NANO
LETTERS
Formation of λ-DNA’s in Parallel- and
Crossed-Line Arrays by Molecular
Combing and Scanning-Probe
Lithography
2006
Vol. 6, No. 7
1334-1338
Minjung Shin,† Chilwoo Kwon,† Seong Kyu Kim,*,† Hyung Jin Kim,‡
Yonghan Roh,‡ Byungyou Hong,‡ Jong Bae Park,†,§ and Haeseong Lee*,§
Department of Chemistry and School of Information and Communication Engineering,
Sungkyunkwan UniVersity, Suwon 440-746, Korea, and Korea Basic Research Institute,
Jeonju 561-756, Korea
Received January 24, 2006; Revised Manuscript Received May 16, 2006
ABSTRACT
With the combination of a molecular combing technique and scanning-probe lithographic patterning, λ-DNA’s were stretched and aligned to
form line array structures on patterned organic monolayer surfaces. The pattern was generated by anodizing a silicon surface using scanningprobe lithography to implant a polar organic layer in the middle of a nonpolar layer. The molecule in the polar layer, (aminopropyl)triethoxysilane
(APS), has a −NH3+ terminal group, which interacts strongly with phosphate backbone of DNA and provides a site for selective attachment
of DNA. When parallel lines of APS were patterned, followed by combing along the lines, a single DNA was attached from the very top of each
line and stretched along the line all the way to the bottom. The DNA−APS interaction was strong enough to withstand the second combing
applied perpendicular to the first one. Thereby, the crossed-line array of DNA’s was formed on the crossed-line array pattern of APS on a
silicon substrate.
For many applications of nanoscale science, positioning
component materials into a desired location on a solid surface
is a crucial technique. Selective binding of the component
materials can be achieved on a functionalized surface that
is patterned with two different chemical layers whose
interaction strengths with the materials are markedly different. Alignment of nanowires or nanotubes along one direction
can be induced if the surface pattern is designed with parallel
lines or if an external force by flow, gravity, gas blowing,
or electric field is exerted along one direction. The reports
on selective binding and alignment of carbon nanotubes,1,2,6
Si, InP and GaP nanowires,3,4 and V2O5 nanowires5,6 are good
examples.
In the case of aligning long DNA’s into a single direction
on surface, techniques generally termed “molecular
combing”7-17 are used widely. They use a surface tensional
force at the air-water interface of a DNA solution translating
on a grafted substrate. When the speed of translation and the
substrate-DNA interaction are balanced, DNA’s are stretched
along the direction normal to the air-water interface. The
translation of the air-water interface is induced by sliding
* Corresponding authors. E-mail: [email protected]; [email protected].
† Department of Chemistry, SKKU.
‡ School of Information and Communication Engineering, SKKU.
§ Korea Basic Research Institute.
10.1021/nl060160u CCC: $33.50
Published on Web 06/14/2006
© 2006 American Chemical Society
a droplet of DNA solution between a pair of glasses,7-13 or
by pulling the substrate out of DNA solution.14-16 In modified
versions of the combing, methods of sucking up17-19 the
droplet using a pipet, blowing it with gas,20,21 or evaporating
it on a spinning substrate22 have been demonstrated. In our
recent (unpublished) work, sliding the DNA droplet down
the tilted substrate produced a similar effect.
The combing method can maximize its capability when
the surface is prepatterned. By specifying the location and
the direction of DNA attachment determined by prepatterning, the substrate can serve as templates for wires in
nanoscale devices or can assist in high-resolution genomic
studies. There are a few examples in which DNA’s were
stretched and aligned across the patterned lines.23,24 However,
if DNA’s are stretched and aligned along parallel lines, the
direction and interval of the attached DNA’s would be
predetermined and formed much more uniformly. We
showed in our recent report25 that constructing such a
structure is realistic; λ-DNA’s were successfully stretched
and aligned by combing along parallel lines of a polar organic
layer produced by self-assembly and the scanning-probe
lithography (SPL) technique. In this letter, we will present
some new characteristics in the parallel-line structures and
will show that the same methodology can be used to produce
crossed-line arrays of long DNA’s.
Figure 1. Summary of DNA alignment in this work. Parts a-e are the procedure and part f depicts the cross-section of part c.
Our strategy is summarized in Figure 1. First, a monolayer
of octadecyltrichlorosilane (OTS) was formed on a cleaned
surface of silicon substrate, cut by 5 × 5 mm2 (Figure 1a).
Then, using an atomic force microscope (AFM, SII SPA400),
SPL patterning by anodizing the silicon surface was performed. The lithographic pattern was an array of either
parallel lines or crossed lines (like a checkerboard pattern).
Lines of oxidized silicon were formed whose width and
height were typically about 150 and 3 nm, respectively.
During this process, the OTS molecules in the oxidized
region must have been stripped out26 (Figure 1b). Then, the
sample was immersed into 0.1 mM (aminopropyl)triethoxysilane (APS) solution of anhydrous toluene. APS is expected
to replace the stripped-out OTS in the oxidized region (Figure
1c) and provides a site for selective attachment of DNA
through a Coulombic interaction between the -NH3+ moiety
of APS and the phosphate backbone of DNA.
For the combing, we used a custom-made dipping/pulling
machine whose translation is directed parallel to the side
edges and the patterned lines of the silicon substrate. The
patterned silicon substrate was translated into a 9-mm-innerdiameter tube that contained 500 µL of λ-DNA solution
(BioLabs, 48.5 kbp, pH 8.0) until the whole substrate was
immersed. The concentration of the λ-DNA solution was 2
µg/mL for the parallel-line array or 0.2 µg/mL for the
crossed-line array. After 3 min, the substrate was translated
out at the same speed. The speed of the substrate translation
was 32 µm/s, corresponding to 50 µm/s for water meniscus.
Nano Lett., Vol. 6, No. 7, 2006
After the combing, the sample was washed with water to
remove solid salts and weakly bound λ-DNA’s. On the first
combing, λ-DNA’s were stretched and attached only on the
APS/SiOx lines (Figure 1d). When the crossed-line array of
λ-DNA’s was to be formed, the sample after the first
combing was dried over hours and the second combing was
applied perpendicular to the first one (Figure 1e). Other
experimental details are found in ref 25 and also in the
Supporting Information.
Figure 2 is an example of fully aligned λ-DNA’s along
the parallel lines of APS/SiOx. Thirteen APS/SiOx parallel
lines 15 µm in length, separated by 1 µm, were manufactured
by SPL, and the combing was applied from the top to the
bottom of Figure 2. Here, we define the direction of combing
as that of the receding water meniscus when the substrate is
pulled out of the solution. Initially, we obtained the tappingmode AFM image of the patterned area with 512 × 512
pixels2 in a 18 × 18 µm2 window (not shown in this letter).
However, because the thickness of DNA is much smaller
than the pixel size, the presence of DNA was not identifiable.
Then, the tapping-mode AFM scans were conducted with
512 × 512 pixels2 in 4 × 4 µm2 windows. Five areas in the
left side of the whole pattern were scanned with the finer
resolution, and their images were then overlapped to
reconstruct the whole picture for the left four lines (the left
image of Figure 2). The black broken lines in Figure 2
indicate the boundaries of the overlaps. The yellow bold lines
represent mostly the outgrown oxidized silicon. The white
1335
Figure 2. Attachment of stretched λ-DNA’s along the parallel APS/SiOx
lines. The left image shows only the left 4 lines of the 13 fabricated lines.
The left image was reconstructed by overlapping five high-resolution images
of the AFM tapping-mode scan. The black broken lines indicate the boundaries of the overlaps. The three images on the right are magnified from
the blocked area of the left image. Single λ-DNA is found on every APS/SiOx
line. More magnified images are available in the Supporting Information.
1336
objects are unwashed solid salts that were originated from
the buffer solution. When Figure 2 is examined closely, a
single λ-DNA is found on every APS/SiOx line. (See the
magnified images on the right side of Figure 2. More
magnified images are available in the Supporting Information.) In fact, a single DNA was attached from the very top
of each line and stretched along the line all the way to the
bottom. In contrast, no DNA is found in the OTS area.
Because the average length of a stretched λ-DNA is estimated (following 2.4 kbp/µm)9,10 to be 21 µm, the tailing part
of it should appear outside the 15 µm APS/SiOx line. This
was observed in our recent work,25 where tailing parts of
λ-DNA’s were clung to the bottom of APS/SiOx lines and
extended into the OTS region. The length of the patterned
lines in the previous work was 20 µm, and that of the
λ-DNA’s extended into the OTS region ranged between 0 and
6 µm, from which we estimated the anchoring positions of the
λ-DNA’s to be within 5 µm from the top of APS/SiOx lines.
In contrast to our previous work, the tailing parts of
λ-DNA’s at the bottom of APS/SiOx lines were not observed
in Figure 2. This is probably because they were cut out by
the AFM tip during the tapping-mode scan. The λ-DNA’s
may be cut at the boundary of APS/SiOx and OTS regions
because an excess force by the tip may be exerted due to
the abrupt height change during the scan. In fact, even in
the previous work25 the cuts were found at several boundaries. The condition of the tapping-mode AFM scan under
which λ-DNA’s are cut at the boundary is currently unknown
and beyond our control.
The absence of λ-DNA’s extended above the top of APS/
SiOx lines may have a different implication from the cut at
the boundary. In most AFM images taken in our laboratory,
λ-DNA’s extended above the top of the lines were rarely
observed while the tailing parts often appeared at the bottom.
It suggests that DNA extremities have higher binding
strengths, in agreement with Bensimon’s model,7-10 and the
anchoring position of λ-DNA was the top of each APS/SiOx
line. This can be explained as follows.
While the substrate is in DNA solution at the incubation
stage of the combing, the water meniscus is in the OTS
environment above the top of APS/SiOx lines. The hydrophobic or van der Waals (vdW) interaction between OTS
and DNA near the substrate surface is so weak that the
tensional force of the moving water meniscus can sweep the
DNA’s off the substrate surface. As the water meniscus
moves down, DNA’s accumulate at the water meniscus.
(There are many evidences of the higher concentration of
DNA at moving meniscus on combing. For example, see
ref 27.) At the moment the water meniscus reaches the top
of APS/SiOx lines, DNA is anchored by the strong Coulombic interaction with APS. Because at this point the surface
concentration of DNA at the top is much higher than that at
the rest part, the anchoring is preferentially made at the top
of the APS/SiOx lines. We think that this effect was small
in our previous work25 because a much lower concentration
of DNA was used for the combing.
After the water meniscus passes by the top of the lines,
the interfacial force stretches DNA’s along the APS lines,
Nano Lett., Vol. 6, No. 7, 2006
while weakly bound DNA’s in the OTS region are washed
away. The speed of combing should be determined such that
the interfacial force is greater than the vdW interaction
between OTS and DNA but less than the Coulombic
interaction between APS and DNA, while DNA’s are in
solution. After the combing is over, water molecules
evaporate away from the substrate and the vdW force
increases as the DNA’s come closer to the substrate surface.
Any DNA’s in the OTS region that were not detached during
the process of combing should be rinsed away before the
vdW force increases too much. After the sample is completely dried, it is very difficult to remove DNA’s from both
APS and OTS regions.
Another remarkable observation in Figure 2 is that a single
DNA is attached in each patterned line. A single DNA in
fact runs through the middle of each patterned line. This
observation is similar to the work by Hong and co-workers,6
where V2O5 nanowires and carbon nanotubes were aligned
along lines of micrometer-thick polar organic layers. Following their idea and experimental evidences, we suggest
that the binding efficiency of APS with DNA across the line
is shaped as a convex lens; the density of APS and the
orientation of the -NH3+ moiety are most favorable in the
middle of a patterned line and become unfavorable towards
the line edges. So even if the width of the patterned APS/
SiOx line was 150 nm, the effective width for binding DNA
was much narrower. And this effect plus the electrostatic
repulsion between multiple DNA’s may have resulted in the
formation of a single DNA per patterned line.
There have been several successful attempts to construct
a two-dimensional (2D) network structure of DNA by
applying combing methods twice,21,28 the second one perpendicular to the first one. However, none have produced
the 2D structure on a crossed-line patterned surface. When
the combing is applied to the crossed-line pattern, there are
a couple of difficulties to be overcome. First, there is a high
possibility of the crossover after the first combing, because
perpendicular lines were drawn in addition to the parallel
lines in the pattern. DNA’s will attach not only along the
parallel lines but also across the perpendicular lines. The
difference is in the binding energies of the two; the DNA’s
aligned along the parallel lines must have much higher
binding energies than the DNA’s hung across the perpendicular lines. The weaker-bound crossover DNA’s can be
washed away during the combing or by extra rinsing. The
difference in the binding energy increases as the width of
the patterned line decreases for a fixed line interval. In our
case, the average width of the patterned lines was about 150
nm while their interval was 1 µm. However, because the
effective area of APS for binding DNA’s is much narrower,
the chance of crossover will be lower than the ratio dictates.
Another aspect to consider for the formation of the
crossed-line structure is whether the structure of the first
combing would be preserved in the process of the second
combing. The image of Figure 2 remained unchanged after
several harsh wash-outs. Therefore, we expect that the
structure after the first combing can be preserved during the
second combing. However, Figure 2 shows one drawback;
Nano Lett., Vol. 6, No. 7, 2006
Figure 3. Attachment of stretched λ-DNA’s on the crossed-line
array of APS/SiOx. (a) Tapping-mode AFM (topographic) image
of the whole patterned area. (b) A higher scan-resolution image.
The attached λ-DNA’s were stained with gold nanoparticles and
appeared as the brightest objects. The two arrows marked with I
and II indicate the directions of the first and the second combing,
respectively. More higher scan-resolution images are available in
the Supporting Information.
there appears to be a large amount of solid salts that reside
in the APS area. The presence of the solid salts would
decrease the binding strength of APS with DNA on the
second combing. The amount of solid salts is high at high
DNA concentration. (For examples, compare Figure 2 with
the figures in ref 25.) This is probably because the ion
concentration in the DNA solution is locally high in the
vicinity of a DNA chain, so more ions would be attracted to
APS when the combing was carried out at a high concentration of DNA solution.
To form a crossed-line array of λ-DNA’s, a 9 × 9 array
of 11-µm-long APS/SiOx lines was patterned. Then, the
combing with λ-DNA was carried out twice; the second one
perpendicular to the first one. When the same concentration
of λ-DNA solution as that in Figure 2 was used, some
λ-DNA’s, after the first combing, were undesirably attached
across the perpendicular lines in addition to the desirable
parallel attachment. Besides, the high concentration produced
too many solid salts in the APS area. Therefore, we reduced
the concentration of λ-DNA solution to 0.2 µg/mL. This time,
1337
to observe the structure in a large window at once, we made
the attached DNA’s thicker by staining with positively
charged gold nanoparticles, synthesized with aniline as a
reducing agent. (See ref 25 or the Supporting Information.)
Figure 3 is the tapping-mode AFM images of the result. The
brightest objects are gold nanoparticles attached to the DNA
backbone. The gold nanoparticles bound very weakly on the
OTS and APS layers were washed away by rinsing, whereas
they attached strongly to the negatively charged DNA
backbone. In Figure 3a, the directions of the first and the
second combings are indicated by the arrows marked with I
and II, respectively. Figure 3b is a finer-scan image with
512 × 512 pixels2 in a 4 × 4 µm2 window, where the brighter
objects forming lines indicate the DNA backbone.
Because the DNA’s were not fully stained with the gold
nanoparticles, we may not give a clear picture of the
structure. However, to the best of our analysis with more
finer-scan images (see the Supporting Information), we judge
the structure as follows. In the direction of the first combing,
each line of 2, 4, 6, 8, and 9 was fully covered with a λ-DNA
from the top to the bottom. On lines 3, 5, and 7, the coverage
within a few micrometers from the top was not clear. Line
1 was covered with multiple DNA’s, one of which was
displaced from the line when the second combing or washing
was carried out. In the direction of the second combing, each
line of A, B, D, E, F, G, and I was fully covered with a
λ-DNA from the top to the bottom. On lines C and H, the
coverage within a few micrometers from the top was not
clear.
The overall quality of the first combing in Figure 3 was
not as good as that in Figure 2. Other than the DNA coverage
on the APS/SiOx lines, DNA’s in the OTS region were not
completely removed. Over each top of lines 2, 4, and 6, a
DNA was extended into the OTS region. The shorter
patterned lines and the lower DNA concentration to avoid
the crossover and the solid salts formation could be reasons
for the poor performance of the first combing. We think that
the translation speed and the incubation time on combing
should yet be optimized for the crossed-line array. Nonetheless, Figure 3 implies that the decent structure of crossedline array of λ-DNA is possible once optimizing the combing
condition is further developed.
In conclusion, we produced a perfect line array structure
of λ-DNA’s with the combination of the organic monolayer
assembly, SPL, and molecular combing. For the crossedline array structure, the optimal condition for combing is
stricter, but the example structure we produced in this work
suggests the methodology is a promising technique for
constructing similar structures for two-dimensional nanoscale
devices.
Acknowledgment. This work was supported by the
National R&D Project for Nano Science and Technology
1338
(Contract No. M10214000110-02B1500-01800) of MOST
and the SRC program (Center for Nanotubes and Nanostructured Composites) of KOSEF.
Supporting Information Available: Experimental details, magnified details of Figure 2, and higher resolution
images of Figure 3. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Liu, J.; Casavant, M. J.; Cox, M.; Walters, D. A. Boul, P.; Lu, W.;
Rimberg, A. J.; Smith, K. A.; Colbert, D. T.; Smalley, R. E. Chem.
Phys. Lett. 1999, 303, 125-129.
(2) Choi, K. H.; Bourgoin, J. P.; Auvray, S. Esteve, D.; Duesberg, G.
S.; Roth, S.; Burghard, M. Surf. Sci. 2000, 462, 195-202.
(3) Huang, Y.; Duan, X.; Wei, Q.; Lieber, C. M. Science 2001, 291,
630-633.
(4) Whang, D.; Jin, S.; Lieber, C. M. Jpn. J. Appl. Phys. 2004, 43, 44654470.
(5) Myung, S.; Lee, M.; Kim, G. T.; Ha, J. S.; Hong, S. AdV. Mater.
2005, 17, 2361.
(6) Myung, S.; Im, J.; Huang, L.; Rao, S. G.; Kim, T.; Lee, D. J.; Hong,
S. J. Phys. Chem. B 2006, 110, 10217-10219.
(7) Bensimon, A.; Simon, A.; Chiffaudel, A.; Croquette, V.; Heslot, F.;
Bensimon, D. Science 1994, 265, 2096-2098.
(8) Bensimon, D.; Simon, A. J.; Croquette V.; Bensimon, A. Phys. ReV.
Lett. 1995, 74, 4754-4757.
(9) Weier, H.-U. G.; Wang, M.; Mullikin, J. C.; Zhu, Y.; Cheng, J.-F.;
Greulich, K. M.; Bensimon, A.; Gray, J. W. Hum. Mol. Genet. 1995,
4, 1903-1910.
(10) Allemand, J.-F.; Bensimon, D.; Jullien, L.; Bensimon, A.; Croquette,
V. Biophys. J. 1997, 73, 2064-2070.
(11) Ouyang, Z.; Hu, J.; Chen, S.; Sun, J.; Li, M. J. Vac. Sci. Technol.,
B 1997, 15, 1385-1387.
(12) Hu, J.; Wang, M.; Weier, H.-U. G.; Frantz, P.; Kolbe, W.; Oglitree,
D. F.; Salmeron, M. Langmuir 1996, 12, 1697-1700.
(13) Woolley, A. T.; Kelly, R. T. Nano Lett. 2001, 1, 345-348.
(14) Michalet, X.; Eklong, R.; Fougerousse, F.; Rousseaux, S.; Schurra,
C.; Hornigold, N.; van Slegtenhorst, M.; Wolfe, J. Povey, S.;
Beckmann, J. S.; Bensimon, A. Science 1997, 277 1518-1523.
(15) Gueroui, Z.; Place, C.; Freyssingeas, E.; Berge, B. Proc. Natl. Acad.
Sci. U.S.A. 2002, 99, 6005-6010.
(16) Petit, C. A. P.; Carbeck, J. D. Nano Lett. 2003, 3, 1141-1146.
(17) Otobe, K.; Ohtani, T. Nucleic Acids Res. 2001, 29, e109.
(18) Nakao, H.; Hayashi, H.; Yoshino, T.; Sugiyama, S.; Otobe, K.;
Ohtani, T. Nano Lett. 2002, 2, 475-479.
(19) Nakao, H.; Shiigi, H.; Yamamoto, Y.; Tokonami, S.; Nagaoka, T.;
Sugiyama, S.; Ohtani, T. Nano Lett. 2003, 3, 1391-1394.
(20) Li, J.; Bai, C.; Wang, C.; Zhu, C.; Lin, Z.; Li, Q.; Cao, E. Nucleic
Acids Res. 1998, 26, 4786-4786.
(21) Deng, Z.; Mao, C. Nano Lett. 2003, 3, 1545-1548.
(22) Yokota, H.; Sunwoo, J.; Sarikaya, M.; van den Engh, G.; Aebersold,
R. Anal. Chem. 1997, 71, 4418-4422.
(23) Nyamjav, D.; Ivanisevic, A. AdV. Mater. 2003, 15, 1805-1809.
(24) Klein, D. C. G.; Gurevich, L.; Janssen, J. W.; Kouwenhoven, L. P.;
Carbeck, J. D.; Sohn, L. L. Appl. Phys. Lett. 2001, 78, 2396-2398.
(25) Shin, M.; Kim, T.; Kwon, C.; Kim, S. K.; Park, J. B.; Lee, H. Jpn.
J. Appl. Phys. 2006, 45, 2076-2081.
(26) Sugimura, H.; Hanji, T.; Hayahi, K.; Takai, O. Ultramiscoscopy 2002,
91, 221-226.
(27) Liu, Y.-Y.; Wang, P.-Y.; Dou, S.-X.; Wang, W.-C.; Xie, P.; Yin,
H.-W.; Zhang, X.-D.; Xi, X. G. J. Chem. Phys. 2004, 121, 43024309.
(28) Hu, J.; Zhang, Y.; Gao, H.; Li, M.; Hartmann, U. Nano Lett. 2002,
2, 55-57.
NL060160U
Nano Lett., Vol. 6, No. 7, 2006