PCCP View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PAPER
Cite this: DOI: 10.1039/c7cp06103j
View Journal
A unified understanding of the direct coordination of NO to first-transition-row metal centers in metal–ligand complexes Hyunjoo Lee and Joongoo Kang
*
The binding of nitric oxide (NO) to heme-proteins is an important biochemical process involved in a variety of physiological functions. Here, using hybrid density-functional calculations, we systematically investigate the adsorption of NO to first-transition-row metal centers in metal–ligand complexes. Through the comparative study for different transition metal (TM) centers, we provide a unified understanding of the microscopic interactions of NO with the TM centers and related chemical trends. We found that as the atomic number of the TM center increases, the binding strength of NO is largely reduced from 207 kJ mol
1
to near zero due to the low d-orbital energies for late TM centers. The
Received 7th September 2017, Accepted 3rd October 2017
intermolecular spin coupling between the localized spins at the TM center and the NO molecule is
DOI: 10.1039/c7cp06103j
way to avoid the energy penalty associated with the electron occupation in the antibonding states of the
generally antiferromagnetic, except for the case of Sc. The spin–spin coupling is determined in such a NO-bound complex. The adsorption strength of NO is generally larger than of CO because the unpaired
rsc.li/pccp
electron of NO occupies the associated bonding state.
Introduction Nitric oxide (NO) and carbon monoxide (CO) are important signaling molecules involved in a variety of physiological processes.1,2 For example, NO plays an important role in synaptic plasticity,3,4 and abnormal NO signaling can contribute to various neurodegenerative pathologies such as strokes, Alzheimer’s disease, multiple sclerosis, and Parkinson’s disease.5 Many physiological functions of NO involve interactions with heme-proteins. One way in which NO interacts with hemeproteins is through the direct coordination of NO with an Fe center to form an Fe–nitrosyl complex. Despite important biological functions mediated by NO and CO, these molecules are also associated as a toxic gas because of their strong binding to hemoglobin and the resulting poisoning. Recently, it was demonstrated that heme-based scavenger molecules, such as a mutant five-coordinate neuroglobin,6 can be used to effectively eliminate CO from CO-saturated hemoglobin. The binding of NO to heme-proteins is a crucial step in NO-mediated physiological processes. Many studies have been done on the binding of NO to a Fe-porphyrin because of its biochemical importance.7–15 Recently, the adsorption study of NO was extended to include non-heme Fe(III) complexes16 and
Department of Emerging Materials Science, DGIST, Daegu 711-873, Korea. E-mail: [email protected]
This journal is © the Owner Societies 2017
other metalloporphyrins such as Mn- and Co-porphyrins.17–20 It remains an important issue to understand the adsorption of NO to first-row transition metal centers for two reasons. First, a comparative study of NO binding for different transition metal (TM) centers in a TM–ligand complex can provide a unified understanding of the microscopic interactions of NO with the TM centers and related chemical trends. Second, such a comparative study can also provide useful insights into designing a platform for reversible NO binding. For application purpose, it is desirable to develop TM–ligand complexes with proper NO binding strengths. NO is endogenously produced by the human body as a signaling molecule. As an alternative way of providing NO to a target in the body, one may consider NO delivery from outside for medical treatment.21–24 Here, we focus on the direct coordination of NO to a TM center in a TM–ligand complex. In this study, two TM–nitrosyl complexes are considered for TM = Sc, Ti, V, Cr, Mn, Fe, Co, and Ni (Fig. 1): (i) TM(P)(NO), where P = porphyrin, and (ii) [TM(Me3-TPADP)(CH3CN)(NO)]2+, where Me3-TPADP = 3,6,9-trimethyl-3,6,9-triaza-1(2,6)-pyridinacyclodecaphane.25 For comparison, CO adsorption is also considered for the same TM–ligand complexes. The rest of the paper is organized as follows. First, we construct a theoretical framework for understanding the mechanisms and overall chemical trends of the NO and CO binding to the TM–ligand complexes. Next, we present our density functional theory (DFT) calculation results for the NO adsorption and discuss how the main results can be understood
Phys. Chem. Chem. Phys.
View Article Online
Paper
PCCP
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Fig. 2 Two types of the p–d orbital interaction responsible for the binding of NO (or CO) to the TM center of a TM-porphyrin: (a) a p-bonding state between a TM dxz and an NO pxpxp* orbital forming a linear TM–N–O bond and (b) a symmetry-allowed bonding state between a TM dz2 and NO ppp* orbital in a bent binding geometry. Fig. 1 Atomic structures of two TM–nitrosyl complexes: (a) TM(P)(NO), where P = porphyrin, and (b) [TM(Me3-TPADP)(CH3CN)(NO)]2+, where Me3-TPADP = 3,6,9-trimethyl-3,6,9-triaza-1(2,6)-pyridinacyclodecaphane. C, gray; H, white; N, blue; O, red; and TM, orange.
within our theoretical framework. Finally, we compare the binding strengths of NO and CO to TM-porphyrins.
Calculation methods To investigate the binding mechanisms of NO and CO to the TM–ligand complexes in Fig. 1, we performed spin-polarized density functional calculations using the Vienna ab initio Simulation Package.26 Our calculations employed the projector augmented wave (PAW) method27,28 with an energy cutoff of 400 eV for the plane wave part of the valence wave functions. For 3d TM with the atomic number N, we used N 18 valence electrons. The core electrons were treated as frozen cores within the PAW method. For the charged molecules such as [TM(Me3TPADP)(CH3CN)(NO)]2+, a uniform background charge was added to maintain the charge neutrality of the system in our plane-wave basis calculations. The energy convergence criterion for the electronic self-consistent calculations was set to 0.01 meV. Atomic structures of TM–ligands complexes were relaxed within 0.05 eV Å 1. Throughout the paper, the results were obtained by using the PBE0 hybrid density functional29–31 except if mentioned otherwise. For the NO adsorption to TM-porphyrins, we also present the results of other hybrid functionals such as B3LYP (ref. 32) and B3LYP* (ref. 33 and 34).
Results and discussion A.
Molecular-orbital theory of NO and CO adsorption
In this section, we develop a molecular-orbital theory for understanding adsorption mechanisms and overall chemical trends in the NO adsorption to the first-row TM centers in TM-porphyrin. The molecular orbitals of NO responsible for its adsorption are the p-antibonding states (ppp*) of the p orbitals of the N and O atoms (Fig. 2).12,14 There are two NO ppp* states having different orientations for the p orbitals. NO interacts with TM-porphyrin through the direct coordination of the N atom with the TM center, because the unpaired electron in the NO ppp* state is distributed more around the N atom than around the O atom. NO has a magnetic moment of 1 mB because of the unpaired electron in the ppp* state. Depending on the
Phys. Chem. Chem. Phys.
spin direction of NO with respect to the net spin of the TM center, the NO adsorption is characterized by either ferromagnetic or antiferromagnetic spin–spin coupling. As we will discuss later, the binding angle of NO determines which TM d-orbitals are able to interact with the NO ppp* states. The energy difference between the pair of the molecular orbitals, in turn, determines the degree of their hybridization and thus the binding strength of NO. By considering the interaction between the TM d-orbital states and the NO ppp* states, we develop a theoretical framework to understand the NO adsorption to TM-porphyrins, with the following expected properties: (i) the adsorption symmetry of NO dictates which TM d-orbitals are symmetry-allowed to interact with the NO ppp* states. (ii) The adsorption strength generally gets reduced as the atomic number of the TM center increases. (iii) The intermolecular spin coupling in the NO adsorption is generally antiferromagnetic except for the case of TM = Sc. (iv) For a TM-porphyrin, the adsorption strength of NO is expected to be larger than of CO. Each of these is discussed below. (i) The adsorption symmetry associated with the binding angle of NO dictates which TM d-orbitals are involved in the NO adsorption. Throughout the paper, the x-axis in Fig. 1 is chosen as the direction along which the NO molecule is bent, and the z-axis is chosen as a normal vector to the TM-porphyrin plane. In general, the TM–N bond is also slightly tilted with respect to the z-axis for greater interorbital coupling.13 The tilt angle of the TM–N bond is much smaller than the binding angle y. Therefore, for simplicity, the tilt angle is not shown in Fig. 1 and 2. For a linear binding of NO with y = 1801, the TM dxz couples to the NO pxpxp* state,12 forming a p-like bonding state and a corresponding antibonding state (see Fig. 2a for the bonding state). Likewise, the TM dyz couples to the NO pypyp* state to form a metal–nitrosyl complex. In contrast, for the TM dz2, the interorbital coupling is symmetry forbidden because the dz2 and pxpxp* (or pypyp*) states have different reflection parities with respect to the yz (or xz) plane (Fig. 2b). The reflection symmetry should thus be broken to enable the orbital coupling for dz2, as in the bent binding in Fig. 2b. Note that when NO is bent along the x-axis, the reflection symmetry with respect to the xz plane is preserved, making the p-like orbital coupling of the dyz state and the NO pypyp* state still possible. Thus, both TM dz2 and dyz can participate in the NO adsorption for the bent binding, while TM dxz and dyz are involved for the linear NO binding. Hereafter, the two different types of the molecular interactions will be referred to as the dz2/dyz–ppp* interaction and the dxz/dyz–ppp* interaction, respectively.
This journal is © the Owner Societies 2017
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PCCP
From the symmetry argument, it is thus expected that when the dxz/dyz–ppp* interaction is dominant over the dz2/dyz–ppp* interaction, the NO adsorption energetically favors a linear binding. Conversely, in the case that the dz2/dyz–ppp* interaction is strong, the NO adsorption takes a bent structure. The previous notion12 is that the bent TM–NO binding is associated with TM(III)–NO , while the linear binding is associated with TM(I)– NO+. However, this formal assignment is misleading from a computational point of view: in both binding structures, we will see that the NO molecule bound to the TM center is largely spinpolarized by its unpaired electron, indicating that it remains close to charge neutral. (ii) The adsorption strength of NO generally gets reduced as the atomic number of the TM site increases. The d-orbital energy of the first-row 3d TM center in a TM-porphyrin decreases as the atomic number increases. It turns out that the d-orbitals of early TMs (e.g., Sc) lie close in energy to the NO ppp* state, while for late TMs (e.g., Fe) the d-orbital energies are much lower than those of the NO ppp* states. For late TMs, the interactions between the TM d-orbitals and the NO ppp* states become relatively weak because of the large separation in their orbital energies. Therefore, the NO adsorption strength would be weak for late TMs, as compared to the case of early TMs. (iii) The intermolecular spin coupling is generally antiferromagnetic except for the case of Sc. For a moment, we only consider the dxz/dyz–ppp* interaction to discuss the spin–spin interaction for a linear adsorption of NO. (The same argument holds for the dz2/dyz–ppp* interaction in a bent adsorption.) With an exception for TM = Sc, the TM dxz/dyz orbitals in the majority (i.e., spin-up) spin state are fully occupied by two electrons of the TM cation. When it comes to the spin state of the NO adsorbate, the NO ppp* states can be occupied either by a spin-up electron or by a spin-down electron. In the former case, the dxz/dyz–ppp* interaction will lead to the two fullyoccupied bonding states in the spin-up state and the two antibonding states occupied by a single spin-up electron. In the latter case, however, the intermolecular interaction for the spin-up state gives rise to the two fully-occupied bonding states and the two empty antibonding states. The spin-down electron of the NO ppp* now can fill the bonding state in the spin-down state, further lowering the energy of the NO-bound complex. Consequently, the antiferromagnetic spin–spin coupling is energetically more favored than the ferromagnetic coupling. In other words, the spin–spin coupling is determined in such a way to avoid the energy penalty associated with the electron occupation in the antibonding states. The ferromagnetic spin–spin coupling in the NO adsorption is possible only for TM = Sc because the dxz/dyz orbitals are now occupied by only a single electron. The dxz/dyz–ppp* interaction in the ferromagnetic spin configuration then leads to the fullyoccupied bonding states and the empty antibonding states in the spin-up state. (iv) For a TM-porphyrin, the adsorption strength of NO is generally larger than of CO. For the adsorption of CO, the empty ppp* states of CO interact with the occupied TM d-orbitals to form the TM–CO bond. Because of the same
This journal is © the Owner Societies 2017
Paper
orbital symmetry (i.e., ppp*) for NO and CO, the types of orbital coupling in the CO adsorption to a TM-porphyrin is also determined by the adsorption symmetry associated with the binding angle of CO. As we will discuss later, the unpaired electron in the NO ppp* state occupies a bonding state of the NO-bound complex, enhancing the binding energy of NO. For the CO adsorption, however, the electron occupation of the bonding states is likely to be reduced by one, because CO has one electron less than NO. Therefore, the adsorption strength of CO is expected to be smaller than of NO. B.
Comparison with the DFT results
In this section, we present our calculation results of the NO and CO adsorption to TM–ligand complexes and discuss how our main results for different TM centers could be understood within the theoretical framework developed in the previous section. In PBE0 calculations, we found that the binding energy of NO to a TM-porphyrin tends to decrease from 207.0 kJ mol 1 for Sc to 0.2 kJ mol 1 for Ni as the atomic number of the 3d TM increases (Table 1). Here, we do not include the results for Cu and Zn, because NO does not directly bind to these TM centers. Fig. 3 shows that the same chemical trend is obtained regardless of the choice of the hybrid functional (see also Table 2). The binding energy of NO to Feporphyrin is calculated to be 80.0 kJ mol 1 (PBE0), 75.0 (B3LYP), and 104.2 (B3LYP*). The B3LYP* value agrees well with the binding energy of 108 kJ mol 1 to Fe-tetraphenylporphyrin (FeTPP+), which was estimated from the radiative association rates.11 As pointed out in the previous section, the interaction between the TM d-orbital and the NO ppp* state becomes relatively weak for late TMs because of the large energy separation in the orbital energies. Therefore, the NO adsorption strength is generally weak for late TMs, relative to the case of early TMs. Fig. 4 compares the NO binding energies to TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ systems in Fig. 1. We found Table 1 Binding energies (EB) for NO and CO molecules to the first-row 3d TM centers in a TM-porphyrin. Selected structural parameters are listed for TM(P)(XO), where X = N or C. The hybrid functional calculations were done with PBE0. The spin–spin coupling in the table denotes the type of the intermolecular spin interaction between the two localized spins at the TM center and the NO molecule, respectively
TM
EB (kJ mol 1)
dTM–XO (Å)
dX–O (Å)
Spin–spin coupling
NO
Sc Ti V Cr Mn Fe Co Ni
207.0 195.1 149.2 60.3 58.8 80.0 37.7 0.2
2.02 1.80 1.77 1.71 1.88 1.79 1.91 1.98
1.20 1.20 1.19 1.19 1.19 1.17 1.16 1.15
FM Mixed Mixed AFM AFM AFM AFM AFM
CO
Sc Ti V Cr Mn Fe Co Ni
68.7 109.5 99.3 2.6 11.7 26.8 24.1 4.7
2.30 2.04 2.00 2.58 2.45 1.73 1.98 2.00
1.14 1.14 1.14 1.13 1.13 1.15 1.13 1.13
— — — — — — — —
Phys. Chem. Chem. Phys.
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Paper
PCCP
Fig. 3 Comparison of the NO binding energies (EB) to TM-porphyrin in different hybrid functionals: PBE0, B3LYP, and B3LYP*.
Fig. 4 Chemical trends in the binding energies (EB) of NO to TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ in PBE0. For comparison, the binding energies of H2O to TM-porphyrin are also presented.
Table 2 Binding energies (EB) and structural parameters in B3LYP and B3LYP* for the TM(P)(NO) with the first-row 3d TM centers as the binding site. The intermolecular spin coupling between the TM center and the NO molecule is the same as in PBE0
TM
EB (kJ mol 1)
dTM–NO (Å)
dN–O (Å)
B3LYP
Sc Ti V Cr Mn Fe Co Ni
212.5 194.3 148.1 58.3 58.0 75.0 36.7 17.2
2.04 1.81 1.78 1.71 1.88 1.80 1.90 1.96
1.21 1.21 1.19 1.19 1.20 1.18 1.17 1.16
B3LYP*
Sc Ti V Cr Mn Fe Co Ni
216.7 214.3 170.9 100.4 103.0 104.2 62.2 12.5
2.04 1.82 1.68 1.68 1.84 1.76 1.84 1.88
1.21 1.21 1.20 1.19 1.20 1.18 1.18 1.16
that the overall chemical trend across the first-row TMs is the same for both TM–ligand complexes. For comparison, we also present the binding energies of H2O to TM-porphyrin. We found that especially for early TMs, the NO binding strength is significantly stronger than that of H2O. In Table 1, we summarize our DFT results for the intermolecular spin interaction for the NO adsorption. For TM = Sc, the spin–spin coupling is ferromagnetic, while for Cr and later TMs, it is antiferromagnetic. For Ti and V in between, the spin– spin interaction has a mixed character, which will be discussed later. To demonstrate how the spin–spin coupling depends on the TM center in a TM-porphyrin, we take four TM centers, TM = Sc, V, Cr, and Co, as examples (Fig. 5). For the Sc center, the intermolecular spin interaction is ferromagnetic. Fig. 5a shows the projected density of states (PDOS) of the NO-bound Sc-porphyrin. The NO adsorption to the Sc center involves the dxz/dyz–ppp* interaction. Therefore, in the PDOS, the molecular orbitals of the NO-bound complex are projected to the Sc dxz/dyz orbitals and the px/py orbitals of the N and O atoms in NO. When NO is far apart from the
Phys. Chem. Chem. Phys.
Fig. 5 Three different types of the intermolecular spin coupling between the localized magnetic moments at the TM center and the NO adsorbate: (a) ferromagnetic spin–spin interaction for TM = Sc, (b) mixed spin–spin interaction for V, and antiferromagnetic spin interaction for (c) Cr and (d) Co. Projected densities of states (PDOS) are shown for the selected p orbitals of NO and TM d orbitals involved in the NO adsorption. Two vertical dashed lines denote the energies of the HOMO and LUMO states. The p–d bonding (or antibonding) states are indicated by black (or green) arrows. The energy is referenced to the average orbital energy of the NO ppp* states. In the spin-density plots on the right panels, the yellow surface is for the majority spin-up charge density, while the blue surface is for the minority spin-down charge density.
This journal is © the Owner Societies 2017
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PCCP
binding site, the Sc dxz/dyz orbitals are found to be slightly higher than the NO ppp* states in the orbital energies. The Sc dxz/dyz orbitals are occupied by an electron because the Sc center is in a 2+ charge state. When NO binds to the Sc center, the resulting p-like bonding states in the spin-up state are fully occupied, while the p-like antibonding states are empty. For the spin-down state, however, both the bonding and antibonding states are empty. The intermolecular spin interaction is ferromagnetic, as confirmed by the spin-density plot in the right panel of Fig. 5a. Note that the spin-up electrons (yellow) are distributed on both the Sc center and the NO molecule. The total magnetic moment of the complex is thus m = 2 mB, which can be formally assigned to the local spin of 1 mB at the Sc center and that of 1 mB at NO. In contrast, the NO adsorption to the Cr center and later TMs is characterized by the antiferromagnetic spin–spin interaction (Table 1). For Cr, the NO adsorption occurs through the dxz/dyz–ppp* orbital interaction. The NO-free Cr-porphyrin is in a high-spin state with m = 4 mB. When NO is adsorbed to the Cr center, however, the spin state of the Cr site changes to an intermediate-spin state with m = 2 mB. Regardless of the spin states, the spin-up dxz/dyz orbitals are fully occupied by two spin-up electrons. However, the spin-down dxz/dyz orbitals are occupied by a single electron only if the Cr center is in the intermediate-spin state. The relatively small binding energy for Cr can be in part attributed to the spin crossover induced by the NO adsorption (Fig. 3). As discussed in the previous section, the fully occupied TM dxz/dyz orbitals in the spin-up state require the antiferromagnetic spin–spin coupling. Indeed, the spindensity plot in Fig. 5c clearly shows the spin-up density around the Cr center and the spin-down density around the NO molecule. As a result, the net magnetic moment of the NO-bound complex is m = 1 mB. The PDOS in Fig. 5c shows that for both spin-up and -down states, the p-like dxz/dyz–ppp* bonding states are fully occupied, while the corresponding antibonding states are empty. Note that the NO p-orbital component in the p-like bonding states is substantially larger for the spin-down state, leading to the spindown density around the NO molecule in the spin-density plot. For the bent NO binding, the spin–spin coupling is also found antiferromagnetic. Fig. 5d shows the results for the bent NO adsorption to the Co center through the dz2/dyz–ppp* interaction. For an NO-free Co-porphyrin, the Co dz2 and dyz orbitals in the spin-up state are both fully occupied, which is a necessary condition for the antiferromagnetic spin–spin interaction. Indeed, the spin-density plot in Fig. 5d shows the antiferromagnetic spin–spin coupling. We found that the NO-free Co-porphyrin is in a low-spin state with m = 1 mB. The net magnetic moment of the NO-bound complex is then reduced to m = 0 mB due to the antiferromagnetic coupling with NO. The bent singlet ground state agrees with the previous DFT result.19 For the Ti and V centers, the spin–spin coupling exhibits a mixed character: the spin-density plot for the V center in Fig. 5b shows that the spin-up density is distributed around the TM center, while around the NO molecule, both the spin-up (yellow) and spin-down (blue) components are found with the orbital characters of pypyp* and pxpxp*, respectively. For both Ti
This journal is © the Owner Societies 2017
Paper
and V, the NO binding is linear with y = 1801. When compared to the case of Sc, the number of the spin-up electrons in the TM dxz/dyz orbitals increases from one to two, while the number of the spin-down electrons in the dxz/dyz orbitals remains zero. Thus, according to the argument in the previous section, the unpaired electron in the NO ppp* orbitals should be in the spin-down state to couple with the empty TM d-orbitals in the spin-down state. For the spin-up state, the p-like bonding states, dxz–pxpxp* and dyz–pypyp*, at around 3 eV are occupied by two electrons (Fig. 5b). For the spin-down state, the dxz–pxpxp* bonding state, which lies at around 2.5 eV, is occupied by the spin-down electron in the NO pxpxp* orbital. In contrast, the dyz–pypyp* bonding state in the spin-down state is unoccupied and lies at around 0 eV. Note that when compared to the case of Cr with the fully-occupied bonding states in both spin states (Fig. 5c), a single electron is lacking in the dyz–pypyp* bonding states in the spin-down state. Therefore, the spin-density distribution for Ti and V has the mixed nature with both the spin-up and spin-down components at NO (Fig. 5b). Fig. 6 shows how the adsorption energy changes with the binding angle y of NO for several TM-porphyrins. As we discussed from the symmetry argument in the previous section, the NO adsorption energetically favors a linear binding when
Fig. 6 Dependence of the adsorption energy (D E) of NO on the TM–N–O angle y in TM(P)(NO): (a) TM = Co and Cr and (b) TM = Mn and Fe in different spin states. In the right panels, the charge densities of representative bonding states of TM(P)(NO) are plotted for the specified TM centers and the binding angles.
Phys. Chem. Chem. Phys.
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Paper
the dxz/dyz–ppp* interaction is dominant over the dz2/dyz–ppp* interaction. Conversely, in the case that the dz2/dyz–ppp* interaction is strong, the NO adsorption takes a bent structure. For example, the NO binding to the Cr center is linear with the energy minimum at y = 1801, indicating that the p-like dxz/dyz–ppp* orbital interaction is responsible for the NO adsorption (Fig. 6a). In the right panel of the figure, the charge-density plot of one of the p-like bonding states is shown to demonstrate the type of the intermolecular orbital coupling. For the bent NO adsorption to Co, however, the minimum energy is obtained at y = 1221. The NO adsorption to the Co center involves the s-like dz2–ppp*, as well as the p-like dyz–pypyp* interaction. For some TM centers such as Mn and Fe, both types of the orbital interactions can compete for the NO adsorption in certain spin states. As the low-energy states of NO-bound Mn-porphyrin, we obtained the following two spin states (Fig. 6b in B3LYP*): the intermediate-spin (IS) triplet state with m = 2 mB and the low-spin (LS) singlet state with m = 0 mB. For the IS state, we obtained a relatively flat energy curve for the angles from y = 1301 to 1801, with the minimum at y = 1421, because of the competition between the two types of the intermolecular interaction. For the LS state, however, the p-like dxz/dyz–ppp* orbital coupling becomes dominant over the s-like dz2–ppp* coupling. As a result, the LS state favors the linear NO adsorption with a large bending stiffness. The same spin state–structure relationship (i.e., the bent triplet and the linear singlet) is obtained for different hybrid functionals. As the previous work pointed out,18,33 however, the energy difference between these two spin states are strongly functional dependent. In a previous work on Fe(II)–S complexes,33 it was shown that the energy difference between the high- and lowspin states linearly depends on the coefficient of exact exchange, because a high-spin state is generally favored in Hartree–Fock-type theories. Indeed, we found that the relative stability of the singlet state is systematically enhanced as the coefficient of the exact exchange in our hybrid functional calculations decreases from 25% (PBE0) to 20% (B3LYP) and to 15% (B3LYP*). In our B3LYP* results, the triplet state is still more stable than the singlet state, although the energy difference is small (Fig. 6b). However, it should be noted that a previous experiment18 at low temperature, combined with DFT calculations, suggests that the singlet state is the true ground state of NO-bound Mn-porphyrin.19 For Fe-porphyrin, we similarly found that there exist the competing orbital interactions for the NO adsorption, which lead to the relatively flat energy curves in Fig. 6b. For the IS ground state with the net spin of m = 1 mB, the NO-bound complex is bent with the angle of 1501, in good agreement with the previous experimental result (y = 149.21).8 Note that the binding angle y for the Fe center is substantially larger than for the Co-porphyrin (y = 1221) because of the competition between the s-like dz2–ppp* and p-like dxz–ppp* orbital couplings for the Fe center. As the second-lowest energy state in B3LYP*, we also identified the linear high-spin (HS) state with the net spin of m = 3 mB for the NO-bound complex. The adsorption energy
Phys. Chem. Chem. Phys.
PCCP
Fig. 7 Comparison of the binding energies (EB) of NO and CO to TM-porphyrin for different first-row TM centers in PBE0 calculations.
curve is nearly flat around y = 1801, relative to the energy curve for the linear LS Mn-porphyrin. Finally, we compare the binding strengths of NO and CO to TM-porphyrins in Fig. 7. The adsorption energy of CO is substantially smaller than of NO, as expected in the previous section. The crucial difference between the two molecules is that CO has one electron less than NO. For early TM centers such as Sc, Ti, and V, one electron is thus less available for the bonding states of the CO adsorption, leading to the relatively weak binding strengths for CO. For the Cr center, the situation is yet more complicated: the NO adsorption induces the spin crossover of the Cr center from the HS state to the IS state to maximize the number of electrons in the associated bonding states, as discussed earlier in this paper. The CO adsorption, however, does not induce the spin crossover, because the energy gain for the CO adsorption with the Cr center in the IS state is less than the energy cost associated with the spin crossover. Then, the Cr center remains in the HS state after the CO adsorption. The number of electrons in the bonding states is only two for the CO adsorption, while it is four for the NO adsorption. Therefore, the reduced number of electrons in the bonding states explains the reduced CO binding strength for TM = Sc, Ti, V, and Cr. It is also expected that the CO ppp* orbitals would lie at higher energy than the NO ppp* orbitals due to the less electronegative C atom in CO, leading to the larger energy separation between the TM d-orbitals and the CO ppp* states for late TMs and thus less strong binding.
Conclusions We performed hybrid density-functional calculations to investigate the direct coordination of NO to first-row TM centers in TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ systems. Through the comparative study for different 3d TM centers, we found that as the atomic number of the TM center increases, the adsorption strength of NO generally gets reduced due to the low d-orbital energies for late TM centers. The NO adsorption occurs through the dxz/dyz–ppp* interaction for the linear NO binding or through the dz2/dyz–ppp* interaction for the bent binding. The intermolecular spin coupling in the NO adsorption is generally
This journal is © the Owner Societies 2017
View Article Online
PCCP
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
antiferromagnetic, except for the case of Sc, to avoid the energy penalty associated with the electron occupation in the antibonding states. Because the unpaired electron of NO occupies the associated bonding state, the binding energy of NO is calculated to be generally larger than of CO. For application purpose (e.g., medical treatment), it is desirable to develop TM–ligand complexes with proper NO binding strengths. This comparative theoretical study may provide new insights into designing NO-delivery materials.
Conflicts of interest There are no conflicts to declare.
Acknowledgements We thank Dr Jaeheung Cho for providing us with the atomic coordinates of [Co(Me3-TPADP)(CH3CN)]2+. This work was supported by the DGIST R&D Programs of the Ministry of Science, ICT and Future Planning (Grants No. 15-BD-0403 and No. 17-BT-02).
Notes and references 1 L. J. Ignarro, G. M. Buga, K. S. Wood and R. E. Byrns, Proc. Natl. Acad. Sci. U. S. A., 1987, 84, 9265. 2 A. Verma, D. J. Hirsch, C. E. Glatt, G. V. Ronnett and S. H. Snyder, Science, 1993, 259, 381–384. 3 N. Hardingham, J. Dachtler and K. Fox, Front. Cell. Neurosci., 2013, 7, 190. 4 A. P. Hunt and N. Lehnert, Acc. Chem. Res., 2015, 48, 2117–2125. 5 J. R. Steinert, T. Chernova and I. D. Forsythe, Neuroscientist, 2010, 16, 435–452. 6 I. Azarov, L. Wang, J. J. Rose, Q. Xu, X. N. Huang, A. Belanger, Y. Wang, L. Guo, C. Liu, K. B. Ucer, C. F. McTiernan, C. P. O’Donnell, S. Shiva, J. Tejero, D. B. KimShapiro and M. T. Gladwin, Sci. Transl. Med., 2016, 8, 368. 7 B. B. Wayland and L. W. Olson, J. Am. Chem. Soc., 1974, 96, 6037–6041. 8 W. R. Scheidt and M. E. Frisse, J. Am. Chem. Soc., 1975, 97, 17–21. 9 C. Rovira, K. Kunc, J. Hutter, P. Ballone and M. Parrinello, J. Phys. Chem. A, 1997, 101, 8914–8925. 10 C. Rovira, K. Kunc, J. Hutter, P. Ballone and M. Parrinello, Int. J. Quantum Chem., 1998, 69, 31–35.
This journal is © the Owner Societies 2017
Paper
11 O. Chen, S. Groh, A. Liechty and D. P. Ridge, J. Am. Chem. Soc., 1999, 121, 11910–11911. 12 K. M. Vogel, P. M. Kozlowski, M. Z. Zgierski and T. G. Spiro, J. Am. Chem. Soc., 1999, 121, 9915–9921. 13 W. R. Scheidt, H. F. Duval, T. J. Neal and M. K. Ellison, J. Am. Chem. Soc., 2000, 122, 4651–4659. 14 T. G. Spiro, M. Z. Zgierski and P. M. Kozolowski, Coord. Chem. Rev., 2001, 219–221, 923–936. 15 L. E. Laverman and P. C. Ford, Chem. Commun., 1999, 1843–1844. 16 M. D. Pluth and S. J. Lippard, Chem. Commun., 2012, 48, 11981–11983. 17 G. G. Martirosyan, A. S. Azizyan, T. S. Kurtikyan and P. C. Ford, Chem. Commun., 2004, 1488–1489. 18 T. S. Kurtikyan, V. A. Hayrapetyan, G. G. Martirosyan, R. K. Ghazaryan, A. V. Iretskii, H. Zhao, K. Pierloot and P. C. Ford, Chem. Commun., 2012, 48, 12088. ´n, Inorg. Chem., 2015, 54, 5634–5645. 19 M. Rado 20 M. Jaworska and P. Lodowski, Struct. Chem., 2012, 23, 1333–1348. 21 J. Kim, G. Saravanakumar, H. W. Choi, D. Park and W. J. Kim, J. Mater. Chem. B, 2014, 2, 341. 22 H.-J. Xiang, L. An, W.-W. Tang, S.-P. Yang and J.-G. Liu, Chem. Commun., 2015, 51, 2555. 23 M. T. Pelegrino, R. B. Weller, X. Chen, J. S. Bernardes and A. B. Seabra, MedChemComm, 2017, 8, 713. 24 Y. Tahara, T. Yoshikawa, H. Sato, Y. Mori, M. H. Zahangir, A. Kishimura, T. Mori and Y. Katayama, MedChemComm, 2017, 8, 415. 25 B. Shin, K. D. Sutherlin, T. Ohta, T. Ogura, E. I. Solomon and J. Cho, Inorg. Chem., 2016, 55, 12391–12399. ¨ller, Phys. Rev. B: Condens. Matter 26 G. Kresse and J. Furthmu Mater. Phys., 1996, 54, 11169–11186. ¨chl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 27 P. E. Blo 50, 17953–17979. 28 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775. 29 K. Burke, M. Ernzerhof and J. P. Perdew, Chem. Phys. Lett., 1997, 265, 115–120. 30 C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158–6170. 31 J. Paier, R. Hirschl, M. Marsman and G. Kresse, J. Chem. Phys., 2005, 122, 234102. 32 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652. 33 M. Reiher, O. Salomon and B. A. Hess, Theor. Chem. Acc., 2001, 107, 48–55. 34 O. Salomon, M. Reiher and B. A. Hess, J. Chem. Phys., 2002, 117, 4729–4737.
Phys. Chem. Chem. Phys.
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PAPER
Cite this: DOI: 10.1039/c7cp06103j
View Journal
A unified understanding of the direct coordination of NO to first-transition-row metal centers in metal–ligand complexes Hyunjoo Lee and Joongoo Kang
*
The binding of nitric oxide (NO) to heme-proteins is an important biochemical process involved in a variety of physiological functions. Here, using hybrid density-functional calculations, we systematically investigate the adsorption of NO to first-transition-row metal centers in metal–ligand complexes. Through the comparative study for different transition metal (TM) centers, we provide a unified understanding of the microscopic interactions of NO with the TM centers and related chemical trends. We found that as the atomic number of the TM center increases, the binding strength of NO is largely reduced from 207 kJ mol
1
to near zero due to the low d-orbital energies for late TM centers. The
Received 7th September 2017, Accepted 3rd October 2017
intermolecular spin coupling between the localized spins at the TM center and the NO molecule is
DOI: 10.1039/c7cp06103j
way to avoid the energy penalty associated with the electron occupation in the antibonding states of the
generally antiferromagnetic, except for the case of Sc. The spin–spin coupling is determined in such a NO-bound complex. The adsorption strength of NO is generally larger than of CO because the unpaired
rsc.li/pccp
electron of NO occupies the associated bonding state.
Introduction Nitric oxide (NO) and carbon monoxide (CO) are important signaling molecules involved in a variety of physiological processes.1,2 For example, NO plays an important role in synaptic plasticity,3,4 and abnormal NO signaling can contribute to various neurodegenerative pathologies such as strokes, Alzheimer’s disease, multiple sclerosis, and Parkinson’s disease.5 Many physiological functions of NO involve interactions with heme-proteins. One way in which NO interacts with hemeproteins is through the direct coordination of NO with an Fe center to form an Fe–nitrosyl complex. Despite important biological functions mediated by NO and CO, these molecules are also associated as a toxic gas because of their strong binding to hemoglobin and the resulting poisoning. Recently, it was demonstrated that heme-based scavenger molecules, such as a mutant five-coordinate neuroglobin,6 can be used to effectively eliminate CO from CO-saturated hemoglobin. The binding of NO to heme-proteins is a crucial step in NO-mediated physiological processes. Many studies have been done on the binding of NO to a Fe-porphyrin because of its biochemical importance.7–15 Recently, the adsorption study of NO was extended to include non-heme Fe(III) complexes16 and
Department of Emerging Materials Science, DGIST, Daegu 711-873, Korea. E-mail: [email protected]
This journal is © the Owner Societies 2017
other metalloporphyrins such as Mn- and Co-porphyrins.17–20 It remains an important issue to understand the adsorption of NO to first-row transition metal centers for two reasons. First, a comparative study of NO binding for different transition metal (TM) centers in a TM–ligand complex can provide a unified understanding of the microscopic interactions of NO with the TM centers and related chemical trends. Second, such a comparative study can also provide useful insights into designing a platform for reversible NO binding. For application purpose, it is desirable to develop TM–ligand complexes with proper NO binding strengths. NO is endogenously produced by the human body as a signaling molecule. As an alternative way of providing NO to a target in the body, one may consider NO delivery from outside for medical treatment.21–24 Here, we focus on the direct coordination of NO to a TM center in a TM–ligand complex. In this study, two TM–nitrosyl complexes are considered for TM = Sc, Ti, V, Cr, Mn, Fe, Co, and Ni (Fig. 1): (i) TM(P)(NO), where P = porphyrin, and (ii) [TM(Me3-TPADP)(CH3CN)(NO)]2+, where Me3-TPADP = 3,6,9-trimethyl-3,6,9-triaza-1(2,6)-pyridinacyclodecaphane.25 For comparison, CO adsorption is also considered for the same TM–ligand complexes. The rest of the paper is organized as follows. First, we construct a theoretical framework for understanding the mechanisms and overall chemical trends of the NO and CO binding to the TM–ligand complexes. Next, we present our density functional theory (DFT) calculation results for the NO adsorption and discuss how the main results can be understood
Phys. Chem. Chem. Phys.
View Article Online
Paper
PCCP
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Fig. 2 Two types of the p–d orbital interaction responsible for the binding of NO (or CO) to the TM center of a TM-porphyrin: (a) a p-bonding state between a TM dxz and an NO pxpxp* orbital forming a linear TM–N–O bond and (b) a symmetry-allowed bonding state between a TM dz2 and NO ppp* orbital in a bent binding geometry. Fig. 1 Atomic structures of two TM–nitrosyl complexes: (a) TM(P)(NO), where P = porphyrin, and (b) [TM(Me3-TPADP)(CH3CN)(NO)]2+, where Me3-TPADP = 3,6,9-trimethyl-3,6,9-triaza-1(2,6)-pyridinacyclodecaphane. C, gray; H, white; N, blue; O, red; and TM, orange.
within our theoretical framework. Finally, we compare the binding strengths of NO and CO to TM-porphyrins.
Calculation methods To investigate the binding mechanisms of NO and CO to the TM–ligand complexes in Fig. 1, we performed spin-polarized density functional calculations using the Vienna ab initio Simulation Package.26 Our calculations employed the projector augmented wave (PAW) method27,28 with an energy cutoff of 400 eV for the plane wave part of the valence wave functions. For 3d TM with the atomic number N, we used N 18 valence electrons. The core electrons were treated as frozen cores within the PAW method. For the charged molecules such as [TM(Me3TPADP)(CH3CN)(NO)]2+, a uniform background charge was added to maintain the charge neutrality of the system in our plane-wave basis calculations. The energy convergence criterion for the electronic self-consistent calculations was set to 0.01 meV. Atomic structures of TM–ligands complexes were relaxed within 0.05 eV Å 1. Throughout the paper, the results were obtained by using the PBE0 hybrid density functional29–31 except if mentioned otherwise. For the NO adsorption to TM-porphyrins, we also present the results of other hybrid functionals such as B3LYP (ref. 32) and B3LYP* (ref. 33 and 34).
Results and discussion A.
Molecular-orbital theory of NO and CO adsorption
In this section, we develop a molecular-orbital theory for understanding adsorption mechanisms and overall chemical trends in the NO adsorption to the first-row TM centers in TM-porphyrin. The molecular orbitals of NO responsible for its adsorption are the p-antibonding states (ppp*) of the p orbitals of the N and O atoms (Fig. 2).12,14 There are two NO ppp* states having different orientations for the p orbitals. NO interacts with TM-porphyrin through the direct coordination of the N atom with the TM center, because the unpaired electron in the NO ppp* state is distributed more around the N atom than around the O atom. NO has a magnetic moment of 1 mB because of the unpaired electron in the ppp* state. Depending on the
Phys. Chem. Chem. Phys.
spin direction of NO with respect to the net spin of the TM center, the NO adsorption is characterized by either ferromagnetic or antiferromagnetic spin–spin coupling. As we will discuss later, the binding angle of NO determines which TM d-orbitals are able to interact with the NO ppp* states. The energy difference between the pair of the molecular orbitals, in turn, determines the degree of their hybridization and thus the binding strength of NO. By considering the interaction between the TM d-orbital states and the NO ppp* states, we develop a theoretical framework to understand the NO adsorption to TM-porphyrins, with the following expected properties: (i) the adsorption symmetry of NO dictates which TM d-orbitals are symmetry-allowed to interact with the NO ppp* states. (ii) The adsorption strength generally gets reduced as the atomic number of the TM center increases. (iii) The intermolecular spin coupling in the NO adsorption is generally antiferromagnetic except for the case of TM = Sc. (iv) For a TM-porphyrin, the adsorption strength of NO is expected to be larger than of CO. Each of these is discussed below. (i) The adsorption symmetry associated with the binding angle of NO dictates which TM d-orbitals are involved in the NO adsorption. Throughout the paper, the x-axis in Fig. 1 is chosen as the direction along which the NO molecule is bent, and the z-axis is chosen as a normal vector to the TM-porphyrin plane. In general, the TM–N bond is also slightly tilted with respect to the z-axis for greater interorbital coupling.13 The tilt angle of the TM–N bond is much smaller than the binding angle y. Therefore, for simplicity, the tilt angle is not shown in Fig. 1 and 2. For a linear binding of NO with y = 1801, the TM dxz couples to the NO pxpxp* state,12 forming a p-like bonding state and a corresponding antibonding state (see Fig. 2a for the bonding state). Likewise, the TM dyz couples to the NO pypyp* state to form a metal–nitrosyl complex. In contrast, for the TM dz2, the interorbital coupling is symmetry forbidden because the dz2 and pxpxp* (or pypyp*) states have different reflection parities with respect to the yz (or xz) plane (Fig. 2b). The reflection symmetry should thus be broken to enable the orbital coupling for dz2, as in the bent binding in Fig. 2b. Note that when NO is bent along the x-axis, the reflection symmetry with respect to the xz plane is preserved, making the p-like orbital coupling of the dyz state and the NO pypyp* state still possible. Thus, both TM dz2 and dyz can participate in the NO adsorption for the bent binding, while TM dxz and dyz are involved for the linear NO binding. Hereafter, the two different types of the molecular interactions will be referred to as the dz2/dyz–ppp* interaction and the dxz/dyz–ppp* interaction, respectively.
This journal is © the Owner Societies 2017
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PCCP
From the symmetry argument, it is thus expected that when the dxz/dyz–ppp* interaction is dominant over the dz2/dyz–ppp* interaction, the NO adsorption energetically favors a linear binding. Conversely, in the case that the dz2/dyz–ppp* interaction is strong, the NO adsorption takes a bent structure. The previous notion12 is that the bent TM–NO binding is associated with TM(III)–NO , while the linear binding is associated with TM(I)– NO+. However, this formal assignment is misleading from a computational point of view: in both binding structures, we will see that the NO molecule bound to the TM center is largely spinpolarized by its unpaired electron, indicating that it remains close to charge neutral. (ii) The adsorption strength of NO generally gets reduced as the atomic number of the TM site increases. The d-orbital energy of the first-row 3d TM center in a TM-porphyrin decreases as the atomic number increases. It turns out that the d-orbitals of early TMs (e.g., Sc) lie close in energy to the NO ppp* state, while for late TMs (e.g., Fe) the d-orbital energies are much lower than those of the NO ppp* states. For late TMs, the interactions between the TM d-orbitals and the NO ppp* states become relatively weak because of the large separation in their orbital energies. Therefore, the NO adsorption strength would be weak for late TMs, as compared to the case of early TMs. (iii) The intermolecular spin coupling is generally antiferromagnetic except for the case of Sc. For a moment, we only consider the dxz/dyz–ppp* interaction to discuss the spin–spin interaction for a linear adsorption of NO. (The same argument holds for the dz2/dyz–ppp* interaction in a bent adsorption.) With an exception for TM = Sc, the TM dxz/dyz orbitals in the majority (i.e., spin-up) spin state are fully occupied by two electrons of the TM cation. When it comes to the spin state of the NO adsorbate, the NO ppp* states can be occupied either by a spin-up electron or by a spin-down electron. In the former case, the dxz/dyz–ppp* interaction will lead to the two fullyoccupied bonding states in the spin-up state and the two antibonding states occupied by a single spin-up electron. In the latter case, however, the intermolecular interaction for the spin-up state gives rise to the two fully-occupied bonding states and the two empty antibonding states. The spin-down electron of the NO ppp* now can fill the bonding state in the spin-down state, further lowering the energy of the NO-bound complex. Consequently, the antiferromagnetic spin–spin coupling is energetically more favored than the ferromagnetic coupling. In other words, the spin–spin coupling is determined in such a way to avoid the energy penalty associated with the electron occupation in the antibonding states. The ferromagnetic spin–spin coupling in the NO adsorption is possible only for TM = Sc because the dxz/dyz orbitals are now occupied by only a single electron. The dxz/dyz–ppp* interaction in the ferromagnetic spin configuration then leads to the fullyoccupied bonding states and the empty antibonding states in the spin-up state. (iv) For a TM-porphyrin, the adsorption strength of NO is generally larger than of CO. For the adsorption of CO, the empty ppp* states of CO interact with the occupied TM d-orbitals to form the TM–CO bond. Because of the same
This journal is © the Owner Societies 2017
Paper
orbital symmetry (i.e., ppp*) for NO and CO, the types of orbital coupling in the CO adsorption to a TM-porphyrin is also determined by the adsorption symmetry associated with the binding angle of CO. As we will discuss later, the unpaired electron in the NO ppp* state occupies a bonding state of the NO-bound complex, enhancing the binding energy of NO. For the CO adsorption, however, the electron occupation of the bonding states is likely to be reduced by one, because CO has one electron less than NO. Therefore, the adsorption strength of CO is expected to be smaller than of NO. B.
Comparison with the DFT results
In this section, we present our calculation results of the NO and CO adsorption to TM–ligand complexes and discuss how our main results for different TM centers could be understood within the theoretical framework developed in the previous section. In PBE0 calculations, we found that the binding energy of NO to a TM-porphyrin tends to decrease from 207.0 kJ mol 1 for Sc to 0.2 kJ mol 1 for Ni as the atomic number of the 3d TM increases (Table 1). Here, we do not include the results for Cu and Zn, because NO does not directly bind to these TM centers. Fig. 3 shows that the same chemical trend is obtained regardless of the choice of the hybrid functional (see also Table 2). The binding energy of NO to Feporphyrin is calculated to be 80.0 kJ mol 1 (PBE0), 75.0 (B3LYP), and 104.2 (B3LYP*). The B3LYP* value agrees well with the binding energy of 108 kJ mol 1 to Fe-tetraphenylporphyrin (FeTPP+), which was estimated from the radiative association rates.11 As pointed out in the previous section, the interaction between the TM d-orbital and the NO ppp* state becomes relatively weak for late TMs because of the large energy separation in the orbital energies. Therefore, the NO adsorption strength is generally weak for late TMs, relative to the case of early TMs. Fig. 4 compares the NO binding energies to TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ systems in Fig. 1. We found Table 1 Binding energies (EB) for NO and CO molecules to the first-row 3d TM centers in a TM-porphyrin. Selected structural parameters are listed for TM(P)(XO), where X = N or C. The hybrid functional calculations were done with PBE0. The spin–spin coupling in the table denotes the type of the intermolecular spin interaction between the two localized spins at the TM center and the NO molecule, respectively
TM
EB (kJ mol 1)
dTM–XO (Å)
dX–O (Å)
Spin–spin coupling
NO
Sc Ti V Cr Mn Fe Co Ni
207.0 195.1 149.2 60.3 58.8 80.0 37.7 0.2
2.02 1.80 1.77 1.71 1.88 1.79 1.91 1.98
1.20 1.20 1.19 1.19 1.19 1.17 1.16 1.15
FM Mixed Mixed AFM AFM AFM AFM AFM
CO
Sc Ti V Cr Mn Fe Co Ni
68.7 109.5 99.3 2.6 11.7 26.8 24.1 4.7
2.30 2.04 2.00 2.58 2.45 1.73 1.98 2.00
1.14 1.14 1.14 1.13 1.13 1.15 1.13 1.13
— — — — — — — —
Phys. Chem. Chem. Phys.
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Paper
PCCP
Fig. 3 Comparison of the NO binding energies (EB) to TM-porphyrin in different hybrid functionals: PBE0, B3LYP, and B3LYP*.
Fig. 4 Chemical trends in the binding energies (EB) of NO to TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ in PBE0. For comparison, the binding energies of H2O to TM-porphyrin are also presented.
Table 2 Binding energies (EB) and structural parameters in B3LYP and B3LYP* for the TM(P)(NO) with the first-row 3d TM centers as the binding site. The intermolecular spin coupling between the TM center and the NO molecule is the same as in PBE0
TM
EB (kJ mol 1)
dTM–NO (Å)
dN–O (Å)
B3LYP
Sc Ti V Cr Mn Fe Co Ni
212.5 194.3 148.1 58.3 58.0 75.0 36.7 17.2
2.04 1.81 1.78 1.71 1.88 1.80 1.90 1.96
1.21 1.21 1.19 1.19 1.20 1.18 1.17 1.16
B3LYP*
Sc Ti V Cr Mn Fe Co Ni
216.7 214.3 170.9 100.4 103.0 104.2 62.2 12.5
2.04 1.82 1.68 1.68 1.84 1.76 1.84 1.88
1.21 1.21 1.20 1.19 1.20 1.18 1.18 1.16
that the overall chemical trend across the first-row TMs is the same for both TM–ligand complexes. For comparison, we also present the binding energies of H2O to TM-porphyrin. We found that especially for early TMs, the NO binding strength is significantly stronger than that of H2O. In Table 1, we summarize our DFT results for the intermolecular spin interaction for the NO adsorption. For TM = Sc, the spin–spin coupling is ferromagnetic, while for Cr and later TMs, it is antiferromagnetic. For Ti and V in between, the spin– spin interaction has a mixed character, which will be discussed later. To demonstrate how the spin–spin coupling depends on the TM center in a TM-porphyrin, we take four TM centers, TM = Sc, V, Cr, and Co, as examples (Fig. 5). For the Sc center, the intermolecular spin interaction is ferromagnetic. Fig. 5a shows the projected density of states (PDOS) of the NO-bound Sc-porphyrin. The NO adsorption to the Sc center involves the dxz/dyz–ppp* interaction. Therefore, in the PDOS, the molecular orbitals of the NO-bound complex are projected to the Sc dxz/dyz orbitals and the px/py orbitals of the N and O atoms in NO. When NO is far apart from the
Phys. Chem. Chem. Phys.
Fig. 5 Three different types of the intermolecular spin coupling between the localized magnetic moments at the TM center and the NO adsorbate: (a) ferromagnetic spin–spin interaction for TM = Sc, (b) mixed spin–spin interaction for V, and antiferromagnetic spin interaction for (c) Cr and (d) Co. Projected densities of states (PDOS) are shown for the selected p orbitals of NO and TM d orbitals involved in the NO adsorption. Two vertical dashed lines denote the energies of the HOMO and LUMO states. The p–d bonding (or antibonding) states are indicated by black (or green) arrows. The energy is referenced to the average orbital energy of the NO ppp* states. In the spin-density plots on the right panels, the yellow surface is for the majority spin-up charge density, while the blue surface is for the minority spin-down charge density.
This journal is © the Owner Societies 2017
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
PCCP
binding site, the Sc dxz/dyz orbitals are found to be slightly higher than the NO ppp* states in the orbital energies. The Sc dxz/dyz orbitals are occupied by an electron because the Sc center is in a 2+ charge state. When NO binds to the Sc center, the resulting p-like bonding states in the spin-up state are fully occupied, while the p-like antibonding states are empty. For the spin-down state, however, both the bonding and antibonding states are empty. The intermolecular spin interaction is ferromagnetic, as confirmed by the spin-density plot in the right panel of Fig. 5a. Note that the spin-up electrons (yellow) are distributed on both the Sc center and the NO molecule. The total magnetic moment of the complex is thus m = 2 mB, which can be formally assigned to the local spin of 1 mB at the Sc center and that of 1 mB at NO. In contrast, the NO adsorption to the Cr center and later TMs is characterized by the antiferromagnetic spin–spin interaction (Table 1). For Cr, the NO adsorption occurs through the dxz/dyz–ppp* orbital interaction. The NO-free Cr-porphyrin is in a high-spin state with m = 4 mB. When NO is adsorbed to the Cr center, however, the spin state of the Cr site changes to an intermediate-spin state with m = 2 mB. Regardless of the spin states, the spin-up dxz/dyz orbitals are fully occupied by two spin-up electrons. However, the spin-down dxz/dyz orbitals are occupied by a single electron only if the Cr center is in the intermediate-spin state. The relatively small binding energy for Cr can be in part attributed to the spin crossover induced by the NO adsorption (Fig. 3). As discussed in the previous section, the fully occupied TM dxz/dyz orbitals in the spin-up state require the antiferromagnetic spin–spin coupling. Indeed, the spindensity plot in Fig. 5c clearly shows the spin-up density around the Cr center and the spin-down density around the NO molecule. As a result, the net magnetic moment of the NO-bound complex is m = 1 mB. The PDOS in Fig. 5c shows that for both spin-up and -down states, the p-like dxz/dyz–ppp* bonding states are fully occupied, while the corresponding antibonding states are empty. Note that the NO p-orbital component in the p-like bonding states is substantially larger for the spin-down state, leading to the spindown density around the NO molecule in the spin-density plot. For the bent NO binding, the spin–spin coupling is also found antiferromagnetic. Fig. 5d shows the results for the bent NO adsorption to the Co center through the dz2/dyz–ppp* interaction. For an NO-free Co-porphyrin, the Co dz2 and dyz orbitals in the spin-up state are both fully occupied, which is a necessary condition for the antiferromagnetic spin–spin interaction. Indeed, the spin-density plot in Fig. 5d shows the antiferromagnetic spin–spin coupling. We found that the NO-free Co-porphyrin is in a low-spin state with m = 1 mB. The net magnetic moment of the NO-bound complex is then reduced to m = 0 mB due to the antiferromagnetic coupling with NO. The bent singlet ground state agrees with the previous DFT result.19 For the Ti and V centers, the spin–spin coupling exhibits a mixed character: the spin-density plot for the V center in Fig. 5b shows that the spin-up density is distributed around the TM center, while around the NO molecule, both the spin-up (yellow) and spin-down (blue) components are found with the orbital characters of pypyp* and pxpxp*, respectively. For both Ti
This journal is © the Owner Societies 2017
Paper
and V, the NO binding is linear with y = 1801. When compared to the case of Sc, the number of the spin-up electrons in the TM dxz/dyz orbitals increases from one to two, while the number of the spin-down electrons in the dxz/dyz orbitals remains zero. Thus, according to the argument in the previous section, the unpaired electron in the NO ppp* orbitals should be in the spin-down state to couple with the empty TM d-orbitals in the spin-down state. For the spin-up state, the p-like bonding states, dxz–pxpxp* and dyz–pypyp*, at around 3 eV are occupied by two electrons (Fig. 5b). For the spin-down state, the dxz–pxpxp* bonding state, which lies at around 2.5 eV, is occupied by the spin-down electron in the NO pxpxp* orbital. In contrast, the dyz–pypyp* bonding state in the spin-down state is unoccupied and lies at around 0 eV. Note that when compared to the case of Cr with the fully-occupied bonding states in both spin states (Fig. 5c), a single electron is lacking in the dyz–pypyp* bonding states in the spin-down state. Therefore, the spin-density distribution for Ti and V has the mixed nature with both the spin-up and spin-down components at NO (Fig. 5b). Fig. 6 shows how the adsorption energy changes with the binding angle y of NO for several TM-porphyrins. As we discussed from the symmetry argument in the previous section, the NO adsorption energetically favors a linear binding when
Fig. 6 Dependence of the adsorption energy (D E) of NO on the TM–N–O angle y in TM(P)(NO): (a) TM = Co and Cr and (b) TM = Mn and Fe in different spin states. In the right panels, the charge densities of representative bonding states of TM(P)(NO) are plotted for the specified TM centers and the binding angles.
Phys. Chem. Chem. Phys.
View Article Online
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
Paper
the dxz/dyz–ppp* interaction is dominant over the dz2/dyz–ppp* interaction. Conversely, in the case that the dz2/dyz–ppp* interaction is strong, the NO adsorption takes a bent structure. For example, the NO binding to the Cr center is linear with the energy minimum at y = 1801, indicating that the p-like dxz/dyz–ppp* orbital interaction is responsible for the NO adsorption (Fig. 6a). In the right panel of the figure, the charge-density plot of one of the p-like bonding states is shown to demonstrate the type of the intermolecular orbital coupling. For the bent NO adsorption to Co, however, the minimum energy is obtained at y = 1221. The NO adsorption to the Co center involves the s-like dz2–ppp*, as well as the p-like dyz–pypyp* interaction. For some TM centers such as Mn and Fe, both types of the orbital interactions can compete for the NO adsorption in certain spin states. As the low-energy states of NO-bound Mn-porphyrin, we obtained the following two spin states (Fig. 6b in B3LYP*): the intermediate-spin (IS) triplet state with m = 2 mB and the low-spin (LS) singlet state with m = 0 mB. For the IS state, we obtained a relatively flat energy curve for the angles from y = 1301 to 1801, with the minimum at y = 1421, because of the competition between the two types of the intermolecular interaction. For the LS state, however, the p-like dxz/dyz–ppp* orbital coupling becomes dominant over the s-like dz2–ppp* coupling. As a result, the LS state favors the linear NO adsorption with a large bending stiffness. The same spin state–structure relationship (i.e., the bent triplet and the linear singlet) is obtained for different hybrid functionals. As the previous work pointed out,18,33 however, the energy difference between these two spin states are strongly functional dependent. In a previous work on Fe(II)–S complexes,33 it was shown that the energy difference between the high- and lowspin states linearly depends on the coefficient of exact exchange, because a high-spin state is generally favored in Hartree–Fock-type theories. Indeed, we found that the relative stability of the singlet state is systematically enhanced as the coefficient of the exact exchange in our hybrid functional calculations decreases from 25% (PBE0) to 20% (B3LYP) and to 15% (B3LYP*). In our B3LYP* results, the triplet state is still more stable than the singlet state, although the energy difference is small (Fig. 6b). However, it should be noted that a previous experiment18 at low temperature, combined with DFT calculations, suggests that the singlet state is the true ground state of NO-bound Mn-porphyrin.19 For Fe-porphyrin, we similarly found that there exist the competing orbital interactions for the NO adsorption, which lead to the relatively flat energy curves in Fig. 6b. For the IS ground state with the net spin of m = 1 mB, the NO-bound complex is bent with the angle of 1501, in good agreement with the previous experimental result (y = 149.21).8 Note that the binding angle y for the Fe center is substantially larger than for the Co-porphyrin (y = 1221) because of the competition between the s-like dz2–ppp* and p-like dxz–ppp* orbital couplings for the Fe center. As the second-lowest energy state in B3LYP*, we also identified the linear high-spin (HS) state with the net spin of m = 3 mB for the NO-bound complex. The adsorption energy
Phys. Chem. Chem. Phys.
PCCP
Fig. 7 Comparison of the binding energies (EB) of NO and CO to TM-porphyrin for different first-row TM centers in PBE0 calculations.
curve is nearly flat around y = 1801, relative to the energy curve for the linear LS Mn-porphyrin. Finally, we compare the binding strengths of NO and CO to TM-porphyrins in Fig. 7. The adsorption energy of CO is substantially smaller than of NO, as expected in the previous section. The crucial difference between the two molecules is that CO has one electron less than NO. For early TM centers such as Sc, Ti, and V, one electron is thus less available for the bonding states of the CO adsorption, leading to the relatively weak binding strengths for CO. For the Cr center, the situation is yet more complicated: the NO adsorption induces the spin crossover of the Cr center from the HS state to the IS state to maximize the number of electrons in the associated bonding states, as discussed earlier in this paper. The CO adsorption, however, does not induce the spin crossover, because the energy gain for the CO adsorption with the Cr center in the IS state is less than the energy cost associated with the spin crossover. Then, the Cr center remains in the HS state after the CO adsorption. The number of electrons in the bonding states is only two for the CO adsorption, while it is four for the NO adsorption. Therefore, the reduced number of electrons in the bonding states explains the reduced CO binding strength for TM = Sc, Ti, V, and Cr. It is also expected that the CO ppp* orbitals would lie at higher energy than the NO ppp* orbitals due to the less electronegative C atom in CO, leading to the larger energy separation between the TM d-orbitals and the CO ppp* states for late TMs and thus less strong binding.
Conclusions We performed hybrid density-functional calculations to investigate the direct coordination of NO to first-row TM centers in TM-porphyrin and [TM(Me3-TPADP)(CH3CN)]2+ systems. Through the comparative study for different 3d TM centers, we found that as the atomic number of the TM center increases, the adsorption strength of NO generally gets reduced due to the low d-orbital energies for late TM centers. The NO adsorption occurs through the dxz/dyz–ppp* interaction for the linear NO binding or through the dz2/dyz–ppp* interaction for the bent binding. The intermolecular spin coupling in the NO adsorption is generally
This journal is © the Owner Societies 2017
View Article Online
PCCP
Published on 04 October 2017. Downloaded by Gazi Universitesi on 11/10/2017 09:47:28.
antiferromagnetic, except for the case of Sc, to avoid the energy penalty associated with the electron occupation in the antibonding states. Because the unpaired electron of NO occupies the associated bonding state, the binding energy of NO is calculated to be generally larger than of CO. For application purpose (e.g., medical treatment), it is desirable to develop TM–ligand complexes with proper NO binding strengths. This comparative theoretical study may provide new insights into designing NO-delivery materials.
Conflicts of interest There are no conflicts to declare.
Acknowledgements We thank Dr Jaeheung Cho for providing us with the atomic coordinates of [Co(Me3-TPADP)(CH3CN)]2+. This work was supported by the DGIST R&D Programs of the Ministry of Science, ICT and Future Planning (Grants No. 15-BD-0403 and No. 17-BT-02).
Notes and references 1 L. J. Ignarro, G. M. Buga, K. S. Wood and R. E. Byrns, Proc. Natl. Acad. Sci. U. S. A., 1987, 84, 9265. 2 A. Verma, D. J. Hirsch, C. E. Glatt, G. V. Ronnett and S. H. Snyder, Science, 1993, 259, 381–384. 3 N. Hardingham, J. Dachtler and K. Fox, Front. Cell. Neurosci., 2013, 7, 190. 4 A. P. Hunt and N. Lehnert, Acc. Chem. Res., 2015, 48, 2117–2125. 5 J. R. Steinert, T. Chernova and I. D. Forsythe, Neuroscientist, 2010, 16, 435–452. 6 I. Azarov, L. Wang, J. J. Rose, Q. Xu, X. N. Huang, A. Belanger, Y. Wang, L. Guo, C. Liu, K. B. Ucer, C. F. McTiernan, C. P. O’Donnell, S. Shiva, J. Tejero, D. B. KimShapiro and M. T. Gladwin, Sci. Transl. Med., 2016, 8, 368. 7 B. B. Wayland and L. W. Olson, J. Am. Chem. Soc., 1974, 96, 6037–6041. 8 W. R. Scheidt and M. E. Frisse, J. Am. Chem. Soc., 1975, 97, 17–21. 9 C. Rovira, K. Kunc, J. Hutter, P. Ballone and M. Parrinello, J. Phys. Chem. A, 1997, 101, 8914–8925. 10 C. Rovira, K. Kunc, J. Hutter, P. Ballone and M. Parrinello, Int. J. Quantum Chem., 1998, 69, 31–35.
This journal is © the Owner Societies 2017
Paper
11 O. Chen, S. Groh, A. Liechty and D. P. Ridge, J. Am. Chem. Soc., 1999, 121, 11910–11911. 12 K. M. Vogel, P. M. Kozlowski, M. Z. Zgierski and T. G. Spiro, J. Am. Chem. Soc., 1999, 121, 9915–9921. 13 W. R. Scheidt, H. F. Duval, T. J. Neal and M. K. Ellison, J. Am. Chem. Soc., 2000, 122, 4651–4659. 14 T. G. Spiro, M. Z. Zgierski and P. M. Kozolowski, Coord. Chem. Rev., 2001, 219–221, 923–936. 15 L. E. Laverman and P. C. Ford, Chem. Commun., 1999, 1843–1844. 16 M. D. Pluth and S. J. Lippard, Chem. Commun., 2012, 48, 11981–11983. 17 G. G. Martirosyan, A. S. Azizyan, T. S. Kurtikyan and P. C. Ford, Chem. Commun., 2004, 1488–1489. 18 T. S. Kurtikyan, V. A. Hayrapetyan, G. G. Martirosyan, R. K. Ghazaryan, A. V. Iretskii, H. Zhao, K. Pierloot and P. C. Ford, Chem. Commun., 2012, 48, 12088. ´n, Inorg. Chem., 2015, 54, 5634–5645. 19 M. Rado 20 M. Jaworska and P. Lodowski, Struct. Chem., 2012, 23, 1333–1348. 21 J. Kim, G. Saravanakumar, H. W. Choi, D. Park and W. J. Kim, J. Mater. Chem. B, 2014, 2, 341. 22 H.-J. Xiang, L. An, W.-W. Tang, S.-P. Yang and J.-G. Liu, Chem. Commun., 2015, 51, 2555. 23 M. T. Pelegrino, R. B. Weller, X. Chen, J. S. Bernardes and A. B. Seabra, MedChemComm, 2017, 8, 713. 24 Y. Tahara, T. Yoshikawa, H. Sato, Y. Mori, M. H. Zahangir, A. Kishimura, T. Mori and Y. Katayama, MedChemComm, 2017, 8, 415. 25 B. Shin, K. D. Sutherlin, T. Ohta, T. Ogura, E. I. Solomon and J. Cho, Inorg. Chem., 2016, 55, 12391–12399. ¨ller, Phys. Rev. B: Condens. Matter 26 G. Kresse and J. Furthmu Mater. Phys., 1996, 54, 11169–11186. ¨chl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 27 P. E. Blo 50, 17953–17979. 28 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775. 29 K. Burke, M. Ernzerhof and J. P. Perdew, Chem. Phys. Lett., 1997, 265, 115–120. 30 C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158–6170. 31 J. Paier, R. Hirschl, M. Marsman and G. Kresse, J. Chem. Phys., 2005, 122, 234102. 32 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652. 33 M. Reiher, O. Salomon and B. A. Hess, Theor. Chem. Acc., 2001, 107, 48–55. 34 O. Salomon, M. Reiher and B. A. Hess, J. Chem. Phys., 2002, 117, 4729–4737.
Phys. Chem. Chem. Phys.
