Structural and Magnetic Investigations for the Doping Effect of

Research Institute for Electronic Science, Hokkaido University, Sapporo 060-0812, Japan, CREST, Japan Science and Technology Corporation (JST), ...
0 downloads 0 Views 384KB Size
Download PDF
J. Phys. Chem. B 2006, 110, 7671-7677

7671

Structural and Magnetic Investigations for the Doping Effect of Nonmagnetic Impurity on the Spin-Peierls-like Transition in a Quasi-One-Dimensional Magnet: 1-(4′-Nitrobenzyl)pyridinium Bis(maleonitriledithiolato)nickelate X. M. Ren,*,† T. Akutagawa,† S. Noro,† S. Nishihara,‡ T. Nakamura,*,† Y. Yoshida,⊥ and K. Inoue⊥ Research Institute for Electronic Science, Hokkaido UniVersity, Sapporo 060-0812, Japan, CREST, Japan Science and Technology Corporation (JST), Kawaguchi 332-0012, Japan, Department of Physical Science, Graduate School of Science, Osaka Prefecture UniVersity, Sakai, Osaka 599-8531, Japan, and Department of Chemistry, Faculty of Science, Hiroshima UniVersity, 1-3-1 Kagamiyama, Higashi Hiroshima, Hiroshima 739-8526, Japan ReceiVed: October 2, 2005; In Final Form: February 7, 2006

A nonmagnetic compound, [NO2BzPy][Au(mnt)2] (NO2BzPy+ ) 1-(4′-nitrobenzyl)pyridinium; mnt2- ) maleonitriledithiolate), was synthesized and characterized structurally, which is isostructural with [NO2BzPy][Ni(mnt)2] that is a quasi-one-dimensional magnet and possesses a spin-Peierls-like transition with J ) 192 K in the gapless state and spin energy gap ) 738 K in the dimerization state, respectively. Further, ten nonmagnetic impurity doped compounds with a formula [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) were prepared and investigated by crystal structural determinations and magnetic susceptibility measurements. The nonmagnetic doping causes the suppression of the spin transition with an average rate of 221(12) K/percentage of dopant concentration. From the plots of χm-T, the transition collapse (the characteristic of the transition is the sudden drop of χm upon cooling, and the disappearance of this characteristic is considered as the criterion for the transition collapse) is estimated at around x > 0.27. In heavier doped system x ) 0.49, the spin gap vanishes and a gapless phase is achieved again.

Introduction

CHART 1

Impurity doping of a low-dimensional magnetic system often results in nonintuitive ground states and transition behavior differing from that of the pure or parent material. For instance, an intriguing low-temperature phase, where an antiferromagnetism coexists with a spin-Peierls dimerization, was observed in the impurity-doped systems, such as an inorganic spin-Peierls compound CuGeO31-3 and the interacting dimer compound TlCuCl3.4 A field-controlled inhomogeneous state occurs in the impurity doped system of CuGeO3.5 Many investigations for doping magnetic impurity with different spins have been done theoretically and experimentally for the low-dimensional inorganic magnetic systems since the first inorganic spin-Peierls compound CuGeO3 was discovered.1-6 In contrast, the examples of an impurity doping on the low-dimensional molecular magnetic system are very rare7-9 even if the molecular spinPeierls transition compounds were known much earlier than the inorganic one.10 One of the crucial reasons for the situation mentioned above is that the molecular structure is usually so complicated that it is not easy to find a suitable dopant that will match with the parent material in both molecular structure and valence. In our previous study, we observed a peculiar spinPeierls-like transition in a quasi-one-dimensional compound, 1-(4′-nitrobenzyl)pyridinium bis(maleonitriledithiolato)nickelate * Address correspondence to these authors. Phone: +81 11-706-2849. Fax: +81 11-706-4972. E-mail: [email protected] and tnaka@ imd.es.hokudai.ac.jp. † Hokkaido University and CREST. ‡ Osaka Prefecture University. ⊥ Hiroshima University.

illustrated in Chart 1(abbreviated as [NO2BzPy][Ni(mnt)2]).11 This compound possesses columnar stacks along the c-axis with cations and anions segregated. The anionic stack is regular, and it can be considered as a uniform magnetic chain with S ) 1/2 since the [Ni(mnt)2]- anion bears a half spin (S ) 1/2). The magnetic exchange nature is antiferromagnetic above the transition temperature, while below that a spin gap is opened with the spin chain dimerization.11 Most recently, the nonmagnetic compound [NO2BzPy][Au(mnt)2] was prepared and characterized structurally, which is isostructural with [NO2BzPy][Ni(mnt)2]. It is important to note that the molecular structure of [Au(mnt)2]- is rather similar to that of [Ni(mnt)2]- and the valence of [Au(mnt)2]- matches well with that of [Ni(mnt)2]-. Consequently, [NO2BzPy][Au(mnt)2] is expected to be one of the best candidates for studying the doping effect on the spinPeierls-like transition in the spin system of [NO2BzPy][Ni(mnt)2], because (1) [Au(mnt)2]- partially substituting [Ni(mnt)2]in an anionic stack does not change the crystal structure of the

10.1021/jp0556106 CCC: $30.25 © 2006 American Chemical Society Published on Web 03/30/2006

7672 J. Phys. Chem. B, Vol. 110, No. 15, 2006

Ren et al.

TABLE 1: Elemental Analyses (C, H and N), Molar Ratio of [Au(mnt)2]- to [Ni(mnt)2]- in the Starting Materials, and the Occupation Factor of Au (Ni) in the Final Crystal Structural Refinement for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) calcd

found

x

C

H

N

C

H

N

Au:Nia

Au:Nib

0.01 0.04 0.11 0.21 0.27 0.35 0.41 0.49 0.57 0.73

43.2 43.0 42.2 41.2 40.6 39.9 39.3 38.6 37.9 36.7

2.00 1.98 1.95 1.90 1.88 1.84 1.81 1.78 1.75 1.69

15.1 15.0 14.8 14.4 14.2 14.0 13.8 13.5 13.3 12.8

43.1 42.8 42.3 41.1 40.5 40.1 39.4 39.3 37.9 36.6

2.13 2.09 2.09 2.03 2.00 1.97 1.95 1.94 1.88 1.75

15.0 14.9 14.8 14.2 14.3 14.2 13.9 13.6 13.3 12.7

0.02:0.98 0.06:0.94 0.11:0.89 0.25:0.75 0.30:0.70 0.34:0.66 0.42:0.58 0.50:0.50 0.59:0.41 0.75:0.25

0.01:0.99 0.04:0.96 0.11:0.89 0.21:0.79 0.27:0.73 0.35:0.65 0.41:0.59 0.49:0.51 0.57:0.43 0.73:0.27

a The molar ratio in the starting materials. b The occupation factor ratio in the final crystal structural refinement.

TABLE 2: Crystallographic Data for [NO2BzPy][Au(mnt)2] at 273 K molecular formula molecular mass space group a/Å b/Å c/Å β/deg V/Å3, Z µ/mm-1 λ/Å F/g cm-3 R1a wR2 a

C20H11AuN6O2S4 692.55 P2(1)/c 12.237(2) 26.762(5) 7.3168(15) 103.60(3) 2329.0(8), 4 6.706 0.71073 1.975 0.0397 0.0878

R1 )∑(||Fo| - |Fc||)/∑|Fo|, wR2 ) [∑w(|Fo|2 - |Fc|2)2/∑w(|Fo|2)2]1/2.

[Ni(mnt)2]- compound overall and (2) the magnetic property for a doped system is still determined by [Ni(mnt)2]- anions. In this contribution, we present the crystal structures and magnetic properties of a series of compounds with the formula [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1 with the molar fraction of Au in a doped system). Experimental Section Preparation of Compounds. [NO2BzPy][Ni(mnt)2] was prepared according to published procedures.11 [NO2BzPy][Au(mnt)2]. The procedure as described in the literature11 for preparing [NO2BzPy]2[Ni(mnt)2], but replacing NiCl2‚6H2O with KAuCl4‚2H2O, was followed to afford

[NO2BzPy][Au(mnt)2] (yield >80%). Single crystals suitable for X-ray structural analysis were obtained by diffusing diethyl ether into an acetonitrile solution of this compound. Elemental Anal. Calcd for C20H11N6O2S4Au: C, 34.7; H, 1.60; N, 12.1. Found: C, 34.6; H, 1.74; N, 12.1. [NO2BzPy][AuxNi1-x(mnt)2]. Each doped compound was prepared by direct mixing of [NO2BzPy][Ni(mnt)2] and [NO2BzPy][Au(mnt)2] in acetonitrile (as little as possible) in a rough molar ratio of (1 - x) to x at room temperature. The final mixture was put in a refrigerator at about -10 °C for 24 h, and needle- or block-shaped crystals were obtained which are suitable for X-ray structural analysis and other measurements. The molar ratio of [NO2BzPy][Ni(mnt)2] to [NO2BzPy][Au(mnt)2] in the starting materials, the results of elemental analyses for C, H, and N, and the occupation factor of Ni (Au) ion in the final crystal structural refinement for each doped system, [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73), are presented in Table 1. From these results, it is noted that (1) the composition of Ni (Au) in the final product has a small difference from that in the starting mixture, especially in the low concentration of dopant systems, and (2) the elemental analysis results match well with that from the crystal structural refinement for each doping system. Magnetic Susceptibility and Heat Capacity Measurements. Magnetic susceptibility measurement for every polycrystalline sample over the range of 2-350 K was carried out with a Quantum Design MPMS-XL superconducting quantum interference device (SQUID) magnetometer under 1.0 T. Heat capacity measurements were performed following the relaxation method with a Quantum Design PPMS system over the range of 2-200 K on the cooling process. A crystal was selected and attached to the sample platform with a small amount of grease. X-ray Structural Analyses. Crystallographic data except for x ) 0.27 at 10 K were collected with a Rigaku Raxis-Rapid diffractometer with Mo KR (λ ) 0.71073 Å) radiation from a graphite monochromator, and data for x ) 0.27 at 10 K were collected with a Bruker SMART-APEX three-circle diffractometer, equipped with a CCD area detector (graphite-monochromated Mo Ka radiation, λ ) 0.71073 Å). Structures were solved by direct methods by using the SHELXL-97 software package.12 The non-H atoms were refined anisotropically with the full-matrix least-squares method on F2. All H atoms were placed at calculated positions (C-H ) 0.930 Å for benzene or pyridine rings and 0.970 Å for methylene) and refined riding on the parent atoms with U(H) ) 1.2Ueq (bonded C atom). Details of the crystal parameters, data collection, and refinement for [NO2BzPy][Au(mnt)2], [NO2BzPy][Ni(mnt)2], as well as

TABLE 3: Crystallographic Data for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 at 10 K, x ) 0.35 and 0.49 at 103 K) molecular formula molecular mass space group a/Å b/Å c/Å R/deg β/deg γ/deg V/Å3, Z µ/mm-1 λ/Å F/g cm-3 R1a wR2 a

x ) 0.27

x ) 0.35

x ) 0.49

C20H11Au0.27Ni0.73N6O2S4 591.29 P-1 7.072(4) 11.974(6) 26.325(14) 89.419(10) 88.557(10) 78.817(9) 2186(2), 4 2.866 0.71073 1.797 0.1588 0.3869

C20H11Au0.35Ni0.65N6O2S4 602.29 P2(1)/c 12.045(2) 26.430(5) 7.1337(14) 90 101.88(3) 90 2222.4(7), 4 3.313 0.71073 1.803 0.1094 0.2220

C20H11Au0.49Ni0.51N6O2S4 621.79 P2(1)/c 12.061(2) 26.487(5) 7.1393(14) 90 102.01(3) 90 2230.7(8), 4 4.054 0.71073 1.850 0.0639 0.1327

R1 )∑(||Fo| - |Fc||)/∑|Fo|, wR2 ) [∑w(|Fo|2 - |Fc|2)2/∑w(|Fo|2)2]1/2.

[NO2BzPy][Au(mnt)2]

J. Phys. Chem. B, Vol. 110, No. 15, 2006 7673

Figure 2. (a) Distances of M‚‚‚M (d1 and d2) and interplane between neighboring anions (h1 and h2) and (b) the central-to-central distance between adjacent benzene rings (c1 and c2).

Figure 1. (a) ORTEP view (with displacement ellipsoids at the 50% possibility level and the H-atoms omitted for clarity) and (b) packing diagram of [NO2BzPy][Au(mnt)2].

each [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) at 273 K are summarized in Table 2, Table S1 (Supporting Information), and Table S2 (Supporting Information), respectively. The corresponding crystallographic data for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 at 10 K, x ) 0.35 and 0.49 at 103 K) are presented in Table 3. Temperature-Dependent Oscillation Photographs. The oscillation photographs at T ) 273 and 103 K for the doped system x ) 0.11, 0.27, 0.35, and 0.41 were taken by Rigaku RAXIS RAPID Imaging Plate Area Detector with graphite monochromated Mo KR radiation (λ ) 0.71073 Å). Results and Discussion Crystal Structure of [NO2BzPy][Au(mnt)2] at 273 K. This compound crystallizes in the monoclinic system with space group P2(1)/c and is isostructural with [NO2BzPy][Ni(mnt)2]. As shown in Figure 1a, the asymmetric unit within a cell is comprised of a coupled [Au(mnt)2]- and NO2BzPy+ with the anion [Au(mnt)2]- possessing roughly a planar geometry.

Figure 3. Variations of interatomic and interplane distances between neighboring anions in a stack (h1 and h2; see Figure 2) as x for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1) at 273 K.

Although the Au-S bonds are expected to be chemically equivalent, the bond lengths of Au-S are all different and range from 2.303(2) to 2.3153(19) Å which are longer than the lengths of Ni-S in [NO2BzPy][Ni(mnt)2] [2.1409(7)-2.1565(6) Å], whereas the bond angles in a chelate ring of S-M-S [90.41(7)° and 90.58(7)°] are lesser than that in [NO2BzPy][Ni(mnt)2] [92.32(3)° and 92.56(3)°]. In [NO2BzPy]+ moiety, the bond lengths and angles as well as the molecular configuration are almost the same as those in [Ni(mnt)2]-. For example, the dihedral angles of the pyridine ring and the benzene ring with

TABLE 4: The Coordinated Bond Lengths (Å) and Angles (deg) in an Anionic Moiety for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1)

x 0a 0.01a 0.04a 0.11a 0.21a 0.27a 0.27c 0.27c 0.35a 0.35b 0.41a 0.49a 0.49b 0.57a 0.73a 1a

M-S1

M-S2

M-S3

M-S4

S1-M-S2

S3-M-S4

2.1565(6) 2.1558(9) 2.1551(18) 2.1650(11) 2.1801(13) 2.1911(9) 2.220(5) 2.166(6) 2.1995(10) 2.193(4) 2.2174(9) 2.2231(14) 2.215(2) 2.2494(12) 2.2708(12) 2.3153(19)

2.1409(7) 2.1418(9) 2.1413(18) 2.1544(11) 2.1735(13) 2.1829(9) 2.224(5) 2.155(6) 2.1914(11) 2.185(4) 2.2209(9) 2.2158(15) 2.211(3) 2.2475(12) 2.2620(13) 2.303(2)

2.1446(7) 2.1463(9) 2.1490(18) 2.1671(11) 2.1910(13) 2.2010(9) 2.149(6) 2.239(6) 2.2110(10) 2.223(3) 2.2087(10) 2.2317(14) 2.255(2) 2.2340(13) 2.2712(12) 2.3111(19)

2.1442(6) 2.1449(9) 2.1490(18) 2.1690(11) 2.1946(13) 2.2068(9) 2.175(6) 2.219(5) 2.2117(10) 2.239(3) 2.1989(10) 2.2331(14) 2.248(2) 2.2422(13) 2.2690(12) 2.3045(19)

92.56(3) 92.54(4) 92.49(7) 92.61(4) 92.51(5) 92.33(4) 93.8(2) 93.8(2) 92.46(4) 93.22(13) 90.77(4) 90.87(5) 90.27(8) 90.71(5) 91.26(5) 90.58(7)

92.32(3) 92.20(4) 92.10(6) 91.72(4) 91.29(5) 91.04(3) 90.43(19) 90.3(2) 91.17(4) 90.35(11) 92.08(4) 91.81(6) 92.57(8) 91.64(5) 90.56(4) 90.41(7)

a Temperature ) 273 K. b Temperature ) 103 K. c Temperature ) 10 K (there are two crystallographically independent anions for x ) 0.27 at 10 K).

7674 J. Phys. Chem. B, Vol. 110, No. 15, 2006

Ren et al.

Figure 4. Plots of χm(T) versus T for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1) in the temperature range of (a) 70-350 and (b) 2-30 K.

the reference plane CAr-CH2-NPy are 87.1° and 94.1° in [Au(mnt)2]- versus 86.9° and 95.0° in [Ni(mnt)2]-, respectively. The dihedral angle between pyridine and benzene rings is 64.2° in [NO2BzPy][Au(mnt)2] and 65.2° in [NO2BzPy][Ni(mnt)2]. Both anions and cations constructed segregated stacks respectively that run parallel to c-axis (Figure 1b). Within anionic stacks, the face-to-face anions slid along the b-axis and eclipse each other. The interplane distances between the adjacent [Au(mnt)2]- anions, which is defined by four coordinated S atoms, are h1 ) 3.542 and h2 ) 3.576 Å (Figure 2a), while the adjacent Au‚‚‚Au distances are exactly identical [d1 ) d2 ) 3.908 Å], even if the Au atoms lie on a general point position.13 In addition, there is only one kind of crystallographic independent anion in an anionic stack. Therefore, the anionic stack can be considered as a regular one from the viewpoint of crystal structure. Likewise, the cationic stack is also regular with a uniform center-to-center distance of 4.301 Å between neighboring benzene rings (cf. Figure 2b). Crystal Structures of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) at 273 K. As shown in Table S1 and Table S2 (Supporting Information), every compound with the formula [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) is isostructural with [NO2BzPy][Au(mnt)2]; moreover, the cell parameters are rather close to those of [NO2BzPy][Au(mnt)2]. It is noteworthy that the [Au(mnt)2]- anions distribute randomly in the stack of [Ni(mnt)2]- for all crystals of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) at 273 K.14 In structural refinement for all crystals, the position occupancy factor of the Ni (Au) ion was figured out without any constraint and was consistent with the elemental analysis results of the corresponding compounds. The molecular and packing structures for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) are rather similar to those of [NO2BzPy][Au(mnt)2], and the bond lengths of M-S and angles of S-M-S in a chelate ring, along with their estimated standard deviations, are presented in Table 4. Like the parent compounds, both anions and cations respectively form the segregated and regular stacks in crystals of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73). With an increase in the doping concentration x, the distances of M‚‚‚M, M‚‚‚S, S‚‚‚S and interplanar distances between the nearest neighbors (h1, h2; refer to Figure 2a) within an anionic stack increase gradually but not monotonically (cf. Figure 3). Magnetic Susceptibilities. The magnetic behavior of each [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1) was determined on its polycrystalline sample from 2 to 350 K. These compounds are isostructural with each other; moreover the molecular structures of [Au(mnt)2]- and [Ni(mnt)2]- are rather similar. Therefore, it is reasonable to suppose that the diamagnetism in each [NO2BzPy][AuxNi1-x(mnt)2] equals roughly the magnetic susceptibility of [NO2BzPy][Au(mnt)2]. The corrected molar

Figure 5. Temperature dependences of magnetic susceptibility for [NO2BzPy][Ni(mnt)2] (open circles, experimental data; solid line, fits; for details see the text).

paramagnetic susceptibility of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-0.73) as well the molar susceptibility of [NO2BzPy][Au(mnt)2] are given in Figure 4 as χm ) f(T). For [NO2BzPy][Ni(mnt)2] (x ) 0), as reported in our previous work,11 a magnetic transition occurs around 181 K, and the phenomenology of the magnetic transition in this compound closely resembles that of a spin-Peierls transition in quasi-onedimensional spin systems. The spin energy gap with the value of 738 K in the dimerized state was estimated from the fits of the magnetic susceptibility data in the low-temperature phase, but the detailed analysis for magnetic susceptibility data in the high-temperature phase was not given there. Herein we further analyze its magnetic behavior in the high-temperature phase with a regular antiferromagnetic Heisenberg model for a linear chain with S ) 1/2;15 the temperature dependences of magnetic susceptibility in the range of 190-350 K were simulated to eq 1: 5

2 1+

χm )

Ng2µB 4kBT

NiX-i ∑ i)1 6

1+

(1)

DiX-i ∑ i)1

where X ) kBT/|J|, in which J is the magnetic exchange constant of the neighboring spins in a magnetic chain, and the coefficients of Ni and Di in the power series are the following: N1 ) -0.053 837 836, N2 ) 0.097 401 365, N3 ) 0.014 467 437, N4 ) 0.001 392 519 3, N5 ) 0.000 113 934 34; D1 ) 0.446 162 16, D2 ) 0.320 482 45, D3 ) 0.133 041 99, D4 ) 0.037 184 126, D5 ) 0.002 813 608 8, D6 ) 0.000 264 676 28, respectively. The best fitting yielded J/kB ) 192(3) K with a g-factor of

[NO2BzPy][Au(mnt)2]

J. Phys. Chem. B, Vol. 110, No. 15, 2006 7675

Figure 6. (a) plots of d(χmT)/dT versus T for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-0.35) and (b) the x dependences of TC.

2.106(10) (cf. Figure 5) and the correlation coefficient of this fitting R ) 0.984. The fitting value of the g-factor is comparable to that from EPR measurement.16 As shown in Figure 4a,b, the doping of [NO2BzPy][Ni(mnt)2] with nonmagnetic [NO2BzPy][Au(mnt)2] causes (1) an onset of a strong paramagnetic background and (2) a damping of the spin-transition. It is easily understandable that doping can induce the strong paramagnetic background since the antiferromagnetic spin chains of [Ni(mnt)2]- are decoupled at the position of nonmagnetic dopants [Au(mnt)2]-; moreover, the spin number in a chain should be randomly even and odd in equal probability. Therefore, the spin at the end of any odd-numbered spin chain becomes uncoupled, giving a contribution to the paramagnetic susceptibility. Obviously, this contribution is probably dominant especially for heavily doped systems. On the other hand, in the doped system, the transition was still observed at x ) 0.27 with the χm value suddenly droping in the plot of χm-T, while the variations of χm with T are smooth at x g 0.35 after passing a broad maximum upon cooling, and it is possible that the transition collapsed at x > 0.27. In the range of x ) 0.210.57, the broad maximum of magnetic susceptibility appears in the plot of χm-T, which is a typical characteristic of a lowered dimensionality of the spin system due to the existence of shortrange order. In the heavily doped system of x ) 0.73, the temperature dependences of magnetic susceptibility follow the Currie-Weiss law, and a reasonable fitting yielded the parameters C ) 5.59(4) × 10-2 emu‚K‚mol-1, θ ) -0.286(22) K, and correlation coefficient R ) 0.999 (Figure S2, Supporting Information). The transition temperature shifts to the lower temperature as the concentration of [NO2BzPy][Au(mnt)2] (x). To determine the reliable transition temperature (TC), we define TC as the peak temperature in the plot of d(χmT)/dT-T. Figure 6a depicts the derivative curves of χmT versus T for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-0.27), and Figure 6b plots the corresponding x dependences of TC. The values of TC(x) linearly reduce up to x e 0.27, which is expressed as TC(K) ) 180(2) - 221(12)x. Crystal Structures of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 at 10 K, x ) 0.35 and 0.49 at 103 K). The origin of the spin-Peierls-like transition is ascribed to the spin-lattice interaction, and the uniform spin chain is dimerized below the transition which is rather similar to the spin-Peierls transition.13a As a consequence, the critical behaviors observed in a doped spinPeierls system are probably discovered in the [NO2BzPy][AuxNi1-x(mnt)2] system. Therefore, we studied the crystal structures for the doped systems of x ) 0.27 at 10 K and x ) 0.35 and 0.49 at 103 K, respectively. [In the plot of χm-T for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.35), the χm value decreases smoothly after passing the broad maximum upon cooling, and

Figure 7. Oscillation photographs at T ) 273 and 103 K for x ) 0.11, 0.27, 0.35, and 0.41 (the arrows indicates the superlattice diffractions).

the characteristic sudden drop due to spin-Peierls-like transition disappears.] For the doped system of x ) 0.27, the space group degrades from P2(1)/c at 273 K (high-temperature phase) to P1h at 10 K (low-temperature phase). Moreover, the components of an asymmetric unit switch from a pair of NO2BzPy+ and [M(mnt)2]in the high-temperature phase into two ionic pairs in the lowtemperature phase. The molecular structures are nearly identical in both low- and high-temperature phases (refer to Table 4).

7676 J. Phys. Chem. B, Vol. 110, No. 15, 2006

Ren et al.

Figure 8. Temperature dependences of the molar capacity for (a) [NO2BzPy][Ni(mnt)2] and (b) [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 and 0.35).

The nonuniform compression and slippage of the magnetic chain (anionic stack) is observed from the high-temperature phase to the low-temperature phase, and the magnetic chain distorts with the alternating distances of the adjacent M‚‚‚M, 3.736 and 3.679 Å, in an anionic stack. For the doped system of x ) 0.35 (x ) 0.49), the crystal structure at 103 K is isostructural with that at 273 K, and the molecular and stacking structures are almost the same. At both temperature points, the anionic stack is uniform, in which the adjacent M‚‚‚M distances are exactly identical with the value of 3.835 Å at 273 K versus 3.762 Å at 103 K (3.851 Å at 273 K versus 3.775 Å at 103 K). The spin chain in such a doped system remains uniform. Super Lattice Diffractions for [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.11, 0.27, 0.35 and 0.41). Local structural fluctuation is a common phenomenon existing in the quasi-one-dimensional conductor or magnet that originates from electron-phonon or magnetoelastic coupling interaction. Figure 7 displays the temperature-dependent diffraction patterns for the doped system x ) 0.11, 0.27, 0.35, and 0.41, from which the super lattice diffractions were observed for x ) 0.11, 0.27, and 0.35 but was not detected for x ) 0.41 at 103 K. Heat Capacity for [NO2BzPy][Ni(mnt)2] and [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 and 0.35). The heat capacity results for the parent compounds [NO2BzPy][Ni(mnt)2] and [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27, 0.35; their concentrations are near transition suppressing) are shown in Figure 8a,b. In the Cp-T plot of [NO2BzPy][Ni(mnt)2], a sharp thermal abnormality with λ-shape was observed at ∼179 K, which coincides approximately with the temperature at which the abrupt drop of magnetic susceptibility occurs. The transition in [NO2BzPy][Ni(mnt)2], therefore, is first order. In the Cp-T plot of [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.27 and 0.35), no sizable thermal abnormality was observed. For the doped system of x ) 0.27, the transition is clearly identified by the crystal structure determination in the low-temperature phase, and such a transition is probably second order. Conclusion and Remarks In summary, we synthesized and characterized structurally ten nonmagnetic doped compounds with the formula [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0.01-0.73) and investigated the nonmagnetic impurity effects on the spin-Peierls-like transition in a quasi-one-dimensional magnet of [NO2BzPy][Ni(mnt)2]. The nonmagnetic impurities suppress the spin-Peierls-like transition, and the transition temperature TC reduces at an average rate of 221(12) K/percentage of nonmagnetic dopant. From the temperature dependences of magnetic susceptibility measurements, the transition collapse is estimated at around x > 0.27. In a heavier doped system, the energy gap originating from dimerization of a spin chain vanishes and a gapless phase is achieved

once again. The heat capacity investigations revealed that the transition is first order in the parent compound, [NO2BzPy][Ni(mnt)2], while it is second order in the heavier doping system. The nonmagnetic doping effects on the spin-Peierls-like transition observed in present work rather resemble that in a spinPeierls system. For example, the doping leads to the critical behavior where the dimerization state remains after the transition collapsed, and the doped system reaches a gapless phase again in the heavier doped system. Whereas the difference from the real spin-Peierls system is also distinct, such as in an inorganic spin-Peierls transition system CuGeO3, the transition is destroyed by nonmagnetic impurity only at around x ) 0.02,17 which is consistent with the theoretical prediction. In this study, the concentration for destroying the transition is x > 0.27. Obviously, it is necessary to explore further both experimentally and theoretically what causes such a big difference. Acknowledgment. This work was partly supported by a Grant-in-Aid for Science Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan. The authors thank Prof. K. Awaga and Dr. W. Fujita for Cp measurements. X.M.R. (JSPS fellow ID No. P03271) thanks the Japan Society for the Promotion of Science for financial support. Supporting Information Available: X-ray crystallographic files for each [NO2BzPy][AuxNi1-x(mnt)2] (x ) 0-1) (CIF) and a detailed description of the magnetic data acquisition and listings of crystallographic data. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Hase, M.; Terasaki, I.; Sasago, Y.; Uchinokura K. Phys. ReV. Lett. 1993, 71, 4059. (2) Oseroff, S, B.; Cheong, S. W.; Aktas, B.; Hundley, M. F.; Fisk, Z.; Rupp, L. W., Jr. Phys. Rev. Lett. 1995, 74, 1450. (3) Masuda, T.; Fujioka, A.; Uchiyama, Y.; Tsukada, I.; Uchinokura, K. Phys. Rev. Lett. 1998, 80, 4566. (4) Osawa, A.; Ono, T.; Tanaka, H. Phys. ReV. B 2002, 66, 020405. (5) (a) Isobe, M.; Ueda, Y. J. Magn. Magn. Mater. 1998, 177-181, 671. (b) Lohmann, M.; Krug von Nidda, H.-A.; Loidl, A.; Morre´, E.; Dischner, M.; Geibel, C. Phys. ReV. B 2000, 61, 9523. (6) Glazkov, V. N.; Smirnov A. I.; Krug von Nidda, H.-A.; Loidl, A.; Uchinokura, K.; Masuda, T. Phys. ReV. Lett. 2005, 94, 057205. (7) Mito, M.; Tanaka, S.; Kawae, T.; Takeda, K.; Yanagimoto, M.; Mukai K. Phys. B 2003, 329-333, 1150. (8) (a) Mukai, K.; Suzuki, K.; Ohara, K.; Jamali, J. B.; Achiwa, N. J. Phys. Soc. Jpn. 1999, 68, 3078. (b) Mukai, K.; Shimobe, Y.; Jamali, J. B.; Achiwa, N. J. Phys. Chem. B 1999, 103, 10876. (9) Ribas, X.; Sironi, A.; Masciocchi, N.; Lopes, E. B.; Almeida, M.; Veciana, J.; Rovira, C. Inorg. Chem. 2005, 44, 2358. (10) (a) Bray, J. W.; Hart, H. R., Jr.; Interrante, L. V.; Jacobs, I. S.; Kasper, J. S.; Watkins, G. D.; Wee, S. H.; Bonner, J. C. Phys. ReV. Lett. 1975, 35, 744. (b) Jacobs, I. S.; Bray, J. W.; Hart, H. R., Jr.; Interrante, L. V.; Kasper, J. S.; Watkins, G. D.; Prober, D. E.; Bonner, J. C. Phys. ReV.

[NO2BzPy][Au(mnt)2] B 1976, 14, 3036. (c) Huizinga, S.; Kommandeur, J.; Sawatzky, G. A.; Thole, B. T.; Kopinga, K.; de Jonge, W. J. M.; Roos, J. Phys. ReV. B 1979, 19, 4723. (11) Ren, X. M.; Meng, Q. J.; Song, Y.; Lu, C. S.; Hu, C. J.; Chen, X. Y. Inorg. Chem. 2002, 41, 5686. (12) Sheldrick, G. M. SHELXL-97, Program for the Refinement of Crystal Structure; University of Go¨ttingen: Go¨ttingen, Germany, 1997. (13) (a) Ren, X. M.; Akutagawa, T.; Nishihara, S.; Nakamura, T.; Fujita W.; Awaga K. J. Phys. Chem. B 2005, 109, 16610. (b) Notes: In fact, the Au(1) atom lies on a general point position, and the two neighboring Au(1)i and Au(1)ii atoms in a stack are generated by the symmetric transforms, where i is x, 0.5 - y, -0.5 + z and ii is x, 0.5 - y, 0.5 + z. Defining the fractional coordinates of Au(1) as (x, y, z), the coordinates of the two neighbors, (x, 0.5 - y, 0.5 + z) of Au(1)ii and (x, 0.5 - y, -0.5 + z) of Au(1)i, are deduced according to the crystal data. Therefore, the Au(1)‚‚‚ Au(1)ii and Au(1)‚‚‚Au(1)i distances are equal to the moduli of the corresponding vectors of (0, 0.5 - 2y, 0.5) and (0, 0.5 - 2y, -0.5), respectively, which coincide on the YZ-plane of the crystallographic coordinate system. The two-dimensional coordinate system (Y-O-Z) is orthogonal in the monoclinic system; accordingly, the lengths of the vectors

J. Phys. Chem. B, Vol. 110, No. 15, 2006 7677 should be given by |Au(1)‚‚‚Au(1)ii| ) |(0, 0.5 - 2y, 0.5)| ) {02 + (0.5 2y)2 + 0.52}1/2 and |Au(1)‚‚‚Au(1)i| ) |(0, 0.5 - 2y, -0.5)| ) {02 + (0.5 - 2y)2 + (-0.5)2}1/2 ) {02+ (0.5 - 2y)2 + 0.52 }1/2. So that |Au(1)‚‚‚ Au(1)i| ) |Au(1)‚‚‚Au(1)ii|. In other words, the interatomic distances of Au(1)‚‚‚Au(1)i and Au(1)‚‚‚Au(1)ii are exactly identical. Additionally, the distance of Au(1)i‚‚‚Au(1)ii ) |Au(1)i‚‚‚Au(1)ii| ) |(0, 0, 1)|, that is, the crystallographic cell length of c-axis. (14) The crystal structural analyses in low temperature for C20H11AuxNi1-xN6O2S4 (x ) 0.35 and 0.49 at 103 K.; x ) 0.27 at 10 K) revealed that the distribution of [Au(mnt)2]- and [Ni(mnt)2]- anions in the crystals is still random. (15) Johnston, D. C.; Kremer, R. K.; Troyer, M.; Wang, X.; Klu¨mper, A.; Bud’ko, S. L.; Panchula, A. F.; Canfield, P. C. Phys. ReV. B 2000, 61, 9558. (16) Ren, X. M.; Kremer, R. K.; Meng, Q. J. J. Magn. Magn. Mater. 2004, 272-276, 924. (17) Wang, Y. J.; Kim, Y. J.; Christianson, R. J.; Lamarra, S. C.; Chou, F. C.; Masuda, T.; Tsukada, I.; Uchinokura, K.; Birgeneau, R. J. J. Phys. Soc. Jpn. 2003, 72, 1544.