21st International Symposium on Plasma Chemistry (ISPC 21)
Sunday 4 August – Friday 9 August 2013
Cairns Convention Centre, Queensland, Australia
Reactive molecular dynamics simulations of the functionalization of
polyethylene surfaces using an Ar/O2 plasma
Johan Minnebo, Erik C. Neyts and Annemie Bogaerts
Research group PLASMANT, University of Antwerp, Belgium
Abstract: In order to study the effects of plasma species on a polymer surface, molecular
dynamics simulations of oxygen atoms impacting on an amorphous polyethylene film were
carried out. Furthermore, the oxygen atom reactivity after impact was investigated. The
results indicate that oxygen atoms with thermal energy react only after impact, and that the
breaking of C-C bonds is the initial reaction of the oxidation process.
Keywords: Molecular dynamics, plasma, polyethylene, surface, functionalization
1. Introduction
Plasma treatment of polymer surfaces is known to be a
widely used process to improve the surface properties of
the material without altering the bulk properties. Polymers
typically have a low surface energy, whereas many
applications require a higher surface energy. Therefore,
plasma surface treatment is commonly used to increase
the hydrophilicity and adhesion properties through the
introduction of polar functional groups [1], for example
for applications in food packaging [2], optical coatings [3]
and biomedical materials [4].
Although a significant amount of experimental
research has already been carried out in this field,
knowledge of the mechanistic details behind plasmasurface interactions is still limited [5], especially when
polymers are involved [6]. Polymers have, in contrast to
inorganic materials, a complex physical and chemical
structure that can be relatively easily modified by various
reactive plasma species. In order to gain a better
understanding of how the plasma species interact at the
plasma-surface interface, a theoretical approach using
molecular dynamics (MD) simulations can be employed.
Molecular dynamics is an atomistic model that can
provide insights in the (reaction) mechanisms involved in
plasma-surface interactions.
In the current study, the effect of impinging oxygen
atoms with thermal energy as well as with an energy of 1
eV on an amorphous polyethylene (PE) surface was
investigated, as well as the reactivity of oxygen atoms
after impact. Thermal oxygen atoms are the main reactive
species in an argon/oxygen plasma created by a dielectric
barrier discharge (DBD) operating at atmospheric
pressure [7]. Therefore, the surface modification obtained
by the simulations can be related to experimental results.
2. Method
2.1 Molecular dynamics simulations
Molecular dynamics is a simulation technique to follow a
set of atoms in time and space, which interact with each
other through an interatomic potential. The negative
gradient of this potential governs the forces on the atoms,
while their motion is calculated by solving the set of
coupled equations of motion. The simulations in this work
utilize the reactive force field for hydrocarbon oxidation,
developed by Chenoweth and coworkers [8]. The oxygen
atom impact simulations were carried out using the
LAMMPS simulation package [9], while the ReaxFF
program [10] was used for the bulk oxidation simulations.
Periodic boundary conditions were imposed on the x, y
and z directions, and the time step was set to 0,25 fs.
2.2 Model polyethylene structure
The polyethylene structures used in this study were
created using a self-written script in the Python
programming
language.
This
script
generates
polyethylene chains through the periodic boundaries in
the x and y directions, and allows for branching. The
structure size was set to 30Åx30Åx20Å in the x, y and z
dimensions, respectively. The minimum chain length was
specified to be ten carbon atoms, while the branching
parameter was chosen so that a side chain of at least one
ethylene unit was placed on approximately one out of
every ten carbon atoms of the primary chain.
Furthermore, a few ether and alcohol functional groups
were added to represent a native oxygen content of 1,2
wt%. The final polyethylene structures had a density
varying from 0,91 to 0,93 g/cm³ and contained around
2150 atoms.
3. Results
3.1 Atomic oxygen impact
In order to simulate oxygen atoms impinging on a
polyethylene surface, the structure was first allowed to
relax for 25 ps, during which an interface layer was
formed with polyethylene chains sticking out of the
surface (Fig. 1). A distinction can be made between a
bulk region, where the change in structure density is
minimal, and an interface region, where the structure
density is significantly lower. The bulk region ranges
from 0 to 19 Å, while the interface region ranges from 19
to 25 Å. Before and during the impact simulations, the
atoms in the bottom 2Å of the structure were kept at fixed
positions to mimic the immobility of the bulk structure.
Three different polyethylene surfaces were created, on
each of which 200 independent atomic oxygen impacts
were simulated at an incident angle normal to the surface.
The impinging oxygen atoms were simulated with
thermal energy as well as with 1 eV of kinetic energy.
species in the plasma can greatly increase the oxygen
content of polymer surfaces [11,12]. Therefore the
simulations were repeated with impacting oxygen atoms
with an energy of 1 eV.
Figure 2: Outcome distribution of atomic oxygen impacts
on amorphous polyethylene after 600 independent
impacts, with thermal energy as well as with an energy of
1 eV.
Table 1: Number of reactive events out of 600
independent impacts of atomic oxygen on amorphous
polyethylene, both with thermal energy as well as with an
energy of 1 eV.
Number of events
Figure 1: Side view of the model polyethylene surface
with, as indicated on the left, a bulk region (1) and an
interface region (2). Hydrogen, carbon and oxygen atoms
are represented in white, gray and red, respectively.
The impact of an oxygen atom with the model
polyethylene surface can have four different outcomes:
firstly it can be reflected from the surface without further
interaction, secondly it can be stopped at the surface and
stay in the interface region, thirdly it can penetrate the
surface to reach the bulk region, and lastly it can react
with the surface to break existing bonds and/or to create
new bonds. These outcomes will be further referred to as
“reflection”,
“adsorption”,
“incorporation”
and
“reaction”, respectively.
In the case of oxygen atoms with thermal energy, Fig.
2 shows that the oxygen atoms are most likely to adsorb
to the surface due to their relatively low impact velocity.
Only some are reflected from the surface, and about 20%
of the oxygen atoms can penetrate the surface upon
impact. Although approximately 4 % of the oxygen atoms
react with the surface, it should be noted that all reactions
are due to a single alcohol moiety in the central part of the
interface region of one of the three polyethylene surface
models (Table 1). The impinging oxygen atom can either
abstract the hydrogen from the alcohol or form a bond
similar to a hydrogen bond, resulting in a stabilized
oxygen atom. No reactions were observed between the
impinging oxygen atoms and the carbohydrate chains.
From experiments, however, it is known that oxygen
Thermal
1 eV
C-C cleavage
0
25
H abstraction
0
3
Alcohol formation
0
2
Ether formation
0
1
H stabilisation
0
6
Alcohol interaction
25
18
Reaction
In the case of oxygen atoms with an energy of 1 eV, Fig.
2 shows that sticking is still the most likely outcome, but
the higher impact velocity leads to a significant decrease
in sticking in favor of reflection, while the probability of
incorporation increases slightly. More importantly, the
oxygen atoms show an increased reactivity towards the
carbohydrate chains. Although the single alcohol moiety
is still responsible for many reactions (Table 1),
approximately 6 % of the impinging oxygen atoms react
with the carbohydrate chains. The cleavage of a C-C bond
is by far the most likely reaction, but hydrogen
abstraction, alcohol formation as well as ether formation
were also observed. When the impacting oxygen atom
binds with a hydrogen atom, but lacks the energy to
abstract it, the oxygen is found to remain bonded, as if it
was stabilized by a metastable hydrogen bond.
As can be seen in Table 2, the penetration depth of
oxygen atoms is rather shallow; the oxygen atoms that
can penetrate the surface reach only an average depth of
2,8 Å with thermal energy and 4,3 Å with an energy of 1
eV. In comparison, experiments have shown that an
increase of oxygen can be found up to several nm in the
polymer surface [13]. This leads to the conclusion that the
impinging oxygen atoms readily diffuse into the surface
before reacting.
Table 2: Average penetration depth of the fraction of
oxygen atoms that are incorporated in the amorphous
polyethylene surface, out of 600 independent impacts,
both with thermal energy as well as with an energy of 1
eV.
Average penetration
depth (Å)
Thermal
2,8 ± 1,6
1 eV
4,3 ± 2,4
In combination with the results from the oxygen atom
reactivity on impact, it seems very likely that the
oxidation of the polymer surface does not happen directly
on impact, but only a certain amount of time after impact.
Further simulations have been carried out to investigate
the reactivity of oxygen atoms that have diffused to the
bulk region of the polymer surface.
3.2 Bulk oxidation
To investigate the oxygen atom reactivity in the bulk of
amorphous polyethylene, 200 oxygen atoms were added
to the model structure corresponding to approximately 24
wt% of oxygen (Fig. 3). Four independent structures were
used to approximate a representative sample of
amorphous bulk polyethylene. After a short
thermalization of 5 ps, already one third of the oxygen
atoms had recombined to form oxygen molecules, leaving
on average 127 oxygen atoms free to interact with the
polyethylene structure. These oxygen atoms were then
followed during a simulation time of 250 ps.
Figure 3: Top view of the model bulk polyethylene, after
the addition of 200 oxygen atoms and thermalization.
Hydrogen, carbon and oxygen atoms are represented in
white, gray and red, respectively.
At room temperature, only eight of the free oxygen atoms
had reacted with the polyethylene chains to form new
functional groups, leading to an average reactivity of 1,5
% after 250 ps. Interestingly, the oxidation process was
more likely to begin at the native ether groups than at the
carbohydrate chains. While there were only four ether
groups present in a polyethylene structure containing
more than 700 carbon atoms, two bond breaking reactions
involving the ether groups were observed, compared to
one reaction involving a carbohydrate chain. Furthermore,
the terminal oxygen radicals created during the initiating
reaction seem to increase the reactivity of the carbon atom
to which they are attached. Of the six terminal oxygen
radicals created during the simulations, one was oxidized
to an aldehyde by hydrogen abstraction. The aldehyde
group was then further oxidized by a free oxygen atom to
form a functional group with two oxygen radicals
attached to the terminal carbon atom, which was
eventually eliminated from the chain by a carbon-carbon
bond breaking reaction of another free oxygen atom.
In order to gather more information on the types of
reactions that take place during bulk oxidation, the
simulations were repeated at a higher temperature of
350K. Because of the increased temperature, 36 of the
free oxygen atoms reacted with the polyethylene chains
during the simulation, leading to an average reactivity of
7,2 %. Just as at room temperature, the initial oxidation
reactions were more likely to involve an ether group than
carbohydrate chains: four and three reactions,
respectively. After the initial formation of a terminal
oxygen radical, several subsequent reactions were
observed, confirming the increased reactivity at the site
where the chain was broken. The prevalent reaction
mechanism, however, was not aldehyde formation but the
breaking of the carbon-carbon bond at the terminal site,
leading to the formation of formaldehyde and a new
terminal oxygen radical. The formaldehyde was quickly
oxidized by oxygen attachment, and could be oxidized
further to eventually form carbon dioxide.
These results suggest that the main reaction pathway
of the oxidation of amorphous polyethylene by oxygen
atoms starts with the scission of the carbon backbone of
the polymer to form terminal oxygen radicals, followed
by the elimination and further oxidation of formaldehyde
as shown in Fig. 4.
Figure 4: Observed main reaction pathway for the
oxidation of amorphous polyethylene by oxygen atoms
4. Conclusion
The effects of oxygen atom impact on an amorphous
polyethylene surface have been investigated, as well as
the oxygen reactivity after impact. With thermal energy,
oxygen atoms are most likely to stick to the surface and
show no reactivity towards the carbohydrate chains,
whereas with an energy of 1 eV, approximately 6% of the
impacting atoms is observed to react, leading almost
exclusively to the breaking of a carbon-carbon bond.
Free oxygen atoms in the bulk of an amorphous
polyethylene structure show a reactivity of 1,5% after a
simulation of 250 ps at room temperature, compared to
7,2% at 350K. The main reaction pathway is found to be
the scission of the backbone chain, followed by
elimination and oxidation of the terminal radical.
5. Acknowledgements
The authors gratefully acknowledge financial support
from the Prime Minister’s Office through IAP VII. This
work was carried out in part using the Turing HPC
infrastructure at the CalcUA core facility of the
Universiteit Antwerpen, a division of the Flemish
Supercomputer Center VSC, funded by the Hercules
Foundation, the Flemish government (department EWI),
and the Universiteit Antwerpen.
6. References
[1] M. Mortazavi, M. Nosonovsky, Applied Surface
Science, 258, 6876 (2012)
[2] M. Ozdemir, C. U. Yurteri, H. Sadikoglu, Crit. Rev.
Food Sci. Nutrition, 39, 457 (1999)
[3] U. Schulz, Appl. Opt., 45, 1608 (2006)
[4] F. Poncin-Epaillard, G. Legeay, J. Biomater. Sci.,
Polym. Ed. 14, 1005 (2003)
[5] K. Chenoweth, A. C. T. van Duin, W. A. Goddard III,
J. Phys. Chem. A, 112, 1040 (2008)
[6] J. J. Végh, D. B. Graves, Plasma Process. Polym., 6,
320 (2009)
[7] M. Moravej, X. Yang, R. F. Hicks, J. Penelon, S. E.
Babayan, Journal Of Applied Physics, 99, 093305 (2006)
[8] S. Plimpton, J Comp Phys, 117, 1 (1995)
[9] A.C.T. van Duin, S. Dasgupta, F. Lorant, F and W.A.
Goddard III, J. Phys. Chem. A, 105, 9396 (2001)
[10] J. J. Végh, D. Nest, D. B. Graves, R. Bruce, S.
Engelmann, T. Kwon, R. J. Phaneuf, G. S. Oehrlein, B. K.
Long, C. G. Willson, Journal Of Applied Physics, 104,
034308 (2008)
[11] Y. Ren, Y. Hong, J. Sun, Y. Qiu, Journal of Applied
Polymer Science, 110, 995 (2008)
[12] H. Drnovská, L. Lapčík Jr., V. Buršíková, J. Zemek,
A. M. Barros-Timmons, Colloid Polym. Sci. 281, 1025
(2003)
[13] I. Banik, K. S. Kim, Y. I. Yun, D. H. Kim, C. M.
Ryu, C. S. Park, G. S. Sur, C. E. Park, Polymer, 44, 1163
(2003)