Giant Dielectric Anisotropy and Relaxor Ferroelectricity Induced by

May 9, 2008 - Faculty of Physics, Adam Mickiewicz UniVersity, Umultowska 85, ... nanoregions.1,2 However, the performance of ferroelectric ... tors, a...
0 downloads 0 Views 1MB Size
J. Phys. Chem. B 2008, 112, 6779–6785

6779

Giant Dielectric Anisotropy and Relaxor Ferroelectricity Induced by Proton Transfers in NH+ · · · N-Bonded Supramolecular Aggregates Marek Szafran´ski*,† and Andrzej Katrusiak‡ Faculty of Physics, Adam Mickiewicz UniVersity, Umultowska 85, 61-614 Poznan´, Poland, and Faculty of Chemistry, Adam Mickiewicz UniVersity, Grunwaldzka 6, 60-780 Poznan´, Poland ReceiVed: February 6, 2008

A huge dielectric effect has been observed in a pure and water-soluble hydrogen-bonded organic crystal, 1,4-diazabicyclo[2.2.2]octane hydroiodide [C6H13N2]+ · I- (dabcoHI). In this structure, the dabco cations are NH+ · · · N bonded into linear aggregates, where the protons are disordered at two nitrogen atoms and the crystal acquires the symmetry of space group P6jm2. This nonpolar crystal exhibits a barely temperaturedependent dielectric constant exceeding 1000 at ambient conditions. The dielectric response is extremely anisotropic, more than 2 orders of magnitude higher along the linear hydrogen bonded chains than in perpendicular directions. The physics underlying this effect originates from proton transfers in the NH+ · · · N bonds, leading to disproportionation defects and formation of polar nanodomains, which, on the macroscopic scale, results in one-dimensional relaxor ferroelectricity. Such properties are unprecedented for the materials with hydrogen bonds highly polarizable due to proton disorder. The proton disordering in dabcoHI is analogous to this in H2O ice, where the hydrogen bonds remain disordered until the lowest temperature. Introduction Miniaturization of electronic devices drives the research on materials exhibiting dielectric constants ε′ greater than 1000, mainly on ferroelectrics showing a long-range polar order and on ferroelectric relaxors where the polar order is confined to nanoregions.1,2 However, the performance of ferroelectric materials is affected by the undesirable Curie–Weiss-type temperature dependence of ε′. In the past decade, there have been reports of several systems for which giant values of electric permittivity persist over broad temperature ranges.3–5 Most of these findings were subsequently shown to be extrinsic effects arising from metal-dielectric contacts, grain boundaries, or layer interfaces.6,7 Thus, ferroelectric materials, where dielectric constants are related intrinsically to the highly polarizable crystal lattices, still remain one of the most attractive classes of materials for microelectronics devices,2 despite the fact that coherent arrangements of electric dipoles (either permanent or due to ionic displacements) are very temperature sensitive. Temperature-independent behavior has been predicted only for the systems with an electronic mechanism underlying ferroelectricity.8 However, so far such, ferroelectric ordering arising from electron correlations has been reported for LuFe2O4 only.9 Of particular interest are the so-called ferroelectric relaxors exhibiting, apart from high dielectric constants, extremely high field-induced strains,10–12 the properties required in small, precise, and powerful manipulators operating at high frequencies with quick response times. In contrast to classical ferroelectrics, relaxor ferroelectrics are characterized by13 (i) a rounded and strongly frequency-dependent dielectric anomaly migrating toward lower temperatures with decreasing frequency of the electric measuring field; (ii) absence of macroscopic spontaneous polarization; and (iii) no crystal-symmetry breaking when cooled to low temperatures. It is commonly accepted that the unusual * Towhomcorrespondenceshouldbeaddressed.E-mail:[email protected]. † Faculty of Physics. ‡ Faculty of Chemistry. E-mail: [email protected].

properties of relaxors originate from the correlated polar regions submerged in the nonpolar crystal matrixes. In mixed perovskites, which are prototypical relaxor ferroelectrics, the local polar order is routinely achieved by a chemical substitution.13 A chemical doping is also applied to induce relaxor behavior in charge-transfer complexes,14 while in polymers, a long-range ferroelectric order can be suppressed, in favor of a local polar order, by irradiation.11 Both approaches apparently rely on an artificially introduced compositional disorder, but there has also been evidence of relaxor properties in pure systems.15–17 Practically all of the ferroelectric materials used in the electronic technology are inorganic perovskite-type compounds, most of them containing lead. For example, of particular importance for practical applications is the PbZrxTi1-xO3 solid solution widely used in piezoelectric applications, microcapacitors, and memory devices. At the same time, research is ongoing for organic electronics. Therefore, new organic functional materials are strongly desirable, which would be environment friendly, easily disposable, and processed at low cost.18,19 The envisaged development of electronics based totally on organic compounds is currently a challenge for the research on highdielectric-constant organic materials,20 which could replace their lead-containing inorganic counterparts. The organic compounds can be designed and engineered into crystal structures of desired properties by modeling the shape and interactions of molecules, and the interactions most often considered and applied for this purpose are hydrogen bonds. The discovery of ferroelectricity in KH2PO4 (KDP) in 19352,21 triggered a general interest in hydrogen-bonded ferroelectrics. These materials, where paraelectric-ferroelectric phase transition is driven by the dynamics of protons in hydrogen bonds, are described as KDP-type ferroelectrics. Presently, the group of KDP-type crystals is structurally very diverse and has representatives among inorganic, organic, and hybrid compounds. The proton dynamics is also responsible for the properties of the so-called proton glasses,22,23 produced of two structurally proximate compounds exhibiting competing ferro-

10.1021/jp801106m CCC: $40.75  2008 American Chemical Society Published on Web 05/09/2008

6780 J. Phys. Chem. B, Vol. 112, No. 22, 2008

Szafran´ski and Katrusiak submerged in this antiparallel matrix. The entanglement of these elements within this structurally simple crystal results in a complex dielectric system exhibiting weak relaxor-like properties apart from the normal ferroelectricity.16 Here, we describe the crystal of dabco hydroiodide, [C6H13N2]+ I- (dabcoHI). The crystal consists of hydrogenbonded chains, analogous to those in dabco ferroelectrics; however, its structure is nonpolar. The protons in NH+ · · · N bonds are inherently disordered in a wide temperature range; however, despite the nonpolar crystal structure, the crystal exhibits very strong and strictly one-dimensional relaxor behavior leading to a giant dielectric response. With respect to the disordering of protons to the lowest temperature and high polarizability of hydrogen bonds, dabcoHI can be considered as an NH+ · · · N-bonded one-dimensional analogue of OH · · · Obonded H2O ice Ih. Experimental Section

Figure 1. The crystal habit and structure of dabcoHI: (a) an ionic pair of the dabcoH+ cation and I- anion as present in the structure at 120 K, with the thermal ellipsoids drawn at the 50% probability level (the atomic labels of the symmetry-independent part of the unit cell have been indicated only); (b) the crystal morphology with indicated direction of chains of the hydrogen-bonded cations in the microscopic structure along [001]; and (c) the space-filling model of the crystal structure of dabcoHI at 120 K with both half-occupied sites of the disordered protons indicated. In drawings (a) and (b) the disordered proton has been drawn only in one site.

electric and antiferroelectric long-range interactions. Most of the studies in the field of H-bonded ferroelectrics and related compounds are concerned with the materials linked by short OH · · · O hydrogen bonds. Only recently, NH · · · N hydrogenbonded ferroelectric analogues of KDP-type structures were reported.24,25 In our previous papers, we have shown that onedimensional NH+ · · · N-bonded aggregates are essential for the properties of a new group of ferroelectrics based on organic 1,4-diazabicyclo[2.2.2]octane (abbreviated dabco; see Figure 1). Presently, this group consists of three crystals of a general formula, [C6H13N2]+A-, where A- stands for tetrahedral anions BF4-, ClO4-, or ReO4-. While in ferroelectric phases of dabcoHBF4 and dabcoHClO4 the polarized polycationic chains are arranged antiparallel, in dabcoHReO4, they are ordered in a parallel manner. Hence, this unique structure of dabcoHReO4 results in a very large spontaneous polarization exceeding 17 µC/cm2, and the polarization reversal requires a concerted proton transfer within the chains. The unique one-dimensional NH+ · · · N-bonded structure of the dabco crystals investigated allowed new dielectric properties to be observed for these materials. In dabcoHBF4, an unusual dielectric response was observed along the hydrogen-bonded chains, that is, perpendicular to the ferroelectric axis.16 Its structural origin could not be reconciled with the crystal symmetry. The dabcoHBF4 structure comprises (i) a long-range ferroelectric order along [010] induced by the ionic displacements; (ii) the bulk matrix of antiparallel-polarized H-bonded chains along [001]; and (iii) the short-range polar regions

Sample Preparation. The dabcoHI compound was synthesized by dissolving stoichiometric amounts of 1,4 diazabicyclo[2.2.2]octane and hydriodic acid in water. The substance obtained was purified by repeated crystallizations. A slow evaporation of the saturated aqueous solution at 298 K yielded colorless pillar-shaped crystals with the largest dimension of 5–20 mm along [001]. The cross sections of the largest crystals varied between 0.5 and 2 mm2. For dielectric measurements, the 0.2–0.5 mm thick crystal plates were cut perpendicular to [001] or [100]. Two different types of electrodes were used, a silver paste deposited by painting on the polished surfaces or gold deposited by evaporation in vacuum. Dielectric, Pyroelectric, and Calorimetric Measurements. Complex dielectric permittivity ε ) ε′ - iε′′ was measured with a Hewlett-Packard 4192A impedance analyzer in the frequency range from 1 kHz to 10 MHz at the ac measuring field of 2–5 V/cm. A closed-cycle cooler CCC1204 (Oxford Instruments) was used for the low-temperature studies. The polarization of the crystal was derived from the pyroelectric charge measured as a function of temperature with a Keithley 6514 electrometer. The samples were cooled under bias electric field to about 90 K, then the field was released, and the pyroelectric charge was measured upon heating the crystal at a rate of 2 K/min. Calorimetric measurements were carried out in the temperature range from 95 to 400 K using a differential scanning calorimeter Q 2000 (TA Instruments). The sample was heated/ cooled at a rate of 10 K/min. X-Ray Diffraction Analyses. The crystals selected for the single-crystal X-ray analysis were small and had regular dimensions; to minimize the effect of crystal absorption, they were hexagonal pillars with truncated vertices and edges. A KUMA KM-4 four-circle diffractometer, equipped with a sealed Mo X-ray tube and a graphite monochromator, was used for structural studies. The temperature was controlled within 0.1 K with a stream of gaseous nitrogen from an Oxford Cryosystem attachment. The data were collected in the bisecting mode and corrected for the Lorentz and polarization effects and for the absorption, consistent with the Ψ scans of individual reflections. The structures were solved by the Patterson method and refined with full-matrix least-squares on reflections intensities (F2’s) using the program SHELXL-97.26 The details of the experiment and crystal structure parameters are listed in Table 1. The crystal data for 120, 300, and 370 K have been deposited in the form of crystal information files (CIF) in the Cambridge Crystal-

Giant Dielectric Anisotropy and Relaxor Ferroelectricity

J. Phys. Chem. B, Vol. 112, No. 22, 2008 6781

TABLE 1: Crystal Data and Structure Refinement for 1,4-Diazabicyclo[2.2.2]octane Hydroiodide at 120, 300, and 370 K temperature [K] empirical formula formula weight wavelength [Å] crystal system space group unit cell [Å]: a, c volume [Å3] Z, Calculated density [g/cm3] absorption coefficient [mm-1] F(000) crystal size [mm] θ range for data collection [°] limiting indices Reflections collected/unique completeness to θ absorption correction refinement method full-matrix least-squares on F2 data/restraints/parameters goodness-of-fit on F2 final indices R1/R2 [I > 2σI] R1/R2 (all data) absolute structure parameter largest diff. peak and hole [e · Å-3]

120 [C6H13N2]+ · I240.08 0.71073 hexagonal P6jm2 7.0339(10), 5.3292(11) 228.34(7) 1, 1.746 3.436 116 0.20 × 0.20 × 0.18 3.34 to 27.05 -8 e h e 7, -8 e k e 7, 0 ele6 590/232 [R(int) ) 0.0244] 100.0% to 27.05° empirical full-matrix least-squares on F2

300 [C6H13N2]+ · I240.08 0.71073 hexagonal P6jm2 7.0930(10), 5.3460(11) 232.93(7) 1, 1.712 3.369 116 0.20 × 0.20 × 0.18 3.32 to 30.07 -9 e h e 9, -9 e k e 9, 0 ele7 1370/309 [R(int) ) 0.0474] 100.0% to 30.07° empirical full-matrix least-squares on F2

370 [C6H13N2]+ · I240.08 0.71073 hexagonal P6jm2 7.1042(10), 5.3600(11) 234.27(7) 1, 1.702 3.349 116 0.20 × 0.20 × 0.18 3.31 to 26.82 -7 e h e 9, -8 e k e 0, -6 e l e 0 586/236 [R(int) ) 0.0305] 100.0% to 26.82° empirical full-matrix least-squares on F2

232/0/18 1.151 0.0107/0.0271 0.0107/0.0271 -0.08(5) 0.241 and -0.362

309/0/18 1.035 0.0153/0.0301 0.0238/0.0333 -0.06(5) 0.287 and -0.658

236/0/18 1.127 0.0194/0.0439 0.0201/0.0441 0.04(8) 0.477 and -0.204

TABLE 2: The NH+ · · · N Hydrogen Bond Dimensions [Å] in the dabcoHI Structure at 120, 300, and 370 Ka 120 K 300 K 370 K

N(1) · · · N(1b)

N(1)-H(1)

H(1) · · · N(1b)

2.800(8) 2.826(9) 2.842(11)

0.96(8) 0.84(9) 0.94(11)

1.84(8) 1.99(9) 1.91(11)

a Owing to the symmetry requirements, the N-H · · · N angle is equal to 180°. b Symmetry code of the transformed atom: x, y, z+1.

lographic Database Centre as supplementary publication No.’s CCDC 617798, CCDC 617799, and CCDC 617800, respectively (The atomic coordinates for these structuree have been deposited with the Cambridge Crystallographic Data Centre. The coordinates can be obtained, upon request, from the Director, Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge, CB2 1EZ, U.K.). The atomic coordinates, temperature parameters, and molecular dimensions have been collected in the Supporting Information. Results and Discussion Average Crystal Structure. The crystal structures determined at 120, 300, and 370 K are all hexagonal, space group P6j m2, with one anion and one cation in the unit cell, both located at D3h-symmetric sites. The cations are linked into linear chains along [001] (Figure 1). The dimensions of the hydrogen bond (Table 2) and, in particular, the N · · · N distances equal to or exceeding 2.8 Å, are consistent with the observation of asymmetric proton sites.24,25,27,28 The N · · · N distance in dabcoHI is comparable to the corresponding distances of 2.841(5) Å in dabcoHClO4 and 2.839(7) Å in dabcoHBF4, determined at 298 K.24,29 Despite the relatively long N · · · N distance, the protons observed in the difference Fourier maps are disordered in the NH+ · · · N bonds, as required by the crystal symmetry. This crystal symmetry and H+ disorder persist from 370 K to low temperatures, which was verified by the statistical tests comparing this model with lower symmetry models, for example, the model in space group P3m2. The topology of linear NH+ · · · N-bonded chains

(Figure 1) is identical to that in the dabco ferroelectrics, but the spherical iodine shape results in the hexagonal arrangement of the chains and suppression of a long-range polar order arising from the anion-cation interplay. Dielectric Response. Despite the nonpolar symmetry ruling out the long-range ferroelectric order, dabcoHI exhibits spectacular dielectric properties. The real part of the complex electric permittivity ε ) ε′ – iε′′, measured along two crystallographic directions [100] and [001] at different frequencies as a function of temperature, is shown in Figure 2a. Along [100], perpendicular to the NH+ · · · N-bonded chains, the ε′ magnitude of 6–7 over the whole temperature range and its negligible frequency dependence in the kilo- and megahertz frequency range are typical of nonpolar dielectrics without any polymorphic transformations. Completely different is the dielectric response of the crystal along the hydrogen bonds down [001]; the huge and weakly temperature-dependent ε′ persists over a broad temperature range, especially in the low-frequency range of the electric measuring fields. Another striking feature of these results is a large frequency dispersion both in the real (Figure 2a) and imaginary (Figure 2b) parts of ε. Several overlapping anomalies in the dielectric response along [001] are distinguishable. The strongly frequency-dependent anomalies between 350 and 300 K, below 200 K, and below 100 K cannot be ascribed to symmetry-breaking phase transitions. Their absence between 100 and 380 K was confirmed by differential scanning calorimetry (DSC) and the structural determinations. Thus, the unusual dielectric properties of dabcoHI can be associated with a short-range order rather than with a long-range one, and the giant dielectric anisotropy clearly indicates the prominent role of the NH+ · · · N bond transformations. The complexity of the ε′(T) and ε′′(T) dependencies indicates that several relaxation processes are involved and contribute to the dielectric response of the crystal. Apparently, the two distinct anomalies observed in Figure 2 below 200 K and below 100 K have essentially a similar character. The both shift toward lower temperatures with decreasing frequency. A reverse tendency is revealed for the anomaly that

6782 J. Phys. Chem. B, Vol. 112, No. 22, 2008

Szafran´ski and Katrusiak

Figure 3. The relaxor behavior of the dielectric anomaly measured for dabcoHI along [001]. The inset illustrates the fulfillment of the Vogel-Fulcher law, and the solid line represents the fit with the parameters described in the text.

Figure 2. Temperature and frequency dependencies of the real ε′ (a) and imaginary ε′′ (b) components of the complex electric permittivity measured along the NH+ · · · N hydrogen bonds ([001] direction). For comparison, ε′(T) measured at a frequency of 150 kHz in the direction perpendicular to the hydrogen bonds is shown; the values of ε′(T) are multiplied by 5 to avoid overlay of the plot with the abscissas axis.

appears in the high-T range, above room temperature. Thus, it seems that the low- and high-temperature phenomena are related, with different underlying physics. A closer analysis of the ε′′(T) dependencies shows that the amplitude of the absorption peak observed in the T range between 100 and 200 K increases with increasing frequency, which is characteristic of ferroelectric relaxors. A further evidence of relaxor features is provided in Figure 3. At higher frequencies, in the megahertz range, the high-temperature contribution to the electric permittivity is suppressed in favor of the processes observed at lower temperatures. As a result, a rounded peak emerges in the ε′(T) dependence. With decreasing frequency, f, the temperature of this peak maximum, Tm, shifts progressively toward lower temperatures, and the peak amplitude grows, just as in classical ferroelectric relaxors. The frequency dispersion of Tm can be modeled with the Vogel-Fulcher equation,30 fulfilled in many relaxor and glassy systems, τ ) τ0 exp[Ea/k(Tm - TVF)], where τ ) 1/(2πf), τ0 is a parameter related to the cutoff frequency of the spectrum of dipolar entities, Ea is the activation energy, k is the Boltzmann constant, and TVF is interpreted as the dipolar freezing temperature. The best fit of the Vogel-Fulcher equation to the experimental data, shown in the inset of Figure 3, yields the following parameters: Ea ) 477 K (0.041

Figure 4. Temperature-induced depolarization of the dabcoHI crystal along [001].

eV), τ0 ) 1.9 × 10-10 s, and TVF ) 143 K. The application of the experimental Vogel-Fulcher equation to ferroelectric relaxors was recently justified theoretically.31 Remnant Polarization along H-Bonded Chains. The experiments with an electric field applied along [001] testify that the field-cooled crystal acquires properties similar to a normal ferroelectric state, such as macroscopic polarization and related breaking of the nonpolar symmetry. The scale of the polar ordering was strongly dependent on the electrical and thermal treatments of the sample. The results shown in Figure 4 were obtained after poling the crystal at 342 K with an electric field of E ) (3.8 kVcm-1 and cooling it under bias to about 90 K. Upon heating the crystal, the decay of the remnant polarization Pr proceeded progressively with the temperature, but the depolarization curve showed some anomalies, which suggests that several mechanisms are involved in the polarization process, as observed in the temperature variation of the electric permittivity. The Pr value, deduced from these measurements, exceeds

Giant Dielectric Anisotropy and Relaxor Ferroelectricity 30 µCcm-2, and its magnitude is comparable to those observed in perovskite ferroelectric relaxors.32 The application of an electric field with opposite sense resulted in a reversal of polarization, as illustrated in the inset of Figure 4. This curve was measured in the four subsequent heating–cooling-heating cycles, with turning points situated at different temperatures. The results show that recooling from the point on the polarization decay path did not lead to a recovery of the low-temperature value of Pr, as one would expect for a normal ferroelectric state. This behavior is characteristic of the polarization mechanism in ferroelectric relaxors.32 It should be stressed that no traces of remnant polarization were detected when the biasing field was applied perpendicular to [001]. Microscopic Model of the Giant Dielectric Response. Structural models revealed by diffraction measurements are averaged over the crystal regions which are smaller or commensurate with the cohesion lengths of the applied radiation,33 up to few hundreds of nm depending on the neutron or X-ray source type. Thus, the X-ray diffraction observation of disordered protons, down to 120 K, in two sites of the long NH+ · · · N bond justifies the conclusion that polar and antipolar nanoregions have to coexist in the dabcoHI structure. The structure of dabcoHI also can be considered as a frustrated intermediate of two extreme cases of a long-range order of polarized chains, one of them being that of antiparallel arrangement realized in the ferroelectric dabcoHBF4 and dabcoHClO4 structures, while the second opposite counterpart is that of parallel chains in dabcoHReO4. The presence of local polar or antipolar regions requires that, apart from the dabcoH+ cations, the H+dabcoH+ dications and neutral molecules be formed at the boundaries within the chains (Figures 5 and 6). Randomly distributed defects connected with the H+ site are prone to form, owing to the low-energy barrier and highly polarizable NH+ · · · N bonds.27,34,35 There is no structural evidence that proton sites in the linear NH+ · · · N bonds are coupled to the ionic displacements. The thermal ellipsoids of the iodine anions are only slightly elongated along [001], testifying that only weak coupling between neighboring H bonds can be realized through electrostatic interactions between the protons and the anions. However, the I- anions are large and highly polarizable, and their electronic structure adjusting to the local fluctuations of electric field, generated by proton transfers, can considerably contribute to the crystal polarization. In the dabcoHI structure, each H bond is surrounded by three symmetry-equivalent anions (Figure 5a), which facilitates the coupling of the proton site with the lattice, but on the other hand, it would be most energetically unfavorable that all three H bonds around one anion be polarized in the same direction, as this would increase the displacement of the anion. The protons within the chain can also “communicate” via the changes in electronic structure of dabco induced by random H+ transfers. Weak correlations between the proton sites are crucial for the formation of a short-range order breaking the average symmetry of the crystal, as schematically shown in Figure 6, and for the observed relaxor behavior. The weak coupling between the structure and the proton site further implies that the lattice vibrations can move the borders of the regions by displacing the protons. The most likely to receive a proton are neutral dabco molecules, and the most likely to reject the proton are the dications. In either case, the proton transfer is likely to propagate in the way of the neutral (associated with a dabco molecule) or dicationic (associated with a dication) solitonic waves through the lattice. Such soliton-like excitations are highly probable in the linear H-bonded aggregates36 and can contribute to the

J. Phys. Chem. B, Vol. 112, No. 22, 2008 6783

Figure 5. Autostereograms of the dabcoHI structure: (a) the van der Waals model of one plane perpendicular to [001] of ionic pairs, in a configuration fulfilling the rule that none of the I- anions are surrounded by identically polarized hydrogen bonds; (b) the ball-and-stick model of a sheet of H-bonded chains perpendicular to [100] and to the plane of ions shown in (a).

dielectric properties of crystals. For example, the generation of protonic solitons was proposed for explaining dielectric response in a quasi-one-dimensional bisquaric acid.37 The solitons propagating along the chains in the dabcoHI crystal, and inherently associated with a transport of charge defects, are consistent with a considerable dielectric absorption (tangent of dielectric loss > 1) observed in the high-frequency range of the dielectric spectrum below 200 K (see Figure 2b). A collision of neutral and dicationic solitons may annihilate the two fronts and the region between them. When a cation receives or rejects a proton, a new polar defect built of the molecule and a dication is formed.

6784 J. Phys. Chem. B, Vol. 112, No. 22, 2008

Szafran´ski and Katrusiak positions exists between the chains. In the three-dimensional H-bonded network of H2O ice, the hydrogen-atom positions are coupled through the Bernal-Fowler ice rules.38 For this reason, short-range polar domains are not formed in ice, and its polarizibility relies only on reorientations of polar water molecules, while in dabcoHI, the formation of polar nanodomains is crucial for the dielectric response of unprecedented magnitude. The strictly anisotropic giant permittivity clearly related to the supramolecular structure of dabcoHI testifies to the paramount role of the polarizability of the NH+ · · · N bonds in this compound. Acknowledgment. This study was supported by Polish Ministry of Education and Science, Grant N202 14631/2707. Supporting Information Available: Tables of atomic coordinates, bond lengths and angles, thermal parameters, and the crystallographic information files for dabcoHI at 120, 300, and 370 K. This material is available free of charge via Internet at http://pubs.acs.org. References and Notes

Figure 6. A schematic representation of one layer of NH+ · · · N-bonded polycationic chains and hydroiodide anions in dabcoHI. The chains are built of neutral, monocationic (in two opposite polarizations), and dicationic dabco, listed and shown below the layer. Within the chains, oppositely polarized regions, built either of H+dabco or dabcoH+ monocations, are separated with neutral dabco at one end and with dabco2+ dications at the other. The polar regions of the NH+ · · · N bonds directed upward and downward are indicated with yellow and green backgrounds, respectively. The ethylene H atoms have been omitted for clarity.

Conclusions It has been shown that relaxor behavior can originate from the intrinsic properties of a pure substance, where the interplay of crystal symmetry with low disproportionation energy and highly polarizable NH+ · · · N hydrogen bonds result in a huge and strictly one-dimensional dielectric response. This evidence of the H-bonded relaxor opens new directions in the search for new types of lead-free organic materials with desired dielectric properties. At the same time, in the terms of the proton disordering until the lowest temperature, the NH+ · · · N-bonded dabcoHI crystals investigated possess a close resemblance to the disordered OH · · · O bonds in the ice structure. Although the microscopic mechanism of the proton disordering is differentsthe protons transfer between the N sites of H-bonded cations in dabcoHI, while the disorder in H2O ice is mainly caused by tumbling motionsin both cases, the persistence of the disorder to the lowest temperatures relies on the very weak correlations of the H sites with the heavy-atoms structure. In the previously investigated dabcoBF4 and dabcoClO4 ferroelectric salts, the proton motion in the NH+ · · · N bonds was coupled to the dynamical disordering of the anions. Moreover, the dabcoHI topology is the first example of a strictly onedimensional system, where no strong coupling of the proton

(1) Scott, J. F. Ferroelectric Memories; Springer: Berlin, Germany, 2000. (2) Lines, M. E.; Glass, A. M. Principles and Applications of Ferroelectrics and Related Materials; Oxford University Press: Oxford, U.K., 1979. (3) Chern, G.; Hsieh, W. K.; Tai, M. F.; Hsung, K. S. Phys. ReV. B 1998, 58, 1252–1260. (4) Ramirez, A. P.; Subramanian, M. A.; Gardel, M.; Blumberg, G.; Li, D.; Vogt, T.; Shapiro, S. M. Solid State Commun. 2000, 115, 217–220. (5) Homes, C. C.; Vogt, T.; Shapiro, S. M.; Wakimoto, S.; Ramirez, A. P. Science 2001, 293, 673–676. (6) Lunkenheimer, P.; Bobnar, V.; Pronin, A. V.; Ritus, A. I.; Volkov, A. A.; Loidl, A. Phys. ReV. B 2002, 66, 052105/1–052105/4. (7) O’Neill, D.; Bowman, R. M.; Gregg, J. M. Appl. Phys. Lett. 2000, 77, 1520–1522. (8) Portengen, T.; Östreich, Th.; Sham, L. J. Phys. ReV. B 1996, 54, 17452–17463. (9) Ikeda, N.; Ohsumi, H.; Ohwada, K.; Ishii, K.; Inami, T.; Kakurai, K.; Murakami, Y.; Yoshii, K.; Mori, S.; Horibe, Y.; Kitô, H. Nature 2005, 436, 1136–1138. (10) Park, S.-E.; Shrout, T. R. J. Appl. Phys. 1997, 82, 1804–1811. (11) Zhang, Q. M.; Bharti, V.; Zhao, X. Science 1998, 280, 2101. (12) Park, S.-E.; Shrout, T. R. Mater. Res. InnoVations 1997, 1, 20–25. (13) Cross, L. E. Ferroelectrics 1994, 151, 305–320. (14) Horiuchi, S.; Kumai, R.; Okimoto, Y.; Tokura, Y. Phys. ReV. Lett. 2000, 85, 5210–5213. (15) Kolpakowa, N. N.; Wiesner, M.; Kugel, G.; Bourson, P. Ferroelectrics 1997, 201, 107–115. (16) Szafran´ski, M.; Katrusiak, A. J. Phys. Chem. B 2004, 108, 15709– 15713. (17) Hemberger, J.; Lunkenheimer, P.; Fichtl, R.; Krug von Nidda, H.A.; Tsurkan, V.; Loidl, A. Nature 2005, 434, 364–367. (18) Bao, Z.; Locklin, J., Eds. Organic Field Effect Transistors; Taylor and Francis Group, LLC: London, 2007. (19) Ling, M. M.; Bao, Z. N. Chem. Mater. 2004, 16, 4824–4840. (20) Yoon, M. H.; Yan, H.; Facchetti, A.; Marks, T. J. J. Am. Chem. Soc. 2005, 127, 10388–10395. (21) Busch, G.; Scherrer, P. Naturwissenschaften 1935, 23, 737. (22) Courtens, E.; Rosenbaum, T. F.; Nagler, S. E.; Horn, P. M. Phy. ReV. B 1984, 29, 515–518. (23) Loidl, A.; Böhmer, R. In Disorder Effects on Relaxational Processes; Richert, R., Blumen, A., Eds.; Springer-Verlag: Berlin, Heidelberg, Germany, 1994; pp 659–696. (24) Katrusiak, A.; Szafran´ski, M. Phys. ReV. Lett. 1999, 82, 576–579. (25) Szafran´ski, M.; Katrusiak, A.; McIntyre, G. J. Phys. ReV. Lett. 2002, 89, 215507/1–215507/4. (26) Sheldrick, G. M. SHELXL-97 Program for Crystal Structure Refinement; University of Goettingen: Goettingen, Germany, 1997. (27) Brzezin´ski, B.; Zundel, G. J. Chem. Soc., Faraday Trans. 2 1983, 79, 1249–1257. (28) Olovsson, I.; Jönsson, P.-G. X-Ray and Neutron Diffraction Studies of Hydrogen Bonded Systems. In The Hydrogen Bond; Schuster, P., Zundel, G., Sandorfy, C., Eds.; North-Holland Publ. Comp: Amsterdam, The Netherlands, 1976; Vol. II, pp 393–456.

Giant Dielectric Anisotropy and Relaxor Ferroelectricity (29) Katrusiak, A.; Ratajczak-Sitarz, M.; Grech, E. J. Mol. Struct. 1999, 474, 135–141. (30) (a) Vogel, H. Phys. Z. 1921, 22, 645–646. (b) Fulcher, G. S. J. Am. Chem. Soc. 1925, 8, 339–355. (31) Pirc, R.; Blinc, R. Phys. ReV. B 2007, 76, 020101/1–020101/3. (32) Viehland, D.; Li, J. F.; Jang, S. J.; Cross, L. E.; Wuttig, M. Phys. ReV. B 1992, 46, 8013–8016. (33) Born, M.; Wolf, E. Principles of Optics; Pergamon Press: London, 1959.

J. Phys. Chem. B, Vol. 112, No. 22, 2008 6785 (34) Rabold, A.; Bauer, R.; Zundel, G. J. Phys. Chem. 1995, 99, 1889– 1895. (35) Katrusiak, A.; Szafran´ski, M. J. Am. Chem. Soc. 2006, 128, 15775– 15785. (36) Davydov, A. S. Solitony V molekularnych systemach [Solitons in Molecular Systems, in Russian]; Naukova Dumka: Kiev, Russia, 1988. (37) Takasu, I.; Sugawara, T.; Mochida, T. J. Phys. Chem. B 2004, 108, 18495–18499. (38) Bernal, J. D.; Fowler, R. H. J. Chem. Phys. 1933, 1, 515–548.

JP801106M