Adsorption of Polycyclic Aromatic Hydrocarbons

0 downloads 0 Views 1MB Size Report
shows that the adsorption strength of adsorbed PAHs onto γ-graphyne nanosheets (GY) is ..... GY-PTCDA system in ~0.4 eV compared to the free adsorbent.

Adsorption of Polycyclic Aromatic Hydrocarbons (PAHs) onto Graphyne: Comparisons with Graphene Diego Cortés-Arriagada* Nucleus Millennium Chemical Processes and Catalysis (CPC); Laboratorio de Química Teórica Computacional (QTC), Pontificia Universidad Católica de Chile, Av. Vicuña Mackenna 4860, Macul, Santiago, Chile. *Corresponding author e-mail: [email protected] Abstract. Density functional theory calculations were implemented to expand the knowledge about graphyne and its interaction with polycyclic aromatic hydrocarbons (PAHs). Due to the porous character of graphyne, the adsorption strength of PAHs onto graphyne surfaces is expected to be lower with respect to graphene (a perfect -extended system). However, there are not quantitative evidences for this assumption. This work shows that the adsorption strength of adsorbed PAHs onto -graphyne nanosheets (GY) is weakened in 1223% with respect to the adsorption onto graphene, with a decrease of 1020% in the dispersive interactions. The adsorption energies (in eV) of the GYPAH

1

systems can be straightforward obtained as Eads/eV0.033NH+0.031NC, where NH and NC is the number of H and C atoms in the aromatic molecule, respectively. This equation predicts the binding energy of graphene-graphyne bilayers with a value of 31 meV/atom. Analysis of the electronic properties shows that PAHs behaves as n-dopants for GY, introducing electrons in GY and also reducing its bandgap in up to 0.5 eV. Strong acceptor or donor substituted PAHs decrease the bandgap of -graphyne in up to 0.8 eV, with changes in its valence or conduction band, depending on the chemical nature of the adsorbate. Finally, these data will serve for future studies related to the bandgap engineering of graphyne surfaces by non-aggressive molecular doping, and for the development of graphyne based materials with potential applications in the removal of persistent aromatic pollutants. Keywords: graphyne, graphene, polycyclic aromatic hydrocarbons, adsorption, bandgap. 1. Introduction The adsorption of polycyclic aromatic hydrocarbons (PAHs) onto low-dimensional carbon allotropes have been of interest, which is due to that PAHs can be used as nonaggressive dopant molecules to tailor the electronic properties of carbon materials1-4. In addition, carbon nanomaterials can act as good adsorbents toward priority PAHs pollutants5-7. Graphene can be considered as a perfect -extended system, which is suitable to analyze the - interactions with aromatic molecules in detail. In this case, PAHs can be adsorbed onto graphene with high stability and adsorption energies due to the additive character of the dispersive interactions2. This property is useful, for instance, to use PAHs in the exfoliation of graphene flakes8. PAHs also can induce the bandgap opening of graphene (on the range of 10240 meV2,4,9,10), which is enough to apply these modified 2

graphene materials as semiconductors at room temperature4,10. Graphene based materials are also excellent adsorbents for the removal of persistent PAHs pollutants by means of solid phase extraction; these materials shows good adsorption and reusability properties in a wide range of pH7,11-13. Furthermore, graphene based materials also allow to improve the separation and pre-concentration of pollutant traces for analytical methods12. In the last years, graphyne surfaces have also emerged as new large and porous two dimensional surfaces analogues to graphene. Graphyne is composed of benzene rings connected by acetylene bonds, this is containing sp and sp2 bonds14-16. Several synthetic efforts have been developed to obtain graphyne nanosheets, which are based on the use of annulene frameworks, and metals as catalysts14,17,18, where the usage of metal substrates appears to be necessary14,19; therefore, the rise of graphyne materials is expected to emerge in the coming years. Graphyne behaves as a semiconductor, with high mechanical stability, and their properties are promising for several applications such as in nanofillers, gas storage, sensors, desalination, Li-ion batteries, transistors, smooth metal-semiconductor interfaces, optoelectronic devices, UV light protection, among others14-16,20-23. Moreover, graphyne is also an emerging adsorbent material. Density functional theory (DFT) studies show that graphyne has good sorption properties towards molecules such as H2CO, NH3, nucleobases, and aromatic aminoacids24-27. To our knowledge, there are no works reporting the interaction of PAHs onto graphyne. The latter could be explained because of the porous character of graphyne surfaces is expected to avoid the strong - stacking with aromatic molecules with respect to graphene. Nonetheless, there are not quantitative evidences for such assumptions. In this framework, and taking as reference the effects of adsorbed PAHs onto graphene, a 3

dispersion corrected DFT study is performed to get insights into the adsorption strength and effects of PAHs molecules onto graphyne. All the calculations were performed with graphyne, which was taken as a representative class of graphyne surfaces. Adsorption energies and electronic properties were also compared with those computed for graphene. 2. Computational Methodology Finite graphene (G: C96H26) and -graphyne (GY: C86H22) models were used in the calculations, where the dangling bonds at the edges were saturated with hydrogen atoms. The selection of these models is based on the criteria of well converged adsorption energies as the surface size increases (see the supporting information). The studied PAHs molecules were: benzene (C6H6), naphthalene (C10H8), pyrene (C16H10), chrysene (C18H12), benzo[a]pyrene (C20H12),

C24H12) and ovalene (C32H14). In the case of PAHs, a

wide range of molecular weights (78.1398.4 g/mol) and C:H ratios (1.00.4) were considered. All the DFT calculations were performed in the ORCA 3.0 program28, without symmetry or geometry constraints. The PBE29 functional was used in combination with the all-electron Def2-SVP basis sets30. The DFT-D3 method (including the Becke-Johnson damping function) was used to include the dispersive energy correction in the SCF energies (ESCF-DFT) and gradients31,32; the latter results in the PBE-D3 method. According with the DFT-D3 method, the total energy (EDFTD3) is expressed as a sum of electronic (ESCF-DFT) and dispersion contributions (Edisp): EDFTD3=ESCFDFT+Edisp. Wavefunction and Atoms-inMolecules (AIM) analyses were performed in the Multiwfn program33 with the .molden output files of the ORCA calculations. 4

Adsorption energies (Eads) were obtained as:

Eads  EG  EPAH  EG  PAH

(1)

where, EG, EPAH and EG-PAH correspond to the total energies of the adsorbent, PAH, and adsorbent-PAH systems, respectively. The more positive values of Eads, the more stable the adsorbed structures are. Basis set superposition errors were corrected with the geometrical counterpoise correction of Kruse and co-workers34. The EvdW term was also analyzed, which is the contribution of dispersive interactions to the adsorption energy. According with the DFT-D3, total adsorption energies can be decomposed into the sum of electronic and dispersion contributions: Eads=Eads-SCF-DFT+EvdW. In this equation, Eads-SCF-DFT is the electronic contribution (the adsorption energy without dispersion corrections), and EvdW is the

dispersion

contribution.

EvdW

is

computed

as:

EvdW=Edisp(adsorbent)+Edisp(diox)Edisp(adsorbentdioxane), where Edisp(i) are the dispersion corrections of the isolated fragments and the adsorbent-adsorbate system. 3. Results and discussion 3.1 Adsorption strength First of all, to test the reliability of the selected methodology (and in order to have our reference values) the interlayer cohesive energy of graphite was computed. This property has a reference experimental value of 525 meV/atom, which was determined by means of thermal desorption analyses35. To obtain this value, the adsorption energies per carbon atom (Eads/C atom) of PAHs adsorbed onto graphene were computed, this is the total adsorption energy (Eads) per number of carbon atoms in the PAH molecule. Fig. 1a

5

shows the computed Eads/C atom values, which are in a linear relation with respect to the H:C ratio in the aromatic molecules (NH:NC). In this way, the extrapolated Eads/C atom value gives the interlayer cohesive energy of graphite when NH:NC=0 (this is when the adsorbed molecule is entirely composed of carbon)35. The predicted value is 45.2 meV with the PBE-D3/Def2-SVP level of theory, showing a deviation of only 7 meV/atom from the experimental value (525 meV/atom). Note that the prediction is in agreement with those values obtained from periodic PBE-D3 and vdW-DF calculations (43-48 meV/atom)36,37. Therefore, the selected methodology appears to be reliable to study the adsorption of PAHs onto graphene as a reference, and subsequently, onto graphyne. In addition, the computed total adsorption energies for the interaction of benzene, naphthalene, coronene and ovalene (Eads) onto graphene are in good agreement with the experimental ones (Table 1 in parenthesis)35.

Figure 1. a) Adsorption energy per carbon atom (Eads/C atom) of PAHs adsorbed onto graphene (G) and graphyne (GY) according to the H:C ratio of the PAH molecules. b) Stack (S) adsorption configuration of some GPAHs systems. c) All the adsorption configurations in the GYpyrene system; all the GPAH and GYPAH systems are in the 6

supporting information (SI). Hydrogen atoms were deleted from the molecular representations. Table 1. Adsorption energies (Eads in eV) and intermolecular distances (dinter in Å) of selected PAHs adsorbed onto graphene (G) and graphyne (GY). (a) Experimental values from Redhead analysis35. Conf is the corresponding adsorption configuration found for each adsorbent-adsorbate system and listed as a, b, c...; each of these adsorption configurations adopt a spatial orientation according to those defined in Fig. 1c, which are S (stack), B (bridge) and R (rotated). All these adsorption configurations are in Fig. S3 of the supporting information. PAH

benzene

graphene (G) Eads EvdW dinter 0.43

0.49

3.44

0.76

3.50

a

(0.5±0.08)

napthalene

0.67 (0.8±0.1)a

pyrene

1.00

1.11

3.53

chrysene

1.14

1.25

3.55

benzo[a]pyrene

1.23

1.37

3.55

1.41

1.56 1.56 2.02

3.56

coronene

(1.3±0.2)a

ovalene

1.79 (2.2±0.2)a

3.57

graphyne (GY) conf

Eads EvdW

dinter

a (S) b (B) c (B) a (B) b (R) c (B) a (B) b (R) c (S) d (S) e (B) a (R) b (B) a (B) b (R) c (S) d (S) a (S) b (B) a (B) b (S) c (B)

0.38 0.33 0.33 0.58 0.58 0.56 0.84 0.83 0.82 0.81 0.81 0.97 0.96 1.03 1.03 1.02 1.00 1.13 1.13 1.44 1.44 1.43

3.46 3.44 3.45 3.40 3.39 3.41 3.42 3.42 3.51 3.54 3.45 3.39 3.39 3.37 3.43 3.52 3.51 3.53 3.58 3.55 3.54 3.54

0.40 0.39 0.39 0.65 0.65 0.65 1.00 0.96 0.92 0.90 0.95 1.13 1.12 1.22 1.20 1.15 1.14 1.29 1.33 1.69 1.74 1.71

7

The reliability of the dispersion corrected PBE-D3 calculations was also compared against the non-local (NL) DFT methodology, where the electron density is used to obtain the long-range dispersive contributions. In the DFT-NL method, the non-local term of the VV10 functional (with refitted parameters) is used in combination with the PBE functional38,39. In this case, the NL-corrected adsorption energies were up to 0.03 eV larger than those determined by the DFT-D3 method, and they almost does not change the predicted value of the interlayer cohesive energy of graphite. The latter indicates the good agreement between the DFT-D3 and DFT-NL methods for the description of the dispersive interactions. Additionally, the B97-D3 functional was also considered for the calculations. However, B97-D3 was found to overestimate the adsorption energies for the interaction of larger PAHs onto graphene, causing an overestimated value of the interlayer cohesive energy of graphite (61.2 meV) (see the supporting information). All of these results show that the selected PBE-D3 methodology is reliable for the present study taking into account accuracy and computational cost. The most stable adsorption configuration of adsorbed PAHs onto graphene is the stacked (S) conformation4,40,41 (Fig. 1b), and they were obtained for comparison purposes. Conversely, PAHs can adopt several adsorption configurations onto GY [S (stack), B (bridge) and R (rotated), Fig. 1c], without observing a trend in which one is preferred with respect to their adsorption energies (Table 1). For instance, the stack configuration is the most stable for the GYbenzene interaction because the - interaction is maximized; but in larger PAHs, bridge and rotated configurations can be preferred. In both graphene and GY, PAHs are adsorbed at intermolecular distances in the range of dinter=3.43.6 Å. At these equilibrium distances, the adsorption of neutral aromatic molecules onto the aromatic 8

and regular graphene lattices (in addition to the graphene/graphyne interactions) is due to the cooperative interplay between , short-range electrostatic Coulombic interactions, Pauli repulsion and van der Waals interactions40-44. In this respect, the porous character of GY is expected to decrease the adsorption strength as a result of its decreased system compared to graphene. Indeed, the adsorption energies (Eads values in Table 1) of the GYPAH systems are lower compared to the GPAH systems. Nonetheless, the decrease in the adsorption strength is of the order of 1223% with respect to the adsorption onto intrinsic graphene. For instance, larger PAHs (such as coronene and ovalene) are adsorbed with high adsorption energies of up to 1.4 eV; thus, PAHs can still be adsorbed with high adsorption energies onto GY. The decrease in the adsorption strength is clearly attributed to the decrease in the long-range interactions as noted above. The contribution of dispersive interactions in the GYPAH systems is decreased in 1021% with respect to the GPAH systems (EvdW values in Table 1), which agrees well with the decrease in the total adsorption energies. Note that EvdW values are always larger than the Eads values, indicating the adsorption is only due to long-range interactions and the electronic part of the adsorption energy is mainly repulsive. Additionally, the different adsorption configurations of the GYPAH systems show differences of up to 0.05 eV between them. In this regard, the diffusion energy paths of aromatic molecules adsorbed onto two dimensional carbon surfaces is very flat (with barriers of up to 15 meV)2,45. Therefore, PAHs are expected to freely diffuse and rotate onto GY at low and room temperature (self-arranging onto GY), but they are still adsorbed with high energies.

9

On the other hand, the adsorption energy per carbon atom (Eads/C atom) of PAH molecules onto GY is displayed in Fig. 1a, which is also a linear relation with respect to the NH:NC ratio like in the case of graphene. In this way, the maximum adsorption energy (in eV) reached by PAHs onto GY is fitted to a straightforward linear equation as: Eads/eV(0.033NH+0.031NC)

(2)

where NH and NC is the number of H and C atoms in the aromatic molecule, respectively. From this equation, it is obtained a straightforward way to estimate the adsorption energy of any neutral and non-substituted PAH adsorbed onto GY. Interestingly, this relation was also successfully applied to PAHs with saturated moieties or five membered rings [such as acenaphthene (C12H10), fluorene (C13H10) and fluoranthene (C16H10)]. The predicted total adsorption energies of these molecules onto GY were of 0.70, 0.73 and 0.83 eV, respectively, by using the equation (2). These values are in excellent agreement with those computed by means of the PBE-D3/Def2-SVP method of 0.69, 0.72 and 0.83 eV, respectively (these systems are in the supporting information). Furthermore, the equation (2) allows to predict the interaction energy of GY with graphene layers by extrapolation to NH=0, which gives a predicted value of 31 meV/atom. This value should be considered in further works related to graphene-graphyne systems, which show potential applications for electronics, for instance as field effect transistors in highly-integrated circuits46. Moreover, the Eads energy difference between the curve of graphene and graphyne (see Fig. 1a) indicates that the maximum decrease in the adsorption strength in GY with respect to graphene, which is of 30% for PAHs with low NH:NC ratio (like graphene) and 12% for PAHs with high NH:NC ratio (like benzene). In this regard, the

10

veracity of the predicted binding energy for the graphene-graphyne interaction [and in general of the equation (2)] is well supported by comparison with the binding energies of graphyne bilayers, which are in the range of 19-24 meV/atom44,47. The interaction strength of the graphyne-graphyne interaction is lower than the graphyne-graphene as expected, and the decrease must be of at least 30% as noted above from the equation (2) and Fig. 1a. By this simple assumption, the graphyne-graphyne binding energy would represent the 70% of the graphyne-graphene interaction, and the graphene-graphyne binding energy must be of the order 28-34 meV/atom by simple mathematical inspection. The latter is in good agreement with the predicted value from the equation (2), indicating the reliability of this relation. 3.2 Electronic properties The effects of the aromatic molecules into the electronic properties of GY is explored in this section. Some relevant electronic properties of the GY-PAH systems are displayed in Table 2: the total Mulliken charge of the PAHs molecule after interaction (QPAH), eigenvalues of the frontier molecular orbitals (HOMO and LUMO), and the HOMO-LUMO energy gaps. As noted in Table 2, the charges of PAHs molecules after interaction with GY are of the order of 0.010.05|e|; thus, the electron transfer occurs in the PAHGY direction. This behavior is similar compared to aromatic molecules adsorbed onto graphene, where the PAH charge after interaction was determined to be of 0.01|e| in agreement with previous DFT studies2. According to these results, PAHs behave as mild n-dopants for the adsorbent, introducing electrons in GY. Note that introduction of holes and electrons should result in 11

enhanced conductive properties. Additionally, Fig. 2a displays the electron density difference () between the electron densities after and before the interaction takes place; the outflow and accumulation of electron density are depicted with red and yellow color, respectively. From the  surfaces, it is noted that th ads bat s p la iz th π-density in graphene and GY by means of the intramolecular charge transfer just below the adsorption site; in this way, the adsorption site turns electron deficient. This charge density distribution is responsible for the contribution of electrostatic interactions in the  stacking of aromatic molecules on graphene. On the basis of these results, the nature of the electrostatic interactions with PAHs is similar in graphene and graphyne. Table 2. Electronic properties of the most stable GYPAH systems: Mulliken charge of the PAH molecule after interaction; HOMO and LUMO energies (HOMO and LUMO), and HOMO-LUMO energy gap (HL). Energies are in eV; charges are in |e|.

PAH

QPAH

HOMO

LUMO

HL

benzene naphthalene pyrene chrysene benzo[a]pyrene coronene ovalene BTQBT PTCDA

0.01 0.03 0.05 0.03 0.05 0.01 0.02 0.11 -0.01

-4.96 -4.94 -4.71 -4.77 -4.53 -4.78 -4.39 -4.10 -5.03

-3.57 -3.54 -3.54 -3.54 -3.53 -3.53 -3.53 -3.50 -4.05

1.39 1.40 1.17 1.23 1.00 1.25 0.86 0.60 0.98

-4.96

-3.57

1.39

GY

12

Figure 2. a) Charge density difference of the GPAH and GYPAH systems; the outflow and accumulation of electron density after the interaction are depicted with red and yellow color, respectively. b) Bond paths connecting adsorbent and adsorbate through intermolecular bond critical points assigned to long-range electrostatic interactions. GYbenzene and GYovalene systems (conformation a) were selected as representatives; the bond paths for all the systems are in the supporting information. An atom-in-molecules (AIM) analyses of the GYPAH systems was also performed to get insights into the nature of their electrostatic interactions. Fig. 2b displays the AIM bond paths connecting the nuclear critical points (NCPs) and bond critical points (BCPs). The attention was focused on the intermolecular critical points connecting adsorbent and adsorbate through the intermolecular BCPs. The AIM inspection of the GPAH and GYPAH systems shows similar patterns in the distribution of intermolecular BCPs, which 13

have low electron density values (