Observation of a magnetic phase transition but absence of an electrical response in a new two-dimensional mixed-valence nickel-bis-dithiolene molecular crystal

Abstract

A new mixed-valence molecular crystal, [C₃-Apy][Ni(dmit)₂]₃ (1) (dmit²⁻ = 2-thioxo-1,3-dithiole-4,5-dithiolate and C₃-Apy⁺ = 4-amino-1-propylpyridinium) was synthesized utilizing a facile solution process and a small size of counter-cation. This is distinct from the general strategy to obtain mixed-valence of [Ni(dmit)₂] molecular crystal in which the large size counter-cation is desirable and the electrocrystallization method was used. The mixed-valence molecular crystal shows a magnetic phase transition around 77 K with a ca. 6 K thermal hysteresis loop. The crystal structure analysis at ambient temperature, the electronic band structure calculation based on the crystal structure, the variable temperature infrared spectra in 6–293 K, and the conductance measurements in 4–295 K were further performed, and the results revealed that 1 is a semiconductor and the magnetic phase transition is probably related to the electronic structure change in the {[Ni(dmit)₂]₃}¹⁻.

---

1. Introduction

The planar [M(dmit)₂] (dmit²⁻ = 2-thioxo-1,3-dithiole-4,5-dithiolate; M = Ni, Pd and Pt), an excellent building block, has been widely used to construct molecular crystals¹ with novel electrical,² photo-conducting³ or magnetic⁴ properties since the 1960's.

The dithiolene unit in dmit²⁻ ligand extends with SR groups, and this leads to the presence of versatile S⋯S interactions between neighboring [Ni(dmit)₂] moieties in the crystal. The face-to-face π-stacking, lateral-to-lateral and head-to-tail intermolecular S⋯S contacts between the neighboring [Ni(dmit)₂] moieties could give rise to the molecular crystals with a one-dimensional (1-D) chain,⁵ 1-D ladder,⁶ two-dimensional (2-D) layer,⁷ or three-dimensional (3-D) lattice⁸ of [Ni(dmit)₂].

The oxidation states of [M(dmit)₂] (M = Ni, Pd, Pt) span from −2 to 0. By contrast, the molecular crystal with only the [M(dmit)₂]⁻ species is frequently obtained, and the reports on mixed-valence [M(dmit)₂] molecular crystals are limited.⁹ To get the mixed-valence [M(dmit)₂] (M = Ni, Pd, Pt) molecular crystal, a large sized counter-cation is generally desirable¹⁰ and the electrocrystallization method is usually used.¹¹ These synthesis strategies produced the mixed-valence molecular crystals with various average oxidation states of [M(dmit)₂]ⁿ⁻, such as n = 2/5,^11b 2/3,¹² 1/3 (ref. 13) and 1/2.¹⁴ In previous works, we got a mixed-valence [C₈-Apy]₂[Ni(dmit)₂]₃ molecular crystal (C₈-Apy⁺ = 4-amino-1-octaylpyridinium) by inducing a small sized C₈-Apy⁺ into the [M(dmit)₂] lattice via the simple solution process. Interestingly, such a mixed-valence molecular crystal shows rapid, clear and stable responses of photoconductivity under UV irradiation.¹⁵

In this paper, we report a new mixed-valence molecular crystal, [C₃-Apy][Ni(dmit)₂]₃ (1; C₃-Apy⁺ = 4-amino-1-propylpyridinium), which shows a magnetic phase transition around 77 K with a ca. 6 K thermal hysteresis loop. We explored the possible origin of the phase transition via the investigation of variable temperature infrared spectra, temperature dependent conductivity, magnetic susceptibility and electronic band structure.

2. Experimental section

2.1. Materials and chemicals

All chemicals and reagents were purchased from commercial sources and used without any further purification. 4,5-Di(thiobenzoyl)-1,3-dithiole-2-thione was prepared according to the published procedure.¹⁶ [C₃-Apy]Br was synthesized using a similar procedure for the preparation of [C₆-Apy]Br in the literature.¹⁷

2.2. Synthesis of [C₃-Apy][Ni(dmit)₂]₃

4,5-Di(thiobenzoyl)-1,3-dithiole-2-thione (812 mg, 2 mmol) suspended in methanol (10 mL) was mixed with a methanol solution (10 mL) containing 184 mg (8 mmol) of sodium under an argon atmosphere at ambient temperature. The mixture was stirred for 30 min to give a wine red solution. NiCl₂·6H₂O (238 mg, 1 mmol) and [C₃-Apy]Br (217 mg, 1 mmol) in methanol (20 mL) were added to the wine red solution with strong stirring, and the precipitant was immediately formed. The solution of I₂ (127 mg, 0.5 mmol) and NaI (150 mg, 1 mmol) in methanol (20 mL) was then added to the above-obtained mixture with strong stirring for 120 minutes. The microcrystalline product was collected by filtration, washed with MeOH and dried in vacuum. Yield: ∼65%. Anal. calc. for C₂₆H₁₃N₂S₃₀Ni₃ (1): C, 20.94; H, 0.88; N, 1.88%. Found: C, 21.16; H, 0.90; N, 1.89%. IR (KBr pellet, cm⁻¹): 1319, 1358 (νC=C), 1076, 1063 (νC=S), 507 and 503 (δS–C–S).

The dark green crystals of 1 suitable for X-ray diffraction analyses were obtained by slow evaporation of the solution of 1 in acetone at ambient temperature over 7–10 days. It is worth noting that we obtained the mixed-valence [Ni(dmit)₂] molecular crystal utilizing a facile solution process and small sized counter-cation, which is distinct from the general strategy in which a large sized counter-cation is desirable and the electrocrystallization method was used.

2.3. Physical measurements

Elemental analyses (C, H and N) were performed on a Vario EL III elemental analyzer. Powder X-ray diffraction (PXRD) data were collected on a Bruker D8 Advance powder diffractometer with Cu Kα radiation (λ = 1.5418 Å). IR spectra were recorded at room temperature on a Bruker VERTEX80V FTIR spectrometer (as KBr discs) under vacuum. Temperature-dependent FTIR spectra were recorded on a Thermo Nicolet 8700 spectrometer with a combination of an Oxford Variox AC-TL optical cryostat instrument. The sample temperature was controlled from 6 to 293 K through an ITC503 digital temperature controller. Magnetic susceptibility data were measured for polycrystalline samples on a Quantum Design MPMS-5S superconducting quantum interference device (SQUID) magnetometer, such a measurement was performed over the temperature range of 1.8–400 K in both cooling and warming modes. The measurement of variable temperature electrical conductivity based on four-probe method was made for two selected thin plate-shaped single crystals using a Keithley 2400 sourcemeter and Agilent 3440 1A digit multimeter, and the temperature varied between 4 and 295 K.

2.4. X-ray single crystallography. Single-crystal X-ray diffraction data were collected on a Bruker Smart Apex II CCD detector with graphite-monochromated Mo Kα radiation (λ = 0.71073 Å) at 296 K using the ω-scan technique. The data were integrated using the SAINT program, which was also used for the intensity corrections for the Lorentz and polarization effects. An empirical absorption correction was applied using the SADABS program.¹⁸ The structure was solved by direct methods using the program SHELXS-97 and all non-hydrogen atoms were anisotropically refined on F² by the full-matrix least-squares technique using the SHELXL-97 crystallographic software package.¹⁹ All hydrogen atoms were placed at the calculated positions and refined as riding on the parent atoms. Details about data collection, structure refinement, and crystallography are summarized in Table 1.

Table 1 Crystallographic data and refinement parameters for 1

a R1 = ∑||Fo| − |Fc||/|Fo|.b wR2 = [∑w(∑Fo^2 − Fc^2)^2/∑w(Fo^2)^2]^1/2.
Temp./K	296(2)
Wavelength/Å	0.71073
Formula	C26H13N2Ni3S30
Formula weight	1491.31
Space group	P1̄
CCDC no.	1027558
Crystal system	Triclinic
a/Å	11.5235(13)
b/Å	12.6902(13)
c/Å	19.485(2)
α/°	74.904(3)°
β/°	75.852(4)°
γ/°	63.906(3)°
V/Å^−3, Z	2442.8(4)/2
ρ/g cm^−1	2.028
μ/mm^−1	2.458
F(000)	1494.0
θ range for data collection (°)	1.09–27.55
Index ranges	−14 ≤ h ≤ 14
	−16 ≤ k ≤ 16
	−25 ≤ l ≤ 24
Rint	0.0912
Independent reflect./restraints/parameters	11 221/0/552
Refinement method	The least square refinement on F^2
Goodness-of-fit on F^2	1.008
R1, wR2^a [I > 2σ(I)]	0.0646, 0.1209
R1, wR2^b [all data]	0.1669, 0.1557
Residual/e Å^−3	1.111/−0.647

---

2.5. Details of density functional theory calculation for electronic band structures and densities of states

Calculation for electronic band structure and densities of states (DOS) was performed using the CASTEP program²⁰ on the non-modelized crystal structure of 1, which was obtained from X-ray single crystal structure analyses at room temperature. The electrons in the orbitals of Ni (3d⁸4s²), S (3s²3p⁴) and C (2s²2p²) are treated as the valence electrons. The generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE)²¹ is used for the exchange–correlation functional. The k-points in the Brilliouin zone were set to be 3 × 3 × 4 according to the Monkhorst–Pack scheme.²² The plane wave basis set cut-off used was 270 eV and the convergence criterion was 2.0 × 10⁻⁶ eV per atom for all relevant calculations, and other calculating parameters were set to be the default values in the CASTEP code.

3. Results and discussion

3.1. Crystal structure

Compound 1 crystallizes in the triclinic space group P1̄. An asymmetric unit is comprised of one C₃-APy⁺ cation together with three planar [Ni(dmit)₂] units (ref. Fig. 1). The bond lengths and angles in C₃-APy⁺ cation fall within the expected values. All atoms are located in the general positions in the [Ni(dmit)₂] moieties containing Ni(1), Ni(2) and Ni(3). The typical bond lengths in three different [Ni(dmit)₂] species are summarized in Table 2. The total charge of three [Ni(dmit)₂] moieties is 1−. Since the cation shows a 1+ charge, the averaged oxidation state of [Ni(dmit)₂] moieties in an asymmetric unit is −1/3. For the [Ni(dmit)₂] species, the bond C=C, Ni–S and C–S lengths are sensitive to its oxidation state. The theoretical analysis based on DFT calculations predicted that the lowest unoccupied molecule orbital (LUMO) of the neutral [Ni(dmit)₂]⁰ consists of the 3d orbitals of Ni and the π orbital of ligand,¹⁵ and is bonding with respect to the π(C=C) bond and antibonding with respect to the π(C–S) and π(Ni–S) bonds. The addition of one electron to the LUMO to give [Ni(dmit)₂]¹⁻ was expected to strengthen the π(C=C) bond and weaken the π(C–S), π(Ni–S) bonds. As a result, the π(C=C) bond distance is reduced, the π(C–S) and π(Ni–S) distances are increased with increasing n value in the series [Ni(dmit)₂]ⁿ⁻ (n = 0, 1 and 2).^11b To inspect carefully the C=C, Ni–S and C–S lengths in 1, it can be concluded that the [Ni(dmit)₂] species containing Ni(2) and Ni(3) are neutral ones, while the [Ni(dmit)₂] species containing Ni(1) has 1− charge (see Table 2). These assignments are further confirmed by theoretical analysis (Fig. 6b–d).

Fig. 1 ORTEP view with non-hydrogen atom labeling and thermal ellipsoids at 30% probability level.

Table 2 Average bond lengths in the [Ni(dmit)₂] moiety of 1

	Ni–S (Å)	S–C (Å)	C=C (Å)	Ref.
[Ni(dmit)2]^2−	2.195(9)	1.733(8)	1.352(3)	23a
[Ni(dmit)2]^1−	2.166(10)	1.715(10)	1.360(6)	23b
[Ni(dmit)2]^0.5−	2.163(3)	1.701(4)	1.382(7)	23c
[Ni(dmit)2]^0	2.143(3)	1.698(5)	1.393(10)	23d
Ni(1) moiety	2.165(13)	1.700(5)	1.375(6)	This work
Ni(2) moiety	2.154(13)	1.697(5)	1.384(6)	This work
Ni(3) moiety	2.152(13)	1.692(4)	1.384(6)	This work

---

As displayed in Fig. 2a, the C₃-APy⁺ cations form a monolayer molecular sheet, and the sheet is parallel to the crystallographic ab-plane. The pyridyl rings of the cations are parallel to each other owing to the restraint of inversion centers between them. The mixed-valence [Ni(dmit)₂] species are stacked in a …ABCABC… manner of …Ni(2)Ni(1)Ni(3)… along the a-axis direction. The neighboring [Ni(dmit)₂] stacking columns are aligned along the b-axis direction to form a mixed-valence [Ni(dmit)₂] layer. The cation layer (C) and mixed-valence [Ni(dmit)₂] layers (M) are parallel to each other, and alternate in a fashion of …MCMC… along c-axis. Within a [Ni(dmit)₂] stack, the [Ni(dmit)₂] species show face-to-face overlapping in the longitudinal offset mode; the dihedral angle and distance between the mean–molecular planes of [Ni(dmit)₂] species, defined by four coordinated S atoms, is 1.30° and 3.517 Å between Ni(1)- and Ni(2)-[Ni(dmit)₂] species, 1.31° and 3.482 Å between Ni(1)- and Ni(3)-[Ni(dmit)₂] species, and 0.82° and 3.625 Å between Ni(2)- and Ni(3)-[Ni(dmit)₂] species. As shown in Fig. 2b, there exist short interatomic contacts, which are less than the sum of van der Waals radii of two S atoms (3.7 Å), within and between the [Ni(dmit)₂] stacks. These typical interatomic distances are listed in Table 3, and they indicate the existence of strong intermolecular interactions in the [Ni(dmit)₂] layer, but weak intermolecular interactions between the [Ni(dmit)₂] layers.

Fig. 2 (a) Alternating layered packing structure viewed along the crystallographic a-axis direction in 1 (b) a monolayer of mixed-valence [Ni(dmit)₂] species and the different mixed-valence [Ni(dmit)₂] species are represented by different colors (Ni1: red; Ni2: blue and Ni3: cyan).

Table 3 Typical short interatomic distances, which are less than the sum of van der Waals radii of two S atoms (3.7 Å) in the crystal of 1^a

a Symmetric codes: #1 = 1 − x, 1 − y, 1 − z; #2 = 1 − x, 1 − y, 1 − z; #3 = 1 − x, 1 − y, 1 − z; #4 = −x, 1 − y, 1 − z; #5 = 1 − x, −y, 1 − z.
S(23)⋯S(14)#1	3.582Å	S(3)⋯S(19)#1	3.597Å
S(8)⋯S(14)#2	3.597Å	S(16)⋯S(30)#3	3.540Å
S(1)⋯S(11)#2	3.504Å	S(27)⋯S(27)#4	3.536Å
S(1)⋯S(1)#2	3.572Å	S(17)⋯S(17)#5	3.313Å

---

3.2. Magnetic property

The magnetic susceptibility (χ_m) of 1 as a function of temperature is displayed in Fig. 3, in which χ_m represents the molar magnetic susceptibility with one [Ni(dmit)₂]¹⁻ species per formula unit. A magnetic phase transition appears around 77 K with a ca. 6 K hysteresis loop (insets in Fig. 3) in the plot of χ_m versus T; above and below the transition temperature, molecular crystal 1 shows Curie–Weiss magnetic behaviors. As a result, we first used the Curie–Weiss law to estimate the average magnetic coupling interaction in crystal 1. The experimental magnetic susceptibility data in high-temperature (HT) and low-temperature (LT) phases were fitted using eqn (1),where the χ₀ term represents the sum of the diamagnetic and possible Van Vleck paramagnetic susceptibilities. The diamagnetism contributes from the atom cores and the Van Vleck paramagnetism is related to the coupling of the ground and excited states through a magnetic field.²⁴ The best fit gave rise to the parameters C = 0.360(3) emu K mol⁻¹, θ = −16.0(9) K and χ₀ = −6.2(1) × 10⁻⁴ emu mol⁻¹ for the magnetic susceptibility data in the range of 96–400 K in HT phase; C = 0.396(25) emu K mol⁻¹, θ = −9.0(5) K and χ₀ = −4.2(5) × 10⁻³ emu mol⁻¹ for the magnetic susceptibility data in the range of 1.8–65 K in LT phase. Although the fitted C values in both the HT and LT phases are close to the spin-only value (0.375 emu K mol⁻¹ when g = 2.0) for an S = 1/2 magnetic system, the yielded diamagnetism χ₀ value in the LT phase seems too large. In addition to this, the antiferromagnetic (AFM) interaction is approximately proportional to the square of the overlap integral of magnetic orbitals between the coupling magnetic centers,²⁵ and lattice shrinking in the LT phase should lead to the overlap of magnetic orbitals being strengthened, namely the AFM interaction should be stronger in the LT phase than in the HT phase. Thus, it is unreasonable that the Weiss constant, representing the AFM interaction in crystal 1, in the LT phase is less than that in the HT phase.

Fig. 3 Temperature dependent χ_m of 1. (a) The squares are the experimental data, and the blue and red lines represent the fits using the Curie–Weiss equation. The inset shows the thermal hysteresis loop. (b) The squares are the experimental data and the red line represents the fits for the magnetic susceptibility data in the HT phase using 1-D uniform chain model.

For [Ni(dmit)₂]-based molecular crystals, the electronic and magnetic interactions between neighboring [Ni(dmit)₂] species are transmittable through long nonbonded interatomic contacts owing to the conjugated system in [Ni(dmit)₂] extending across the entire molecule. The crystal structural analysis demonstrated the existence of three types of contacts between neighboring [Ni(dmit)₂] species, (1) the face-to-face stacking within a [Ni(dmit)₂] column along a-axis direction, where the π-orbitals overlap is available, (2) the lateral-to-lateral S⋯S contacts between the adjacent [Ni(dmit)₂] columns along the b-axis direction and (3) the head-to-tail S⋯S contacts between the adjacent [Ni(dmit)₂] columns along the long molecular axis direction of [Ni(dmit)₂] (see Fig. 2). Obviously, the magnetic coupling is stronger via the π-orbitals overlap than those via the lateral-to-lateral or head-to-tail S⋯S contacts. From this point, the analysis can be simplified as a 1-D spin system for the magnetic behavior of the molecular crystal 1 in the HT phase.

In the HT phase, the neighboring S = 1/2 [Ni(dmit)₂]¹⁻ anions are separated by two neutral [Ni(dmit)₂] species (one Ni(2)–[Ni(dmit)₂] and one Ni(3)–[Ni(dmit)₂] species) within a stack (ref. Fig. 2b). Thus, a columnar [Ni(dmit)₂] stack can be thought as an S = 1/2 Heisenberg uniform linear chain. The magnetic susceptibility as a function of temperature for an S = 1/2 alternating chain system, derived from the spin Hamiltonian eqn (2), is expressed as eqn (3),

where J or αJ is the exchange integral between a spin and its right or left neighbor, respectively.

In eqn (3), A–F represent a set of parameters that depend on the alternating constant (α), X = |J|/k_BT (J < 0 and 0 ≤ α ≤ 1). Extremely, when α = 0 the alternating linear chain model is simplified to the dimer model with pairwise interactions and when α = 1 the alternating linear chain model is simplified to the uniform linear-chain model.²⁶ If the diamagnetic and possible Van Vleck paramagnetic susceptibilities are further considered, the experimental molar magnetic susceptibility is given in eqn (4) form,

χ_m = χ_m(chain) + χ₀ (4)

The best fit was performed for the magnetic susceptibility data in the range of 96–400 K using eqn (3) together with eqn (4) to give |J|/k_B = 17.5(2) K with g = 2.0 and χ₀ = −6.2(1) × 10⁻⁴ emu mol⁻¹ fixed or |J|/k_B = 13.2(2) K, g = 1.96(3) with χ₀ = −6.2(1) × 10⁻⁴ emu mol⁻¹ fixed. The theoretically reproduced χ_m–T plot is displayed in Fig. 3b using the fitted parameters |J|/k_B = 17.5(2) K, g = 2.0 and χ₀ = −6.2(1) × 10⁻⁴ emu mol⁻¹. The small |J| value indicated the existence of weakly AFM coupling within a [Ni(dmit)₂] stack in the HT phase.

It is no possible to further analyze the exact magnetic coupling nature in the LT phase owing to lack of accessible crystal structure data.

3.3. Temperature dependent IR spectra

The temperature-dependent IR spectra of 1 in 6–293 K are shown in Fig. S2,† indicating that most of the vibrational bands show no sizable change in both intensity and the central position. The main alterations as the temperature changes concern the vibrational bands in four spectroscopy regions, which are respectively displayed in Fig. 4a–d. The band centered around 1319 cm⁻¹ and 1358 cm⁻¹ at 293 K are assigned to νC=C of the dmit²⁻ ligands; two bands are related to the neutral (1319 cm⁻¹) and negative (1358 cm⁻¹) [Ni(dmit)₂] species, respectively.²⁷ As the temperature drops, the band centered at 1319 cm⁻¹ shifts towards a higher frequency, and a reflected point appeared at ca. 80 K if we connect sequentially the band maxima (Fig. 4b). This reflected point temperature is close to that at which the magnetic transition occurs. In addition to this, the shoulder around 1358 cm⁻¹ becomes a clear peak (see Fig. 4a). The bands due to the νC=S mode of the dmit²⁻ ligands localize at 1063 cm⁻¹ and 1076 cm⁻¹ for [Ni(dmit)₂]⁻ at ambient temperature. As demonstrated in Fig. 4c, the stretching vibration bands shift to a higher frequency upon cooling.

Fig. 4 Variable-temperature IR spectra of 1 in the temperature range 6–293 K, showing changes in the shape and position of characteristic bands: (a) and (b) the νC=C bands of the dmit²⁻ ligands, (c) the νC=S bands and (d) the νNi–S bands.

The νNi-S bands shift to high frequency in the [Ni(dithiolene)₂]ⁿ⁻ (n = 0–2) with the n value decreasing because the LUMO in the neutral [Ni(dithiolene)₂] species is antibonding with respect to the σ(Ni–S) bonds. Nakamoto et al. theoretically analyzed the IR spectra of [Ni(dithiolene)₂]ⁿ⁻, and as indicated in Table 4, the change trend of νNi-S band frequency with n value is in good agreement with the observations.²⁸ As shown in Fig. 4d, two weak bands centered at 485 and 455 cm⁻¹ are observed at 293 K, which probably originated from the Ni–S stretching vibrations in [Ni(dmit)₂] and [Ni(dmit)₂]⁻ species, respectively. The band around 485 cm⁻¹ becomes weak while the band around 455 cm⁻¹ develops into the visible and shifts towards a high frequency upon cooling. Moreover, a reflected point also appeared at ca. 80 K if we connect sequentially the band maxima in the low frequency side.

Table 4 Comparison of νNi–S band frequencies (cm⁻¹) in bis(dithiolato)nickel²⁸

Species	B2u	B3u	Species	B2u	B3u
[Ni(S2C2H2)2]^0	428	398	[Ni(S2C2(CF3)2)2]^0	465	425
[Ni(S2C2H2)2]^−	411	385	[Ni(S2C2(CF3)2)2]^−	449	415
[Ni(S2C2(C6H5)2)2]^0	475, 408	454	[Ni(S2C2(CF3)2)2]^2−	422	394
[Ni(S2C2(C6H5)2)2]^−	465, 406	428	[Ni(S2C2(CN)2)2]^−	468, 396	365
[Ni(S2C2(C6H5)2)2]^2−	450, 401	418	[Ni(S2C2(CN)2)2]^2−	457, 365	357

---

Vibrational band shifts indicate changes in the electronic state of the molecule because the vibration frequency is directly related to the force constant of the bond. As mentioned above, the changes of bands from the intramolecular vibration modes (νC=C, νC=S and νNi–S) within the [Ni(dmit)₂] and [Ni(dmit)₂]⁻ species indicate that electron-vibration couplings within the mixed-valence [Ni(dmit)₂] species cooperate with the magnetic phase transition. In fact, it was known that the intramolecular vibrations played an important role in the occurrence of the superconductivity of organic materials three decades ago, for which the molecular superconductivity was caused by electron–phonon interactions.^29,30 Upon cooling, the νC=C and νC=S bands shift to high frequency, representing that the C=C and C=S bonds are stronger in the LT phase than those in the HT phase; the bands from νNi–S of [Ni(dmit)₂] and [Ni(dmit)₂]⁻ species become invisible, and meanwhile a new band appears in their middle spectroscopy region, disclosing that the Ni–S bonds in [Ni(dmit)₂]⁻ species are stronger in the LT phase than those in the HT phase. On the contrary, the Ni–S bonds in [Ni(dmit)₂] species are weaker in the LT phase than those in the HT phase. The above observation suggests that the electron localized in the [Ni(dmit)₂]⁻ species in the HT phase probably delocalizes between the neutral and negative [Ni(dmit)₂] species within a [Ni(dmit)₂]⁰/[Ni(dmit)₂]¹⁻/[Ni(dmit)₂]⁰ unit in the LT phase. This results from the lattice shrinking in the LT phase, leading to more efficient overlap of the frontier π-orbitals between [Ni(dmit)₂] and [Ni(dmit)₂]⁻ species. In a previous study, we investigated the temperature dependent single crystal EPR spectra of 1-D molecular alloy [NO₂–BzPy][Ni_xAu_1−x(mnt)₂] systems, where NO₂–BzPy⁺ is 4-nitrobenzyl-1-pyridinium, and we found the phenomenon that the electron delocalizes between the neighboring [Ni(mnt)₂]⁻ and [Au(mnt)₂]⁻ species within the [Ni_xAu_1−x(mnt)₂]⁻ columnar stack via the frontier π-orbitals overlap.³¹ Most recently, Stoddart and coworkers also discovered the electron delocalization in a rigid cofacial naphthalene-1, 8:4, 5-bis(dicarboximide) π–dimer via π–π stacking interaction.³²

3.4. Conductivity

The thin plate-shaped single crystal was selected and the plate area is about 0.3 × 0.4 mm², and as shown in Fig. 5a, the four-probe method was used for the measurement of electrical conductivity. The resistance as a function of temperature is displayed in Fig. 5b. Upon cooling, the resistance increases from ca. 2 kohm at 295 K to ca. 10 Mohms at 94 K, and then suddenly drops. The resistance's abrupt decrease below 94 K is due to that too high of a resistance of the crystal results in the current being below the instrument limitation. In order to investigate the conducting nature of molecular crystal 1 below 94 K, two 1 Mohms carbon film resistors were respectively connected in parallel between probes 1 and 2 as well as between probes 3 and 4 of the crystal (ref. Fig. 5a), and the corresponding equivalent circuit is illustrated in Fig. 5c. Extremely, in this case, the total resistance is zero when the crystal is a superconductor and the total resistance is 1 Mohms when the crystal is an insulator. The measured resistances against temperature are plotted in Fig. 5d. It is worth noting that there is absence of the anomaly around 77 K where a magnetic phase transition undergoes. On the other hand, we further fitted the temperature-dependent resistance in the range of 127–295 K to estimate the activation energy, E_a, using the Arrhenius equation below,

Fig. 5 (a) Image of a thin plate-shaped single crystal with four-probes used in the variable temperature conductivity measurements. (b) Temperature dependence of the resistance of 1, where the black line represents the measured data, and the red line is fitted plot using Arrhenius equation at 127–295 K. (c and d) Illustration for the equivalent circuit when two 1 Mohms carbon film resistors were respectively connected in parallel between probes 1 and 2 as well as between probes 3 and 4 of the crystal and the corresponding plot of resistance against temperature.

In eqn (5), R₀, E_a and k_B correspond to the pre-exponential factor, activation energy and Boltzmann constant, respectively. The best fit gave E_a = 1147.5(3) K (=0.099 eV), and the theoretically reproduced plot in the temperature range of 127–295 K is shown in Fig. 5b.

3.5. Electronic band structure

The calculated bands near to the Fermi level along the high symmetry directions of the Brillouin zone are displayed in Fig. 6a. By contrast, a small dispersion is observed along the direction from G to Z in k-space, which is approximately parallel to the c-axis direction; this is indicative of weak orbital interactions between the mixed-valence [Ni(dmit)₂] layers and a 2-D electronic nature of this molecular crystal. The relatively sharp dispersion appears along the direction from G to Y and from G to X in k-space, indicating the existence of stronger orbital interactions within the mixed-valence [Ni(dmit)₂] layer. These results are in agreement with the crystal structure analysis that (1) the neighboring mixed-valence [Ni(dmit)₂] layers are separated by the cation layer, and there exist shorter head-to-tail S⋯S contacts between inter-layers; (2) there exist strongly π–π stacking interactions within a mixed-valence [Ni(dmit)₂] column and larger amounts of shorter lateral-to-lateral S⋯S contacts between the inter-columns in a mixed-valence [Ni(dmit)₂] layer. In addition, as displayed in Fig. 6b–d, the Partial DOS (PDOS) of the species with Ni1 is different from other two [Ni(dmit)₂] species, whereas the PDOS of the species with Ni2 and Ni3 is quite analogous to each other near the Fermi level, representing that the electric state of Ni1 species is extremely different from those of Ni2 and Ni3 (the differences can be found in p-based band in Fig. 6b–d), and they further confirm that the [Ni(dmit)₂] species with Ni1 is monoanion and other two are neutral species. This finding is in good agreement with the analysis of crystal structure.

Fig. 6 (a) Several highest occupied bands and the lowest unoccupied bands in the HT phase of 1, where the Fermi levels are shown by dashed lines, k-points: G = (0, 0, 0), X = (0.5, 0, 0), Y = (0, 0.5, 0), Z = (0, 0, 0.5) and Q = (0, 0.5, 0.5) and (b–d) PDOS of three crystallographically different [Ni(dmit)₂] species.

It was noted that the calculated energy gap between the highest occupied and the lowest unoccupied bands is 0.003 eV. However, this molecular crystal shows a semi-conductive behavior in the 4–295 K region. For this discrepancy between the theoretical and experimental results is understandable because (i) it is well-known that the DFT method does not accurately describe the eigenvalues of the electronic states, causing the quantitative underestimation of band gaps;³³ (ii) for a system with a partially filled band, the normal metals and the magnetic insulator possess similar partially filled bands when the calculated width of the partially filled band is narrow. In a so-called magnetic insulator, the effective on-site Coulomb repulsion energy (U_eff) is much larger than the hopping energy (t) that an electron transfers between two sites (the frontier orbitals between the neighboring molecules). However, the magnetic insulating state may become more stable than the metallic state owing to electron–electron repulsion.³⁴ Both the analyses of crystal structural and PDOS revealed that the electron localizes in the [Ni(dmit)₂]⁻ species in the HT phase, thus molecular crystal 1 should be a magnetic insulator in the HT phase.

4. Conclusion

In summary, we achieved a new mixed-valence molecular crystal [C₃-Apy][Ni(dmit)₂]3 using a facile solution process. The mixed-valence molecular crystal shows a hysteretic magnetic transition with ca. 6 K thermal hysteresis loop. The crystal structure and electronic band structure analyses revealed that the electron is localized in the [Ni(dmit)₂]⁻ monoanion in the HT phase, and probably delocalized within …[Ni(dmit)₂]⁰/[Ni(dmit)₂]¹⁻/[Ni(dmit)₂]⁰… in the LT phase owing to lattice shrinking, which leads to the frontier π-orbital overlap being strengthened within a [Ni(dmit)₂]⁰/[Ni(dmit)₂]¹⁻/[Ni(dmit)₂]⁰ unit. The study on the variable temperature conductivity disclosed that this mixed-valence molecular crystal is a semiconductor.

Acknowledgements

Authors thank the Priority Academic Program Development of Jiangsu Higher Education Institutions and the National Nature Science Foundation of China (Grant nos: 91122011 and 21271103) for financial support.

References

1. (a) R. Kato, Chem. Rev., 2004, 104, 5319; (b) P. Cassoux, L. Valade, H. Kobayashi, A. Kobayashi, R. A. Clark and A. E. Underhill, Coord. Chem. Rev., 1991, 110, 115; (c) T. Akutagawa and T. Nakamura, Coord. Chem. Rev., 2000, 198, 297; (d) A. Kobayashi, E. Fusjiwara and H. Kobayashi, Chem. Rev., 2004, 104, 5243.
2. (a) T. Nakamura, T. Akutagawa, K. Honda, A. E. Underhill, A. T. Coomber and R. H. Friend, Nature, 1998, 394, 159; (b) T. Akutagawa, H. Koshinaka, D. Sato, S. Takeda, S.-I. Noro, H. Takahashi, R. Kumai, Y. Tokura and T. Nakamura, Nat. Mater., 2009, 8, 342.
3. (a) T. Naito, T. Karasudani, S. Mori, K. Ohara, K. Konishi, T. Takano, Y. Takahashi, T. Inabe, S. Nishihara and K. Inoue, J. Am. Chem. Soc., 2012, 134, 18656; (b) T. Ishikawa, N. Fukazawa, Y. Matsubara, R. Nakajima, K. Onda, Y. Okimoto, S. Koshihara, M. Lorenc, E. Collet, M. Tamura and R. Kato, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 80, 115108; (c) N. Fukazawa, T. Tanaka, T. Ishikawa, Y. Okimoto, S. Koshihara, T. Yamamoto, M. Tamura, R. Kato and K. Onda, J. Phys. Chem. C, 2013, 117, 13187.
4. T. Kusamoto, H. M. Yamamoto, N. Tajima, Y. Oshima, S. Yamashita and R. Kato, Inorg. Chem., 2013, 52, 4759.
5. N. Takamatsu, T. Akutagawa, T. Hasegawa, T. Nakamura, T. Inabe, W. Fujita and K. Awaga, Inorg. Chem., 2000, 39, 870.
6. (a) T. Akutagawa, D. Sato, Q. Ye, T. Endo, S. I. Noro, S. Takeda and T. Nakamura, Dalton Trans., 2010, 39, 8219; (b) H. Imai, T. Otsuka, T. Naito, K. Awaga and T. Inabe, J. Am. Chem. Soc., 1999, 121, 8098.
7. Y. Kosaka, H. M. Yamamoto, A. Nakao, M. Tamura and R. Kato, J. Am. Chem. Soc., 2007, 129, 3054.
8. D. J. Reefman, J. P. Cornelissen, J. G. Haasnoot, R. A. G. de Graaf and J. Reedijk, Inorg. Chem., 1990, 29, 3933.
9. (a) K. Takahashi, H. B. Cui, Y. Okano, H. Kobayashi, Y. Einaga and O. Sato, Inorg. Chem., 2006, 45, 5739; (b) T. Akutagawa, D. Sato, H. Koshinaka, M. Aonuma, S. I. Noro, S. Takeda and T. Nakamura, Inorg. Chem., 2008, 47, 5951; (c) T. Akutagawa, K. Shitagami, S. Nishihara, S. Takeda, T. Hasegawa, T. Nakamura, Y. Hosokoshi, K. Inoue, S. Ikeuchi, Y. Miyazaki and K. Saito, J. Am. Chem. Soc., 2005, 127, 4397.
10. (a) A. E. Pullen, H. L. Lui, D. B. Tanner, K. A. Abboud and J. R. Reynolds, J. Mater. Chem., 1997, 7, 377; (b) T. Nakamura, A. E. Underhill, A. T. Coomber, R. H. Friend, H. Tajima, A. Kobayashi and H. Kobayashi, Synth. Met., 1995, 70, 1061; (c) Y. S. J. Veldhuizen, J. G. Haasnoot and J. Reedijk, Synth. Met., 1997, 86, 1827; (d) Y. S. J. Veldhuizen, W. J. J. Smeets, N. Veldman, A. L. Spek, C. Faulmann, P. Auban-Senzier, D. Jérome, P. M. Paulus, J. G. Haasnoot and J. Reedijk, Inorg. Chem., 1997, 36, 4930.
11. (a) T. Akutagawa, T. Nakamura, T. Inabe and A. E. Underhill, Thin Solid Films, 1998, 331, 264; (b) J. Lieffrig, O. Jeannin, P. Auban-Senzier and M. Fourmigué, Inorg. Chem., 2012, 51, 7144.
12. (a) A. Miyazaki, A. Izuoka and T. Sugawara, Bull. Chem. Soc. Jpn., 1993, 66, 2832; (b) M.-L. Doublet, E. Canadell, B. Garreau, J.-P. Legros, L. Brossard, P. Cassoux and J.-P. Pouget, J. Phys.: Condens. Matter, 1995, 7, 4673.
13. (a) A. E. Pullen, H.-L. Liu, D. B. Tanner, K. A. Abboud and J. R. Reynolds, J. Mater. Chem., 1997, 7, 377; (b) Y. S. J. Veldhuizen, J. G. Haasnoot and J. Reedijk, Synth. Met., 1997, 86, 1827.
14. (a) B. Pomaréde, B. Garreau, I. Malfant, L. Valade, P. Cassoux, J.-P. Legros, A. Audouard, L. Brossard, J.-P. Ulmet, M.-L. Doublet and E. Canadell, Inorg. Chem., 1994, 33, 3401; (b) H. Kim, A. Kobayashi, Y. Sasaki, R. Kato and H. Kobayashi, Chem. Lett., 1987, 1799.
15. H. Yang, J. L. Liu, L. C. Zhou and X. M. Ren, Inorg. Chem. Front., 2014, 1, 426.
16. G. Steimecke, H. J. Sieler, R. Kirmse and E. Hoyer, Phosphorus Sulfur Relat. Elem., 1979, 7, 49.
17. H. B. Duan, X. M. Ren, L. J. Shen, W. Q. Jin, Q. J. Meng, Z. F. Tian and S. M. Zhou, Dalton Trans., 2011, 40, 3622.
18. G. M. Sheldrick, SADABS, Program for Bruker Area Detector Absorption Correction, University of Göttingen, Göttingen, Germany, 1997.
19. G. M. Sheldrick, SHELXS-97, Program for Crystal Structure Solution, University of Göttingen, Göttingen, Germany, 1997.
20. (a) M. D. Segall, P. J. D. Lindan, M. J. Probert, C. J. Pickard, P. J. Hasnip, S. J. Clark and M. C. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717; (b) V. Milman, B. Winkler, J. A. White, C. J. Pickard, M. C. Payne, E. V. Akhmatskaya and R. H. Nobes, Int. J. Quantum Chem., 2000, 77, 895.
21. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
22. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Condens. Matter Mater. Phys., 1976, 13, 5188.
23. (a) H. K. Fun, K. Sivakumar, J. L. Zuo, T. M. Yao and X. Z. You, Acta Crystallogr., Sect. C: Cryst. Struct. Commun., 1996, 52, 312; (b) M. J. J. Mulder, H. Kooijman, A. L. Spek, J. G. Haasnoot and J. Reedijk, J. Chem. Crystallogr., 2002, 32, 347; (c) J. P. Cornelissen, R. L. Loux, J. Jansen, J. G. Haasnoot, J. Reedijk, E. Horn, A. L. Speck, B. Pomarede, J. P. Legros and D. Reefman, J. Chem. Soc., Dalton Trans., 1992, 2911; (d) N. Kushch, C. Faulmann, P. Cassoux, L. Valade, I. Malfant, J. P. Legros, C. Bowlas, A. Errami, A. Kobayashi and H. Kobayashi, Mol. Cryst. Liq. Cryst. Sci. Technol., Sect. A, 1996, 284, 247.
24. J. H. van Vleck, The Theory of Electric and Magnetic Susceptibilities, Oxford, London, 1932.
25. M.-H. Whangbo, H.-J. Koo and D. Dai, J. Solid State Chem., 2003, 176, 417.
26. (a) J. Bonner and M. E. Fisher, Phys. Rev. A, 1976, 235, 640; (b) J. W. Hall, W. E. Marsh, R. R. Weller and W. E. Hatfield, Inorg. Chem., 1981, 20, 1033; (c) D. A. Tucker, P. S. White, K. L. Trojan, M. L. Kirk and W. E. Hatfield, Inorg. Chem., 1991, 30, 823.
27. (a) K. I. Pokhodnya, C. Faulmann, I. Malfant, R. Andreu-Solano, P. Cassoux, A. Mlayah, D. Smirnov and J. Leotin, Synth. Met., 1999, 103, 2016; (b) B. Cai, J. L. Liu, X. L. Sheng and X. M. Ren, Inorg. Chem. Commun., 2011, 14, 1971; (c) M. K. Johnson, Dithiolene chemistry: synthesis, properties, and applications, Prog. Inorg. Chem., John Wiley&Sons, USA, 2004, ch. 4, vol. 52.
28. C. W. Schläpfer and K. Nakamoto, Inorg. Chem., 1975, 14, 1338.
29. E. Demiralp, S. Dasgupta and W. A. Goddard, J. Am. Chem. Soc., 1995, 117, 8154.
30. T. Kato and T. Yamabe, J. Chem. Phys., 2003, 119, 5680.
31. X. M. Ren, Y. X. Sui, G. X. Liu and J. L. Xie, J. Phys. Chem. A, 2008, 112, 8009.
32. Y. Wu, M. Frasconi, D. M. Gardner, P. R. McGonigal, S. T. Schneebeli, M. R. Wasielewski and J. F. Stoddart, Angew. Chem., Int. Ed., 2014, 126, 9630.
33. (a) R. W. Godby, M. Schluther and L. Sham, J. Phys. Rev. B: Condens. Matter Mater. Phys., 1987, 36, 6497; (b) C. M. I. Okoye, J. Phys.: Condens. Matter, 2003, 15, 5945; (c) R. Terki, G. Bertrand and H. Aourag, Microelectron. Eng., 2005, 81, 514.
34. (a) M. H. Whangbo, J. M. Williams, P. C. W. Leung, M. A. Beno, T. J. Emge, H. H. Wang, K. D. Carlson and G. W. Crabtree, J. Am. Chem. Soc., 1985, 107, 581; (b) M. H. Whangbo and R. Hoffmann, J. Am. Chem. Soc., 1978, 100, 6093.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 1027558. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4ra13354d

---