2
amidoamine (PAMAM) dendrimers, which vary from 0.6
nm for generation 0 (G0) to 6 nm (G10) [10, 11, 13, 15].
Experiments and theory have shown that the capacitance
in EDLCs is significantly enhanced when the pore size of
the electrode material matches the size of the electrolyte
ions [16, 17].
Interfacial adsorption of dendrimers on electrodes dy-
namically exposes their charge groups to the electrolyte,
which facilitates the formation of unique electric double
layer structures. This has been demonstrated by exper-
iments by Guo et al. [18], who reported that hyper-
branched polymers, like dendrimers, exhibit very low
losses in the dielectric response function, even at high
operational frequencies (∼1 MHz). Freire et al. also
showed that the presence of dendrimer can screen re-
pulsive contacts between different counter ion molecules
and favored ionic conductivity [19]. These interesting
properties of dendrimers make them potentially excellent
candidates as electrolytes in EDLCs, however, a thorough
microscopic examination of the interfacial behaviour, and
the resulting effect on electrochemical performance, have
not yet been elaborated.
Computer simulations employing molecular dynamics
(MD) based quantum mechanically derived accurate po-
tentials (force-fields) is a common tool for elucidating
the microscopic nature of interfacial systems, and would
be well suited for exploring the role of dendrimers in
modulating the EDL structure and ultimately the perfor-
mance of EDLCs. The key here is an accurate description
of the interaction parameters, coarse-grained simulations
revealed significant force-field dependence in the binding
strength of dendrimers to graphene electrodes, as the sys-
tem’s pH [11] is varied [20]. To more clearly understand
the nature of the interaction of the dendrimer-graphene
composites requires us to go beyond coarse-grained force-
fields and perform fully atomistic simulations. Yet, fully
atomistic simulations of dendrimers at interfaces, where
the dendrimer is being used as an electrolyte, are rel-
atively rare to the best of our knowledge. In contrast,
there have been several experimental [21, 22] and simula-
tion studies [23, 24] that reported carbon-based electrode
materials and studied the capacitance values with ionic
liquids being the electrolytes [25]. For example, Trigue-
rio et al. [21] reported that the dendrimer functionalised
carbon nanotubes can improve the nanotube’s perfor-
mance as an electrode. Another study by Chandra et al.
[26] reported that dendrimer functionalised nanoparticles
coated on an electrode surface can enhance the surface
area available to the electrolyte atoms, thereby achieving
efficient charge transfer and low contact resistances. In
another experimental work, Liu et al. [27] used den-
drimer functionalized graphene-oxide sheet as a coating
on the sulfur electrode of a Li-S battery and achieved long
cycle life (up to 500 cycles). When considering common
electrolytes in EDLCs, various computational studies on
RTILs have been reported, including work by Yeh et al.
[28] which discussed the effect of periodic boundary con-
ditions (BCs) in EDLC simulations.
Recently, we used atomistic MD simulations to study
the structural deformations [13] and the free-energy of
the binding [29] of PAMAM dendrimers at a charge
neutral graphene/water interface, as a function of the
protonation state of the dendrimer. We found that the
van der Waals interactions play a pivotal role in driv-
ing dendrimer adsorption. We also found that mod-
erately charged (neutral pH) dendrimers achieve maxi-
mum surface wetting as compared to the non-protonated
(high pH) and fully protonated (low pH) dendrimers. We
showed that lower generation dendrimers tended to de-
form and form flat, disk-like architectures, with good sur-
face accessibility, at the graphene/water interface [13].
These observations suggest that the lower generation PA-
MAM at neutral pH condition as an ideal choice for
achieving maximum charge densities in PAMAM-based
supercapacitors. Therefore, in this work, we consid-
ered a G2 PAMAM dendrimer at neutral pH, and eluci-
date the electrochemical performance in graphene/water
based EDLCs. Beyond accurate simulations of dendrimer
based systems, we are also concerned with modelling bi-
ased nanoscale interfaces, as a means of probing elec-
trochemical effects in EDLCs. Here, there are two main
computation methods commonly employed for doing this:
1) the constant charge method (CCM) – the charges of
the electrode atoms are fixed, and 2) the constant poten-
tial method (CPM) – a grand canonical statistical me-
chanical ensemble is defined by means of a fictitious bath
that exchanges electrons with the electrodes to maintain
a constant electrode potential (the number of electrons
and chemical potential are conjugate pairs) [30, 31]. The
CPM approach is generally preferred as it enables simula-
tions that are more directly comparable to experiments.
However, it is somewhat restricted in its applicability due
to significant additional computational demands. To this
end, Wang et al. [30] compared both approaches for
a LiClO4-acetonitrile/graphite EDLC and showed that
both the approaches lead to similar ion and solvent den-
sity profiles for voltages less than 2V. Comparing the
performance of both approaches for a more complicated
electrode/electrolyte morphology is one of the aims of
this study. Moreover, to address the computational chal-
lenges, Reed and coworkerset al. [32] have developed an
efficient approach for simulating cells within CPM [30] in
the LAMMPS simulation engine, which we employ here.
The manuscript is organized as follows. In section
II, we provide the model building and the simulation
methodologies adopted in this work. In section III, we
present our results on the electrostatic potential, charge
density profiles and the capacitance values, obtained
from the CCM and the CPM approaches. Finally, in
section IV, we summarize our findings and conclude with
the key insights from our study to provide future research
directions.