J Mol Model (2017) 23: 187

DOI 10.1007/s00894-017-3356-2

ORIGINAL PAPER

DFT study of CO2 and H2O co-adsorption on carbon

models of coal surface
Zhengyang Gao 1 & Yi Ding 1

Received: 27 February 2017 / Accepted: 24 April 2017 / Published online: 12 May 2017
# Springer-Verlag Berlin Heidelberg 2017

Abstract The moisture content of coal affects the adsorption micropore structure and functional group, it is shown that the
capacity of CO2 on the coal surface. Since the hydrogen bonds adsorbed H2O can promote CO2 adsorption on the coal sur-
are formed between H2O and oxygen functional group, the face. These results demonstrate that the micropore factor plays
H2O cluster more easily adsorbs on the coal micropore than a dominant role in affecting CO2 adsorption capacity, the at-
CO2 molecule. The coal micropores are occupied by H2O tractive interaction of adsorbed H2O to CO2 makes little
molecules that cannot provide extra space for CO2 adsorption, contribution.
which may leads to the reduction of CO2 adsorption capacity.
However, without considering factors of micropore and oxy-
Keywords CO2 adsorption . H2O adsorption . DFT
gen functional groups, the co-adsorption mechanisms of CO2
simulation . Electrostatic potential . Atoms in molecules .
and adsorbed H2O molecule are not clear. Density functional
Reduced density gradient . Energy decomposition analysis
theory (DFT) calculations were performed to elucidate the
effect of adsorbed H2O to CO2 adsorption. This study reports
some typical coal-H2O···CO2 complexes, along with a de-
tailed analysis of the geometry, energy, electrostatic potential Introduction
(ESP), atoms in molecules (AIM), reduced density gradient
(RDG), and energy decomposition analysis (EDA). The re- Carbon dioxide (CO2) as a major greenhouse gases can make
sults show that H2O molecule can more stably adsorb on the the average global surface temperature rise [1]. In recent years,
aromatic ring surface than CO2 molecule, and the absolute the vast usage of primary energy has caused an increase in the
values of local ESP maximum and minimum of H2O cluster concentration of CO2 gas [2–5]. Reducing the release of
are greater than CO2. AIM analysis shows a detailed interac- greenhouse gas CO2 into the environment has been one of
tion path and strength between atoms in CO2 and H2O, and the most urgent issues in earth environmental protection. To
RDG analysis shows that the interactions among CO2, H2O, control the emission of CO2, there have been lots of investi-
and coal model belong to weak van der Waals force. EDA gations about capturing and storing CO2 gas by metal-organic
indicates that electrostatic and long-range dispersion terms frameworks (MOFs) [6, 7], activated carbon [8–10], metal
play a primary role in the co-adsorption of CO2 and H2O. oxide [11, 12], and sequestration in coal seams [13–16].
According to the DFT calculated results without considering Since abandoned coal seams with large space have a great
potential for CO2 sequestration, this has been considered an
Electronic supplementary material The online version of this article important and valuable strategy to alleviate greenhouse envi-
(doi:10.1007/s00894-017-3356-2) contains supplementary material,
ronment problems. In the abandoned coal seams, little CH4
which is available to authorized users.
gas remains which formed by coalification. In order to en-
* Yi Ding
hance the CH4 gas recovery, injection of CO2 gas into coal
dyi0508@163.com seams has been considered a reliable strategy [17].
Experimental research of competitive adsorption of CO2 and
1
The department of Power Engineering, North China Electric Power CH4 on coal surface have shown that the adsorption capacity
University, Baoding 071003, China of CO2 molecules in carbonaceous pores is stronger than that
187 Page 2 of 17 J Mol Model (2017) 23: 187

of CH4 molecules, and CO2 injected into coal seam can pro- adsorption energy, electrostatic surface potential, AIM analy-
mote the recovery of CH4 [18–21]. sis, RDG analysis, and energy decomposition analysis were
In coal seams, the adsorption capacity and stability of the selected.
adsorbed CO2 are affected by factors of pressure, temperature,
coal rank, moisture content [22], and so on [23]. There have
been some investigations to clarify the effect of moisture on
the adsorption capacity of CO 2 gas on coal [24, 25]. Computational details
Experiments [23] show that there has been a competitive ad-
sorption between CO2 and H2O, that the amount of adsorbed All geometries involved in this study were optimized by
CO2 in a coal seam appears to increase as more water is re- BLYP density functional theory [39] with long-range disper-
moved, and thus the presence of water significantly decreases sion correction DFT-D3 using Becke-Johnson damping func-
the adsorption capacities of CO2 gas in coal. One possible tion [40] and Grimme’s gCP for BSSE correction [41], and
explanation is that the accessible micropore volume of coal basis set def2-TZVP was employed. Single-point energy of
with high moisture is less than dry coal [26], the water mole- optimized conformations was evaluated by BLYP density
cule or cluster occupies the micropore via the hydrogen inter- functional theory with DFT-D3(BJ), and basis set def2-
action with oxygen containing functional group in coal, which QZVPP(−g,−f) was adopted [42]. BLYP-D3 is an ideal meth-
leads to the reduction of CO2 adsorption. Liu et al. [27] stud- od for studying an organic large system because it balances
ied the effect of monovacancy with dissociated H2O of graph- computational efficiency and accuracy, and this method has
ite surface on the adsorption of CO2, the result indicated that proven many times to provide satisfactory results for non-
CO2 is more likely to react indirectly via a dissociated H2O covalent interactions [43–47]. Geometries optimization and
molecule than directly with the monovacancy. Lopes et al.’s single-point energy calculation were carried out by ORCA
[28] experimental results indicated that the adsorption capac- 3.0.1 [48], and in order to save the computational time,
ity of H2O is stronger than CO2 on the activated carbon resolution-of-identity (RI) technique [49] was used for accel-
surface. erating self-consistent field (SCF), and the error of accuracy
Coal is a highly complex organic polymer with various due to RI technique is negligible [50].
oxygen containing functional groups, sulfur containing func- When a single molecule adsorbed on the aromatic ring
tional groups, aromatic hydrocarbon, aliphatic hydrocarbon, surface, the adsorption energy was calculated by Eq. (1):
and moisture in pores [29–31]. In the coal seams, the poly-
meric structure of aromatic rings is mainly involved in the E ads ¼ E aromaticþmolecule −Earomatic −Emolecule ð1Þ
adsorption of coal-bed gas on the coal surface [32–35]. Xu
et al. [32] studied the competitive adsorption of CH4 and CO2 Where Earomatic + molecule is the total energy of adsorption
on 4×4 six-ring aromatic cluster. Lee et al. [35] discussed the complex of aromatic ring and gas-phase molecule, Earomatic
chemisorption and physisorption of CO2 on grapheme at var- and Emolecule are the energy of aromatic ring and gas-phase
ious levels of DFT theory. Wu et al. [36] studied the adsorp- molecule, respectively. Negative value of Eads reflects the ad-
tion of H2O on a single aromatic ring with different oxygen sorption is an exothermic progress, and higher negative values
containing functional groups based on DFT. Cabrera-Sanfelix of Eads correspond to the stronger adsorption capacity of aro-
et al. [37] investigated the adsorption of water on perfect and matic rings onto gas phase molecule [32].
defected graphite (0001) surface. Gao et al. [38] studied dif- When two kinds of molecules adsorbed on the aromatic
ferent H2O cluster adsorption on lignite model surface by ring surface, the adsorption energy was calculated as follows:
DFT. These researchers successfully revealed the adsorption E ads−A ¼ EaromaticþAþB −EaromaticþB ð2Þ
mechanisms of CO2 or H2O molecule on graphite surface at
the molecular level. Where Eads − A is the co-adsorption energy of molecule A
However, quantum chemical studies about the competitive on the aromatic ring surface which has adsorbed molecule B,
adsorption mechanisms of CO2 and H2O on coal surface at the Earomatic + A + B is the total energy of the optimized complex of
molecular level are still scarce. Without considering the influ- aromatic rings containing molecules A and B, and Earomatic + B
ence factors of micropore and oxygen containing functional is the total energy of the adsorption complex of molecule B on
group, the co-adsorption mechanisms of CO2 and H2O cluster aromatic ring surface.
have not been studied, where adsorbed H2O molecules may Electrostatic potential (ESP) analysis [51], atoms in mole-
promote or restrain CO2 adsorption on the aromatic ring sur- cules (AIM) topological analysis [52–54], and reduced densi-
face. According to the research findings, the degree of impor- ty gradient (RDG) analysis [55] were evaluated by wave func-
tance for some factors can be commendably determined. In tion analysis program of Multiwfn 3.2.1 [56]. In this paper, all
this study, to analyze the co-adsorption behavior of CO2 and isosurface maps of adsorption complexes were plotted by
H 2 O molecule, the optimized configuration structures, VMD 1.9.1 program [57]. LMO-EDA energy decomposition
J Mol Model (2017) 23: 187 Page 3 of 17 187

analysis [58] was carried out by GAMESS-US program [59] (H2O)n (n = 2, 3, 4, 8) cluster adsorbs horizontally on the
at BLYP/6-311G(d,p) level. C48H18 model surface, and the adsorption energies of these
In this study, two kinds of aromatic rings models have been configurations are similar to each other and slightly greater
used to simulate the local surface of coal, as shown in Fig. 1. than that of a single H2O molecule. The energy values of
Pyrene (C16H10) model proved in the previous study that it H2O molecule adsorption on graphene model surface are con-
can well provide the result for the adsorption of CH4 molecule sistent with the DFT calculated result by Wu [36] and exper-
on the coal surface [47]. The average length of the C-C bond imental result by Salame [61].
of 4×4 six-ring aromatic cluster (C48H18) is close to the struc- The adsorption energy curves as a function of the vertical
tural parameters of perfect graphite [32]. height of CO2 and H2O from graphene model surface are
shown in Fig. 6(a) and (b), respectively. From Fig. 6(a), it
can be found that the CO2 adsorption on C48H18 model sur-
Results and discussion face is more stable than that on C16H10 model at different
vertical distances. As shown in Fig. 6(b), curves of C16H10···
CO2 and H2O molecule adsorption on the graphene model H2O and C48H18···H2O are coincident in changing tendency,
surface and the most stable adsorption energy of different configura-
tions is similar. Along with the increasing of vertical distance,
The interaction of single CO2 on the surface of C48H18 cluster the adsorption energy decreases rapidly for C48H18···nH2O
model has been studied in detail by Xu et al. [32], they studied (n = 2, 3, 4, 8) configurations, which indicates that H2O cluster
the adsorption behavior of CO2 in two direction at three ad- is going to desorb at a relative far distance from the C48H18
sorption sites. In order to avoid repeating the same work and model surface.
pay attention to the adsorption mechanisms of H2O cluster on
the surface of grapheme, only one adsorption behavior of CO2 CO2 and H2O co-adsorption on the graphene model
in parallel direction at Bbridge^ site has been considered, as surface
shown in Figs. 2 and 3. The optimized adsorption configura-
tions of single H2O on C16H10 model surface and single H2O, The adsorption behaviors of CO2 on the C16H10-H2O and
H2O dimer, H2O trimer, H2O tetramer, H2O octamer on C48H18-nH2O (n = 1, 2, 3, 4, 8) model surface which had
C48H18 model surface have been studied, as shown in Figs 4 already adsorbed a single H2O molecule or H2O clusters have
and 5, respectively, the corresponding adsorption energies of been used to study the co-adsorption mechanism of CO2 and
these complexes are shown in Table 1. H2O. For C16H10 graphene model, co-adsorption behavior of
As shown in Table 1, the adsorption energies of CO2 mol- CO2 and single H2O has been discussed. For C48H18 graphene
ecule on C16H10 and C48H18 model are −15.6 and −17.9 kJ model, co-adsorption behaviors of CO2 and single H2O, H2O
mol−1, respectively, and the vertical adsorption distances from dimer, H2O trimer, H2O tetramer, and H2O octamer have been
CO2 to the model slab are 3.22 Å and 3.14 Å. These results are studied.
consistent with the theoretical calculation of Xu [32]. The CO2 and H2O adsorption on C16H10-H2O and C16H10-CO2
adsorption distance of CO2 on C48H18 is shorter than that of model surface at different vertical distances have been studied,
C16H10, and the adsorption energy of the former is larger than respectively. The stable adsorption configuration is shown in
the latter. CO2 adsorbs on the graphene models in horizontal Fig. 7, and the relationship of adsorption energy versus verti-
direction, and two oxygen atoms of CO2 located above the cal distance has been shown in Fig. 9(a). As shown in
centers of two adjacent benzene rings. Fig. 7(b), H2O and CO2 adsorbed vertically and horizontally
The adsorption energy of H2O molecule on C16H10 model on C16H10 model, respectively. The adsorption energies of
surface is −16.8 kJ mol−1, and on C48H18 model surface is H2O and CO2 on C16H10-CO2 and C16H10-H2O model are
−18.0 kJ mol−1. A single H2O molecule adsorbs on C16H10 −26.7 and −23.5 kJ mol−1, respectively. From Fig. 9(a), it is
and C48H18 model in the vertical direction, two H atoms of shown that the adsorption energy of H2O on the C16H10-CO2
H2O close to the two adjacent benzene rings. Since the inter- model, ranged from 3.1 to −7.9 kJ mol−1, corresponds to the
action is between the electrophilic H atom and nucleophilic π- vertical distance ranged from 2.63 to 5.63 Å. When H2O mol-
system of benzene ring, the configurations of H2O molecule ecule gets too close to C16H10-CO2 model surface, the inter-
adsorbed vertically on graphene slab are stable. The adsorp- action between H2O and C16H10-CO2 model is repulsive.
tion distance from single H2O to C48H18 slab surface is shorter With the H2O molecule far away from C16H10-CO2 model
than C16H10, and the adsorption energy of H2O on C48H18 surface the interaction changes from repulsion to attraction,
slab is slightly larger than that on C16H10 slab. Based on the with the most stable vertical distance of 3.23 Å; and the same
comparison, the adsorption energy of H2O on graphene model change regulation as shown in the adsorption of CO2 on
is greater than that of CO2, indicating that the adsorption of C16H10-H2O model surface. Compared with two adsorption
H2O on graphene model surface is more stable. Intuitively, the energy curves in Fig. 9(a), the co-adsorption energies of H2O
187 Page 4 of 17 J Mol Model (2017) 23: 187

Fig. 1 Graphene model (a)

C16H10 and (b) C48H18 to
simulate the local surface of coal

(a) (b)

on C16H10 model are greater than that of CO2 at different and the most stable adsorption vertical distance of CO2 ranges
vertical distances. Based on the comparison between co- from 3.16 to 3.27 Å and the adsorption energy ranges from
adsorption and individual adsorption of H2O and CO2, it can −20.2 to −37.8 kJ mol−1. As shown in Fig. 9(b), the adsorption
be found that the co-adsorption energy of H2O and CO2 re- energy curves of CO2 adsorbed on C48H18-2H2O, C48H18-
spectively on C16H10-CO2 and C16H10-H2O model is greater 3H2O, C48H18-4H2O, and C48H18-8H2O model surface are
than individual adsorption energy, indicating that single H2O similar in changing tendency. Comparing these configura-
or CO2 is stable in the co-adsorption system, and adsorbed tions, it can be found that the adsorption energy of CO2
CO2 or H2O can promote the adsorption of single H2O or adsorbed on C48H18-2H2O and C48H18-4H2O model surface
CO2, respectively. Similarly, the co-adsorption mechanisms are the largest and the least, respectively. In C48H18-2H2O···
of CO2 and single H2O on C48H18 graphene model have been CO2 configuration, CO2 interact with H2O not only through
studied. As shown in Fig. 8(a), the most stable configuration the attraction between C atom in CO2 and O atom in H2O, but
has the vertical distance 3.12 Å of H2O and 3.24 Å of CO2. also O atom in CO2 and H atom in H2O approximate and
The co-adsorption energies of H2O and CO2 on C48H18 attract each other by van der Waals forces. The configuration
graphene model are a little greater than on C16H10 model. (C48H18-2H2O)···CO2 is shown in Fig. 8(a2) and (b2), CO2
Comparing co-adsorption and individual adsorption energy molecule and H2O dimer at the same level, with one H atom in
of H2O and CO2 on C48H18 model, it is shown that the former H2O close to the O atom in CO2, in addition, O-H bond in
energy is greater than the latter energy. The results also show H2O aligns with O atom in CO2. Therefore, these two inter-
that adsorbed CO2 or H2O on C48H18 model can promote the action paths may lead to the stable adsorption of CO2.
adsorption of a single H2O or CO2 molecule, respectively. Whereas, in C48H18-4H2O···CO2 configuration, C atom in
This may be due to the attractive interaction between the O CO2 is far away from the O atom in H2O, and O atom in
atom in H2O and the C atom in CO2, where the O atom and C CO2 is relatively close to H atom in H2O, the interaction
atom represent electronegativity and electropositive, may just exist between O atom of CO2 and H atom of H2O
respectively. and thus the adsorption energy is relatively small. Based on
The adsorption configurations of CO2 on C48H18-nH2O the comparison of co-adsorption and individual adsorption
model are shown in Fig. 8. In these configurations, CO2 mol- energy of CO2 molecule on C16H10 and C48H18 models, it is
ecule adsorbed horizontally on C48H18 graphene model sur- shown that the former energy is greater than the latter energy,
face and located at almost the same level with H2O cluster, which indicates that CO2 adsorbed stably on co-adsorption

Fig. 2 The optimized adsorption

configuration of CO2 onto the
surface of C16H10 model

(a) (b)
J Mol Model (2017) 23: 187 Page 5 of 17 187

Fig. 3 The optimized adsorption

configuration of CO2 onto the
surface of C48H18 model

(a) (b)

system and adsorbed water clusters can promote the adsorp- ESP minimum with the value of −38.94 kcal mol−1 on the
tion of CO2 molecule on the coal surface without oxygen configuration surface. Also it is confirmed that the electrostat-
functional groups. ic potential of the π-cloud of benzene has less electronegativ-
ity than that of the lone pair of oxygen atom. The H atom in
Electrostatic potential (ESP) analysis H2O represents the global ESP maximum with the value of
22.15 kcal mol−1, and two H atoms in H2O with large positive
Electrostatic potential (ESP) on molecular surface is vital for charge are attracted by benzene ring which represents negative
studying and predicting intermolecular interaction [62, 63], charge. Since the electron transfer of carbon atom to oxygen
and it has been used to describe various properties of some atom in CO2, the C atom represents positive charge that can be
chemical systems [64, 65]. ESP calculations have been widely attracted by O atom in H2O. Similarly, as shown in the other
used to locate electrophilic and nucleophilic attack sites on three figures, the attraction of adsorbed H2O to CO2 is mainly
molecules [66]. The ESP value provides a method to quantify by the electrical attraction between C atom with positive
reactivity at several specified sites in a molecule or cluster. In charge in CO2 and O atom with negative charge in H2O.
this paper, the ESP of some typical adsorption configurations Since the H atom in adsorbed H2O and O atom in CO2 repre-
is employed for a deeper understanding of the behavior of gas- sent different electrical property, the electrical attraction may
phase molecules adsorption on the graphene model surface. be presented between these two atoms. Adsorbed H2O cluster
The ESP on molecular surface and extreme points of leads to production of more ESP maxima and minima, which
C48H18-H2O···CO2, C48H18-2H2O···CO2, C48H18-3H2O··· changes the distribution of ESP values on the molecular con-
CO2 and C48H18-4H2O···CO2 configurations are shown in figuration surface. Also, the adsorbed H2O cluster provides
Fig. 10, and the surface area in different ESP ranges are plot- more adsorption sites that can promote the adsorption of other
ted as Fig. 11. Color-filled ESP map of other configurations molecules. In addition, the absolute values of local ESP max-
are shown in Supplementary materials. Intuitively, these con- ima and minima on H2O cluster surface are greater than that of
figurations are seen to have several surface regions with neg- CO2, which indicates that the electrical attraction strength of
ative electrostatic potential, and these regions are associated H2O cluster on graphene model is stronger compared with
with the lone pairs of oxygen and π electrons of the benzene CO2 adsorbs on graphene model surface.
ring. On the contrary, the regions with positive electrostatic The molecular vdW surface of C48H18-nH2O···CO2 (n = 1,
potential are related to the hydrogen in H2O and carbon in 2, 3, 4) configurations have been partitioned into multiple
CO2. In particular, the C atom in benzene and CO2 represent fragments based on different ESP value ranges, which can
negative and positive charge, respectively. As shown in be used to analyze the characteristic of ESP distribution of
Fig. 10(a), lone pair of O atom in H2O leads to the global these configurations. As shown in Fig. 11, for C48H18-H2O···

Fig. 4 The optimized adsorption

configuration of single H2O onto
the surface of C16H10 model

(a) (b)
187 Page 6 of 17 J Mol Model (2017) 23: 187

Fig. 5 The optimized adsorption

configuration of H2O cluster onto
the surface of C48H18 model

(a1) (b1)

(a2) (b2)

(a3) (b3)

(a4) (b4)

(a5) (b5)

CO2 configuration, the largest surface area is 184.50 Å2 with the largest surface area of ESP range is associated with ben-
the occupancy of 32.97% of overall surface, corresponding to zene ring, and ESP ranges with the minimum and maximum
the ESP value ranges from −11 to −6 kcal mol−1. The molec- ESP value are related to the O and H atom in H2O, respec-
ular vdW surface areas of the minimum and maximum ESP tively. Along with the greater H2O adsorption on grahene
range are 3.58 and 2.35 Å2, respectively, which occupy 0.64% model surface, the range with largest surface area moves right
and 0.42% of overall surface. According to the ESP value and when the ESP values increase, indicating the configuration is
molecular vdW surface area value, it can be determined that relatively less electronegative. Since the H atom in H2O
J Mol Model (2017) 23: 187 Page 7 of 17 187

Table 1 Adsorption distance (Dads) and energy (Eads) of equilibrium surface of configuration, and provides more adsorption sites
configurations
for CO2 molecule.
Configurations Dads (Å) Eads (kJ mol−1)

CO2 C16H10···CO2 3.22 −15.6 Atoms in molecules (AIM) analysis

C48H18···CO2 3.14 −17.9
DFT (PBE/DNP)a 3.20 −15.8 Atoms in molecules theory has been successfully and widely
Expermientb −17.2 used to study the properties of shared-shell (covalent) and
H2O C16H10···H2O 3.31 −16.8 closed-shell bonding [66–69]. Characteristics of a bond can
C48H18···H2O 3.22 −18.0 be well evaluated by the Hessian matrix of the charge density.
C48H18···2H2O 3.20 −19.1 The three eigenvalues of Hessian matrix are λ1, λ2, and λ3
C48H18···3H2O 3.24 −20.2 (λ1 < λ2 < λ3), and two valuable parameters of the bond critical
C48H18···4H2O 3.20 −19.5 point (BCP) are the laplacian of the charge density ∇2ρ(r)
C48H18···8H2O 3.17 −21.2 (which is the sum of λ1, λ2, and λ3) and the ellipticity
DFT (M062X-D3/6–311 + G**)c −17.1 ε = λ1/λ2−1 [70]. At the bond critical point position, the sign
Expermientd −21 ± 5 of laplacian can be used to distinguish closed-shell and cova-
lent interaction, the ellipticity ε can be used to measure the
a
Taken from ref [32]. The theoretical data was calculated based on the extent of charge accumulation in a given plane [71], and sub-
C48H18 graphene model.
b
stantial bond elliipticities reflect structural instability [72].
Taken from ref [60]. The expermiental data was obtained from the
Carroll and Bader [73–75], Koch and Popelier [70] had
physisorption of CO2 on single-walled carbon nanotubes.
c proposed some criteria based on charge density topology.
Taken from ref [36]. The theoretical data was calculated based on the
adsorption of H2O on benzene. The first important criterion is the presence of a bond critical
d
Taken from ref [61]. The expermiental data was obtained from the point (BCP) and the bond path between atoms A and B. The
isosteric heats of water adsorption on an activated carbon surface. laplacian value of the charge density of the BCP is zero
(∑3i¼1 λi ¼ 0 ), and the bond path is linear. The second criteri-
represents positive charge, the adsorption of more H2O mole- on is the linear positive correlation between charge density at
cules leads to the production of the molecular vdW surface BCP and the binding energy for H-bond, ionic bond, and
range with large positive value (> 12 kcal mol−1). covalent bond. The third criterion is the relationship between
Stated thus, in these configurations, the increasing of elec- the sign of laplacian of the charge density at the BCP and the
tronegativity is in going from π-cloud over oxygen in CO2 to interaction type of bond. The positive and negative sign of
oxygen in H2O. The electropositivity of hydrogen in H2O is laplacian at the BCP are indicative of closed-shell interaction
greater than carbon in CO2. The ESP values of maxima of and covalent interaction between two monomers, respectively.
H2O cluster are greater than that of CO2, which indicates the However, Cremer et al. [76] suggested that the analysis of
electrical attraction between graphene model and H2O mole- characteristics of bond just based on the Laplacian of charge
cule is stronger than that between graphene model and CO2. In density is unreliability and even may lead to erroneous con-
addition, the adsorbed H2O cluster leads to the presence of clusions in some structures. They suggested that the potential
more ESP maximum and minimum on the molecular vdW energy density V(r), Lagrangian kinetic energy G(r), and total

Fig. 6 Adsorption energy curves

as a function of the height of the
(a) CO2 and (b) H2O cluster from
the C16H10 and C48H18 model
surface

(a) (b)
187 Page 8 of 17 J Mol Model (2017) 23: 187

Fig. 7 The optimized adsorption

configuration of CO2 onto the
surface of C16H10-H2O model

(a) (b)

energy density H(r) (H = V + G) at the BCP should been The electron density at the BCP can be used to measure the
considered. For covalent interaction, the potential energy den- strength of interaction bond, the greater the electron density,
sity V(r) plays a dominant role, where |V(r)| > G(r) and H(r) is the stronger the interaction between two atoms. From Table 2,
negative. On the contrary, for non-shared interaction, the it can be found that the electron density ρ values at BCP of
Lagrangian kinetic energy G(r) plays a primary role, where C(CO2)···C bond in (C48H18-nH2O)···CO2 (n = 2, 3, 4, 8)
|V(r)| < G(r) and H(r) is positive. Espinoza et al. [77] proposed complexes are approximate, and (C48H18-2H2O)···CO2 com-
a similar criterion based on the ratio |V(r)|/G(r), that |V(r)|/ plex has the greatest value of electron density. The ρ value at
G(r) < 1 for non-shared interaction and |V(r)|/G(r) > 2 for BCP of C(CO2)···C bond in (C16H10-H2O)···CO2 complex is
covalent interaction. relatively small. Furthermore, the ρ values of C(CO2)···
Electron density topologies of the C16H10···CO2, C48H18··· O(H 2 O) bond are even two times larger than that of
CO2, (C16H10-H2O)···CO2 and (C48H18-H2O)···CO2 complex C(CO2)···C bond in (C48H18-nH2O)···CO2 (n = 2, 3, 4, 8)
are shown in Fig. 12, and other complexes are shown in complexes. In particular, the ρ value of O(CO2)···H(H2O)
Supplementary materials. In Fig. 12(a) and (b), a bond critical bond in (C48H18-2H2O)···CO2 complex is greater than other
point (BCP) is found between the C atom of CO2 and the C-C bonds. Similarly, the ρ values at BCP of C(CO2)···C-C bond in
bond of graphene model, which confirms the presence of C16H10···CO2 and C48H18···CO2 complexes are almost similar
C16H10···C(CO2) and C48H18···C(CO2) bonds in these com- to the ρ value of C(CO2)···C bond in other complexes. These
plexes. Similarly, as shown in Fig. 12(c) and (d), BCPs are results indicate that the interaction strength of CO2 molecule
found between C(CO2) atom and C(graphene model) atom, and graphene model slab is almost similar, but weaker than the
O(CO2) atom and C(graphene model) atom, C(CO2) atom and interaction between C atom in CO2 and O atom in adsorbed
O(H2O) atom, H(H2O) atom and C(graphene model) atom, H2O. Therefore, the adsorbed H2O cluster enhances the ad-
which indicates the presence of C16H10···X (X = C/O/H) and sorption of CO2 on graphene model. The greater ρ values of
C48H18···X (X = C/O/H) bonds in these two complexes. interaction bonds in (C48H18-2H2O)···CO2 complex reveal
Electron density values ρ(r), Laplacian of electron density that the CO2 can stably adsorb on graphene model with strong
values ∇2ρ(r), Lagrangian kinetic energy G(r), potential ener- interaction, which is consistent with the greater calculated
gy density V(r), total energy density H(r), the eigenvalues adsorption energy in CO2 and H2O co-adsorption on the
(λ1 < λ2 < 0) of the Hessian of the electron density and the graphene model surface.
ellipticity at the BCP of some typical configurations are The bond ellipticity ε provides a measure for its stability.
shown in Table 2. For all the configurations, the ∇2ρ values From Table 2 the ε value at BCP of C(CO2)···C-C bond in
are positive, which is characteristic of non-shared interaction C48H18···CO2 configuration is a little greater than that of
based on Koch criteria. According to the Cremer criteria, C16H10···CO2 configuration, indicating that the former bond
|V(r)| < G(r) and positive H(r) of primary interaction bonds is more stable than the latter. This result is consistent with the
of all complexes, thereby confirming the presence of non- order of value size of electron density ρ and adsorption energy.
shared interaction between single CO2 and graphene model A similar analysis is conducted for other type bonds, the de-
or water cluster adsorbed graphene model. Similarly, the creasing order of ellipticity ε for C(CO2)···C bond in different
values of |V(r)|/G(r) at the BCP of interaction bonds are less configurations is: (C16H10-H2O)···CO2 > (C48H18-4H2O)···
than 1, which confirms that the bonding is a non-shared inter- CO 2 > (C 4 8 H 1 8 -8H 2 O)···CO 2 > (C 4 8 H 1 8 -3H 2 O)···
action based on Espinoza criteria. These results indicate that CO2 > (C48H18-2H2O)···CO2 configuration, for C(CO2)···
the adsorption of CO2 on graphene slab belongs to the non- O(H2O) bond is: (C48H18-8H2O)···CO2 > (C48H18-3H2O)···
shared interaction, and van der Waals force plays a primary CO2 > (C16H10-H2O)···CO2 > (C48H18-H2O)···
role in the adsorption. CO2 > (C48H18-2H2O)···CO2 configuration, for O(CO2)···
J Mol Model (2017) 23: 187 Page 9 of 17 187

Fig. 8 The optimized adsorption

configuration of CO2 onto the
surface of C48H18-nH2O (n = 1, 2,
3, 4, 8) cluster model

(a1) (b1)

(a2) (b2)

(a3) (b3)

(a4) (b4)

(a5) (b5)
187 Page 10 of 17 J Mol Model (2017) 23: 187

Fig. 9 Adsorption energy curves

as a function of the height of the
single H2O or CO2 from the (a)
C16H10-H2O/CO2 and (b)
C48H18-H2O/CO2 complex model
surface

(a) (b)

H(H2O) bond is: (C48H18-3H2O)···CO2 > (C48H18-4H2O)··· The bonding of C(CO2)···graphene model is relatively weaker
CO2 > (C48H18-2H2O)···CO2 configuration. According to than that of C(CO2)···O(H2O) in the same configuration.
the order of configurations, the relative stability of interaction
bond can be determined, where the greater ε corresponds to Reduced density gradient (RDG) analysis
the relatively weak bond. In Table 2 it can be seen that the
ellipticity of O(CO2)···H(H2O) bond in (C48H18-2H2O)···CO2 As mentioned in Atoms in molecules (AIM) analysis section,
configuration is less than that of (C48H18-3H2O)···CO2 con- the sign of Laplacian of the charge density ∇2ρ(r) can be
figuration, conversely, the electron density value of the former applied to distinguish between different kinds of interactions.
is greater than the latter, confirming that the latter bond is Furthermore, the sign of eigenvalue λ2 can be used to distin-
relatively weak. guish bonded (λ2 < 0) from nonbonded (λ2 > 0) interaction,
Stated thus, based on the analysis of properties of BCP in and the charge density ρ provides information about their
interaction bond, it can be concluded that the adsorption of strength [55]. The reduced density gradient (RDG) analysis
CO2 on graphene model is a non-shared interaction, and spe- method was widely used to make a deeper study and give a
cific interaction paths between different atoms are determined. more intuitive graph for closed-shell interactions, like hydro-
In addition, the detailed relative strength of different bonds in gen bonding and non-covalent bonding. In the theory of RDG
configurations has been presented. For different adsorption analysis, which was proposed by Johnson et al. [55], the re-
configurations, the interaction strength between C atom in duced density gradient (s = 1/(2(3π2)1/3)|∇ρ|/ρ4/3) plays a role
CO2 and in graphene model is approximate to each other. in characterizing the deviation from a homogeneous electron

Fig. 10 Color-filled ESP map of

(a) C48H18-H2O···CO2, (b)
C48H18-2H2O···CO2, (c) C48H18-
3H2O···CO2 and (d) C48H18-
4H2O···CO2 configuration. The
unit is in kcal mol-1. The local
maximum and minimum points of
ESP are represented as orange and
cyan spheres. The global
maximum and minimum values
are highlighted by bold and italic
font (a) (b)

(c) (d)
J Mol Model (2017) 23: 187 Page 11 of 17 187

presented in the space among CO2, H2O molecules, and

graphene model, which indicates that the interaction of CO2
and graphene model is weak, and van der Waals weak inter-
action exists between the CO2 and H2O molecules. Red color
isosurfaces represent the space steric effect in the center of
benzene rings. The scatter plots in Fig. 13(a2) show below
the isoline with the RDG value of 0.5, two primary
sign(λ2)ρ spikes and another spike lying at near zero and
0.022 value of axis, respectively, corresponding to the van
der Waals weak interaction and steric. Similarly as shown in
Fig. 13(b1), the RDG color-filled map shows that the interac-
tion among CO2, H2O, and C48H18 model is weak, and the
attraction between CO2 and H2O is through the weak interac-
tion of C atom in CO2 and O atom in H2O. Since the steric has
Fig. 11 Surface area in each ESP range on the vdW surface of C48H18- not been considered in this configuration, the corresponding
nH2O···CO2 (n = 1, 2, 3, 4) configurations scatter plots do not show sign(λ2)ρ spike with large positive
value of axis. As shown in Fig. 13(c1) and (d1), blue color
distribution [78], and large negative and positive values of isosurfaces are presented between the H2O molecules, which
sign(λ2)ρ are indicative of attractive and repulsive interac- indicates the interaction between H2O molecules is hydrogen
tions, respectively. In this paper, RDG analysis was carried bond. In the corresponding scatter plots, the sign(λ2)ρ spikes
out to visualize the interactions between gas-phase molecules lying at relative large negative value of axis (< −0.02) confirm
and graphene model. the existence of hydrogen bond in these configurations.
Color-filled RDG map and corresponding scatter plots of Compared with other configurations, especially in
(C16H10-H2O)···CO2, (C48H18-H2O)···CO2, (C48H18-2H2O)··· Fig. 13(c1), green color isosurface is presented in the space
CO2, and (C48H18-3H2O)···CO2 configuration are shown in between O atom in CO2 and H atom in H2O, which indicates
Fig. 13, other configurations are shown in Supplementary that the H2O cluster attracts CO2 molecule through the inter-
materials. In the scatter plots, the sign(λ2)ρ spikes lying at action not only between C atom in CO2 and O atom in H2O,
large negative, near zero, and large positive values of axis, but also between O atom in CO2 and H atom in H2O. Stated
which indicates strong attraction (hydrogen bond), weak in- thus, these results confirm that the adsorbed H2O molecule or
teraction (van der Waals force), and strong repulsion (steric), cluster enhances the adsorption of CO2 via van der Waals
respectively. From Fig. 13(a1), green color isosurfaces are weak force.

Fig. 12 Electron density

topologies for (a) C16H10···CO2,
(b) C48H18···CO2, (c) (C16H10-
H2O)···CO2 and (d) (C48H18-
H2O)···CO2 complex. Bond
critical points (BCP) are shown in
orange, paths are shown in blue

(a) (b)

(c) (d)
187 Page 12 of 17 J Mol Model (2017) 23: 187

Table 2 Topological parameters of the BCP of bond in adsorption configurations

Configurations Bond a ρ ∇2ρ G V H V/G λ1 λ2 εb

C16H10···CO2 C(CO2)···C-C 0.0057 0.0184 0.0037 −0.0028 0.0009 0.7621 −0.0020 −0.0013 0.5353
C48H18···CO2 C(CO2)···C-C 0.0065 0.0198 0.0040 −0.0031 0.0009 0.7726 −0.0019 −0.0013 0.4650
(C16H10-H2O)···CO2 C(CO2)···C 0.0042 0.0147 0.0029 −0.0021 0.0008 0.7270 −0.0024 −0.0011 1.1818
C(CO2)···O(H2O) 0.0098 0.0348 0.0072 −0.0057 0.0015 0.7900 −0.0061 −0.0034 0.7941
(C48H18-H2O)···CO2 O(CO2)···C 0.0051 0.0154 0.0032 −0.0025 0.0007 0.7822 −0.0031 −0.0015 1.0667
C(CO2)···O(H2O) 0.0107 0.0397 0.0085 −0.0071 0.0014 0.8347 −0.0086 −0.0062 0.3871
(C48H18-2H2O)···CO2 C(CO2)···C 0.0058 0.0182 0.0037 −0.0028 0.0009 0.7657 −0.0022 −0.0013 0.6923
O(CO2)···H(H2O) 0.0119 0.0493 0.0096 −0.0069 0.0027 0.7197 −0.0138 −0.0132 0.0426
C(CO2)···O(H2O) 0.0110 0.0376 0.0078 −0.0063 0.0016 0.7990 −0.0068 −0.0052 0.3077
(C48H18-3H2O)···CO2 C(CO2)···C 0.0056 0.0187 0.0038 −0.0029 0.0009 0.7708 −0.0023 −0.0012 0.8544
O(CO2)···H(H2O) 0.0037 0.0138 0.0026 −0.0018 0.0008 0.6788 −0.0029 −0.0013 1.1834
C(CO2)···O(H2O) 0.0087 0.0259 0.0052 −0.0039 0.0013 0.7549 −0.0067 −0.0034 0.9706
(C48H18-4H2O)···CO2 C(CO2)···C 0.0052 0.0193 0.0037 −0.0025 0.0012 0.6856 −0.0029 −0.0014 1.0714
O(CO2)···C 0.0056 0.0179 0.0038 −0.0031 0.0007 0.8204 −0.0035 −0.0016 1.1139
O(CO2)···H(H2O) 0.0051 0.0215 0.0044 −0.0034 0.0010 0.7736 −0.0064 −0.0033 0.9394
(C48H18-8H2O)···CO2 C(CO2)···C 0.0054 0.0199 0.0037 −0.0025 0.0013 0.6651 −0.0031 −0.0016 0.9375
O(CO2)···C 0.0054 0.0168 0.0035 −0.0029 0.0007 0.8115 −0.0035 −0.0014 1.5213
C(CO2)···O(H2O) 0.0079 0.0267 0.0054 −0.0041 0.0013 0.7612 −0.0071 −0.0034 1.0882
a
In bond column, C atom without notes is in C16H10 or C48H18 model.
b
ε = λ1/λ2–1 denotes the ellipticity at the BCP

Energy decomposition analysis (EDA) the calculated results of adsorption energy via EDA are accu-
rate. In addition, the values of electrostatic, repulsion, polari-
The energy decomposition analysis (EDA) method is often zation, and dispersion energy terms decrease along with in-
used to study the physical origins of different energy compo- creasing vertical distance. The values of exchange energy
nents of total intermolecular interaction energy, and it can change from negative to positive and gradually tend to zero,
separate the total interaction energy into various terms such that is the most stable configuration has the greatest value of
as electrostatic, exchange repulsion, polarization, and disper- exchange term. These results indicate that the adsorption dis-
sion energy. Electrostatic interaction results from the homoge- tance affects the value of every energy term, the shorter verti-
neous electron distribution of two monomers, and this inter- cal distance the greatest absolute value of energy terms except
action term may be either attractive or repulsive. Whereas exchange energy. The values of electrostatic term change sig-
exchange repulsion interaction is caused by exchange of elec- nificantly along with the increasing of vertical distances.
trons between two monomers, and it belongs to short-range Dispersion energy with the greatest negative value indicates
repulsive interaction. Polarization interaction results from the that this interaction is attractive and dominant, and this long-
effect of distorted electron distribution of two monomers, and range interaction plays a primary role in the adsorption of H2O
this interaction is always attractive [79]. Dispersion is caused on C16H10 model surface. Electrostatic energy with the nega-
by the interaction of two neutral monomers with a non- tive value shows that this interaction plays an attractive role in
overlapping electron density, and this interaction is always the adsorption. The repulsion exchange energy term is invari-
attractive [80]. ably positive, similarly, this interaction with positive value
Energy decomposition analysis was used to study the ad- plays a repulsive role in the adsorption. The polarization en-
sorption of H2O and CO2 on the C16H10 model surface, and ergy term with the least negative value means that this
the results are shown in Fig. 14 (a) and (b), respectively. The
results of co-adsorption energy decomposition analysis of
Fig. 13 Color-filled reduced density gradient (RDG) map of optimized„
H2O on C16H10-CO2 model and CO2 on C16H10-H2O model conformations and corresponding scatter plots: (a) (C16H10-H2O)···CO2,
are shown in Fig. 14 (c) and (d). From Fig. 14 (a), it can be (b) (C 48 H 18 -H 2 O)···CO 2 , (c) (C 48 H 18 -2H 2O)···CO 2 , (d) (C 48 H 18 -
found that the values of Eads (calculated at BLYP-D3/def2- 3H2O)···CO2. The value of isosurface of RDG is set to 0.50. The value
of sign(λ2)ρ on the surfaces is represented by filling color according to the
QZVPP(−g,−f) level) are almost consistent with total energy
color bar at bottom. Scatter plots of reduced density gradient (RDG)
(calculated via the sum of every energy terms based on EDA) versus electron density (ρ) multiplied by the sign of the second Hessian
at different vertical distances of H2O molecule, indicating that eigenvalue (λ2)
J Mol Model (2017) 23: 187 Page 13 of 17 187

(a1) (a2)

(b1) (b2)

(c1) (c2)

(d1) (d2)
187 Page 14 of 17 J Mol Model (2017) 23: 187

(a) (b)

(c) (d)
Fig. 14 Various terms of interaction energy curves as a function of the at BLYP/6-311G(d,p) level is attributable to other energy. Accurate
height of the (a) H2O and (b) CO2 from the C16H10 model surface, (c) dispersion energy is calculated by ORCA program at BLYP-D3/def2-
H2O from the C16H10-CO2 model surface, (d) CO2 from the C16H10-H2O QZVPP(−g,-f). Total (EDA) energy is the sum of electrostatic,
model surface. Since the BLYP functional is unable to calculate the exchange, repulsion, polarization, other, and dispersion energy. Eads is
dispersion effect, ‘dispersion energy’ obtained via LMO-EDA program calculated via ORCA program at BLYP-D3/def2-QZVPP(−g,-f) level

interaction makes less contribution to the attraction. Similar contribute to the adsorption. Similar change regulations have
change regulation for CO2 adsorption on C16H10 model is been found based on the comparison of Fig. 14(d) and (b). In
shown in Fig. 14(b), the energy curve of total energy fits well Fig. 14(d), it is shown that the values of the electrostatic term
to Eads. A comparison of Fig. 14(a) and (b) shows that the are almost the same as the dispersion term.
values of energy term of H2O adsorption are almost similar to Stated thus, the electrostatic, polarization, and dispersion
the CO2 adsorption except dispersion energy. As is evident, energy terms play an attractive role, on the contrary, the ex-
the dispersion energy term of CO2 adsorption on C16H10 mod- change and repulsion energy terms play a repulsive role in the
el are smaller than that of H2O adsorption. Since H2O is a adsorption. The absolute values of energy terms decrease
polar molecule, the long-range interaction of dispersion force along with the increase of vertical distance, especially for
between H2O and graphene slab is stronger than that between the electrostatic energy term which changes most significant-
CO2 and graphene model. ly. Long-range interaction of dispersion force plays an impor-
Comparing Fig. 14(c) with (a), the energy terms of H2O tant role in the adsorption of H2O and CO2 on graphene model
adsorbed on C16H10-CO2 model surface at different vertical slab.
distances are little greater than that of H2O adsorbed on
C 16 H 10 model except exchange term. Additionally, in
Fig. 14(c), the values of the electrostatic energy term are the Conclusions
greatest compared with other energy terms. These results in-
dicate that adsorbed CO2 not only influence the adsorption Abandoned coal seams with large space have a great potential
behavior of H2O, but also affect the values of interaction en- for CO2 sequestration, but the presence of moisture may affect
ergy terms. In particular, the electrostatic term plays a domi- the adsorption of CO2 molecule. To afford more knowledge of
nant role in the attractive interaction, and the values of ex- the co-adsorption mechanisms of CO2 and H2O molecules on
change term close to zero in this interaction almost do not coal surface, and identify the important degree of influence
J Mol Model (2017) 23: 187 Page 15 of 17 187

factors of adsorbed H2O to CO2 adsorption, some typical ad- interaction between O atom in CO2 and H atom in H2O is
sorption configurations have been investigated by DFT in this vdW force, hydrogen bond is not formed.
study. Electrostatic surface potential (ESP) is evaluated to EDA shows that electrostatic, polarization, dispersion
study the electric property of the molecular vdW surface of terms and exchange, repulsion terms play an attractive and
some configurations. Atoms in molecules (AIM) theory is repulsive role in the adsorption, respectively. Along with the
applied to study the properties of specific interaction paths increasing of the vertical distance of CO2 molecule from
between interactional atoms. Reduced density gradient C16H10 model surface, all energy terms are decreasing and
(RDG) is used to analyze and visualize the interaction among are gradually close to zero. For CO2 or H2O molecule that
the CO2, H2O, and graphene models. Energy decomposition individually adsorbs on graphene model, long-range interac-
analysis (EDA) is applied to decompose the interaction energy tion of dispersion force plays a primary role in the attraction;
into several terms and analyze the contribution of each energy and for CO2 and H2O co-adsorption on graphene model, elec-
term. trostatic and dispersion terms play a major role in the
DFT study indicates that, CO2 molecule can stably co- adsorption.
adsorb with H2O on the graphene model surface. The adsorp- The co-adsorption mechanisms of CO2 and H2O on aro-
tion of CO2 on C48H18 model is stronger than that on C16H10 matic ring surface indicate that the adsorbed H2O molecule
model surface, similarly, single H2O molecule adsorption on can enhance the adsorption of CO2 via the van der Waals
C48H18 model is more stable. H2O molecule can more stably attraction interaction. Whereas in macroscopical experiments
adsorb on C16H10 and C48H18 model surface than CO2 mole- of CO2 adsorption on coal with different moisture content, the
cule. Co-adsorption energies of CO2 on C48H18-nH2O (n = 1, CO2 adsorption capacity of moist coal is lower than that of dry
2, 3, 4, 8) model are greater than CO2 individual adsorption on coal, intuitively, the DFT result is adverse to the experiment
C48H18 model, which indicates that adsorbed H2O cluster can result. For actual CO2 adsorption on moist coal seam, the
enhance the adsorption of CO2. The adsorption energy of CO2 factors of coal micropore, oxygen containing functional group
on C48H18-2H2O model is the largest compared with other and interaction between H2O and CO2 affect the CO2 adsorp-
configurations, in C48H18-2H2O···CO2 configuration, the in- tion. In detail, H2O more easily adsorbs on oxygen functional
teractions are presented not only between C atom in CO2 and groups via hydrogen bond, and then more H2O molecules
O atom in H2O, but also between O atom in CO2 and H atom participate in the adsorption, finally, H2O clusters are formed
in H2O dimer, indicating that the more atoms participate in the on the coal surface and may occupy the micropores. In addi-
interaction, the more stable for CO2 adsorption on graphene tion, the interaction between nonpolar CO2 and coal surface is
model surface. relatively weak, the micropores with specific structure can
ESP analysis shows that the π-cloud of benzene ring sur- enhance CO2 storage, however, large numbers of micropoers
face represents negative charge, carbon atom in CO2 repre- have been occupied by H2O clusters that do not provide extra
sents positive charge, hydrogen and oxygen atom in H2O space for adsorbing CO2. Therefore, the CO2 adsorption ca-
cluster represent global maximum positive and minimum neg- pacity of coal seam decreases along with increasing H2O mol-
ative electrostatic potential, respectively. The adsorption of ecules adsorption. The experiment results may can be ex-
H2O cluster on graphene model surface is mainly through plained by above elaboration, but without considering the fac-
the electric attraction between hydrogen atoms in H2O and tors of micropoers and functional groups, the DFT study
graphene model. Carbon atom in CO2 with positive charge shows that the adsorbed H2O cluster can promote the CO2
can be attracted by oxygen atom in CO2 and the π-cloud of adsorption on aromatic ring surface, which indicates that the
benzene ring. The adsorption of H2O cluster leads to a more micropore and functional group factors play a more important
electropositive molecular vdW surface of configurations. and primary role in adsorbing CO2, the attraction interaction
AIM analysis shows that the interactions among CO2, H2O of adsorbed H2O to CO2 makes little contribution. In future
molecule, and graphene model belong to non-shared interac- meticulous study, the micropore and oxygen containing func-
tion. In the same configuration, the interaction between carbon tional group factors can be considered to study co-adsorption
atom in CO2 and graphene model is weaker than that of of CO2 and H2O on different rank coal model surface at the
C(CO2)···O(H2O) bond, confirming the adsorbed H2O cluster molecular level.
enhances the CO2 adsorption on graphene model surface. For
different adsorption configurations, the interactions between
carbon atom in CO2 and graphene model are similar in
strength. References
RDG analysis indicates that the interactions among CO2,
H2O, and graphene model are mainly weak van der Waals 1. Service RF (2004) The carbon conundrum. Science 305:962-963
force, which is in accord with the relatively small adsorption 2. Pérez-Lombard L, Ortiz J, Pout C (2008) A review on buildings
energies of different equilibrium configurations. The energy consumption information. Energ Buildings 40:394–398
187 Page 16 of 17 J Mol Model (2017) 23: 187

3. Roca J, Alcántara V (2001) Energy intensity, CO 2 emissions and 25. Day S, Sakurovs R, Weir S (2008) Supercritical gas sorption on
the environmental Kuznets curve. The Spanish case. Energ Policy moist coals. Int J Coal Geol 74:203–214
29:553–556 26. Krooss BV, Van Bergen F, Gensterblum Y, Siemons N, Pagnier H,
4. Marland G, Turhollow AF (1991) CO2 emissions from the produc- David P (2002) High-pressure methane and carbon dioxide adsorp-
tion and combustion of fuel ethanol from corn. Energy 16:1307– tion on dry and moisture-equilibrated Pennsylvanian coals. Int J
1316 Coal Geol 51:69–92
5. Berndes G, Hoogwijk M, van den Broek R (2003) The contribution 27. Liu Y, Wilcox J (2012) Effects of surface heterogeneity on the
of biomass in the future global energy supply: a review of 17 stud- adsorption of CO2 in microporous carbons. Environ Sci Technol
ies. Biomass Bioenergy 25:1–28 46:1940–1947
6. Cui P, Ma Y, Li H, Zhao B, Li J, Cheng P, Balbuena PB, Zhou H 28. Lopes FV, Grande CA, Ribeiro AM, Loureiro JM, Evaggelos O,
(2012) Multipoint interactions enhanced CO2 uptake: a zeolite-like Nikolakis V, Rodrigues AE (2009) Adsorption of H2, CO2, CH4,
zinc–tetrazole framework with 24-nuclear zinc cages. J Am Chem CO, N2 and H2O in activated carbon and zeolite for hydrogen
Soc 134:18892–18895 production. Sep Sci Technol 44:1045–1073
7. Li J, Sculley J, Zhou H (2011) Metal–organic frameworks for sep- 29. Haenel MW (1992) Recent progress in coal structure research. Fuel
arations. Chem Rev 112:869–932 71:1211–1223
8. Jiang J, Sandler SI (2005) Separation of CO2 and N2 by adsorption 30. Yu J, Tahmasebi A, Han Y, Yin F, Li X (2013) A review on water in
in C168 schwarzite: a combination of quantum mechanics and mo- low rank coals: the existence, interaction with coal structure and
lecular simulation study. J Am Chem Soc 127:11989–11997 effects on coal utilization. Fuel Process Technol 106:9–20
9. Guo B, Chang L, Xie K (2006) Adsorption of carbon dioxide on 31. Salmas CE, Tsetsekou AH, Hatzilyberis KS, Androutsopoulos GP
activated carbon. J Nat Gas Chem 15:223–229 (2001) Evolution lignite mesopore structure during drying. Effect of
10. Somy A, Mehrnia MR, Amrei HD, Ghanizadeh A, Safari M (2009) temperature and heating time. Dry Technol 19:35–64
Adsorption of carbon dioxide using impregnated activated carbon 32. Xu H, Chu W, Huang X, Sun W, Jiang C, Liu Z (2016) CO 2
promoted by zinc. Int J Greenhouse Gas Control 3:249–254 adsorption-assisted CH 4 desorption on carbon models of coal sur-
11. Horiuchi T, Hidaka H, Fukui T, Kubo Y, Horio M, Suzuki K, Mori face: a DFT study. Appl Surf Sci 375:196–206
T (1998) Effect of added basic metal oxides on CO 2 adsorption on 33. Cabrera-Sanfelix P (2008) Adsorption and reactivity of CO2 on
alumina at elevated temperatures. Appl Catal A Gen 167:195–202 defective graphene sheets. J Phys Chem A 113:493–498
12. Baltrusaitis J, Schuttlefield J, Zeitler E, Grassian VH (2011) Carbon 34. Yang S, Ouyang L, Phillips JM, Ching WY (2006) Density-
dioxide adsorption on oxide nanoparticle surfaces. Chem Eng J
functional calculation of methane adsorption on graphite (0001).
170:471–481
Phys Rev B 73:165407
13. Lackner KS (2003) A guide to CO2 sequestration. Science 300:
35. Lee K, Kim S (2013) Theoretical investigation of CO2 adsorption
1677–1678
on graphene. Bull Kor Chem Soc 34:3022–3026
14. Huijgen WJ, Witkamp G, Comans RN (2005) Mineral CO2 seques-
36. Wu J, Wang J, Liu J, Yang Y, Cheng J, Wang Z, Zhou J, Cen K
tration by steel slag carbonation. Environ Sci Technol 39:9676–
(2017) Moisture removal mechanism of low-rank coal by hydro-
9682
thermal dewatering: physicochemical property analysis and DFT
15. Benson SM, Cole DR (2008) CO2 sequestration in deep sedimen-
calculation. Fuel 187:242–249
tary formations. Elements 4:325–331
37. Cabrera-Sanfelix P, Darling GR (2007) Dissociative adsorption of
16. Gale J, Freund P (2001) Coal-bed methane enhancement with CO2
water at vacancy defects in graphite. J Phys Chem C 111:18258–
sequestration worldwide potential. Environ. Geosci. 8:210–217
18263
17. Stevens SH, Spector D, Riemer P (1998) Enhanced coalbed meth-
ane recovery using CO2 injection: worldwide resource and CO2 38. Gao Z, Ding Y, Yang W, Han W (2017) DFT study of water ad-
sequestration potential. SPE International Oil and Gas Conference sorption on lignite molecule surface. J Mol Model 23:27
and Exhibition in China Society of Petroleum Engineers 39. Stephens PJ, Devlin FJ, Chabalowski C, Frisch MJ (1994) Ab initio
18. Shi J, Mazumder S, Wolf K, Durucan S (2008) Competitive meth- calculation of vibrational absorption and circular dichroism spectra
ane desorption by supercritical CO2 injection in coal. Transp using density functional force fields. J Phys Chem 98:11623–11627
Porous Media 75:35–54 40. Grimme S, Ehrlich S, Goerigk L (2011) Effect of the damping
19. Kowalczyk P, Gauden PA, Terzyk AP, Furmaniak S, Harris PJ function in dispersion corrected density functional theory. J
(2012) Displacement of methane by coadsorbed carbon dioxide is Comput Chem 32:1456–1465
facilitated in narrow carbon nanopores. J Phys Chem C 116:13640– 41. Kruse H, Grimme S (2012) A geometrical correction for the inter-
13649 and intra-molecular basis set superposition error in Hartree-Fock
20. Busch A, Gensterblum Y, Krooss BM, Siemons N (2006) and density functional theory calculations for large systems. J
Investigation of high-pressure selective adsorption/desorption be- Chem Phys 136:154101
haviour of CO 2 and CH 4 on coals: an experimental study. Int J 42. Weigend F, Ahlrichs R (2005) Balanced basis sets of split valence,
Coal Geol 66:53–68 triple zeta valence and quadruple zeta valence quality for H to Rn:
21. Ceglarska-Stefańska G, Zarębska K (2002) The competitive sorp- design and assessment of accuracy. Phys Chem Chem Phys 7:
tion of CO 2 and CH 4 with regard to the release of methane from 3297–3305
coal. Fuel Process Technol 77:423–429 43. Sagan F, Filas R, Mitoraj MP (2016) Non-covalent interactions in
22. Švábová M, Weishauptová Z, Přibyl O (2011) Water vapour ad- hydrogen storage materials LiN (CH3) 2BH3 and KN (CH3)
sorption on coal. Fuel 90:1892–1899 2BH3. Crystals 6:28
23. Ozdemir E, Schroeder K (2009) Effect of moisture on adsorption 44. Tian Z, Dai S, Jiang DE (2016) What can molecular simulation do
isotherms and adsorption capacities of CO2 on coals. Energy Fuel for global warming? Comput Mol Sci doi: 10.1002/wcms.1241
23:2821–2831 45. Mo J, Xue Y, Liu X, Qiu N, Chu W, Xie H (2013) Quantum chem-
24. Goodman AL, Busch A, Bustin RM, Chikatamarla L, Day S, Duffy ical studies on adsorption of CO 2 on nitrogen-containing molecular
GJ, Fitzgerald JE, Gasem K, Gensterblum Y, Hartman C (2007) segment models of coal. Surf Sci 616:85–92
Inter-laboratory comparison II: CO 2 isotherms measured on 46. Gao W, Feng H, Xuan X, Chen L (2012) The assessment and
moisture-equilibrated Argonne premium coals at 55 C and up to application of an approach to noncovalent interactions: the energy
15 MPa. Int J Coal Geol 72:153–164 decomposition analysis (EDA) in combination with DFT of revised
J Mol Model (2017) 23: 187 Page 17 of 17 187

dispersion correction (DFT-D3) with slater-type orbital (STO) basis 64. Hohenstein EG, Sherrill CD (2009) Effects of heteroatoms on aro-
set. J Mol Model 18:4577–4589 matic π− π interactions: benzene−pyridine and pyridine dimer. J
47. Qiu N, Xue Y, Guo Y, Sun W, Chu W (2012) Adsorption of meth- Phys Chem A 113:878–886
ane on carbon models of coal surface studied by the density func- 65. Suresh CH, Gadre SR (2007) Electrostatic potential minimum of
tional theory including dispersion correction (DFT-D3). Comput the aromatic ring as a measure of substituent constant. J Phys Chem
Theor Chem 992:37–47 A 111:710–714
48. Neese F (2012) The ORCA program system. Comput Mol Sci 2: 66. Mani D, Arunan E (2013) The X–C⋯ Y (X= O/F, Y= O/S/F/cl/Br/
73–78 N/P)‘carbon bond’and hydrophobic interactions. Phys Chem Chem
49. Weigend F (2002) A fully direct RI-HF algorithm: implementation, Phys 15:14377–14383
optimised auxiliary basis sets, demonstration of accuracy and effi- 67. Cheeseman JR, Carroll MT, Bader R (1988) The mechanics of
ciency. Phys Chem Chem Phys 4:4285–4291 hydrogen bond formation in conjugated systems. Chem Phys Lett
50. Goerigk L, Grimme S (2011) A thorough benchmark of density 143:450–458
functional methods for general main group thermochemistry, kinet- 68. Popelier P, Bader R (1994) Effect of twisting a polypeptide on its
ics, and noncovalent interactions. Phys Chem Chem Phys 13:6670– geometry and electron distribution. J Phys Chem 98:4473–4481
6688 69. Tang T, Hu W, Yan D, Cui Y (1990) A quantum chemical study on
selected π-type hydrogen-bonded systems. J Mol Struct
51. Murray JS, Politzer P (1998) Electrostatic potentials: chemical ap-
THEOCHEM 207:319–326
plications. Encyclopedia of Computational Chemistry. doi:
70. Koch U, Popelier P (1995) Characterization of CHO hydrogen
10.1002/0470845015.cca014
bonds on the basis of the charge density. J Phys Chem 99:9747–
52. Tavakol H, Mollaei-Renani A (2014) DFT, AIM, and NBO study of 9754
the interaction of simple and sulfur-doped graphenes with molecu- 71. Bader R, Slee TS, Cremer D, Kraka E (1983) Description of con-
lar halogens, CH3OH, CH3SH, H2O, and H2S. Struct Chem 25: jugation and hyperconjugation in terms of electron distributions. J
1659–1667 Am Chem Soc 105:5061–5068
53. Becke A, Matta CF, Boyd RJ (2007) The quantum theory of atoms 72. Popelier P (1998) Characterization of a dihydrogen bond on the
in molecules: from solid state to DNA and drug design. Wiley, New basis of the electron density. J Phys Chem A 102:1873–1878
York 73. Carroll MT, Bader RF (1988) An analysis of the hydrogen bond in
54. Bader RF (1991) A quantum theory of molecular structure and its BASE-HF complexes using the theory of atoms in molecules. Mol
applications. Chem Rev 91:893–928 Phys 65:695–722
55. Johnson ER, Keinan S, Mori-Sanchez P, Contreras-Garcia J, Cohen 74. Carroll MT, Chang C, Bader RF (1988) Prediction of the structures
AJ, Yang W (2010) Revealing noncovalent interactions. J Am of hydrogen-bonded complexes using the Laplacian of the charge
Chem Soc 132:6498–6506 density. Mol Phys 63:387–405
56. Lu T, Chen F (2012) Multiwfn: a multifunctional wavefunction 75. Shahi A, Arunan E (2014) Hydrogen bonding, halogen bonding
analyzer. J Comput Chem 33:580–592 and lithium bonding: an atoms in molecules and natural bond or-
57. Humphrey W, Dalke A, Schulten K (1996) VMD: visual molecular bital perspective towards conservation of total bond order, inter-and
dynamics. J Mol Graph 14:33–38 intra-molecular bonding. Phys Chem Chem Phys 16:22935–22952
58. Su P, Li H (2009) Energy decomposition analysis of covalent bonds 76. Cremer D, Kraka E (1984) Chemical bonds without bonding elec-
and intermolecular interactions. J Chem Phys 131:14102 tron density—does the difference electron-density analysis suffice
59. Schmidt MW, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, for a description of the chemical bond? Angew Chem Int Ed Engl
Jensen JH, Koseki S, Matsunaga N, Nguyen KA, Su S (1993) 23:627–628
General atomic and molecular electronic structure system. J 77. Espinosa E, Alkorta I, Elguero J, Molins E (2002) From weak to
Comput Chem 14:1347–1363 strong interactions: a comprehensive analysis of the topological and
60. Cinke M, Li J, Bauschlicher CW, Ricca A, Meyyappan M (2003) energetic properties of the electron density distribution involving
CO 2 adsorption in single-walled carbon nanotubes. Chem Phys X–H⋯ F–Y systems. J Chem Phys 117:5529–5542
Lett 376:761–766 78. Cohen AJ, Mori-Sánchez P, Yang W (2008) Insights into current
limitations of density functional theory. Science 321:792–794
61. Salame II, Bandosz TJ (1999) Experimental study of water adsorp-
79. Umeyama H, Morokuma K (1977) The origin of hydrogen bond-
tion on activated carbons. Langmuir 15:587–593
ing. An energy decomposition study. J Am Chem Soc 99:1316–
62. Lu T, Manzetti S (2014) Wavefunction and reactivity study of 1332
benzo [a] pyrene diol epoxide and its enantiomeric forms. Struct 80. Zhechkov L, Heine T, Patchkovskii S, Seifert G, Duarte HA (2005)
Chem 25:1521–1533 An efficient a posteriori treatment for dispersion interaction in
63. Murray JS, Politzer P (2011) The electrostatic potential: an over- density-functional-based tight binding. J Chem Theory Comput 1:
view. Comput Mol Sci 1:153–163 841–847