Structural and electronic properties of covalently functionalized 2-aminoethoxy-metallophthalocyanine–graphene hybrid materials: a computational study

Abstract

The formation of a strong covalent bond between graphene and 2-aminoethoxy metallophthalocyanine (AEMPc) with the metal atom (M) being Zn, Fe, and Ni is established from density functional theory (DFT) based calculations at the B3LYP/6-31G(d)/LANL2DZ level of theory. The optimized structures of the hybrid complexes, represented by AEMPc–graphene, are reported. The projected density of states (PDOS) spectrum of each molecule has been calculated to explore the change in the HOMO–LUMO gap due to anchoring of AEMPc to graphene. The IR peak-positions and intensities obtained for the newly formed C–H and C–N bonds confirm the covalent link between the two moieties. The computed Raman spectra of the hybrid complexes show some changes in the relative intensities of D and G bands of graphene in accordance to those observed experimentally in a similar graphene based hybrid material. TDDFT calculations are carried out to study their absorption spectra in DMF solvent. For all three metal atoms in the composite molecules, there appears a charge transfer band in the range 600–630 nm. Three long-range corrected functionals such as M06-2X, CAM-B3LYP, and wB97XD are used to compare the results with those of the hybrid B3LYP functional.

---

1. Introduction

The structure of phthalocyanine (Pc) is similar to that of chlorophyll and hemoglobin. Over the years a large number of metallophthalocyanines (MPcs) have been synthesized and characterized. MPcs are very stable macrocyclic compounds with a high π-electron delocalization and efficient light absorption property in the red and visible region. These are used as photoconducting materials and photosensitizers in photodynamic therapy,^1–6 as liquid crystals^7,8 and in nonlinear optics.^9,10 Due to high thermal and chemical stabilities, MPc complexes containing first-row transition metals have drawn a considerable interest in a wide range of technological applications such as semiconductor devices, thin-film transistors, solar cells, light-emitting diodes, and even as molecular spintronics.^11–14 Moreover, these transition-metal phthalocyanines, when attached covalently or non-covalently with electron acceptor materials such as C₆₀ and carbon nanotubes, make them promising candidates in the emerging field of organic solar cells as well as artificial photosynthetic systems.^15–18 Because of their biological and photochemical significances, and less complicated synthesis, a large number of experimental and theoretical studies have been made to investigate various electronic properties of MPcs. Liao and Scheiner¹⁹ have shown that the valence electronic structures of MPcs containing Fe and Co are significantly different from those containing Ni and Zn. A number of energetically close-lying electronic states arising out of (n − 1)d open shells in both Fe and Co have been identified. In previous studies,^20,21 charge transfer transitions have been identified in the spectra of FePc and CoPc. It is also known that Pcs can generate singlet oxygen which is the cytotoxic species responsible for the destruction of tumor in photodynamic therapy.²² In this context, ZnPc has a special significance as the Zn atom provides the Pc ring an important fluorescence and singlet oxygen production characteristics. Several spectroscopic studies were carried out on ZnPc in gas, solid, and solution phases.^23–27 Time dependent DFT (TDDFT) calculations on ZnPc²⁸ provided an accurate spectral data which matched excellently with the experimental results.

Graphene has a high electrical conductivity due to its honeycomb lattice structure. It is known to be a zero-gap semiconductor with its valence and conduction band touching in k-space. It has been used to prepare transparent conducting electrodes for energy conversion and storage system.^29,30 Despite the potential applications, graphene possesses chemically inert surfaces for physical adsorption. This feature weakens its competitive strength in the field of semiconductors and biomedical applications. Like carbon nanotubes, graphene sheets are also insoluble in aqueous media and in most organic solvents. Moreover, graphene has a tendency to form irreversible aggregates or restack by means of van der Waals interactions and π–π stacking. As a consequence, the conjugation of graphene with active biological components becomes a daunting challenge. To overcome these problems, exfoliation techniques are routinely employed to fabricate graphene based hybrid materials in which the unique properties of graphene retain. Two sides of graphene's plane are free to undergo chemical reactions which would control its electronic properties.^31,32 ZnPc–poly(p-phenylene vinylene) supramolecularly interacting with graphene has produced a hybrid material that is used as a component in a photovoltaic cell.³³ Recently,³⁴ (2-aminoethoxy)(tri-tert-butyl) ZnPc–graphene hybrid material was prepared by graphene exfoliation upon tip sonication in o-dichlorobenzene and characterized for its photosensitivity. This ZnPc–graphene hybrid material having charge separated transient species, when integrated into an optical transparent electrode, showed stable and reproducible photocurrent responses. Tao et al.³⁵ carried out TDDFT calculations on tetra-thiafulvalene annulated zinc porphyrazine for determining vertical excitation energies and comparing with the observed UV-vis spectra.

Liquid phase exfoliation and sonication methods have also been adopted recently to functionalize few-layer graphene and hexagonal boron nitride (hBN) nanosheets with NiPc in solution. Electrical and optical properties of 2D graphene and hBN are controlled by such functionalization with NiPc as verified by different spectroscopic measurements including Raman, XPS, EDX etc.³⁶ These complexes are found to be good materials in photonic and optoelectronic applications. The adsorptions of cobalt phthalocyanine (CoPc) and cobalt tetra-amine phthalocyanine on graphene functionalized with COO⁻ and CO are investigated at the DFT level to determine their electrical characteristics.³⁷ Studies suggest that the electron transfer takes place from phthalocyanine to the graphene moiety and the interaction energy depends on the relative positions of the functionalized graphene and CoPc planes. Defects and vacancies in the graphene sheet are found to affect such electron transfer.

In this article, we have quantum chemically studied the interaction between 2-aminoethoxy-MPc (AEMPc) with graphene by means of DFT calculations. FePc is known to have a different valence electronic structure from NiPc and ZnPc, so for the sake of comparison, three metal atoms namely, Fe, Ni, and Zn are chosen in the present study. Absorption spectra of each hybrid complex in the gas phase as well as in DMF solvent are computed at the TDDFT level. IR and Raman intensities are also estimated and compared with those of the experimentally synthesized and similarly functionalized ZnPc–graphene hybrid material.

2. Computational details

Geometries of pristine graphene, AEMPc, and AEMPc–graphene (M = Fe, Ni, Zn) molecules in their ground states were optimized using the ab initio based DFT calculations with B3LYP hybrid functional which involves the Becke three-parameter exchange and Lee–Yang–Parr correlation functional.³⁸ In the DFT calculations, we have employed 6-31G(d) basis sets for the lighter atoms namely, C, H, N, and O, while LANL2DZ basis sets and effective core potentials are used for the heavier Fe, Ni, and Zn atoms. No constraint was imposed during the geometry optimization. Throughout the calculations, the SCF convergence limit was set to 10⁻⁶ a.u. on energy and density. For each system, electronic states with different spin multiplicities are considered to determine the ground state geometry. All the stationary points on the potential energy surfaces were confirmed from the harmonic vibrational frequency analysis at 298.15 K and 1 atm using the ideal gas approximation. DFT calculations have been performed using Gaussian 09 suite of program.³⁹

In order to simulate the absorption spectra, optical transitions of all AEMPcs, graphene, and their hybrid molecules are studied by performing TDDFT calculations at their ground-state optimized geometries. The vertical single excitation energies and oscillator strengths are computed for a maximum of 100 dipole and spin allowed excited states. Solvent effects are included through the self-consistent reaction field (SCRF) calculations using the polarizable continuum model (PCM)⁴⁰ as implemented in Gaussian 09. In the present study, N,N-dimethylformamide (DMF) is used as the solvent. However, the accurate prediction of vertical excitation energies of the low-lying charge transfer transitions is a major problem in the conventional TDDFT method. To alleviate this problem, a number of theoretical investigations have been made to calculate the charge transfer states of phthalocyanines using hybrid meta-GGA functional which depends not only on the spin density and its reduced gradient but also on the non-interacting spin-component kinetic energy densities.^35,41,42 In the present study, the charge transfer transition states have been assessed by computing electronic excitation energies with three different long-range corrected meta-hybrid functionals such as M06-2X,⁴³ CAM-B3LYP,⁴⁴ and wB97XD.⁴⁵

The M06-2X functional, which is a parameterized semiempirical hybrid functional with double amount of nonlocal exchange energy, performs well for the prediction of molecular geometries, vibrational frequencies, valence and Rydberg excitations, Raman intensities, thermochemistry and kinetics, non-covalent interactions etc.⁴⁶ The long-range corrected functionals like CAM-B3LYP and wB97XD also produce satisfactory results to reproduce the spectral properties as well as the non-linear optical properties.^47,48 The CAM-B3LYP functional is a long-range corrected version of the hybrid B3LYP functional using the Coulomb attenuating method. On the other hand, wB97XD corresponds to a modified B97 exchange and correlation functionals, and also involves the long-range correction with empirical atom–atom dispersion.

To assess the performance of the hybrid and meta-hybrid functionals, we have compared several molecular properties like HOMO–LUMO energy gap, polarizabilities, IR stretching frequencies, and the UV-visible spectra in the solvent phase. The computed Raman intensities using the values of the scattering activities obtained from different DFT methods have also been analyzed, because these intensities depend on derivatives of molecular polarizabilities with respect to the nuclear coordinates.

3. Results and discussion

3.1. Geometries

Electronic structures of MPcs (M = Fe, Ni, Zn) are known from several previous studies.^19,49–52 Earlier DFT calculations suggest that their structures belong to nearly a D_4h point group and in each case the metal atom binds strongly with the Pc ring. The shortest computed M–N bond distances in these molecules lie between 1.90 and 2.02 Å, which are comparable with the X-ray diffraction data. The metal–Pc binding energies for FePc and NiPc are found to be around 230.6 kcal mol⁻¹, while for ZnPc the magnitude is nearly half. The Pc rings in NiPc and ZnPc show high electron delocalization with the HOMO localized mainly on the ring, while in FePc, it is mostly a metal 3d orbital. In the present study, the 2-aminoethoxy group is attached to the periphery of the Pc ring so that its covalent functionalization with the pristine graphene is facilitated. Geometries of AEMPc molecules are optimized at the B3LYP/6-31G(d)/LANL2DZ level. The molecular axis of the 2-aminoethoxy group, when attached to one of the four benzene rings of the Pc molecule, remains in the plane of the Pc ring. The average optimized M–N bond distances in AEMPcs are 1.953, 1.924, and 2.014 Å for Fe, Ni, and Zn, respectively. The HOMOs of all three organometallic compounds are localized on four indole rings of Pc almost identically and they are of π type involving carbon atoms only. The LUMO of AEFePc is a pure 3d orbital of Fe, while the LUMOs of the corresponding Ni and Zn complexes do not have any metal contribution. These are mostly π type orbitals comprising of all N atoms surrounding the metal atom.

The diminished contribution of M-3d orbitals can be understood from the analyses of the molecular orbital energy levels of the metal substituted and unsubstituted Pcs. Generally, energies of M orbitals decrease from Fe to Zn. The HOMOs of AEMPcs are mostly contributed by Pc, and hence their energies are not affected by the metal substitution. In case of AEFePc, several 3d-like orbitals are situated near the LUMO of Pc so that the LUMO of AEFePc has a Fe-3d_π character. This stems from the fact that the empty d_x²−y² orbital of Fe lies just beneath the LUMO of Pc. However, in case of AENiPc, the 3d orbitals of Ni continue to be lowered in energy. Subsequently in AEZnPc, the 3d subshell of Zn is filled and lowered enough to form rather pure MOs. As a result, the metal d orbitals sink below the HOMO of Pc, thus making the latter as the HOMO of the complex.

The functionalized metal–phthalocyanine complexes, AEMPcs are then allowed to interact with a pristine graphene sheet containing 20 hexagonal rings. The amine group of AEMPc anchors covalently to the graphene sheet through a nucleophilic addition. Consequently, there is a local deformation of the graphene plane.^53–56 The optimized structures of AEMPc–graphene hybrid molecules at the B3LYP/6-31G(d)/LANL2DZ level are represented in Fig. 1. For all three hybrid complexes in the ground state, the covalent C–N bond distance is calculated to be 1.493 Å. Three C–C bonds of the graphene sheet surrounding the point of grafting elongate to 1.532, 1.522, and 1.519 Å distances. The extent of such bond relaxation is more or less same for all three metal phthalocyanines. The computed bond lengths are very close to the previously reported values obtained for the functionalization of graphene monolayers with –NH₂.⁵⁷ The newly formed C–H bond has a length of 1.109 Å. For all three AEMPc–graphene complexes, the lowest energy structures on the potential energy surface correspond to that geometry in which AEMPc is oriented perpendicular to the graphene nanosheet. The orientation of the functionalized MPc plane above the graphene sheet in the hybrid molecule matches well with the X-ray crystal structure of the recently synthesized (2-aminoethoxy)(tri-tert-butyl)-ZnPc–graphene hybrid material.³⁴ The average binding energy is computed to be about 79.3 kcal mol⁻¹.

Fig. 1 A representative ground-state structure of 2-aminoethoxy-MPc–graphene hybrid molecule.

A graphene sheet having about 500 carbon atoms is predicted to have an almost zero band gap. Earlier calculations have shown that the HOMO–LUMO energy gap decreases with the increase in the number of carbon atoms.⁵⁸ In the present study, we have taken a graphene sheet of 58 C atoms with H termination at the zigzag boundary edges and the computed HOMO–LUMO gap amounts to 0.9 eV. It is expected that the gap will reduce if a larger graphene cluster is taken. Because, for a large graphene cluster, the quantum effect becomes less important, thereby its electronic properties would resemble with those of the pure graphene having a zero band gap. The density of states (DOS) spectra of AEMPcs before being covalently grafted with graphene indicate that the HOMO–LUMO gap for AEFePc is about 1.53 eV, while for AENiPc and AEZnPc the values are 2.18 and 2.14 eV, respectively. Fig. 2 shows how the covalent grafting of the functionalized MPc on the graphene sheet perturbs the DOS pattern. The HOMO–LUMO energy gaps of three hybrid molecules are marginally less than 0.9 eV which is the computed gap in pure graphene sheet. This feature supports the fact that the properties of graphene would remain unchanged after the covalent grafting with the functionalized MPcs.

Fig. 2 PDOS spectra of graphene, functionalized MPc, and their hybrid molecules: (a) M = Fe, (b) M = Ni, (c) M = Zn.

Fig. 3 displays the relative energies of HOMO and LUMO and their respective band gaps for the individual graphene and AEMPc as well as the composite systems. It demonstrates that the LUMOs of AENiPc and AEZnPc molecules are at sufficiently higher position than the LUMO of graphene. Hence, the charge transfer takes place from the phthalocyanine derivative to the graphene molecule. Due to charge transfer interactions between the donor and acceptor, the relative ordering of the individual HOMO and LUMO in AEMPc alters in the composite system. The band alignments in such complexes and the graphene molecule resemble closely, which corroborates with the DOS spectra. The HOMO and LUMO isosurfaces of AEFePc–graphene show that the HOMO is localized on graphene and the LUMO on FePc with a major contribution of the Fe atom. Thus in AEFePc–graphene, the graphene acts as an electron donor, while AEFePc is the electron withdrawing group. On the other hand, both the HOMO and LUMO in AEMPc–graphene (M = Ni, Zn) are mainly localized on the graphene moiety with contributions from the carbon p_z and s atomic orbitals.

Fig. 3 Relative positions of HOMOs and LUMOs of graphene (I), AEFePc (II), AENiPc (III), AEZnPc (IV), AEFePc–graphene (V), AENiPc–graphene (VI), AEZnPc–graphene (VII).

The interaction between AEMPc and graphene is further analyzed from the electron density distribution maps. The color-filled map of the electron density of AEZnPc–graphene in the binding region as calculated from the molecular orbital is displayed in ESI Fig. S1.† The localized orbital locator (LOL)^59,60 analysis has been used for locating high localization regions in the conjugation of AEZnPc to the graphene sheet. The shaded surface map with the projection of LOL in the plane of grafting of AEZnPc–graphene composite is shown in Fig. 4. The covalent regions have high LOL values which mean more confinement of the electronic motion within it, thereby reflecting a strong binding between the N atom of the aminoethoxy group of AEZnPc and the C atom of graphene. The electron depletion region between the valence and inner shell is shown by the blue circles around the nuclei.

Fig. 4 Shaded surface map with projection of the localized orbital locator in the plane containing the conjugated atoms of AEZnPc–graphene.

The chemical potential represented by (ε_HOMO + ε_LUMO)/2, is enhanced by 2% due to the covalent grafting of AEMPc to the graphene sheet. At the same time, the polarities of the hybrid molecules are increased to make these compounds more soluble in some suitable solvents. In AEZnPc–graphene, the calculated NBO charges on the C atom of graphene and the N atom of the amine group of AEZnPc at the point of contact are about +0.14 and −0.69, respectively showing an enhanced polarization. Moreover, polarizabilities and hyperpolarizabilities of these hybrid complexes would rationalize the dynamical response of these systems in an applied electric field. The computed dipole moments (μ), isotropic polarizabilities (α) as defined by (α_xx + α_yy + α_zz)/3, and average hyperpolarizabilities (β) for all the bare and composite systems are reported in Table 1. It shows that the anchoring of AEMPc to the graphene sheet increases both α and β, thereby enhancing the nonlinear optical properties of the composite systems.

Table 1 Computed dipole moments, μ (in Debye), isotropic polarizabilities, α (in a.u.), and average hyperpolarizabilities, β (in a.u.) of graphene, AEMPc (M = Fe, Ni, Zn) and their hybrid complexes at the B3LYP/6-31G(d)/LANL2DZ level of theory

System	μ	α	β
Graphene	0.00	1105.1	1.9
AEFePc	2.15	648.4	2048.9
AENiPc	2.17	647.9	2106.0
AEZnPc	2.17	659.3	2207.7
AEFePc–graphene	2.98	1730.8	2705.1
AENiPc–graphene	2.86	1727.6	2764.5
AEZnPc–graphene	3.03	1741.6	2954.6

---

3.2. Spectral features

The computed IR intensities of the hydrogen terminated graphene, AEZnPc, and the composite molecule are shown in Fig. 5(a)–(c). It is known that there exists no significant vibration in graphite or graphene. However, intensities shown in Fig. 5(a) are due to C–H stretching and some bending vibrations because of the added H atoms at the edges of the graphene sheet to satisfy the valence. In the IR spectra of AEZnPc, the characteristic fingerprint bands are located in the range 755–1670 cm⁻¹. Even after the functionalized ZnPc is covalently bonded to graphene, the IR band in the fingerprint region remains almost unchanged as seen in Fig. 5(c). This compares well with the fingerprint region 1045–1610 cm⁻¹ observed in the IR spectra of the (2-aminoethoxy)(tri-tert-butyl) ZnPc–graphene hybrid material prepared recently.³⁴ All the C–H stretching vibrations are in the frequency band 3000–3250 cm⁻¹. However, the newly formed C–H bond on the adjacent carbon atom of the graphene sheet where the phthalocyanine is anchored has a stretching vibrational frequency of 2916 cm⁻¹. The C–N bond, which links the two moieties, has vibrational frequencies in the range 1100–1200 cm⁻¹. A similar pattern of the IR band is observed for the other two hybrid complexes.

Fig. 5 Computed IR spectra of (a) graphene, (b) AEZnPc, (c) AEZnPc–graphene.

The conjugated double bonds in the exfoliated graphene or pristine graphite are precisely characterized by intense Raman signals. There exists a D band at 1338 cm⁻¹, a G band at 1578 cm⁻¹, a D′ band at 1615 cm⁻¹, and a 2D band at 2668 cm⁻¹.⁶¹ Fig. 6 shows how the computed Raman activities of the graphene sheet change when AEZnPc is covalently anchored on it. The computed bands of graphene at 1364 and 1613 cm⁻¹ resemble the experimental D and G bands. However, the D′ band did not appear, while 2D band is red shifted at 3207 cm⁻¹. An additional band at 1204 cm⁻¹ may be due to the presence of several C–H bonds at the edges of the graphene moiety. Because of the covalent link with the AEZnPc molecule, there is a formation of a sp³ carbon atom on the graphene sheet and the intensity ratio of the D and G band (I_D/I_G) is enhanced from 0.43 to 0.89 as revealed from the present computed data. This shift compares well with the magnitude of the enhancement of such intensity ratio reported from the experimental Raman spectra of the exfoliated graphene and (2-aminoethoxy)(tri-tert-butyl) ZnPc–graphene.³⁴ Raman activities of AEFePc–graphene and AENiPc–graphene complexes are very close to that of the AEZnPc–graphene system.

Fig. 6 Computed Raman spectra of graphene, AEZnPc, and AEZnPc–graphene.

The vertical excitation energies and oscillator strengths of graphene, AEZnPc, and AEZnPc–graphene obtained from the TDDFT calculations in DMF solvent are given in Table 2. The low-energy absorption band due to the graphene moiety is blue shifted in the composite molecule by 27 nm and its intensity is enhanced to a small extent. The strongest absorption of graphene at 573 nm is largely suppressed and shifted when it is covalently grafted with AEZnPc. At the same time, a weak band appears at 684 nm in the absorption spectra of the hybrid molecule and the composition shows that it is due to H−3 → L and H → L+4 transitions which involve only π orbitals of graphene.

Table 2 Calculated excitation energies (E_ex), oscillator strength (f), and contribution of molecular orbitals of electronic excitations for graphene, AEZnPc, and AEZnPc–graphene molecules in the DMF solvent

System	Eex, eV (nm)	f	Assignment (%)
Graphene	1.03 (1199)	0.6128	H → L(97)
	2.01 (617)	0.1370	H → L+2(48), H−2 → L(44)
	2.17 (573)	1.5994	H−1 → L+1(82), H → L(12)
	2.64 (470)	0.3012	H−3 → L(74), H → L+3(21)
	3.35 (370)	0.4662	H−1 → L+4(38), H−4 → L+1(31), H → L+9(11)
	3.50 (354)	0.4822	H−6 → L+1(48), H−2 → L+2(26), H−9 → L(14)
	3.63 (342)	0.6925	H−2 → L+2(50), H → L+10(17), H−6 → L+1(14)
	4.09 (303)	0.3184	H−3 → L+2(69), H−2 → L+3(17)
AEZnPc	1.97 (628)	0.5734	H → L(96)
	2.01 (618)	0.6588	H → L+1(96)
	2.97 (417)	0.1463	H−1 → L(64), H−1 → L+1(33)
	3.60 (344)	0.2378	H−6 → L+1(49), H−5 → L+1(33)
	3.70 (335)	0.6581	H−8 → L(42), H−5 → L+1(24)
	3.71 (334)	0.9406	H−5 → L(52)
	3.78 (328)	0.1284	H−9 → L(27), H−7 → L+1(18), H−8 → L+1(14), H−8 → L(12), H−9 → L+1(10)
AEZnPc–graphene	1.06 (1172)	0.7051	H → L(100)
	1.81 (684)	0.2573	H−3 → L(62), H → L+4(35)
	1.97 (628)	0.5371	H−2 → L+1(94)
	2.00 (618)	0.6586	H−2 → L+2(93)
	2.27 (546)	0.2217	H−1 → L+4(49), H−1 → L+3(31), H−3 → L+3(12)
	2.35 (527)	0.5542	H−1 → L+3(46), H−3 → L+3(33), H−1 → L+4(10)
	2.49 (497)	0.3875	H−3 → L+3(48), H−1 → L+4(35)
	2.72 (455)	0.2514	H−3 → L+4(79)
	2.97 (418)	0.1605	H−5 → L+1(57), H−5 → L+2(29)
	3.48 (356)	0.2006	H−1 → L+7(37), H−6 → L+3(31)
	3.60 (344)	0.2275	H−14 → L+2(47), H−12 → L+2(32)
	3.65 (339)	0.1981	H−4 → L+5(29), H−8 → L+3(27)
	3.69 (336)	0.1247	H−1 → L+9(24), H−8 → L+3(21), H−4 → L+5(18)
	3.70 (335)	0.6571	H−16 → L+1(35), H−12 → L+2(22)
	3.71 (334)	0.8644	H−12 → L+1(49), H−17 → L+2(10)
	3.75 (331)	0.2603	H−3 → L+7(39), H−1 → L+9(18), H−6 → L+4(16)
	3.80 (326)	0.4359	H−23 → L(18), H−1 → L+9(12), H−1 → L+11(12)

---

In AEZnPc, there are two almost equally strong absorption peaks centered at 628 and 618 nm. After the formation of the covalent link with the graphene sheet, the positions of these two absorption bands remain unaltered and their oscillator strengths change only marginally. But the transitions involved with these two peaks are predominantly H−2 → L+1 (94%) and H−2 → L+2 (93%), respectively. The analysis of the frontier orbitals as depicted in Fig. 7 indicates that these transitions involve charge transfer interaction between AEZnPc and graphene. The H−2 MO is fully localized on the graphene molecule, while both L+1 and L+2 MOs are described by the AEZnPc molecule. The new band around 546 nm is mainly attributed to two major transitions namely, H−1 → L+4 and H−1 → L+3. The nature of MOs suggests that there is a transfer of electron density from ZnPc to the graphene moiety associated with this band. The pure graphene absorption peak at 470 nm is blue shifted to 455 nm, while the peak at 417 nm for the functionalized ZnPc is red shifted in the hybrid molecule by 1 nm only. As seen from Table 2, two strong absorption peaks at 334 and 335 nm of AEZnPc are not perturbed by the covalent anchoring in the hybrid molecule, and their oscillator strengths do not change significantly.

Fig. 7 Some important occupied and unoccupied molecular orbitals of AEZnPc–graphene.

Table 3 displays the positions of absorption peaks mostly in the visible region for AEFePc, AENiPc, and their graphene anchored hybrid molecules in DMF solvent. Earlier experimental studies³⁶ suggest the presence of two main absorption regions in the spectrum of NiPc. There is a B band or Soret band in between 250 and 500 nm due to π → d transitions involving the Ni atom, while the Q band is centered around 650 nm due to π → π* transitions. This feature is more or less retained in the computed absorption spectra of AENiPc (ESI Fig. S2†). For both the Fe and Ni complexes, a pair of strong absorption peaks exists in the wavelength region 600–630 nm similar to those of AEZnPc–graphene. In case of AEFePc–graphene, these two peaks are due to two main transitions: H−1 → L+2 and H−1 → L+3, while for the nickel composite, these are H−2 → L+1 and H−2 → L+2. The nature of the molecular orbitals confirms that both the transitions are due to the interaction between AEMPc and graphene. So, excitations of the AEMPc–graphene hybrid molecules in the frequency range 600–630 nm result in a charge transfer between graphene and AEMPc. Two absorption peaks at 1172 and 684 nm make a signature of the graphene moiety in all three AEMPc–graphene hybrid molecules. Thus, the characteristics of graphene are expected to retain in the composite systems.

Table 3 Calculated E_ex, f, and contribution of molecular orbitals of electronic excitations for the functionalized FePc, NiPc, and their hybrid molecules with graphene in the DMF solvent

System	Eex, eV (nm)	f	Assignment (%)
AEFePc	2.02 (615)	0.5554	H → L+1(97)
	2.05 (605)	0.6372	H → L+2(97)
AEFePc–graphene	1.06 (1172)	0.7054	H → L(100)
	1.81 (684)	0.2608	H−3 → L(62), H → L+5(35)
	2.02 (615)	0.4357	H−1 → L+2(91)
	2.05 (605)	0.5538	H−1 → L+3(91)
	2.27 (546)	0.2220	H−2 → L+5(49), H−2 → L+4(32), H−3 → L+4(12)
	2.35 (527)	0.5559	H−2 → L+4(48), H−3 → L+4(33), H−2 → L+5(10)
	2.49 (497)	0.3875	H−3 → L+4(48), H−2 → L+5(35)
	2.72 (455)	0.3377	H−3 → L+5(88)
AENiPc	1.97 (629)	0.5222	H → L(97)
	2.00 (618)	0.6089	H → L+1(97)
AENiPc–graphene	1.06 (1173)	0.7053	H → L(100)
	1.81 (684)	0.2575	H−3 → L(62), H → L+4(35)
	1.97 (629)	0.4892	H−2 → L+1(97)
	2.00 (619)	0.6101	H−2 → L+2(97)
	2.27 (546)	0.2214	H−1 → L+4(50), H−1 → L+3(33), H−3 → L+3(12)
	2.35 (527)	0.5542	H−1 → L+3(49), H−3 → L+3(33), H−1 → L+4(11)
	2.49 (497)	0.3850	H−3 → L+3(49), H−1 → L+4(36)
	2.72 (455)	0.3192	H−3 → L+4(89)

---

3.3. Effects of long range corrected functionals

Table 4 compares HOMO–LUMO gaps of graphene, AEZnPc, and their hybrid molecules using three different long range corrected functionals, namely, M06-2X, CAM-B3LYP, and wB97XD computed at the B3LYP optimized geometries. The long-range corrected band gaps are considerably larger than the B3LYP values. The largest gaps are found when the wB97XD functional is used. This is mainly attributed to the inclusion of full Hartree–Fock exchange term in the wB97XD functional at long distances. However, when AEZnPc is anchored to graphene, the gaps are reduced by 0.05–0.07 eV only. The DFT-computed HOMO–LUMO energy gaps for graphene, AEZnPc, and their hybrid complex are further verified using molecular orbital based methods like HF and MP2. The predicted band gaps by both HF⁶² and MP2 ⁶³ methods are overestimated by 2.33–2.75 eV compared to DFT based calculations using B3LYP hybrid functional. A similar deviation for ZnPc has been reported by Ueno et al.⁶⁴ It may be mentioned that the virtual orbital energies estimated by DFT method are lower than those computed by HF/MP2 methods, while for the occupied orbitals, the predicted orbital energies are higher in case of hybrid DFT method. For the long-range corrected functional, the HOMO energy is further lowered compared to the B3LYP hybrid functional. These two contrasting factors lead to the variation in the HOMO–LUMO energy gap. The smallest band gap obtained for each system using B3LYP may be due to the incorrect asymptotic behavior shown by the common exchange functional.⁶⁵

Table 4 Computed HOMO–LUMO energy gap (E_g) using different hybrid and meta-hybrid functionals for graphene, AEZnPc, and AEZnPc–graphene molecules. All parameters are in eV

Functional	Eg
	Graphene	AEZnPc	AEZnPc–graphene
B3LYP	0.90	2.14	0.89
M06-2X	1.72	3.39	1.67
CAM-B3LYP	1.95	3.66	1.89
wB97XD	2.65	4.61	2.58

---

Although hybrid functionals like B3LYP are quite popular to determine the molecular geometries and many electronic properties, long-range corrected meta-hybrid functionals have been proved to be relatively better for the calculation of the non-linear optical properties. The computed α and β values for graphene, AEZnPc, and the complex using three meta-hybrid functional are reported in ESI Table S1.† The β values derived from B3LYP are found to be larger than those from M06-2X, CAM-B3LYP, and wB97XD. However, the long-range corrected functionals predict augmented α values compared to the B3LYP functional. This is in accordance with the previous findings.^66–68

The IR intensities of AEZnPc–graphene using meta-hybrid M06-2X, CAM-B3LYP, and wB97XD functionals are compared in ESI Fig. S3.† The computed IR spectra display a similar finger print band in the region 758–1694 cm⁻¹. The Raman activities of the same hybrid molecule for three functionals are given in ESI Fig. S4.† The positions of all important G and D bands do not change much. However, an additional strong Raman activity is noted around 1070 cm⁻¹, which is absent in the B3LYP result. The 2D band around 3200 cm⁻¹ remains unchanged. Fig. 8 displays the comparative absorbance spectra of AEZnPc–graphene computed at the B3LYP and three meta-hybrid functionals, while Table 5 shows the magnitudes of the excitation energies and oscillator strengths of the corresponding peaks of the charge transfer band (610–660 nm) and the low energy band (1100–1400 nm) due to graphene. In case of the M06-2X functional, these bands are blue shifted by 6–8 nm only, while for the other two long-range corrected functionals the bands are red shifted by about 60–200 nm. The computed oscillator strengths of the graphene band using meta-hybrid functionals are relatively large. However, the nature of transitions and their contributions remain unchanged. In the low energy band, the calculated excitation energies involving H → L as the main transition and using the asymptotically corrected CAM-B3LYP and wB97XD functionals are somewhat smaller than those obtained from B3LYP and M06-2X functionals. This is mainly attributed to the inclusion of a larger percentage of long-range HF exchange in CAM-B3LYP and wB97XD. Moreover, with the increase of HF exchange the optical band gap reduces significantly resulting in the red shift of the H → L transition peak. This suggests that the computed HOMO–LUMO gaps using hybrid DFT methods are overestimated compared to the TDDFT derived gaps. This is in accordance with the previous findings which demonstrated that the calculated HOMO–LUMO gaps by TDDFT using all types of functionals match well with the experimental data.⁶⁹ It may be noted that the computed transition energies using B3LYP for the charge transfer bands are also underestimated compared to CAM-B3LYP and wB97XD because the former functional is not asymptotically correct to describe the charge transfer states. Therefore, the performance of the DFT exchange correlation functional on the charge transfer excitation energy strongly depends on Koopmans' orbital energies of the donor and the acceptor.

Fig. 8 Comparison of the absorption spectra of AEZnPc–graphene computed by using different functionals in DMF solvent.

Table 5 Comparison of E_ex and f of AEZnPc–graphene hybrid molecule in the DMF solvent for different hybrid and meta-hybrid functionals

Functional	Eex, eV (nm)	f	Assignment (%)
B3LYP	1.06 (1172)	0.7049	H → L(100)
	1.97 (628)	0.5371	H−2 → L+1(94)
	2.00 (618)	0.6586	H−2 → L+2(93)
M06-2X	1.06 (1165)	1.0892	H → L (99)
	2.00 (620)	0.6250	H−1 → L+1(95)
	2.03 (612)	0.7020	H−1 → L+2(94)
CAM-B3LYP	1.00 (1235)	1.1284	H → L(98)
	1.92 (645)	0.6265	H−1 → L+1(95)
	1.95 (636)	0.7018	H−1 → L+2(94)
wB97XD	0.90 (1380)	1.2047	H → L(96)
	1.88 (660)	0.6337	H−1 → L+1(94)
	1.91 (650)	0.7030	H−1 → L+2(93)

---

4. Conclusion

2-Aminoethoxy metallophthalocyanines with Fe, Ni, and Zn as the metal atom can be covalently anchored to a graphene sheet to form stable AEMPc–graphene composites, which retain the characteristics of graphene. These complex molecules may serve as the model for the recently synthesized (2-aminoethoxy)(tri-tert-butyl) zinc phthalocyanine–graphene hybrid material which is used in photo electrochemical cell. In all three complexes, there is a strong covalent bond through the amino group of the functionalized MPc. The computed energy of interaction between the graphene and AEMPc at the B3LYP/6-31G(d)/LANL2DZ level is 79.3 kcal mol⁻¹ for all three metal atoms. Dipole moments and polarizabilities are improved in the hybrid system compared to those of the graphene itself. The strong binding between the two moieties is justified by the appearance of a stretching vibrational frequency at 2916 cm⁻¹ due to the newly formed C–H bond directed above the graphene plane. Other vibrational frequencies associated with the C–N bond at the point of contact lie in the range 1100–1200 cm⁻¹. The intense Raman G and D bands of graphene are retained in the Raman spectra of the AEZnPc–graphene hybrid molecule and the ratio of their intensities (I_D/I_G) follows a similar shift as observed experimentally for a similar hybrid material. Two strong absorption peaks of AEMPc–graphene located in the range 600–630 nm for all three metal atoms are associated with the electron transfer between graphene and AEMPc as revealed from the nature of the important transitions. The analyses of the corresponding molecular orbitals confirm the charge transfer bands in all three cases. Due to large oscillator strengths in both UV-vis and IR regions, and nonlinear optical properties, these covalently grafted hybrid complexes can be used as potential candidates for optoelectronic devices. The overall findings obtained from the calculations using the B3LYP functional are found to be comparable to those computed using the long range meta-hybrid functionals like M06-2X, CAM-B3LYP, and wB97XD.

Acknowledgements

PNS would like to thank CSIR, Govt. of India for Research Associateship.

References

1. Phthalocyanines: Properties and Applications, ed. C. C. Leznoff and A. B. P. Lever, VCH Publishers, New York, 1990–1996, vol. 1–4.
2. R. W. Wagner, J. S. Lindsey, J. Seth, V. Palaniappan and D. F. Bocian, J. Am. Chem. Soc., 1996, 118, 3996–3997.
3. D. A. Fernández, J. Awruch and L. E. Dicelio, Photochem. Photobiol., 1996, 63, 784–792.
4. D. S. Lawrence and D. G. Whitten, Photochem. Photobiol., 1996, 64, 923–935.
5. H. Ali and J. E. van Lier, Chem. Rev., 1999, 99, 2379–2450.
6. E. Ben-Hur and W.-S. Chan, Phthalocyanines in Photobiology and Their Medical Applications, in The Porphyrin Handbook, ed. K. M. Kadish, K. M. Smith and R. Guilard, Academic Press, Boston, 2003, vol. 19, pp. 1–35.
7. M. Ince, M. V. Martinez-Diaz, J. Barberá and T. Torres, J. Mater. Chem., 2011, 21, 1531–1536.
8. H. Hayashi, W. Nihashi, T. Umeyama, Y. Matano, S. Seki, Y. Shimizu and H. Imahori, J. Am. Chem. Soc., 2011, 133, 10736–10739.
9. G. de la Torre, P. Vázquez, F. Agulló-López and T. Torres, Chem. Rev., 2004, 104, 3723–3750.
10. Y. Chen, M. Hanack, W. J. Blau, D. Dini, Y. Liu, Y. Lin and J. Bai, J. Mater. Sci., 2006, 41, 2169–2185.
11. J. Simon and J.-J. André, Molecular Semiconductors, Springer, Berlin, 1985.
12. K. M. Kadish, K. M. Smith and R. Guilard, Applications of Phthalocyanines, in The Porphyrine Handbook, Academic Press, San Diego, 2003, vol. 19.
13. S. Sanvito, J. Mater. Chem., 2007, 17, 4455–4459.
14. L. Bogani and W. Werndorfer, Nat. Mater., 2008, 7, 179–186.
15. A. Gouloumis, S. G. Liu, A. Sastre, P. Vazquez, L. Echegoyen and T. Torres, Chem.–Eur. J., 2000, 6, 3600–3607.
16. L. Martin-Gomis, K. Ohkubo, F. Fernandez-Lazaro, S. Fukuzumi and A. Sastre-Santos, Org. Lett., 2007, 9, 3441–3444.
17. J. R. Pinzón, C. M. Cardona, M. A. Herranz, M. E. Plonska-Brzezinska, A. Palkar, A. J. Athans, N. Martin, A. Rodriguez-Fortea, J. M. Poblet, G. Bottari, T. Torres, S. S. Gayathri, D. M. Guldi and L. Echegoyen, Chem.–Eur. J., 2009, 15, 864–877.
18. S. Campidelli, B. Ballesteros, A. Filoramo, D. Diaz, G. de la Torre, T. Torres, G. M. A. Rahman, C. Ehli, D. Kiessling, F. Werner, V. Sgobba, D. M. Guldi, C. Cioffi, M. Prato and J. P. Bourgoin, J. Am. Chem. Soc., 2008, 130, 11503–11509.
19. M.-S. Liao and S. Scheiner, J. Chem. Phys., 2001, 114, 9780–9791.
20. A. B. P. Lever, S. R. Pickens, P. C. Minor, S. Licoccia, B. S. Ramaswamy and K. Magnell, J. Am. Chem. Soc., 1981, 103, 6800–6806.
21. P. C. Minor, M. Gouterman and A. B. P. Lever, Inorg. Chem., 1985, 24, 1894–1900.
22. R. Berera, I. H. M. van Stokkum, G. Kodis, A. E. Keirstead, S. Pillai, C. Herrero, R. E. Palacios, M. Vengris, R. van Grondelle, D. Gust, T. A. Moore, A. L. Moore and J. T. M. Kennis, J. Phys. Chem. B, 2007, 111, 6868–6877.
23. L. Edwards, D. H. Dolphin and M. Gouterman, J. Mol. Spectrosc., 1970, 35, 90–109.
24. T. N. Nyokong, Z. Gasyna and M. J. Stillman, Inorg. Chem., 1987, 26, 1087–1095.
25. J. Mack and M. J. Stillman, J. Phys. Chem., 1995, 99, 7935–7945.
26. T. C. VanCott, J. L. Rose, G. C. Misener, B. E. Williamson, A. E. Schrimpf, M. E. Boyle and P. N. Schatz, J. Phys. Chem., 1989, 93, 2999–3011.
27. D. H. Metcalf, T. C. VanCott, S. W. Snyder, P. N. Schatz and B. E. Williamson, J. Phys. Chem., 1990, 94, 2828–2832.
28. G. Ricciardi, A. Rosa and E. J. Baerends, J. Phys. Chem. A, 2001, 105, 5242–5254.
29. H. Bai, C. Li and G. Shi, Adv. Mater., 2011, 23, 1089–1115.
30. P. V. Kamat, J. Phys. Chem. Lett., 2011, 2, 242–251.
31. K. P. Loh, Q. Bao, P. K. Ang and J. Yang, J. Mater. Chem., 2010, 20, 2277–2289.
32. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228–240.
33. J. Malig, N. Jux, D. Kiessling, J.-J. Cid, P. Vázquez, T. Torres and D. M. Guldi, Angew. Chem., Int. Ed., 2011, 50, 3561–3565.
34. N. Karousis, J. Ortiz, K. Ohkubo, T. Hasobe, S. Fukuzumi, Á. Sastre-Santos and N. Tagmatarchis, J. Phys. Chem. C, 2012, 116, 20564–20573.
35. W. Tao, Y.-H. Kan, S.-X. Wu, H.-B. Li, L.-K. Yan, S.-L. Sun and Z.-M. Su, J. Mol. Graphics Modell., 2012, 33, 26–34.
36. J. Thompson, A. Crossley, P. D. Nellist and V. Nicolosi, J. Mater. Chem., 2012, 22, 23246–23253.
37. G. I. Cárdenas-Jirón, P. Leon-Plata, D. Cortes-Arriagada and J. M. Seminario, J. Phys. Chem. C, 2011, 115, 16052–16062.
38. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789.
39. M. J. Frisch, et al., Gaussian 09 Revision A.02, Gaussian, Inc., Wallingford, CT, 2009.
40. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
41. N. Marom and L. Kronik, Appl. Phys. A, 2009, 95, 159–163.
42. N. Marom and L. Kronik, Appl. Phys. A, 2009, 95, 165–172.
43. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
44. T. Yanai, D. P. Tew and N. C. Handy, Chem. Phys. Lett., 2004, 393, 51–57.
45. J.-D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008, 10, 6615–6620.
46. Y. Zhao and D. G. Truhlar, Chem. Phys. Lett., 2011, 502, 1–13.
47. Y. Tawada, T. Tsuneda, S. Yanagisawa, T. Yanai and K. Hirao, J. Chem. Phys., 2004, 120, 8425–8433.
48. M. Kamiya, H. Sekino, T. Tsuneda and K. Hirao, J. Chem. Phys., 2005, 122, 234111.
49. J. M. Robertson and I. Woodward, J. Chem. Soc., 1937, 219–230.
50. J. F. Kirner, W. Dow and W. R. Scheidt, Inorg. Chem., 1976, 15, 1685–1690.
51. W. R. Scheidt and W. Dow, J. Am. Chem. Soc., 1977, 99, 1101–1104.
52. C.-Y. Ruan, V. Mastryukov and M. Fink, J. Chem. Phys., 1999, 111, 3035–3041.
53. V. Georgakilas, M. Otyepka, A. B. Bourlinos, V. Chandra, N. Kim, K. C. Kemp, P. Hobza, R. Zboril and K. S. Kim, Chem. Rev., 2012, 112, 6156–6214.
54. M. Baraket, R. Stine, W. K. Lee, J. T. Robinson, C. R. Tamanaha, P. E. Sheehan and S. G. Walton, Appl. Phys. Lett., 2012, 100, 233123.
55. D. W. Boukhvalov and M. I. Katsnelson, J. Phys.: Condens. Matter, 2009, 21, 344205.
56. E. C. Anota, A. T. Soto and G. H. Cocoletzi, Appl. Nanosci., 2014, 4, 911–918.
57. K. Z. Milowska and J. A. Majewski, Phys. Status Solidi B, 2013, 250, 1474–1477.
58. N. Kheirabadi and A. Shafiekhani, Physica E, 2013, 47, 309–315.
59. T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580–592.
60. H. L. Schmider and A. D. Becke, J. Mol. Struct.: THEOCHEM, 2000, 527, 51–61.
61. K. S. Rao, J. Senthilnathan, Y.-F. Liu and M. Yoshimura, Sci. Rep., 2014, 4, 4237–4242.
62. C. C. J. Roothaan, Rev. Mod. Phys., 1951, 23, 69–89.
63. C. Møller and M. S. Plesset, Phys. Rev., 1934, 46, 618–622.
64. L. T. Ueno, A. E. H. Machado and F. B. C. Machado, J. Mol. Struct.: THEOCHEM, 2009, 899, 71–78.
65. N. M. Speirs, W. J. Ebenezer and A. C. Jones, Photochem. Photobiol., 2002, 76, 247–251.
66. A. J. Garza, G. E. Scuseria, S. B. Khan and A. M. Asiri, Chem. Phys. Lett., 2013, 575, 122–125.
67. B. Champagne, E. A. Perpète, D. Jacquemin, S. J. A. van Gisberger, E.-J. Baerends, C. Soubra-Ghaoui, K. A. Robins and B. Kirtman, J. Phys. Chem. A, 2000, 104, 4755–4763.
68. O. Loboda, R. Zaleśny, A. Avramopoulos, J.-M. Luis, B. Kirtman, N. Tagmatarchis, H. Reis and M. G. Papadopoulos, J. Phys. Chem. A, 2009, 113, 1159–1170.
69. G. Zhang and C. B. Musgrave, J. Phys. Chem. A, 2007, 111, 1554–1561.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra14628c

---