REVIEW ARTICLE

Fluid flow in carbon nanotubes

and nanopipes
Nanoscale carbon tubes and pipes can be readily fabricated using self-assembly techniques and
they have useful electrical, optical and mechanical properties. The transport of liquids along their
central pores is now of considerable interest both for testing classical theories of fluid flow at the
nanoscale and for potential nanofluidic device applications. In this review we consider evidence for
novel fluid flow in carbon nanotubes and pipes that approaches frictionless transport. Methods for
controlling such flow and for creating functional device architectures are described and possible
applications are discussed.

M. WHITBY AND N. QUIRKE
Chemistry Department, Imperial College, South Kensington,
London SW7 2AZ, UK.
e-mail: n.quirke@imperial.ac.uk
Understanding and controlling the flow of liquids at the nanoscale
is currently a subject of great interest. More than 100 papers
describing work in nanofluidics have been published in the past
five years. Recent experimental measurement of fluid transport
through channels with widths of a few nanometres has confirmed
predictions derived from computer simulations that such flow
occurs at a much greater rate than expected. This discovery has
significant implications, both for our understanding of how fluids
behave at very small length scales and for the design of nanofluidic
devices. Here we review selected papers with an emphasis on fluid
flow through the central pores of carbon tubes.
The practical investigation of fluid flow in nanoscale channels has
been facilitated by the availability of tubular carbon structures with
open central pore diameters in the range of one to several hundred
nanometres. Carbon offers a number of attractive features for the
fabrication of nanofluidic devices including: the existence of several
well studied processes for self-assembly of tubular structures; useful
electronic properties for sensing and signalling; biocompatibility,
ready chemical modification to allow functionalization; and low-
friction transport thanks to graphitic surfaces.
Two types of nanoscale carbon tubes have been used
experimentally, each with distinct properties and methods
of synthesis. The first are carbon nanotubes, both single and
multiwalled, with the fullerene molecular structure reported by
Ijima in 1991 (ref. 1). The second are carbon nanopipes produced
by chemical vapour deposition (CVD) of amorphous carbon in
alumina templates with a honeycomb pore morphology (Fig. 1).
Throughout this review we reserve the term ‘nanotube’ to refer
to true molecular carbon structures. The term ‘nanopipe’ will
be used to describe carbon pipes produced in templates using
CVD. The latter are generally larger and composed mainly of
amorphous rather than well ordered graphitic carbon. The
properties and fabrication of both types of carbon tube, and their
arrays when assembled as nanoporous membranes, are discussed
in later sections.
Our understanding of the static properties of liquids in
macroscopic capillaries is based on the nineteenth century work
of Laplace, Poisson and Young2. Steady-state flow of simple
incompressible fluids in a channel width 2h, driven for example
by gravity ρg or a pressure gradient dP/dy, can be described by
the Navier Stokes equation3. The solution for the velocity in the
direction of flow, y, as a function of the distance from the wall,
z, has a parabolic profile (Fig. 2) given by:
Uy(z) = (ρg/2η)[(δ + h)2 – z2]
where η is the viscosity and δ, the slip length, is the distance into
the wall at which the velocity extrapolates to zero. Conventionally,
the slip length is assumed to be zero. However, if it is not then
we are said to have slip boundary conditions at the wall, which
for very small channels can significantly enhance the fluid flow.
Integrating the velocity profile over the cross-sectional area of the
pipe gives the Hagen-Poiseuille law for the flux through a pipe.
This law states that the flow rate of a fluid passing through a tube is
directly proportional to the pressure difference between the tube
ends and to the fourth power of the tube’s internal radius. The
flow rate is inversely proportional to the tube length and to the
viscosity of the fluid.
The dynamics of capillary filling were elucidated by Washburn4,5
using the Hagen-Poiseuille law with the driving force for flow given
by the Laplace equation for the pressure difference across the
invading liquid meniscus. Thus the penetration length L at time t of
a fluid in a capillary of radius R is given by:
L2 = (Rγ/2η)t
with γ being the liquid/vapour surface tension and η the shear
viscosity of the invading liquid. In deriving the Washburn equation
it is assumed that there is no fluid motion at the wall (that is, stick
boundary conditions). The Washburn equation is very successful
on time scales sufficient to establish viscous flow but breaks
down for short times. A more general solution can be found for

nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology 87

REVIEW ARTICLE
b a
20 μm
1 μm
500 nm
Figure 1 Scanning electron microscope images of carbon nanopipes produced by the authors using standard chemical vapour deposition. a, Nanopipes partially
released from an anodic aluminium oxide template following sonication in NaOH. b, Cross section of intact carbon coated membrane. c, Higher magnification view of
individual aligned carbon pipes. d, Surface of carbon membrane showing open pores (diameter ~160 nm). The growth procedures are described in the section on “Flow
through nanoscale channels including carbon nanopipes”.
non-steady viscous incompressible flow (the Bosanquet equation6),
which approaches the Washburn equation at long times.
For nanoscale capillaries the flow behaviour is dominated not
by bulk properties (γ, η) but by the interaction of the fluid with
the capillary walls. In this regard it is important to include the
role of surface friction due to molecular corrugation, especially
for graphitic surfaces where the high surface density of atoms
leads to a very low corrugation. A convenient parameter here is
the Maxwell coefficient7 α, which represents the fraction of the
molecular collisions with the capillary walls that undergo diffuse
scattering and/or trapping–desorption, the remainder being
specularly reflected (no energy loss). The Maxwell coefficient
has been shown to provide a useful description of molecular
friction8 for a variety of materials9. However, although it is clear
+h
Uy(z)
z adsorption, for fluids such as water considerable uncertainty has
y
–h
Figure 2 Parabolic velocity profile Uy(z) for Poiseuille flow in a capillary of width 2h.
88
2 μm
that the smaller the value of α, the greater the slip length, a direct
relationship between the two only exists at low density10.
SIMULATION AND THEORY OF NANOSCALE FLOW
A central question in nanofluidics concerns the extent to which
the classical equations describing the way fluids interact with
nanomaterials hold at the nanoscale. Molecular simulation is
ideally suited to shed light on this problem and has shown that
static properties such as surface tensions and contact angles
of simple fluids obey classical relations down to almost single
nanometre dimensions11. Steady-state Poiseuille flows maintain a
parabolic velocity profile down to widths of several nanometres12.
Deviations from this behaviour can typically be linked to density
inhomogeneities in the confining walls.
In performing simulations relevant to experimental work on
flow in nanopores, a major challenge has been to choose realistic
potentials between the molecules in the fluid and in the pore walls.
Although it has proved possible to be reasonably confident about,
say, determining alkane/carbon potentials by fitting to heats of
arisen. For example, reports of the measured contact angle of water on
a graphitic surface differ by as much as a factor of two13. The choice of
intermolecular potential determines whether a model nanotube can
be filled with water14,15 or not. However recent work has shown how it
may be possible to calibrate solvent–nanotube potentials using Raman
shifts in solution16. The uncertainty in water potentials notwithstanding,
some general features of nanoscale flow have been determined from
molecular dynamics using simple fluids relevant to the experimental
work. For example Sokhan et al.8,17 were the first to show that fluids
flowing through (slit) carbon nanopores experience very low surface
friction, that is very small Maxwell coefficients (~0.01) or equivalently
large slip boundary conditions, whereas pores made from other
materials show higher values and stick boundary conditions.
nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology

Time = 60 ps
Figure 3 Imbibition of decane in a (7,7) carbon nanotube (diameter = 0.951 nm). Left: Simulated imbibition, showing a snapshot of the fluid after 60 ps50. Right: plot of the
imbibition velocity obtained from the analytic model using parameters derived from nanoscale simulation, the upper (blue) line shows results for a carbon nanotube, the
bottom (black) line, a non-carbon tube50.
Subsequent theoretical work shows that in a cylindrical geometry,
decreasing the radius can reduce the Maxwell coefficients by orders
of magnitude compared with a slit (~0.000002 for decane in a
(7,7) nanotube, diameter 0.75 nm) and that carbon nanotubes are
predicted to be essentially frictionless pipes18,19. Skoulidas et al.20
compared transport diffusivities for carbon nanotubes with zeolites
using molecular dynamics and found exceptionally high transport
rates for the nanotubes due to the smooth carbon surface. We
should expect therefore that crystalline carbon channels (as opposed
perhaps to amorphous surfaces) show significantly increased flow.
Another consequence is the extremely rapid imbibition of wetting
fluids (decane is predicted to fill (7,7) nanotubes, at 800 m s–1; ref. 18)
observed in non-equilibrium molecular dynamics. Note that the
same model predicts external wetting to be much slower by factors
of roughly 30 for a (13,13) tube with diameter 1.765 nm. This has
implications for nanotube arrays, where flow may occur along either
the external or internal surfaces of the tubes.
From a Bosanquet-like equation (where we approximate the
surface friction with an expression involving the Maxwell coefficient
and the surface tension and the driving force is represented by the
wall-wetting liquid tension), it is possible to obtain a solution for the
imbibition velocity and penetration length L at all times18. There is an
L ∝ t dependence for short times tending to t1/2 at long times (as in the
Washburn equation), the timescale for which depends on the value
of the Maxwell coefficient. Figure 3 shows how the imbibition speed
for decane varies as a function of time with ultrafast uptake for t<μs
falling to cm s–1 at longer times.
The situation is more complex for hydrogen bonding fluids such as
water. In very narrow tubes the water molecules form a one-dimensional
chain with linking hydrogen bonds. At equilibrium, conduction
through such tubes occurs in bursts14 and this behaviour is attributed
to density fluctuations at the openings of the tube.
FLOW THROUGH NANOSCALE CHANNELS INCLUDING CARBON NANOPIPES
Initial indications of rapid fluid flow through very small channels
in an experimental system were published by Pfahler et al. in
1990 (ref. 21). Working with rectangular channels etched in a
silicon substrate, they found an n-propanol flux approximately
three times greater than expected in the smallest channels (height
0.8 μm and width 100 μm). Following up on this observation in 2002,
Cheng and Gordanio22 reported measurements for Newtonian
fluids flowing under pressure through nanoscale channels produced
nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology
REVIEW ARTICLE
900
800
700
600
V m s–1
500
400
300
200
100
0
0 2 4 6 8 10 12
log t (ps)
lithographically. The height of the channels varied from 2.7 μm down
to 40 nm. The results show a marked and progressive departure from
Poiseuille flow as the channel height falls below 200 nm. This effect
was observed most strongly for hexadecane and is also evident for
decane and for silicone oil. Slip lengths were estimated at 25 to 30 nm,
~9 nm and ~14 nm respectively. The effect was not seen for water at
the length scales investigated.
Several groups have recently reported results from experiments
involving flow of liquids and gases through membranes with
nanoscale pores composed of carbon. Two principal methods have
been used to construct suitable membranes. The first involves CVD
of carbon in an anodic aluminium oxide (AAO) template. The second
is based on production of polymer composites containing an aligned
array of carbon nanotubes spanning the thickness of the membrane.
Both techniques allow liquid flow through the central pores of carbon
nanotubes and nanopipes to be studied without the complication of
external wetting.
The first method for creating carbon nanomembranes takes
advantage of the self-assembly of highly-ordered nanopores arranged
in a dense hexagonal pattern in alumina when aluminium foil is
etched in acid23,24. Using this approach, monodisperse-sized pores with
densities as high as 1011 cm–2 are possible. To line the nanopores, the
AAO template is placed in a furnace and a hydrocarbon gas is flowed
for several hours, whereupon carbon is deposited on the hot surface of
the alumina25. Carbon pipes with an outside diameter as small as 10 nm
have been formed in this way although larger 100 to 200 nm nanopipes
are more usual. The length of these pipes is typically 50 to 100 μm.
It is important to appreciate that, as synthesized, the typically
20-nm-thick walls of carbon pipes produced by the AAO templating
method will be at least partly amorphous. Che et al.26 reported electron
diffraction studies in a transmission electron microscope (TEM),
which produced a pattern characteristic of non-aligned graphite
fragments. There are reports that high-temperature annealing can
improve the order of the tube walls through graphitization27,28,
which may enhance transport properties. Miller et al.25 confirmed
electro-osmotic flow through these carbon nanomembranes and
reported a linear relationship between electro-osmotic flow velocity
and applied current density.
FLUID FLOW THROUGH CARBON NANOTUBES
Hinds et al.29 published the first account of experiments involving fluid
flow through a membrane composed of true molecular nanotubes. This
89
REVIEW ARTICLE
a closed by end caps or plugged by residual iron catalyst. To overcome
50 μm
b and alkanes being measured after a fixed time period. Their results
50 μm
c by Holt et al. was to use fabrication techniques adapted from the
Figure 4 Carbon nanotube arrays and membranes. a, An as-grown, dense, multiwalled
carbon nanotube array produced with an Fe-catalysed chemical vapour deposition
process. b, The cleaved edge of the nanotube-polystyrene membrane after exposure
to H2O plasma oxidation. The polystyrene matrix is slightly removed to contrast the
alignment of the nanotubes across the membrane. c, Schematic of the target membrane
structure. With a polymer embedded between the nanotubes, a viable membrane
structure can be readily produced, with the pore being the rigid inner-tube diameter of
the nanotube. Reproduced with permission from ref. 29. Copyright (2004) AAAS.
group used dense, well aligned arrays of multiwalled carbon nanotubes
(MWNT) grown using CVD on a flat quartz substrate (Fig. 4). One
advantage of this approach is that the pore diameters are much smaller
(4.3 ± 2.3 nm) than with the AAO template method. Another is that
the crystallinity of the inner carbon nanotube walls is greater, with
consequent benefits for low friction fluid flow.
As synthesized MWNT arrays are permeable not only through the
central pores of the nanotubes, but also extensively through the gaps
between the individual tubes. Furthermore, many nanotubes will be
90
these problems, a polystyrene solution was spin coated onto the array
to fill the gaps. Then plasma etching was used to remove the top layer
of the composite and to open the nanotube central pores. The pore
density was estimated to be 6 (±3) × 1010 cm–2.
Using gas flow apparatus, the MWNT composite array showed
a linear relationship between flow rate and pressure drop across
the membrane. The authors also studied diffusion of aqueous ionic
species (Ru(NH3)63+ diameter = 0.55 nm) and noted that the enhanced
diffusion coefficient was near the bulk aqueous-solution diffusion for
the Ru cation. They infer only limited interaction between the ion and
the nanotube tip and walls.
Majumder et al.30 published the results of further experiments by
the same group on fluid transport through the MWNT composite
membranes. A simple pressure-driven flow apparatus was used, with
the accumulated mass of transported fluids including water, ethanol
demonstrate dramatically enhanced flow through ~7-nm-diameter
nanotube cores compared with conventional fluid flow theory.
Observed flow rates were four to five orders of magnitude greater
than predicted by hydrodynamics based on macroscale behaviour
(Table 1). The authors conclude that the implied slip lengths
(3 to 70 μm), which are much greater than the tube diameter, are
consistent with a nearly frictionless interface.
In the studies reviewed so far, increasingly interesting fluid flow
results were found at progressively smaller length scales. In 2006,
Holt et al.31 achieved a milestone in measuring water flow through
the central pores of double walled carbon nanotubes (DWNT). These
tubes had inner diameters less than 2 nm with defect-free graphitic
walls that are expected from simulation (see previous section) to
present a low friction surface to transported fluids.
The practical challenges involved in studying transport through
very small carbon nanotubes are formidable. The approach used
semiconductor industry. A dense array of carbon nanotubes were first
grown on a silicon substrate onto which metal catalyst particles had
been deposited. The spaces between the tubes were then completely
filled with silicon nitride from low-pressure chemical vapour. Finally
the closed ends of the tubes and catalyst particles were removed by
a series of etching and ion milling steps. The resulting membranes
had a thickness in the range of 2.0 to 3.0 μm and pore densities
≤0.25 × 1012 cm–2. Pore diameters, calibrated by passage of gold
nanoparticles, were in the range of 1.3 to 2.0 nm.
Gas flow and water flow measurements were taken in flow cells
sealed with o-rings. Five hydrocarbon and eight non-hydrocarbon
gases were tested to determine flow rates and to demonstrate
molecular weight selectivity compared with helium. Water flow was
pressure-driven at 0.82 atm and measured by following the level of
the meniscus in a feed tube. The results for both gas and liquid show
dramatic enhancements over flux rates predicted with continuum flow
models. Gas flow rates were between 16 and 120 times that expected
according to the Knudsen diffusion model in which fluid molecule–
wall collisions dominate the flow. Water flow rates were 560 to 8,400
times greater than those calculated according to the Hagen–Poiseuille
equation. Minimum slip lengths are estimated in the range of 140 to
1,400 nm. To express these findings in a practical context, the nanotube
membranes showed flow rates several orders of magnitude greater
than those of conventional membranes, despite having pore sizes an
order of magnitude smaller. The authors of these studies point out the
significance of this finding for separation applications.
MICROSCOPE STUDIES OF FLUID FLOW IN NANOTUBES
Permeation experiments allow quantitative measurement of
fluid transport through nanoscale carbon pores. A more qualitative
nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology

approach is to observe fluid interaction using optical and electron
microscopes. It should be noted that the structures studied are
generally large (tube diameter > 200 nm). The results are therefore
unlikely to show the more unusual surface-dominated flow
behaviour expected in true nanoscale channels.
In 2004, Rossi et al.32 reported a series of observations of liquid/
carbon-pipe interactions made in an environmental scanning
electron microscope (ESEM). The advantage of using an ESEM
is that in situ experiments involving fluids can be undertaken.
Furthermore, by carefully varying the pressure in the sample
chamber and by controlling the temperature using a Peltier cooling
stage, it is possible to cause condensation and evaporation to occur
at will in the sample. Using this method, the team captured the
behaviour of liquid menisci inside 200 to 300 nm diameter CVD
grown carbon pipes (Fig. 5).
A key result was that water condensed on the internal walls of
the carbon tubes in preference to the external steel surface of the
stage. The authors concluded that the disordered walls of the AAO
template-grown tubes are hydrophilic and they measured contact
angles with water between 5 and 20°. They also reported a slight
deformation of tubes containing plugs of water, indicating negative
pressure inside the channel due to tensile capillary forces. Carbon
tubes of this type may be useful for guiding aqueous fluids to and
from specific locations at the nanoscale and for collecting attolitre
and picolitre samples.
Also in 2004, Kim et al.33 reported work on liquid interactions
with similar carbon pipes and plotted filling length against time.
Working with ≥200 nm nanopipes, they found a good fit with
x(t) = At1/2 (the Washburn equation) as expected for the long (ms)
observation times.
A year later, the same group published a follow-on paper34
reporting on imbibition experiments that involved introducing
fluorescent nanoparticles (diameter ~50 nm) into a variety of
liquids. The suspended particles were taken up by carbon pipes
with 500-nm diameters through a combination of capillary action
and evaporation forces. The introduced particles could be observed
moving along the tubes as their fluorescence was visible through
the tube walls. The authors speculate that this technique may be
useful for observing the interactions of biological macromolecules
in the vacuum environment of an electron microscope.
CONTROLLING LIQUID FLOW THROUGH NANOTUBES
In the previously discussed study by Miller et al.25, electrochemical
derivitization was used to change the surface functionalization
of carbon nanopipes. Using this technique, both the magnitude
and the direction of electro-osmotic flow could be modified.
Simple application of a potential across the membrane can also
act as an immediate transport switch as reported by several
researchers29,35,36.
Table 1 Comparison of observed flow velocity with predictions for pressure-driven flow through aligned MWNT membranes30. These results indicate liquid flow
rates four to five orders of magnitude faster than would be predicted from conventional fluid-flow theory. The slip lengths are calculated from observed data. *
Units: cm3 per cm2 at minimum pressure. †Flow velocities in cm s–1 at 1 bar.
Liquid Initial permeability* Observed flow velocity†
Water 0.58 25
1.01 43.9
0.72 9.5
Ethanol 0.35 4.5
Iso-Propanol 0.088 1.12
Hexane 0.44 5.6
Decane 0.053 0.67
nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology
REVIEW ARTICLE
Miller and Martin35 describe a more sophisticated redox
modulation of ionic transport, which they used to switch both the
rate and direction of electro-osmotic flow. Working again with
carbon nanopipes, they coated the carbon surfaces with the redox
polymer poly(vinylferrocene) (PVFc). A variable potential was used
to totally oxidize, totally reduce or to set the redox state of the PVFc
to an intermediate value. Once set, the membrane would remember
its state until altered. The authors compare their approach to the field-
effect concept in semiconductors.
Similar findings are reported by Majumder et al.37 using their
finer MWNT composite membranes (described earlier). Instead
of functionalizing the entire length of the nanopores, the authors
selectively modified just the tip region of the nanotubes. The tips were
exposed by plasma etching and functionalized using a carbodiimide
mediated coupling. They showed that different chemical end groups
attached near the entrance to the channels could selectively modulate
ionic flux.
Babu et al.38 report a technique for guiding the entry of water
into carbon nanopipes which were selectively modified by the
electrodeposition of polypyrrole on their tips. The authors conclude
that this may be useful for constructing devices capable of controlling
liquid flows at the nanoscale.
Molecular dynamics simulations of water in the pores of single-
walled carbon nanotubes39 show that an electric field can also modulate
filling through dipole interactions. This effect might also be exploited
to achieve selective control of liquid flow at the nanoscale and may
play a part in natural biological pore transport mechanisms.
CHALLENGES IN BUILDING NANOFLUIDIC DEVICES
Many of the papers reviewed above describe novel phenomena and
methods that will be directly applicable to the design and fabrication
of nanofluidic devices with practical application in medical
diagnosis, sensing and materials processing. Before these ambitions
can be achieved, a considerable number of practical hurdles will
need to be overcome. In this section some approaches are considered
that may have useful architectural or procedural application in the
implementation of a viable nanofluidics technology.
Microfluidics is a rapidly advancing field with emerging
analytical and synthetic applications40. Cao et al.41 report
one of the few experimental investigations of the problem of
microfluidic to nanofluidic interfacing. Liquid samples for
characterization or processing using nanoscale components are
likely to start out as macroscale droplets. These must be guided
into progressively smaller channels for ultimate delivery to the
nanoscale device components. A number of potential problems
arise. One is channel blocking due to large macromolecules
or insoluble debris. This can already be a limiting hazard with
microscale lab-on-a-chip systems. Work is currently taking
place to understand the dynamics of particulate transport at the
Expected flow velocity† Slip length (µm)
0.00057 54
0.00057 68
0.00015 39
0.00014 28
0.00077 13
0.00052 9.5
0.00017 3.4
91
REVIEW ARTICLE
a b

~15º

~5º Water

Vapour 200 nm
200 nm

c d

200 nm 200 nm

e f
Vapour

Water

200 nm 50 nm

Figure 5 ESEM images of the dynamic behaviour of a water plug in a carbon nanopipe. a–e, The meniscus shape changes when the stage temperature is constant but the
vapour pressure of the water in the chamber is changed: a, 5.5 Torr, b, 5.8 Torr, c, 6.0 Torr, d, 5.8 Torr and e, 5.7 Torr (where the meniscus returns to the shape seen in a). The
asymmetrical shape of the meniscus, especially the complex shape of the meniscus on the right side in a and e is a result of the difference in the vapour pressure caused by
the open left end and closed right end of the tube. Estimated contact angles between the meniscus and the tube wall are indicated at two locations. f, TEM image showing a
similar plug shape in a closed carbon nanotube under pressure. Reproduced with permission from ref. 32. Copyright (2004) ACS.

nanoscale and to find low friction coatings or channel materials
that help to reduce blocking42.
Another problem is more fundamental: the large size of polymers,
including biologically relevant molecules such as DNA, which are
often tangled and tightly folded in vivo. As Cao et al.41 point out, a
typical DNA molecule from a virus has a length of 100 to 200 kilobases
and will form a random coil with a radius of some 700 nm in aqueous
solution at 20 °C. This is several times greater than the pore diameter
of even large carbon pipes and two orders of magnitude greater than
the diameter of a single walled carbon nanotube such as might be
functionalized to detect the presence of specific base sequences43. DNA
molecules will fit into the central channel of even small nanotubes,

92 nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology

 REVIEW ARTICLE
but only when unravelled and fed into the pore opening lengthwise. a
The entropic barrier to achieve this from the disordered state is very
high and therefore such long molecules are normally excluded.
To overcome this problem of interfacing the microscale to
the nanoscale, Cao et al. used optical lithography to fabricate an
array of microchannels that form a gradient from wide to narrow
fluid passages (Fig. 6a)41. They used a novel modified form of
diffraction gradient lithography involving a photosensitive
blocking mask resist on a silicon wafer substrate. The technique
10 μm
is inherently parallel and much faster and more efficient than
using electron-beam lithography. The end result is a massive
array of microposts, with a continuous reduction in the gaps,
which form fluidic channels as the chip is traversed from one
side to the other. b
To test the device, long DNA strands stained with a fluorescent
dye were introduced on the microscale side (right-hand side of
Fig. 6b). Diffusion of the molecules was then observed under
ultraviolet light and captured on video. Still frames from the
recording show individual DNA molecules straightening out and
moving through the interface in an extended configuration towards
the nanoscale region (left-hand side of Fig. 6b). The authors report
transport of the stretched DNA molecules with significantly greater 10 μm
efficiency compared with random diffusion through comparable
nanoscale pores in the absence of any gradient interface.
Melechko et al.36 working at Oak Ridge National laboratory in
Tennessee report a general method for creating patterned arrays of
silica nanopipes precisely positioned over pores in a silicon nitride
membrane on a silicon substrate. The method requires costly Figure 6 Interfacing microfluidics with nanofluidics. a, Optical image of fluidic
equipment and top-down fabrication techniques, but allows a high channels defined by microposts on a chip (after photoresist development). A
degree of control over device architecture. Critically, precise control of continuous reduction of the gaps (light areas) between microposts occurs in the
nanopipe location is achieved by deterministic positioning of catalyst gradient zone. b, Integrated CCD video images of DNA flow through the channels.
particles for CVD growth. The right-hand side of the image shows partially stretched long DNA molecules
Bau et al.44 describe another method for constructing a nanofluidic passing through the micropost array and the gradient zone and continuously entering
device comprising a single carbon nanopipe (diameter 250 nm) the nanochannels on the left-hand side, becoming fully stretched in the process.
connecting two fluid reservoirs. Dielectrophoresis was used to orient Reproduced with permission from ref. 41. Copyright (2002) AIP.
and manipulate the nanopipe into position prior to depositing a
dividing wall on top. These general approaches enable prototyping
of specific functional devices for research and testing. Once effective
architectures have been developed, less expensive methods can be across the tubes and also by functionalizing tube entrances34,35. The
investigated for large-scale manufacture. problem of the transition from micro to nanoflows is being tackled,
Within the past two years, a number of papers have started to for example, by the creation of patterned surfaces, which guide
appear reporting experimentally realized nanofluidic devices. For biomolecules through progressively smaller apertures36. The more
example Flashbart and co-workers45 describe fabrication using contact control that can be engineered into nanopore arrays, the greater the
printing of a microfluidic array with nanochannel interconnects. This prospect of using them in applications requiring high selectivity and
allows their device to take advantage of nanocapillary effects such fast throughput.
as in situ size-based separation, enhanced mixing and molecular It is significant that the size range of nanopores discussed
concentration. Han et al.46 have used surface micromachining in this review is essentially the size range of many important
on silicon nitride, silica and polysilicon in combination with biological entities (antibodies, viruses). Larger molecules such as
micromoulded polydimethylsiloxane to fabricate flexible nanofluidic DNA can be uncoiled36 to fit. Thus carbon nanopipes are potential
device architectures. They have demonstrated attolitre electrokinetic conduits, collimators, sensors, encapsulators and probes for medical
injection using this technique. Finally, Wang et al.47 have created a applications. Clearly there are still many challenges ahead before
nanoscale preconcentration device using standard photolithography such devices become viable including: controlling the mechanical
and etching methods, which has achieved factors in the range of strength of nanoelements in contact with cells and tissue; methods for
106 to 108. They exploit the electrokinetic trapping effect found in the assembly of huge numbers of very small components; fouling of
nanofluidic filters. the nanopipes and surfaces; management of defects in components;
fluid/biomolecule/disease marker input and output; and managing
CONCLUSIONS AND FUTURE DIRECTIONS the information flow from large arrays of nanoscale sensors to the
outside world.
Molecular dynamics simulations16,17,48 have suggested that carbon In a biological context, it is worth noting the vital importance of
nanotubes have very low surface friction with respect to fluid flow. nanoscale pores in natural systems, particularly the channels in cell
This is now being confirmed by direct experiment on timescales membranes involved in ion and water transport. Aquaporin proteins,
which show fast flow corresponding to large slip lengths18,28,25. Water for the elucidation of which Peter Agre earned the Nobel Chemistry
appears to wet carbon nanopipes with flow governed by the Washburn Prize in 2003 (ref. 49), are one example. As is often the case, science
equation, but not for subnanosecond timescales. Control of liquid is only now starting to investigate and appreciate long-established
uptake has been achieved by applying a potential difference27,32,33 natural mechanisms.

nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology 93

REVIEW ARTICLE
doi: 10.1038/nnano.2006.175
References
1. Iijima, S., Helical microtubules of graphitic carbon. Nature 354, 56–58 (1991).
2. Rowlinson, J. S. & Widom, B. Molecular Theory of Capillarity. Vol. Section 1.3. (Oxford Univ. Press,
Oxford, 1982).
3. Palmer, S. B. & Rogalski, M. S. Advanced University Physics (Gordon and Breach, New York, 1996).
4. Washburn, E. W. The dynamics of capillary flow. Phys. Rev. 17, 273–278 (1921).
5. de Gennes, P.-G., Brochard-Wyart, F. & Quere, D. Capillarity and Wetting Phenomena.
(Springer, Berlin, 2004).
6. Bosanquet, C. H., The flow of liquids into capillary tubes. Phil. Mag. 6, 525–531 (1923).
7. Maxwell, J. C. On stresses in rarified gases arising from inequalities of temperature. Philos. Trans. R.
Soc. 170, 231–256 (1879).
8. Sokhan, V. P., Nicholson, D. & Quirke, N. Fluid flow in nanopores: an examination of hydrodynamic
boundary conditions. J. Chem. Phys. 115, 3878–3887 (2001).
9. Quirke, N. (ed.) Adsorption and Transport at the Nanoscale (CRC Press, Boca Raton,
Florida, 2005).
10. Lauga, E., Brenner, M. & Stone, H. The no-slip boundary condition: a review in The Handbook Of
Experimental Fluid Dynamics (Springer, 2005).
11. Powell, C., Fenwick, N., Bresme, F. & Quirke, N. Wetting of nanoparticles and nanoparticle arrays.
Colloid. Surface. A 206, 241–251 (2002).
12. Travis, K. P., Todd, B. D. & Evans, D. J. Departure from Navier–Stokes hydrodynamics in confined
liquids. Phys. Rev. E 55, 4288–4295 (1997).
13. Werder, T., Walther, J. H., Jaffe, R. L., Halicioglu, T. & Koumoutsakos, P. On the water-carbon
interaction for use in molecular dynamics simulations of graphite and carbon nanotubes. J. Phys.
Chem. B 107, 1345–1352 (2003).
14. Hummer, G., J. C. Rasaiah, and J. P. Noworyta, Water conduction through the hydrophobic channel
of a carbon nanotube. Nature 414, 188–190 (2001).
15. Waghe, A., Rasaiah, J. C. & Hummer, G. Filling and emptying kinetics of carbon nanotubes in water.
J. Chem. Phys. 117, 10789–10795 (2002).
16. Longhurst, M. & Quirke, N. Environmental effects on the radial breathing modes of carbon
nanotubes in water. J. Chem. Phys. 124, 234708 (2006).
17. Sokhan, V. P., Nicholson, D. & Quirke N. Fluid flow in nanopores: accurate boundary conditions for
carbon nanotubes. J. Chem. Phys. 117, 8531–8539 (2002).
18. Supple, S. & Quirke, N. Molecular dynamics of transient oil flows in nanopores I: Imbibition speeds
for single wall carbon nanotubes. J. Chem. Phys. 121, 8571–8579 (2004).
19. Supple, S. & Quirke, N. Rapid imbibition of fluids in carbon nanotubes. Phys. Rev. Lett. 90,
214501 (2003).
20. Skoulidas, A. I. et al. Rapid transport of gases in carbon nanotubes. Phys. Rev. Lett. 89,
185901 (2002).
21. Pfahler, J., Harley, J, Bau, H. & Zemel, J. Liquid transport in micron and submicron channels. Sensor.
Actuat. A 22, 431–434 (1990).
22. Cheng, J. T. & Giordano N. Fluid flow through nanometer-scale channels. Phys. Rev. E 65,
214501 (2002).
23. Masuda, H., Hasegwa, F. & Ono, S. Self-ordering of cell arrangement of anodic porous alumina
formed in sulfuric acid solution. J. Electrochem. Soc. 144, L127–L130 (1997).
24. Routkevitch, D., Bigioni, T., Moskovits, M. & Xu, J. M. Electrochemical fabrication of CdS nanowire
arrays in porous anodic aluminum oxide templates. J. Phys. Chem. 100, 14037–14047 (1996).
25. Miller, S. A., Young, V. Y. & Martin, C. R. Electroosmotic flow in template-prepared carbon nanotube
membranes. J. Am. Chem. Soc. 123, 12335–12342 (2001).
26. Che, G., Lakshmi, B. B., Fisher, E. R. & Martin, C. R. Carbon nanotubule membranes for
electrochemical energy storage and production. Nature 393, 346–349 (1998).
94
27. Kyotani, T., Tsai, L. F. & Tomita, A. Preparation of ultrafine carbon tubes in nanochannels of an
anodic aluminum oxide film. Chem. Mater. 8, 2109–2113 (2006).
28. Mattia, D., et al. Effect of graphitization on the wettability and electrical conductivity of CVD-
carbon nanotubes and films. J. Phys. Chem. B 110, 9850–9855 (2006).
29. Hinds, B. J. et al. Aligned multiwalled carbon nanotube membranes. Science 303, 62–65 (2004).
30. Majumder, M., Chopra, N., Andrews, R. & Hinds, B. J. Nanoscale hydrodynamics: enhanced flow in
carbon nanotubes. Nature 438, 44 (2005).
31. Holt, J. K. et al. Fast mass transport through sub-2-nanometer carbon nanotubes. Science 312,
1034–1037 (2006).
32. Rossi, M. P. et al. Environmental scanning electron microscopy study of water in carbon nanopipes.
Nano Lett. 4, 989–993 (2004).
33. Kim, B. M., Sinha, S. & Bau, H. H. Optical microscope study of liquid transport in carbon
nanotubes. Nano Lett. 4, 2203–2208 (2004).
34. Kim, B. M., Murray, T. & Bau, H. H. The fabrication of integrated carbon pipes with sub-micron
diameters. Nanotechnology 16, 1317–1320 (2005).
35. Miller, S. A. & Martin, C. R. Redox modulation of electroosmotic flow in a carbon nanotube
membrane. J. Am. Chem. Soc. 126, 6226–6227 (2004).
36. Melechko, A. V., McKnight, T. E., Guillorn, M. A. & Merkulov, V. I. Vertically aligned carbon
nanofibers as sacrificial templates for nanofluidic structures. Appl. Phys. Lett. 82, 976–978 (2003).
37. Majumder, M., Chopra, N. & Hinds, B. J. Effect of tip functionalization on transport through
vertically oriented carbon nanotube membranes. J. Am. Chem. Soc. 127, 9062–9070 (2005).
38. Babu, S., Ndungu, P., Bradley, J.-C., Rossi, M. P. & Gogotsi, Y. Guiding water into carbon nanopipes
with the aid of bipolar electrochemistry. Microfluid. Nanofluid. 1, 284–288 (2005).
39. Vaitheeswaran, S., Rasaiah, J. C. & Hummer, G. Electric field and temperature effects on water in the
narrow nonpolar pores of carbon nanotubes. J. Chem. Phys. 121, 7955–7965 (2004).
40. de Mello, A. J. Control and detection of chemical reactions in microfluidic systems. Nature 442,
394–402 (2006).
41. Cao, H., Tegenfeldt, J. O., Austin, R. H. & Chou, S. Y. Gradient nanostructures for interfacing
microfluidics and nanofluidics. Appl. Phys. Lett. 81, 3058–3060 (2002).
42. Kenis, P. J. A. & Stroock, A. D. Materials for micro- and nanofluidics. MRS Bull. 31,
87–94 (2006).
43. Heller, D. A. et al. Optical detection of DNA conformational polymorphism on single-walled carbon
nanotubes. Science 311, 508–511 (2006).
44. Bau, H. H., Sinha, S., Kim, B. & Riegelman, M. Fabrication of nanofluidic devices and the study of
fluid transport through them in Proc. SPIE. Vol. 5592 (eds Lai, W. Y., Pau, S., Lopez, O. D.) 201–213
(SPIE, Bellington, Washington, 2005).
45. Flachsbart, B. R., et al. Design and fabrication of a multilayered polymer microfluidic chip with
nanofluidic interconnects via adhesive contact printing. Lab Chip 6, 667–674 (2006).
46. Han, A. P., de Rooij, N. F. & Staufer, U. Design and fabrication of nanofluidic devices by surface
micromachining. Nanotechnology 17, 2498–2503 (2006).
47. Wang, Y. C., Stevens, A. L. & Han, J. Y. Million-fold preconcentration of proteins and peptides by
nanofluidic filter. Anal. Chem. 77, 4293–4299 (2005).
48. Skoulidas, A. I., Ackerman, D. M., Johnson, J. K. & Sholl, D. S. Rapid transport of gases in carbon
nanotubes. Phys. Rev. Lett. 89, 185901 (2002).
49. Agre, P. Aquaporin water channels (Nobel lecture). Angew. Chem. Int. Edn 43, 4278–4290 (2004).
50. Supple, S. Molecular dynamics simulation of capillary flow at the nanoscale. PhD Thesis, Imperial
College London (2005).
Acknowledgements
M. Whitby acknowledges support from the EPSRC through a Doctoral Training Award and thanks
F. Barclay and T. Cotter at RGB Research for helpful discussions. The authors are members of the EU
network Inside POReS.
nature nanotechnology | VOL 2 | FEBRUARY 2007 | www.nature.com/naturenanotechnology