|
(1) Non-Equilibrium Transport
of CH4 and O2
|
(a) Setup of System
The figure on the
right shows a snapshot from the simulation of O2 transport
through an opened, hydrogen-terminated, (10,0) single-walled nanotube.
Graphite sheets are set up at the right and left ends of the gas
reservoir. The graphite sheet at the interface of the gas reservoir and
the nanotube is open.
|
(b) Evolution
It is found from
the figure on the left that the rate of transport of the O2
molecules entering the nanotube from the gas reservoir does not depend
on the helical symmetry of the nanotube, but does strongly depend on
its pore size. Up to 50 ps, the molecular densities in the various
nanotubes increase at almost the same rate.
In this study, the
transport of the gas molecules down nanotubes that are about 100
Å for 400 ps can be classified into 3 stages.
In the first
stage, gas molecules enter the nanotube and move down its axis. To
enter the nanotube, a molecule must have a larger net repulsive force
between the gas molecule and the other gas molecules in the gas
reservoir than a net repulsive force between the gas molecule and
nanotube tip atoms, or have enough kinetic energy to allow decelerated
molecules to keep moving into the nanotube. This stage is dominated by
non-equilibrium transport.
When the molecule
reaches the nanotube end closest to posterior region, the molecule is
affected by strong attractive van der Waals interactions with the
nanotube atoms. As the diameter of the nanotube decreases, this
attractive interaction increases because more nanotube tip atoms can
get close to the molecule. The gas molecule that cannot overcome the
attractive force changes its direction of motion and moves backwards
into the nanotube.
The second stage
of transport is dominated by chaotic flow, as some molecules in the
nanotubes move towards the posterior region, and some molecules move
towards the gas reservoir. The molecules moving forward and backward
disturb the movements of the molecules that are just entering the
nanotube by colliding with them, which initiates molecular "bouncing"
motions inside the nanotube.
As the simulation
progresses, molecules continuously enter the nanotube from the gas
reservoir moving towards the posterior region. The buildup of molecular
density in the nanotube overcomes the molecular attraction with the
nanotube and forces molecules that are within the nanotube and near the
posterior region to exit. This is the third stage, which is
characterized by near steady-state transport of molecules from the gas
reservoir, into the nanotube, and out to the posterior region. In this
stage there is still some molecular motion in the opposite direction,
but it is much less dominant than it was in the second stage and the
molecular motion is consequently significantly less chaotic. The movie
on the right shows the change from the first to third stage.
|
(c) Diffusion Modes
Diffusion in pores can occur via
several different modes, including Knudsen,
molecular
(or Fickian), transition,
surface,
hydrodynamic
flow, capillary
condensation, viscous
flow, normal-mode,
single-file
diffusion, and anomalous
diffusion (subdiffusion and superdiffusion).
Figures on the left show plots of
log mean squared displacement (MSD) versus log of time for CH4
molecules that enter the (10,0), and (17,0) zigzag nanotubes at
different times after 200 ps (third stage) of simulation time. Again,
comparable behavior was seen for the similarly sized armchair (6,6),
and (10,10) nanotubes. In the case of the (10,0) nanotube, normal
diffusion line (α = 1) does not fit the data well. In contrast, the
anomalous diffusion line (α= 1.74) fits all the data very well. When we
extended this simulation beyond 250 ps by adding more molecules to the
gas reservoir and allowed the simulation to continue under
near-equilibrium conditions, the anomalous diffusion line reproducibly
fit the simulation data. |
(d) Interactions
Figures on the right clearly show
that the interaction between the arbitrary CH4 molecule and
the other CH4 molecules in the nanotubes is more repulsive,
and has more severe fluctuations, than the interaction between the CH4
molecule and the nanotube wall. However, as the nanotube diameters
increase, the repulsive character of the CH4 interaction
with other methane molecules decreases significantly, indicating fewer
violent and elastic molecule-molecule collisions, while the fluctuation
in the interaction of the CH4 with the nanotube increases,
reflecting the more dispersed distribution of molecules throughout the
larger nanotube interior. Significantly, the fluctuation in the CH4-nanotube
interaction for the smallest diameter (6,6) nanotube (estimated to be
0.02 - 0.04 eV, on average) is larger than both kBT/2
at 300 K (0.0125 eV) and the mean interaction of one CH4
molecule with the other CH4 molecules in the nanotube. |
(2) Equilibrium Transport of O2
(a) Setup of System
An
equilibrium-state system consists of an opened, hydrogen-terminated
nanotube in a self-contained, three-dimensional "box" that is filled
with oxygen molecules.
|
|
(b) Radial Distribution of O2
in Nanotubes
Figures on the
right are projection views of equilibrium transport systems consisting
of a 107 Å long (10,0), (14,0), and (17,0) opened CNTs at 500 ps.
O2 molecules are separated from the nanotube by about 3.5
Å both inside and outside the (10,0) nanotube. The projection
view for a (14,0) nanotube shows O2 molecules in the
nanotube form a ring. Although the molecules in the (14,0) nanotube
occupy more space as the nanotube diameter is increased by 1.6 Å,
the wall-O2 distance is slightly decreased by 0.2 Å,
and standard deviation is still small. For a larger (17,0) nanotube,
the spatial distribution of O2 molecules changes to a
different pattern. Some O2 molecules are located at the
center of the nanotube, and they look separated from the other
molecules that still form a layer apart from the wall.
|
(c) Diffusion Modes
In the figures on the left, the
relationship of the MSD with time shows a change in behavior as the
nanotube diameter increases. The MSD for the smallest nanotube smoothly
fluctuates above the ln <z2> = ln t
line, and abruptly drops down around 250 ps. The fluctuation of the
curve for the (14,0) nanotube is less radical. The curve for the (17,0)
nanotube is almost parallel to the dashed line for the range from 15 ps
to 250 ps. That is, the equilibrium transport of O2
molecules in nanotubes as large as the (17,0) nanotube is close to
normal diffusion mode. |
|
(3) Transport in
Shorter Nanotubes for Shorter Time
(a) Single Nanotubes:
50 Å long; 100 ps
If the molecular
structure is spherical, molecular transport behavior can be clearly
distinguished as either normal-mode or single-file mode in time scales
of about 100 ps. For methane, no changes in diffusion mode and minimal
changes in diffusion coefficient are predicted as a function of the
helical symmetry of the nanotube.
If the molecular
shape is asymmetrical, transition-mode diffusion occurs when molecules
are able to pass each other if they are perfectly aligned parallel to
the nanotube axis, but are not able to pass each other if they have
undergone small-angle rotational motion during transport. In this case,
molecules can pass each other some of the time but not at other times.
For ethane and ethylene, molecular transport is predicted to follow a
spiral path around the circumference of the nanotube that matches the
helical symmetry of the system. This only occurs for nanotubes with
diameters between 16 and 22 Å at very low molecular
concentrations. The driving force is the strong interaction energy
between C-C molecular and nanotube wall bonds; nanotubes try to
continue this alignment as the molecules move through the nanotube.
(b) Diffusion in Nanotube
Bundles: Bundle of (10,0) nanotubes; methane
- No molecules are predicted to diffuse into the
interstitial sites in the bundle.
- Interaction energy = - 0.24 eV/molecule inside
the nanotubes.
- Interaction energy = - 0.16 eV/molecule between
the nanotubes.
- For a single nanotube that is not part of a
bundle, the interaction energy is - 0.31 eV/molecule.
|
|
(c) Molecular Separations from
Mixtures
| |
(10,0) |
(8,8) |
(10,10) |
| CH4/n-C4H10 |
CH4 : normal mode
n-C4H10 : no motion |
CH4 : normal mode
t-C4H10 : no motion |
both molecules diffuse
CH4: normal mode
C2H6: transition
throughput of CH4/C2H6 is 2:1 |
| CH4/t-C4H10 |
CH4 : normal mode
n-C4H10 : single-file
throughput of CH4/ n-C4H10 is
18/1 |
CH4 : normal mode
t-C4H10 : no motion |
both molecules diffuse
both molecules are normal mode
throughput of CH4/C2H6 is 3:2 |
| CH4/C2H6 |
CH4 : normal mode
n-C4H10 : single-file
throughput of CH4/ n-C4H10 is
10/1 |
CH4 : normal mode
n-C4H10 : single-file
throughput of CH4/ t-C4H10 is
15/1 |
both molecules diffuse
both molecules are normal mode
throughput of CH4/C2H6 is 4:3 |
- As the diameters of the nanotubes decreases, the amount of
separation of the molecular species increases.
- As the differences in the relative size of the molecules
increases, the amount of separation of the molecular species increases.
- Separation was predicted for all cases considered.
|