Structure, dielectric relaxation and electrical conductivity of 2,3,7,8-tetramethoxychalcogenanthrene-2,3-dichloro-5,6-dicyano=l,4-benzoquinone1 :1 charge-transfer complexes? Ulrich Behrens,' Ricardo Diaz Calleja,*b Mark Dotze,b Ursula Franke,' Walter GunDer,' Gunter Klar,*" Jens Kudnig," Falk Olbrich," Enrique Sanchez Martinez,d Maria J. Sanchisb and Barbel Zimmer' "Institut fur Anorganische und Angewandte Chemie der Universitat Hamburg, Martin-Luther-King-Platz 6, 0-201 46 Hamburg, Germany Departamento de Termodinhmica Aplicada, E. T.S.I.I., Universidad Politkcnica de Valencia, Camino de Vera, s/n, E-46071 Valencia, Spain 'Institut fur Physikalische Chemie der Universitat Hamburg, Bundesstraj'e 45, 0-20146 Hamburg, Germany Departamento de Ingenieria Electrbnica, E.T.S.I.T., Universidad Politkcnica de Valencia, Camino de Vera, s/n, E-46071 Valencia, Spain ( 5,10-chalcogena-cyclo-diveratrylenes,2,3,7,8-Tetramethoxychalcogenanthrenes 'Vn2E2', E = S, Se) form isotypical 1 : 1 charge-transfer (CT) complexes with 2,3-dichloro-5,6-dicyano-174-benzoquinone (DDQ). X-ray analysis of Vn, S2-DDQ shows the compound to have a columnar structure with segregated stacks of donors and acceptors. The donors are virtually planar in accordance with a formulation of [VnzE2] '[DDQ] -. Donor cations and acceptor anions are equidistant in their respective stacks, but in each case they inclined to the stacking axis, nevertheless guaranteeing an optimum overlap of the half-filled frontier orbitals which are of n-type character according to MNDO calculations.Dielectric ac measurements of permittivity E' and loss factor E" clearly reveal two processes, a dielectric one at low temperatures and a conductive one at high temperatures. The dielectric process can be described by the Havriliak-Negami (HN) and the Kohlrausch-Williams-Watts (KWW) model, and the conductive process by a Debye-type plot. Using these methods, the relevant parameters are evaluated. The dc conductivities of polycrystalline samples moulded at lo8 Pa show a temperature dependence in the plots of In t~ us. T-l, which is typical of semiconductors. Two slopes are found; that in the low-temperature region (<285 K) is explained by an easy-path model (intragrain conductivity with low activation energies), whereas in the high-temperature region conduction across the grain boundaries (with higher activation energies) is becoming predominant.The activation energies for the intrinsic conductivities obtained by the ac and dc measurements are similar. Despite the columnar structure with segregated stacks, due to stoichiometric oxidation states of the components, the absolute values of conductivity are low (ca. S cm-' at 293 K), though higher (by a factor of ca. lo3)than those of compounds like Vn,E, * TCNQ with stacks in which donor and acceptor molecules alternate. The electron-rich 2,3,7,8-tetramethoxychalcogenanthrenes ( 5,lO-dichalcogena-cyclo-diveratrylenes, 'Vn, E2'; E = S, Se) act as donors in charge-transfer (CT) complexes. With tetra- cyanoethane (TCNE) the complexes 2Vn,Ez * TCNE are formed which in their solid states possess stacks of Vn,E, I TCNE I Vn, E, units., With 7,7,8,8-tetracyanoquinodi-methane (TCNQ), 1 : 1 complexes are obtained in the stacks in which donor and acceptor molecules alternate.3 As a conse- quence of their crystral structures both types of compounds show only poor electrical cond~ctivity.~*~*~ Using 2,3-dicyano-5,6-dichloro-1,4-benzoquinone(DDQ), complexes are formed which are isotypical for E = S and E = Se according to scanning electron microscope (SEM) measure- menk6 The dark blue compounds were easily prepared by combining the hot solutions in acetonitrile.In the case of the sulfur compound, single crystals suitable for a crystal structure determination could also be obtained.Crystal Structure Determination of Vn2S2 DDQ A Siemens P4 four-circle diffractometer (Mo-Ka radiation with II = 71.073 pm, 03-20 scan mode, Lorentz and polarization corrections) and the program Siemens SHELXTL-PLUS (VMS)7 were used for X-ray analysis. The structure was determined by direct methods. Fourier syntheses allowed the positions of all non-hydrogen atoms to be determined; these ?Part 7 of a series entitled Self-stacking Systems. For Part 6, see ref. 1. OaEnO0 E 0 I I H3C CH3 Vn2E2(E = S,Se) NCYCN 0 atom positions were refined with anisotropic temperature factors. The positions of the hydrogen atoms were calculated with fixed distances of 105 pm and isotropic temperature factors.The results are given in Tables 1 and 2.1 Descriptionof the structure Molecular structure. The asymmetric unit of Vn,S2 DDQ contains half a molecule of each component, the molecular $Supplementary data available from the Cambridge Crystallographic Data Centre: see Information for Authors, J. Muter. Chem., 1996, Issue 1. J. Muter. Chem., 1996, 6(4), 547-553 547 Table 1 Crystal structure parameters of Vn,S, * DDQ empirical formula c16 H1604 sZ * c8 c12 N2 OZ crystal system monoclinic space group p2/C alpm 370 8(2) blpm 1306 9(6) clpm 2355 7( 11) fljdegrees 93 50(4) z 2 Mlg mol-' 563 4 V/m3 1139 x lo6 Dlgcy 1 642 Plcm 5 16 scan range/degrees 5 < 28 < 50 independent reflections 2037 reflections with I F,I 2 44 1 Fol) 1548 refined parameters 165 R 0 0473 Rw 0 0436 Table 2 Atomic parameters of Vn,S, * DDQ 0 3579(4) 0 5294( 1) 0 4307( 1) 0 4485 (11) 0 4088(2) 04581(1) 0 5774( 11) 0 3853(2) 0 5138( 1) 0 6589(10) 0 2839(3) 0 5287( 1) 0 6126(9) 0 2072( 3) 0 4892( 1) 0 4765( 10) 0 2307(2) 0 4327( 1) 0 3924( 10) 0 3298 (2) 0 4181( 1) 0 6890( 7) 0 1078(2) 0 4991( 1) 0 8389(11) 0 0823( 3) 0 5550(1) 0 4368( 7) 0 1496( 2) 0 3968( 1) 02780(11) 0 1699( 3) 0 3410( 1) 0 2668(3) 0 9085( 1) 0 1914( 1) 0 2658( 10) 0 7019(3) 0 1969(1) 0 3949(9) 0 7968 (2) 0 2251( 1) 0 3878(10) 06091(2) 0 2254( 2) 0 0632( 7) 0 7022( 2) 0 1533( 1) 0 2696(10) 0 5141(3) 0 2005(1) 0 1723(9) 0 4370( 2) 0 1821(1) structures of which are shown in Fig 1 The outstanding feature of the donor is its planarity, although the thermal parameters of the sulfur atoms indicate that they may be disordered with deviations of up to +20 pm from the best planes of the two aryl rings From this a formulation for the +complex of [Vn, S,] [DDQ] -is suggested Namely, whereas the neutral Vn,S2 molecule is folded at the SS axis [angle of fold (defined as the angle between the normals to the best Q Fig. 1 Molecular structures of the components of Vn,S, -DDQ with atom numbering scheme 548 J Muter Chem, 1996, 6(4), 547-553 Table 3 Mean values of bond lengths (pm) and angles (degrees) in the donor molecule of Vn,S, -DDQ and comparable compounds Vn,S, * DDQ [Vn, S,] + [SbCl,] a Vn, Szb s-c 0-C (ar) 0-C (alk) C(S)-C(S) C(S)--C(H) CW-(30) C(O)-C(O)c-s-cc-0-c s-C(S)- C( S) S -C(S)-C( H) 0-c(0)-C(0) 172 6( 3) 135 2(4) 143 4(4) 140 3(5) 140 3( 5) 137 l(4) 142 8(5) 107O( 2) 116 8(3) 126 3(3) 113 9(2) 115 5(3) 172 3(5) 1349(5) 143 3(5) 139 5(6) 140 O(6) 136 2(6) 143 O( 7) 107 4( 3) 118 O(4) 126 O(4) 114 9(4) 114 5( 5) 177 6( 7) 136 9(8) 142 5( 9) 138 2( 10) 139 7(9) 138 O( 10) 141O( 10) 100 2( 3) 117 6( 5) 121 l(5) 119 O( 5) 115 2(6) 0-C(0)-C(H) C(S)-c(S)-C(0) C(S)-C( H)-C(0) C(H)-C(0)-C (0) 124 7( 3) 119 7(3) 120 4( 3) 119 9( 3) 125 9(5) 119 5(4) 120 O( 5) 119 6(5) 125 2(6) 119 8(6) 120 5(6) 119 6( 6) "Ref 10 bRef 8 planes of the aryl rings), 4 = 131 and 128" for the monoclinic' and orthorhombic' forms, respectively], its monocation [Vn,S,]+ is planar (4 = 180°)10 and any partially oxidized form [Vn2S2Ix+ (0 < x < l), e g 2Vn,S, -TCNE or Vn,S2 TCNQ, has angles between these values These observations correlate well with the oxidation potentials, El, of these acceptors (+0 51 V for DDQ, +0 15 V for TCNE and l2+O 17 V for TCNQ, each vs SCE) In the levelling of the donor molecule the methoxy substitu- ents are included, ze each pair of ortho-standing methoxy groups is coplanar with its aryl ring, both in exo positions, thus facilitating the formation of stacks in the crystal For other derivatives of Vn,S2 this behaviour has also been found,, the consequences with respect to bond lengths and angles have already been discussed in connection with the structure of Vn,S2 itself' (Table 3) A charge transfer will affect the bond lengths and angles of both the donor and the acceptor molecules Indeed, the data of Tables 3 and 4 confirm the formulation of the complex as [Vn,S,]+[DDQ]-since they agree well with those of the corresponding radical ions and differ significantly from those of the neutral molecules This is shown for the donor by the CS and CO distances and the angles in the central dithiin ring, for the acceptor by the CO distance and the beginning of equalization of the distances within the six-membered ring Crystal structure.From the unit cell of Vn,S,.DDQ in Fig 2 it can be seen that a columnar structure with segregated stacks of donor and acceptor radical ions is formed in the crystal The constituents of each stack are coplanar and equidistant, their molecular planes being inclined to the stack- ing axes (Fig 3) The inclination differs in the donor and acceptor columns, thus leading to different interplanar dis- tances, namely 359 and 308 pm for the donor and acceptor stacks, respectively, both distances being shorter than the van der Waals distance (half thickness of an aromatic nucleus, 185 prnI5) In the crystals of [Vn, S,] [SbCl,] * CH, CN'' the radical cations [Vn,S2]+, which are also planar, form different kinds of stacks There are two orientations of the coplanar ions in an alternating sequence ABAB (angle between the molecu- lar axes, 35") The interplanar distance (353 pm) agrees well with that in Vn,S2 DDQ Molecular Orbital (MO) Calculations In order to arrive at a better understanding of the CT interactions in Vn,S, -DDQ, MO calculations were carried out As shown already,I6 good results can be obtained by Table 4 Mean values of bond lengths (pm) and angles (degrees) in the acceptor molecule of Vn,S, -DDQ and comparable compounds Vn, S, -DDQ [NEtJ+[DDQ]-a DDQ~ c-Clc-0 171.5(3) 123.6(4) C-N C-C(N) C( CN)- C(CN) C( CN)- C(0) N-C-C C( Cl) -C( C1) c(C1) -c(0) cl-c-c(cl)c1-c-C(0) 0-c-C(C1) 114.7( 5) 143.0( 5) 136.8( 6) 138.6( 7) 147.2( 5) 144.5( 5) 178.0( 4) 121.6( 1) 115.8(2) 122.5( 3) 0-C-C(CN) C( N)-C- C( CN) C(N) -C-C(0) C (C1)- C( 0)-C( CN) C(C1)-C(CN)-C(CN) C( C1)- C( C1) -C(0) 123.0( 3) 119.7(2) 11 7.4( 3) 122.6 (2) 114.5( 3) 122.9( 2) "Ref.13. bRef.14. , f i(I b Fig. 2 Unit cell of Vn,S, * DDQ HAM317*'s calculations, whereas the MND019?,' method nor- mally gives less reliable values for unoccupied orbital^,^'-^^ i.e. too high energy differences between the HOMOS of the donor and the LUMOs of the acceptor molecules are found by MNDO calculations. Nevertheless, with respect to the orbital symmetries, the same energy orders are obtained from both methods, although the absolute values of the orbital energies differ. Since HAM3 parameters have not yet been determined for elements beyond the second row of the periodic table we decided in favour of MNDO calculations. Furthermore, 2,3,7,8-tetrahydroxythianthrene was taken as a model compound of Vn,S,.This seemed tenable, because the expenditure of calcu- 171.6( 3) 169.7( 3) 124.6 (4) 120.3( 3) 114.0(4) 113.4(4) 143.0(5) 143.6(4) 136.3( 4) 133.9( 4) 138.6(4) 134.3(4) 146.3 (4) 148.2( 4) 144.4 (4) 149.7( 4) 178.4(4) 178.5( 3) 121.8 (2) 122.8( 2) 115.7( 2) 115.5(2) 122.6( 3) 123.3( 2) 122.6( 3) 119.8(2) 120.7( 3) 122.8(2) 116.8(3) 115.9(2) 122.6( 3) 121.7( 2) 114.9( 3) 117.0( 2) 122.6( 3) 121.3(2) lations for the tetramethoxy derivative would be out of all proportion to the attainable improvement in the results. The frontier orbitals in question, i.e. the HOMO of Vn,S2 and the LUMO of DDQ, are z-type orbitals and therefore very suitable for CT interactions between parallel planar molecules.These MOs are semi-occupied in the radical ions [Vn,S2]+ and [DDQI-. In each column only orbitals of the same type interact and an optimum overlap would be guaran- teed when the molecular planes are perpendicular to the stacking axes. However, probably to achieve closer packing, the molecular planes are inclined to the axes, but apparently in such a way that the MOs overlap quite well in these arrangements also (Fig. 4). Dielectric Relaxation of the CT Complex Experimental Dielectric relaxation measurements by the conventional ac technique were carried out with a Genrad 1689M bridge at 20 frequencies between lop2 and 102kHz in the tempera- ture range 167-360 K for Vn,S2 -DDQ and 220-420 K for Vn,Se, -DDQ, with steps of 5 "C.Each sample was moulded into a disc-shaped pill of diameter 1cm and thickness 1mm. Dielectric ac measurements of permittivity, E', and loss factor, E", clearly show two processes: at low temperatures a dielectric process and at high temperatures a conductive one. In Fig. 5, plots of E', E", us. T are shown for various frequencies in the range of the dielectric low-temperature process for both compounds. From an Arrhenius plot, lnf (at the maximum E") us. T-' (Fig. 6) we obtain values of 0.22 and 0.19 eV for the energies of activation, E,, of Vn2S2 -DDQ and Fig. 3 Segregated stacks of donors and acceptors in Vn,S, DDQ: (a) top view; (b) side view J. Muter. Chem., 1996, 6(4), 547-553 549 Fig. 4 (a) Relative positions of two successive donors and acceptors in the stacks of Vn,S,-DDQ, (b) overlap of the corresponding frontier orbitals Vn,Se, -DDQ, respectively In order to obtain more detailed information about this relaxation we have applied the Eynng equation /=-exp(g)kT 27th where k, h and R are the Boltzmann, Planck and gas constants, respectively, and AG is the Gibbs free energy of the barrier to the relaxation process, which is related to the activation enthalpy AH and activation entropy AS by AG=AH-TAS This leads to AS AHIn -f = In-k + ---T 2nh R RT where AH and the activation energy, E,, given by the Arrhenius equation are related by E, =AH +RT The values of AH and AS were determined directly from ln(f/T) us 1/T plots (Fig 6) and the results obtained are summarized in Table 5 Stark~eather~~held simple relaxations responsible for the processes of low activation entropy The dielectnc relaxation process is customarily represented in terms of E" us E' (Cole-Cole plots) Whereas for Debye-type peaks the curves are semicircles, the complex diagram plots representing the dielectic results associated with the dipolar relaxation are skewed arcs, which in many cases approach the real axis through a straight line, descnbed by the Havriliak- Negami (HN) equation 25 (3) where E~ and E, represent the relaxed and unrelaxed dielectric permittivity, respectively, of the relaxation process, zo is the relaxation time and a,y are parameters related to the shape and skewness of the complex dielectric plot (a is a parameter characterizing a symmetrical broadening of the distribution of relaxation times and y characterizes an asymmetrical one) A non-linear squares regression (NLSR)26 was used to improve the data fit The equivalent electric circuit (in series configuration involving a condenser, C, and an HN-type impedance, Z,, = [1+(~ozo)ol]y/zo[Co-C,]) shown in Fig 7 Table 5 Eynng equation parameters Compound AHfeV EaIeV ASlmeV K-' Vn, S2 DDQ 0 2037 0 2212 0 0028 Vn2Se, * DDQ 0 1772 0 1946 0 0027 550 J Muter Chem, 1996, 6(4), 547-553 150 200 250 300 I50 f ex %* xx X8 04 150 200 250 300 350 TI K X. :x0.xxe 0. % 4, 10 0 Fig. 5 Temperature dependence of the dielectnc permittivity E and loss E of (a) Vn,S2 * DDQ and (b) Vn,Se, DDQ at various frequencies (a,100Hzl x, 50Hz, U, 20Hz1 +, 10Hz, +, 5Hz) was employed in order to fit the experimental data, and the best set of parameters obtained for both compounds at different temperatures is given in Table6 The quality of the fit is demonstrated in Fig 8 The application of the HN equation to the dielectric loss data, ~"(o),provides an adequate functional form for the calculation of the dipolar correlation function, #(z) This is due to the fact that the dielectric permittivity is related to the dipole moment time correlation function, #(z), by a one-sided Founer or pure imaginary Laplace transformation 8-4-Cls a -4 --8 ! .I I 3 4 5 T-*/K-l Fig.6 Temperature dependence of fErr, (a) according to an Arrhenius equation (In f,~,, (b) according to an Eyring equation for [In us. TI); (f,,t/T)us. TI]for Vn,S,.DDQ (0,0)and Vn,Se,-DDQ (0,m) Fig. 7 Electrical circuit representing the dielectric process at low temperatures 0 20 40 60 80 6’ Fig. 8 Cole-Cole plot at 220 K for Vn, S2* DDQ (0,experimental data; a,calculated data) where o (=2nf)is the angular frequency and the total dipole moment time correlation function, #(z), is given by:27 where pk is the dipolar moment of each entity (dipole or dipole groups), assumed to be the same for all the entities. The time correlation function d(z) is obtained by a half- sided cosine Fourier transformation: E”O) cos wt dw where Ae =go -E, is the relaxation strength.The numerical evaluation of the preceding equation can be performed by depicting the dipolar process using the HN equation, for which: ~”(0)=AEr -sin yY r2= 1 +2(0z)” cos (a -3+ (OZ)~~ (7) (ozysin (a9) tan Y = I. 1 +(Ozy cos (a ;) where a, y, zo are parameters obtained with the fit of E* to the HN equation. In order to carry out the fit, we chose the Kohlrausch- Williams-Watts (KWW)28,29 function, given by: where zKWw,(a characteristic relaxation time) and ,8 (0 < P < l), (a parameter that describes the non-exponential character of the correlation function) are summarized in Table 6. Note that whereas the P parameters obtained from the KWW function remain practically constant over the entire temperature range, the distribution parameters, a and y, of the HN equation vary significantly with temperature.This is a consequence of the fact that different pairs of a and y values correspond to each /3 value because KWW is a single-parameter function whereas HN is a two-parameter one. The parameters of both functions are related by eqn. (9): by)’ =P (9) with c =0.95 and 1.28 for Vn2S2 DDQ and Vn2Se2 DDQ, respectively. Curves describing the evolution of this function with time at several temperatures are shown in Fig. 9. In the high-temperature process, E” continuously increases with decreasing frequency and no relaxation peak is seen. For this reason, the data were modelled using the electrical modulus Table 6 Havriliak-Negami equation parameters (A&,a, y, z~~)and Kohlrausch-Williams-Watts parameters (B,zKWw)for the CT complexes Vn, S2* DDQ 190 3.08 118.83 121.91 0.79 0.72 2.15 x 10-4 0.584 1.45 x 10-4 200 2.65 117.32 119.97 0.79 0.69 1.17 x 10-4 0.569 7.38 x 10-5 210 2.67 118.31 120.98 0.76 0.71 7.02 x lo-’ 0.557 4.78 x 10-5 220 2.79 121.48 124.27 0.78 0.70 4.75 x 10-5 0.561 3.08 x lo-’ Vn2Se2* DDQ 220 6.36 75.12 81.49 0.72 0.90 2.58 x lo-’ 0.570 2.72 x lo-’ 230 8.29 65.53 73.82 0.77 0.96 1.38 x lo-’ 0.669 1.47 x 10-5 240 9.88 62.48 72.36 0.79 0.87 6.42 x 0.655 5.89 x J.Muter. Chem., 1996, 6(4), 547-553 551 formalism according to M* =(&*)-l,where M'= &' (&')2 + (&")2 M" = E (&')2 + In the complex electric plot of M" us.M' a nearly exact semicircle was obtained as can be seen in Fig. 10. This proves I .00 T2 0.60 a 0.20 1 -15.00 -10.00 -5.00 0.00 5.00 w Fig. 9 Normalized correlation function calculated according to HN equation and KWW equation for Vn,S,.DDQ (0,0) and Vn,Se, * DDQ (0,U)(open symbols, experimental data; filled sym- bols, calculated data) 0.00) * 0 0-0.m mr 1 9 ?o o.oO0 ! 1I 1 I O.Oo0 0.002 0.004 0.006 QOOB M' Fig. 10 M' us. M" (Cole-Cole) plot at 360 K for Vn,S, -DDQ (0 experimental data, 0,calculated data) Fig. 11 Electrical circuit representing the conductive process at high temperatures that the process at high temperature is purely conductive. Accordingly, and following impedance spectroscopy tech-niques, we tried to fit the experimental data to the electrical model circuit given in Fig.11 and Table 7. Electrical Conductivity of the CT Complexes Experimental The electrical conductivity of Vn,S2 DDQ and Vn2Se2 DDQ were measured with an HP-4329-A electrometer together with a Guildline 6500 teraohmmeter by using silver-coated elec- trodes on samples moulded at ca. lo8Pa. The temperature range was 140-370K; each temperature at which measure- ments were taken was kept constant by use of a Pt-100 resistance thermometer and an Eurotherm 820 controller. Results The results of the conductivity measurements are given in Fig. 12. The absolute values of the conductivities are relatively low, but in both cases the temperature dependence of the conductivity can be described by eqn.(11): CJ =o0exp(-Ea/2kT) (11) b = -24 --201 Fig. 12 Arrhenius plot of In gac7In ad"vs. 1/T for Vn,S, -DDQ (0,0) and Vn,Se, -DDQ (El, U)(open symbols, experimental data; filled symbols, calculated data) Table 7 Parameters for the model of Fig. 11 giving the best fit with the experimental data for both CT complexes at different temperatures TIK Vn, S, -DDQ Vn, Se, * DDQ R cx 109 Ma R cx 1o'O MOJ 3 30 1.32 x 107 0.1287 0.00777 340 8.84 x lo6 0.1288 0.00776 1.98 x 107 0.9132 0.01095 350 8.55 x lo6 0.1314 0.0076 1 1.58 x 107 0.8974 0.01 114 360 6.43 x lo6 0.1327 0.00754 1.47 x 107 0.8870 0.01127 3 70 1.43 x 107 0.8708 0.01 148 552 J.Muter. Chew., 1996, 6(4), 547-553 Table 8 Conductivity data for some Vn,E, derivatives Vn,S, * DDQ Vn, Se, -DDQ 0.58 0.32 2.09 0.012 4.88 x 8.60 x Vn,S, * TCNQ" Vn,Se2 -TCNQ" 0.78 0.84 0.04 0.012 6.12 x lo-'' 8.20 x 10-9 2VnzSz * TCNEb 1.27 0.95 1.11 x lo-" 2Vn2 Se, -TCNE~ 1.17 0.96 8.32 x lo-'' Vn,Se, * Br; 1.07 0.0025 4.3 x 10-5 DNDTd 0.96 738 7.0 x lo-', "Ref. 3. bRef.2. 'Ref. 31. dDNDT=dinaphtho[2,3-b;2',3'-e][ 1,4]-dithiin-5,7,12,14-tetra0ne.~' where k = Boltzmann's constant, E, =gap energy, i.e. acti- vation energy of the electrical conductivity, and go= pre-exponential factor, i.e. conductivity at infinite temperature. This equation describes behaviour typical of semiconductors.The activation energies, E,, obtained from the slope of the Arrhenius plot (In CT us. 1/T) of the two conductivities (ac, dc) for both compounds are given in Table 8. We can see in Fig. 12 that the representation of dc conductivity displays two regions (at approximately 285 K the slope for both compounds changes significantly): a low-temperature region where the two activation energies are similar (Ea(ac)=0.50 eV, Ea(dC)= =0.58 eV for Vn2S2 -DDQ, Ea(ac)=0.31 eV, Ea(dC) 0.32 eV for Vn2Se2 DDQ) and a high-temperature region where E, rises to approximately 0.83 eV and 0.73 eV for Vn2S2 DDQ and Vn, Se, -DDQ, respectively. The first region may be explained in terms of an easy-path model and the second region is reached when the conduction through the grain boundaries (having a higher activation energy) exceeds the intragranular conduction.This result (two regions in Arrhenius plot) has also been observed by other authors for polycrystallinematerial^.^^*^^ The E, values (intrinsic conductivity) obtained for both compounds are of the same magnitude as the values of similar compounds such as Vn2S2 TCNE (0.78 eV) and 2Vn2S2-TCNE (0.95 eV). Owing to the fact that the charges of the donor and acceptor molecules are stoichiometric (+1 and -1, respectively), the conductivities of the compounds Vn2E2 DDQ (E = S, Se) are not very high, although they are significantly higher than in other compounds, such as Vn, E2 TCNQ and 2Vn, E2 -TCNE (Table 8).The central problem with ac measurements arises from the interpretation of the data. This is because the sample and electrode arrangement is electrically a 'black box', the equival- ent circuit of which (i.e.its representation by some combination of R and C elements) is often unknown. The crux of the problem in analysing ac data is (a) to determine the appropriate equivalent circuit for the cell and (b)to evaluate the various R and C components in the electric circuit. In polycrystalline materials the overall sample resistance may be a combination of the intragranular resistance (or bulk crystal resistance) and the intergranular (or grain boundary) resistance. Both resist- ances are parallel with an associated capacitance, and each parallel RC element gives rise to a semicircle in the complex plane z*.In the frequency range in which we worked, we observed only one semicircle associated with grain interior conductivity (intrinsic conductivity); for this reason, we employed the simple parallel RC element (Fig. 11). This work was supported by the Acci6n Integrada Hispano- Alemana (no. HA94-103, 322-ai-e-dv). We also thank the Volkswagenstiftung and the Fonds der Chemischen Industrie for financial support. M.D. gratefully acknowledges the grant by DAAD (NATO-fonds) M.J.S. acknowledges the Conselleria de Educacion y Ciencia de la Generalitat Valenciana for a grant. References 1 J. Behrens, W. Hinrichs, T. Link, C. Schiffling and G. Klar, Phosphorus Sulfur Silicon, 1995,101,235.2 P. Berges, J. Kudnig, G. Klar, E. Sanchez Martinez and R. Diaz Calleja, 2.Naturforsch., Teil B, 1989,44,211. 3 W. Hinrichs and G. Klar, J. Chem. Res., 1982, (S)336; (M)3540. 4 W. Hinrichs, P. Berges, G. Klar, E. Sanchez Martinez and W. GunDer, Synth. Met., 1987,20,357. 5 E. Sanchez Martinez, R. Diaz Calleja, W. GunDer, P. Berges and G. Klar, Synth. Met., 1989,30,67. 6 W. Gunoer, J. H. Henning, G. Klar and E. Sanchez Martinez, Ber. Bunsenges. Phys. Chem., 1989,93, 1370. 7 G. M. Sheldrick, SHELXTL-PLUS, Release 4.21/0, Siemens Analytical X-Ray Instruments, 1990. 8 W. Hinrichs, H-J. Riedel and G. Klar, J. Chem. Res., 1982, (S) 344; (M)3501. 9 H. Bock, A. Rauschenbach, C. Nather, Z. Havlas, A. Gavezzotti and G.Filippini, Angew. Chem., 1995,107, 120; Angew. Chem., Int. Ed. Engl., 1995,34,76. 10 W. Hinrichs, P. Berges and G. Klar, 2.Naturforsch., Teil By1987, 42, 169. 11 M. E. Peover, J. Chem. Soc., 1962,4540. 12 R. C. Wheland and J. L. Gillson, J. Am. Chem. SOC.,1976,98,3916. 13 G. Zanotti, A. Del Pra and R. Bozio, Acta Crystallogr., Sect. B, 1982,38,1225. 14 G. Zanotti, R. Bardi and A. Del Pra, Acta Crystallogr., Sect. B, 1980,36, 168. 15 Handbook of Chemistry and Physics, ed. R. C. Weast, CRC Press, Cleveland, OH, 1977-1978,58th edn., p. D-178. 16 E. Sanchez Martinez, R. Diaz Calleja, P. Berges, J. Kudnig and G. Klar, Synth. Met., 1989,32,79. 17 L. Asbrink, C. Fridh and E. Lindholm, Chem. Phys. Lett., 1977,52, 63; 69, 72. 18 L. Asbrink, C.Fridh and E. Lindholm, QCPE Bull., 1980,12,393. 19 M. J. S. Dewar and W. Thiel, J. Am. Chem. SOC., 1977, 99, 4899; 4907. 20 W. -Thiel, QCPE Bull., 1982, 2,438. 21 L. Asbrink, C. Fridh and E. Lindholm, Chem. Phys., 1978,27,169. 22 C. Fridh, L. Asbrink and E.oLindholm, Chem. Phys., 1978,27,159. 23 E. Lindholm, G. Bieri, L. Asbrink and C. Fridh, Int. J. Quantum Chem., 1978,14,737. 24 H. W. Starkweather, Jr., Macromolecules, 1981, 14, 1277; 1990, 23, 328. 25 S. Havriliak and S. Negami, Polymer, 1967,8, 161. 26 J. Ross McDonald, Complex Nonlinear Least Squares Immitance Fitting Program, LE VM6, 1993; Impedance Spectroscopy, Wiley-Interscience, New York, 1987. 27 G. Williams, Chem. SOC. Rev., 1978, 7, 89. 28 G. Williams and D. C. Watts, Trans. Faraday SOC.,1970,66,80. 29 R. Kohlrausch, Pogg. Ann. Phys. Chem., 1884,91,179. 30 A. R. West, Solid State Chemistry and its Applications, Wiley, Chichester, 1984, ch. 13. 31 E. Sanchez Martinez, R. Diaz Calleja and G. Klar, Synth. Met., 1990,38,93. 32 E. Sanchez Martinez, R. Diaz Calleja, J. Behrens, P. Berges, J. Kudnig, N. Wislki and G. Klar, J. Chem. Res., 1991, (S) 246, (M)2379. Paper 5/06484H; Received 2nd October, 1995 J. Muter. Chem., 1996, 6(4), 547-553 553