Abstract
Organic charge-transfer (CT) co-crystals in a mixed stack system are disclosed, wherein a donor molecule (D) and an acceptor molecule (A) occupy alternating positions (DADADA) along the CT axis. A platform is provided which amplifies the molecular recognition of donors and acceptors and produces co-crystals at ambient conditions, wherein the platform comprises (i) a molecular design of the first constituent (-complement), (ii) a molecular design of the second compound (-complement), and (iii) a solvent system that promotes co-crystallization.
Claims
1. An organic charge-transfer (CT) co-crystal consisting essentially of an electron acceptor molecule (A) and an electron donor molecule (D), such that one of (A) and (D) is incorporated into the other of (A) and (D) through molecular linkages in a solvent system to form a co-crystalline supramolecular network, the co-crystal characterized by having a crystal superstructure comprising a mixed stack lattice (DADADA) and a topological hydrogen-bonded network, said organic CT co-crystal wherein A is selected from the group consisting of ##STR00002## and D is selected from the group consisting of ##STR00003##
2. The organic CT co-crystal according to claim 1, wherein A is a diimide.
3. The organic CT co-crystal according to claim 1, wherein the solvent system is one or more organic solvents.
4. The organic CT co-crystal according to claim 3, wherein the solvent system is selected from the group consisting of dichloroethane/diethyl ether and N-methylpyrrolidone.
5. The organic CT co-crystal according to claim 1, wherein one of the (A) and (D) has a substituent that is a diethylene glycol moiety.
6. The organic CT co-crystal according to claim 1, wherein the hydrogen-bonded network comprises interstack and intrastack hydrogen bonds.
7. The organic CT co-crystal according to claim 1, wherein the co-crystal is devoid or substantially devoid of solvent.
8. The organic CT co-crystal according to claim 1, wherein one of the (A) and the (D) has a substituent that has one or more hydrogen-bonding recognition sites.
9. The organic CT co-crystal according to claim 8, wherein the one or more hydrogen-bonding recognition sites is independently selected from the group consisting of amino, carbonyl, ether and hydroxyl moieties.
10. The organic CT co-crystal according to claim 1, wherein the (A) has at least four substituents.
11. The organic CT co-crystal according to claim 10, wherein the (D) has at least one substituent.
12. The organic CT co-crystal according to claim 1, wherein the co-crystal is grown in the dark, under ambient conditions.
13. The organic CT co-crystal according to claim 12, wherein the co-crystal is grown using liquid diffusion.
14. The organic CT co-crystal according to claim 1 selected from the group consisting of ##STR00004## ##STR00005## ##STR00006##
15. The organic CT co-crystal according to claim 1, wherein the co-crystal has a packing motif that is a crossed stack.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
(1) FIG. 1 provides the structural formulas of the lock-arm supramolecular (LASO) ordering electron donors and electron acceptors used in the self-assembly of ten network solids.
(2) FIGS. 2a-c provide images of the crystal morphology for ten LASO crystals and the experimental controls for the crystallization; (a) ten LASO co-crystals grown from four -complements and eight -complements; (b) image taken after 5 days showing the co-crystals of 19 grown by diffusion 1-chlorobutane into 1,2-dichloroethane/diethyl ether, contaminated with traces of H.sub.2O; (c) image shows experimental controls for the co-crystallizations grown to show that -arm and -arms are essential for self-assembly of LASO materials.
(3) FIG. 3 is a diagram illustrating the use adaptive intermolecular recognition in the system 19.
(4) FIG. 4 provides diagrams that slice through the crystal structures of the ten LASO network solids.
(5) FIGS. 5a-c provide noncovalent connectivity diagrams depicting the global topology of the hydrogen bonded network for systems 19, 110, and 112; (a) diagram showing the network topology of the system 19; (b) diagram showing the network topology of the system 110; (c) diagram showing the network topology of the system 112.
(6) FIGS. 6a-d (a) molecular structures of electron donor (1) and electron acceptor (9, 10 and 12) molecules; (b) co-crystal 19 belonging to the P1 space group with unit cell spacings of 9.5063(4) , 12.1715(6) , and 12.8872(6) , and two unique DA pairs (1.sub.A9.sub.A, 1.sub.B9.sub.B) with spacings of 3.25 and 3.26 between the closest CC contact; (c) co-crystal 110 belongs to the Pn space group with unit cell spacings of 6.9937(2) , 11.8675(2) , and 17.5154(3) ; the co-crystal has a CC distance of 3.30 for the closest contact; (d) co-crystal 112 has a P2.sub.1 space group with unit cell spacings of 11.909(6) , 6.959(3) , and 16.711(7) , and a CC spacing of 3.48 for the closest contact in the DA dimer.
(7) FIGS. 7a-c provide noncovalent connectivity diagrams through Hirshfeld surface analysis of (a) co-crystal 19; (b) co-crystal 110; and (c) co-crystal 112.
(8) FIGS. 8a-f provide images showing the growth of LASO networks after 48 hours for (a) co-crystal 19; (b) co-crystal 110; and (c) co-crystal 112; and optical microscopy of very thin (>10 m) co-crystals with linearly polarized white light for (d) co-crystal 19; (e) co-crystal 110; and (f) co-crystal 112.
(9) FIGS. 9a-c provide linear dichroism of the LASO networks for (a) co-crystal 19; (b) co-crystal 110; and (c) co-crystal 112.
(10) FIGS. 10a-c show the temperature variation of the dielectric constant of LASO (a) co-crystal 19; (b) co-crystal 110; and (c) co-crystal 112.
(11) FIGS. 11a-g provide polarisation hysteresis curves for complex 19 measured at (a) 300 K and (b) 74 K; polarisation hysteresis curves for complex 110 measured at (c) and (d) 300 K and at (e) 7 K; polarisation hysteresis curves measured for complex 112 at (f) 300 K and (g) 7 K.
(12) FIGS. 12a-b are (a) a .sup.1H NMR spectrum of 1 recorded at 25 C. in CD.sub.3COCD.sub.3 and (b) a .sup.13C NMR spectrum of 1 recorded at 25 C. in CD.sub.3COCD.sub.3.
(13) FIG. 13 is a polarized vibrational spectra (300 K) of the co-crystal 17 showing coincident Raman and IR modes.
(14) FIG. 14 is a polarized vibrational spectra (300 K) of the co-crystal 18 showing coincident Raman and IR modes.
(15) FIG. 15 is a polarized vibrational spectra (300 K) of the co-crystal 19 showing coincident Raman and IR modes.
(16) FIG. 16 is a polarized vibrational spectra (300 K) of the co-crystal 110 showing coincident Raman and IR modes.
(17) FIG. 17 is a polarized vibrational spectra (300 K) of the co-crystal 112 showing coincident Raman and IR modes.
(18) FIG. 18 is a polarized vibrational spectra (300 K) of the co-crystal 29 showing coincident Raman and IR modes.
(19) FIG. 19 is a polarized vibrational spectra (300 K) of the co-crystal 211 showing coincident Raman and IR modes.
(20) FIG. 20 is a polarized vibrational spectra (300 K) of the co-crystal 53 showing coincident Raman and IR modes.
(21) FIG. 21 is a polarized vibrational spectra (300 K) of the co-crystal 54 showing coincident Raman and IR modes.
(22) FIG. 22 is a polarized vibrational spectra (300 K) of the co-crystal 63 showing coincident Raman and IR modes.
(23) FIGS. 23a-b provide linear graphs used to determine the ionicity based on the shifting of the ungerade modes in the IR spectra; (a) follows the shift of (CC) for 12.sub.; (b) follows the shift of (CN) for 4.sub..
(24) FIG. 24 is a flow chart illustrating IR modes used to determine for the LASO co-crystals.
(25) FIGS. 25a-d provide linear graphs used to determine the ionicity, based on the shifting of the ungerade modes in the IR spectra; (a) follows the shift of (CO) for 1.sub.; (b) follows the shift of (CC) for 9.sub.; (c) follows the shift of (CO) for 2.sub.; and (d) follows the shift of (CO) for 3.sub..
(26) FIG. 26 is an Oak Ridge Thermal Ellipsoid Plot (ORTEP) drawing of 17.
(27) FIG. 27 is an ORTEP drawing of 18.
(28) FIG. 28 is an ORTEP drawing of 19.
(29) FIG. 29 is an ORTEP drawing of 110.
(30) FIG. 30 is an ORTEP drawing of 112.
(31) FIG. 31 is an ORTEP drawing of 29.
(32) FIG. 32 is an ORTEP drawing of 211.
(33) FIG. 33 is an ORTEP drawing of 53.
(34) FIG. 34 is an ORTEP drawing of 54.
(35) FIG. 35 is an ORTEP drawing of 63.
(36) FIGS. 36a-b provide linear graphs used to determine the ionicity based on the shifting of the ungerade modes in the IR spectra; (a) small crystals of complex TCNQ10 grown and for this complex used as a reference to estimate the ionicity of the remaining co-crystals; (b) ionicity for 19, 110 and 112 determined to be 0.67, 89, and 0.43, respectively.
(37) FIG. 37 are images of complex 19 and 110 growing along the glass of a crystallizing container.
(38) FIG. 38 is a graph of the temperature dependent dielectric constant of TTF-CA (control).
(39) FIGS. 39a-b provide graphs for magnetic hysteresis of LASO (a) co-crystal 19; and (b) co-crystal 112.
(40) FIGS. 40a-b provide graphs for zero-field cool and field cool magnetisation versus temperature plots for LASO(a) co-crystal 19; and (b) co-crystal 112.
(41) FIG. 41 shows the iron analysis of solvents used in LASO co-crystallisation.
(42) FIG. 42 shows the iron analysis of untreated compounds 1 and 9.
(43) FIGS. 43a-d (a) is an image of mixed powder of compounds 1 and 9; M(H) curves at 7 K for mixed powder of compounds (b) 1 and (c) 9; (d) iron analysis report for mixed compound 1 and 9.
(44) FIG. 44 shows the iron analysis for co-crystal 19 and co-crystal 112.
(45) FIGS. 45a-b are a magnetic force microscopy imaging in Lift-off Mode for co-crystal 112; (a) plots of the height (nm); and (b) frequency shift at 80 nm.
(46) FIGS. 46a-c (a) are a magnetic force microscopy, Z-spectroscopy plot of repulsive and attractive potentials across the surface of crystal 112; (b) is a plot of frequency shift versus height (nm) illustrating an attractive potential; (c) is a plot of frequency shift versus height (nm) characteristic of a repulsive potential.
(47) FIGS. 47a-c are polarization hysteresis of compound 110 at (a) 7 K; (b) 150 K; and (c) 300 K.
(48) FIGS. 48a-b are polarization hysteresis of compound 112 at (a) 7 K and up to (b) 35 K.
(49) FIGS. 49a-b provide optical images of (a) ferroelectric device with gold paste and wire; and (b) ferroelectric device with gold paste and wire at voltage >900V and higher temperatures, wherein crystals melted.
DETAILED DESCRIPTION OF THE INVENTION
(50) Accordingly, a non-limiting embodiment of the invention is an organic charge-transfer (CT) co-crystal in a mixed stack system, wherein the co-crystal consists essentially of an electron acceptor molecule (A) and an electron donor molecule (D), wherein one of A and D is an -complement and the other one of A and D is a -complement, such that the -complement is incorporated into the -complement through molecular linkages in a solvent system to form a co-crystalline supramolecular network, wherein one or more of the molecular linkages between the -complement and the -complement use adaptive intermolecular recognition to form the one or more molecular linkages, the co-crystal characterized by having a crystal superstructure comprising a mixed stack lattice (DADADA) and a topologically intricate hydrogen-bonded network.
(51) In an embodiment of the invention, the -complement makes use of a binding motif referred to as adaptive intermolecular recognition. Adaptive intermolecular recognition is defined as the use of conformational isomerism (flexibility) by a molecule to alter the spatial distribution of its recognition sites to achieve energetically stable intermolecular binding in a solid supramolecular network (Aakeroy, C. B., et al. Crystengcomm 2010, 12, 22-43 and Moulton, B., et al. Chem. Rev. 2001, 101, 1629-1658, both incorporated herein by reference). The structure of a molecule that exhibits adaptive intermolecular recognition has two distinguishing criteria, (i) conformational flexibility (this term excludes hydrogen atoms as well as small distortions of molecular geomtery associated with lattice packing, as for example, bond lengthing or the deviation of an aromatic ring from planarity); and (ii) recognition sites that have a unique distribution of distances relative to the molecular centroid for every conformational isomer (FIG. 5). The use of conformational flexibility to achieve intermolecular binding (adaptive intermolecular recognition) in the lattice distinguishes the co-crystals described herein from network solids utilizing rigid molecules whose recognition sites have a fixed distance with respect to the molecular centroid, e.g. coordination polymers and metal organic frameworks.
(52) As described herein, 1, 2, 5, and 6, have the conformational flexibility needed for adaptive intermolecular recognition. 1 is a good illustration of how adaptive intermolecular recognition is used in the systems described herein. It has two flexible diethylene glycol appendages (FIG. 1). Each appendage (or -arm) has two recognition sites for hydrogen bond (H-bond) formation, i.e., an ether oxygen atom and a hydroxyl group. In co-crystals which contain 1, the molecules that neighbor 1 also have supramolecular recognition sites (Desiraju, G. R. Angew. Chem. Int. Ed. 1995, 34, 2311-2327, incorporated herein by reference) (CO, NH.sub.2, OH), and the glycol appendages reach for them to form the most thermodynamically stable H-bonds (FIG. 3 and FIG. 4). As a result, the conformations of the glycol chains adapt to the spatial positions of local recognition sites with the conformational changes, depending on the donor used in the co-crystal (FIG. 5).
(53) To initiate the self-assembly, the -complements () are paired with a smaller CT partner (donor or acceptor) with one or more rigid H-bond recognition sites such as amino, carbonyl, ether and hydroxyl moieties (FIG. 2 and FIG. 3) called the -complement (). The only that is capable of adaptive intermolecular recognition is 12. This fact makes 112 unique since both constiuents engage in adaptive intermolecular recognition. In the crystallizing solution, the and associate through CT. It is believed, however, that the complementary H-bonding interactions between the flexible and rigid moieties on and drive the molecules to self-assemble into an energetically stable network solid.
(54) In another embodiment of the invention, the methods used to produce the co-crystals of the invention should preferably exercise a strong influence on the self-assembly of and . When a mixture of and is combined in the optimal solvent system, expedient self-assembly under ambient conditions should be observed. The dependence of amplified co-crystal growth on the solvent suggests that the crystallization solution promotes the self-assembly of the constituents. The solvent system can be one solvent or a mixture of solvents. The solvent(s) is, preferably, an organic solvent. In a more preferred embodiment, Table 1 provides examples of solvent systems employed by the methods for producing the co-crystals.
(55) TABLE-US-00001 TABLE 1 Molar Growth Co- Concentration.sup.a Ratio Time Crystal Solvent System (mg/mL) (:) (d) 1.Math.9 Dichloroethane/ 2 2 3 Diethyl Ether (200/1) 1.Math.10 Dichloroethane/ 1.5 2 3 Diethyl Ether (200/1) 1.Math.12 Dichloroethane/ 1 2 3 Diethyl Ether (200/1) 1.Math.7 Dichloroethane/ 1.5 2 3 Diethyl Ether (200/1) 1.Math.8 Dichloroethane/ 1.5 2 3 Diethyl Ether (200/1) 2.Math.9 Dichloroethane/ 2 15.sup.b 3 Diethyl Ether (200/1) 2.Math.11 Dichloroethane/ 1 2 3 Diethyl Ether (200/1) 6.Math.3 N- 2 2 5 Methylpyrrolidone 5.Math.3 Dichloroethane/ 2 2 3 Diethyl Ether (200/1) 5.Math.4 N- 2 2 5 Methylpyrrolidone .sup.aConcentration of the electron acceptor ( or ) in the crystallizing solution only. .sup.bConcentration needed to initiate self-assembly.
(56) As used herein, the self-assembly platform producing the co-crystals is referred to as Lock-Arm Supramolecular Ordering (LASO). A LASO network solid is defined as a crystalline supramolecular network wherein one or more of the molecular linkages use adaptive intermolecular recognition to bind to its neighboring molecules. The crystal superstructures (FIG. 4) of the LASO networks used herein constitute a combination of a mixed stack lattice and a topologically intricate H-bonded network. Preferably, the s and s are tightly packed and the crystals are devoid or substantially devoid of solvent. Also preferably, the use of the LASO platform is applied to diimide acceptors.
(57) The D and A components used in co-crystal systems are shown in FIG. 1. As used herein, the ring system of a molecule is the frame (FIG. 3), and any appendage extending from the frame that has a supramolecular binding site is an arm. For simplicity, aromatic hydrogen atoms on the frame will not be classified as arms. In addition to arms, the recognition sites, such as the sulfur atoms in 12, are incorporated into the frame and are referred to as sites. The nitrogen atoms in the diimide molecule are sterically hindered hydrogen bond acceptors and are therefore not considered sites. In a LASO co-crystal, both complementary molecules, i.e., and (FIG. 4) have arms (FIG. 3). The is the molecule that contains the longest conformationally flexible arm extending from the frame (measured by the number of atoms). This appendage is called the -arm. The can have one or more, and preferably two or more, arms, and even more preferably four or more arms, of which at least one, and preferably two, is an -arm. All other arms are denoted without the prime symbol, e.g., the shorter -arms on the or the rigid CO -arms extending from the frame (FIG. 1 and FIG. 3). The is the molecule with one or more, preferably at least two, shorter arms that is incorporated into the supramolecular framework of a LASO network solid through molecular recognition with . The structures in FIG. 1 are the s and s that are used to create the LASO systems (FIG. 2a and FIG. 4) described herein. In the specific non-limiting examples of FIG. 1, there are four s, two six-arm electron acceptors (1, 2) and two, two-arm electron donors (5, 6). All the s have a pair of similar -arms based on a diethylene glycol moiety, but in addition to the -arm, 1 and 2 also have four shorter carbonyl -arms. For the s, two electron acceptors (3, 4) and six electron donors (7-12) are used. Acceptors 3 and 4 are six-arm and four-arm s, respectively. Donors 7-10 are all two-arm s, while 11 is a one-arm . The donor molecule 12 is the most distinct of the s. It has two RCH.sub.2OH -arms, and four sulfur sites making it a two-arm, four-site .
(58) Turning to the ferroelectric behavior of specific embodiments, the CT crystals disclosed herein are based on complexes between donors that are, for example, derivatives of a pyromellitic diimide-based acceptor (FIG. 6a). These CT pairs are co-crystallised (FIG. 8a-8c) under ambient conditions and the resulting solid state structures characterized by X-ray crystallography (FIG. 6b-6d). The basic structural details for each lattice are summarized in Table 2. In these networks, the assembly of the acceptor 1 and donor (9, 10, 12) components in the lattice are stabilized by four primary supramolecular interactions: (i) charge transfer (CT), (ii) hydrogen bonding, (iii) - stacking, and (iv) van der Waals forces. Diimide 1 is functionalized with diethylene glycol arms that are capable of acting as both hydrogen bond donors and acceptors. Electron-rich compounds 9, 10, 12 can also interact through hydrogen bonding since their shorter arms are terminated by hydroxyl or amino groups. An extensive hydrogen bonded network comprised of interstack and intrastack hydrogen bonds (FIG. 6b-6d) is formed during the self-assembly process. This LASO enables complementary molecules to crystallize rapidly into functional networks from solution under ambient conditions (FIG. 7a-7c). The overall motif for LASO structures requires a hierarchical organization based on noncovalent bonding interactions that bridge distances from Angstroms to nanometers and are considerably stronger than van der Waals forces. Locally, the CT and n-n stacking interactions are directed along a single dimension parallel to the mixed stack axis, while the hydrogen bonds extend into three dimensions. This panoply of supramolecular interactions leads to a tightly packed network of mixed stacks locked over larger length scales by hydrogen bonds, - stacking, and CT.
(59) TABLE-US-00002 TABLE 2 19 110 112 formula C.sub.28H.sub.30N.sub.4O.sub.8 C.sub.34H.sub.32N.sub.4O.sub.8 C.sub.26H.sub.28N.sub.2O.sub.10S.sub.4 M 550.56 624.64 656.77 Crystal triclinic monoclinic monoclinic system Space P1 Pn P2.sub.1 Group a () 9.5063 (4) 6.9937 (2) 11.9236 (4) b () 12.1715 (6) 11.8675 (2) 6.9553 (3) c () 12.8872 (6) 17.5154 (3) 16.7123 (5) (deg.) 61.896 (3) 90.00 90.00 (deg.) 89.095 (3) 100.896 (1) 104.157 (2) (deg.) 76.689 (3) 90.00 90.00 V (.sup.3) 1272.50 (1) 1427.53 (5) 1343.89 (8) Z 2 2 2 T (K) 84 100 85
(60) In order to illustrate the connectivity of the hydrogen bonded network in the interstitial regions between the stacks, Hirshfeld surface analysis (McKinnon, J. J., et al. Acta Crystallogr., Sect. B: Struct. Sci 2004, 60, 627-668, incorporated herein by reference) is used to measure the distribution of close contact interactions (FIG. 7). The Hirshfeld surface is a graphical tool that compares the atomically averaged electron density of a molecule to the electron density of the entire crystal and partitions the lattice into molecular surfaces which map the spatial contacts. In FIG. 7, white regions represent molecular contacts at the van der Waals distance, while grey and black portions represent lengths shorter and longer than the van der Waals distance, respectively. This type of analysis helps identify which interactions are the most dominant among neighbouring stacks. The bar graphs show that [O . . . H] interactions responsible for hydrogen bonding make up 37-41% of all interstack contacts while [H . . . H] interactions are 42-50% of such contacts, revealing the close packed nature of the structure (FIG. 7). The lengths shorter than van der Waals are significant because they arise from short [O . . . H] and [H . . . H] distances when interstack hydrogen bonds are formed. These interstack interactions are between neighboring arms and allow the stacks to pack tightly into a supramolecular network.
(61) Since the electron transfer occurs along the stacking axis, ionicity (), the extent of CT, is characterized to investigate how its magnitude affects ferroelectric behavior. Polarized vibrational spectroscopy (FT-IR) is used to determine for each compound. The ungerade modes are used to calculate because they are not influenced by electron-molecular vibration interactions. At room temperature, for 19, 110, and 112 is determined by following the linear shift of the CO stretch (1728-1716 cm.sup.1) polarized parallel to the DA stack. Compounds 19 and 110 are measured to be ionic with =0.68 and 0.89 (see Table 3 below), respectively, while 112 lies close to the neutral-ionic border (=0.5) with =0.43. Therefore, the polar nature of the crystal enables the LASO network to be ferroelectric. Along with significant electron transfer, a violation of the mutual exclusion rule between the IR and Raman modes exists in all three systems at 300 K, indicating a non-centrosymmetric lattice. This behavior in mixed stack crystals, comprised of symmetric molecules, demonstrates that LASO networks fulfill the requirements for a ferroelectric system, namely, DA dimerization and a polar lattice.
(62) Polarized UV-Vis transmission spectroscopy (Kuwatagonokami, M. et al. Nature 1994, 367, 47-48, incorporated herein by reference) (FIG. 9a-9c) and polarized optical microscopy (FIG. 8d-8f) are employed to elucidate the anisotropy of the CT in LASO networks with regard to the crystal axis. When the linear polarization of white light is oriented parallel to the polar mixed stack (long axis), the system absorbs intensely. Conversely, when the polarization is oriented perpendicular to the stack axis, there is a clear lack of color. The absorbance bands (FIG. 9a-9c) associated with this color change shows a maximum in absorbance when the polarization is aligned with the direction of the stacks for 19, 110, and 112. These transitions (FIG. 9a-9c, inset) located between 1.38-1.50 eV (800-900 nm) are attributed to the lowest intra-dimer CT exciton state (Meneghetti, M., et al. J. Chem. Phys. 1996, 105, 397-407, incorporated herein by reference) (DA.sup.+A.sup. [D.sup.+A.sup.]*). Based on this pronounced dichroism, it is possible to establish unequivocally that the polar axis of the material is aligned with the long axis of the crystal.
(63) To determine the ferroelectric Curie temperature of each structure, the dielectric constant (.sub.r) is measured as a function of temperature along the ferroelectric axis. For co-crystals 19, 110, and 112, no characteristic discontinuity is observed between 5 K and 400 K (FIG. 10a-10c). These results suggest that the ferroelectric phase exists at room temperature, an observation which is consistent with spectroscopic and crystallographic evidence. The temperature dependent dielectric constant measurements are performed on TTF-CA as a control (FIG. 38). This measurement shows a T.sub.c of 84 K, consistent with previously reported studies of ferroelectric phase transitions. Further evidence for room temperature ferroelectricity is obtained by measuring hysteresis curves of electric displacement versus electric field (D-E) along the ferroelectric axis (FIGS. 11a, 11c, and 11f). Polarization hysteresis curves for co-crystal 19 are measured at 300 K (FIG. 11a) and 74 K (FIG. 11b). Polarization hysteresis curves for complex 110 are measured at 300 K (FIGS. 11c and 11d) and at 7 K (FIG. 11e). Polarization hysteresis curves are measured for complex 112 at 300 K (FIG. 11f) and 7 K (FIG. 11g). Room temperature hysteresis curves for LASO complexes are underpolarized because of leakage currents at high voltage. Ferroelectric network 110 shows hysteresis similar to 19 and 112 at small electric fields (FIG. 11d). At larger electric fields, 110 demonstrates larger hysteresis loops (FIG. 11c). The unexpectedly large remnant polarisation in 112 observed at low temperatures is attributed to a combination of charge transfer exchange and proton dynamics within the lattice. Hysteresis curve measurements are performed at f=0.1 Hz for FIGS. 11a-e, g, and f=1 Hz for FIG. 11f. Polarization hysteresis of co-crystals 19, 110, and 112 are observed at 300 K with remnant polarizations (P.sub.r) exceeding 1 C/cm.sup.2. Attempts to observe saturation by applying higher electric fields results in dielectric breakdown and crystal melting. Larger polarization hysteresis loops are obtained at lower temperatures down to 7 K where leakage currents are minimized (FIGS. 11b, 11d, 11e and 11g). At low temperature, D-E curves for co-crystal 112 are unexpectedly large. Surprisingly, the P.sub.r for this network is found to be approximately 55 C/cm.sup.2, much larger than compound 19 or 110. This large polarization can result from the combination of the CT process and proton dynamics within the crystal (Horiuchi, S., et al. Nature Mater. 2008, 7, 357-366 and Horiuchi, S., et al. Nature 2010, 463, 789-797, both incorporated herein by reference). The P.sub.r of 111 at 7 K is among the highest reported for organic ferroelectrics based on charge transfer, hydrogen bonding, liquid crystalline or polymeric materials. The resistivity of all LASO systems investigated, however, is found to be very high (>10.sup.9 /cm) at room temperature.
(64) The ferroelectric curves obtained at room temperature are biased at a lower electric field compared to cryogenic temperatures. At high electric fields at room temperature, dielectric leakage and joule heating prevents the measurement of saturating polarization hysteresis loops. Curves measured at 300 K are obtained by applying a smaller electric field than required for saturation. As a result, these systems are inherently under-polarized and have smaller remnant polarizations than saturated loops.
(65) Larger hysteresis loops are obtained in compound 110 (FIG. 11d) at room temperature because this network is able to withstand higher voltages. It is interesting to note that this hydrogen bonded network has a higher ionicity (0.89) than co-crystals 19 or 112. The only CT ferroelectric that demonstrates polarization bistability is TTF-BA with a T.sub.c, of 53 K. TTF-BA also has a very large ionicity with 0.9, similar to 110. As pointed out by Torrance, J. B. Accounts of Chemical Research 1979, 12, 79-86, incorporated herein by reference, high ionicities inhibit current flow in CT crystals because of Coulombic interactions. Negative and positive ions in a lattice can behave as ionic impurities that actively scatter moving electrons. Network 110 and TTF-BA have large ionicities and may therefore mitigate leakage current to some degree. In this context, developing networks of CT complexes with large ionicities may be a useful design rule for ferroelectricity at room temperature and above.
(66) Ferroelectric networks 19 and 112 are characterized by SQUID magnetometry and revealed magnetic hysteresis loops. Extensive elemental analysis described in detail below shows that any magnetic impurities present have to be below the detection limit of currently available instruments for inductively coupled plasma atomic emission spectroscopy (ICP-AES). Other measurements (Magnetic Force Microscopy) described below attempt to verify ferromagnetic behavior.
(67) Materials and Methods
(68) All compounds are purchased from commercial vendors (Sigma Aldrich and VWR) and are used as supplied without further purification. For the synthesis of 1, 2, 5, 6, and 12 see: Bevers, S., et al. J. Am. Chem. Soc. 2000, 122, 5905-5915; Sue, C. H., et al. Chem Sci 2010, 1, 119-125; Asakawa, M., et al. Journal of Organic Chemistry 1996, 61, 9591-9595; and Saha, S., et al. Chem. Eur. J. 2005, 11, 6846-6858, all incorporated herein by reference.
(69) All the crystals are grown in the dark, under ambient conditions using liquid diffusion. Two distinct solvent systems are found to promote expedient crystal growth. The molar ratio : and the total concentration of + are optimized to achieve the best crystal size and growth rate (Table 1). The eight co-crystals which do not contain 3 (pyromellitic diimide) are grown from liquid diffusion of anhydrous non-protic solvents 1-chlorobutane into a 1,2-dichloroethane and diethyl ether mixture (FIG. 2a and FIG. 3). As H.sub.2O is found to disrupt the self-assembly of these eight co-crystals, anhydrous conditions for the materials and solvents are ensured before the crystals are grown. Owing to the insolubility of 3 in common organic solvents, a different type of solvent combination is used for the two complexes containing . The liquid diffusion of protic H.sub.2O into polar N-methylpyrrolidone is found to produce the most pronounced crystal growth. EG/SiC samples are produced in a UHV chamber with a base pressure below 110.sup.10 Torr unless otherwise noted. The SiC is resistively heated by passing current through the SiC while temperatures are monitored using an optical pyrometer (Cyclops) at an emissivity of 0.85. The SiC is degassed overnight at 600 C. and then annealed for 2 minutes at 1000 C. The SiC is then flashed 3 times at 1100 C. for 2 minutes each. After each flash, the sample is allowed to cool for 10 minutes. Finally, the SiC is graphitized at 1300 C. for 2 flashes and then 10 flashes at 1350 C. for 1 minute apiece.
(70) Thin layer chromatography (TLC) is performed on silica gel 60 F254 (E. Merck). Nuclear magnetic resonance (NMR) spectra are recorded at 25 C. on Varian Inova 500 spectrometers, with working frequencies of 500 MHz for .sup.1H, and 125 MHz for .sup.13C nuclei (see FIG. 12a and FIG. 12b for .sup.1H and .sup.13C NMR, respectively, of 1). The chemical shifts are listed in ppm on the scale and coupling constants are recorded in Hertz (Hz). The following abbreviations are used to explain the multiplicities: s, singlet; d, doublet; t, triplet; b, broad peaks; m, multiplet or overlapping peaks. High resolution electrospray ionization (HR ESI) mass spectra are measured on a Micromass Q-TOF Ultima mass spectrometer. All FT-IR spectra are collected by a Perkin Elmer Spectrum Spotlight Imaging System utilizing liquid nitrogen cooled single element mercury-cadmium-telluride detector in single-point reflectance mode with an aperture setting of 5050 m.sup.2. The sample is presented as a single crystal placed on a Si plate. Raman spectra are obtained with either an Advantage 532 Raman Spectrometer using a 532 nm excitation line or a 633 nm HeNe laser (Research Electro-Optics Inc.) that collects the scattered light into a spectrograph (PI Acton SP2500i) equipped with a 600 g/mm grating blazed at 750 nm and a liquid-N.sub.2 cooled CCD detector. Polarization of the HeNe laser is achieved using an adjustable air spaced achromatic half-wave waveplate (CVI Melles Griot ACWP-400-700-10-2). A baseline correction is performed to allow for interpretation of the spectra. Single crystals of the complexes are mounted in oil (InfineumV8512) on a glass fiber under a nitrogen cold stream at 83(2) K. X-Ray diffraction data are collected on a Bruker Kappa diffractometer, equipped with a CuK or MoK sealed-tubesource and an APEX II CCD detector. Data are collected, integrated and corrected for decay and Lp effects using BrukerAPEX II software. Final unit cell parameters are obtained through a refinement of all observed reflections during data integration. A multi-scan absorption correction is performed using SADABS. The structures are solved and refined using the SHELXTL suite of software. The absorption spectra are taken using a polarization microscopy setup. A Nikon TE2000 inverted microscope and Prior ProScan II stage are used to manipulate the sample position. The microscope halogen lamp is used as the source for the absorption spectra. Spectra are recorded using an Ocean Optics USB 2000 miniature spectrometer. The polarization dependence is varied using a thin film polariser. Cyclic voltammetry (CV) and square-wave differential pulse voltammetry (SWDPV) experiments are performed at room temperature in argon-purged solutions of N,N-dimethylformamide (DMF) with a Gamry Multipurpose instrument (Reference 600) interfaced to a PC. The CV and SWDPV experiments in each case are performed using a glassy carbon working electrode (0.071 cm.sup.2). The surface of this electrode is polished routinely with 0.05 m alumina-water slurry on a felt surface immediately before each run. The counter electrode is a Pt coil and the reference electrode is a saturated calomel electrode (SCE) for both CV and SWDPV experiments. The concentration of the sample and supporting electrolyte tetrabutylammonium hexafluorophosphate (TBAPF.sub.6) are 1.010.sup.3 mol.Math.L.sup.1 and 0.1 mol.Math.L.sup.1, respectively. The scan rate for CV experiments is set to 200 mV.Math.s.sup.1. Experimental errors: potential values, 10 mV for CV and 1 for SWDPV.
(71) Ferroelectric structures are mechanically robust and can be handled with vacuum tweezers. Gold wire electrodes (12.5 m) are attached on either end using gold paint (Ted Pella Gold Paste). The resulting devices are tested in a QuantumDesign PPMS 6000 under an inert atmosphere. The dielectric constant of LASO complexes is determined by capacitance-voltage measurements at 10 V with a 1, 5, or 10 kHz frequency. These measurements are performed using an Agilent E4980A LCR meter. Polarization hysteresis is measured using a ferroelectric tester at 0.1 Hz or 1 Hz frequency (Radiant Technologies Precision LC with Trek amplifier).
(72) Vibrational spectroscopy data helps elucidate the lattice symmetry of the co-crystals at ambient conditions. Six of the LASO crystals are refined in centrosymmetric space groups. The remaining four co-crystals (19, 110, 112, 53), however, are found to adopt a non-centrosymmetric lattice. These network solids have the spectroscopic signature of a mixed stack crystal that has undergone a polar phase transition where the donors and acceptors have dimerized (D.sup.0 A.sup.0 D.sup.0 A.sup.0 D.sup.+A.sup. D.sup.+A.sup.) along the charge transfer (CT) axis.
(73) Employing IR and Raman spectroscopic techniques, the details of the ground state for a CT crystal are experimentally accessible (See FIGS. 13-23). In segregated stacks, totally symmetric (ts) modes can be used to measure the value of the ionicity () for the material (Girlando, A., et al. Synth. Met. 2004, 141, 129-138, incorporated herein by reference), but in mixed stack systems the ts modes are perturbed by the electron-molecular vibration interaction (Girlando, A., et al. J. Chem. Phys. 1983, 79, 1075-1085, incorporated herein by reference) and are not useful for this purpose. However, because the ts Raman modes of molecules in crystals are subject to the selection rules governed by site symmetry, these bands are good probes for the loss of centrosymmetry. When molecules do not occupy a center of inversion, the ts vibrations can violate the principle of mutual exclusion (Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds, 6th ed., Wiley [Oxford Wiley-Blackwell, distributor]: Hoboken, N.J., 2009, incorporated herein by reference) and appear in the IR spectrum. Away from the center of inversion the molecular dipole moment can change with ts vibrations about its equilibrium. In a polar mixed stack system with symmetric molecules, the dimerization of the donor and acceptor breaks the inversion symmetry in the lattice, and the donor and acceptor no longer reside on inversion centers. The ts modes are now capable of producing an asymmetric charge distribution (dipole moment) and strongly coupling to the CT along the stack. As a result, the ts modes can become coincident in their IR and Raman spectra and are strongly polarized in the direction of the CT axis. Most of the co-crystals have some coincident bands ((CH.sub.2)) in the Raman and IR spectra, but only four (FIG. 5, FIG. 6, FIG. 7 and FIG. 10) of the LASO co-crystals display this spectroscopic signature in regions ((CO), (CC)) that are sensitive to the CT interaction.
(74) Referring to FIGS. 13-23, provided are the polarized vibrational spectra (300 K) of co-crystals showing coincident Raman and IR modes. The lowest plot is the unpolarized Raman spectrum. The symbols and indicate linear polarization of the IR radiation with the electric field oriented parallel and perpendicular to the CT stack, respectively. The 17 co-crystal is a crossed stack system (FIG. 13); the 18 co-crystal is a crossed stack system, and the lattice also contains and asymmetric (FIG. 14); the violation of the rule of mutual exclusion, a spectroscopic signature of dimerization in co-crystal 19 (FIG. 15); the violation of the rule of mutual exclusion, a spectroscopic signature of dimerization in co-crystal 110 (FIG. 16); the violation of the rule of mutual exclusion, a spectroscopic signature of dimerization in co-crystal 112 (FIG. 17); co-crystal 29 shows very few coincident peaks, indicating a lack of dimerization between the donor and acceptor, possibly caused by the positional disorder found in the lattice (FIG. 18); co-crystal 211 contains an asymmetric that could result in the overlap of modes between the IR and Raman spectra (FIG. 19); the violation of the rule of mutual exclusion, a spectroscopic signature of dimerization in co-crystal 53 (FIG. 20); co-crystal 54 shows very few coincident peaks, indicating a lack of dimerization between the donor and acceptor, possibly caused by the positional disorder found in the lattice (FIG. 21); co-crystal 63 shows very few coincident peaks, indicating a lack of dimerization between the donor and acceptor, possibly caused by the positional disorder found in the lattice (FIG. 22).
(75) Since Girlando, A., et al. J. Chem. Phys. 1983, 79, 1075-1085 establishes that the degree of CT () can be probed through the shifts of ungerade fundamental modes in the vibrational spectra, IR spectroscopy is used extensively for this purpose. Ungerade modes are the best choice for determining , because the shift is not affected by electron-phonon coupling. As long as the crystal is not close to the Curie temperature of a phase transition, the relationship between the shifts in these modes and changes in p are nearly linear. Linear interpolation between the peak positions of the neutral and fully charged molecular speciesdonor or acceptoryields a reliable estimate of .
(76) To determine in the LASO co-crystals, a method based on the linear shifting of the ungerade modes is also used (FIG. 23 and FIG. 36). However, a complex between 4 (tetracyanoquinodimethane, TCNQ) and the tetrathiafulvalene (TTF) derivative 12 is grown and used as a reference to estimate for the rest of the LASO crystals. The affect of CT interactions on the vibrational spectra of TCNQ are well understood (Bozio, R., et al. J. Chem. Soc., Faraday Trans. II 1978, 74, 235-248; and Meneghetti, M., et al. J. Chem. Phys. 1985, 83, 3134-3145, incorporated herein by reference) and the ionicity of a co-crystal can be determined by using the shift of the (CN) mode (FIG. 23b). This procedure is used to estimate for TCNQ4 (FIG. 23a and FIG. 36a), =1.0. In the IR spectra of the LASO co-crystals, it is found that the (CO) mode in 1 is sensitive to changes in (Horiuchi, S., et al. Nat. Mater. 2005, 4, 163-166, incorporated herein by reference). Using extrapolation in the plot of vs. (CO) of 1 and 112 (FIG. 36b), the ionicity of 19 and 110 is determined to be =0.68 and =0.89, respectively. FIG. 24 is a flow chart illustrating which IR modes are used to determine for the LASO co-crystals described herein. FIG. 25 contains linear graphs used to determine the ionicity based on shifting the ungerade modes in the IR spectra. FIG. 25a is a graph that follows the shift of (CO) for 1; FIG. 25b follows the shift of (CC) for 9; FIG. 25c follows the shift of (CO) for 2; and FIG. 25d follows the shift of
(77) (CO) for 3.
(78) X-ray crystallographic data is obtained for co-crystals and are as follows.
(79) A) 17: C.sub.46H.sub.48N.sub.4O.sub.18, M=472.44, triclinic, a=6.7868(1), b=10.8904(2), c=15.7778(2) , =70.880(1), =81.554(1), =83.156(1), V=1086.72(3) .sup.3, T=100(2) K, space group P1, Z=1, =1.44 g.Math.cm.sup.3, (Mo.sub.K)=0.11 mm.sup.1, 10415 independent observed reflections, 6720 reflections with I>2(I), R.sub.int=0.040, R[F.sup.2>2(F.sup.2)]=0.052, wR(F.sup.2)=0.132.
(80) FIG. 26 is the Oak Ridge Thermal Ellipsoid Plot Program (ORTEP) for co-crystal 17. The hydrogen atoms are omitted for clarity. This co-crystal is crossed stack. The ratio of the PMDI-based to the naphthalene-based is 2:1 acceptor in the unit cell. There is -face-to--face packing and edge-to--face packing between the and . All ellipsoids are displayed at the 50% probability level.
(81) B) 18: C.sub.46H.sub.49N.sub.5O.sub.17, M=471.95, triclinic, a=6.7603(7), b=10.8522(11), c=15.785(2) , =71.882(9), =81.810(9), =84.105(8), V=1087.3(2) .sup.3, T=84(2) K, space group P1, Z=1, =1.44 g.Math.cm.sup.3, (Mo.sub.K)=0.94 mm.sup.1, 4368 independent observed reflections, 3222 reflections with I>2(I), R.sub.int=0.047, R[F.sup.2>2(F.sup.2)]=0.051, wR(F.sup.2)=0.147.
(82) FIG. 27 is the ORTEP for co-crystal 18. The hydrogen atoms are omitted for clarity. Co-crystal 18 is crossed stack. The ratio of the PMDI-based to the naphthalene-based is 2:1 acceptor in the unit cell. There is -face-to--face packing and edge-to--face packing between the and . All ellipsoids are displayed at the 50% probability level.
(83) C) 19: C.sub.28H.sub.30N.sub.4O.sub.8, M=550.56, triclinic, a=9.5063(4), b=12.1715(6), c=12.8872(6) , =61.896(3), =89.095(3), =76.689(3), V=1272.50(10) .sup.3, T=84(2) K, space group P1, Z=2, =1.44 g.Math.cm.sup.3, (Mo.sub.K)=0.11 mm.sup.1, 10748 independent observed reflections, 7435 reflections with I>2(I), R.sub.int=0.038, R[F.sup.2>2(F.sup.2)]=0.052, wR(F.sup.2)=0.146.
(84) FIG. 28 is an ORTEP drawing of co-crystal 19. The hydrogen atoms are omitted for clarity, co-crystal 19 is crossed stack. The ratio of the PMDI-based to the naphthalene-based is 2:1 acceptor in the unit cell. There is -face-to--face packing and edge-to--face packing between the and . The is asymmetric with amino and hydroxy arms occupying the 1,5-positions. This asymmetry results in substitutional disorder at the 1,5-positions in the lattice. All ellipsoids are displayed at the 50% probability level.
(85) D) 1.sub.10.sub.: C.sub.34H.sub.32N.sub.4O.sub.8, M=624.64, monoclinic, a=6.9937(2), b=11.8675(2), c=17.5154(3) , =100.896(1), V=1427.53(5) .sup.3, T=100(2) K, space group Pn, Z=2, =1.46 g.Math.cm.sup.3, (CU.sub.K)=0.87 mm.sup.1, 3273 independent observed reflections, 2998 reflections with I>2(I), R.sub.int=0.024, R[F.sup.2>2(F.sup.2)]=0.048, wR(F.sup.2)=0.142.
(86) FIG. 29 is an ORTEP drawing of co-crystal 110. The hydrogen atoms are omitted for clarity. Co-crystal 110 is crossed stack. The ratio of the PMDI-based to the naphthalene-based is 2:1 acceptor in the unit cell. There is -face-to--face packing and edge-to--face packing between the and . All ellipsoids are displayed at the 50% probability level.
(87) E) 112: C.sub.26H.sub.28N.sub.2O.sub.10S.sub.4, M=656.74, monoclinic, a=11.9236(4), b=6.9553(3), c=16.7123(5) , =104.227(4), V=1348.26(1) .sup.3, T=85(2) K, space group P2.sub.1, Z=2, =1.62 g.Math.cm.sup.3, (CU.sub.K)=3.81 mm.sup.1, 3213 independent observed reflections, 2680 reflections with I>2(I), R.sub.int=0.040, R[F.sup.2>2(F.sup.2)]=0.045, wR(F.sup.2)=0.130.
(88) FIG. 30 is an ORTEP drawing of co-crystal 112. The hydrogen atoms are omitted for clarity. There is one TTF-based and one PMDI-based in the unit cell. All ellipsoids are displayed at the 50% probability level.
(89) F) 29: C.sub.32H.sub.28N.sub.4O.sub.8 M=596.58, triclinic, a=6.9510(2), b 8.6966(2), c=12.1281(3) , =72.093(2), =76.054(2), =80.941(2), V=674.30(3) .sup.3, T=100(2) K, space group PI, Z=2, =1.47 g.Math.cm.sup.3, (Mo.sub.K)=0.11 mm.sup.1, 3880 independent observed reflections, 2075 reflections with I>2(I), R.sub.int=0.057, R[F.sup.2>2(F.sup.2)]=0.076, wR(F.sup.2)=0.255.
(90) FIG. 31 is an ORTEP drawing of co-crystal 29. The hydrogen atoms are omitted for clarity. There is one naphthalene-based and one naphthalene diimide-based (NPDI) in the unit cell. There is positional disorder in the glycol -arm and the ring system of . All ellipsoids are displayed at the 50% probability level.
(91) G) 211: C.sub.38H.sub.33N.sub.3O.sub.8, M=659.67, triclinic, a=10.9811(8), b=12.4287(8), c=12.9441(9) , =94.620(5), =112.518(5), =109.840(5), V=1489.33(18) .sup.3, T=85(2) K, space group PI, Z=2, =1.47 g.Math.cm.sup.3, (Cu.sub.K)=0.86 mm.sup.1, 4883 independent observed reflections, 3441 reflections with I>2(I), R.sub.int=0.042, R[F.sup.2>2(F.sup.2)]=0.058 wR(F.sup.2)=0.176.
(92) FIG. 32 is an ORTEP drawing of co-crystal 211. The hydrogen atoms are omitted for clarity. There is one asymmetric pyrene-based and two crystallographically unique NPDI s in the unit cell. The has one amino arm at the 1-position. The asymmetry of the results in substitutional disorder at the 1,8-positions within the lattice. All ellipsoids are displayed at the 50% probability level.
(93) H) 53: C.sub.28H.sub.30N.sub.4O.sub.8, M=550.56, monoclinic, a=6.6667(3), b=23.3906(10), c=8.3455(3) , =104.657(3), V=1259.03(9) .sup.3, T=84(2) K, space group Pc, Z=2, =1.45 g.Math.cm.sup.3, (CU.sub.K)=0.11 mm.sup.1, 6787 independent observed reflections, 3829 reflections with I>2(I), R.sub.int=0.080, R[F.sup.2>2(F.sup.2)]=0.059, wR(F.sup.2)=0.141.
(94) FIG. 33 is an ORTEP drawing of co-crystal 53. The hydrogen atoms are omitted for clarity. There is one PMDI and one napthalene-based in the unit cell. All ellipsoids are displayed at the 50% probability level.
(95) I) 5.sub.4.sub.: C.sub.30H.sub.30N.sub.6O.sub.4, M=538.60, triclinic, a=6.8961(2), b=8.0293(3), c=12.3435(4) , =89.213(1), =83.730(2), =73.487(2), V=651.25(4) .sup.3, T=100(2) K, space group PI, Z=1, =1.37 g.Math.cm.sup.3, (Cu.sub.K)=0.76 mm.sup.1, 2233 independent observed reflections, 2176 reflections with I>2(I), R.sub.int=0.018, R[F.sup.2>2(F.sup.2)]=0.031, wR(F.sup.2)=0.084.
(96) FIG. 34 is and ORTEP drawing of co-crystal 54. The hydrogen atoms are omitted for clarity. There is one TCNQ and one naphthalene-based in the unit cell. All ellipsoids are displayed at the 50% probability level.
(97) J) 63: C.sub.28H.sub.28N.sub.2O.sub.10, M=552.52, monoclinic, a=6.6836(2), b=23.4173(6), c=8.3689(2) , =106.174(2), V=1257.99(6) .sup.3, T=100(2) K, space group P2.sub.1/c, Z=2, =1.46 g.Math.cm.sup.3, (CU.sub.K)=0.11 mm.sup.1, 3828 independent observed reflections, 2322 reflections with I>2(I), R.sub.int=0.097, R[F.sup.2>2(F.sup.2)]=0.051, wR(F.sup.2)=0.123.
(98) FIG. 35 is an ORTEP drawing of co-crystal 63. The hydrogen atoms are omitted for clarity. There is one PMDI and one napthalene-based in the unit cell. All ellipsoids are displayed at the 50% probability level.
(99) The magnetic properties of LASO co-crystals are characterised by SQUID magnetometry (FIG. 39). Magnetic hysteresis is observed at low temperature for compounds 19 and 112. Samples for SQUID magnetometry are prepared by packing LASO crystals into a non-magnetic gel capsule. Masses for the sample range from 10 mg to 30 mg. The magnetism as a function of temperature is also measured for both co-crystals (FIG. 40). Great care is taken during the preparation of LASO samples to prevent the introduction of magnetic impurities. Compounds 1, 9, 10, and 12 are recrystallised multiple times and stored in a drybox prior to use. No metal instruments or containers are used during the process. Samples for SQUID magnetometry are prepared in a laminar flow hood. ICP-AES is performed by Galbraith Laboratories to directly address concerns over the possible presence of iron impurities. FIG. 41 are analysis reports of iron content in solvents used for the co-crystallisation of 19, 110, and 112: 1-chlorobutane, 1,2 dichloroethane, and diethyl ether. FIG. 44 are analysis reports of iron content on co-crystals 19 and 112. The overall iron content is very low. The less than symbol (<) indicates that the iron content is below the sensitivity limit of the instrument. Iron analysis of both independent compounds, simple mixes of D-A molecules, and charge transfer crystals are also performed. As an example, to illustrate the overall low iron content in LASO starting materials, compounds 1 and 9 are analyzed (FIG. 42). Compounds 1 and 9 are coarsely mixed together to form a charge transfer powder and measured by both SQUID magnetometry and iron analysis (FIG. 43). In FIG. 43, the black is indicative of charge transfer and complexation. M(H) curves at 7 K for mixed powder of compound 1 and 9. No hysteresis is observed, solely diamagnetic contribution from the powder and capsule container. Both graphs are of the same sample, with different x-axis scale to illustrate the lack of ferromagnetic hysteresis. Iron analysis report for the above mixed compound powder is provided (bottom image).
(100) The low magnetic saturation signal observed indicates the possibility of extrinsic magnetisation. To further quantify whether magnetism is derived from LASO networks or a tertiary impurity, low-temperature magnetic force microscopy (MFM) is performed.
(101) These MFM measurements are performed at 45 K in two modes: lift-off and z-spectroscopy. In the case of the former, the topography of compound 112 is first measured (FIG. 45). A second scan of the same region with a magnetised cantilever is performed while approximately 10, 20, 30, 50, and 80 nm above the surface. At all heights, topography is still visible in the measurement while in phase-locked loop mode (FIG. 45).
(102) The influence of topography is evident in the scan 80 nm above the crystal surface. Therefore, Z-spectroscopy is performed to ascertain whether LASO materials are inherently magnetic (FIG. 46). The presence of both attractive and repulsive interactions with a magnetized cantilever proves that magnetic domains exist within the material. Though a prominent magnetic hysteresis loop is observed by SQUID magnetometry, the lack of a characteristic repulsive force in Z-spectroscopy suggests that LASO networks are not inherently magnetic.
(103) Samples studied by low temperature MFM are also charge compensated at 800 mV. Kelvin-probe measurements on the surface of compound 112 show some differences in charge distribution, possibly due to ferroelectric domains.
(104) Ferroelectric hysteresis loops are obtained at low temperatures for compounds 19, 110 and 112. Large voltages are needed to achieve polarization saturation (>1 kV). This large voltage results in crystal melting and dielectric breakdown at higher temperatures. Thus, to obtain hysteresis at room temperature, LASO materials are under-polarized to obtain hysteresis.
(105) LASO materials based on Pyrene demonstrated hysteresis from low temperature (7 K, 150 K) upto room temperature (FIG. 47). Compounds based on tetrathiafulvalene showed large hysteresis at low temperatures with nice saturation upto 1.6 kV (FIG. 48). This unexpectedly large polarization is likely the result of CT processes and proton dynamics.
(106) Challenges with higher temperature and higher voltage measurements prevent the recording of saturated hysteresis curves. The main issue is likely avalance breakdown and crystal melting at high voltages (>1 kV). Devices that begin with long needles (FIG. 49a) are destroyed and emerge as re-crystallized solids on the electrode (FIG. 49b). Crystal networks can be formed into ferroelectric devices with gold paste and wire. At high voltages (>900V) and higher temperatures, crystals frequently melt likely due to high leakage currents and avalanche breakdown.
(107) Example 1
(108) 1 (Scheme 1): Pyromellitic dianhydride (5.00 g, 22.9 mmol) is added to a 40 mL pressure tube containing 2-(2-aminoethoxy)ethanol (4.7 mL, 46.8 mmol). The reaction mixture is heated to 160 C. and stirred for 24 hours. After cooling to ambient temperatures, the solid is dissolved in trifluoroacetic acid (100 mL), and the resulting solution is stirred for 24 hours. The mixture is then neutralized with a saturated NaHCO.sub.2 aqueous solution, and the precipitate is filtered and washed with H.sub.2O (3100 mL).
(109) The crude product is recrystallized twice from THF and Et.sub.2O to yield a sticky white solid (5.49 g, 61%) of 1. .sup.1H NMR (500 MHz, CD.sub.3COCD.sub.3, 298 K): =8.21 (s, 2H), 3.90 (t, J=5.9 Hz, 4H), 3.75 (t, J=5.9 Hz, 4H), 3.55 (m, 8H). .sup.13C NMR (125 MHz, CD.sub.3COCD.sub.3, 298 K): =38.6, 61.9, 67.7, 71.6, 119.3, 137.4, 166.9. HR ESI: calcd for [M+H].sup.+ m/z=393.1298; found m/z=393.1299.
(110) ##STR00001##
Discussion
(111) Using the LASO platform, ten co-crystals are grown (FIG. 2a) from 12 s and s. As the co-crystals grew, three distinguishing features emerged: (i) growth rate, (ii) size, and (iii) morphology. The growth rate is controlled by adjusting the total concentration of and in the crystallizing solution, and crystal size can be changed with growth time. The morphology of the materials (FIG. 2a) does not change with subsequent crystallizations. Only one system (17) is found to contain a small amount of polymorphic material. Seven of the co-crystals grew as long prisms (19, 110, 112, 29, 211, 63, 53). The system 54 also grew as a prism but assumed a shorter cuboidal shape. Complexes 17 and 18 have the most unusual morphology. These materials grew rapidly in two dimensions, giving them the appearance of thin sheets.
(112) Using the concentration of and (Table 1), the growth time of high-quality single crystals (cm length-scale) is optimized to a period of several days (FIG. 2a). The LASO systems 63 and 53 are crystallized from H.sub.2O/N-methylpyrrolidone and takes 5 days to reach the maximum size. The eight remaining co-crystals finish growing slightly faster (3 days) from anhydrous 1-chlorobutane/1,2-dichloroethane/diethyl ether. Notably, many of the pairs show visible co-crystals 2 hours after being mixed in the crystallizing solution, indicating the strong drive for self-assembly. Although LASO materials are capable of expedient growth, strict adherence to the optimized solvent conditions is crucial for successful co-crystallization. FIG. 2b shows the prevalent out-growth of 19 when the crystallizing solution is contaminated with trace amounts of H.sub.2O, indicating the strong influence solvent conditions have on the crystal morphology.
(113) FIG. 37 provides images of complex 19 and 110 growing along the glass of the crystallising container. This type of growth indicates the possibility of controlled growth of LASO on surfaces.
(114) Two co-crystallizations of DA pairs are used as controls in order to demonstrate that arms are crucial to the self-assembly of LASO materials (FIG. 2c). The first experimental control is the growth of 39, a co-crystal with two s. This co-crystal is equivalent to LASO systems 19 and 63, except it lacks a -arm. Without an -arm, neither component in 39 can engage in adaptive intermolecular recognition, the phenomenon which is the key attribute in the LASO platform. Co-crystal 39 is grown from diffusion of 1-chlorobutane into 1,2-dichloroethane/diethyl ether/N-methylpyrrolidone under anhydrous conditions. For the crystallization, 3 and 9 are used in 1:2 ratio at a concentration of 2 mg/mL of 3. 39 co-crystals grew as black in color, similar to the color of both 19 and 63, yet the crystal size is significantly smaller and the quality is not adequate for structure determination by X-ray crystallography (FIG. 2c). This result suggests that the -arm, used for adaptive intermolecular recognition by the s, is crucial for self-assembly in LASO systems.
(115) The second control experiment is to co-crystallize tetrathiafulvalene (TTF) with 1. A crystal of TTF12 is the co-crystal equivalent of 112 but lacking the -arms. This control demonstrates that the -arms are important for crystal growth. Diffusion of 1-chlorobutane into 1,2-dichloroethane/diethyl ether is used as the solvent system and, after 14 days at 22 C., only discolored crystals of 1 are found to grow. For the crystallization, 1 and TTF are used in 1:2 ratio at a concentration of 2 mg/mL of 1. This result shows that the -arm is not a passive observer since these appendages also promote the co-crystallization of and . From these control experiments, it appears that both the conformationally flexible -arm and the rigid -arm must be present for self-assembly to occur. These arms appear to enhance the supramolecular affinity between CT pairs and provide stability (G of the lattice ground state) to the LASO network solid through interstack H-bonding.
(116) FIG. 3 illustrates the use of adaptive intermolecular recognition by an inside a LASO network solid. In 19, the flexible -arms of 1 adopt two distinct conformations in the lattice (acceptor 1 and acceptor 2) to drive the system to an energetically favorable network topology. The arm conformation of the two as mold to adjoining intermolecular recognition sites. The conformational isomer of can vary significantly (FIG. 4 and FIG. 5) depending upon the that is used to generate a LASO network solid. In the local network structure 19, - and - H-bonds (FIG. 3) bind neighboring as together, and the - and - arm interactions form intermolecular H-bonds between adjacent s and s. The resulting topology is a highly interconnected three-dimensional (3-D) supramolecular network (FIG. 5a).
(117) The crystal superstructures for all ten LASO materials are elucidated from single-crystal X-ray diffraction data (FIG. 4). Each of the systems has stacks of alternating donors and acceptors (DADA) and a multidimensional H-bonded network (1D-3D). The space groups of the LASO co-crystals are listed in Table 3. While eight of the systems exhibit the typical 1-D packing of a mixed stack, two co-crystals are distinct from the rest. Systems 17 and 18 (FIG. 2a), display a previously unknown packing motif for CT crystals (FIG. 5). In these systems, the ratio of (acceptor) to (donor) is 2:1 where the s engage in CT with the through -face-to--face (ff) and edge-to--face stacking (ef), respectively. As a result of the ff and ef packing, there are two CT axes (bidirectional CT) in these systems (17 and 18) that intersect at an angle of 90. As referred to herein, this packing motif is a crossed stack. The resulting arrangement is a checker board-pattern of s and s which lie parallel to the plane [001].
(118) TABLE-US-00003 TABLE 3 Co-Crystal Space Group .sup.b 1 .Math. 9 P1 0.68 1 .Math. 10 Pn 0.89 1 .Math. 12 P2.sub.1 0.43 1 .Math. 7 P1 0.53.sub.ff, 0.47.sub.ef 1 .Math. 8 P1 0.57.sub.ff, 0.42.sub.ef 2 .Math. 9 PT 0.58 2 .Math. 11 PT 0.44 6 .Math. 3 P2.sub.1/c 0.27 5 .Math. 3 PT 0.45 5 .Math. 4 Pc 0.12
(119) Importantly for self-assembly, the intermolecular H-bonds between the arms establish the pattern of local connectivity for neighboring molecules. It is the global topology, however, of the H-bonded network that is a primary distinguishing feature of a LASO co-crystal. In FIG. 5, a noncovalent connectivity diagram (Etter, M. C. Acc. Chem. Res. 1990, 23, 120-126; Etter, M. C., et al. Acta Crystallogr., Sect. B: Struct. Sci 1990, 46, 256-262; Bernstein, J., et al. Angew. Chem. Int. Ed. 1995, 34, 1555-1573; Motherwell, W. D. S., et al. Acta Crystallogr., Sect. B: Struct. Sci 1999, 55, 1044-1056; Grell, J., et al. Acta Crystallogr., Sect. B: Struct. Sci 1999, 55, 1030-1043; Grell, J., et al. Acta Crystallogr., Sect. B: Struct. Sci 2000, 56, 166-166, all incorporated herein by reference) for three systems (19, 110, 112) is used to illustrate the diverse global topology of the H-bonded network in a LASO platform. This structural feature is partially ascribed to the ability of 1 to engage in adaptive intermolecular recognition. By varying the identity of in these three co-crystals, the binding conformation of 1 (inset of FIG. 5) along with the network topology changes. FIG. 5 shows that topologically, the H-bonding in 19 can be represented by a single 3-D network (FIG. 5a). By contrast, co-crystals 110 and 112 each consist of interpenetrating network topologies. The former is found to be fashioned from a pair of distinct 3-D networks (FIG. 5b), and the latter can be represented as a 3-D network entwined with a series of repeating 2-D networks, respectively.
(120) In addition to the structural characteristics, the choice of the corresponding DA pairs () has a significant effect on the value of (ionicity) for each system (Eddaoudi, M., et al. Acc. Chem. Res. 2001, 34, 319-330, incorporated herein by reference). In Table 3, for the ten LASO co-crystals is shown to vary =0.12-0.89, another result that highlights the modularity of LASO network solids. Three of the systems can be classified as ionic (19, 110, 29), five co-crystals are mixed valent (17, 18, 112, 211, 53), and the remaining two materials are neutral (54, 63). For convenience, the crossed stack systems 17 and 18 are grouped into the mixed valence category. The two crossing stacks (ff and ef) are found, however, to have different values for (Table 3). This dichotomy makes the crossed stack systems a mixed valent/neutral hybrid co-crystal. Four of the eight mixed stack systems (1.sub.9.sub., 1.sub.10.sub., 1.sub.12.sub., 5.sub.3.sub.) display the spectroscopic signature of an asymmetric lattice caused by the dimerization of donors and acceptors. In one-dimensional CT systems, this phenomenon is the result of quantum instabilities, e.g., the Peierls (Torrance, J. B., et al. Phys. Rev. Lett. 1981, 46, 253-257; Torrance, J. B., et al. Phys. Rev. Lett. 1981, 47, 1747-1750; Iwasa, Y., et al. Phys. Rev. B: Condens. Matter 1990, 42, 2374-2377; Bruinsma, R., et al. Phys. Rev. B: Condens. Matter 1983, 27, 456-466; Girlando, A., et al. J. Chem. Phys. 1983, 79, 1075-1085; Masino, M., et al. Phys. Chem. Chem. Phys. 2001, 3, 1904-1910; Horiuchi, S., et al. Science 2003, 299, 229-232; Tokura, Y., et al. Solid State Commun. 1986, 57, 607-610; Girlando, A., et al. Solid State Commun. 1986, 57, 891-896; Collet, E., et al. Science 2003, 300, 612-615; Koshihara, S., et al. Phys. Rev. B: Condens. Matter 1990, 42, 6853-6856; Mitani, T., et al. Phys. Rev. Lett. 1984, 53, 842-845; Tokura, Y., et al. Phys. Rev. B: Condens. Matter 1988, 38, 2215-2218; and Iwasa, Y., et al. Phys. Rev. B: Condens. Matter 1989, 39, 10441-10444, all incorporated herein by reference) and Spin-Peierls transitions (Girlando, A., et al. Solid State Commun. 1985, 54, 753-759; Hughes, R. C., et al. J. Chem. Phys. 1968, 48, 1066-1076; Huizinga, S., et al. Phys. Rev. B: Condens. Matter 1979, 19, 4723-4732; Hasegawa, T., et al. Solid State Commun. 1997, 103, 489-493; Kagawa, F., et al. Nature Phys. 2010, 6, 169-172, all incorporated herein by reference).
(121) Several LASO materials are found to contain structural disorder in the lattice. Systems 18 and 211 both contain asymmetric s that exhibit substitutional disorder. In 29, the -arm of 2 and the aromatic ring system of 9 are found to have positional disorder. The lack of long-range periodicity has a noticeable effect on the self-assembly since 15 equivalents of 9 (Table 1) are needed in the crystallizing solution before co-crystals begin to grow.
(122) The redox potentials of the first electron transfer processes of the individual species 1,9,10 and 12 are recorded in DMF at 298 K using both CV and SWDPV (Table 4). In the case of an irreversible oxidation process for compounds 9 and 10, the redox potential determined by CV is estimated, assuming a one-electron process based on the expected separation between anodic and cathodic peaks for a Nernstian process.
(123) TABLE-US-00004 TABLE 4 Compound CV.sup.a/V C SWDPV.sup.a/V 1.sup.b 0.73 0.73 9.sup.c +0.80.sup.d +0.82 10.sup.c +0.83.sup.d +0.81 12.sup.c +0.40 +0.40 .sup.aCollected at 298 K in argon-purged DMF .sup.bFirst reduction process .sup.cFirst oxidation process .sup.dIrreversible process, estimated assuming a one-eletron process
(124) The LASO platform presented herein is a type of molecular recognition that can amplify the growth of donor and acceptor co-crystals (cm length-scale) under ambient conditions in 3-5 days. The LASO strategy has three components which work cooperatively to promote growth of the network solid. The main constituent is a donor/acceptor (-complement) that uses flexible appendages (diethylene glycol, -arm) to form intermolecular H-bonds in the crystal. The second constituent is a smaller CT partner (-complement) with short and relatively rigid H-bonding functionalities (CO, NH.sub.2, OH, -arm) that is incorporated into the LASO network through molecular recognition with the -complement. The final element is the solvent system that promotes the co-crystallization of the -complement and -complement. With this platform, supramolecular architectures combine the H-bonded network and stacks of alternating donors and acceptors. Not only does the LASO strategy produce network solids that are capable of amplified co-crystal growth, but it creates an entirely new 2-D donor and acceptor packing motif made of two perpendicular CT axes, i.e., the crossed stack.
(125) As further described herein, a molecular design that allows donor and acceptor molecules to self-assemble into CT ferroelectric networks at ambient temperatures is afforded. The co-crystals solve the long-standing challenge that DA mixed stack materials can exhibit a ferroelectric T.sub.c above room temperature. The demonstration of ferroelectric properties in an organic network enables new opportunities to produce these systems into new forms with exciting function such as electrically addressable hydrogels, ferroelectric catalysts, and CT-based sensitizers for photovoltaics, among others. The combination of donor-acceptor interactions with hydrogen bonded networks offers a promising supramolecular platform to design novel organic electronic structures.
(126) The disclosures of all articles and references, including patents, are incorporated herein by reference. The invention and the manner and process of making and using it are now described in such full, clear, concise and exact terms as to enable any person skilled in the art to which it pertains, to make and use the same. All references cited in this specification are incorporated herein by reference. It is to be understood that the foregoing describes preferred embodiments of the present invention and that modifications may be made therein without departing from the spirit or scope of the present invention.
