crystengcomm - riken · this paper, we report some supramolecular ni(dmit) 2 radical anion salts...

Cite this: CrystEngComm, 2013, 15,3200

Supramolecular Ni(dmit)2 salts with halopyridiniumcations -development of multifunctional molecularconductors with the use of competing supramolecularinteractions3

Received 12th November 2012,Accepted 2nd January 2013

DOI: 10.1039/c2ce26841h

www.rsc.org/crystengcomm

Yosuke Kosaka,a Hiroshi M. Yamamoto,*ab Akiko Tajima,a Akiko Nakao,a

Hengbo Cuia and Reizo Kato*a

Halopyridinium cations with multiple halogen and hydrogen bonding donor sites have been used as the

counter ions for the Ni(dmit)2 anion (dmit = 1,3-dithiole-2-thione-4,5-dithiolate). Because of the

competing supramolecular interactions, such as halogen and hydrogen bondings, some of the Ni(dmit)2

salts showed complicated structures with multiple functionality that originates from the intriguing

molecular arrangements. In these multifunctional salts, Ni(dmit)2 molecules play two roles, conducting and

magnetic, depending on the molecular arrangement in the layers that they belong to. The physical

properties of these salts have been examined by conductivity measurements, with or without hydrostatic/

uniaxial pressure, as well as magnetic susceptibility and band calculations. By analysing the low

temperature conductivity, it can be concluded that the itinerant electrons in the conducting layer have

magnetic coupling with the localized spins in the magnetic layers to result in a Kondo singlet formation.

Introduction

Multifunctional materials are of wide and intense interest inmaterial science because interference among different physi-cal properties sometimes provides a source of fast, low-dissipation and/or high-density electronic devices that cannotbe achieved without them. In particular, the interactionbetween the charge and spin of an electron is one of the mostimportant issues in spintronics,1 heavy fermions2 and high-TC

superconductors.3 Though the quest for inorganic multi-functional materials is preceding that of organic materials,the organic counterparts are also under considerable investi-gation in this regard. The number and variety of organicmultifunctional materials are, however, still limited andexpansion is needed to enrich the technology and science oforganic devices in the next era. Since the properties of organicmaterials are severely dependent on the molecular arrange-ments inside them, we have been studying a supramolecularstrategy to control the conduction and magnetic properties ofmolecular conductors by introducing halogen interactions.4 Inthis paper, we report some supramolecular Ni(dmit)2 radicalanion salts with bi-functionality based on the molecular

arrangement. Many halogen and hydrogen bonds are foundin the crystal structures and the competition and cooperationamong these supramolecular interactions play an importantrole in constructing the unique and complex crystal structures.Such a special structure resulted in a coexistence of conduct-ing and magnetic activities in the crystal, in which one kind ofmolecule plays two contrasting roles. The nature of thethioketone–halogen/hydrogen interactions will be also dis-cussed.

Results and discussion

Electrochemical oxidation of (nBu4N)[Ni(dmit)2] in the pre-sence of the cations listed in Scheme 1 afforded five Ni(dmit)2

anion-radical salts with supramolecular networks: (Me-3,5-DBP)[Ni(dmit)2]2 (1), (Me-3,5-DIP)[Ni(dmit)2]3 (2), a-(Me-3,5-DIP)[Ni(dmit)2]2 (3) b-(Me-3,5-DIP)[Ni(dmit)2]2 (4) and (Me-3,5-BIP)[Ni(dmit)2]2 (5). Because of the high directionality ofhalogen bonding and hydrogen bonding, the crystal structure

aRIKEN, Hirosawa, Wako, Saitama, 351-0198, Japan. E-mail: [email protected];

[email protected]; Fax: +81 48 462 4661; Tel: +81 48 467 9410bInstitute for Molecular Science, Myodaiji, Okazaki, Aichi, 444-8585 Japan.

Fax: +81 564 55 7325; Tel: +81 564 55 7334

3 CCDC 645114, 910061–910062, 910623 and 915357. For crystallographic data inCIF or other electronic format see DOI: 10.1039/c2ce26841h Scheme 1 Molecular structures.

CrystEngComm

PAPER

3200 | CrystEngComm, 2013, 15, 3200–3211 This journal is � The Royal Society of Chemistry 2013

http://dx.doi.org/10.1039/c2ce26841h

http://dx.doi.org/10.1039/c2ce26841h

is restricted to exhibit asymmetry to some extent. Thisasymmetry is relaxed in crystals 1 and 2 by the introductionof disorder, while highly complex structures that accommo-date this asymmetry are realized in crystals 3 and 5. Fromthese results, we anticipate that supramolecular frustrationintroduced by multiple halogen and hydrogen bondingformations can be a tool to induce unconventional compli-cated structures. Because such structures resulted in alternatestacking of conducting and magnetic 2D (2D = two-dimen-sional) layers in 3 and 5, this seems a good strategy for theconstruction of bi-functional materials to bring in suchcompeting supramolecular interactions, although rationaldesign is not possible as yet.

Crystal structures and band calculations

(Me-3,5-DBP)[Ni(dmit)2]2 (1). The crystal structure of 1 isshown in Fig. 1 and 2. Two Ni(dmit)2 anions (A/B) and onecation are crystallographically independent, both of whichform 2D layers that stack alternately to one another. TheNi(dmit)2 anions form conventional columnar structureswhere the anions stack along the a–b direction with theirmolecular planes parallel to each other. A repeating unitcontains four molecules (A, A9, B9 and B), two pairs of whichare interrelated by an inversion center (A–A9 and B–B9). Theinterplanar distances are 3.57 (A–B), 3.68 (A–A9) and 3.51 Å(B–B). The cation (Me-3,5-DBP)+ is orientationally disordered andpossesses two possible positions (Sites A and B) with anoccupancy ratio of 60 : 40. The Br atom of the cation and the Satom of the terminal thioketone groups of the dmit ligandform halogen bonding Br…S interactions, whose lengths are3.08 and 3.24 Å (Site A) and 3.51 and 3.40 Å (Site B),respectively. In addition to these halogen bonding interac-tions, there seem to be hydrogen bondings, although thepositions of the hydrogen atoms are not refined anddetermined by calculations. For example, the hydrogen atomof the pyridine ring (at the 2 position) interacts with thethioketone group to form H…S interactions with lengths of2.82 (Site A) and 2.78 Å (Site B), respectively. These are muchshorter than the sum of the van der Waals radii of 3.65 Å (for

Br…S, Br: 1.85 Å, S: 1.80 Å) or 3.00 Å (for H…S, H: 1.20 Å),respectively, which builds up the supramolecular networkthrough these strong interactions. In each site, through theseinteractions, one cation molecule bridges two Ni(dmit)2 layerswith supramolecular interactions and a 1D supramolecularchain along the b + c direction is formed (Fig. 1). The yellowellipses mark the molecules that do not participate in theinfinite supramolecular chain but interact with the cation toform a branch. The coexistence of sites A and B seems to be aresult of competition between the hydrogen and halogeninteractions, where a strong halogen bond (3.08/3.24 Å) and aweak hydrogen bond (2.82 Å) are formed in site A while a weakhalogen bond (3.40/3.51 Å) and a strong hydrogen bond(2.78 Å) are formed in site B. Although this kind of competitionexists, the whole crystal structure is still a product of thecooperation of these two interactions. The calculated overlapintegral for p3 is much larger than p1 and p2 (Fig. 3). Thissuggests strong dimerization of the Ni(dmit)2 anions, whichreflects the short interplanar distance of p3. Every inter-columnar interaction (s1–s3, r1 and r2) is rather small becauseof the symmetry of the LUMO of the Ni(dmit)2 molecule.5

Since the weaker intracolumnar and the intercolumnarinteractions are comparable, this system should have a 2Dcharacter with a narrow dispersion. The tight-binding bandcalculations imply a large separation between the highest andlowest sub-bands associated with the strong dimerization.Every sub-band has a narrow width due to the weak inter-

Fig. 1 Supramolecular network of 1 viewed along its perpendicular direction(upper; Site A, lower; Site B). Supramolecular Br…S and H…S interactions areindicated by dotted lines.

Fig. 2 Crystal structure of 1 viewed along the longitudinal axis (left: Cation islocated in site A, right: Cation is located in site B). Supramolecular Br…S andH…S interactions are indicated by dotted lines.

Fig. 3 End-on projection of Ni(dmit)2 anions (left) and the band dispersion(right) for 1. Overlap integrals among the LUMOs (61023) are as follows: p1 =3.39, p2 = 1.57, p3 = 18.0, s1 = 0.096, s2 = 22.04, s3 = 1.77, r1 = 20.36, r2 =20.46; (interlayer) c = 20.34. The inset in the right panel shows the Brillouinzone. EF denotes the Fermi energy.

This journal is � The Royal Society of Chemistry 2013 CrystEngComm, 2013, 15, 3200–3211 | 3201

CrystEngComm Paper

dimer and intermolecular interactions. With a band filling of1/4, this system would be a band insulator with a band gap Eg

of y100 meV. Indeed, the temperature dependence of thiscrystal shows an activation-type behavior with a 55 meVactivation energy, as will be discussed later.

(Me-3,5-DIP)[Ni(dmit)2]3 (2). The crystal structure of 2 isshown in Fig. 4. One and a half of the Ni(dmit)2 anions (A andB) and a half of the cation Me-3,5-DIP are crystallographicallyindependent. The Ni(dmit)2 anion A is located on theinversion center. The cation is located on the two-fold rotationaxis and occupies two possible positions (Sites A and B) due tothe orientational disorder. The occupancy ratio of Sites A andB is 50 : 50. Their positions, however, cannot be exactlydetermined because these sites are very close to one another.The unit cell contains two crystallographically equivalentlayers of Ni(dmit)2 anions. In each layer, the Ni(dmit)2 anionsform two columns, which are interrelated to one another bythe glide plane. The Ni(dmit)2 molecules are arranged in aherringbone manner. The interplanar distances are 3.48 (A–Band A–B9) and 3.68 Å (B–B9). The cation and Ni(dmit)2 anionsform supramolecular I…S and H…S interactions, whoselengths are 3.22, 3.52 Å (I…S; Sites A and B, respectively) and2.91 Å (H…S), respectively. These are shorter than the sum ofvan der Waals radii (3.78 Å for I…S, I: 1.98 Å, S: 1.80 Å) and asupramolecular network is formed with these interactions.The supramolecular patterns based on the cations at Sites Aand B, however, are different from one another. Infinite 1Dchains run along the a + 2c direction and are associated withthe cation in Site A (Fig. 4 upper panel). On the other hand, thecations in Site B form 1D zig-zag chains (Fig. 4 lower panel).

Calculated overlap integrals suggest that the Ni(dmit)2

anions exhibit weak trimerization in the column. Everyintercolumnar interaction (r1–r3 and s1–s3) is considerablyweaker than intracolumnar interactions (Fig. 5). This column,therefore, should have a q1D (=quasi-one dimensional)character with weak degeneracy. The tight-binding calculation

affords energy bands. Six branches are divided into threegroups (Fig. 5) due to the following three aspects: 1) there areno interlayer interactions between the two crystallographicallyequivalent Ni(dmit)2 anion layers in the unit cell and thus, thesub-bands associated with Layers I and II are degenerate toeach other, 2) two equivalent 3-fold columns within the anionlayer afford three groups and 3) small intercolumnar interac-tions weakly break the degeneracy within each subgroup. Thisband structure indicates a q1D metallic character. However,since the lower sub-band group is effectively half-filled and theband width is narrow, this system would be a Mott insulator.

a-(Me-3,5-DIP)[Ni(dmit)2]2 (3). This crystal can be selectivelyyielded by electrochemical crystallization with current switch-ing. The crystal structure of 3 is shown in Fig. 6. Two Ni(dmit)2

anions (A and B) and one cation are crystallographicallyindependent. The crystal comprises four stacked repeatingNi(dmit)2 anion layers, two of which are crystallographicallyindependent (Layers I and II). The internal structures of LayersI and II are considerably different from one another. Theselayers are repeated alternately along the c-axis. The terminalthioketone groups of the Ni(dmit)2 anion in both layers areassociated with the iodine and hydrogen atoms of the cationsthrough supramolecular interactions. The I…S lengths are 3.28and 3.49 Å and the S…H lengths are 2.80 and 2.86 Å,respectively. These are rather shorter than the sum of vander Waals radii (rI…S: 3.78 Å and rS…H: 3.00 Å). Thesupramolecular networks run along the c-axis direction(Fig. 7). In Layer I, the Ni(dmit)2 anions stack along the a–bdirection. The calculated overlap integrals among the LUMOsare shown in Fig. 8 and listed in Table 1. These suggest strongdimerization of the Ni(dmit)2 units. The intradimer interac-tion, p1, is approximately 2 orders of magnitude larger than all

Fig. 4 Supramolecular network of 2 for Site A (upper panel) and B (lower panel).Supramolecular I…S and H…S interactions are indicated by dotted lines.

Fig. 5 End-on projection of the Ni(dmit)2 layer, band dispersion and Fermisurface for 2. Overlap integrals among the LUMOs (61023) are as follows: b1 =26.63, b2 = 23.76, s1 = 20.84, s2 = 0.64, s3 = 0.079, r1 = 20.053, r2 = 0.66, r3 =0.82. EF denotes the Fermi energy.

Fig. 6 Crystal structure of 3 (side view). Supramolecular I…S and H…Sinteractions are indicated by dotted lines.


Paper CrystEngComm

the other interdimer interactions in Layer I (p2, s1–s2, and b1).The interplanar distances are 3.45 Å within the dimer and3.82 Å between the dimers. The Raman spectra indicate thatthe formal charge of the Ni(dmit)2 anions in Layers I and II arethe same. The tight-binding band calculation suggests a largeseparation (y400 meV) between the upper and lower sub-bands, each of which has an extremely narrow bandwidth, W,of y20 meV (Fig. 9). HOMO–LUMO band inversion does notoccur in every layer because the HOMO–LUMO energydifference is 0.6–0.8 eV.6 The effective on-site Coulombrepulsion in the dimer, Ueff, is expressed by the equation interms of the intradimer transfer integral, |tdimer|, and theintramolecular Coulomb repulsion, Ubare, and is obviouslygreater than W (Ueff » W) for the half-filled lower sub-band.7

Layer I, therefore, should be in a Mott insulating state, which

displays paramagnetism based on a localized spin on thedimer unit [Ni(dmit)2]2

2.On the other hand, in Layer II, the Ni(dmit)2 anions form a

non-columnar structure that includes two overlapping modes(Modes I and II), as shown in Fig. 10. In Mode I, one moleculeoverlaps with two molecules, which is called the ‘‘spanning-overlap’’.5 Meanwhile, one molecule overlaps with only onemolecule in Mode II. The band calculation affords a two-dimensional Fermi surface with an elliptical cross-section(Fig. 11), suggesting that Layer II exhibits 2D metallicconduction. Since the Ni(dmit)2 anion in Layer II has a chargeof 20.5, the 2D Fermi surface has 50% of the area of the firstBrillouin zone. This calculation is consistent with a result fromquantum (Schubnikov–de Haas) oscillation where a Fermisurface of 51% of the first Brillouin zone is observed.8 It isnotable that the band filling for both Layer I and II is the samedespite the Fermi level and should be different (due to thedifferent band structure). Such a situation can be explained bya strong on-site Coulomb repulsion in Layer I that excludes anaddition of excess electrons from Layer II, which is reminis-cent of the relationship between the s- and d-electrons inKondo insulators.2

b-(Me-3,5-DIP)[Ni(dmit)2]2 (4). This compound was obtainedin the same batch as crystals 2 and 3. The crystal structure isshown in Fig. 12 and 13. Two Ni(dmit)2 anions and one cation

Fig. 7 Supramolecular network in 3. Supramolecular I…S and H…S interactionsare indicated by dotted lines.

Fig. 8 End-on projection of the Ni(dmit)2 anions in Layers I (left) and II (right).

Table 1 Calculated overlap integrals among LUMOs (61023) for 3 and 5

Crystal p1 p2 b1 s1 s2 p3 p4 p5 b2 ca

3 23.4 20.38 0.41 20.66 0.67 213.3 20.50 28.80 0.68 0.305 24.4 21.52 0.30 20.69 1.23 212.3 20.51 29.69 0.74 0.18

a Interlayer interaction (Fig. 6)

Fig. 9 Calculated band dispersions for Layer I in 3 (left) and 5 (right).

Fig. 10 Molecular arrangement for Layer II in 3.

Fig. 11 Fermi surface (left) and band dispersion (right) for Layer II in 3. (aP=(a +b)/2, bP = b, cP = c).


CrystEngComm Paper

are crystallographically independent. The molecular arrange-ment of Ni(dmit)2 is a columnar structure similar to that in 1but the supramolecular interactions between the cations andNi(dmit)2 anions are different. The lengths of the supramole-cular I…S interactions are 3.27 and 3.31 Å (I…S). In addition,the 2-position H atom of the pyridine ring and the thioketoneare associated to form supramolecular H…S interactions,whose lengths are 2.97 Å. With these interactions, one cationmolecule bridges two Ni(dmit)2 columns within a Ni(dmit)2

layer and a supramolecular 2D network is constructed in the bcplane (Fig. 13). The calculated overlap integrals p1 and p2 aresignificantly stronger than p3 (Fig. 14). This suggests atetramerization, which is supported by the interplanardistances between the Ni(dmit)2 anions of 3.56, 3.53 and3.61 Å (corresponding to p1, p2 and p3, respectively). Thetight-binding calculation affords the band structure, where theenergy band associated with the tetramer splits into four sub-bands. Since p2 is larger than p1, the highest and lowest sub-bands are widely separated and the other two sub-bands arelocated near the original LUMO level of the Ni(dmit)2. Due tothe band filling of 1/4, this system is expected to be a bandinsulator. The band gap can be estimated to be ca. 60 meV.

(Me-3,5-BIP)[Ni(dmit)2]2 (5). The crystal structure is isomor-phous to that of 3. The substitution of a bromine for an iodineof the pyridine ring leads to about a 2% volume reduction (seethe Experimental section). Compared to the cell parameters of

3, the b-axis is nearly comparable, while the a-axis and c-axisshorten by y0.1 and y0.4 Å, respectively. The cation isdisordered and the iodine and bromine atoms share the samepositions (Sites A and B) with an occupancy ratio of 67 : 33.The unit cell holds the columnar structure (Layer I) and thespanning-overlap (Layer II) of the Ni(dmit)2 anions. In Layer I,the interplanar distances are 3.80 (interdimer) and 3.44 Å(intradimer), which are shorter than those in 3. Thiscorresponds to the shortening of the a-axis. The cations andNi(dmit)2 anions form supramolecular X…S interactions withlengths of 3.26 (Mode A; I…S, Mode B; Br…S) and 3.50 Å (ModeA; Br…S, Mode B; I…S). These are rather shorter than the sumof the van der Waals radii. Moreover, supramolecular H…Sinteractions are formed in the same manner as in 3. Theselengths are 2.80 and 2.88 Å.

The calculated overlap integrals among the LUMOs arelisted in Table 1. In Layer I, interdimer interactions, inparticular the values of p2 and s2, are enhanced compared tothose in 3. This can be due to the ‘‘chemical pressure’’ effectalong the a-axis direction. The tight-binding band calculationsfor Layer I still afford two widely separated sub-bands.However, the lower sub-band has a wider band width, W, ofy50 meV than that of 3, which is due to the enhancement ofthe interdimer interactions. The effective on-site Coulombrepulsion in the dimer, Ueff, is still greater than W (Ueff . W)for the half-filled lower sub-band (Fig. 9, right panel) and LayerI would be in a Mott insulating state. On the other hand, everyoverlap integral is almost comparable to those of 3 in Layer II.The calculated Fermi surface, therefore, implies that Layer IIexhibits 2D metallic character (Fig. 15).

Overview of the crystal structures. A metal dithiolatecomplex M(dmit)2 (M = Ni, Pd, Pt, etc.) can form conductinganion radical salts with various cation species.5 Since side-by-side interactions between Ni(dmit)2 anions tend to becancelled out due to the b2g symmetry of the LUMO, many oftheir conduction properties have 1D character.(Me4N)[Ni(dmit)2]2, for example, has a columnar structurewhich is conventional among the Ni(dmit)2 anion radical salts.This system has two pairs of q1D Fermi surfaces that originatefrom the ‘‘solid crossing columnar structure’’: two crystal-lographically equivalent q1D columns are aligned in a crossdirection to each other.9 On the other hand, in a-

Fig. 12 Crystal structure of 4. Intermolecular I…S and H…S interactions areindicated by dotted lines.

Fig. 13 Supramolecular network in 4. Intermolecular I…S and H…S interactionsare indicated by dotted lines.

Fig. 14 End-on projection of the Ni(dmit)2 layer (left panel) and band dispersion(right panel) in 4. Overlap integrals among the LUMOs (61023) are as follows:p1 = 9.85, p2 = 14.7, p3 = 2.27, s1 = 20.47, s2 = 4.19, s3 = 22.67, r1 = 20.39, r2= 20.58. The inset in the right panel shows the Brillouin zone. EF denotes theFermi energy.


Paper CrystEngComm

(Et2Me2N)[Ni(dmit)2]2, the Ni(dmit)2 anions no longer form acolumnar structure. One molecule overlaps with two mole-cules, which is called ‘‘spanning-overlap’’. This arrangementcompensates for weak side-by-side interactions, which leads to2D conduction properties. In addition, Ni(dmit)2 anions canbe arranged in various ways such, as the herringbone10 andface-to-edge arrangement manners11 and so on. The cations’bulkiness and shape along with weak unintended interactionsare the origin of this variety. In the present work, on the otherhand, the molecular arrangement in the crystals is stronglyaffected by the formation of a supramolecular network so thatthe coexistence of both q1D and 2D electron systems arerealized in 3 and 5 (the physical properties will be describedlater). Here, we overview the ways in which the supramolecularsynthons affect the arrangement of Ni(dmit)2 in the crystal.

Schematic supramolecular coordination manners for 1–5are depicted in Fig. 16. Supramolecular 1D chains withbranches, comprising the cations and Ni(dmit)2 anions, areformed in 1 for both cation sites (A and B). In compound 2, thecations in Site A form 1D 2-leg chains with the Ni(dmit)2

molecules. On the other hand, the cations in Site B form 1Dchains with branches of the Ni(dmit)2 anions. Compounds 3and 5 have closely linked supramolecular grids comprised ofthe cations and Ni(dmit)2 anions to construct the 2D network.Compound 4 has a 2D supramolecular mesh. Three kinds ofarrangements of the Ni(dmit)2 anions, columnar, herringboneand spanning-overlap structures, have been observed in thepresent five compounds 1–5. These arrangements are alsofound in the conventional conducting Ni(dmit)2 anion radicalsalts but many differences are observed in these cases. Forexample, compounds 1 and 4 have similar 4-fold columnarstructures of the Ni(dmit)2 anions. However, the Ni(dmit)2

anions are dimerized within the column in 1, while those in 4are tetramerized. This seems to be due to the differencebetween the supramolecular Br…S and I…S interactions. In 1,the H atoms of the cations interact with one kind ofthioketone group to form significantly short supramolecularH…S interactions. On the other hand, one of the two Br atomsof the cations interacts with two kinds of thioketone groups(Sites A and B) due to the disorder. This orientational disorderof the cation seems to originate from the weakness of the Br…Shalogen bonding. In 4, on the other hand, short supramole-cular I…S interactions are formed and the H…S interactionsare relatively longer than those in 1. This could mean adifference in priority of these interactions: the Br…S interac-tions are comparable to or weaker than the H…S interactions,whereas the I…S interaction is stronger than the H…Sinteractions. This competition between the halogen andhydrogen bondings seems to result in the difference in thedetails of the stacking motifs. In addition to these examples,the coexistence of columnar and spanning-overlap structuresfound in crystals 3 and 5 is quite rare in Ni(dmit)2 salts.Because the lengths of the C–I and C–H covalent bonds in Me-3,5-DIP are quite different, the Ni(dmit)2 molecules need totwist their principle axes from the normal stacking positionsin order to satisfy all the halogen and hydrogen bodinggeometries. This competition and cooperation among themultiple interactions are the origin of such a unique structure.

The distances and angles of the X…S and H…S interactionsbetween the cations and Ni(dmit)2 anions are summarized inTable 2. The shortest distances of the supramolecular X…Sinteractions are 3.08 and 3.22 Å (for rBr…S and rI…S,respectively), which are ca. 15% shorter than the sum of thevan der Waals radii. Furthermore, the H atoms of the pyridinerings (2 and 5 positions) form supramolecular H…S interac-tions with the Ni(dmit)2 anions. The shortest distance of 2.78 Åis ca. 7% shorter than the sum of van der Waals radii. The C–X…S angles are almost restricted to ca. 155u–170u, except forthe interactions with long distances. This is because thecoordination angles in the C–X…S interaction are governed bythe geometry of s-hole of the X atom, which is projected in theopposite direction to the C–X bond. On the other hand, Fig. 17depicts a schematic of the sp2 orbitals of the CLS bond. The Q

and h represent the directions of the interactions. Taking onlythe short interactions into account, the Q values are in therange of 100–150u (for Br…S) and 105–125u (for I…S), while theh values have no specific angular limitation (Table 3).

Generally, the directions of halogen bonding interactionsare mainly given by electrostatic effects.12 For an ideal sp2

Fig. 16 Supramolecular coordination manners (panels on the left) and theirschematic views (panels on the right) for 1, 2, 3/5 and 4.

Fig. 15 Fermi surface (left) and band dispersion (right) for Layer II in 5. (aP=(a +b)/2, bP = b, cP = c).


CrystEngComm Paper

orbital, the Q and h values should be 120u (or 240u) and 90u,respectively. Most of the Q values are situated within a narrowrange around 120u, while the h values are distributed in a widerange. These correspond to previously reported data forhalogen bonding with ketones.13 It seems that both lone pairsof the ketonic sulfur and the p electrons of the CLS bond areparticipating in the halogen bonding.

Physical properties

The coexistence of conduction and magnetic electrons leads tonumerous intriguing phenomena. In the field of inorganicsystems, for example, some metal oxides can exhibit a steepdecrease in the electrical resistivity when a magnetic field isapplied. This is called the giant magnetoresistive (GMR)effect.14 Moreover, competition between the Ruderman–Kittel–Kasuya–Yoshida (RKKY) interactions and the Kondosinglet state leads to the enhancement of the effective mass,called the heavy fermion state.2,15 Molecular conductors haveprovided a wide variety of materials that exhibit a coexistenceof metallic conduction and magnetic properties. In mostmagnetic molecular conductors, however, conducting ionradicals and insulating closed-shell counter ions form theirown layered structures which align alternately. When theinsulating layer contains localized d-moments, the system canexhibit both conducting and magnetic properties. Severalcompounds, b99-(BEDT–TTF)4(H2O)Fe(C2O4)3?C6H5CN (super-

conductor with paramagnetism),16 (ET)3[MnCr(C2O4)3] (metalwith ferromagnetism),17 l-(BETS)2FeX4 (X = Br, Cl; magnetic-field-induced superconductor)18 and k-(BETS)2FeCl4 (super-conductor with antiferromagnetism)19 exhibiting ‘bi-function-ality’ have been reported as typical examples. They are knownas the ‘pp–d system’.

In contrast to these pp–d systems, both itinerant andlocalized electrons are surely derived from molecular p-elec-trons of the Ni(dmit)2 anion in the present case. Magneticmolecular conductors without localized 3d moments havebeen scarcely reported and the sole example is thea-Per2M(mnt)2 series (Per = perylene; mnt = maleonitrile-dithiolate; M = Ni, Pd, Pt, Fe, Cu, Au and Co). This systemexhibits one-dimensional metallic behavior and paramagnet-ism with antiferromagnetic (AF) interactions, which arederived from the molecular p-electrons of Per2+ andM(mnt)2

2, respectively.20 In the present 3 and 5 system,however, ‘‘one’’ molecule plays ‘‘two’’ contrastive roles.

Conductivity. The electrical resistivities of 1, 2 and 4 areshown in Fig. 18. The room-temperature resistivities, rrt, of 1and 4 are 3.2 and 0.60 V cm, respectively and increasemonotonically with decreasing temperature. This supportsthat they are band insulators, as anticipated from the tight-binding band calculation. Compound 2 also shows semicon-ductive behavior with a rrt of 0.16 V cm, which indicates that 2is a Mott insulator, as is also expected from the bandcalculation.

The anisotropic temperature dependence of the electricalresistivities for 3 and 5 are depicted in Fig. 19. As anticipatedfrom the band calculations, compound 3 exhibits 2D metallicbehavior in the ab-plane. The resistivity along the b-axis

Table 2 Selected bond lengths and angles for compounds 1–5. The values in parentheses represent the lengths of relatively longer interactions

1 2

3 4 5A B A B

X(1)…S (Å) 3.240(6) 3.401(8) 3.5192(19) 3.2166(17) 3.2771(19) 3.265(3) 3.256(2)X(2)…S (Å) 3.082(3) (3.510(5)) — — 3.491(2) 3.313(3) (3.503(2))H(1)…S (Å) 2.816 2.778 (2.911) (2.911) 2.795 (2.969) 2.796H(3)…S (Å) — — — — 2.863 — 2.880C–X(1)…S (u) 171.33 158.77 157.92 165.12 162.91 167.36 163.65C–X(2)…S (u) 168.98 (143.04) — — 154.31 172.57 152.32

For the numbering of the atoms, see Fig. 17.

Fig. 17 Numbering of the atoms (upper panel) and schematic orbitals for theCLS bond (lower panel).

Table 3 Selected Q and h angles (u) for compounds 1–5. The values inparentheses show the angles of relatively longer interactions

1

2 3 4 5A B

CLS…X(1) Q 100 105 110 105 125 105h 10 15 30 65 30 60

CLS…X(2) Q 150 (95) — 105 125 (105)h 90 (15) — 20 55 (20)

CLS…H(1) Q 125 130 — 130 (140) 130h 90 90 — 80 (90) 85

CLS…H(3) Q — — — 125 — 125h — — — 45 — 45


Paper CrystEngComm

decreases monotonically with decreasing temperature. Theresistivity along the a-axis also shows essentially metallicbehavior down to 4.2 K, although it has a broad minimum anda broad maximum around 155 K and 72 K, respectively. Incontrast, for the interlayer direction (along the c-axis), theresistivity increases with decreasing temperature. A broadmaximum and a broad minimum are observed at about 65 and20 K, respectively. In addition, an anomaly appears at about185 K. At 4.2 K, the resistivity along this direction is larger thanalong the a- and b-axis directions by more than four orders ofmagnitude. These features clearly indicate that the metalliclayer (Layer II) and the thick insulating layer (Layer I; y19 Å)alternate. Compound 5 shows 2D metallic behavior in the ab-plane similar to that of 3. The resistivity along the b-axisdecreases monotonically, whereas along the a-axis, itdecreases with a broad minimum (y145 K) and maximum(y52 K). Below 10 K, an increase in the resistivity is observedalong both directions. On the other hand, the resistivity alongthe c-axis increases with a broad maximum (y50 K) and abroad minimum (y10 K) with decreasing temperature. All thetemperatures where these resistivity anomalies appear for 5are lower than those for 3. This appears to be due to the‘‘chemical pressure’’ effect, which is consistent with theresistivity measurement under hydrostatic pressure. At 4.2 K,the resistivity anisotropy between the parallel and perpendi-cular directions to the c-axis is more than three orders ofmagnitude, suggesting strong 2D character for this system.

The absolute values for the in-plane resistivity (Ia and Ib),however, are larger than those of 3, although both compoundshave essentially the same Fermi surface. This would be due tothe cation disorder in 5.

The first differential curves of the electrical resistivities for 3are shown in Fig. 20. All the inflection points in the threedirections appear at the same temperatures in each axisdirection, around 46, 102 and 175 K. This means that theresistivity for each axis has the same tendency towardstemperature dependence and the electronic states are coupledto one another.

In order to further investigate the transport properties whereitinerant electrons and localized spins are associated, theelectrical resistivity was measured under hydrostatic pressureor uniaxial strain. Fig. 21 shows the electrical resistivity (rIa,rIb and rIc) for 3 under hydrostatic pressure. For the a- andb- axis directions, an application of pressure lowers thetemperatures where the broad maximum and minimumappear (Fig. 21). These effects correspond to the chemicalpressure effect in 5. At 15 kbar, this system exhibits normal 2Dmetallic behavior in the temperature region of about 80–300 K.On the other hand, for the c-axis direction, the broadmaximum is diminished and the resistivity increases with anincrease of the applied pressure. Although there are differ-ences in the high temperature region, there is a commonfeature at low temperature. In each case, the resistivityincreases with decreasing temperature and can be fit with2logT. This type of temperature dependence is typical for theKondo effect,21 which means that the itinerant and localizedelectrons form a singlet bound state to exhibit a heaviereffective mass. These behaviors are enhanced by the applica-tion of pressure (Fig. 21). This suggests that the interactionsbetween localized (Layer I) and conduction (Layer II) electronsare increased by the pressure and result in an enhancement ofthe Kondo temperature.

Fig. 22 shows the electrical resistivity (rIa and rIb) for 3under uniaxial strain along the interlayer direction (PIc). In alow pressure region, each resistivity for the a- and b-axesexhibits normal metallic behavior without any observedanomalies under hydrostatic pressure. With further applica-tion of the strain, on the other hand, the resistivity increaseswith decreasing temperature and starts to obey 2logTdependence in the low temperature region.

In the pp–d system, the difference in energy levels betweenthe itinerant electron (the HOMO of the donor molecule) andthe localized electron (d orbitals) is about 1 eV,22 while in thepresent system, the conduction and magnetic electrons shouldhave almost the same energy levels because both are electrons

Fig. 19 Anisotropic temperature dependence of the electrical resistivity for 3(left) and 5 (right).

Fig. 20 Anisotropic temperature dependence of the first differential curve ofthe electrical resistivity for the a- (left) b- (center) and c-axis (right) for 3.

Fig. 18 Temperature dependence of the electrical resistivity for 1, 2 and 4.


CrystEngComm Paper

from the same molecular orbitals. Moreover, for pp-electrons,localized and itinerant states are convertible to each other.These situations can provide a strong interaction betweenconduction and magnetic electrons, potentially leading to a

new hybrid phenomena associated with conduction andmagnetism. Indeed, the physical measurements, such asSchubnikov–de Haas oscillation9 and NMR,23 as well as thepresent results suggest interplay of the conduction andmagnetic electrons.

Magnetism. For 3 and 5, magnetic susceptibility (x) isplotted against temperature in Fig. 23. Both of them showparamagnetic behavior at high temperature, which arises fromthe Ni(dmit)2 anions because the cation (Me-3,5-DIP)+ isdiamagnetic (closed shell). The x–T curve in the temperaturerange of 100–300 K can be fitted by a linear combination of theCurrie–Weiss term (xCW = C/(T 2 h); C = 0.375 emu K mol21

(fixed)) and a constant, xconst, for the Pauli paramagneticcontribution from the conduction electrons in Layer II, xtotal =(xCW + xconst)/2. For 3, the Weiss temperature, h, is estimated tobe 25.3 K, indicating that the spins interact antiferromagne-tically with each other. An estimated xconst of 7.2 6 1024 emumol21 is close to that of conventional Pauli paramagneticmolecular metals (ca. 5 6 1024 emu mol21). The x–T curve hasa broad maximum at 20 K, which is due to short-rangeordering of antiferromagnetic spins, and anomalies at 6 and10 K (Fig. 23, inset of the left panel). However, no anisotropy ofx can be observed. The 13C NMR23 and magnetic torque8

measurements indicate that no magnetic transition occurs atthese temperatures.

On the other hand, h and xconst for 5 are estimated to be210.7 K and 7.5 6 1024 emu mol21, respectively. The increaseof h is due to the enhancement of the interdimer interactions.Although the antiferromagnetic interactions are enhanced, x

steeply increases below 10 K (Fig. 23, right panel). The M–Hcurve indicates free spin like behavior below this temperature(Fig. 24), but the absolute value of M is much smaller than thatof an ideal free spin of 1/2 (saturated at ~5.5 6 103 emumol21). Although the cation disorder appears to prevent theantiferromagnetic ordering and lead to the generation ofpartially free spins, the mechanism remains an open question.

X-ray diffraction at low temperature. X-ray diffractionpatterns for 3 in a low temperature region are shown inFig. 25. On decreasing the temperature, superlattice spots withquasi-four-fold periodicity gradually appear from 170 K. Asmarked by the red circles, they correspond to the a* + b* anda* 2 b* directions. A further decrease of the temperature leadsto modulation of the period (from four-fold to three-fold, as

Fig. 22 Electrical resistivity of 3 under uniaxial strain (P//c) for the a- (upper twopanels) and b-axis (lower two panels) with logr-T (left) and r-logT (right) scales,respectively. The dashed lines represent the fitting curve of 2logT.

Fig. 23 Magnetic susceptibility for 3 (a) and 5 (b). The solid lines represent theCurrie–Weiss model with an additional constant term, xconst. The data below 30K (3) and 40 K (5) are expanded in the inset, where the arrows mark anomaliesat 6 and 10 K for 3 and 10 K for 5.

Fig. 21 Electrical resistivity of 3 under hydrostatic pressure for the a- (upper twopanels), b- (middle two panels) and c-axis (lower two panels) with logr-T (left)and r-logT (right) scales, respectively. The dashed lines represent the fittingcurve of 2logT.


Paper CrystEngComm

shown in Fig. 25). Quasi-three-fold superlattice spots appearbelow 110 K. Compound 5, on the other hand, provides adifferent diffraction pattern. On decreasing the temperature,quasi-four-fold superlattice spots along the a* + b* and a* 2 b*directions gradually appear from about 235 K and they remaindown to 6 K. The three-fold spots do not appear. Since the half-width at half maximum (HWHM) of these superlattice spotsare much wider than those of normal spots, they are not Braggspots but diffuse spots. This indicates that the first-orderstructural phase transition does not occur. This is alsosupported by the fact that the resistivity and magneticsusceptibility do not exhibit hysteresis in the correspondingtemperature region. Although these structure modulationsmay provide small perturbations to the system, the electricalstructure seems not to be affected by these small structuralfluctuations because the section area of the Fermi surface,determined by quantum oscillation methods, matches wellwith the calculated value of 50%.8

Conclusions

Preparation, crystal structure analysis and physical propertymeasurements have been done to understand the role ofhalogen and hydrogen bonding in forming the uniquestructures that resulted in a coexistence of conducting andmagnetic properties. It is known that many organic–inorganic

hybrid materials form halogen bond.24 In crystals 3 and 5,however, almost all types of chemical bonds are used toconstruct the crystal structure: covalent, metal-coordination,ionic, van der Waals, metallic, hydrogen, chalcogen andhalogen bonds are included in these crystals.

The most interesting point in these materials is theinterplay between their conduction and magnetic properties.Since one kind of molecule, Ni(dmit)2, constitutes two kinds oflayers with completely different properties, coexistence as wellas interplay between the itinerant and localized electrons arerealized in a single molecular crystal. Such an interplay hasresulted in the formation of a Kondo singlet state, the researchof which is attracting recent attention.

Experimental

Synthesis

All syntheses were performed under Ar. THF was freshlydistilled and other solvents were used without furtherpurification. (nBu4N)[Ni(dmit)2] was synthesized according tothe literature method.25

3,5-Diiodopyridine. A mixture of 3,5-dibromopyridine (5.0 g,21.1 mmol), KI (114.7 g, 691 mmol) and CuI (10.5 g, 55.1mmol) was placed in a three-necked flask (1 L). DMF (450 mL)was added and the flask was heated at 180 uC. After 1 week, theflask was allowed to cool to room temperature and the reactionmixture was poured onto ice-water. The precipitate wascollected on a G4 glass filter by filtration and dissolved inCH2Cl2–pyridine (5 : 1). A saturated Na–EDTA solution wasadded to remove any Cu ions from the organic layer and themixture was stirred overnight. The organic layer was separatedand dried over MgSO4. After filtration, the residue was purifiedwith column chromatography on silica gel with CH2Cl2. Thesame procedure was repeated three times until two bromineatoms were completely replaced by iodine atoms. Finalpurification yielded 3,5-diiodopyridine (5.42 g, 77.6%) as awhite powder. MS (EI): 331.

5-Bromo-3-iodopyridine26. A THF solution of iPrMgCl (5.1mL, 10.2 mmol) was added dropwise to a solution of 3,5-dibromopyridine (2.01 g, 8.55 mmol) dissolved in THF (20 mL)in a three-necked flask (300 mL) at 0 uC. After stirring for 15min whilst maintaining the temperature, I2 (2.54 mg, 10.2mmol) dissolved in THF (15 mL) was added slowly. Thereaction mixture was then allowed to warm to roomtemperature and stirred continuously for 1 h. A 10% NH4Clsolution was poured onto the reaction mixture and theprecipitate was extracted with Et2O. The organic layer waswashed with a 10% Na2SO3 solution and brine and dried overNa2SO4. After removing the solvent, the residue was purifiedwith column chromatography on silica gel with CH2Cl2–n-hexane (1 : 1). 5-Bromo-3-iodopyridine (2.43 g, 67.5%) wasafforded as a white powder. MS (EI): 283, 285.

(Me-3,5-DIP)BF4. A mixture of 3,5-diiodopyridine (2.20 g,14.9 mmol) and Me3OBF4 (2.15 g, 6.20 mmol) was placed in athree-necked flask (100 mL). Toluene (100 mL) was added andthe solution was heated at 100 uC. After 7 hours, EtOH (20 mL)was added at room temperature and the mixture was filtered

Fig. 25 X-ray diffraction patterns for 3 at 160 K (left panel) and 50 K (rightpanel).

Fig. 24 M–H curves at 2, 4, 6 and 8 K for 5.


CrystEngComm Paper

thorough a G4 glass filter. The precipitate was washed withbenzene and dissolved in MeCN. After removal of MeCN underreduced pressure, the residue was recrystallized from EtOH–CH3CN–n-hexane to yield (Me-3,5-DIP)BF4 (0.92 g, 34.3%) ascolorless crystals. MS (ESI+): 346.

(Me-3,5-DBP)BF4. An MeCN solution of 3,5-dibromopyridinewas dropped onto a Me3OBF4 in MeCN in a three-necked flaskand the solution was stirred at room temperature. The solventwas then removed to a small volume in vacuo and then excessEt2O was added. The precipitate was recrystallized fromMeCN–Et2O to yield (Me-3,5-DBP)BF4 as colorless crystals(93.9% yield). (Me-3,5-BIP)BF4, was obtained by the samemethod.

(cation)x[Ni(dmit)2]y. Anion radical salts,(cation)x[Ni(dmit)2]y, were obtained by the galvanostaticelectrolysis of the corresponding solution of(nBu4N)[Ni(dmit)2] and the (cation)BF4 (supporting electrolyte)with Pt electrodes (1 mm diameter) under Ar. The electro-chemical crystallization afforded five salts: (Me-3,5-DBP)[Ni(dmit)2]2 (1), (Me-3,5-DIP)[Ni(dmit)2]3 (2), a-(Me-3,5-DIP)[Ni(dmit)2]2 (3), b-(Me-3,5-DIP)[Ni(dmit)2]2 (4) and (Me-3,5-BIP)[Ni(dmit)2]2 (5). Current switching (applied for compounds3 and 5) was performed according to the literature method.27

Crystal structure

Crystal data were collected at room temperature using anautomatic CCD diffractometer (RIGAKU/MSC Mercury) for 1, 2and 4 or an imaging plate detector (RIGAKU/R-AXIS) for 3 and5 with graphite-monochromated Mo–Ka (l = 0.71070 Å)radiation. The crystal structures were solved by the directmethod and refined by the full-matrix least-squares procedure.Non-hydrogen atoms were refined anisotropically and hydro-gen atoms were placed at geometrically calculated positions.For all the salts, calculations were performed using the teXsancrystallographic software package from Molecular StructureCo.28 Application of the numerical absorption correction

based on the azimuthal scans of several reflections tocompounds 1–4 yielded transmission factors of 0.5862–0.9453, 0.5685–0.8541, 0.432–0.896, and 0.316–0.452, respec-tively. An empirical absorption correction based on theazimuthal scans of several reflections of compound 5 yieldedtransmission factors of 0.337–0.647. The crystallographic datafor 1–5 are summarized in Table 4 (CCDC 910061–910062,915357, 645114, and 910623).

Band calculation

Overlap integrals (S) between the LUMOs of adjacent Ni(dmit)2

molecules were calculated on the basis of the obtained crystalstructures using the extended Huckel MO method andsemiempirical parameters for Slater-type atomic orbitals.29

Double-f orbitals were used for Ni. Band structures wereobtained by tight-binding band calculations using the transferintegrals (t) with an approximation of t y eS, where e is aconstant corresponding to the order of the energy level of theHOMO (y210 eV).

Conductivity measurements

Measurements of the temperature dependence of the electricalresistivity under ambient pressure were carried out by astandard four-probe dc method. Gold leads (15 mm diameter)were attached to the sample with carbon paste. For theresistivity measurement under hydrostatic pressure, thesample was put in a Teflon capsule filled with a pressuremedium (DN-oil 7373) and then the capsule was set in a clamptype pressure cell made of BeCu. For the resistivity measure-ment under uniaxial strain, the sample was immersed in anepoxy liquid (stycast 1266/A). Then the epoxy containing thesample was molded into a cylinder and placed in the cylinderof a clamp-type pressure cell. The relation between thepressure applied on the epoxy and strain was checked with astrain gauge. For the resistivity measurement for 3, aconventional four-probe dc method was performed. Thecurrent was applied along the a-, b- and c-axis directions for

Table 4 Crystallographic data for compounds 1–5

Crystals 1 2 3 4 5

Empirical formula C18H6Br2NNi2S20 C24H6I2NNi3S30 C18H6I2NNi2S20 C18H6I2NNi2S20 C18H6BrINNi2S20

Formula weight 1154.66 1700.03 1248.66 1248.66 1201.66Crystal system Triclinic Monoclinic Monoclinic Triclinic MonoclinicSpace group P1 C2/c C2/c P1 C2/ca/Å 7.632(5) 35.368(7) 14.3470(7) 7.427(2) 14.218(4)b/Å 11.716(7) 11.088(2) 6.4710(3) 11.680(3) 6.436(1)c/Å 20.19(1) 13.129(3) 76.556(3) 21.199(7) 76.15(2)a/u 98.87(1) — — 81.935(10) —b/u 93.400(8) 103.317(3) 92.998(4) 96.41(1) 93.836(9)c/u 101.19(1) — — 99.94(1) —V/Å3 1742.4 5010.2 7096.2 1786.2 6952.7Z 2 4 8 2 8rcalc/g cm23 2.201 2.254 2.337 2.322 2.297m/cm21 4.600 3.627 3.998 3.972 4.348Measured reflections 7856 5731 8552 7748 7838Independent reflections 5632 3981 6205 5857 5884s limit 2.0 2.0 2.0 2.0 2.0GOF 1.060 0.976 1.048 1.075 1.121Ra, Rwb 0.0533, 0.1805 0.0389, 0.1034 0.0635, 0.1857 0.0558, 0.1620 0.0549, 0.1734

a R = S||Fo| 2 |Fc||/S|Fo|. b Rw= [Sw(|Fo| 2 |Fc|)2/SwFo2]1/2.


Paper CrystEngComm

the measurements under hydrostatic pressure and along the a-and b-axis directions for the measurements under uniaxialstrain.

Magnetic measurements

The temperature-dependent magnetic susceptibility of thepolycrystalline samples was measured using a SQUID magnet-ometer (Quantum Design MPMS XL) over a temperature rangeof 2.0–300 K under an applied magnetic field of 1 T. Thediamagnetic contributions of the cations and Ni(dmit)2 anionswere subtracted with Pascal’s Law.

Low temperature X-ray

Crystal data for 3 and 5 were collected using an imaging platedetector (RIGAKU/R-AXIS) with graphite-monochromated Mo–Ka (l = 0.71070 Å) radiation. The temperature was controlledwith a temperature controller (Lake Shore/Model 331) in atemperature range of 6–300 K.

Acknowledgements

This work was partially supported by Grant-in-Aid for ScientificResearch (S) (No.22224006) from the Japan Society for thePromotion of Science (JSPS).

Notes and references

1 (a) A. Ando, R. Antos, S. E. Barnes, G. E. W. Bauer,A. Brataas, A. Hirohata, T. Jungwirth, P. J. Kelly, T. Kimura,S. Maekawa, S. Murakami, N. Nagaosa, J. Nitta, T. Ono,Y. Otani, E. Saitoh, Y. Sakuraba, J. Sinova, Y. Suzuki,S. Takahashi, Y. Tserkovnyak, K. Uchida, S. O. Valenzuela,J. Wunderlich, T. Yokoyama, L. P. Zarbo and S. Zhang, SpinCurrent (Semiconductor Science and Technology), ed. S.Maekawa, S. O. Valenzuela, E. Saitoh and T. Kimura,Oxford Univ Press, Oxford, 2012; (b) V. A. Dediu, L.E. Hueso, I. Bergenti and C. Taliani, Nat. Mater., 2009, 8,707.

2 H. Tsunetsugu, M. Sigrist and K. Ueda, Rev. Mod. Phys.,1997, 69, 809.

3 N. Plakida, High-Temperature Cuprate Superconductors,Springer Series in Solid-State Sciences, Heidelberg, 2010,vol. 166.

4 (a) Y. Kosaka, H. M. Yamamoto, A. Nakao, M. Tamura andR. Kato, J. Am. Chem. Soc., 2007, 129, 3054; (b) H.M. Yamamoto, Y. Kosaka, R. Maeda, J. Yamaura,A. Nakao, T. Nakamura and R. Kato, ACS Nano, 2008, 2,143.

5 (a) R. Kato, Chem. Rev., 2004, 104, 5319.6 T. Miyazaki and T. Ohno, Phys. Rev. B: Condens. Matter,

1999, 59, R5269.7 M. Tamura and R. Kato, J. Phys. Soc. Jpn., 2004, 73,

3108–3110.8 (a) K. Hazama, Y. Takahide, M. Kimata, T. Terashima,

S. Uji, Y. Kosaka, H. M. Yamamoto and R. Kato, J. Phys.:Conf. Ser., 2010, 150, 022025; (b) K. Hazama, S. Uji,Y. Takahide, M. Kimata, H. Satsukawa, A. Harada,

T. Terashima, Y. Kosaka, H. M. Yamamoto and R. Kato,Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 83, 165129.

9 H. Kim, A. Kobayashi, Y. Sasaki, R. Kato and H. Kobayashi,Chem. Lett., 1987, 1799–1802.

10 L. Valade, M. Bousseau, A. Gleizes and P. Cassoux, Chem.Commun., 1983, 110.

11 T. Nakamura, A. E. Underhill, A. T. Coomber, R. H. Friend,H. Tajima, A. Kobayashi and H. Kobayashi, Inorg. Chem.,1995, 34, 870.

12 (a) P. Politzer, J. S. Murraya and T. Clark, Phys. Chem.Chem. Phys., 2010, 12, 7748; (b) G. Espallargas, L. Brammerand P. Sherwood, Angew. Chem., Int. Ed., 2006, 45, 435; (c)F. F. Awwadi, R. D. Willett, K. Peterson and B. Twamley, J.Phys. Chem. A, 2007, 111, 2319.

13 V. Cody, J. Mol. Struct., 1984, 112, 189.14 Y. Tokura, A. Urushibara, Y. Moritomo, T. Arima, A. Asamitsu,

G. Kido and N. Furukawa, J. Phys. Soc. Jpn., 1994, 63, 3931.15 (a) M. A. Ruderman and C. Kittel, Phys. Rev., 1954, 96, 99;

(b) T. Kasuya, Prog. Theor. Phys., 1956, 16, 45; (c)K. Yoshida, Phys. Rev., 1957, 106, 893; (d) F. Steglich,J. Aarts, C. D. Bredl, W. Lieke, D. Meschede, W. Franz andH. Schafer, Phys. Rev. Lett., 1979, 43, 1892.

16 M. Kurmoo, A. W. Graham, P. Day, S. J. Coles, M. B. Hursthouse,J. L. Caulfield, J. Singleton, F. L. Pratt, W. Hayes, L. Ducasse andP. Guionneau, J. Am. Chem. Soc., 1995, 117, 12209.

17 E. Coronado, J. R. Galan-Mascaros, C. J. Gomez-Garcıa andV. N. Laukhin, Nature, 2000, 408, 447.

18 (a) S. Uji, H. Shinagawa, T. Terashima, T. Yakabe, Y. Terai,M. Tokumoto, A. Kobayashi, H. Tanaka and H. Kobayashi,Nature, 2001, 410, 908; (b) S. Uji, H. Kobayashi, L. Balicasand J. S. Brooks, Adv. Mater., 2002, 14, 243–245.

19 (a) E. Ojima, H. Fujiwara, K. Kato, H. Kobayashi,H. Tanaka, A. Kobayashi, M. Tokumoto and P. Cassoux,J. Am. Chem. Soc., 1999, 121, 5581; (b) T. Konoike, S. Uji,T. Terashima, M. Nishimura, S. Yasuzuka, K. Enomoto,H. Fujiwara, B. Zhang and H. Kobayashi, Phys. Rev. B:Condens. Matter Mater. Phys., 2004, 70, 094514.

20 (a) R. T. Henriques, L. Alcacer, J. P. Pouget and D. Jerome, J.Phys. C: Solid State Phys., 1984, 17, 5197; (b) V. d. Gama, R.T. Henriques and M. Almeida, J. Phys. Chem., 1994, 98, 997.

21 J. Kondo, Prog. Theor. Phys., 1964, 32, 37–49.22 T. Mori and M. Katsuhara, J. Phys. Soc. Jpn., 2002, 71,

826–844.23 S. Fujiyama, A. Shitade, K. Kanoda, Y. Kosaka, H.

M. Yamamoto and R. Kato, Phys. Rev. B: Condens. MatterMater. Phys., 2008, 77, 060403.

24 (a) F. Zordan, L. Brammer and P. Sherwood, J. Am. Chem.Soc., 2005, 127, 5979; (b) R. D. Willett, F. Awwadi andR. Butcher, Cryst. Growth Des., 2003, 3, 301.

25 G. Steimecke, H. J. Sieler, R. Krimse and E. Hoyer,Phosphorus Sulfur Relat. Elem., 1979, 7, 49.

26 F. Cottet and M. Schlosser, Eur. J. Org. Chem., 2002, 327.27 S. Hunig, M. Kemmer, H. Meixner, K. Sinzger, H. Wenner,

T. Bauer, E. Tillmanns, F. R. Lux, M. Hollstein, H.-G. Gross,U. Langohr, H.-P. Werner, J. U. von Schutz and H.-C. Wolf,Eur. J. Inorg. Chem., 1999, 1999, 899.

28 teXsan: Crystal Structure Analysis Package, MolecularStructure Co., The Woodlands, TX, 1985 and 1999.

29 (a) A. J. Berlinsky, J. F. Carolan and L. Weiler, Solid StateCommun., 1974, 15, 795; (b) R. H. Summerville and R.J. Hoffman, J. Am. Chem. Soc., 1976, 98, 7240.


CrystEngComm Paper

crystengcomm - riken · this paper, we report some supramolecular ni(dmit) 2 radical anion salts...

Documents