Size-exclusion chromatography using self-organized nanopores in anodic porous alumina

Article abstract of DOI:10.1063/1.1629379

The band separation of DNA molecules was attempted using an anodic alumina as a separation matrix. This chip worked as an SEC-type device, and the channel was never clogged.

Full text of DOI:10.1063/1.1629379

Size-exclusion chromatography using self-organized nanopores in anodic porous
alumina
Tohru Sano, Noriyuki Iguchi, Kazuhiro Iida, Toshitsugu Sakamoto, Masakazu Baba, and Hisao Kawaura
Citation: Applied Physics Letters 83, 4438 (2003); doi: 10.1063/1.1629379
View online: http://dx.doi.org/10.1063/1.1629379
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/83/21?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
MultiLevel Self Organization Process For A Parallel Fabrication Of Aligned Metal Structures In Microelectrode
Gaps Using DNA And Metal Nanoparticles
AIP Conf. Proc. 725, 9 (2004); 10.1063/1.1805373
Fabrication of nanohole array on Si using self-organized porous alumina mask
J. Vac. Sci. Technol. B 19, 1901 (2001); 10.1116/1.1403442
Self-repair of ordered pattern of nanometer dimensions based on self-compensation properties of anodic porous
alumina
Appl. Phys. Lett. 78, 826 (2001); 10.1063/1.1344575
Low-energy electron beam focusing in self-organized porous alumina vacuum windows
Appl. Phys. Lett. 76, 3635 (2000); 10.1063/1.126731
Field emitter arrays based on natural self-organized porous anodic alumina
J. Vac. Sci. Technol. B 16, 1222 (1998); 10.1116/1.589988
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.64.11.153 On: Wed, 08 Oct 2014 05:55:26

APPLIED PHYSICS LETTERS
VOLUME 83, NUMBER 21
24 NOVEMBER 2003
Size-exclusion chromatography using self-organized nanopores in anodic
porous alumina
Tohru Sano,a) Noriyuki Iguchi, Kazuhiro Iida, Toshitsugu Sakamoto, Masakazu Baba,
and Hisao Kawaura
Fundamental Research Laboratories, NEC Corporation, 34, Miyukigaoka, Tsukuba,
Ibaraki 305-8501, Japan
͑Received 20 February 2003; accepted 25 September 2003͒
We have achieved the size separation of DNA molecules on a biochip using an anodic porous
alumina as a separation matrix in electrophoresis. The bottom of the microfluid channel on the chip
was made of a porous alumina membrane that has uniform nanoscale pores on its surface. Since
smaller molecules are much more frequently trapped by nanopores than larger ones, larger
molecules elute earlier than smaller molecules. This type of separation is the same process as that
resulting from size-exclusion chromatography. When two kinds of DNA samples ͑3.2 and 0.3
kilobases͒ were applied to the biochip, they were clearly separated into two bands, with the larger
molecules ͑3.2 kilobases͒ eluting earlier. Since nanopore formation does not require any
nanolithography, the fabrication is easy and inexpensive. The matrix can be integrated with other
biomicroelectromechanical systems reactors on a micro total analysis system chip. © 2003
American Institute of Physics. ͓DOI: 10.1063/1.1629379͔
A precisely controlled two-dimensional ͑2-D͒ nanostruc-
ture is a promising separation matrix for a micro total analy-
sis system ͑␮TAS͒ chip, and it can be widely used for bio-
logical applications such as proteomics and diagnosis.
Volkmuth et al. proposed the concept of an artificial matrix.1
Turner et al. presented a nanometer-scale pillar structure to
sieve DNA molecules, and their processing techniques were
adapted from a semiconductor electronics fabrication process
that allowed precise control and good uniformity in the pillar
gap.2 Clear separation results have not been shown to date,
and there was a problem with the DNA clogging at the
boundary of the dense pillar region.2 In addition, the device
is difficult to fabricate because it requires nanolithography.
In order to solve these two problems, we propose a de-
sign for a nanostructured biochip using an anodic porous
alumina as a separation matrix in electrophoresis. This chip
does not need any filling matrices such as photopolymerized
porous resin or cross-linked polymer. The bottom of the mi-
crofluid channel was made of a porous alumina membrane
that has uniform nanoscale pores on its surface. Unlike larger
molecules, smaller molecules are frequently trapped by the
pores. Smaller molecules spend a lot of time in the nanop-
ores while larger molecules have less loss time, resulting in
larger molecules moving more rapidly in the channel ͓see
Fig. 1͑b͔͒. This type of separation is the same as effected by
size-exclusion chromatography ͑SEC͒. This structure solves
the clogging problem because large molecules, which are
bigger than the exclusion limit, are not trapped and the depth
of the separation channel is much greater than the size of the
larger molecules. Since nanopore formation is a self-
organized process and it does not require any nanolithogra-
phy, fabrication is easy. Thus, this biochip has both advan-
tages of nonclogging separation and easy fabrication.
The channels cross each other at one point on the nanochip,
hereinafter referred to as the cross area. Reservoirs A, B, C,
and D were made, and glass tubes are attached at each res-
ervoir. The longest channel, running from A to D, is 40 mm
long. We observed samples at a point that is 2 mm from the
cross area. The magnified illustration of the separation chan-
nel is shown in Fig. 1͑b͒. There are many uniform nanopores
at the bottom of the 80-␮m-wide channels, which are made
of insulator material, Al2O3 ͑alumina͒. The radius of each
pore is about 80 nm, and this is larger than the radius of
gyration (Rg) of the 0.3 kilobases ͑kb͒ DNA ͑28 nm͒ and
smaller than that of 3.2 kb DNA ͑105 nm͒.3 Because the
channel is 1 ␮m deep and the depth is narrower than the
FIG. 1. ͑a͒ Schematic illustration of the chip. The chip has four reservoirs:
A, B, C, and D. ͑b͒ Magnified view of the separation channel. The radius of
each pore was controlled to allow smaller molecules to enter the pores while
stopping larger molecules from entering. ͑c͒ SEM micrograph of the nanop-
ores.
Figure 1͑a͒ shows a schematic illustration of the chip.
a͒
Electronic mail: sano@frl.cl.nec.co.jp
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
0003-6951/2003/83(21)/4438/3/$20.00 4438 © 2003 American Institute of Physics
130.64.11.153 On: Wed, 08 Oct 2014 05:55:26

Appl. Phys. Lett., Vol. 83, No. 21, 24 November 2003
Sano et al.
4439
FIG. 3. Electropherogram of the separation. Luminescent intensity was
measured by PMT at a point 2 mm from the cross area.
flow. The 1ϫTBE containing MPC was used in the follow-
ing experiments.
We used two kinds of DNA during electrophoresis. The
first DNA was a pUC118 vector DNA digested by EcoRI
͑3162 base pairs; 3.2 kb DNA͒, and the other was a poly-
merase chain reaction product amplified from C. elegans
tpa-1 gene6 ͑300 base pairs; 0.3 kb DNA͒. Sample DNA was
labeled using a luminescent dye, YOYO-1 ͑Molecular
Probes Inc.͒. The sample DNA was loaded at reservoir B
͓Fig. 1͑a͔͒. After the channel running between reservoirs B
and C was electrokinetically filled with the sample, a sample
injection was performed, and the sample at the cross area
was injected into the separation channel running between the
cross area and reservoir D. We measured the luminescent
intensity of the labeled DNA using a photomultiplier tube
͑PMT͒ attached to the fluorescent microscope at a point 2
mm from the cross area. We also observed DNA molecules
FIG. 2. Fabrication procedures. Top-view images ͑left͒ and A–A cross-
section images ͑right͒.
Brownian motion range of each molecule, almost all of the
molecules repeatedly touched to the bottom during electro-
phoresis. The 0.3 kb DNA was more frequently trapped by
the nanopores than the 3.2 kb DNA. Figure 1͑c͒ shows a
scanning electron microscope ͑SEM͒ micrograph of the na-
nopores.
Figure 2 shows the fabrication procedure. Anodization of
Al was performed to form self-organized uniform nanopores,
as previously reported.4 We carried out a two-step reaction
after a 300-nm-thick deposition of Al on a glass slide by
sputtering. In the first reaction, the Al was anodized under a
constant 140 V in a 0.3 M H3PO4 solution at 2 °C. In the
second reaction, the anodic porous alumina was etched in a 3
wt % H3PO4 solution at 30 °C for 60 min. This anodization
of the Al formed uniform pores about 80 nm in radius on the
whole surface of the alumina membrane. Microcapillaries
were then fabricated on the porous alumina membrane. To
prevent the sample solution from leaking out of the channels,
we spin-coated a poly͑methyl methacrylate͒ ͑PMMA͒ resist
onto the anodic porous alumina. The PMMA was directly
patterned on a micrometer-scale channel design using elec-
tron beam lithography. After the patterning step, a glass
cover with four holes was adhered to the PMMA at 160 °C
and a glass tube was attached at each hole. To make electri-
cal contacts with the channels, a Pt wire was dipped into
each reservoir.
using
a cooled CCD camera and image intensifier
͑Hamamatsu photonics K.K.͒ attached to the fluorescent mi-
croscope. Electrophoresis was performed at 150 V in the
channel running between reservoirs A and D ͑37.5 V/cm͒.
The elapsed time for the homemade electrophoresis system
to completely separate the DNA molecules began with the
sample injection.
Figure 3 is the electropherogram of the separation, and
two peaks were clearly detected. The top of the first peak
was at 52 s and the second peak was at 65 s. Larger mol-
ecules retain a much deeper dye than smaller molecules
when we use samples of equal concentration. Since the in-
tensity of the first peak was stronger than the second one, the
first peak was thought to be 3.2 kb DNA and the second one
was to be 0.3 kb DNA. We were able to observe the 3.2 kb
DNA as a particle, while 0.3 kb DNA was observed as clouds
in a fluorescent microscope with a 100ϫ objective lens. We
also observed the particles followed by clouds at the same
position of PMT during electrophoresis. These results re-
vealed that the first peak corresponded to the 3.2 kb DNA
and the second peak corresponded to the 0.3 kb DNA. Our
results thus show that this type of device can produce the
same results as SEC because the larger molecules move
faster than the smaller molecules in the separation channel.
Because the diameter of uniform nanopores can be precisely
controlled ranging from 5 ͑native protein scale͒ to 450 nm
͑large DNA scale͒,7 the matrix should be tuned according to
the size of each molecule.
The channels made of heterogeneous materials were
filled with a 1ϫ tris-borate-ethylenediamine tetra-acetic acid
͑TBE͒. When voltage was applied, DNA molecules near
PMMA ͑side walls͒ and glass ͑ceiling͒ surfaces moved from
cathode to anode, whereas molecules near alumina ͑bottom͒
moved in the opposite direction. After we added a 0.5 wt %
2-methacryloyloxy
ethyl
phosphorylcholine
͑MPC͒
chemical5 ͑Lipidure-HM, NOF Co.͒ into the TBE, all mol-
ecules moved in one direction from cathode to anode under
electrophoretic migration. We assume that the MPC coating,
left to stand for over 30 min, formed homogeneous surfaces
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
on the channels and reduced the effect of electric osmotic
We have demonstrated band separation of DNA mol-
130.64.11.153 On: Wed, 08 Oct 2014 05:55:26

4440
Appl. Phys. Lett., Vol. 83, No. 21, 24 November 2003
Sano et al.
ecules using an anodic porous alumina as a separation ma-
trix. This chip worked as an SEC-type device, and the chan-
nel was never clogged. Since formation of nanopores does
not require any nanolithography, the fabrication is easy and
inexpensive. The matrix can be easily integrated with other
biomicroelectromechanical systems reactors on a ␮TAS chip
because it does not require any filling material. The inte-
grated ␮TAS chip will meet developing demands of biologi-
cal applications such as proteomics and diagnosis.
to thank to Dr. Jun-ichi Sone and Dr. Toshio Baba for their
continuous encouragement.
1 W. D. Volkmuth and R. H. Austin, Nature ͑London͒ 358, 600 ͑1992͒.
2 S. W. Turner, A. M. Perez, A. Lopez, and H. G. Craighead, J. Vac. Sci.
Technol. B 16, 3835 ͑1998͒.
3 P. G. Righetti, Electrophoresis in Analytical Biotechnology ͑CRC Press,
Boca Raton, FL, 1996͒.
4 H. Masuda and K. Fukuda, Science 268, 1466 ͑1995͒.
5 K. Ishihara, H. Oshida, Y. Endo, T. Ueda, A. Watanabe, and N. Nakaba-
yashi, J. Biomed. Mater. Res. 26, 1543 ͑1992͒.
6 T. Sano, Y. Tabuse, K. Nishiwaki, and J. Miwa, J. Mol. Biol. 251, 477
͑1995͒.
We would like to thank Hiroko Someya and Toshie
Miyamoto for their technical assistance. We would also like
7 H. Masuda, K. Yada, and A. Osaka, Jpn. J. Appl. Phys. 37, L1340 ͑1998͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
130.64.11.153 On: Wed, 08 Oct 2014 05:55:26