Antiferromagnetic Ordering in the Single-Component Molecular

Jul 15, 2016 - Single-component molecular material hosting antiferromagnetic and spin-gapped Mott subsystems. Rina Takagi , Takamasa Hamai , Hiro ...
0 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/IC

Antiferromagnetic Ordering in the Single-Component Molecular Conductor [Pd(tmdt)2] Satomi Ogura,† Yuki Idobata,† Biao Zhou,*,† Akiko Kobayashi,*,† Rina Takagi,‡,§ Kazuya Miyagawa,‡ Kazushi Kanoda,‡ Hidetaka Kasai,∥ Eiji Nishibori,∥ Chikatoshi Satoko,⊥ and Bernard Delley# †

Department of Chemistry, College of Humanities and Sciences, Nihon University, Setagaya-ku, Tokyo 156-8550, Japan Department of Applied Physics, University of Tokyo, Bunkyo-ku, Tokyo 113-8656, Japan § Center for Emergent Matter Science (CEMS), RIKEN, Wako-shi, Saitama 351-0198, Japan ∥ Faculty of Pure and Applied Science, Center for Integrated Research in Fundamental Science and Engineering (CIRFSE), and Tsukuba Research Center for Interdisciplinary Materials Science (TIMS), University of Tsukuba, Tsukuba 305-8571, Japan ⊥ Department of Integrated Sciences in Physics and Biology, College of Humanities and Sciences, Nihon University, Setagaya-ku, Tokyo 156-8550, Japan # Condensed Matter Theory Group, Paul Scherrer Institut, CH-5232 Villigen PSI, Switzerland ‡

S Supporting Information *

ABSTRACT: Crystals of [Pd(tmdt)2] (tmdt = trimethylenetetrathiafulvalenedithiolate) were prepared in order to investigate their physical properties. The electrical resistivity of [Pd(tmdt)2] was measured on single crystals using two-probe methods and showed that the roomtemperature conductivity was 100 S·cm−1. The resistivity behaviors implied that [Pd(tmdt)2] was a semimetal at approximately room temperature and became narrow-gap semiconducting as the temperature was decreased to the lowest temperature. X-ray structural studies on small single crystals of [Pd(tmdt)2] at temperatures of 20−300 K performed using synchrotron radiation at SPring-8 showed no distinct structural change over this temperature region. However, small anomalies were observed at approximately 100 K. Electron spin resonance (ESR) spectra were measured over the temperature range of 2.7−301 K. The ESR intensity increased as the temperature decreased to 100 K and then decreased linearly as the temperature was further decreased to 50 K, where an abrupt decrease in the intensity was observed. To investigate the magnetic state, 1H nuclear magnetic resonance (NMR) measurements were performed in the temperature range of 2.5−271 K, revealing broadening below 100 K. The NMR relaxation rate gradually increased below 100 K and formed a broad peak at approximately 50 K, followed by a gradual decrease down to the lowest temperature. These results suggest that most of the sample undergoes the antiferromagnetic transition at approximately 50 K with the magnetic ordering temperatures distributed over a wide range up to 100 K. These electric and magnetic properties of [Pd(tmdt)2] are quite different from those of the singlecomponent molecular (semi)metals [Ni(tmdt)2] and [Pt(tmdt)2], which retain their stable metallic states down to extremely low temperatures. The experimental results and the band structure calculations at the density functional theory level showed that [Pd(tmdt)2] may be an antiferromagnetic Mott insulator with a strong electron correlation.

■

exhaustively investigated.1,5−7 Interestingly, [M(tmdt)2] systems are multiorbital correlated electron systems and exhibit a variety of electronic properties when different transition metals take the place of the central metal atom (M). Singlecomponent molecular (semi)metals with even numbers of total electrons, [Ni(tmdt)2] and [Pt(tmdt)2], have been shown to possess very high conductivity and metallic behavior down to extremely low temperatures.1,5 In contrast, [Au(tmdt)2] and [Cu(tmdt)2], which have odd numbers of total electrons, exhibit magnetic phase transitions.6,7 The electromagnetic

INTRODUCTION Development of new functional molecular systems has a possibility to create a new concept. The single-component molecular metal [Ni(tmdt)2] (tmdt = trimethylenetetrathiafulvalene dithiolate) developed in 2001 has disclosed a new field of conducting materials composed of one kind of molecule.1,2 In 2014, even single-component molecular superconductor [Ni(hfdt) 2 ] (hfdt = bis(trifluoromethyl)tetrathiafulvalenedithiolate) was identified by high-pressure resistivity measurements (7.5−8.7 GPa with a maximum Tc [onset temperature] of 5.5 K).3,4 Among the single-component molecular conductors available, a series of isostructural [M(tmdt)2] (M = Ni, Au, Pt, or Cu) systems has been most © XXXX American Chemical Society

Received: May 12, 2016

A

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 1a

properties of [Au(tmdt)2] are chiefly determined by π-like frontier molecular orbitals. It undergoes an antiferromagnetic transition at 110 K and retains its metallic state down to low temperatures.6 [Cu(tmdt)2] is a novel single-component multifrontier π-d system with a new type of magnetic molecular conductor. 7 This system behaves as an unanticipated conducting, one-dimensional antiferromagnetic Heisenberg system with magnetic ordering at 13 K. The one-dimensional Heisenberg behaviors originate from the intermolecular overlap of the d-like orbitals, confirming the coexistence of one localized spin (S = 1/2) on the central {CuS4} part of each Cu(tmdt)2 molecule and π-like conduction bands. This finding suggests that a large π-d coupling, which has very rarely been realized in traditional molecular conductors, should be possible in single-component molecular conductors with a mixed alloy system of Ni and Cu complexes. A molecular (semi)metal with diluted magnetic moments [Ni1−xCux(tmdt)2] (x = 0.1−0.27) was synthesized by alloying [Ni(tmdt)2] and [Cu(tmdt)2], which are isostructural.8 Electrical resistivity and magnetoresistance measurements showed that this mixed alloy is a “molecular Kondo system” with a Kondo temperature of ∼5 K and an estimated π-d exchange interaction, Jπ‑d, of ∼30 meV. This value is approximately 20 times larger than that of a typical π-d organic system, λ-(BETS)2FeCl4, in which magnetic anions are separated from the organic conduction layers.9,10 We believe that [Ni1−xCux(tmdt)2] will exhibit unprecedentedly strong π-d interactions, as expected for a single-component molecular conductor.10 Multiorbital systems, such as [M(tmdt)2], can be studied using nuclear magnetic resonance (NMR), which is a powerful microscopic experimental technique that is able to probe the spin state in a site- and orbital-selective manner.11 13C NMR and 1H NMR measurements were performed to probe the electronic states and molecular dynamics of [Ni(tmdt)2] and [Pt(tmdt)2].12 The 13C Knight shift and nuclear spin−lattice relaxation rate confirmed that both compounds exhibited paramagnetic metallic states. Interestingly, the temperature and frequency dependences of the relaxation rate revealed unusual molecular dynamics, which presumably arise from tmdt’s twisting degrees of freedom or a novel type of molecular motion specific to the M(tmdt)2 (M = Ni or Pt) family of materials.12 Recently, we prepared a single-component molecular conductor [Pd(tmdt)2] with an even number of total electrons. [Pd(tmdt)2] is isostructural to [Ni(tmdt)2] and [Pt(tmdt)2], but its physical properties are quite different. Herein, we report the low metallicity and unusual magnetic properties of this newly developed [Pd(tmdt)2].

■

a

Reagents and conditions: (i) 25-wt% Me4N·OH/MeOH (4 equiv), dry THF, room temperature, 30 min, (ii) [PdCl2(PhCN)2]/MeOH (0.5 equiv), −78 °C to room temperature, overnight, (iii) cation exchange in nBu4N·PF6/MeCN solution, (iv) electrochemical oxidation in nBu4N·PF6/MeCN solution, ∼0.4 μA for 4 weeks. −78 °C (dry ice/MeOH bath), a solution of [PdIICl2(PhCN)2] (60.4 mg; 0.157 mmol) in MeOH (5.0 mL) was added dropwise to the reaction mixture. Then, the reaction mixture was warmed to room temperature by stirring overnight. The resultant solids were collected by filtration, and the grayish-pink crystals of (nBu4N)2[Pd(tmdt)2] were obtained from a MeCN solution of nBu4N·PF6 at room temperature. Electrochemical Synthesis of [Pd(tmdt)2]. (nBu4N)2[Pd(tmdt)2] (∼15 mg; 0.012 mmol) and nBu4N·PF6 (120 mg; 0.31 mmol) as a supporting electrolyte were poured into a standard Hshaped glass cells containing Pt electrodes and dissolved in MeCN (20.0 mL) under an argon atmosphere. Air-stable black crystals of [Pd(tmdt)2] grew on the Pt electrode, after a constant current of 0.4 μA was applied, within approximately 4 weeks at room temperature. Anal. Calcd for C18H12S12Pd: C, 30.05; H, 1.68; N, 0.00. Found: C, 30.61; H, 1.98; N, 0.00. Compared to the single-component molecular conductors with bulky ligands, such as [Ni(L)2] (L = e.g., hfdt or ptdt (= propylenedithiotetrathiafulvalenedithiolate)),3,4,15,16 it was difficult to grow high-quality single crystals of [M(tmdt)2] (M = Ni, Au, Pt, or Cu). The synthesis of [Pd(tmdt)2] yielded twinned microcrystals with sizes less than 70 μm. X-ray Structural Studies of (nBu4N)2[Pd(tmdt)2]. Single-crystal X-ray structure determination was performed using the crystals of (nBu4N)2[Pd(tmdt)2] at 180 K. The crystal data and experimental details of the crystal structure determination are listed in Table S1. The X-ray intensity measurements of (nBu4N)2[Pd(tmdt)2] were conducted on a Rigaku Micro7HFM-VariMax Saturn 724R CCD system equipped with graphite monochromated Mo−Kα radiation (λ = 0.7107 Å) and a confocal X-ray mirror. The crystal structures were refined with anisotropic displacement parameters for all non-hydrogen atoms. The hydrogen atoms were included in the final calculations, but were not refined. All the calculations were performed using the Crystal Structure crystallographic software package of the Molecular Structure Corporation (MSC) and Rigaku Corporation.17 Electrical Resistivity Measurements of [Pd(tmdt)2]. Fourprobe resistivity measurements were performed on compressed pellets of polycrystalline samples of [Pd(tmdt)2] in the temperature range of 4−300 K. Two-probe resistivity measurements were also conducted using very small and thin single crystals with a size of approximately 70 μm. Annealed Au wires (10 μm in diameter) bonded to the sample by Au paint were used as leads. Static Magnetic Susceptibility Measurements. Static magnetic susceptibility measurements of [Pd(tmdt)2] (8.67 mg; 1.21 × 10−2 mmol) were collected with a Quantum Design MPMS-7XL super-

EXPERIMENTAL SECTION

General Methods. As shown in Scheme 1, the syntheses of [Pd(tmdt)2] were similar to that of the analogues, [M(tmdt)2] (M = Ni, Pt, or Cu), reported previously.1,5,7 The tmdt ligand13 and the palladium source, trans-bis(acetonitrile)dichloropalladium(II) [PdIICl2(PhCN)2],14 were prepared according to procedures reported in the literature. All solvents were dehydrated and freshly distilled under an argon atmosphere. Other reagents were of analytical grade and used without purification. Synthesis of (nBu4N)2[Pd(tmdt)2]. Under an argon atmosphere, the tmdt ligand (121.0 mg; 0.300 mmol) was dissolved in 10.0 mL of THF solution. This solution was hydrolyzed with a 25 wt % MeOH solution of tetramethylammonium hydroxide (Me4N·OH) (440 mg; 1.2 mmol) at room temperature. The solution was stirred for 30 min, and its color changed from orange to purple. After being cooled to B

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry conducting quantum interference device (SQUID) magnetometer in the temperature range of 2.0−300 K. The applied magnetic field was 5000 Oe. The samples were wrapped in clean aluminum foil whose magnetic susceptibility was separately measured and subtracted. The diamagnetic contribution was estimated from Pascal’s constants to be −3.25 × 10−4 emu mol−1. Electron Spin Resonance (ESR) Measurements. The ESR spectra of [Pd(tmdt)2] (5.05 mg; 7.04 × 10−3 mmol) were measured using a JEOL JES-FA200 ESR spectrometer and a low-temperature unit (JEOL ES-CT470) in the temperature range of 2.7−301 K. The applied microwave power was 1.0 mW, and the modulation field was approximately 20 G. NMR Measurements. 1H NMR measurements were conducted using an assembly of microcrystals at a magnetic field of 3.66 T. The samples were packed into a Teflon tube. The NMR spectra were obtained by Fourier transformation of the so-called solid-echo signals following a (π/2)x-(π/2)y pulse sequence.18 The typical width of the π/2 pulse was 1.4 μs, which is much smaller than the inverse of the spectral width, indicating that the nuclear spins are exhaustively driven into resonance. Below 100 K, where the spectra progressively broadened, the pulse width was squeezed down to 1.1 μs to capture the whole spectra. Synchrotron Radiation (SR) X-ray Structural Studies of [Pd(tmdt)2]. X-ray structural studies of small, single crystals of [Pd(tmdt)2] at low temperatures were successfully performed using SR at SPring-8. The SR X-ray diffraction experiment was conducted on the BL02B1 beamline of SPring-8 in Japan. Data were collected at 20, 70, 110, 150, 200, and 300 K using SR with a wavelength of 0.4136 Å. The single-crystal X-ray diffraction pattern was measured with a large cylindrical image-plate camera.19 Diffraction data were used with a resolution of d > 0.6 Å and an average completeness of 1.0. The parameters of the unit cell were determined by least-squares fitting. The crystal data are shown in Table S2. The integrated intensities of all Bragg reflections were obtained using the software RAPIDAUTO,20 which was also used to apply the Lorentz−polarization corrections. Data scaling and merging were performed with SORTAV.21 The crystal structure was solved with SIR92,22 and the refinement was performed using SHELXL2013.23 The crystal structures were refined with anisotropic temperature factors for nonH atoms, and all the H atoms were refined isotropically. The X-ray powder diffraction and single-crystal diffraction experiments confirmed that no polymorphism or reflections from impurities in [Pd(tmdt)2] crystals existed. The X-ray powder diffraction data, and the results of the Rietveld analysis are shown in Figure S1, panels a and b, respectively. Calculation of the Molecular Orbital Overlap Integrals of [Pd(tmdt)2]. The extended Hückel molecular orbital calculations were performed using Slater-type atomic orbitals. The exponents (ζ) and ionization potentials (Ei) of the atomic orbitals are listed in Table S3. Molecular Orbital, Band Structure, and Fermi Surface Calculations of [M(tmdt)2] (M = Ni, Pd, or Pt). The molecular orbital, band structure, and Fermi surface calculations were performed at the density functional theory (DFT) level based on the GGA-BLYP functional using the DMol3 module as implemented in Materials Studio v5.5 (Accelrys, San Diego, CA, USA).24,25

Figure 1. (a) Molecular structure of the dianion complex [Pd(tmdt)2]2− in (nBu4N)2[Pd(tmdt)2] at 180 K. (b) The packing diagram of (nBu4N)2[Pd(tmdt)2] viewed along the a axis.

not symmetrical: One ligand was somewhat bent at the positions of the S atoms of tetrathiafulvalene (TTF) with a dihedral angle of 16.0°, and the other was bent with a dihedral angle of 6.0°. The CC distances were 1.324−1.344 Å, and no significant difference was found for either ligand. Electrical Properties of [Pd(tmdt)2]. The temperature dependence of the electrical resistivity of [Pd(tmdt)2] was measured on single crystals using two-probe methods and is shown in Figure 2. The maximum crystal size was

Figure 2. Temperature dependence of electrical resistivity of [Pd(tmdt)]. The insets indicate the existence of some transition from a highly conducting semiconducting state to a lower conducting semiconducting state near 100 K.

approximately 70 μm. The room-temperature conductivity was 100 S·cm−1, which was almost the same as that reported for [Cu(dmdt)2] (dmdt = dimethyltetrathiafulvalenedithiolate).26 The temperature dependence was weakly metallic near room temperature. Between 270 and 150 K, the resistivity (R) was approximately constant. It subsequently increased very slowly as the temperature decreased (R(100 K)/R(270 K) ≈ 1.7 and R(4 K)/R(270 K) ≈38). (d ln R/dT−1) showed a maximum near 100 K (see the inset of Figure 2), indicating that the temperature dependence of the activation energy changed at approximately 100 K. Thus, [Pd(tmdt)2] was semimetallic near room temperature and became semiconducting with small activation energies (3−9 meV, in the temperature range of 4− 150 K). A comparative study using four-probe resistivity measurements was performed on compressed pellets of

■

DISCUSSION Crystal Structure of (nBu4N)2[Pd(tmdt)2]. Figure 1a shows the molecular structure of the dianion complex [Pd(tmdt)2]2− in (nBu4N)2[Pd(tmdt)2] at 180 K, and Figure 1b shows the packing diagram viewed along the a axis. The unit cell contains one crystallographically independent [Pd(tmdt)2]2− anion and two nBu4N+ cations. The coordination geometry of [Pd(tmdt)2]2− was slightly distorted from planar with an average Pd−S distance of 2.319 Å. The average S−Pd− S angle was 91.20°. The dihedral angle between the two leastsquares planes of the PdS2C2 five-membered rings was 8.76°. The geometry of the two tmdt ligands around the Pd atom was C

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry polycrystalline samples of [Pd(tmdt)2], which showed a behavior similar to single crystals using two-probe methods (Figure S2). ESR of [Pd(tmdt)2]. The ESR spectra of polycrystalline samples of [Pd(tmdt)2] were measured at 2.7−300 K. Figure 3a

Figure 4. Temperature dependence of the magnetic susceptibility of [Pd(tmdt)]. The blue squares show the susceptibility subtracted with impurities.

mol−1. After impurity corrections, the magnetic susceptibility was 1.4 × 10−4 emu mol−1 at 300 K and decreased slowly as the temperature decreased to approximately 180 K; however, below 150 K, the magnetic susceptibility increased slightly. Judging from the behavior of the magnetic susceptibility after impurity corrections, the fine magnetic behavior appears to have been concealed by the small quantity of the samples, the magnetic susceptibility of the aluminum foil, and the impurity spins. In the high-temperature region, nonlocalized conduction electrons show weak Pauli paramagnetic behavior, similar to that of [M(tmdt)2] (M = Ni or Pt). Below approximately 100 K, the magnetic susceptibility was observed to increase, likely because of the existence of the magnetic moments of nonsubtracted impurity spins. Unfortunately, the magnetic susceptibility showed no anomaly (shoulder) peak at approximately 50 K. 1 H NMR Measurements of [Pd(tmdt)2]. As mentioned above, the ESR signal intensity decreased below 100 K, implying the occurrence of a magnetic transition. To further investigate the magnetic state, we performed 1H NMR measurements at 2.5−271 K. Typical NMR spectra are shown in Figure 5. The spectral width was insensitive to

Figure 3. (a) Dysonian-like ESR signal at 180 K and temperature dependence of the IA/IB values of [Pd(tmdt)2]. IA and IB are the integrated intensities of the ESR signal. (b) Temperature dependence of the spin susceptibility of [Pd(tmdt)2]. (c) Temperature dependences of the peak-to-peak line width and g-value of [Pd(tmdt)2]. The green symbols show the ESR signal due to the paramagnetic impurities.

shows a Dysonian-like ESR signal at 300 K and the temperature dependence of the IA/IB values of [Pd(tmdt)2]. IA and IB are the integrated intensities of the ESR signal and are indicated in the inset of Figure 3a. The temperature dependence observed for IA/IB values exceeding 1.0 indicated that the semimetallic region of [Pd(tmdt)2] existed between room temperature and approximately 180 K. At 100−180 K, the IA and IB signals were almost equal. Remarkably, the ESR signals between 50 and 100 K were disordered, likely because of the fluctuation of the anomalous magnetic phase transition at 50 K. Figure 3b shows the temperature dependence of the ESR intensity. The ESR intensity increased as the temperature decreased to 100 K and then decreased gradually as the temperature was further decreased to 50 K, at which point an abrupt decrease in the intensity occurred. Below 50 K, as the intensity of the signal was less than 1% of that of the maximum peak at 100 K, it was thought to be produced by the presence of paramagnetic impurities. These results indicate that an anomalous magnetic phase transition (antiferromagnetic phase transition) occurred at approximately 50 K. Figure 3c shows the peak-to-peak line widths (ΔHpp) and g values. At approximately 300 K, the peak-to-peak line width exceeded 200 G, and it decreased as the temperature decreased to 50 K, where it became ∼75 G. Interestingly, the g value was 1.990 at approximately 300 K and increased slightly as the temperature decreased, becoming 2.008 at approximately 50 K. Magnetic Susceptibility of [Pd(tmdt)2]. The temperature dependence of the magnetic susceptibility is shown in Figure 4. The increase in susceptibility at low temperature was ascribed to the 0.8% paramagnetic (S1/2) impurities (C = 3.04 × 10−3 emu·K·mol−1, θ = −2.11 K). The diamagnetic contribution was estimated using Pascal’s constants to be −3.25 × 10−4 emu

Figure 5. 1H NMR spectra of [Pd(tmdt)2] measured at 2.5 and 156 K.

temperature above 100 K. The full-width at half-maximum at 156 K was approximately 50 kHz, which can be explained by the nuclear-dipole interactions between protons in the trimethylene group. The spectra showed broadening below 100 K and extended over a frequency range of ±300 kHz or more at the lowest temperature. This result constitutes clear evidence for the emergence of internal magnetic fields associated with magnetic ordering in this sample. The NMR relaxation rate gradually increased below 100 K and exhibited a broad peak at approximately 50 K, followed by a gradual decrease down to the lowest temperature. This behavior suggests that the magnetic ordering temperature is distributed over a wide range up to 100 K and that a major part of the sample undergoes a magnetic transition at approximately 50 K. This behavior is consistent with the observation that the ESR intensity decreases below 100 K and reaches a minimum at D

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry 50 K. The distribution of the ordering temperature might be attributable to the existence of some inhomogeneity in the sample. X-ray Structural Studies of [Pd(tmdt)2] at Low Temperatures. To improve our understanding of the physical properties of [Pd(tmdt)2], single-crystal X-ray structure determinations were performed using SR at SPring-8 in the temperature range of 20−300 K. [Pd(tmdt)2] is isostructural to [M(tmdt)2] (M = Ni, Pt, Au, or Cu). The molecular and crystal structures at 20 K are shown in Figure 6a−c.

Figure 7. Temperature dependences of unit cell volume and lattice constants of [Pd(tmdt)2].

ordering in α-(BEDT-TTF)2I3 was clarified by this method via SR X-ray single-crystal diffraction.29 When the chemically equivalent C−S and CC bond lengths of the Pd(tmdt)2 molecule in Figure 8 are averaged, the

Figure 6. Molecular structure and crystal structure of [Pd(tmdt)2] determined at 20 K. (a) Molecular plane and side view of the molecule. (b) Crystal structure of [Pd(tmdt)2] viewed along the a axis. (c) Crystal structure of [Pd(tmdt)2] viewed along the molecular long axis.

The temperature dependences of the lattice constants and the cell volume are shown in Figure 7, and the intermolecular S···S close contacts shorter than the van der Waals distance (3.70 Å) along the [100], [111], [001], and [101] directions are shown in Figure S3, panels a, b, c, and d, respectively. These crystal structure determinations at six different temperatures between 20 and 300 K showed that no drastic crystal structural change occurred. However, below approximately 100 K, the temperature dependences of the lattice constants diminished, and the intermolecular S···S close contacts were slightly shortened. The bond lengths of Pd(tmdt)2 were precisely examined, in order to know whether an intramolecular electron transfer occurs between the central metal atom and the two πligands. It is well-known that the valence of TTF (= tetrathiafulvalene) related molecules, such as TTF, BEDT-TTF (= bis(ethylenedithio)tetrathiafulvalene), and TCNQ (= 7,7′,8,8′-tetracyanoquinodimethane), can be estimated by an empirical method based on the bond lengths obtained by the accurate crystal structure determination.27,28 Indeed, the charge

Figure 8. Temperature dependences of δ (= d̅C−S − d̅CC) and γ (d̅CC/d̅C−S). δ and γ are thought to be proportional to the valence of the Pd(tmdt)2 molecule. d̅C−S and d̅CC are the average bond lengths of 10 C−S and 3 CC, respectively.

values d̅C−S and d̅CC are obtained. Here, δ and γ are defined as δ = d̅C−S−d̅CC and γ = d̅CC/d̅C−S and are thought to be proportional to the valence of the molecule. We calculated the temperature dependences of δ and γ, which are shown in Figure 8. Figure 9 presents the temperature dependence of the averaged Pd−S distance. The averaged Pd−S distance (2.303 Å at 300 K) was almost constant down to approximately 150 K and then became slightly shorter (2.298 Å at 110 K) (the standard deviations of the Pd−S distances are less than 0.003 Å). Although very small anomalies were observed at approximately 100 K, considering the very small values of the changes in δ, γ, and the Pd−S distances, Figures 8 and 9 predict that the valence of Pd(tmdt)2 will be almost constant between 20 and 300 K. Therefore, in [Pd(tmdt)2], no drastic structural E

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

(Figure 11). Our most interesting point of research is clarifying the reason why [Pd(tmdt)2]’s physical properties are so different from those of [Ni(tmdt)2] and [Pt(tmdt)2] crystals.

Figure 9. Temperature dependence of the averaged Pd−S distance.

change and no intramolecular electron transfer between the central Pd atom and the two π ligands occur. To examine the temperature dependences of the intermolecular interactions, the highest occupied molecular orbital (HOMO)−HOMO, lowest unoccupied molecular orbital (LUMO)−LUMO, and HOMO−LUMO overlap integrals of [Pd(tmdt)2] were calculated. The extended Hückel molecular orbital calculations were performed using the Slater-type atomic orbitals shown in Table S3, and the overlap integrals calculated based on the crystal structures are shown in Figure 10 and

Figure 11. Molecular orbitals and the energy levels of M(tmdt)2 (M = Ni, Pd, or Pt) molecules.

The electronic band structure calculations of [M(tmdt)2] (M = Ni, Pd, or Pt) were performed at the DFT level based on the GGA-BLYP functional using the DMol3 module.24,25 Only the valence electrons were considered in the calculation, with the core being replaced by norm-conserving scalar relativistic pseudopotentials.30,31 Figure 12 presents the band energies in the vicinity of the Fermi level for M(tmdt)2 (M = Ni, Pd, or Pt). We carefully investigated the band energies near the Fermi level, which appear fairly similar. However, in [Pd(tmdt)2], a slight difference near M(1/2a*,1/2b*,0) was observed, which corresponded to the disappearance of the hole Fermi surface. This disappearance was produced by the splitting of the Fermi pocket at M(1/2a*,1/2b*, 0) into two separate pockets because of a minor structural change, as predicted by the [Ni(tmdt)2] band structure calculation.30 Figure 13 shows the Fermi surfaces of [M(tmdt)2] (M = Ni, Pd, or Pt). The electron and hole Fermi surfaces of [Pd(tmdt)2] indicate that it is a typical semimetallic crystal. The band widths of [Pd(tmdt)2] are 0.75 eV for the HOMO band and 0.67 eV for the LUMO band, which are comparable to those of [Ni(tmdt)2] (0.66 eV for the HOMO band and 0.54 eV for the LUMO band) obtained by Rovira et al.30 The overlap of the two bands is 0.24 eV, which is slightly larger than that of [Ni(tmdt)2] (0.19 eV). Generally, a narrow bandwidth of approximately 0.7 eV for the upper and lower bands implies that localization phenomena occur readily in strong electron correlation systems.12 The narrow-band semimetals are thus prone to various types of instabilities. The band structure calculations at 20 K shown in Figure S4 indicate that the Fermi surface was almost the same as that at 300 K. At the present stage, a nesting vector which derives a magnetic phase transition has not been found. Considering the semiconducting nature of [Pd(tmdt)2] at low temperatures, [Pd(tmdt)2] is believed to be an antiferromagnetic Mott insulator with a strong electron correlation. The semiconducting temperature dependence of the resistivity down to the lowest temperature and the decreasing antiferromagnetic ordering temperature observed in the present system are known to occur when the Mott insulating state residing near the Mott transition has quenched disorder.32 The presence of this disorder is consistent with the inhomogeneous spin state of this system. To more precisely elucidate the origin of [Pd(tmdt)2]’s unusual physical properties, further experiments, including examinations of the Hall effect, the thermoelectric power, 13C NMR, and the electrical conductivity at high pressure, and band structure calculations with a strong electron

Figure 10. Temperature dependences of overlap integrals of HOMO− HOMO (a), LUMO−LUMO (b), and HOMO−LUMO (c) calculated along the [100], [111], [001], and [101] directions (d).

Table S4. The temperature dependences of the absolute values of overlap integrals increased as the temperature decreased to 20 K. Similar to the temperature dependences of the S···S short contacts, the intermolecular interactions became stronger as the temperature decreased. However, at approximately 100 K, small anomalies were observed in both, suggesting that the electric and magnetic properties of [Pd(tmdt)2] may have changed somewhat. Band Structure Calculations and Electronic Structures of [Pd(tmdt)2]. [Pd(tmdt)2] was in a semimetallic state near room temperature, and it became narrow-gap semiconducting as the temperature decreased. An unanticipated magnetic phase transition occurred at approximately 50 K; this was not observed for [Ni(tmdt)2] and [Pt(tmdt)2], which retain their metallic properties to extremely low temperatures. This magnetic phase transition is thought to be an antiferromagnetic one based on the ESR and NMR results. In simple DFT molecular orbital calculations of M(tmdt)2 (M = Ni, Pd, or Pt),5,30 the energy level pdσ(−) was much higher than those of sym-Lπ and asym-Lπ(d), which form stable 3D metal bands F

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 12. Band energy dispersion curve of [M(tmdt)2] (M = Ni, Pd or Pt) around the Fermi energies. The symbols Γ, X, Y, Z, M, N, P, and Q represent the following positions in the reciprocal space: Γ (0,0,0), X (1/2a*,0,0), Y (0, 1/2b*,0), Z (0,0, 1/2c*), M (1/2a*,1/2b*,0), N (1/2a*,0, 1/2c*), P (0, 1/2b*,1/2c*) and Q (1/2a*,1/2b*,1/2c*).

Figure 13. Fermi surfaces of [M(tmdt)2] (M = Ni, Pd, or Pt) viewed along the (a) a* axis and (b) c* axis.

correlation, including the spin−orbit coupling, are necessary and will be performed in the future.

[Pd(tmdt)2]; additional X-ray crystallographic data including the intermolecular short S···S distances of [Pd(tmdt)2]; the electrical resistivity of polycrystalline samples of [Pd(tmdt)2]; the detail of extended Hückel molecular orbital calculations; the band energy dispersion curve and the Fermi surfaces of [Pd(tmdt)2] at 20 and 300 K (PDF) X-ray crystallographic data in CIF format (CIF1, CIF2, CIF3, CIF4, CIF5, CIF6, CIF7)

■

CONCLUSION A new single-component molecular conductor, [Pd(tmdt)2], was synthesized. The crystal structure determination of [Pd(tmdt)2] revealed that it is isostructural to [Ni(tmdt)2] and [Pt(tmdt)2], and simple DFT MO and band structure calculations predicted that it had a semimetallic band roughly similar to those of [Ni(tmdt)2] and [Pt(tmdt)2]. However, we found that its physical properties are quite different. [Ni(tmdt)2] and [Pt(tmdt)2] are (semi)metallic down to very low temperatures, whereas [Pd(tmdt)2] is a semimetal near room temperature and then becomes semiconducting as the temperature decreases to the lowest temperature. Additionally, it exhibited an unanticipated magnetic phase transition at approximately 50 K. Although antiferromagnetic ordering is observed in M(tmdt)2 (M = Au or Cu), which has an odd number of total electrons, [Pd(tmdt)2] is the only singlecomponent molecular conductor with an even number of total electrons to exhibit antiferromagnetic ordering.

■

■

AUTHOR INFORMATION

Corresponding Authors

*(A.K.) E-mail: [email protected]. *(B.Z.) E-mail: [email protected]. Author Contributions

The syntheses and characterizations [Pd(tmdt)2] (electrical resistivity, ESR, magnetic susceptibility measurements) were performed by S.O., Y.I., B.Z., and A.K. NMR measurements and their analyses were done by R.T., K.M., and K.K. X-ray crystal structure determinations by using synchrotron radiation at SPring-8 were done by H.K. and E.N. The MO and band structure calculations were done by B.Z., C.S., and B.D. All authors have given approval to the final version of the manuscript.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.6b01166. X-ray crystallographic data of (nBu4N)2[Pd(tmdt)2]; Xray powder diffraction data and their Rietvelt analysis of

Notes

The authors declare no competing financial interest. G

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

■

(14) Anderson, G. K.; Lin, M. Inorg. Synth. 1990, 28, 60−63. (15) Kobayashi, A.; Tanaka, H.; Kumasaki, M.; Torii, H.; Narymbetov, B.; Adachi, T. J. Am. Chem. Soc. 1999, 121, 10763− 10771. (16) Zhou, B.; Yajima, H.; Idobata, Y.; Kobayashi, A.; Kobayashi, T.; Nishibori, E.; Sawa, H.; Kobayashi, H. Chem. Lett. 2012, 41, 154−156. (17) CrystalStructure 3.8.0; Crystal Structure Analysis Package; Rigaku and Rigaku/MSC: The Woodlands, TX, 2000−2006. (18) Slichter, C. P. Principles of Magnetic Resonance; Springer: Berlin, Germany, 1998. (19) Sugimoto, K.; Ohsumi, H.; Aoyagi, S.; Nishibori, E.; Moriyoshi, C.; Kuroiwa, Y.; Sawa, H.; Takata, M. AIP Conf. Proc. 2010, 1234, 887−890. (20) RAPID-AUTO; Rigaku Corporation: Tokyo, Japan, 2004. (21) Blessing, R. H. Crystallogr. Rev. 1987, 1, 3−58. (22) Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.; Burla, M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27, 435. (23) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, 64, 112−122. (24) (a) Delley, B. J. Chem. Phys. 1990, 92, 508−517. (b) Delley, B. J. Chem. Phys. 2000, 113, 7756−7764. (c) DMol3: Accelrys Material Studio 5.5; Accelrys Software Inc.: San Diego, CA, 2011. (25) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865−3868. (26) Zhou, B.; Idobata, Y.; Kobayashi, A.; Cui, H.; Kato, R.; Takagi, R.; Miyagawa, K.; Kanoda, K.; Kobayashi, H. J. Am. Chem. Soc. 2012, 134, 12724−12731. (27) Umland, T. C.; Allie, S.; Kuhlmann, T.; Coppens, P. J. Phys. Chem. 1988, 92, 6456−6460. (28) (a) Guionneau, P.; Kepert, C. J.; Bravic, G.; Chasseau, D.; Truter, M. R.; Kurmoo, M.; Day, P. Synth. Met. 1997, 86, 1973−1974. (b) Kobayashi, H.; Kobayashi, A.; Sasaki, Y.; Saito, G.; Inokuchi, H. Chem. Lett. 1984, 183−186. (29) Kakiuchi, T.; Wakabayashi, Y.; Sawa, H.; Takahashi, T.; Nakamura, T. J. Phys. Soc. Jpn. 2007, 76, 113702. (30) Rovira, C.; Novoa, J. J.; Mozos, J.-L.; Ordejon, P.; Canadell, E. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 65, 081104. (31) (a) Delley, B. Int. J. Quantum Chem. 1998, 69, 423−433. (b) Delley, B. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 66, 155125. (32) Sasaki, T. Crystals 2012, 2, 374−392.

ACKNOWLEDGMENTS Authors thank Prof. M. Tokumoto (National Defence Academy of Japan) for the preliminary susceptibility measurements. Authors thank Prof. H. Kobayashi (Nihon University), Dr. S. Ishibashi (National Institute of Advanced Industrial Science and Technology), and Prof. K. Terakura (Japan Advanced Institute of Science and Technology) for valuable discussions. This study was financially supported by Grant-in-Aid on Innovative Areas (Nos. 20110002 and 20110003) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. The study was also supported by the “Ministry of Education, Culture, Sports, Science and Technology-Supported Program for the Strategic Research Foundation at Private Universities, 2009-2013 (No. S0901022)). The synchrotron radiation experiment was performed at the BL02B1 of SPring-8 (Proposal No. 2015A0078) with the approval of the Japan Synchrotron Radiation Research Institute (JASRI).

■

REFERENCES

(1) (a) Tanaka, H.; Okano, Y.; Kobayashi, H.; Suzuki, W.; Kobayashi, H. Science 2001, 291, 285−287. (b) Kobayashi, A.; Fujiwara, E.; Kobayashi, H. Chem. Rev. 2004, 104, 5243−5264. (2) (a) Belo, D.; Alves, H.; Branco Lopes, E. B.; Duarte, M. T.; Gama, V.; Teives Henriques, R. T.; Almeida, M.; Perez-Benitez, A.; Rovira, C.; Veciana, J. Chem. - Eur. J. 2001, 7, 511−519. (b) Tenn, N.; Bellec, N.; Jeannin, O.; Piekara-Sady, L.; Auban-Senzier, P.; Iniguez, J.; Canadell, E.; Lorcy, D. J. Am. Chem. Soc. 2009, 131, 16961−16967. (c) Tse, J. S.; Leitch, A. A.; Yu, X.; Bao, X.; Zhang, S.; Liu, Q.; Jin, C.; Secco, A.; Desgreniers, S.; Ohishi, Y.; Oakley, R. T. J. Am. Chem. Soc. 2010, 132, 4876−4886. (d) Cui, H.; Tsumuraya, T.; Miyazaki, T.; Okano, Y.; Kato, R. Eur. J. Inorg. Chem. 2014, 2014, 3837−3840. (3) Cui, H.; Kobayashi, H.; Ishibashi, S.; Sasa, M.; Iwase, F.; Kato, R.; Kobayashi, A. J. Am. Chem. Soc. 2014, 136, 7619−7622. (4) Sasa, M.; Fujiwara, E.; Kobayashi, A.; Ishibashi, S.; Terakura, K.; Okano, Y.; Fujiwara, H.; Kobayashi, H. J. Mater. Chem. 2005, 15, 155− 163. (5) Zhou, B.; Kobayashi, A.; Okano, Y.; Nakashima, T.; Aoyagi, S.; Nishibori, E.; Sakata, M.; Tokumoto, M.; Kobayashi, H. Adv. Mater. 2009, 21, 3596−3600. (6) (a) Suzuki, W.; Fujiwara, E.; Kobayashi, A.; Fujishiro, Y.; Nishibori, E.; Takata, M.; Sakata, M.; Fujiwara, H.; Kobayashi, H. J. Am. Chem. Soc. 2003, 125, 1486−1487. (b) Zhou, B.; Shimamura, M.; Fujiwara, E.; Kobayashi, A.; Higashi, T.; Nishibori, E.; Sakata, M.; Cui, H.; Takahashi, K.; Kobayashi, H. J. Am. Chem. Soc. 2006, 128, 3872− 3873. (c) Zhou, B.; Kobayashi, A.; Okano, Y.; Cui, H.; Graf, D.; Brooks, J. S.; Nakashima, T.; Aoyagi, S.; Nishibori, E.; Sakata, M.; Kobayashi, H. Inorg. Chem. 2009, 48, 10151−10157. (7) Zhou, B.; Yajima, H.; Kobayashi, A.; Okano, Y.; Tanaka, H.; Kumashiro, T.; Nishibori, E.; Sawa, H.; Kobayashi, H. Inorg. Chem. 2010, 49, 6740−6747. (8) Idobata, Y.; Zhou, B.; Kobayashi, A.; Kobayashi, H. J. Am. Chem. Soc. 2012, 134, 871−874. (9) Kobayashi, H.; Cui, H.; Kobayashi, A. Chem. Rev. 2004, 104, 5265−5288. (10) Yasuzuka, S.; Idobata, Y.; Zhou, B.; Kobayashi, A.; Katoh, K.; Cui, H.; Kato, R.; Tokumoto, M.; Kobayashi, H. J. Phys. Soc. Jpn. 2014, 83, 074701. (11) Seo, H.; Ishibashi, S.; Otsuka, Y.; Fukuyama, H.; Terakura, K. J. Phys. Soc. Jpn. 2013, 82, 054711. (12) Takagi, R.; Miyagawa, K.; Yoshimura, M.; Gangi, H.; Kanoda, K.; Zhou, B.; Idobata, Y.; Kobayashi, A. Phys. Rev. B: Condens. Matter Mater. Phys. 2016, 93, 024403. (13) (a) Binet, L.; Fabre, J. M.; Montginoul, C.; Simonsen, K. B.; Becher, J. J. Chem. Soc., Perkin Trans. 1 1996, 783−788. (b) Binet, L.; Montginoul, C.; Fabre, J. M.; Ouahab, L.; Golhen, S.; Becher, J. Synth. Met. 1997, 86, 1825−1826. (c) Fujiwara, E.; Kobayashi, A.; Fujiwara, H.; Kobayashi, H. Inorg. Chem. 2004, 43, 1122−1129. H

DOI: 10.1021/acs.inorgchem.6b01166 Inorg. Chem. XXXX, XXX, XXX−XXX