An ab initio study of hydroxylated graphane

Francesco Buonocore, Andrea Capasso, and Nicola Lisi

Citation: The Journal of Chemical Physics 147, 104705 (2017); doi: 10.1063/1.4986858
View online: http://dx.doi.org/10.1063/1.4986858
View Table of Contents: http://aip.scitation.org/toc/jcp/147/10
Published by the American Institute of Physics
 THE JOURNAL OF CHEMICAL PHYSICS 147, 104705 (2017)

An ab initio study of hydroxylated graphane

Francesco Buonocore,1,a) Andrea Capasso,2 and Nicola Lisi1
1 ENEA, Casaccia Research Centre, I-00123 Rome, Italy
2 Istituto Italiano di Tecnologia, Graphene Labs, Via Morego 30, Genova 16163, Italy
(Received 7 June 2017; accepted 25 August 2017; published online 12 September 2017)

Graphene-based derivatives with covalent functionalization and well-defined stoichiometry are highly
desirable in view of their application as functional surfaces. Here, we have evaluated by ab initio calcu-
lations the energy of formation and the phase diagram of hydroxylated graphane structures, i.e., fully
functionalized graphene derivatives coordinated with –H and –OH groups. We compared these struc-
tures to different hydrogenated and non-hydrogenated graphene oxide derivatives, with high level of
epoxide and hydroxyl groups functionalization. Based on our calculations, stable phases of hydroxy-
lated graphane with low and high contents of hydrogen are demonstrated for high oxygen and hydrogen
partial pressure, respectively. Stable phases of graphene oxide with a mixed carbon hybridization
are also found. Notably, the synthesis of hydroxylated graphane has been recently reported in the
literature. Published by AIP Publishing. [http://dx.doi.org/10.1063/1.4986858]

I. INTRODUCTION hydrogen spillover, hydrogenated GO, referred hereafter as

Graphene is a two-dimensional crystal of sp2 hybridized
carbon atoms forming a honeycomb structure that exhibits
many superlative properties. Chemical modification of
graphene is a crucially important means of extending these
properties1 and unlock features of interest for several other
applications, ultimately creating the organic functional all-
surface material par excellence. For example, graphene deriva-
tives are interesting as thin interlayers in organic photovoltaics
both in hydrogenated2 and oxidized form.3 For this reason,
in recent years graphene oxide (GO)—the single layer of
graphite oxide—has attracted a large interest from the scien-
tific community. Graphite oxide is synthesized in solution by
the reaction of graphite with oxidants, following different syn-
thesis routes.4 Despite being known for almost 150 years, the
exact chemical structure of graphite oxide is still questioned
since its composition can strongly depend on many factors,
such as the raw material employed (the often inconstant nat-
ural graphite), the type of oxidant, the reaction condition,
etc.4–6 However, it is generally accepted that the ideal GO is
a functionalized form of graphene that contains only epoxide
and hydroxyl functional groups disorderly distributed along
a defect-free and two-dimensional carbon lattice.4–14 These
groups are covalently attached out of the plane to carbon atoms
bonded into the lattice in sp3 hybridization. Conversely, ketone
and carboxylic groups can only be distributed along edges
or at defected lattice sites since the occurrence of their dou-
ble bonds on the sp3 plane implies the breaking of the lattice
continuity with sp2 points or edges. Covalently bonded hydro-
gen cannot be expected because GO is commonly prepared by
strong oxidative chemical routes, such as Hummers’ method.5
Nevertheless, using a post-processing technique, such as
a)Authorto whom correspondence should be addressed: francesco.
buonocore@enea.it.
HGO, can be synthesized.14 However, the possibility that
graphene derivatives could be synthesized by other chemical-
physical methods (e.g., by chemical vapor deposition—
CVD—with suitable gas precursors and conditions) eventu-
ally adding plasma activation of selected radicals cannot be
ruled out. All these methods could provide a route to the for-
mation of hydroxylated graphane (HyGH) structures, i.e., a
fully functionalized graphene derivative with –H, and –OH
groups bonded to the graphene network and full sp3 hybridiza-
tion. Notably, the synthesis of such graphene derivatives, with
a well-defined stoichiometry, has been reported in a recent
work 15 where the synthesis of HyGH and graphol (highly
hydroxylated graphene) was achieved through hydroboration
of GO followed by protonation. These compounds showed cat-
alytic properties towards the oxidation of biomarkers and even
towards hydrogen evolution. Moreover, Tuček and co-authors
were able to prepare room temperature hydroxofluorographene
magnets by replacing fluorine atoms in fluorographene with
hydroxyl groups, obtaining very high magnetic moments.16
Graphene functionalized with H and O atoms is the very
analogue of molecules found in organic chemistry, where the
skeleton (either a chain or ring) of carbon atoms supports the
functional groups that regulate the chemical interactions with
other molecules and surfaces, whether organic or not. The dif-
ferent arrangements of hydroxide and epoxide groups in the
GO structures change the electronic and optical properties,
such as the work function and the bandgap, and the chemi-
cal interactions. For example, it is well known that graphene
presents a hydrophobic behavior while GO is hydrophilic.17,18
By adding the hydrogen and hydroxyl functionalization, the
HyGH structures proposed in this work would instead allow a
fine control of the chemical interface with different solids and
liquids. Both the residual sp2 alkene bonds and the –OH groups
can finally act as sites available for further covalent func-
tionalization processes. Moreover, dosing the hydrophilic OH
bonds on these graphene derivatives could have tremendous

0021-9606/2017/147(10)/104705/8/$30.00 147, 104705-1 Published by AIP Publishing.

104705-2 Buonocore, Capasso, and Lisi
implications on many application fields, both by conferring
specific functions to surfaces for making devices and by
allowing new wet processes.
Several theoretical studies have investigated the distribu-
tion of oxygen species (either periodic or randomly distributed)
in the two-dimensional carbon geometry.19–26 Yan and co-
authors23 have found that stable GO structures are constituted
by hydroxyl and epoxide groups that aggregate together to
form specific kinds of strips separated by sp2 carbon regions.
On the other hand, Wang and co-authors25 have shown the
stability of fully covered, thermodynamically stable GO struc-
tures, without the presence of any sp2 carbons. The opposing
experimental evidence of sp2 carbons in GO was explained as
a metastable phase. However, no one has investigated so far the
thermodynamic stability of the HyGH structures. As a remark,
it is well known, and also quite intuitive, that H termination can
stabilize sp3 carbons. Therefore, the occurrence and stability
of fully functionalized structures with simple arrangement of
hydroxyl/epoxide groups and hydrogenated carbon is of great
interest.
The core of the present paper is the investigation of the
hydroxylation of graphane by the progressive substitution of
H with –OH groups. Since the chair conformer is the most
stable graphane, the natural choice was to build up the super-
cell of our models using the equivalent hexagonal unit cell
of graphene. Our model is based on an 8 atom periodic lattice
cell, therefore only highly functionalized structures are consid-
ered and explored. We investigated the stability of such HyGH
structures and compared them to different hydrogenated and
non-hydrogenated graphene oxide derivatives, with epoxide
and hydroxyl groups functionalization. Our study is aimed at
determining which structures, among those sharing the same
supercell, have stable phases, and are thus favored with respect
to the competing formation of water by the available H and O
atoms. We build up the phase diagram of HyGH as a func-
tion of the atomic chemical potentials of H and O by using
first-principles calculations.25
II. THEORETICAL METHODS
The computational approach was based on a pseudo-
potential plane-wave method using PWSCF code as imple-
mented in the QUANTUM-ESPRESSO package.27 We used
the generalized gradient approximation (GGA) with the
Perdew, Burke, and Ernzerhof (PBE) exchange-correlation
functional,28 taking into account spin-polarisation for all cal-
culations. The pseudo-potential plane-wave calculations were
performed using Vanderbilt ultra-soft pseudo-potentials.29
All geometry optimizations were performed with cutoff
for the wave functions of 40 Ry and cutoff for the charge
EB = (ETOT − EG − nEPO EO − nHYD EH − nh EOH )/(nEPO + nHYD + nh ),
where EG , EO , EH , and EOH are the total energy of
graphene, O atom, H atom, and OH group, respec-
tively. Following definition (2), negative (positive) binding
J. Chem. Phys. 147, 104705 (2017)
density of 400 Ry and 12 × 12 × 1 Monkhorst-Pack grid. The
systems were fully relaxed with a convergence threshold of
0.001 Ry/Å on the inter-atomic forces. We had previously opti-
mized the unit cell of graphane, a fully hydrogenated graphene
structure,30 by imposing that the stress on the cell is less than
0.04 GPa, and we are here using the same cell for each of the
functionalized graphene structures. The final self-consistent
calculations of the optimized structure properties made use
of a cutoff for the wave functions and charge density of 60
and 600 Ry, respectively, allowing a convergence of the total
energy below 0.002 eV/atom, compared with the calculations
with 70 Ry of cutoff and 15 × 15 × 1 Monkhorst-Pack grid.
In all of the examined structures, we have added O and/or H
atoms to the same carbon honeycomb supercell built by using
the in-plane hexagonal 2 × 2 unit cell of the graphene that
contains 8 C atoms, vacuum gap between each monolayer and
its image was set to 20 Å.
We considered two different kinds of functionalization
that we distinguish with the symbols GO H and GO : the GO H
structures have all C atoms sp3 hybridized as in graphane;
the GO structures have a mix of sp2 and sp3 hybridization as
in partially functionalized graphene. In this way, for instance,
GO H(nh OH nEPO O) indicates that nh hydroxide and nEPO epox-
ide groups have been added to the ideal graphene supercell,
and then all of the remaining (8-nh -2nEPO ) C atoms have
been hydrogenated; GO (nh OH nEPO O nHYD H) would indicate
that nh (nEPO ) hydroxide (epoxide) groups and nHYD hydro-
gen atoms (those that are not in the hydroxide group) have
been added to C atoms. The functionalized (nh +2nEPO +nHYD )
C atoms are sp3 hybridized and the remaining (8-nh -2nEPO -
nHYD ) C atoms in the supercell are sp2 hybridized. All the
GO H structures without epoxide groups, e.g., GO H(nh OH), are
coincident with HyGH, while the GO H structures containing
epoxide groups, e.g., GO H(nh OH nEPO O), are more in general
HGO.
The formation energy per functionalized C atom is defined
as
!
1 1
EF = ETOT − EG − nO EO2 − nH EH 2 /nF , (1)
2 2
where ETOT and EG are the total electronic energy of the GO H
or GO and graphene systems, nO and nH are the number of
O and H atoms, nF is the number of functionalized C atoms,
and EO2 and EH2 are the total energy of O2 and H2 molecules,
respectively. The module of formation energy indicates the
extent of stability or instability of the system. Lower formation
energies (i.e., more negative values) correspond to more stable
systems.
We calculated the binding energy per functional group in
accord with the following definition:
(2)
energy corresponds to bound (unbound) system and its
module gives a measure of the average strength of the
bonds.
104705-3 Buonocore, Capasso, and Lisi
The key quantity used to calculate the GO and GO H total
energy phase diagram is the free energy of formation defined
as
∆GF = ETOT − EG − nO µO − nH µH . (3)
In Eq. (3), we omitted to take into account the vibrational
Helmholtz free energy that should be evaluated by phonon cal-
culations under the quasi-harmonic approximation.31 This is
beyond the scope of the present paper since the accurate esti-
mation of the vibrational free energy would require a careful
study of the phonon dispersion for the numerous structures
examined, most of which have not been investigated in the
previous literature. This makes it impractical to include calcu-
lations of the vibrational modes in our stability study. There-
fore, the energy of the graphene derivatives, at 0 K and not
considering the zero point energy (ZPE), is employed to infer
stability instead, as it is typically done in the crystal structure
search.32
The chemical potentials of O and H atoms in the gaseous
phases of O2 and H2 , at partial pressures PO2 and PH2 ,
respectively, are defined as follows:33
!
1 1 1 PO2
µO = EO2 + ∆µO2 + KB T ln , (4)
2 2 2 P0
! groups. However, structures that were equivalent by symme-
1 1 1 PH2
µH = EH2 + ∆µH2 + KB T ln , (5)
2 2 2 P0
where P0 is the atmospheric pressure. The ∆µ quantities are the
entropy and enthalpy contributions to the chemical potential
of the gaseous phase defined as
∆µ = ∆H(T) − T∆S(T), (6)
where ∆H and ∆S are the enthalpy and entropy contribu-
tions at atmospheric pressure and absolute temperature T,
as given from the thermochemical tables.34 They result to
be
∆H(T) = [H(T) − H◦ (Tr )] − [H(0) − H◦ (Tr )], (7)
∆S(T) = S◦ (T). (8)
We have taken into account that EH2 and EO2 in Eqs. (4)
and (5) are the total energies of the isolated molecules at
J. Chem. Phys. 147, 104705 (2017)
T = 0 K, therefore Eqs. (7) and (8) are the difference of the gas
enthalpy and entropy, respectively, between T = 0 K and the
temperature of interest at the reference pressure.
It is evident here that we can only evaluate the relative
stability of structures within the set of configurations we have
considered. Therefore, other structures, also with periodicity
larger than the in-plane 2 × 2 unit cell of the graphene, can be
more stable than those investigated here.
III. RESULTS AND DISCUSSION
A. Hydroxylated graphane
We have performed a systematic study of hydroxyl func-
tionalized GO H structures, where all of the C atoms not bound
to O were passivated with hydrogen so that every C atom is sp3
hybridized as in graphane in such a way to form HyGH struc-
tures. We have considered chair conformers only. We found
that the chair conformer of graphane is 0.1 eV per CH group
more stable than the corresponding boat conformer, while the
chair conformer of GO H(8OH) is 0.25 eV per COH group
more stable. We built up the starting geometries by the system-
atic substitution of graphane’s hydrogen atoms with hydroxide
try have been considered as one structure and discarded. The
relaxed geometries are shown in Fig. 1, where the two lowest
energy structures for each configuration, when available, are
reported together with the total energy differences between the
two isomers. The average geometrical parameters are reported
in Table I, where it can be seen that by substituting the H
atoms of graphane by hydroxyl groups, the average C–C bond
length increases so that the covalent bond gets weaker. On
the other side, the increase of the OH –H bond length can be
related to the formation of hydrogen bonds when the density of
hydroxyl groups increases. Here, OH indicates the O atom of
the hydroxyl group to avoid any ambiguity with the epoxy O.
The average C–OH bond length starts to decrease only when
one of the two sides is fully functionalized and then the forma-
tion of the hydrogen bonds chain is completed. The absolute
value of the binding energy, reported in Fig. 2, decreases by
FIG. 1. Relaxed geometries of selected configurations
for hydroxylated graphane. The total energy difference
between the lowest and second lowest energy isomer is
reported.
104705-4 Buonocore, Capasso, and Lisi J. Chem. Phys. 147, 104705 (2017)

TABLE I. Average geometrical parameters for hydrogenated graphene with

hydroxyl functional groups and number of functionalized C atoms nF . Bond
lengths are in Angstrom (Å).

System d(C–OH ) d(C–H) d(OH –H) d(C–C) nF

GO H(OH) 1.454 1.110 0.976 1.538 8

GO H(2OH) 1.454 1.080 0.998 1.539 8
GO H(3OH) 1.454 1.104 0.991 1.541 8
GO H(4OH) 1.452 1.100 0.998 1.540 8
GO H(5OH) 1.446 1.103 1.099 1.546 8
GO H(6OH) 1.439 1.098 1.008 1.550 8
GO H(7OH) 1.433 1.096 1.004 1.550 8
GO H(8OH) 1.431 ... 1.013 1.553 8

increasing the number of hydroxyl groups, implying the weak-

ening of the bond strength, in agreement with the bond lengths
reported in Table I. Besides, the formation energy, shown in
Fig. 3 (triangle points), decreases by increasing nh , so that
by replacing H atoms with hydroxyl groups the function-
alized graphene becomes more stable. We observe that the
GO H(nh OH) structures are even more stable than graphane,
which has the same total number of functionalized C atoms.
GO H(8OH), which is full passivated by hydroxyl groups
structure whose stability has been investigated by Wang,25 is
more stable than graphane by as much as 1.23 eV per func-
tionalized C atom. However, we highlight that GO H(8OH)
is the only structure with all C atoms sp3 hybridized con-
sidered in this paper that does not present atomic hydrogen
functionalization.
We have also considered the GO (OH H), GO (2OH 2H),
and GO (3OH H) structures that are derived from GO H(OH),
GO H(2OH), and GO H(3OH) but which contain only the mini-
mal number of H atoms necessary to passivate the C dangling
bonds due to unpaired electrons, i.e., non-graphane structures
but HGO. The resulting structures are passivated by one, two,
and one H atom, respectively, such that the tri-coordinated
C atoms are sp2 hybridized (as in graphene) and the
FIG. 3. Formation energy per functionalized C atom for hydroxylated
graphane (GO H), or graphene functionalized with hydroxyl groups mixed
sp2 and sp3 hybridization (GO ), versus the number of hydroxyl groups.
tetra-coordinated C atoms are sp3 hybridized (as in graphane).
The relaxed structures are shown in Fig. 4 and the average geo-
metrical parameters are reported in Table II. Among these three
structures, GO (2OH 2H) is the one with the shortest average
C–OH bond length and the highest absolute value of bind-
ing energy. We observe from Fig. 3 (round points) that the
free energy of formation decreases, i.e., stability increases, by
rising the number of hydroxyl groups.
B. Hydrogenated graphene with hydroxyl
and epoxy functional groups
We compare the HyGH structures to HGO structures
functionalized with both hydroxyl and epoxy groups. We con-
sidered the two limit structures GO (O OH) and GO (2O 2OH),
corresponding to low degree of oxidation and saturated oxida-
tion, respectively. The low energy isomers were investigated,
without studying the stability, by Lahaye and co-authors22
and are shown in Fig. 7. In this section we investigate the
hydrogenated structures GO H(O OH) and GO H(2O 2OH),
fully covered by adding five and two H atoms, respec-
tively. Moreover, we investigated GO (O 4H), corresponding
to graphene with one epoxy group adsorbed and passivated
with four H atoms, where two tri-coordinated C atoms in
the GO (O 4H) structure are sp2 hybridized. The relaxed
structures of GO H(O OH), GO H(2O 2OH), and GO (O 4H)
are shown in Fig. 4 and the average geometrical parame-
ters are reported in Table III. GO H(O OH) has the short-
est average C–O bond length and is also the structure with
more C atoms hydrogen-passivated. The GO H(O OH) and
GO H(2O 2OH) are the structures with the lowest value
of the binding energy and formation energy of 0.38 and
FIG. 2. Binding energy per functional group for hydroxylated graphane
(GO H), or graphene functionalized with hydroxyl groups mixed sp2 and sp3
hybridization (GO ), versus the number of hydroxyl groups.
0.50 eV per functionalized C atom, respectively, as reported
in Figs. 5 and 6. However, the two structures, epoxide-
and hydroxyl-functionalized, are both found to be less sta-
ble than GO H(nh OH) HyGH structures for nh > 1. On the
other side, GO (O 4H) is less stable with high formation
energy.
104705-5 Buonocore, Capasso, and Lisi J. Chem. Phys. 147, 104705 (2017)

FIG. 4. Relaxed geometries of selected configurations

for hydrogenated graphene functionalized with hydroxyl
and epoxy groups.

C. Non-hydrogenated graphene with hydroxyl non-hydrogenated GO (O OH) and GO (O) structures have high
and epoxy functional groups
The non-hydrogenated GO (O OH), GO (2O 2OH), and
GO (O) structures were also investigated to be compared
with HyGH structures. The relaxed structures are shown in
Fig. 7 and the average geometrical parameters are reported
in Table III. The GO (2O 2OH) structure has short average
C–OH and C–O lengths and it is also the one characterized
by the highest absolute value of the binding energy per func-
tionalized C atom. The GO (2O 2OH) structure is found to
be the most stable among them with formation energy of
formation energy of 0.010 and 0.408 eV per functionalized
C atom. In particular, the low stability of GO (O OH) is related
to the presence of a single unpaired electron in the unit cell
so that it reacts easily with functional groups in the environ-
ment. The unpaired electron results in a magnetic moment
of 1.0 bohr magneton per unit cell and a ferromagnetic-non
magnetic (FM-NM) states energy splitting of 0.16 eV.
The GO (nh OH) structures with nh = 1, 2, 3, 5, 6, and
7, not shown here, obtained from the most stable structures
found in Sec. III A by removing the hydrogen atoms, have
0.50 eV per functionalized C atom. On the other hand, the
TABLE II. Average geometrical parameters for partially hydrogenated
graphene with hydroxyl and epoxy functional groups, as well as number of
functionalized C atoms nF . Bond lengths are in Angstrom (Å). OH indicates
the oxygen atom of the hydroxyl group.
unpaired electrons and are very reactive, and so they are
unstable. Indeed, the GO (nh OH) structures with nh = 1, 5,
and 7 exhibit magnetic properties with 1.0 bohr magneton,
i.e., a single unpaired electron, per unit cell of eight carbon
atoms, and FM-NM states energy splitting of 0.14, 0.16, and
0.29 eV, respectively. The GO (nh OH) structures with nh = 2

System d(C–OH ) d(C–H) d(OH –H) d(C–O) d(C–C) nF

GO (OH H) 1.488 1.117 0.981 ... 1.480 2

GO (2OH 2H) 1.451 1.108 0.999 ... 1.510 4
GO (3OH H) 1.469 1.109 0.992 ... 1.507 4
GO H(2O 2OH) 1.457 1.1025 0.993 1.444 1.530 8
GO (O 4H) ... 1.1095 ... 1.468 1.513 6
GO H(O OH) 1.450 1.1058 0.977 1.461 1.532 8

TABLE III. Average geometrical parameters for graphene with hydroxyl and
epoxy functional groups, as well as number of functionalized C atoms nF . Bond
lengths are in Angstrom (Å). OH indicates the oxygen atom of the hydroxyl
group.

System d(C–OH ) d(C–O) d(OH –H) d(C–C) nF

GO (4OH) 1.463 ... 1.110 1.509 4

GO (O) ... 1.468 ... 1.470 2
GO (O OH) 1.472 1.461 0.981 1.485 3 FIG. 5. Binding energy per functional group for graphene functionalized with
GO (O 2OH) 1.501 1.464 1.001 1.496 4 epoxide and hydroxyl groups and full sp3 hybridization (GO H), or mixed sp2
GO (2O 2OH) 1.454 1.444 0.996 1.509 6 and sp3 hybridization (GO ), versus the total number of hydroxyl and epoxy
groups.
104705-6 Buonocore, Capasso, and Lisi
FIG. 6. Formation energy per functionalized C atom for graphene function-
alized with epoxide and hydroxyl groups and full sp3 hybridization (GO H),
or mixed sp2 and sp3 hybridization (GO ), versus the total number of hydroxyl
and epoxy groups.
and 3 exhibit 1.68 and 0.76 Bohr magneton per unit cell and
FM-NM states energy splitting of 0.15 and 0.015 eV, respec-
tively, while GO (6OH) exhibits 2.0 Bohr magneton per unit cell
(i.e., two unpaired electrons localized on two different carbon
atoms) and FM-NM states energy splitting of 0.37 eV. We were
not able to find anti-ferromagnetic states. Instead, GO (4OH)
has no unpaired electrons, and the average C–OH bonds are
longer, while the binding energies is higher and the formation
energy is lower than those of GO H(4OH), as shown in Figs. 2
and 3.
D. Phase diagram of hydrogenated graphene
with hydroxyl and epoxy functional groups
The formation energy as defined in Eq. (1) defines the
stability of the structures at T = 0 K. Instead, the phase dia-
gram serves to find the thermodynamic conditions that allow
the existence of those structures, which could be the result of
a growth process. It shows the dependence on the chemical
J. Chem. Phys. 147, 104705 (2017)
potentials of H2 and O2 , which themselves depend on their
partial pressures and thus implicitly on the temperature. A
structure is stable when the free energy of formation defined
in Eq. (3) is lower than zero with the condition that µO + 2µH
is equal or less than the formation energy of water Ef (H2 O).
In this way the chemical potentials of the atoms do not favour
the formation of water molecules. Here, µO and µH are defined
taking as reference the half of the total energy of O2 and H2
molecules, respectively. As analogues, µO and µH are requested
to be less than zero, otherwise O and H would form O2 and H2
molecules, respectively. High µO and µH in the phase diagram
shown in Fig. 8 correspond to high partial pressures, as also
low µO and µH correspond to low partial pressures.
Due to the given constraints, we have stable phases of
GO and HyGH with low contents of hydrogen functionaliza-
tion only for high µO (O-rich conditions) and low µH (H-poor
conditions), while we have stable phases of graphane and
HyGH with high contents of hydrogen functionalization only
for low µO (O-poor conditions) and high µH (H-rich condi-
tions). Indeed, at high µH and by increasing µO we move from
stable phase of graphane to the stable phases of GO H(1OH)
and GO H(2OH) HyGH structures, where one and two H atoms
of graphane are substituted by hydroxyl groups, respectively.
At high µO and low µH we have the stable phase of GO H(8OH)
HyGH, where all H atoms of graphane are substituted by
hydroxyl groups. The same holds for GO (2O 2OH), in agree-
ment with the finding of Yan and co-authors23 that the mix
of hydroxyl and epoxide groups disposed on opposite sides of
the graphene plane can minimize the energy and form stable
structures.
At greater µH we have the stable phases of GO H(7OH)
and GO H(6OH) HyGH, and by further increasing µO we have
the stable phase of GO (4OH) GO. The GO H(nh OH) HyGH
structures with 3 ≤ nh ≤ 5, GO H(O OH), GO H(2O 2OH), and
GO (O 4H) are not stable phases and dissociate to graphene,
component gas, and water. In other words, the decomposition
to water of each of the previous structures is an exothermic
process, as it can be verified by the calculation of enthalpy
of this reaction found to be negative for those structures. The
FIG. 7. Relaxed geometries of selected configurations
for graphene with hydroxyl and epoxy functional groups.
104705-7 Buonocore, Capasso, and Lisi
FIG. 8. Phase diagram of hydrogenated and non-hydrogenated graphene
functionalized with hydroxyl and epoxy groups vs. the chemical potentials
of oxygen and hydrogen reported in eV.
triple point where graphane, graphene, and water coexist is
J. Chem. Phys. 147, 104705 (2017)
promote a favourable kinetic to the formation not only of
HyGH but also of HGO. Indeed the CVD growth of carbona-
ceous films occurs by heterogeneous reactions of gases over
reactive surfaces and it usually requires the decomposition of
precursors over the surface of a metal at high temperature (gen-
erally above 700 K). The presence of the catalyst reduces the
barriers35 and can selectively promote the local exchange of
H and O atoms from the gas phase and the growing carbon
film. It is well known that nickel is a very active hydro-
genation catalyst for olefin and benzene double bond.36 For
example, a mechanism of GO reduction and hydrogenation
by hydrogen spillover on nickel has been proposed.37 More
recently, the easy hydrogenation of GO at room temperature
has been established using nickel nanoparticles and through
in situ generated hydrogen and enhanced spillover, producing
HGO.14
Since the diffusion of species plays a dominant role in
the kinetics of agglomeration, we observe that the formation
of fully epoxide covered graphene is very unlikely because
of the large energy barriers to the migration of the isolated
epoxide groups.23 Therefore, we have not investigated such a
structure, whose stability has been anyway studied by Wang
and co-authors.25
We also notice that by using plasmas it is possible to fur-
ther modify the chemical potentials. Plasmas can dissociate
H2 and O2 molecules into their atomic counterparts while
also increasing the chemical potential of selected chemical
species at selected surfaces by promoting the flux of charged
at (µH , µO ) = (0.248, 1.931) corresponding, at T = 400 K, ions and radicals, favouring their insertion in the growing
to PH2 = 10 atm and PO2 = 10 38 atm, while the triple point lattice.38
where GO H(8OH), graphene, and water coexist is at (µH , µO )
= (0.958, 0.514) corresponding, at T = 400 K to PH2 = 10 17
atm and PO2 = 0.02 atm.
Wang25 investigated GO structures for supercells other
than those considered in the present work, consisting of
fully epoxide- and hydroxyl-functionalized GOs, which are
not included in our phase diagram. We have found that the
4(h2 )I 1×2 structure of Wang has the same total energy of the
GO H(8OH) structure, although the orientation of the O–H
bonds is different, so that 4(h2 )I 1×2 and GO H(8OH) are equiv-
alent isomers. The shape of the GO H(8OH) phase region is dif-
ferent from that of 4(h2 )I 1×2 for two reasons: (i) the presence of
more favourite phases not considered by Wang and co-authors;
(ii) the absence of the phases 2(h2 e2 )I 1×3 and 2(h2 e)II 2×1 con-
sidered by them. Considering Fig. 8, the latter two phases
would be located on the left-hand side of the GO H(7OH)
phase, and their inclusion in the phase diagram would result
in the reduction of the regions belonging to the GO H(8OH)
and GO (2O 2OH) phases. We conclude that by increasing the
chemical potential of oxygen and at the same time decreas-
ing the chemical potential of hydrogen, the same stable phases
of the fully oxidized structures considered by Wang could be
found.
We point out that the existence of stable phases does not
imply directly that the homogeneous growth is a viable solu-
tion. Indeed, the growth not only requires the stable phases
to belong to the stability region of the phase diagram, but
it also needs to overcome the reaction barriers and have
a favourable kinetic. However, the use of catalysts could
IV. CONCLUSIONS
By using first principles calculations, we have estimated
the energy of formation of HyGH structures, i.e., graphene-
based derivatives with carbon sp3 hybridization as in graphane,
but with hydroxide groups in lieu of hydrogen atoms. We
compared the corresponding results to those of different
hydrogenated and non-hydrogenated graphene oxide deriva-
tives, functionalized with epoxide and hydroxyl groups. The
graphene oxide derivatives have been modelled by using a
hexagonal periodic cell containing 8 carbon atoms. We have
found that hydroxylation increases the average strength of the
functionalization bonds and gives rise to stable phases. The
stability of the HyGH structures had not been investigated in
previous studies.
The formation energy of HyGH [distinguished as
GO H(nh OH)] with fully sp3 hybridized C atoms decreases
towards more stable configurations by increasing nh , i.e., by
substituting H atoms with hydroxide groups. Fully HyGH,
GO H(8OH), was found the most stable among the investigated
systems, and even more stable than graphane. The HGO struc-
tures functionalized with hydroxyl and epoxy groups GO H(O
OH) and GO H(2O 2OH) have also been examined for compar-
ison. We found that they have low formation energy, but they
are less stable than most of HyGH structures. Instead, among
the non-hydrogenated GO structures with mixed sp2 and sp3
hybridization of C atoms, GO (4OH) has low formation energy,
as well as GO (2O 2OH).
104705-8 Buonocore, Capasso, and Lisi
The stability regions of the most stable structures were
also investigated in the phase diagram. It was found that stable
phases are possible only for extreme values of the chemical
potentials. The GO H(8OH) as well as GO H(nh OH) with nh
= 1, 2, 6, and 7 HyGH structures have stable phases. The
GO (2O 2OH) and GO (4OH) GO structures as well as the
GO H(6OH), GO H(7OH), and GO H(8OH) HyGH structures
with low contents of hydrogen are formed for high µO , i.e., high
O2 partial pressure, while graphane as well as GO H(1OH) and
GO H(2OH) HyGH structures with high contents of hydrogen
functionalization are formed for high µH , i.e., high H2 partial
pressure.
GO H(O OH) and GO H(2O 2OH) (the HGO structures),
do not have stable phases due to competing water forma-
tion mechanism. Instead, both GO (2O 2OH) and GO (4OH) are
GO structures with a mixed hybridization and stable phases.
Therefore, we have established that non fully covered GO
structure, mixing sp2 and sp3 hybridization of C atoms, can
form stable phases, as a different from the findings of Wang
and co-authors.25 However, those are found only at high µO
chemical potential, i.e., high partial pressure PO2 .
The stability regions of HyGH in the phase diagram
are found either for very high or very low partial pres-
sures. However, the use of catalysts could promote the for-
mation of HyGH. We observe that none of the HGO struc-
tures different from HyGH forms examined in this work has
been found to have stable phases. Remarkably, the synthe-
sis of HyGH and graphol, achieved through hydroboration of
GO followed by protonation, has been reported in a recent
work.15
The present work is limited to graphene derivatives with
a regular distribution of functional groups, but it can in prin-
ciple be extended to systems with larger periodicity, where
stable patches of sp3 bonded carbons can coexist with a sp2
lattice and can be seen as islands of stable functionalization.
Moreover, vibrational energies may affect the relative stability
of structures with small energy differences. Nonetheless, their
effects should be assessed in the next works for more accurate
determination of the thermodynamic stability.
In analogy to organic molecules, the proposed HyGH
structures will allow a fine control of the chemical interac-
tions and interfacing properties with other molecules, liq-
uids, and solids. Several fields of application, from solar
organic to biochemistry to electrochemical energy would then
benefit from the experimental development of these surface
materials.
ACKNOWLEDGMENTS
The computing resources and the related techni-
cal support used for this work have been provided by
CRESCO/ENEAGRID High Performance Computing infras-
tructure and its staff.39 CRESCO/ENEAGRID High Perfor-
mance Computing infrastructure is funded by ENEA, the
Italian National Agency for New Technologies, Energy and
Sustainable Economic Development, and by Italian and Euro-
pean research programs, see http://www.cresco.enea.it/english
for information.
J. Chem. Phys. 147, 104705 (2017)
1 M. F. Craciun, I. Khrapach, M. D. Barnes, and S. Russo, J. Phys.: Condens.
Matter 25, 423201 (2013).
2 A. Capasso, L. Salamandra, G. Faggio, T. Dikonimos, F. Buonocore,
V. Morandi, L. Ortolani, and N. Lisi, ACS Appl. Mater. Interfaces 8, 23844
(2016).
3 J. Liu, M. Durstock, and L. Dai, Energy Environ. Sci. 7, 1297 (2014).
4 D. R. Dreyer, S. Park, C. W. Bielawski, and R. S. Ruoff, Chem. Soc. Rev.
39, 228 (2010).
5 W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc. 80, 1339 (1958).
6 T. Szabo, O. Berkesi, P. Forgo, K. Josepovits, Y. Sanakis, D. Petridis, and
I. Dékány, Chem. Mater. 18, 2740 (2006).
7 Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De,
I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun’ko et al., Nat.
Nanotechnol. 3, 563 (2008).
8 H. C. Schniepp, J.-L. Li, M. J. McAllister, H. Sai, M. Herrera-Alonson,
D. H. Adamson, R. K. Prud’homme, R. Car, D. A. Seville, and I. A. Aksay,
J. Phys. Chem. B 110, 8535 (2006).
9 M. Acik, G. Lee, C. Mattevi, M. Chhowalla, K. Cho, and Y. J. Chabal, Nat.
Mater. 9, 840 (2010).
10 J. T. Robinson, F. K. Perkins, E. S. Snow, Z. Q. Wei, and P. E. Sheehan,
Nano Lett. 8, 3137 (2008).
11 S. Park, K.-S. Lee, G. Bozoklu, W. Cai, S. T. Nguyen, and R. S. Ruoff, ACS
Nano 2, 572 (2008).
12 J. T. Robinson, M. Zalalutdinov, J. W. Baldwin, E. S. Snow, Z. Wei,
P. Sheehan, and B. H. Houston, Nano Lett. 8, 3441 (2008).
13 A. Lerf, H. He, M. Forster, and J. Klinowski, J. Phys. Chem. B 102, 4477
(1998).
14 R. Krishna, E. Titus, O. Okhay, J. C. Gil, J. Ventura, E. V. Ramana, and
J. J. A. Gracio, Int. J. Electrochem. Sci. 9, 4054 (2014), ISSN: 1452-3981.
15 H. L. Poh, Z. Sofer, P. Šimek, I. Tomandl, and M. Pumera, Chem. Eur. J.
21, 8130 (2015).
16 J. Tuček, K. Holá, A. B. Bourlinos, P. Błoński, A. Bakandritsos, J. Ugolotti,
M. Dubecký, F. Karlický, V. Ranc, K. Čépe, M. Otyepka, and R. Zbořil,
Nat. Commun. 8, 14525 (2017).
17 F. Taherian, V. Marcon, and N. F. A. van der Vegt, Langmuir 29, 1457
(2013).
18 N. Wei, C. Lv, and Z. Xu, Langmuir 30, 3572 (2014).
19 W. Boukhvalov and M. I. Katsnelson, J. Am. Chem. Soc. 130, 10697 (2008).
20 A. F. Fonseca, H. Zhang, and K. Cho, Carbon 84, 365 (2015).
21 S. Zhou and A. Bongiorno, Sci. Rep. 3, 2484 (2013).
22 R. J. W. E. Lahaye, H. K. Jeong, C. Y. Park, and Y. H. Lee, Phys. Rev. B
79, 125435 (2009).
23 J.-A. Yan, L. Xian, and M. Y. Chou, Phys. Rev. Lett. 103, 086802 (2009).
24 N. Ghaderi and M. Peressi, J. Phys. Chem. C 114, 21625 (2010).
25 L. Wang, Y. Y. Sun, K. Lee, D. West, Z. F. Chen, J. J. Zhao, and S. B. Zhang,
Phys. Rev. B 82, 161406(R) (2010).
26 B. Huang, H. Xiang, Q. Xu, and S.-H. Wei, Phys. Rev. Lett. 110, 085501
(2013).
27 P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni,
D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo et al., J. Phys.: Condens.
Matter 21, 395502 (2009).
28 J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996).
29 D. Vanderbilt, Phys. Rev. B 41, 7892 (1990).
30 J. O. Sofo, A. S. Chaudhari, and G. D. Barber, Phys. Rev. B 75, 153401
(2007).
31 A. A. Maradudin, E. W. Montroll, and G. H. Weiss, Theory of Lattice
Dynamics in the Harmonic Approximation, 2nd ed. (Academic, New York,
1971).
32 E. Zurek, Rev. Comput. Chem. 29, 274 (2016).
33 K. Reuter and M. Scheffler, Phys. Rev. B 65, 035406 (2001).
34 M. W. Chase, Jr., NIST-JANAF Thermochemical Tables, 4th ed. (American
Institute of Physics, New York, 1998).
35 W. Zhang, P. Wu, Z. Li, and J. Yang, J. Phys. Chem. C 115, 17782
(2011).
36 T. W. Solomons and C. B. Fryhle, Organic Chemistry (Wiley, New York,
2004).
37 R. Krishna, E. Titus, L. C. Costa, J. C. J. M. D. S. Menezes, M. R. P. Correia,
S. Pinto, J. Ventura, J. P. Araújo, J. A. S. Cavaleiro, and J. J. A. Gracio,
J. Mater. Chem. 22, 10457 (2012).
38 Y. Wang, X. Xu, J. Lu, M. Lin, Q. Bao, B. Özyilmaz, and K. P. Loh, ACS
Nano 4, 6146 (2010).
39 G. Ponti, F. Palombi, D. Abate, F. Ambrosino, G. Aprea, T. Bastianelli,
F. Beone, R. Bertini, G. Bracco, M. Caporicci et al., Int. Conf. HPCS 2014,
14600240.