Electronic, Magnetic, and Transport Properties of Fen-bis(n-acene

Using density functional theory (DFT), we have investigated structural distortion of Fe(benzene)2 and Fe2(naphthalene)2 considering their different sp...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/JPCC

Electronic, Magnetic, and Transport Properties of Fen‑bis(n‑acene) and Fen‑bis(n‑BNacene) [n = 1,2,∝]: A Theoretical Study Dibyajyoti Ghosh,† Prakash Parida,‡ and Swapan K Pati*,‡,§ †

Chemistry and Physics of Materials Unit, ‡Theoretical Sciences Unit, and §New Chemistry Unit, Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore, 560064, India ABSTRACT: Using density functional theory (DFT), we have investigated structural distortion of Fe(benzene)2 and Fe2(naphthalene)2 considering their different spin states. It turns out that the structure is less distorted and more stable in its high-spin state than the corresponding low-spin state. We also have studied the structural, electronic, and magnetic properties of infinitely extended system, that is, Fe∝bis(polyacene) and BN-analogue of it, Fe∝-bis(BNpolyacene). The band structure calculations show that the electronic structures of these two systems are quite dissimilar; that is, Fe∝-bis(polyacene) shows metallic behavior, whereas its BNanalogue shows robust half-metallic behavior. We have also studied the transport characteristic of finite size Fe3-bis(3BNacene) where it has been coupled to gold electrodes on either side. From the transport calculations, we predict that under finite bias, Fe3-bis(3-BNacene) shows efficient spin filter behavior, which possibly can have a huge application in spintronic devices.

■

INTRODUCTION As we are in the final stage of the silicon-device era,1,2 a huge attention has been given to develop new devices where charge as well as intrinsic spin of electrons can be used to store information.3 Actually, spin-based devices are more efficient than the charge-based devices in terms of energy, size, response time, and so on.3−6 In the initial stage of spintronic research, metals and inorganic semiconductors were mainly explored to have efficient spintronic devices.7−9 However, recently, organic materials came into the limelight for their certain advantages, like long spin-relaxation length and time, cost, portability, flexibility, and so on compared with other materials,3,10−14 and for such kind of organic spintronic devices, half-metals15 which usually have nearly 100% spin polarization in a particular orientation at Fermi surfaces, can be used as a perfect ingredient.16 The search for organic half-metallic system got an enormous success when some of the 1D organometallic sandwich molecular wires (SMWs) exhibited half-metallicity and spin filter effect.13,17−19 A number of SMWs having the same20 or different types21 of transition metals (TMs) that are sandwiched between the layers of organic molecules have been well-investigated theoretically.22 The development of laser vaporization technique brings a new possibility to synthesize finite-size clusters of these SMWs.23,24 Several properties of these SMWs have also been investigated extensively.25−28 These systematic studies illustrate the importance of the presence of aromatic ligands to determine the stable structure29 as well as nature of magnetic interaction between metal atoms.30,31 Also, some attentions have been given to replace all © 2012 American Chemical Society

carbon systems by their BN-analogue (B = boron and N = nitrogen).13,32 Experimental as well as theoretical findings illustrate that system properties can change abruptly by changing all carbon systems to their BN-analogues.33−35 In all of the previous studies, SMWs are periodic along the direction of sandwich, that is, along the direction [ligand− metal−ligand−metal]∝. As previously reported, in the case of [TM-cyclic aromatic ligand]∝, the size of the ligand can even be varied from a five membered ring (cyclopentadiene) to a three fused six member rings (anthracene)30,31. Among all of this, TM-benzene sandwiched complexes are one of the most extensively studied class of complexes.36,37 It has also been suggested theoretically how one can stabilize such an organometallic sandwich complex.38 However, a major problem with this type of organometallic sandwiched complexes is that late 3d TMs (Fe−Zn) cannot form proper stacked complexes because of their large number of valence electrons, which is much greater than 18,39 hindering the formation of complexes. A possible way to overcome this problem and to make a proper stacking of the sandwich complexes of these late TM is to decrease the number of valence electrons in the system, and that can be done by constructing a structure where the ligands can have less coordination (i.e., less than η6). To this regard, it is interesting to ask a question: can the aromatic ligand, benzene, be extended to polyacene to realize a properly stacked wire considering late 3d TM, which is periodic Received: May 11, 2012 Revised: July 23, 2012 Published: August 9, 2012 18487

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

the basis set 6-31+G(d) for carbon and hydrogen and LanL2DZ45−47 for iron. (See Figure 1.) We have carried out

along the direction of polyacene instead of the direction of sandwich? At this junction, we consider a Fe-based sandwiched complex consisting of two polyacene chain and an infinite chain of Fe atoms, in which Fe-chain remains sandwiched in between the two polyacene chains (see later). In this manner, we would be able to create a η4 coordination in between Fe and each polyacene chain where the complex contains 16 valence electrons (16-electron system), giving proper stacking. For further discussion, we have nomenclatured the infinite Fenbis(n-acene) and Fen-bis(n-BNacene) as Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene), respectively. We have explored the structural, electronic, magnetic and transport properties of Fe∝bis(polyacene) and its BN-analogue, Fe∝-bis(BNpolyacene). From band structure calculation, it can be found that Fe∝bis(polyacene) is metallic in nature, whereas its BN-analogue is a half-metal. The transport calculations show the efficient spinfilter property for Fe3-bis(3-BNacene).

■

COMPUTATIONAL DETAILS To investigate properties of molecules, we have used Gaussian 09 package.40 We have used spin-unrestricted density functional theory (DFT) as implemented in the SIESTA41 package for geometry optimization and electric structure calculations of infinite chain systems. We have used a double-ζ polarized (DZP) basis set for all atoms, and real mesh cutoff is chosen as 300 Ry. For exchange-correlation, we have used Perdew− Burke−Ernzerhof (PBE) functional within the generalized gradient approximation42(GGA). For 1D SMWs, a unit cell of 20 × 20 × c Ang,3 where c is the length of unit cell in periodic direction, has been used to avoid spurious interaction in nonperiodic directions. To find the magnetic ground state of the molecular wire, we have taken a supercell of two unit cells. For periodic calculation, we have used a Monkhorst-Pack 1 × 1 × 60 k-point grid (total number of k-points is 31) to sample the 1D Brillouin zone. For all of the geometry optimization, interatomic forces are relaxed up to 0.04 eV/Å. To validate the results obtained using localized basis, we have also done a few electronic structure calculations within the plane-wave basis set using PWscf package, as implemented in the QuantumESPRESSO.43 For transport property calculation, we have used nonequilibrium Green's function (NEGF) methodology extended for spin-polarized systems, as implemented in the TranSIESTA package.44 The transmission function is calculated using the formula T (E) = Tr[ΓL(E)G(E)ΓR (E)G†(E)]

Figure 1. Optimized structures of Fe(Bz)2 in their (a) singlet and (b) triplet spin states. Structures (c) and (d) represent the singlet and pentet spin states of Fe2(Napthalene)2. All structures shown here are calculated using wB97XD functionals.

the calculations using DFT as well as Hartree−Fock (HF) methods. We have used four different exchange-correlation functionals, OPBE,48 B3LYP,49−51 wB97X,52 and wB97XD53 for the DFT calculations. We have considered two possible spin multiplicities, that is, singlet and triplet. We have tabulated the stabilization energy (ΔE = Etriplet − Esinglet) of the triplet state for various functionals (Table 1). Table 1. Stabilization Energy (ΔE, ΔED6h) and Fe−Benzene Distance (d) for the Triplet State of Fe(Bz)2

∫μ

μR L

[f (E , μL ) − f (E , μR )]T (E) dE

HF

OPBE

B3LYP

wB97X

wB97XD

ΔE (eV) ΔED6h (eV)

−1.02 −5.41

−0.08 −1.74

−0.97 −2.00

−0.83 −2.08

−0.86 −2.11

2.06

1.72

1.83

1.78

1.77

d (Å)

The spin-splitting energy (ΔE) corresponds to the geometry optimization without considering any constraint, whereas ΔED6h corresponds to the geometry-optimization fixing the symmetry of the system to D6h. It is well known that the spin-ground state of iron complexes largely depends on the choice of functionals.54−58 Therefore, we have used several functionals for the calculations. Our calculations show that the HF method excessively favors the triplet state. Considering the HF method, the absence of electronic correlation between the unlike spins makes the highspin state much more stable than the low-spin state.54,59 Under DFT method, the pure functionals like LDA,60 PBE,42 BLYP,49,61 and so on overstabilize the low-spin state, whereas the hybrid functionals like B3LYP, PBEO,62 and so on predict high-spin state as ground state for iron complexes, in particular.54,55 Several studies pointed out that functionals based on Handy and Cohen’s optimized exchange functional,63 such as OPBE, correctly calculate the ground state of iron

(1)

where the G(E) is Green’s function and it has been calculated from the Hamiltonian and self-energies of the central region. Γα(E) is (−2 times) the imaginary part of the self-energies of the left and right electrodes (α = L, R). The current is calculated using the Landauer−Buttiker formula I = (G0 /e)

methods/functionals

(2)

where μL and μR are the chemical potential of left and right electrode, respectively, and G0 is the quantum of conductance and f(E) represents the Fermi−Dirac distributions at the two electrodes with electrochemical potentials μL and μR.

■

RESULTS AND DISCUSSION First, we have discussed finite-size cluster, Fe(Bz)2. Using Gaussian 09 package, we have optimized the geometry within 18488

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

complexes.54,55 Considering OPBE functionals, we also predict the triplet spin-state as the ground state even though the ΔE value is one order less than the other functionals. However, the calculated ΔE value (−83 meV) is even larger than the room temperature to stabilize the triplet state. We note that although Youn et al.64 reported that singlet and triplet states are almost energetically the same, the ground state is predicted to be in the triplet state using various functionls within our calculations. We also have used long-range corrected hybrid density functionals, wB97X, which can effectively capture the noncovalent interaction between benzene rings and iron atom. As the dispersion interaction between benzene molecules is very prominent for benzene-dimers, 65,66 we also have used dispersion-corrected functionals, wB97XD. For the functionals, wB97X and wB97XD, the triplet spin state appears to be the ground spin state. We have also calculated the distance between Fe and benzene ring for triplet spin state and tabulated those in Table 1. Except HF and B3LYP, other functionals give the distance value, which is in agreement with previous study.22 We note that two benzene rings tend to stay away from each other in its singlet state. As Fe(Bz)2 system is a 20-electron environment, one of the benzene rings needs to be distorted so that the electron donation from one ligand gets reduced and hence the total number of valence electrons approaches 18 electrons to form stacked complexes. We also find the same distorted structures of Fe(Bz)2 in its singlet state. (See Figure 1a.) For the triplet state, two benzene rings remain planar and stacked properly on each other. (See Figure 1b.) However, we note that C atoms of one benzene ring are not exactly on top of the C atoms of the other ring (i.e., not fully eclipsed); rather, they are slightly staggered with a small dihedral angle of 5.56°. As the staggering is very little, we still find that the Fe orbital splitting is same as that obtained due to the D6h ligand field. For the triplet state, two unpaired electrons get accommodated in the degenerate antibonding orbitals produced by dxz and dyz orbitals of Fe and π2 orbitals of benzene. (See Figure 2b.) Now,

causes large strain in the structure and hence it gets distorted severely to make the singlet state. (See Figures 1a and 2a.) To have a better picture about idea of this strain, we also optimize Fe(Bz)2 constraining both singlet and triplet state to D6h symmetry. We find that triplet state is highly energetically stable compared by singlet state under D6h symmetry. (See Table 1.) Comparing the values of ΔE and ΔED6h, we find that the singlet state is highly unstable in D6h symmetry. Thus, the benzene rings get distorted in the singlet state to form its lowenergetic structure during full-optimization. Next, we have focused on the magnetic and molecular structure of Fe2(napthalene)2. To determine the spin ground state, we have considered all possible spin multiplicities: singlet, triplet, and pentet. Optimizing the molecular geometry for each spin-state, the highest spin (HS) state (i.e., pentet) turns out to be the most stable, whereas the lowest spin (LS) state (i.e., singlet) appears to be the least stable. The relative stability of the spin states has been estimated by calculating the spin stabilization energy (ΔEs), which can be formulated as ΔEs = EHS − ELS, where EHS and ELS denote the energy of HS and LS states, respectively. For this complex, the HS state gets stabilized over the LS state by an energy ΔEs = −1.379 eV. To have an estimation of exchange coupling, we have modeled the spin−spin interactions between two spin-centers using the ⎯ → ⎯ → Heisenberg Hamiltonian, H = −J12S1· S2 , where J12 is the → ⎯ exchange coupling and Si is the spin at ith site. In our case, as each Fe (3d84s0) site contains two unpaired electrons, it comes out as S1 = S2 = 1. To calculate the coupling constant, we have considered the energy difference between triplet and pentet states. The interaction appears to be ferromagnetic (FM) in nature with the exchange coupling, J12= −0.325 eV. Looking at the optimized geometry in the low spin state, one of the naphthalene ring gets distorted like the benzene ring in Fe(Bz)2. (See Figure 1c.) However, the distortion of naphthalene ring is less compared with that of the benzene ring. (See Figure 1a.) This is because with the increase in ligand size from benzene to naphthalene the coordination approaches η4-type, whereas, in the high spin state, unlike Fe(Bz)2, two naphthalene rings in this complex stack properly without any staggering. (See Figure 1d.) Thus, we are motivated to extend naphthalene to polyacene for proper stacking of these kinds of Fe-based sandwich complexes. In Figure 3, we present optimized geometries of Fe∝bis(polyacene) and Fe∝-bis(BNpolyacene). For Fe∝-bis(BNpolyacene), we have considered two tautomeric eclipsed structures: eclipsed-1, the structure in which B atoms are on top of B atoms of two BNpolyacene, and eclipsed-2, where B or N atoms are on top of the counter atoms (i.e., N or B atoms, respectively; see Figure 2b). We find that eclipsed-2 is far more stable than eclipsed-1. This stabilization can be understood on the basis of favorable charge transfer between B and N atoms. In the rest of our discussions, we have considered the stable conformation, eclipsed-2. (See Figure 3.) Vertical distance between Fe chain and polyacene in Fe∝-bis(polyacene) is found to be 1.77 Å, whereas for Fe∝-bis(BNpolyacene), it is 1.88 Å. The Fe−Fe distances in these two are 2.51 and 2.56 Å, respectively. To examine the stability of these sandwiched complexes, we have calculated the binding energy (BE) of the corresponding unit cell using the following formulas

Figure 2. (a) HOMO plot of Fe(Bz)2 in its singlet spin state. (b) One of the SOMO plots of Fe(Bz)2 in its triplet spin state.

to generate the singlet state, the degeneracy of these antibonding states has to be lifted off. Therefore, one of the orbital stays in lower energy and forms HOMO (see Figure 2a), whereas the other one goes to higher energy and forms LUMO. However, within the D6h ligand field, the pairing of electrons (i.e., formation of singlet) will no longer be allowed, and hence the structure tends to adopt some other symmetry. Therefore, this pairing of electrons in one of the antibonding orbitals 18489

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

Figure 3. Optimized structure of infinite chain of (a) Fe∝-bis(polyacene) and (b) Fe∝-bis(BNpolyacene). For panel b, dark blue and light pink are nitrogen and boron atoms, respectively.

Figure 4. Spin-resolved band structure of Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene) and the wave functions of some selected bands. The plot is scaled for EF to lie at 0 eV.

ligands. Large positive magnetic moments of 1.924 μB and 2.224 μB are found on the Fe atom in Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene), respectively. Small negative moments of −0.264 μB and −0.224 μB are localized on polyacene and BNpolyacene chains, respectively. Because of the chargeseparated states in BNpolyacene chain, the magnetic moment that arises from the ligand is nonuniformly distributed over B and N atoms, whereas for the polyacene chain, the magnetic moment is homogeneously distributed over all edge carbon atoms. Like graphene nanoribbons,67 we find that C and B atoms at edges (which are hydrogen-passivated) mainly contribute to the magnetic moment of polyacene and BNpolyacene, respectively. To get a better picture of orbitals that contribute to the net magnetic moment of the unit cell in these structures, we present the spin-polarized band structure along with the respective wave function of the bands in Figure 4. Because of strong hybridization, 4s levels are pushed well above the Fermi level. As a result, the effective electronic configuration of Fe becomes 3d84s0. For both systems, the bands (bands 1 and 2) derived from dz2 and dx2−y2 orbitals (see the wave function plots in Figure 4) remain far below the Fermi energy and hence are completely filled up for majority as well as minority spin electrons. Hence, these two orbitals do not contribute to the net magnetic moment and in fact remain silent in both transport and magnetic behaviors. The bands (bands 3−5) derived from dxy, dyz, and dxz (see Figure 4) remain close to the Fermi level, and the dispersion of these bands mainly determines the transport and magnetic properties of these systems. For Fe∝-bis(polyacene), the top of the valence band (band 5) that has a major contribution from the 3dxz orbital just crosses the Fermi level in the majority spin channel,

E b([Fe(C4 H 2)2 ]∝ ) = E([Fe(C4 H 2)2 ]∝ )) − E(Fe)∝ − 2E(C4 H 2)∝ E b([Fe(B2N2H 2)2 ]∝ ) = E([Fe(B2N2H 2)2 ]∝ ) − E(Fe)∝ − E(B2N2H 2)2 )∝ (5)

The binding energies for unit cell of Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene) are found to be −2.99 and −0.97 eV, respectively. The less BE for Fe∝-bis(BNpolyacene) is due to the charge-separated state between boron and nitrogen atoms, where π-electrons get less delocalized. In fact, less availability of π electrons causes weak η4 interaction of Fe chain with BNpolyacene chain compared with polyacene chain.31 To investigate the type of magnetic interaction acting among Fe atoms, we have taken a supercell of 1 × 1 × 2 and have calculated the local magnetic moments. We find all the systems to be ferromagnetically stabilized and thus have a magnetically polarized ground state. To estimate the FM stability, we have calculated the energy difference (ΔE) between the FM and the antiferromagnetic (AFM) states using the formula, ΔE = EFM − EAFM. Note that the negative sign of ΔE indicates the stabilization of the FM state over the AFM state. These stabilization energies are ΔE = −527.1 meV for Fe∝bis(polyacene) and ΔE = −36.7 meV for Fe ∝ -bis(BNpolyacene) per corresponding unit cells. The large magnitude of ΔE designates the high stability of the FM state even at room temperature. From our calculation, we find that net magnetic moment of the unit cell is 1.66 μB for Fe∝bis(polyacene) and 2.00 μB for Fe∝-bis(BNpolyacene). To obtain a quantitative picture of magnetic moment distribution, we have calculated local magnetic moment on Fe atoms and the 18490

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

Figure 5. Spin-resolved DOS for Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene). Solid green portion represents the pDOS of Fe ion. Top panel shows the results obtains from localized basis (SIESTA), and bottom panel shows results from plane-wave basis (Quantum Espresso). Up-arrow and down-arrow show the results for majority spins and minority spins, respectively.

system metallic in nature, whereas for Fe∝-bis(BNpolyacene) a finite number of states at the Fermi level are available to take part in transport for minority spin channel, whereas a semiconducting gap of 1.5 eV opens up for the majority spin channel, making the system half-metallic. Because SIESTA uses the localized basis, in principle, one should verify the obtained electronic structures using the planewave basis set. For that, we have calculated the electronic density of states of both the systems using plane-wave basis and have plotted in Figure 5a′,b′. It can be seen clearly that both the basis sets show qualitatively similar behavior with the same conclusion about the overall electronic states of the systems. In particular, our main findings of half-metallicity retain true irrespective of the types of basis sets used for calculations. Although the positions of peaks and energy gap are not same with different basis set calculations, the overall electronic nature of the systems is robust against the details of the calculations. To explore the possibility of device applications, we have investigated the transport properties of these systems. Because the band structure of Fe∝-bis(BNpolyacene) shows halfmetallic behavior, we have chosen a finite fragment of Fe∝bis(BNpolyacene) for our transport study. We consider Fe3bis(3-BNacene) consisting of three Fe atoms coupled to gold (Au) electrodes on either side (i.e., electrode−molecule− electrode arrangement) for our transport calculation. We consider nonmagnetic 4 × 4 × 4 bulk gold electrodes (Au (111) plane) containing 48 gold atoms for the transport calculations. To have a strong coupling between molecular fragment and Au electrodes, we have used thiol (−SH) as an anchoring group. First, we have optimized the geometry of molecular fragment (with thiol groups), Fe3-bis(3-BNacene)(SH)4, and gold electrodes individually. Then, we have removed hydrogen atoms of −SH groups and placed the molecule between the two gold electrodes and optimized the

whereas the bands 3 and 4 that originate effectively from 3dyz and 3dxz cross the Fermi level in the minority spin channel, leading to the metallic behavior of Fe∝-bis(polyacene). Because the band 5 is (almost) filled for majority spin electrons and is completely empty for minority spin electrons, it contributes 1 μB/per magnetic ion to the net moment. The remaining moment of 0.92 μB originates from the partially filled bands 3 and 4. In the case of Fe∝-bis(BNpolyacene), all five d-bands are completely filled for majority spin, and consequently it opens up a semiconducting gap of 1.5 eV (at X point) in the majority spin channel. The dispersive bands, bands 3 and 4, which are in a mixed characters between dxz and dyz (see wave function plots in Figure 4) just cross the Fermi energy, making the system metallic in minority spin channel. The important point here is that we have a coexistence of the metallic and semiconducting nature for electrons in minority and majority spin channels, respectively, leading to a half-metallic behavior for the Fe∝bis(BNpolyacene). Because the band 5, which is derived from dxy orbital (see Figure 4), is completely filled for majority spin and is completely empty for minority spin channel, it contributes 1 μB/magnetic ion to the total moment (2.22 μB). Also, band 4, which is completely filled for majority spin and (almost) empty for minority spin channel, contributes a moment of 1 μB/magnetic ion. The remaining moment of 0.22 μB arises from the small unoccupied portion of top of the valence band (band 3) in the minority spin channel. To analyze further the metallic and half-metallic character of Fe∝-bis(polyacene) and Fe∝-bis(BNpolyacene), respectively, in Figure 5, we have presented total density of states (DOS) and projected density of states (pDOS) for Fe atoms. As can be seen clearly, the density of states near the Fermi level mainly arise from Fe atoms. For Fe∝-bis(polyacene), a finite density of states for both the spin channels at the Fermi level makes the 18491

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

peaks correspond to the resonant transmission through molecular states. Transmission shows a peak only when molecular states resonate with the states of the electrodes. In a low-energy window around the Fermi energy, the vanishing contribution of molecular states to the eigenstates of the system leads to a case of weaker resonance, which in turn makes the transmission almost negligible for majority spin electrons, whereas due to strong resonance, minority spin electrons show strong transmission peaks near the Fermi level. Looking once again at the plots of HOMO and LUMO for majority and minority spin electrons in Figure 7, it can be seen that due to weak coupling of molecular orbitals and incident states from the electrode, the HOMO and LUMO rarely contribute to the transmission for majority spin, whereas the HOMO and LUMO for the minority spins show a strong transmission peaks because of the delocalization of molecular states with the electrodes. Because the actual device works at a finite bias, in Figure 8, we have presented the spin-polarized I−

whole system freezing coordinates of the gold electrodes. In Figure 6, we have plotted the spin -polarized DOS and zero-

Figure 6. Spin-resolved DOS and T(E) plot for Au-[Fe3-bis(3BNacene)(S4)]-Au. Solid green represents the pDOS of Fe ion. The plot is scaled for EF to lie at 0 eV.

bias transmission functions. It is quite clear that within a small energy window near the Fermi level (−2 to 2 eV) only minority spin electrons show a strong transmission peak at the Fermi level (EF) and hence take part in transport, whereas a transport gap opens up for majority spin electrons. To understand this, we have plotted the total DOS and its projection onto the molecular states. As can been seen, the transmission spectra show a series of peaks with strong correlation between transmission and pDOS spectra, especially with regard to the location of their peaks. (See Figure 6.) These transmission

Figure 8. I−V characteristics for Au-[Fe3-bis(3-BNacene)(S4)]-Au. Code: black and red solid lines show I−V for majority and minority spins, respectively. Inset: Spin-polarized current (IS%) versus voltage has been plotted.

Figure 7. Wave function plot of HOMO and LUMO orbitals for majority and minority spin channels of Au-[ Fe3-bis(3-BNacene)(S4)]-Au. 18492

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

(6) Wolf, S. A.; Awschalom, D. D.; Buhrman, R. A.; Daughton, J. M.; von Molnar, S.; Roukes, M. L.; Chtchelkanova, A. Y.; Treger, D. M. Science 2001, 294, 1488. (7) Ohno, H.; Chiba, D.; Matsukura, F.; Omiya, T.; Abe, E.; Dietl, T.; Ohno, Y.; Ohtani, K. Nature 2000, 408, 944. (8) Yuasa, S.; Nagahama, T.; Fukushima, A.; Suzuki, Y.; Ando, K. Nat. Mater. 2004, 3, 868. (9) Zutic, I.; Fabian, J.; Das Sarma, S. Rev. Mod. Phys. 2004, 76, 323. (10) Dediu, V.; Murgia, M.; Matacotta, F. C.; Taliani, C.; Barbanera, S. Solid State Commun. 2002, 122, 181. (11) Henkelman, G.; Arnaldsson, A.; Jónsson, H. Comput. Mater. Sci. 2006, 36, 354. (12) Xiong, Z. H.; Wu, D.; Vardeny, Z. V.; Shi, J. Nature 2004, 427, 821. (13) Han, W.-Q.; Yu, H.-G.; Zhi, C.; Wang, J.; Liu, Z.; Sekiguchi, T.; Bando, Y. Nano Lett. 2008, 8, 491. (14) Herrmann, C.; Solomon, G. C.; Ratner, M. A. J. Am. Chem. Soc. 2010, 132, 3682. (15) de Groot, R. A.; M., F. M.; van Engen, P. G.; Buschow, K. H. J. Phys. Rev. Lett. 1983, 50, 2024. (16) Bea, H.; Bibes, M.; Sirena, M.; Herranz, G.; Bouzehouane, K.; Jacquet, E.; Fusil, S.; Paruch, P.; Dawber, M.; Contour, J. P.; Barthelemy, A. Appl. Phys. Lett. 2006, 88, 062502. (17) Maslyuk, V. V.; Bagrets, A.; Meded, V.; Arnold, A.; Evers, F.; Brandbyge, M.; Bredow, T.; Mertig, I. Phys. Rev. Lett. 2006, 97, 097201. (18) Liu, R.; Ke, S.-H.; Yang, W.; Baranger, H. U. J. Chem. Phys. 2007, 127, 141104. (19) Parida, P.; Basheer, E. A.; Pati, S. K. J. Mater. Chem. 2012, 22, 14916. (20) Zhou, L.; Yang, S.-W.; Ng, M.-F.; Sullivan, M. B.; Tan, V. B. C.; Shen, L. J. Am. Chem. Soc. 2008, 130, 4023. (21) Jiang, S. D.; Wang, B. W.; Su, G.; Wang, Z. M.; Gao, S. Angew. Chem., Int. Ed. 2010, 49, 7448. (22) Pandey, R.; Rao, B. K.; Jena, P.; Blanco, M. A. J. Am. Chem. Soc. 2001, 123, 3799. (23) Weis, P.; Kemper, P. R.; Bowers, M. T. J. Phys. Chem. A 1997, 101, 8207. (24) Hoshino, K.; Kurikawa, T.; Takeda, H.; Nakajima, A.; Kaya, K. J. Phys. Chem. 1995, 99, 3053. (25) Miyajima, K.; Nakajima, A.; Yabushita, S.; Knickelbein, M. B.; Kaya, K. J. Am. Chem. Soc. 2004, 126, 13202. (26) Miyajima, K.; Yabushita, S.; Knickelbein, M. B.; Nakajima, A. J. Am. Chem. Soc. 2007, 129, 8473. (27) Miyajima, K.; Knickelbein, M. B.; Nakajima, A. J. Phys. Chem. A 2008, 112, 366. (28) Kurikawa, T.; Negishi, Y.; Hayakawa, F.; Nagao, S.; Miyajima, K.; Nakajima, A.; Kaya, K. J. Am. Chem. Soc. 1998, 120, 11766. (29) Dutta, S.; Mandal, T. K.; Datta, A.; Pati, S. K. Chem. Phys. Lett. 2009, 479, 133. (30) Zhang, Z.; Wu, X.; Guo, W.; Zeng, X. C. J. Am. Chem. Soc. 2010, 132, 10215. (31) Parida, P.; Kundu, A.; Pati, S. K. Phys. Chem. Chem. Phys. 2010, 12, 6924. (32) Rao, C. N. R.; Satishkumar, B. C.; Govindaraj, A.; Nath, M. ChemPhysChem 2001, 2, 78. (33) Zheng, F.; Zhou, G.; Liu, Z.; Wu, J.; Duan, W.; Gu, B.-L.; Zhang, S. B. Phys. Rev. B 2008, 78, 205415. (34) Yu, Z.; Hu, M. L.; Zhang, C. X.; He, C. Y.; Sun, L. Z.; Zhong, J. J. Phys. Chem. C 2011, 115, 10836. (35) Du, A.; Chen, Y.; Zhu, Z.; Amal, R.; Lu, G. Q.; Smith, S. C. J. Am. Chem. Soc. 2009, 131, 17354. (36) Shen, L.; Yang, S.-W.; Ng, M.-F.; Ligatchev, V.; Zhou, L.; Feng, Y. J. Am. Chem. Soc. 2008, 130, 13956. (37) Xiang, H. J.; Yang, J. L.; Hou, J. G.; Zhu, Q. S. J. Am. Chem. Soc. 2006, 128, 2310. (38) Mallajosyula, S. S.; Pati, S. K. J. Phys. Chem B 2008, 112, 16982. (39) Youn, I. S.; D., Y. K.; Singh, N. J.; Sung, W. P.; Jihee, Y.; Kwang, S. K. J. Chem. Theory Comput. 2012, 8, 99.

V characteristic of the system. When bias is applied, the current reaches up to several microamperes for minority spin component, however, remains almost vanishing for the majority spin component. To get a quantitative picture of spin polarization, we have shown spin-resolved current (IS) as a function of bias in the inset of Figure 8. IS is formulated as IS(V ) =

(Imin(V ) − Imaj(V )) (Imin(V ) + Imaj(V ))

(6)

where Imin(max)(V) corresponds to current due to minority (majority) spin electrons at bias V. It can be clearly seen that IS reaches a high value ∼96% at low bias. The IS value reduces as the bias increases. Although the dominant channel for conduction is the minority spin channel at low bias, after ∼1.5 eV, the majority spin channel also starts conducting, as a result of which, the spin-polarization nature of the total current gets reduced. This IS value (96%) is much higher in comparison with the IS values found in most of the similar kind of other complexes.17,19,20,68−70 Because the finite-size cluster of Fe∝bis(BNpolyacene) allows only one kind of spin to pass through, it can possibly be realized as a good candidate for spin-filtering component in spintronics.

■

CONCLUSIONS In conclusion, we have explored how different spin states cause the structural distortion in Fe(benzene)2 and Fe2(naphthalene)2. On the basis of the crystal-field splitting and 18-electron rule, we conjecture that the structures are more stable and less distorted in their high-spin state compared with the corresponding low-spin state. We have evidently illustrated a possible way to make the proper stacking of these Fe-based sandwiched complexes by packing one infinite chain of Fe between two polyacene and BNpolyacene ligands. From the electronic structure calculation, whereas Fe∝-bis(polyacene) is an FM metal, Fe∝-bis(BNpolyacene) shows strong FM halfmetallic behavior. Owing to the high-spin ground state and efficient spin filter behavior of Fe3-bis(3-BNacene), we believe that this system possibly has a huge impetus in spintronics applications.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS P.P. acknowledges the CSIR for a research fellowship and S.K.P. acknowledges research support from the CSIR and DST, Government of India, and AOARD, U.S. Airforce, for the research grants.

■

REFERENCES

(1) Schulz, M. Nature 1999, 399, 729. (2) Vashchenko V. A., S. V. F. Physical Limitations of Semiconductor Devices; Springer, 2008. (3) Rocha, A. R.; Garcia-Suarez, V. M.; Bailey, S. W.; Lambert, C. J.; Ferrer, J.; Sanvito, S. Nat. Mater. 2005, 4, 335. (4) Dash, S. P.; Sharma, S.; Patel, R. S.; de Jong, M. P.; Jansen, R. Nature 2009, 462, 491. (5) Allwood, D. A.; Xiong, G.; Faulkner, C. C.; Atkinson, D.; Petit, D.; Cowburn, R. P. Science 2005, 309, 1688. 18493

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494

The Journal of Physical Chemistry C

Article

(69) Wang, L.; Gao, X.; Yan, X.; Zhou, J.; Gao, Z.; Nagase, S.; Sanvito, S.; Maeda, Y.; Akasaka, T.; Mei, W. N.; Lu, J. J. Phys. Chem. C 2010, 114, 1893. (70) Wang, L.; Cai, Z.; Wang, J.; Lu, J.; Luo, G.; Lai, L.; Zhou, J.; Qin, R.; Gao, Z.; Yu, D.; Li, G.; Mei, W. N.; Sanvito, S. Nano Lett. 2008, 8, 3640.

(40) Frisch M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö ; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009. (41) Soler, J. M.; Artacho, E.; Gale, J. D.; Garcia, A.; Junquera, J.; Ordejon, P.; Sanchez-Portal, D. J. Phys.: Condens. Matter 2002, 14, 2745. (42) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. (43) Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.; Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; Dal Corso, A.; de Gironcoliaaa, S.; Fabris, S.; Fratesi, G.; Gebauer, R.; Gerstmann, U.; Gougoussis, C.; Kokalj, A.; Lazzeri, M.; Martin-Samos, L.; Marzari, N.; Mauri, F.; Mazzarello, R.; Paolini, S.; Pasquarello, A.; Paulatto, L.; Sbraccia, C.; Scandolo, S.; Sclauzero, G.; Seitsonen, A. P.; Smogunov, A.; Umari, P.; Wentzcovitch, R. M. J. Phys.: Condens. Matter 2009, 21, 395502. (44) Brandbyge, M.; Mozos, J. L.; Ordejon, P.; Taylor, J.; Stokbro, K. Phys. Rev. B 2002, 65, 165401. (45) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (46) Wadt, W. R.; Hay, P. J. J. Chem. Phys. 1985, 82, 284. (47) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299. (48) Swart, M.; Ehlers, A. W.; Lammertsma, K. Mol. Phys. 2004, 102, 2467. (49) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (50) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200. (51) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (52) Chai, J.-D.; Head-Gordon, M. J. Chem. Phys. 2008, 128, 084106. (53) Chai, J.-D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615. (54) Swart, M. J. Chem. Theory Comput. 2008, 4, 2057. (55) Swart, M.; Groenhof, A. R.; Ehlers, A. W.; Lammertsma, K. J. Phys. Chem. A 2004, 108, 5479. (56) Reiher, M.; Salomon, O.; Artur Hess, B. Theor. Chem. Acc. 2001, 107, 48. (57) Chang, C. H.; Boone, A. J.; Bartlett, R. J.; Richards, N. G. J. Inorg. Chem. 2003, 43, 458. (58) Boone, A. J.; Chang, C. H.; Greene, S. N.; Herz, T.; Richards, N. G. J. Coord. Chem. Rev. 2003, 238−239, 291. (59) Swart, M. Inorg. Chim. Acta 2007, 360, 179. (60) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200. (61) Becke, A. D. Phys. Rev. A 1988, 38, 3098. (62) Perdew, J. P.; Ernzerhof, M.; Burke, K. J. Chem. Phys. 1996, 105, 9982. (63) Handy, N. C.; Cohen, A. J. Mol. Phys. 2001, 99, 403. (64) Youn, I. S.; Kim, D. Y.; Singh, N. J.; Park, S. W.; Youn, J.; Kim, K. S. J. Chem. Theory Comput. 2011, 8, 99. (65) Johnson, E. R.; Mackie, I. D.; DiLabio, G. A. J. Phys. Org. Chem. 2009, 22, 1127. (66) Sherrill, C. D.; Takatani, T.; Hohenstein, E. G. J. Phys. Chem. A 2009, 113, 10146. (67) Yamashiro, A.; Shimoi, Y.; Harigaya, K.; Wakabayashi, K. Phys. Rev. B 2003, 68, 193410. (68) Koleini, M.; Paulsson, M.; Brandbyge, M. Phys. Rev. Lett. 2007, 98, 197202. 18494

dx.doi.org/10.1021/jp304592h | J. Phys. Chem. C 2012, 116, 18487−18494