DNA Transport in
Hierarchically-Structured
Colloidal-Nanoparticle Porous-Wall
Nanochannels
Deying Xia,† Thomas C. Gamble,‡ Edgar A. Mendoza,§ Steven J. Koch,†,|
Xiang He,† Gabriel P. Lopez,‡ and S. R. J. Brueck*,†,⊥
Center for High Technology Materials, UniVersity of New Mexico, 1313 Goddard, SE,
Albuquerque, New Mexico 87106, Center for Biomedical Engineering and Department
of Chemical and Nuclear Engineering, Department of Physics and Astronomy,
Department of Electrical and Computer Engineering, UniVersity of New Mexico,
Albuquerque, New Mexico 87131, Redondo Optics Inc., 811 N. Catalina AVenue, Suite
1100, Redondo Beach, California 90277
Received January 20, 2008; Revised Manuscript Received April 1, 2008
ABSTRACT
We report a simple approach to the formation of 3D colloidal nanoparticle structures incorporating enclosed mesoscopic structures through
a simple process of spin-coating-driven directed self-assembly onto lithographically defined polymer templates. Removal of the buried polymer
patterns by high temperature calcination results in the formation of hierarchically enclosed channels, continuous networks, isolated cavities,
and multilayered structures with high stability and environmental resistance. These channels are used to investigate the transport of DNA
molecules in constrained geometries.
An exciting direction for nanoparticle structures is the
fabrication of complex three-dimensional (3D) hierarchical
structures with potential applications to multilayer photonic
crystals, chemical sensors, catalysis, and biotechnology.
Micro- and nanoparticles have been used as templates for
preparation of porous metallic nanostructures and monodis-
perse colloidal crystals.1,2 Directed self-assembly of metal
nanoparticles (e.g., Au) into alumina membranes forms
porous metallic tubes.3 The fabrication of spherical, cylindri-
cal, and hollow colloidal crystals using spray and microcap-
illaries has also been studied using template-assisted ap-
proaches.4–7 Three-dimensional periodic structures with
micrometer-scale periods have been fabricated with nano-
particle inks.8 Well-defined 3D and multiplayer nanostruc-
tures with deposited metal (Au) have been formed by a
nanotransfer printing approach.9,10 Even though nanochannel
structures for nanofluidic applications have been fabricated
using traditional etching and thermal oxidation approaches
using both interferometric and nanoimprint lithographies,11–13
there have been relatively few reports of the fabrication of
enclosed mesoscopic structures using nanoparticles as build-
ing blocks, thereby providing porous confinement layers
around the voids.14 These structures have evident biomimetic
organization (similar to trees, skin, hair, etc.) with long
transport channels interconnected by porous media along with
hierarchical structures for fluid access that will be important
both for laboratory studies of biological processes and for
biomimetic applications.
Directed self-assembly, the combination of bottom-up self-
assembly and top-down pattern definition, is a promising
approach for the fabrication of complex micro- and nano-
structures.15,16 Generally, and in this work, mesoscopic
enclosed structures (.particle diameter) have been defined
by lithographic techniques while the nanoscale interparticle
separations giving rise to the porosity (,particle diameter)
are generated through self-assembly approaches.
In previous work, we demonstrated 1D and 2D patterned
nanoparticle structures on flat and patterned surfaces using
directed self-assembly.17,18 Here, we present an approach for
fabrication of enclosed mesoscopic silica nanoparticle struc-
tures on planar Si surfaces using simple and inexpensive
* Corresponding author. E-mail: brueck@chtm.unm.edu.
† Center for High Technology Materials, University of New Mexico.
‡ Center for Biomedical Engineering and Department of Chemical and
Nuclear Engineering.
§ Redondo Optics Inc.
| Department of Physics and Astronomy, University of New Mexico.
⊥ Department of Electrical and Computer Engineering, University of New
Mexico.
NANO
LETTERS
2008
Vol. 8, No. 6
1610-1618
10.1021/nl080190s CCC: $40.75  2008 American Chemical Society
Published on Web 05/07/2008

approaches including spin-coating and interferometric li-
thography (IL). This approach provides a flexible, rapid, and
inexpensive fabrication suite, using only low-cost and large-
area lithography techniques while avoiding costly and
complex processing steps such as etching and metallization.
Photoresist (PR) patterns defined by interferometric lithog-
raphy were used as a soft template around which silica
nanoparticles were self-assembled. Additional structure at
larger scales can be added, either in the same or different
resist levels, using traditional mix-and-match lithographic
processes. Multiple layer stacks, as high as four photoresist
layers, are demonstrated, allowing complex functionality.
High-temperature calcination was used to remove the PR
and enhance the structural stability of the nanoparticle
assembly by strengthening the bonding between adjacent
particles (localized sintering or necking). Using this approach,
the fabrication of complex 3D structures incorporating
defined length scales over a 107 scale range from ∼5 nm to
2 cm is demonstrated. The hierarchical integrated organiza-
tion provides for facile interfacing between macro- and
nanoscales.
IL is a powerful technique for fabrication of a wide range
of samples.19 IL is a parallel optical lithography approach
that provides an inexpensive, large-area capability (typical
samples in this work were ∼2 × 2 cm2). In conventional
photolithography, the time and cost required to fabricate a
photomask can be a significant issue, particularly for dense
nanoscale features patterned by e-beam lithography. IL does
not require a photomask, with the tradeoff of reduced pattern
flexibility (periodic patterns) for individual layers.15 Mix-
and-match lithographic approaches with lower resolution
tools can provide some of the requisite pattern flexibility.
There are many active research areas for the applications of
IL technology such as fabrication of photonic crystals for
the visible spectrum20 and metamaterial structures exhibiting
a negative permeability and negative refractive index.21,22
Silica nanoparticles are commercially available, easy to
prepare, compatible with both silicon microfabrication and
biological species, and easy to functionalize. In addition, the
use of silica nanoparticles for providing the walls (and tops)
of enclosed channels is very attractive for biological ap-
plications due to their extended optical transparency range
(UV-THz).11
Several examples of 1D enclosed nanostructures are shown
in Figure 1. A colloidal suspension of 50 nm diameter silica
nanoparticles was used to form 1D enclosed channels with
an ∼1 µm period as shown in Figure 1a-d. Well-defined,
1D stacked PR/antireflection coating (ARC) polymer patterns
were generated with IL using negative PR and a developable
ARC (Figure 1a). After six cycles of spin-coating with a
dilute aqueous silica suspension (2 wt %) and removal of
the PR and ARC by high temperature calcination, relatively
large channels (∼700 nm wide) were fabricated with high
fidelity to the PR/ARC templates (Figure 1b-c). Smaller
cross section nanochannels are readily available, as will be
demonstrated below. One important advantage of this ap-
proach for forming the enclosed nanostructures is the
uniformity over a large area. The typical size of a Si wafer
Figure 1. One-dimensional enclosed channels with silica particles. (a) SEM image of 1D channel with negative PR 500P. (b-c) 1000 nm
period channels formed with 50 nm silica nanoparticles. (d) Photograph of a typical, 2 × 2 cm2, 1D channel sample showing uniformity.
(e-f) SEM images of 1D channels with negative PR 500P, 1000 nm period, and 160 nm silica particle. (g) SEM image of enclosed channel
structures with 50 nm silica particles on all sides, with 500 nm periods atop a two-layer silica nanoparticle blanket film: (top left) monolayer
film of 50 nm silica nanoparticles; (top right) PR/wet-I patterns atop a two-layer 50 nm silica nanoparticle film. (h) SEM image of enclosed
channels with 500 nm periods on heterogeneous 50/160/50 nm silica particle films.
Nano Lett., Vol. 8, No. 6, 2008 1611

sample was 2 × 2 cm2 as shown in Figure 1d. A uniform
yellow diffraction color was observed with white-light, tilted-
angle illumination, indicating a uniform array of nanostruc-
tures; detailed SEM measurements confirm this result.
Larger diameter (∼160 nm) silica particles were used to
form enclosed channels with a ∼1 µm period using negative
PR 500P. The channels have a 600 × 500 nm2 cross section
(Figure 1e). The silica particle array structure with these
larger and more nearly monodisperse particles is hexagonal
close-packed. The top seal layer has roughly five layers of
silica particles (Figure 1f). We can extend this approach to
particle sizes large enough to directly form photonic crystals.
The lithography methods used for this purpose can be IL or
conventional photolithography. Fabrication of buried linear
and three-dimensional extrinsic defects within self-assembled
colloidal photonic crystals has been demonstrated recently
via a directed self-assembly strategy involving a combination
of top-down photolithography and bottom-up colloidal as-
sembly.23,24
Furthermore, enclosed channel structures with nanopar-
ticles on all four sides can be fabricated (Figure 1g-h). To
form this structure, a blanket film (monolayer or multiple
layer) of silica particles was first assembled on a flat sample
surface using spin-coating deposition with appropriate diluted
silica particle suspensions. A monolayer film of ∼50 nm
diameter silica nanoparticles was deposited on a planar Si
surface with ∼2 wt % suspension (top left inset in Figure
1g). Therefore, approximately two-layer-deep silica particle
arrays were formed with a two-step deposition under
these same conditions. The IL defined PR/ARC patterns were
formed atop the self-assembled silica nanoparticle films.25,26
In this case, using a positive PR (top right insert in Figure
1g). Well-defined enclosed channels with silica particles on
all sides are clearly observed with high fidelity to the
PR/ARC patterns. Free-standing channel structures can also
be fabricated with a metal layer, such as Cr, or other re-
lease layer between the Si substrate and the silica particle
layer.27
Finally, heterostructure colloidal crystal films can be
formed by varying the particle properties (size, composition)
within the multiple colloidal nanoparticle deposition stages.
This possibility is demonstrated by Figure 1h. The structure
on the bottom is composed of a heterogeneous colloidal
crystal (50/160/50 nm diameters) on a Si substrate. Two
layers of 160 nm diameter silica particles are sandwiched
between 50 nm diameter silica colloidal films. Similar
composite structures (Figure 1h) have been obtained previ-
ously by convective deposition using two different sphere
sizes.28 These channel structures with silica nanoparticles on
all sides confirm the extensibility and suitability of this
approach to the fabrication of linear extrinsic defects
(waveguides) for colloidal nanoparticle photonic crystal
applications.24
High temperature calcination (∼700-1000 °C in air
ambient) was used to remove the sacrificial PR and ARC
patterns and to enhance the mechanical strength of the
resultant channel structures by sintering of the silica spheres
in the immediate vicinity of the nanoparticles contact points.
The ordering and porosity of silica nanoparticle films
decrease when going from the low to high temperature.7 At
the highest temperatures investigated, coalescence of the
nanoparticles occurs. The silica nanoparticle melting point
was reduced substantially from that of an amorphous bulk
material (1986 K).29
The samples treated at T < 700 °C present a much lower
mechanical stability than those annealed at T ∼ 950 °C. The
formation of necks between the spheres caused by an
incipient sintering process starts taking place at T ∼ 800 °C
through viscous flow and is responsible for the sample
strengthening.30 When samples are annealed at T > 950 °C,
the nanoparticle spheres deform, losing their spherical shape.
Investigating the temperature effects on enclosed channel
structures, we draw a similar conclusion to that of Miguez31
that the mechanical strength of channel structures was low
after calcination at T j 700 °C. We also tested the samples
for calcinations above 1000 °C for 2 h; the channel structures
collapsed at this high temperature. Therefore, we treated the
samples at the range of 700-900 °C in a room air ambient
for ∼2 h. SEM images of samples calcined at 115 and
800 °C, respectively, are shown in Figure 2. There are no
apparent structural differences between the two samples. The
silica particles remain spherical, and the channel structures
are well-defined and preserved for treatment at 800 °C for
2 h in air (Figure 2b).
For application in micro- and nanofluidics, the environ-
mental stability of these 1D enclosed channels becomes
important. The prepared channeled samples have remarkable
mechanical stability, with no apparent degradation with
storage in laboratory ambient for long periods (∼months).
We also tested the as-prepared channel structures with
immersion in water and acetone. Examination of structures
before and after immersion in water and acetone (see Figure
S1 of the Supporting Information) indicate that the channel
structures are robust with no observable degradation for
immersion in solvents for a month. In addition, the inner
surfaces of the channels are hydrophilic due to the silica
particle surface chemistry and the calcination processing.
This property will be of benefit for micro- and nanofluidic
applications with aqueous-based fluids. The stability of these
one-dimensional enclosed structures is more than adequate
for many micro- and nanofluidic applications.
Enclosed air cavities and continuous, covered open
networks (with nanoparticle pillars holding up nanoparticles
Figure 2. High-resolution SEM images of 50 nm silica nanoparticle
film and channel structure. (a) Films after baking in air at 115 °C
for 3 min. (b) Channel structure after 800 °C 2 h calcination (bottom
right insert: large area image).
1612 Nano Lett., Vol. 8, No. 6, 2008

roofs) were also easily prepared using this approach, with
double exposures in the IL step to produce 2D photoresist
patterns as shown in Figure 3. PR posts and holes were
produced using positive and negative PR, respectively, in
the IL step (see Figure S2 of the Supporting Information).
Enclosed isolated air cavities were formed after deposition
of silica particles on 2D positive PR/ARC patterns (posts)
and removal of PR/ARC by calcination. The PR/ARC post
shapes were preserved as air cavities (Figure 3a-b). The
existence of isolated air cavities was easily verified by SEM
examination of cleaved samples. The left side of the image
does not exhibit air cavities, as the cleavage line was not
exactly along the center line between the cavities (Figure
3a).
The inverse pattern, a continuous network structure with
an upper sealing layer supported by isolated posts, was
fabricated with negative PR as shown in Figures 3c-d. As
the period of the 2D air-hole patterns increased, the sample
fabrication became more successful (see Figure S3 of the
Supporting Information for 500 nm period). Images in Figure
3c-d show continuous sealing with isolated posts with a 1
µm period. After spin-coating of silica particle suspension
and calcination, the final enclosed structures have broad posts
and thick sealing. Even though some posts did not contact
the bottom (blue arrow in Figure 3c), the surface flatness
and occurrence of defects in the enclosed structures are much
improved compared with a 500 nm period structure. In both
500 and 1000 nm period structures, the post heights are
shorter than the total thickness of PR/ARC, suggesting that
the holes in the PR/ARC pattern were not completely filled
during the spin-coating. We investigated thin PR patterns
(∼250 nm) with a large period (∼1000 nm) (see Figure S3d
of the Supporting Information) to overcome the low unifor-
mity, defects and inconsistency of heights of post and PR/
ARC. We investigated the use of smaller silica nanoparticles
(∼15 nm diameter), of more dilute silica nanoparticle
suspensions, and adjustment of the spin-coating parameters
to reduce the defect density without much impact. We
examined filled structures after spin-coating and before
calcination. The silica nanoparticles only partially filled the
holes in the PR/ARC matrix, which resulted in an inconsis-
tency between the heights of the silica posts and PR/ARC
(see Figure S4 of the Supporting Information). This vari-
ability in filling the PR/ARC holes over a large area likely
contributes to defects and nonuniformity of the final film.
The partial filling of the holes may be a result of surface
hydrophobicity of the isolated air holes in the PR/ARC layer.
There is a complex interplay between surface tension and
fluid forces during the spin-coating that results in this partial
filling of the holes. Compared to the enclosed air cavity
structures, the continuous network structure has more defects
(for example, shortened posts: blue arrow in Figure 3c) and
poorer surface flatness of the top sealing layer (green arrow).
The uniformity of the final enclosed structures depends
on the surface wetting properties of the PR pattern surface,
PR structures, and silica particle suspension. The success
rate and process latitude in forming large-area uniform
patterns, from easy to more difficult, is: 1D channel, 2D
isolated air cavity, and 2D continuous network. For large
spaces around the PR patterns, it is relatively easy for the
nanoparticle suspension to flow around the hydrophobic PR
obstructions and provide a complete pattern, while small
isolated features in the polymer template are more difficult
to fill.
The surface wetting properties depend on the micro- and
nanostructures and have a complex relationship with the
detailed PR structures. For the 1D PR patterned samples,
the macroscopic behavior of the colloidal fluid provides a
simple diagnostic of the level of filling. For partially filled
structures, e.g., colloidal particles below the top of the PR
Figure 3. Two-dimensional enclosed structures fabricated with 50 nm silica nanoparticles. (a-b) SEM images of 2D isolated air cavity
with 500 nm period and positive PR. (c-d) SEM images of 2D continuous air network with 1000 nm period and negative PR.
Nano Lett., Vol. 8, No. 6, 2008 1613

lines, the colloidal fluid formed an elongated drop along the
direction of the PR lines before spinning. This anisotropic
wetting phenomenon was also observed on 1D macropat-
terned surfaces and explained by Takahara et al.32 As
expected, the 2D symmetry PR patterned surface has more
nearly isotropic wetting properties, although faceting along
specific symmetry directions of the underlying pattern is still
evident for partially filled patterns. We can make use of the
anisotropic wetting phenomenon on 1D PR patterns for
potential applications in water harvesting surfaces, drug
release coatings, open-air nanochannel devices, and labora-
tory-on-chip integrated structures.33 The elongation of the
drop for the 1D case and the faceting of the edge of the
drop along the symmetry lines of the 2D pattern are
illustrated in the Supporting Information (see Figure S5 of
the Supporting Information). The static contact angles also
indicate that the surface wetting properties depend on the
period, PR line:space ratio, and PR pattern symmetry. More
detailed investigation of the relationship between the PR
structures and static contact angles is underway. A thorough
understanding of the relationship between PR structures and
surface wetting properties will be helpful in improving the
uniformity of the resultant enclosed structures prepared with
aqueous particle suspensions.
This fabrication approach can be extended to multilayer
structures. Previously, we have demonstrated two layer
channel samples, fabricated with 50 nm silica nanoparticles,
through repeated processing on a single substrate with
independent calcinations for each layer.14 In fabricating
multiple-layered enclosed structures, it is more effective to
perform a single calcination step following deposition of the
complete stack of PR and nanoparticle films. There are two
main advantages to the single-step calcination: (1) reduced
fabrication time with fewer processing steps and (2) uniform
sintering across the multiple layers.
To verify the practicality of single-step calcination, we
first fabricated a three-layered parallel channel structure.
Three steps of IL and spin-coating of silica particle suspen-
sion were followed by a single calcination. We also
intentionally deleted one IL step and while still performing
the spin-coating between the second and third layer so that
the final channel structures are four layers high with three
parallel channel layers and a thicker interlevel spacer between
the top two channel layers. SEM images of final multiple-
layered structures are shown in Figure 4a-c. Three-layered
parallel structures with one-dimensional enclosed channels
are clearly observed. Even though the same IL parameters
were employed in all three steps, the final channel profiles
in these three layers were different, mainly due to the varying
surface condition of each layer for IL. The upper layer has
smaller channel cross sections compared to those in the initial
layer. The thicker spacer layer between second and third
channel layers is clearly evident.
We also fabricated multiple-layered structures with four
channel layers with alternating, orthogonal orientations using
a single calcination as shown in Figure 4d-h. The perpen-
dicular channels are observed in Figure 4d-f. The smooth
inside wall of channels is obvious from Figure 4e. The open
channel cross section over a long length is evident in both
directions (Figure 4g-h shows orthogonal perspectives). It
is also easy to fabricate multiple-layered stacks with hybrid
enclosed structures such as one-dimensional channels in some
Figure 4. Multiple-layered channels with 500 nm period and 50 nm silica nanoparticles using single-step calcination. (a-c) SEM images of
three-parallel layered channels using single-step calcination: (a) enlarged image of bottom two layers; (b) overview of three-layered channels; (c)
large area image of three-layered channels. (d-f) SEM images of four perpendicular-layered channels: (d) image from one direction; (e) enlarged
image of part of (a); (f) image from orthogonal direction. (g) Large area image of (d). (h) Large area of (f).
1614 Nano Lett., Vol. 8, No. 6, 2008

layers and two-dimensional air cavities in other layers using
single-step calcination.
The as-prepared enclosed structures with nanoparticles
have optical applications such as photonic crystals. For
example, photonic crystal structures could be fabricated with
high refractive index nanoparticles arranged in multiple-
layered “Lincoln-log” structures.
Enclosed nanoparticle structures provide the opportunity
to fabricate fluidic systems with heterogeneous, hierarchical
porosities. We believe enclosed structures will have many
applications for catalysis, chemical/biological sensing, and
biomolecular separations. Additionally, they may find utility
for interconnecting laboratory-on-chip arrays of microfluidic
and microreactor devices that perform chemical and catalytic
transformations as well as ion, molecule, and macromolecule
separations. The continuous network structure with an upper
sealing layer supported by isolated posts has potential
applications in laboratory-on-a-chip devices; for example,
the separation of biological components/molecules and
particles by size can be performed through such a connected
network with post arrays using laminar fluidic flow.34 The
multiple-layered enclosed structures have applications in
separation and detection of biomolecules, efficient bioanalysis
by improving separation efficiency (layered enclosed struc-
tures), and biomolecular detection sensitivity (large surface
area of porous nanochannel structures) as well as concentrat-
ing bacteria and viruses in bioanalytical chemistry.
Multiple-layered nanochannels are biomimetic structures
akin to the cellular organization of trees, animal skin, bird
feathers, and hair.35,36 The study of nanofluidics in enclosed
porous nanochannels will allow investigations of biological
processes as well as potentially leading to biomimetic
applications.
As an initial illustration of these potential applications,
we observed DNA transport in 1D porous nanochannel
structures, using capillary action (hydrophilic surface tension)
Figure 5. Fluorescent image and DNA extension in 1D channels (500 nm × 600 nm channels, 1200 nm pitch). (a) Schematic of channeled
sample with inlet well. (b) Fluorescent dye image in a region far from inlet well. (c) Confocal microscope image of Hind-III digested
lambda DNA in region A. (d) Microscope image of DNA in region B just at the edge of well, some channels are blocked (marked by red
arrow). (e-f) Microscope image of DNA in region A with 3 s interval. (g-k) Microscope image of DNA using diluted DNA buffer
solution (100 times less than in (e) and (f)) in region A with 0.15 s interval (motion of individual DNA molecules marked with colored
dashed circles). In all cases, the fluid flow is bottom to top of the image.
Nano Lett., Vol. 8, No. 6, 2008 1615

as the driving force. Many reports have explored the use of
micro- and nanochannel structures to separate, manipulate,
and elongate DNA,13,37–40 and fabrication techniques for
nanofluidic channel devices have been discussed.41 Compared
to previous fabrication approaches, our method has the
advantages of ease of fabrication and, importantly for the
first time, the provision of interconnected porous structures.
We compare the DNA transport in 500 and 130 nm wide
nanochannels. First experiments used 500 nm wide porous
channels composed of 50 nm silica nanoparticles with a 1200
nm period to test the DNA transport. A circular well was
etched for injecting fluorescent dye and/or DNA solution
(Figure 5a). To demonstrate the channel continuity over long
distances, the 1D channels were filled with DI water
containing fluorescent dye by capillary action. Figure 5b
shows the fluorescent images captured in a region far from
the injecting well. The liquid filled the channels uniformly
and continuously. All of the images in Figures 5 and 6 are
oriented so that the flow is from bottom to top of the image.
In initial investigations, we observed directly a new nanof-
luidic phenomenon: oscillatory drying/filling in the advancing
water fronts in these enclosed porous channels. We attribute
this phenomenon to the combination of evaporation through
pores between nanoparticles in the sealing layer and capillary
driving forces. The oscillatory motion exhibits group be-
havior, “waves,” resulting from communication between
adjacent nanochannels by virtue of the porosity.
Fluorescently stained DNA molecules (λ-phage DNA 48.5
kbp, contour length L ) 16.5 µm, New England Biolabs)
were stained with YOYO-1 dye (Molecular Probes) and were
readily transported into 500 nm wide porous channels using
capillary filling. Figure 5c shows a confocal microscopy
image of aligned DNA filling the individual nanochannels
in long streaks. This image was taken in region A, indicated
in Figure 5a. The optical resolution is insufficient to resolve
any transverse structure. The uniform intensity and similar
lengths are reminiscent of aligned DNA molecules.13 How-
ever, the apparent lengths (20-50 µm) are longer than the
known contour length of these polymers in channels of
comparable size (∼2 µm),42 so there must be some clustering
of the DNA molecules (perhaps concatamers of the self-
cDNA molecules). Without extra forces, the capillary driven
motion of DNA reaches as far as several hundreds of
micrometers from the well, limited by the evaporation of
the fluid through the porous tops. Figure 5d shows a video
frame capture of the fluorescence just at the interface between
the well and the nanochannels (region B in Figure 5a). This
image was taken after the initial filling of the nanochannels.
Continuous drying by evaporation from the porous tops of
the nanochannels and refilling from the reservoir provides
continuous additions and flow of DNA into and within the
channels. There are several interesting observations: (1) there
is clearly a cooperative effect between channels because the
DNA just at the channel edges proceeds in a “wave” that is
consistent across many channels, and this is likely a result
of the porous sidewalls that ensure a uniform fluid pressure
across the channels; (2) the fluorescent intensity varies
significantly within different nanochannels, again suggesting
some aggregation of individual DNA molecules in these
relatively wide channels; (3) there is a clearly discernible,
but unresolved, fluorescent glow just at the entrance to the
nanochannels from additional DNA molecules that are
“stacked up” at the channel entrances; (4) the length of the
fluorescent streaks associated with highly mobile DNA, the
short streaks just inside the channels, is much shorter than
the relatively immobile fluorescence sources corresponding
Figure 6. Fluorescent image and DNA extension in narrowed 1D channels (130 nm × 400 nm channels, 800 nm pitch). (a-b) SEM images
of channels. (c) Confocal microscope image of DNA in nanochannels. (d) Microscope image of DNA in nanochannels; the fluid flow is
bottom to top of the image.
1616 Nano Lett., Vol. 8, No. 6, 2008

to the long streaks in Figure 5d, which are likely the same
as those in Figure 5c; (5) there are some blocked channels
with no fluorescence, and some of these are indicated by
the red arrows in the Figure 5d.
The DNA velocities are clearly a function of position and
of previous filling cycles. The DNA motion on initial filling
of dry channels is so rapid that we could not resolve the
individual DNA molecules with standard frame rates. After
several cycles of filling, the DNA molecules accumulate into
bright lines in the top half of the images (further from the
well), reflecting increased concentration of the DNA. As the
DNA motion continues, the edge of the region of bright lines
moves down as seen in Figure 5f. Using a diluted DNA
buffer solution, we easily observe the motion of individual
DNA molecules as shown in Figure 5g-k. The dotted lines
follow individual DNA molecules across four consecutive
frames. Clearly, the velocities are slowing as the particles
reach the top of the field, representing a clear concentration
enhancement. Even though we still observe the accumulation
of DNA toward the middle of each image, the density and
brightness are much lower. An important observation is that
we can readily view the motion of individual DNA molecules
as marked with the colored circles. The two lowest DNA
molecules (orange and green circles) move up toward the
“stationary region” at the same speed, while the motion of
the DNA molecules in the middle of the image (red circles)
is slower, and it is imperceptible at the top of the image
(blue circles). The length of observed single DNA in images
is around 2 µm, similar to previous literature reports in the
same scale of nanochannels.42,43 The sidewall porosity among
the 1D channels provides the communication between the
channels as shown in the “wave” in Figure 5d and in the
correlated motion within different channels shown in Figure
5g-k.
With current fabrication approaches, we are able to readily
fabricate porous nanochannels with smaller cross sections.
Smaller silica nanoparticles (e.g., 15 nm) can be employed
to fabricate porous nanochannels with widths less than 100
nm.44 Even with 50 nm silica nanoparticles, we were able
to fabricate porous nanochannels with widths of ∼100 nm
as shown in Figure 6. The images in parts a and b of Figure
6 show as-prepared nanochannels (fabricated with positive
photoresist, which enables narrower photoresist widths) with
a 130 nm wide and 400 nm high cross section and a 800 nm
pitch. DNA transport behavior was investigated in these
samples as well. The DNA was less likely to enter these
smaller channels and moved more slowly. With only
capillary forces, the DNA molecules are transported only a
short distance into the channels as compared to the above
500 nm wide channels. The apparent length of the concata-
mer DNA arrays (15-25 µm) is shorter in these channels
as shown in Figure 6c compared to Figure 5c. The dimen-
sions of the channels clearly affected the static and dynamical
behaviors of the DNA.43 The real-time microscope image
further confirmed the streaks of DNA confined in nanochan-
nels (Figure 6d). The fluorescence from individual DNA
molecules in images captured in a video frame is rather faint
(see Figure S6 of the Supporting Information) as a result of
bleaching, the rapid motion of the DNA, and instrumental
limitations. With the successful fabrication of porous
nanochannels over a wide size range, further efforts are
underway to investigate the detailed transport mechanisms,
effect of channels size and DNA types, and DNA solution
environments such as ionic strength.43
The fabrication of hierarchically enclosed nanostructures
in colloidal particle films opens new directions in the
preparation of unique, biomimetic porous enclosed struc-
tures with many potential applications. Various enclosed
structures, which include channels, isolated air cavities,
continuous networks, and multiple-layered channels buried
in colloidal particle films, were successfully demonstrated
with this simple approach. The resultant enclosed struc-
tures have excellent mechanical strength, high environ-
mental stability, good uniformity, and proven ability to
transport aqueous solutions and large organic macromol-
ecules such as DNA. The multiple layered channels are
biomimetic structures akin to the cellular organization of
trees, animal skin, feathers, and hair. They are an attractive
platform for the study of natural processes and can be
engineered for many biomimetic applications including
fluidics, biomolecular separation and sensing, detection,
and catalysis.
Acknowledgment. Partial support for this work was
provided by the NSF (IIS 0515684 and CTS 0404124) and
by the Army Research Office under a subcontract from
Redondo Optics, Inc.
Supporting Information Available: Experimental details
and supporting SEM images. This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(1) Velev, O. D.; Tessier, P. M.; Lenhoff, A. M.; Kaler, E. W. Nature
1999, 401, 548.
(2) Jiang, P.; J. Bertone, F.; Colvin, V. A. Science 2001, 291, 453–457.
(3) Lahav, M.; Sehayek, T.; Vaskevich, A.; Rubinstein, I. Angew. Chem.,
Int. Ed. 2003, 42, 5576–5579.
(4) Hong, S. H.; Moon, J. H.; Lim, J.-M.; Kim., S.-H.; Yang, S.-M.
Langmuir 2005, 21, 10416–10421.
(5) Moon, J. H.; Kim, S.; Yi, G.-R.; Lee, Y.-H.; Yang, S.-M. Langmuir
2004, 20, 2033–2035.
(6) Yi, G.-R.; Manaharan, V. N.; Klein, S.; Brzezinska, K. R.; Pine, D. J.;
Lange, F. F.; Yang, S.-M. AdV. Mater. 2002, 14, 1137–1140.
(7) Iskandar, F.; Mikrajuddin; Okuyama, K. Nano Lett. 2002, 2, 389–
392.
(8) Li, Q.; Lewis, J. A. AdV. Mater. 2003, 15, 1639–1643.
(9) Zaumseil, J.; Eitl, M. A.; Hsu, J. E. P.; Acharya, B. R.; Baldwin, K. W.;
Loo, L.-Y.; Roger, J. A. Nano Lett. 2003, 3, 1223–1227.
(10) Jeon, S.; Menard, E.; Park, J.-U.; Maria, J.; Meitl, M.; Zaumseil, J.;
Roger, J. A. AdV. Mater. 2004, 16, 1369–1373.
(11) Lee, C.; Yang, E. H.; Myung, N. V.; George, T. A. Nano Lett. 2003,
3, 1339–1340.
(12) Brien, J. O., II; Bisong, P.; Ista, L. K.; Rabinovich, E. M.; Garcia,
A. L.; Sibbett, S. S.; Lopez, G. P.; Brueck, S. R. J. J. Vac. Sci. Technol.,
B 2003, 21, 2941–2945.
(13) Cao, H.; Yu, Z.; Wang, J.; Tegenfeldt, J. O.; Austin, R. H.; Wu, W.;
Chou, S. Y. Appl. Phys. Lett. 2002, 81, 174–176.
(14) Xia, D.; Brueck, S. R. J. J. Vac. Sci. Technol., B 2005, 23, 2694–
2699.
(15) Gates, B.; Xu, Q.; Stewart, M.; Ryan, D.; Willson, C. G.; Whitesides,
G. M. Chem. ReV. 2005, 105, 1171–1196.
(16) Wang, D.; Mohwald, H. J. Mater. Chem. 2004, 14, 459–468.
(17) Xia, D.; Biswas, A.; Li, D.; Brueck, S. R. J. AdV. Mater. 2004, 16,
1427–1432.
Nano Lett., Vol. 8, No. 6, 2008 1617

(18) Xia, D.; Brueck, S. R. J. Nano Lett. 2004, 4, 1295–1299.
(19) Brueck, S. R. J. Proc. IEEE 2005, 93, 1704–1721.
(20) Campbell, M.; Sharp, D. N.; Harrison, M. T.; Denning, R. G.;
Turberfield, A. J. Nature 2000, 404, 53–56.
(21) Zhang, S.; Fan, W.; Frauenglass, A.; Minhas, B.; Malloy, K. J.; Brueck,
S. R. J. Phys. ReV. Lett. 2005, 94, 037402.
(22) Zhang, S.; Fan, W.; Panoiu, N. C.; Malloy, K. J.; Osgood, R. M.;
Brueck, S. R. J. Phys. ReV. Lett. 2005, 95, 137404.
(23) Yan, Q.; Zhou, Z.; Zhao, X. S.; Chua, S. J. AdV. Mater. 2005, 17,
1917–1920.
(24) Vekris, E.; Kitaev, V.; von Freymann, G.; Perovic, D. D.; Aitchison,
J. S.; Ozin, G. A. AdV. Mater. 2005, 17, 1269–1272.
(25) Xia, D.; Li, D.; Luo, Y.; Brueck, S. R. J. AdV. Mater. 2006, 18, 930–
933.
(26) Xia, D.; Li, D.; Ku, Z.; Luo, Y.; Brueck, S. R. J. Langmuir 2007, 23,
5377–5385.
(27) McLellan, J. M.; Geissler, M.; Xia, Y. J. Am. Chem. Soc. 2004, 126,
10830–10831.
(28) Wong, S.; Kitaev, V.; Ozin, G. A. J. Am. Chem. Soc. 2003, 125,
15589–15598.
(29) Schweigert, I. V.; Lehtinen, K. E. J.; Carrier, M. J.; Zachariah, M. R.
Phys. ReV. B 2002, 65, 235410.
(30) Sacks, M. D.; Seng, T. Y. J. Am. Ceram. Soc. 1984, 67, 532–537.
(31) Miguez, H.; Meseguer, F.; Lopez, C.; Blanco, A.; Moya, J. S.;
Requena, J.; Mifsud, A.; Fornes, V. AdV. Mater. 1998, 10, 480–483.
(32) Morita, M.; Koga, T.; Otsuka, H.; Takahara, A. Langmuir 2005, 21,
911–918.
(33) Zhai, L.; Berg, M. C.; Cebeci, F. C.; Kim, Y.; Miwid, J. M.; Rubner,
M. F.; Cohen, R. E. Nano Lett. 2006, 6, 1213–1217.
(34) Davis, J. A.; Inglis, D. W.; Morton, K. J.; Lawrence, D. A.; Huang,
L. R.; Chou, S. Y.; Sturm, J. C.; Austin, R. H. Proc. Natl. Acad. Sci.
U. S. A. 2006, 103, 14779–14784.
(35) Hungate, R. E. Plant Physiol. D 1934, 9, 783–794.
(36) Zhao, Y.; Cao, X.; Zhang, L. J. Am. Chem. Soc. 2007, 129, 764–765.
(37) Han, J.; Graighead, H. G. Science 2000, 288, 1026–1029.
(38) Dukkipati, V. R.; Pang, S. W. Appl. Phys. Lett. 2007, 90, 083901.
(39) Huh, D.; Mills, K. L.; Zhu, X.; Burns, M. A.; Thouless, M. D.;
Takayama, S. Nat. Mater. 2007, 6, 424–428.
(40) Liang, X.; Morton, K. J.; Austin, R. H.; Chou, S. Y. Nano Lett. 2007,
7, 3774–3780.
(41) Mijatovic, D.; Eijkel, J. C. T.; van den Berg, A. Lab Chip 2005, 5,
492–500.
(42) Reisner, W.; Morton, K. J.; Riegn, R.; Wang, Y. M.; Yu, Z.; Rosen,
M.; Sturm, J. C.; Chou, S. Y.; Frey, E.; Austin, R. H. Phys. ReV. Lett.
2005, 94, 196101.
(43) Reisner, W.; Beech, J. P.; Larsen, N. B.; Flyvbjerg, H.; Kristensen,
A.; Tegenfieldt, J. O. Phys. ReV. Lett. 2007, 99, 058302.
(44) Xia, D.; Brueck, S. R. J. J. Vac. Sci. Technol., B 2004, 22, 3415–3420.
NL080190S
1618 Nano Lett., Vol. 8, No. 6, 2008