146
K.-I. Chen et al.
Figure 10 (a) A schematic diagram of the RT-PCR product of DEN-2 hybridized to a PNA-functionalized SiNW-FET sensor. (b)
Specificity of the PNA-functionalized SiNW-FET to the RT-PCR product of DEN-2. The purified RT-PCR product was applied to
the complementary (black) and non-complementary (red) PNA-functionalized SiNW-FETs. (R − R0 )/R0 (%) represents the resistance change in percentage calculated from [(resistance after hybridization − resistance before hybridization)/resistance before
hybridization] × 100. (c) Resistance change vs. various concentrations of the RT-PCR product of DEN-2 from 100 fM (blue) to 1 fM
(black). A negative RT-PCR product was used as a control (light blue). (d) Schematic illustration of the label-free direct hybridization assay established for the ultrasensitive detection of miRNA. (e) Hybridization specificity demonstrated by the responses of the
PNA-functionalized SiNW-FET to fully complementary (black), one-base mismatched (red), and non-complementary (green) miRNA
sequences. (f) Comparison of the responses of the PNA-functionalized SiNW-FET (black) and a DNA-functionalized SiNW-FET (red)
to the complementary miRNA.
Reprinted from [93,94].
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
147
Figure 11 (a) A scheme of covalent attachment of peptides to a SiNW-FET. (b) Characterization of the bare SOI (red),
amine-terminated (green), and peptide-coupled (blue) surfaces by X-ray photoelectron spectroscopy. (Inset) Water contact angle
goniometric measurements of the surfaces. (c) Electrical response of an AcOH recognition peptide/SiNW-FET (blue) and an amineterminated peptide-free/SiNW-FET (green) to 1000 ppm acetone (introduced at time 5 min) and 100 ppm AcOH (introduced at time
20 min). The black curve is the differential response, obtained by subtracting the green curve from the blue curve. (d) Electrical
responses of an AcOH recognition peptide/SiNW-FET (blue) and an NH3 recognition peptide/SiNW-FET (red) to sequential influxes
of 6% CO2 , 100 ppm AcOH, and 100 ppm NH3 , introduced at the times indicated.
Reprinted from [95].
Biomarker detection
A biomarker is generally defined as something that can be
used as an indicator for a particular disease state or some
other biological state of an organism. For that reason, the
detection of specific biomarkers can be applied to disease
screening. For example, prostate-specific antigen (PSA) has
already been wildly applied to prostate cancer diagnosis
[96]. However, these biomarkers usually exist in the blood in
extremely low concentrations. Therefore, finding a method
to rapidly and precisely detect these biomarkers is an important issue in clinical diagnoses.
Recently, Zheng et al. utilized a SiNW-FET array for the
detection of multiple cancer markers [97]. This SiNW-FET
array, shown in Fig. 12(a), is composed of three independent
SiNW-FET devices on which different antibodies were immobilized. The antibodies used here were against prostate
specific antigen (PSA), carcinoembryonic antigen (CEA), and
mucin-1, respectively, each of which is a clinically confirmed
cancer marker [98,99]. Fig. 12(b) demonstrates the realtime detection of the bindings of the three cancer markers
(PSA, CEA, and mucin-1) to the specific SiNW-FETs. The conductance vs. time measurements showed the simultaneous
recording of the PSA, CEA, and mucin-1 solutions, which
were delivered sequentially to the SiNW-FET arrays. As displayed in Fig. 12(b), these three cancer markers induced
significant signals caused by their binding with the cognate
antibodies. By virtue of the ultrasensitive SiNW-FET, the
detection limit for these three cancer markers has advanced
to the pg/mL scale.
Although the measurement of the cancer markers using
the SiNW-FET turned out to be very successful [97], challenges remain in using the SiNW-FET devices to detect
these markers from a whole blood sample. The reasons
stem not only from biofoulings and/or non-specific bindings that might occur in the electrical measurements, but
also because the ionic strength of the whole blood could
cause a very short D to severely limit the FET signals [80].
148
K.-I. Chen et al.
Figure 12 (a and b) Illustration for the detection of multiple cancer markers (PSA, CEA, and mucin-1) with SiNW-FET arrays. (a)
The illustration of the structure of the SiNW-FET array, which is composed of three independent devices. Devices 1—3 are differentiated with different antibodies (1, red; 2, green; 3, blue) that are specific to the three different cancer markers (PSA, CEA, and
mucin-1). (b) Real-time detection of the bindings of three cancer markers (PSA, CEA, and mucin-1) to the specific SiNW-FET devices
modified with antibodies for PSA (NW1), CEA (NW2), and mucin-1 (NW3), respectively. The solutions were delivered to the SiNW-FET
arrays sequentially as follows: (1) 0.9 ng/mL PSA, (2) 1.4 pg/mL PSA, (3) 0.2 ng/mL CEA, (4) 2 pg/mL CEA, (5) 0.5 ng/mL mucin-1, (6)
5 pg/mL mucin-1. (The solutions were injected following the points indicated by black arrows.) The PSA, CEA, and mucin-1 cancer
markers induced signals only to the specific SiNW-FET modified with the cognate antibodies. (c) Illustration for detecting biomarkers
from whole blood with a MPC-FET system. The MPC-FET system is composed of a microfluidic purification chip (larger gray block),
a valve (pink), and a FET sensing cell (smaller gray block). (i) Primary antibodies (anti-PSA and anti-CA15.3) are immobilized on
the MPC with a photocleavable cross-linker. The valve (pink) directs fluid flow to exit the MPC to either the waste or the SiNW-FET
sensing cell (smaller gray block). (ii) Whole blood is injected into the MPC (with the valve set to the waste compartment; black
arrow shows the direction of fluid flow); meanwhile, biomarkers could bind to their cognate antibodies. (iii) After washing, the MPC
was exposed to UV irradiation (orange arrows). During the UV exposure, the photolabile cross-linker cleaves, allowing the release
of the antibody—antigen complexes into solution. (iv) The valve was now set to the SiNW-FET sensing cell (black arrow indicating
the direction of fluid flow) and the antibody—antigen complexes are transferred to the cell for label-free sensing by the SiNW-FET
arrays. (d) Response of an anti-PSA-functionalized SiNW-FET to an MPC-purified blood sample initially containing 2.5 ng/mL PSA.
(e) Response of an anti-CA15.3-functionalized SiNW-FET to an MPC-purified blood sample initially containing 30 U/mL CA15.3.
Reprinted from [97,100].
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
Therefore, pre-purifying the physiological fluid sample and
changing the buffer to low salt conditions before the sample
flows into the SiNW-FET sensor could substantially improve
the sensing measurements. Recently, Stern et al. developed
a microfluidic purification chip (MPC) system to pre-isolate
the target molecules, followed by using SiNW-FET arrays to
analyze the pre-purified sample [100]. As a demonstrative
example, prostate specific antigen (PSA) and carbohydrate
antigen 15.3 (CA15.3) were chosen to be analyzed with
the MPC-FET system; these samples are standard clinical
markers of prostate [98,99] and breast cancer [101,102],
respectively.
Fig. 12(c) illustrates schematically the structure and
operation of the MPC-FET system, which comprises a
microfluidic purification chip, a SiNW-FET sensing chamber,
and a valve that controls whether fluid exits from the MPC
to the waste or to the SiNW-FET sensing cell [99]. As shown
in Fig. 12(c), the operation procedures are as follows: (i)
the MPC is immobilized with primary antibodies (anti-PSA
and anti-CA15.3 in this case) via a photocleavable crosslinker. (ii) The whole blood is injected into the chip with the
valve set on the waste compartment. In this step, biomarkers could bind to their cognate antibodies in the MPC. (iii)
The MPC is washed before being exposed under UV irradiation to cleave the photolabile cross-linker. (iv) The valve
is now set to the FET sensing chamber; meanwhile, the
antibody—antigen complexes are transferred to the cell for
label-free sensing by the SiNW-FET arrays. With the help
of the MPC to pre-purify the sample, PSA of 2.5 ng/mL and
CA15.3 of 30 U/mL were able to be detected by SiNW-FET
from a whole blood sample as revealed in Fig. 12(d) and (e),
respectively.
Recording electrical and transmitter signals from
cells
Using nano- and neuro-technologies to couple electrical
interfacing with neural systems has great potential to unveil
many details of neuron studies [103]. In the past few
years, SiNW-FETs and CNT-FETs have been applied for electrophysiological measurements by recording signals from
neuron cells and tissues [22,23,38,104—107], e.g., recording the electrical signal from a single neuron [22] and
cardiomyocyte cells [104] and detecting the released neurotransmitter of CgA from living neurons [38]. In this section,
we will introduce some current studies that demonstrate
how SiNW-FET has been used to record these cell signals. Patolsky et al. have reported that hybrid SiNW-FET
arrays integrated with individual axons and/or dendrites
are capable of recording electrical signals from a single
neuron cell [22]. On the designed SiNW-FET array structures, they defined the adhesive zones where poly-lysine
was patterned for neuron cell growth. The strategy of
preparing these neuron-SiNW-FET devices is illustrated in
Fig. 13(a), where the square regions with 30—60 ␮m on the
edge allow cell body adhesion and ∼2 ␮m-wide lines support dendrite growth. Several challenges were undertaken
in the device-array fabrication to prevent the source and
drain electrodes from corrosion under the harsh conditions
of cell culture and the subsequent electrical measurements. A single-step lithography process combined with an
149
undercut multilayer resist was well designed to deposit
isotropic silicon nitride on the metal electrodes as a passivation layer. Devices prepared in this way were able to
survive under continuous cell-culture conditions at 37 ◦ C
for at least 10 days. With this strategy of using poly-lysine
to guide neuron growth, the goal of one-neuron/one-SiNW
device from an array has been reached, as shown in
Fig. 13(b).
After the successful growth of a single neuron cell, a
linear array with a multiple SiNW-FETs system (Fig. 13(c))
was exploited to simultaneously investigate the propagation and back-propagation of the action potential spikes in
axons and dendrites separately. As shown in Fig. 13(d), signal
propagation rates of 0.16 m/s for dendrite and 0.43 m/s for
axon were determined from a single neuron measurement.
In total, the signal propagation rates obtained from different
neurons revealed Gaussian distributions of 0.15 ± 0.04 (±SD)
and 0.46 ± 0.06 m/s for dendrites and axons, respectively
(Fig. 13(e)). These results are comparable to the reported
propagation rates measured by conventional electrophysiological and optical methods [108,109]. Furthermore,
SiNW-axon junction arrays were integrated to examine the
neuronal excitability at a level of 50 ‘‘artificial synapses’’
per neuron. The structure was selected to demonstrate the
capability of nanoelectronic devices for single-cell hybrid
structures at much higher densities. The optical image
shown in Fig. 13(f) represents the well-aligned neuron
growth across these 50 SiNW-FET devices. It is impressive
to see in Fig. 13(g) that the action potentials stimulated
intracellularly in the soma produced a mapping of the spike
propagation detected by the 43 functional devices over the
∼500 ␮m-long axon. The data obtained from these SiNW-FET
devices showed little decay in peak amplitude from NW1
to NW49, which is consistent with the active propagation
process.
In addition to the direct monitoring of the electrical signals from cells, SiNW-FETs and CNT-FETs have also been
used to detect the neuron transmitter [38] and adenosine
triphosphate (ATP) [110] released from living cells. Timko
et al. [23] lately used both planar and flexible polymeric
substrate-based SiNW-FET arrays to simultaneously record
the voltage-calibrated signals from the different parts of
embryonic chicken heart. All of these performances have
demonstrated that the nanoscale FET devices are promising for many more fields of basic and clinical studies of
cardiology in the future.
Three-dimensional localized bioprobes
Although a multitude of SiNW-FETs and CNT-FETs have
been exploited to record extracellular electrical signals
[22,23,38,104—107], these devices are normally created on
planar substrates, making it difficult for the devices to
detect the signals from arbitrary localization in three dimensions (3D). Therefore, it is highly desirable to develop a
movable 3D nano-FET, containing the necessary source (S)
and drain (D) electrical connections, that can be moved
to contact a cell and even into the cell. Very recently,
Tian et al. made a movable 3D nano-FET through the synthetic integration of kinked SiNWs [46,111]. As shown in
Fig. 14(a) and (b), these kinked SiNWs (either p-type or
150
K.-I. Chen et al.
Figure 13 Using SiNW-FETs to record neuronal axon signals. (a) A general schematic of the aligned NW-neuron device array. The
open blue rectangle highlights a single SiNW-neuron device. (b) Optical image of a single cortex neuron aligned across a single SiNW
device and (red box) a magnified image of the area. (c) Optical image of a cortex neuron with the axon and dendrite aligned on a
multi-SiNW structures. (d) Plot showing latency time as a function of distance from NW1 and NW6 for the axon (blue) and dendrite
(red), respectively, for a single neuron. (e) Histogram of propagation speed through axons (blue) and dendrites (red). (f) Optical
image of an aligned axon crossing an array of 50 SiNW devices with a 10-␮m inter-device spacing. (g) Electrical data from the 50
device arrays. The yield of the functional devices is 86% (43 devices were capable of conducting measurements). The peak latency
from NW1 (top arrow) to NW49 (bottom arrow) was estimated to be 1060 ␮s.
Reprinted from [22].
n-type) can be synthesized in a CVD reaction via the VLS
mechanism by varying the reactant pressure and the component of reactant gases during the SiNW growth, the
results of which showed that the lengths, doping components, and growth direction of the as-synthesized SiNWs
can be well controlled. Fig. 14(c) displays the image
of this 3D nano-FET probe, where the FET is located
at the tip of an acute-angle kinked SiNW that connects with the nanostructure arms composed of multilayer
photoresists.
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
151
Figure 14 (a—c) Schematic structure and SEM image of a 3D nano-FET probe. (a) Schematic of the nano-FET probe, which is
composed of a doubly kinked SiNW with a cis-configuration. The blue and pink regions respectively designate the source/drain (S/D)
and nanoscale FET channel. The dimensions of the nano-FET segment (pink) are about 80 nm × 80 nm × 200 nm. (b) SEM image of a
doubly kinked SiNW with a cis-configuration. The scale bar is 200 nm. (c) SEM image of the 3D nano-FET probe. The orange arrow
indicates the nano-FET and the purple star marks the flexible device support, which is composed of poly(methylmethacrylate) and
SU-8 photoresist. The scale bar is 5 ␮m. (d—f) Illustrations of cellular entry by the 3D nano-FET probe. (d) Schematics of lipid-coated
nano-FET probe entrance into a cell. The stepwise processes of the nano-FET probe invasion into the cell (blue) are illustrated from
I to III. The nano-FET (pink) is coated with the phospholipid bilayers (dark purple) before the cellular entry. (e) Microscopy images
of the differential interference contrast of an HL-1 cell and 3D nano-FET probe. Images include the processes as the cell approaches
(I), contacts and internalizes (II), and is retracted from (III) the nano-FET probe. A pulled-glass micropipette (inner tip diameter
∼5 ␮m) at the lower right corner (I—III) was used to manipulate and voltage-clamp the HL-1 cell. The scale bars are 5 ␮m. (f)
Illustration of the electrical recording at the differential interference contrast of nano-FET and HL-1 cell. The potential vs. time
from the nano-FET probe shows a sharp potential drop (∼50 mV within 250 ms) while the probe tip is within the cell (II), then returns
to the baseline when the cell was detached from the nanowire probe end (III). Green and blue arrows mark the beginnings of cell
penetration and withdrawal, respectively.
Reprinted from [111].
This 3D nano-FET probe (n-type) with a sensitivity of
4—8 ␮S/V in 1× PBS solution has been successfully inserted
into a single cell to take an intracellular electrical signal.
As illustrated schematically in Fig. 14(d), the phospholipid
bilayers coated on the nano-FET surface, which were embellished by fusion with micro-phospholipid vesicles, allow
the nano-FET to fuse with the cell membrane [112—114]
and enter the cell [111]. Fig. 14(e) shows optical images
of the real process of the HL-1 cell approached by the
phospholipid-modified nano-FET. The diagram in Fig. 14(f)
shows the potential changes of the phospholipid-modified
nano-FET in contact with the isolated HL-1 cell at differential interference contrasts. The potential vs. time from
the nano-FET probe shows a sharp potential drop (∼50 mV
within 250 ms) while the probe tip is within the cell (II), and
then returns to the baseline when the cell was detached
from the nanowire probe end (III). These experimental
results have shown that the phospholipid-modified nano-FET
probe surface is critical for assisting a rigid probe access to
the intracellular region for intracellular electrical recording
[111,115].
Summary
Over the past years, the invention of biosensors has become
increasingly important for many biological purposes. Biosensors constructed based on traditional methods usually
require massive sample preparations or have low detection
sensitivity. In contrast, nanoscale techniques have attracted
more attention because of their abilities to overcome these
difficulties. In this review, we described the fabrications
and properties of SiNW-FET devices. In addition, we discussed the potential applications of SiNW-FETs in the fields
of biomedical sciences.
From the very beginning, the most frequent applications of SiNW-FET sensors were focused on monitoring
protein—protein interactions, such as the binding between
an antibody and antigen. However, without a reusable
system, the practical uses of SiNW-FETs for screening interacting proteins would be limited. A way to solve this problem
is to apply a GSH/GST-tag or Ni2+ /His6 -tag reversible system
to the SiNW-FET surface, which makes SiNW-FETs capable of
acting as a reusable biosensor.
152
In addition to protein—protein interactions, SiNW-FETs
were also applied to detect specific DNA by PNA—DNA
hybridization. It has been proven that the PNA probe is much
more sensitive than a single-strand DNA-modified device for
detecting DNA or RNA sequences. This strategy provides
a platform not only to diagnose virus infection, but also
to monitor early cancer symptoms. Moreover, the peptidemodified SiNW-FETs with high sensitivity could be employed
to detect specific targets, e.g., the noninvasive breath monitoring for molecular disease indicators, or the discrimination
of chemical odorants from interfering gas mixtures. The
ability of detecting target biomolecules from a cocktail
solution, such as blood, enables SiNW-FETs to become a
novel biosensor for the future medical applications. Furthermore, the capability of SiNW-FETs to record the electrical
response from live neurons suggests this nanoscale device
to be implanted in living systems for clinical uses. The successful construction of 3D nano-FET probes for recording
electrical signals from single cells makes SiNW-FET sensors
achieve a high goal in the design of nanobiosensors.
Apart from other sensing techniques, several significant
advantages of SiNW-FETs can be addressed into direct, labelfree, real-time electrical detection, ultrahigh sensitivity,
excellent selectivity, and the potential for multiple sensing
arrays. In addition, SiNW-FETs can be prepared in high-yield
with a reproducible device characteristic, which is suitable
for the integration of large-scale and complex NW sensor
arrays in different possible applications. However, some limitations to the FET-based sensors need to be concerned in
biosensing measurements, such as the effect of sample solution’s ionic strength on the detection sensitivity of a FET
device. For example, a blood serum sample with high ionic
strength requires a desalting step before analysis to achieve
the highest sensitivity [97]. Notwithstanding the restriction,
we believe that SiNW-FETs will play a significant role in
the development of biomedical sensors in the future. The
progress on this development will be an exciting issue from
both science and technology perspectives. In the long run,
we envision that the demand of this nanobiosensing technique will be focused on how to advance this novel sensory
gadget at a commercial level. For example, to a convenient
purpose, minimizing the size of a SiNW-FET device to an
easily carried electronic sensor will benefit people requiring
immediate diagnosis. Moreover, studies of the compatibility
between SiNW-FET sensors and human body, e.g., for disease
diagnosis or blood testing, should be an important topic for
the future biomedical applications.
Acknowledgment
This work was supported by the National Science Council of
Taiwan under contract number NSC 99-2627-M-002-001.
References
[1] M. Farre, D. Barcelo, TrAC Trends Anal. Chem. 22 (2003) 299.
[2] Y. Wang, H. Xu, J.M. Zhang, G. Li, Sensors 8 (2008) 2043.
[3] S. Viswanathan, H. Radecka, J. Radecki, Monatsh. Chem. 140
(2009) 891.
[4] S. Rodriguez-Mozaz, M.J.L. de Alda, D. Barcelo, Anal. Bioanal.
Chem. 386 (2006) 1025.
K.-I. Chen et al.
[5] D.R. Thevenot, K. Toth, R.A. Durst, G.S. Wilson, Biosens. Bioelectron. 16 (2001) 121.
[6] D. Grieshaber, R. MacKenzie, J. Voros, E. Reimhult, Sensors 8
(2008) 1400.
[7] F.M. Raymo, I. Yildiz, Phys. Chem. Chem. Phys. 9 (2007) 2036.
[8] N.C. Tansil, Z.Q. Gao, Nano Today 1 (2006) 28.
[9] A.M. Morales, C.M. Lieber, Science 279 (1998) 208.
[10] J.T. Hu, T.W. Odom, C.M. Lieber, Acc. Chem. Res. 32 (1999)
435.
[11] X.F. Duan, Y. Huang, Y. Cui, J.F. Wang, C.M. Lieber, Nature 409
(2001) 66.
[12] N.L. Rosi, C.A. Mirkin, Chem. Rev. 105 (2005) 1547.
[13] X.G. Liang, S.Y. Chou, Nano Lett. 8 (2008) 1472.
[14] M. Curreli, R. Zhang, F.N. Ishikawa, H.K. Chang, R.J. Cote,
C. Zhou, M.E. Thompson, IEEE Trans. Nanotechnol. 7 (2008)
651.
[15] A. Poghossian, S. Ingebrandt, M.H. Abouzar, M.J. Schoning,
Appl. Phys. A: Mater. Sci. Process. 87 (2007) 517.
[16] F. Patolsky, G.F. Zheng, C.M. Lieber, Nanomedicine 1 (2006)
51.
[17] I.Y. Park, Z.Y. Li, X.M. Li, A.P. Pisano, R.S. Williams, Biosens.
Bioelectron. 22 (2007) 2065.
[18] G. Gruner, Anal. Bioanal. Chem. 384 (2006) 322.
[19] E. Katz, I. Willner, ChemPhysChem 5 (2004) 1085.
[20] R.J. Chen, H.C. Choi, S. Bangsaruntip, E. Yenilmez, X.W. Tang,
Q. Wang, Y.L. Chang, H.J. Dai, J. Am. Chem. Soc. 126 (2004)
1563.
[21] A.B. Kharitonov, M. Zayats, A. Lichtenstein, E. Katz, I. Willner,
Sens. Actuators B 70 (2000) 222.
[22] F. Patolsky, B.P. Timko, G.H. Yu, Y. Fang, A.B. Greytak, G.F.
Zheng, C.M. Lieber, Science 313 (2006) 1100.
[23] B.P. Timko, T. Cohen-Karni, G.H. Yu, Q. Qing, B.Z. Tian, C.M.
Lieber, Nano Lett. 9 (2009) 914.
[24] W.R. Yang, K.R. Ratinac, S.P. Ringer, P. Thordarson, J.J. Gooding, F. Braet, Angew. Chem.—–Int. Ed. 49 (2010) 2114.
[25] Y. Ohno, K. Maehashi, Y. Yamashiro, K. Matsumoto, Nano Lett.
9 (2009) 3318—3322.
[26] T. Cohen-Karni, Q. Qing, Q. Li, Y. Fang, C.M. Lieber, Nano Lett.
10 (2010) 1098—1102.
[27] C. Li, M. Curreli, H. Lin, B. Lei, F.N. Ishikawa, R. Datar, R.J.
Cote, M.E. Thompson, C.W. Zhou, J. Am. Chem. Soc. 127
(2005) 12484.
[28] A. Choi, K. Kim, H.I. Jung, S.Y. Lee, Sens. Actuators B 148
(2010) 577—582.
[29] B.L. Allen, P.D. Kichambare, A. Star, Adv. Mater. 19 (2007)
1439.
[30] P.A. Hu, J. Zhang, L. Li, Z.L. Wang, W. O’Neill, P. Estrela,
Sensors 10 (2010) 5133.
[31] D.R. Kauffman, A. Star, Chem. Soc. Rev. 37 (2008) 1197.
[32] A. Star, J.C.P. Gabriel, K. Bradley, G. Gruner, Nano Lett. 3
(2003) 459.
[33] R.J. Chen, S. Bangsaruntip, K.A. Drouvalakis, N.W.S. Kam, M.
Shim, Y.M. Li, W. Kim, P.J. Utz, H.J. Dai, Proc. Natl. Acad. Sci.
U. S. A. 100 (2003) 4984.
[34] H.M. So, K. Won, Y.H. Kim, B.K. Kim, B.H. Ryu, P.S. Na, H.
Kim, J.O. Lee, J. Am. Chem. Soc. 127 (2005) 11906.
[35] A. Star, E. Tu, J. Niemann, J.C.P. Gabriel, C.S. Joiner, C. Valcke, Proc. Natl. Acad. Sci. U. S. A. 103 (2006) 921.
[36] X.W. Tang, S. Bansaruntip, N. Nakayama, E. Yenilmez, Y.L.
Chang, Q. Wang, Nano Lett. 6 (2006) 1632.
[37] K. Besteman, J.O. Lee, F.G.M. Wiertz, H.A. Heering, C.
Dekker, Nano Lett. 3 (2003) 727.
[38] C.W. Wang, C.Y. Pan, H.C. Wu, P.Y. Shih, C.C. Tsai, K.T. Liao,
L.L. Lu, W.H. Hsieh, C.D. Chen, Y.T. Chen, Small 3 (2007) 1350.
[39] C.C. Tsai, C.C. Yang, P.Y. Shih, C.S. Wu, C.D. Chen, C.Y. Pan,
Y.T. Chen, J. Phys. Chem. B 112 (2008) 9165.
[40] R. Martel, T. Schmidt, H.R. Shea, T. Hertel, P. Avouris, Appl.
Phys. Lett. 73 (1998) 2447.
Silicon nanowire field-effect transistor-based biosensors for biomedical diagnosis and cellular recording investigation
[41] S. Heinze, J. Tersoff, R. Martel, V. Derycke, J. Appenzeller, P.
Avouris, Phys. Rev. Lett. 89 (2002) 4.
[42] H.R. Byon, H.C. Choi, J. Am. Chem. Soc. 128 (2006)
2188—2189.
[43] F. Patolsky, G.F. Zheng, C.M. Lieber, Nat. Protoc. 1 (2006)
1711.
[44] Y. Wu, Y. Cui, L. Huynh, C.J. Barrelet, D.C. Bell, C.M. Lieber,
Nano Lett. 4 (2004) 433.
[45] D.D.D. Ma, C.S. Lee, F.C.K. Au, S.Y. Tong, S.T. Lee, Science
299 (2003) 1874.
[46] B.Z. Tian, P. Xie, T.J. Kempa, D.C. Bell, C.M. Lieber, Nat.
Nanotechnol. 4 (2009) 824.
[47] Y. Cui, X.F. Duan, J.T. Hu, C.M. Lieber, J. Phys. Chem. B 104
(2000) 5213.
[48] M.C. Lin, C.J. Chu, L.C. Tsai, H.Y. Lin, C.S. Wu, Y.P. Wu,
Y.N. Wu, D.B. Shieh, Y.W. Su, C.D. Chen, Nano Lett. 7 (2007)
3656—3661.
[49] E. Stern, J.F. Klemic, D.A. Routenberg, P.N. Wyrembak, D.B.
Turner-Evans, A.D. Hamilton, D.A. LaVan, T.M. Fahmy, M.A.
Reed, Nature 445 (2007) 519.
[50] Y.B. Zhang, Y. Xiong, X.A. Yang, Y. Wang, W.H. Han, F.H. Yang,
J. Nanosci. Nanotechnol. 10 (2010) 7113—7116.
[51] S.P. Lin, C.Y. Pan, K.C. Tseng, M.C. Lin, C.D. Chen, C.C. Tsai,
S.H. Yu, Y.C. Sun, T.W. Lin, Y.T. Chen, Nano Today 4 (2009)
235.
[52] T. Kudo, T. Kasama, T. Ikeda, Y. Hata, S. Tokonami, S.
Yokoyama, T. Kikkawa, H. Sunami, T. Ishikawa, M. Suzuki,
K. Okuyama, T. Tabei, K. Ohkura, Y. Kayaba, Y. Tanushi, Y.
Amemiya, Y. Cho, T. Monzen, Y. Murakami, A. Kuroda, A. Nakajima, Jpn. J. Appl. Phys. 48 (2009).
[53] O. Knopfmacher, A. Tarasov, W.Y. Fu, M. Wipf, B. Niesen,
M. Calame, C. Schonenberger, Nano Lett. 10 (2010) 2268—
2274.
[54] N. Elfstrom, R. Juhasz, I. Sychugov, T. Engfeldt, A.E. Karlstrom, J. Linnros, Nano Lett. 7 (2007) 2608.
[55] G. Pennelli, Microelectron. Eng. 86 (2009) 2139.
[56] X.T. Vu, J.F. Eschermann, R. Stockmann, R. GhoshMoulick, A.
Offenhausser, S. Ingebrandt, Phys. Status Solidi A 206 (2009)
426—434.
[57] R.S. Wagner, W.C. Ellis, Appl. Phys. Lett. 4 (1964) 89.
[58] Y. Wu, J. Xiang, C. Yang, W. Lu, C.M. Lieber, Nature 430 (2004)
61.
[59] W. Lu, C.M. Lieber, J. Phys. D: Appl. Phys. 39 (2006) R387.
[60] Y. Huang, X.F. Duan, Q.Q. Wei, C.M. Lieber, Science 291 (2001)
630.
[61] D. Whang, S. Jin, Y. Wu, C.M. Lieber, Nano Lett. 3 (2003)
1255.
[62] S. Jin, D.M. Whang, M.C. McAlpine, R.S. Friedman, Y. Wu, C.M.
Lieber, Nano Lett. 4 (2004) 915.
[63] X.L. Li, L. Zhang, X.R. Wang, I. Shimoyama, X.M. Sun, W.S.
Seo, H.J. Dai, J. Am. Chem. Soc. 129 (2007) 4890.
[64] A. Tao, F. Kim, C. Hess, J. Goldberger, R.R. He, Y.G. Sun, Y.N.
Xia, P.D. Yang, Nano Lett. 3 (2003) 1229.
[65] G.H. Yu, A.Y. Cao, C.M. Lieber, Nat. Nanotechnol. 2 (2007)
372.
[66] E.M. Freer, O. Grachev, D.P. Stumbo, Nat. Nanotechnol. 5
(2010) 525.
[67] L.F. Dong, J. Bush, V. Chirayos, R. Solanki, J. Jiao, Nano Lett.
5 (2005) 2112.
[68] A. Motayed, M.Q. He, A.V. Davydov, J. Melngailis, S.N. Mohammad, J. Appl. Phys. 100 (2006) 114310.
[69] S. Raychaudhuri, S.A. Dayeh, D.L. Wang, E.T. Yu, Nano Lett.
9 (2009) 2260.
[70] Z.Y. Fan, J.C. Ho, Z.A. Jacobson, R. Yerushalmi, R.L. Alley, H.
Razavi, A. Javey, Nano Lett. 8 (2008) 20.
[71] R. Yerushalmi, Z.A. Jacobson, J.C. Ho, Z. Fan, A. Javey, Appl.
Phys. Lett. 91 (2007) 203104.
[72] Y.K. Chang, F.C.N. Hong, Nanotechnology 20 (2009) 195302.
153
[73] Y.K. Kim, P.S. Kang, D.I. Kim, G. Shin, G.T. Kim, J.S. Ha, Small
5 (2009) 727.
[74] P.E. Sheehan, L.J. Whitman, Nano Lett. 5 (2005) 803—807.
[75] J.H. Chua, R.E. Chee, A. Agarwal, S.M. Wong, G.J. Zhang,
Anal. Chem. 81 (2009) 6266—6271.
[76] T.W. Lin, P.J. Hsieh, C.L. Lin, Y.Y. Fang, J.X. Yang, C.C. Tsai,
P.L. Chiang, C.Y. Pan, Y.T. Chen, Proc. Natl. Acad. Sci. U. S.
A. 107 (2010) 1047.
[77] E. Kang, J.W. Park, S.J. McClellan, J.M. Kim, D.P. Holland,
G.U. Lee, E.I. Franses, K. Park, D.H. Thompson, Langmuir 23
(2007) 6281.
[78] Y.C.C. Liu, N. Rieben, L. Iversen, B.S. Sorensen, J. Park, J.
Nygard, K.L. Martinez, Nanotechnology 21 (2010) 245105.
[79] Y.S. Lo, D.H. Nam, H.M. So, H. Chang, J.J. Kim, Y.H. Kim, J.O.
Lee, ACS Nano 3 (2009) 3649.
[80] E. Stern, R. Wagner, F.J. Sigworth, R. Breaker, T.M. Fahmy,
M.A. Reed, Nano Lett. 7 (2007) 3405.
[81] G.J. Zhang, G. Zhang, J.H. Chua, R.E. Chee, E.H. Wong, A.
Agarwal, K.D. Buddharaju, N. Singh, Z.Q. Gao, N. Balasubramanian, Nano Lett. 8 (2008) 1066.
[82] X.P.A. Gao, G.F. Zheng, C.M. Lieber, Nano Lett. 10 (2010)
547—552.
[83] C.H. Lin, H.Y. Chen, C.J. Yu, P.L. Lu, C.H. Hsieh, B.Y. Hsieh,
Y.F. Chang, C. Chou, Anal. Biochem. 385 (2009) 224.
[84] H.S. Ro, B.H. Koh, S.O. Jung, H.K. Park, Y.B. Shin, M.G. Kim,
B.H. Chung, Proteomics 6 (2006) 2108.
[85] M. Lee, D.K. Kang, H.K. Yang, K.H. Park, S.Y. Choe, C.
Kang, S.I. Chang, M.H. Han, I.C. Kang, Proteomics 6 (2006)
1094.
[86] G. Pampalakis, S.O. Kelley, Analyst 134 (2009) 447.
[87] K. Maehashi, K. Matsumoto, Y. Takamura, E. Tamiya, Electroanalysis 21 (2009) 1285.
[88] Y. Cui, Q.Q. Wei, H.K. Park, C.M. Lieber, Science 293 (2001)
1289.
[89] L. Movileanu, S. Howorka, O. Braha, H. Bayley, Nat. Biotechnol. 18 (2000) 1091.
[90] H.Y. Liang, C.D. DeMaria, M.G. Erickson, M.X. Mori, B.A.
Alseikhan, D.T. Yue, Neuron 39 (2003) 951.
[91] G.F. Zheng, X.P.A. Gao, C.M. Lieber, Nano Lett. 10 (2010)
3179—3183.
[92] J. Hahm, C.M. Lieber, Nano Lett. 4 (2004) 51.
[93] G.J. Zhang, L. Zhang, M.J. Huang, Z.H.H. Luo, G.K.I. Tay,
E.J.A. Lim, T.G. Kang, Y. Chen, Sens. Actuators B 146 (2010)
138.
[94] G.J. Zhang, J.H. Chua, R.E. Chee, A. Agarwal, S.M. Wong,
Biosens. Bioelectron. 24 (2009) 2504.
[95] M.C. McAlpine, H.D. Agnew, R.D. Rohde, M. Blanco, H. Ahmad,
A.D. Stuparu, W.A. Goddard, J.R. Heath, J. Am. Chem. Soc.
130 (2008) 9583—9589.
[96] J. Hernandez, I.M. Thompson, Cancer 101 (2004) 894.
[97] G.F. Zheng, F. Patolsky, Y. Cui, W.U. Wang, C.M. Lieber, Nat.
Biotechnol. 23 (2005) 1294.
[98] A.J. Vickers, C. Savage, M.F. O’Brien, H. Lilja, J. Clin. Oncol.
27 (2009) 398.
[99] S.F. Shariat, P.T. Scardino, H. Lija, Can. J. Urol. 15 (2008)
4363.
[100] E. Stern, A. Vacic, N.K. Rajan, J.M. Criscione, J. Park, B.R.
Ilic, D.J. Mooney, M.A. Reed, T.M. Fahmy, Nat. Nanotechnol.
5 (2010) 138.
[101] M. Rubach, J.J. Szymendera, J. Kaminska, M. Kowalska, Int.
J. Biol. Markers. 12 (1997) 168.
[102] M. Uehara, T. Kinoshita, T. Hojo, S. Akashi-Tanaka, E.
Iwamoto, T. Fukutomi, Int. J. Clin. Oncol. 13 (2008) 447.
[103] W.L.C. Rutten, Annu. Rev. Biomed. Eng. 4 (2002) 407.
[104] T. Cohen-Karni, B.P. Timko, L.E. Weiss, C.M. Lieber, Proc. Natl.
Acad. Sci. U. S. A. 106 (2009) 7309—7313.
[105] T.S. Pui, A. Agarwal, F. Ye, N. Balasubramanian, P. Chen, Small
5 (2009) 208.
154
K.-I. Chen et al.
[106] G. Cellot, E. Cilia, S. Cipollone, V. Rancic, A. Sucapane, S.
Giordani, L. Gambazzi, H. Markram, M. Grandolfo, D. Scaini,
F. Gelain, L. Casalis, M. Prato, M. Giugliano, L. Ballerini, Nat.
Nanotechnol. 4 (2009) 126.
[107] J.F. Eschermann, R. Stockmann, M. Hueske, X.T. Vu, S.
Ingebrandt, A. Offenhausser, Appl. Phys. Lett. 95 (2009)
083703.
[108] L.G. Nowak, J. Bullier, Exp. Brain. Res. 118 (1998) 489.
[109] S. Antic, G. Major, D. Zecevic, J. Neurophysiol. 82 (1999)
1615.
[110] Y.X. Huang, H.G. Sudibya, D.L. Fu, R.H. Xue, X.C. Dong, L.J.
Li, P. Chen, Biosens. Bioelectron. 24 (2009) 2716.
[111] B.Z. Tian, T. Cohen-Karni, Q.A. Qing, X.J. Duan, P. Xie, C.M.
Lieber, Science 329 (2010) 830.
[112] L.V. Chernomordik, M.M. Kozlov, Nat. Struct. Mol. Biol. 15
(2008) 675.
[113] X.J. Zhou, J.M. Moran-Mirabal, H.G. Craighead, P.L. McEuen,
Nat. Nanotechnol. 2 (2007) 185.
[114] N. Misra, J.A. Martinez, S.C.J. Huang, Y.M. Wang, P. Stroeve,
C.P. Grigoropoulos, A. Noy, Proc. Natl. Acad. Sci. U. S. A. 106
(2009) 13780.
[115] B.D. Almquist, N.A. Melosh, Proc. Natl. Acad. Sci. U. S. A. 107
(2010) 5815.
Kuan-I Chen was born in Kaohsiung, Taiwan
in 1981. She received B.S. in Chemistry at the
National Taiwan Normal University in 2004.
She then graduated from the University of
Queensland, Australia for her Ph.D. degree
in biochemistry (2009). From 2009 to date,
she works as a post-doctoral researcher at the
Institute of Atomic and Molecular Sciences,
Academia Sinica, Taiwan. Her interests are
focused on mediated electrocatalysis and the
study of protein—RNA interaction by silicon
nanowire field-effect transistor.
Bor-Ran Li is currently a postdoctoral
research fellow in the National Taiwan University under the supervision of Professor
Yit-Tsong Chen. He received his B.S. (1999)
and M.Sc. (2001) degree from the Department
of Applied Chemistry, National Chiao-Tung
University in Taiwan. Following this, he serviced as a national compulsory research
assistant at Academia Sinica (2001—2005),
and then moved to Scotland and obtained
his Ph.D. degree in chemistry from Edinburgh
University, UK (2009). His current interests are focused on the design
and the fabrication of nanowire field-effect transistors and also
applications of field-effect transistors in biochemical and biological
research fields.
Yit-Tsong Chen is a professor in the Chemistry
Department of National Taiwan University.
He is also an adjunct Research Fellow at
the Institute of Atomic and Molecular Sciences, Academia Sinica in Taiwan. He joined
the faculties in 1991 after receiving his
B.S. degree from National Taiwan University
(1980) and Ph.D. degree from the University of Chicago (1988) and having postdoctoral
research training at MIT (1989—1991). In his
early research career, he was interested in
studies of the superfine structures, vibrational dynamics, and Rydberg states of fundamental molecules using a variety of nonlinear
laser optics. In the late 1990s, fascinated by the nanoscale world,
he shifted part of his research to catalytic growth and spectroscopic characterization of nanoparticles and nanowires. His current
interests are focused on applying field-effect transistors, scanning
probe microscopy, and optical microscopy as biosensors to study
protein—protein interactions, extracellular ionic fluctuation, cellular exo-endocytosis, and neuron—neuron interactions.