Pure & Appl. Chem., Vol. 62, No. 3, pp. 395406, 1990. Printed in Great Britain. @ 1990 IUPAC

Aromatic/quinoid systems: principles and a p pl ica t io ns

Siegfried Hunig

Inetitut fur Organieche Chemie, Universitat Wunburg, D-8700 Wunburg, Federal Repu- blic of Germany

- A general structural principle b prsllented for reversible two step redox system which contain, on an intermediate oxidation level, radical ions of high thermodynamic stability. Although this principle is of much broader Kope, the paper concentrates on (hetero)aromatic/quinoid Wumter and Weitc type syskms and some biasulenes. These system can be characterized in solution by their redox potentials and applied e. g. as electron transfer catalysb. In addition it is demonstrated that moet organic electric con- ducting materials belong to theae general redox system, from which a new type of acceptor of N,N'-dicyanoquinonediiminea (DCNQI's) has been developed. DCNQI'e are obtained in great variety from substituted quinones in a one pot reaction. Typical examples for CT- complexea and aspeeidly radical anion dta of very high and partly metallic conductivity are diecuesed.

1 A STRUCTURAL PRINCIPLE FOR REVERSIBLE TWO-STEP ELECTRON TRANSFER SYSTEMS

Baeed on earlier suggestions of Weitr [l], Hunig [2] proposed a general structural principle for reversible organic redox system where the redox process (i) c o ~ i e t a of electron transfer only and (ii) occurs stepwise with an intermediate radical ion of considerable thermodynamic stability (Scheme 1). The redox steps RED/SEM and SEM/OX can normally be charackrised by the potentials El and Ea from which the eemiquinone formation constants K ~ E M can be calculated.

ox -e + e

SEM - e RED ==

+e El €2

*

Scheme 7 General structurea for reversible two step redox system.

Systematic inwtigat iom have demonstrated that this principle ir valid a very broad range of etructural variations (3,4]. It can even be extended to certain polyen-. In addition, t h b principle explains the otherwise unexpected redox properties of many rather odd examplea reported in the literature [3]. Furthermore this structural principle includm the fact that within the three oxidation stagea the m a t delocalized system is connected with the radical ion which therefore dieplays the longset wavelength absorption, often shifted to the IR region. This paper, however, will be reatricted to those ryetern, which show aromaticity in the forms RED or OX, whereas the correaponding oxidation levels OX and RED are "quinoid". By t h b way, two general classea can be deduced from Scheme 1 [3]:

1. Wursler f y p e syslemr, consisting of rings, built from the vinylene groups, to which the end groups X or Y

2. Weitr type 8y8tem4 coneisting of r i n g into which the end groups X and the vinylene groups (at least are attached and which are aromatic in the reduced form.

partly) have been incorporated and which are aromatic in the oxidised form,

395

396 s. HUNIG

solvent:

Both system are repreaented by the clwsical axamplea of pphenyknediaminea together with p-benquinone in Scheme 2 and by the quaternary Sate of 4,4'-bipyridine ("viologenea" [l,S]) in Scheme 3.

RED €1 SEM E2 OX

\jJ0 \ * / \ 0 +$sQ 0 k 0" 0 \

Scheme 2 Example0 of Wureter type redox SyStCm.

RED €1 SEM €2 OX

lox + RED Kwr 2 SEMI

Scheme 3 Example of a Weitz type redox system.

2 VARIATIONS OF THE WEITZ-TYPE REDOX SYSTEMS

Weitz type redox system have been systematically studied over the laat 25 years [2] and are still under active investigation [el. Here the numerow variations will be demonstrated only by three typical examplea.

1. Replacement of nitrogen by sulfur or oxygen in the above mentioned viologenea has two consequences aa shown in Scheme 4: K ~ B M decreaaea from lo' to lo' and the potentials move to the positive side of the voltage scale. Thie meam, that the viologenea are air stable in their oxidized form whereas the mrrreaponding pyron and thiopyron derivative are ieolated in their reduced form.

2. Replacement of two central methine groups by an aze or mino group, reapectively, yields redox systems in which all three oxidation fo rm can easily be isolated. The wide variability of the heterocyclic end group and the high stability of the radical cations is demonstrated by Scheme 5.

9 3 x 1 0 ' H

2 . i x i 0 4 H

-1.0 -0.5 0.0 +0*5 + O A

Scheme4 Dependence of redox potentiah El (RED/SEM, left bat), Ea (SEM/OX, right bar) and eamiquinone formation conrtanb KsgM on different heteroatom.

blue green

Scheme 5 Radical catione of two step redox system with central NN-moiety and different heterocyclic end groups.

Arornatic/quinoid systems: principles and applications 397

99% 7x10'

3.

0

So far, all the end g r o u p X and Y act by their lone pairs in the reduced form. It should be possible, however, to substitute them by certain r-system. By this way, one might obtain unsaturated hydrocarbons, which should undergo either two step oxidation or reduction. A suitable arrangement is accomplished by two azulene moietiea which are connected in p i t i o n s 1,4 or 6 by vinylene groups (n = 0-4). Scheme 6 demonstrates the redox situation for bisazulenes, linked by one vinylene group in positions 1 or 6. In accordance with empirical knowledge and backed by the calculated resonance energies [7] for different oxidation levels of C6As and C7H7, the redox behaviour of the bisazulenes is expected to depend strongly on the position of linkage.

95% 90% 8 1 % 72%

f i RE@) -0.569 +0.049 +0.667

1x103 1 4x102 2 8x10' 3 3x10' 4

w @ 0 @ +0.203

RE(B) -0.243 +0.202 +0.647

Scheme 6 Poeeible two electron oxidations and reductions of 1,l'- and 6,6'-biazulenyCethenes together with calculated resonance energies of five and seven membered rings.

Indeed, the 6,6'-derivativea can only be stepwise reduced and cannot be reversibly oxidized due to strong anti- aromaticity of the cyclopentadknyl cation. By constraat, the much lower antiaromaticity of the cyclohepiatrienyl anion d w not prevent 1,l'-linked biazulene syyrtem, which are easily oxidized, frpm also being reversibly reduced. Thus, a pure hydrocarbon can exist in five different oxidation levels, as can be seen from Scheme 7. Again the thermodynamic stabilities of both the radical cations and anions are remarkably high.

Ef + e 1 -e SEM;@

T - -e +e

- -

1 x 103 95% i 9 x 1 0 ° ( 59%

l x l O o 31%

Scheme 7 for reduction (K,) and oxidation (K.).

Redox reactions of vinylogous 1,l'-biazulenea together with semiquinone formation constants

It should be mentioned that reversible four step system can even be derived from polyenes with croeecoqjugated end g r o u p [8].

398 s. HUNIG

3 ELECTRON TRANSFER CATALYSIS BY REVERSIBLE TWO-STEP REDOX SYSTEMS

Since the electron t r d e r betwean the different oxidation stages of the eyetems under discussion occurs at nearly diffusion controlled ratea (91, and becauee the potential can be modeled to cover a broad range, electron transfer catalyeia should be possible. Inded, important applications are found in rather different fields: E. g. unselective herbicidea (ParaquatR, DiquatR), activation of developing agente in photography [3], and reductive debromination of 1,2-dibromidea [lo]. The m a t striking reported case t the reduction of a-ketoacids or oxidation of a-hydroxyacids with complete enantiodectivity by micro organisms [ll). According to Scheme 8, viologenea with R = Me act mediators for reduction (electrons provided by a cathode, carbon monoxide or formic acid), whereas the more easily reducible viologenea with R = CH2-CONH2 mediate the enrymatic oxidation (electrones removed by an anode or dioxygen).

R - -CHI 4

Et - -440 mV

RED-2H SEM

R F W - R - -

Proteus vulgaris

ox

ox

/O R - -H&-C< 4 - -295 mV

NHZ

Scheme 8 Encymatic electron transfer reactions mediated by viologenes. Reduction and oxidation depend on the potentials of the mediatore.

These glimpeea on a variety of two e k p redox eyetern already give a taste of their fruitful chemistry, which, however, has to be studied and applied throughout in solution. In the meantime, electron transfer reactions of theee revedble redox sys tem in the eolid state are becoming even more important, as will be demonstrated in the following chapters.

4 REVERSIBLE TWO-STEP REDOX SYSTEMS - T H E BASIS FOR CONDUCTING ORGANIC MATERIALS

With very few exceptions the majority of organic materials are electric insulators. Therefore, the discovery of the CT-complex from 'ITF and TCNQ, which rhowed high metallic conductivity, rising to - lo4 Scm-I around 60 K, at which point a metal-semiconductor transition occured, immediately created an extremely rapid growing field of research [12).

TCNQ

K S E ~ 2x10'

Scheme 9 CT-complexTCNQ/TTF together with the eemiquinone formation conrtante K ~ E M and the corre- sponding redox potentiah.

Aromatidquinoid systems: principles and applications 399

As expected, a plethora of similar system hw been teetad for electric conductivity, but perhaps not all chemists have realized that both the powerful donors and acceptors repreeent examplea of the general structural principle for two step redox systems: Extremely high thermodynamic stability of the radical ions - expressed by K S E M - is observed for TCNQ (Wurster type - Ig K ~ E M 9.2, [13], an well as for T T F (Weitz type - Ig K ~ E M 6.2,

In the meantime well eatablihed rules [16] have provided guidelines for the design of organic metals: E. g. the organic components should be planar r-systems which in CT-complexes crystallise in segregated stacks. Electric conductivity along these stacb b rendered by partial electron transfer, provided by a difference of 0.25 to 0.34 V [16,17,18) between the potentials OX/SEM of the acceptor and RED/SEM of the donor. Moat of the variations reported in the literature concern the donor T T F (substituents, Se and Te instead of S) from which even superconducting radical cation salts [IS] with T, up to 11 K ([20]) have been isolated. By contrast, very few substituted TCNQ’s were checked, presumably for the following reasons: (i) The Y-shaped and rigid =C(CN)2 group toleretea only very small substituents to preserve the planarity of the molecule; (ii) interstack interactions which are prerequisite for the striking properties of radical salts based on TTF-derivatives (e. g. BEDT-TTF, 121,221) were not expected for TCNQ’a. Nevertheless, we think that still larger variations of the acceptor are highly desirable and may even lead to radical anion salts of outstanding properties.

~ 4 1 ) .

5 DCNQl’s - A N E W CLASS OF ACCEPTORS

Since we already knew that a certain analogy between the groups =C(CN), and =N-CN had been proposed [23,24], we set out to create the 80 far unknown class of N,N’-dicyanoquinonediiminee (DCNQl’s, [25]) which, for the following reaeona, immediately caught the interest of several groups:

1. The possibility of a one pot syntheais starting from 1,4-benzo- and naphthoquinones as well as 9 , l h n t h r a - quinones and bie(trimethyleilyl)carbodiimide in the presence of titanium-tetrachloride produces DCNQI’s in good to excellent yield [26,27,28]. By this route a broad spectrum of DCNQl’s can easily be obtained.

+R1-R4 (H;IC)3Si - N-C-N -Si(CH3)3

Tic14

CN

NC‘ :$: qCN+ RS

CN ‘C N \

I N D E A ‘CN NC’ B

R-Me,Et,nPr,iPr,tBu,Ph,F,CI,Br,I,OMe,SMe

Scheme 10 General synthesis of DCNQI’s and typical examples A - E

2. Since the =N-CN group ia no longer Y-shaped and i b bond angle is somewhat flexible, planarity is even preeerved on tetrwuktitution (C, E, [la]). Disubetitution M in A, B and D gives rise to definite anti and syn configuration, respectively, of the CN group, whereas on tetrasubstitution (e. g. C ) a low energy conversion between the two configurations (ca. 13 kcal/mol) [30] occura in solution.

3. Because of t h b variability, the acceptor strength for the unsubstituted DCNQI which is similar to TCNQ can be modified between ca. -0.1 V to +0.8 V (Ag/AgCI,CB&I2) [31].

From these DCNQI’s many CT-complexes and especially radical anion salts have been prepared whoa conducting propertiea equal or even surpass t h o a of the corresponding TCNQ derivatiw.

400 s. HUNIG

6 CT-COMPLEXES FROM DCNQl's A N D TTF

So far more than two dozens of DCNQI-TTF-complexes have been obtained. M a t of them show powder conductivities of lo-' - lo-' Scm-' ([28]). The segregated stacks are arranged in two different patterns as demonstrated by the following two examples (Figs. 1,2). In Fig. 1 a chess board like arrangement [32] is realized in which the water molecules Beem to occupy the place of a cyan0 group in the corresponding TCNQsTTF complex.

Fig. 1: DCNQI.TTF.2H20 space group Fig. -2: DCNNQsTTF space group C2/c; a,c-projection P21 /c; b,c-projection (a= 10 Scm- I ) (u=25 Scm-')

By contrast, the DCNNQ.TTF complex [33] is compoeed of two parallel rows of both DCNNQ and TTF mole- cules. Both stacks are skewed in the same direction, but with a slightly different angle, thereby allowing two sulfur a t o m of one TTF molecule to approach the CN groups of two stacked DCNNQ moleculea as cloee 318 and 323 pm, reapectively. The conductivity of the complex increws down to T, = 140 K, below which it drops sharply due to a phaee transition.

7 DCNQI RADICAL A N I O N SALTS WITH ORGANIC CATIONS

Both sp3-nitrogen (N+Me4) and sp'-nitrogen (N-methylquindinium ion (NMQ)) can serve ( ~ 8 counter ions in conducting DCNQI-radical anion salts with stoichiomtries 2:l and 3:2. Fig. 3 preeents the crystal packing of [2,5-Cla-DCNQlja[NMq] which shows sig sag stacks of equidistant acceptor molecules [34]. The cations appear to be situated in the holea of the crystal lattice without any specific interaction with the radical anions.

Fig. 3: Crystal packing of [2,5-Cla-DCNQI],[NMe4] space group C 2/m; (u = Scm-')

A different pattern is observed with [2,5-Br,-DCNQlj3[NMQ]a according to Fig. 4 [35]. Slightly shifted trimers of DCNQI's form stacks which are aeparated by stacks of equidistant N-methylquinolinium (NMQ) cations.

Aromatidquinoid systems: principles and applications 40 1

1

3 --'2

Fig. 4: Crystal packing of [~,~-B~z-DCNQI'JS[NMQ]Z space group Pi (u = 0.1 Scm-' (powder))

8 DCNQI RADICAL ANION SALTS WITH METAL CATIONS

This clam of compounds is the most interesting one, especially dab of 2,5diiubatituted DCNQI'e to which this report will be reatricted [34,36,37]. An one can judge from llbble 1, high conductivities are achieved with a variety of metal monocations culminating at 1000 Scm-' for copper dab. It is worthwhile to note that high conductivity ie not restricted to cationr which strongly coordinate with cyan0 groups (Cu+,Ag+). hrthermore, the 2,5-substituente CM be varied within wide limits. Even disorder by two different 2,S-substituente is tolerated and only bulky substituente, e. g. iodine, wem to reduce the conductivity.

Table 1: Room Temperature Powder Conductivitiea of (2-R',5-RS-DCNQI)~M Anion Redical Salta (*Single Crystal, +for similar results aee (36,371)

It turns out that all d b poeeeee the same crystal structure (141/a or a cloeely related space group), allowing for the first time to study a aeries of the radical sdta with the m e crystal lattice and to correlate the obmrved effects with the structural variations. T h e e sslb introduce a new stacking motif into organic conducting materials M

one can see from Figs. 6 and 6. According to Fig. 6, the cations are arranged like a string of pearl, being surrounded by four stacks of DCNQI ligands in such a way that tetrahedral coordination ie allowed at each cation. The dbtancea between the cations (378-397 pm) do not correlate with the nature of the metal since they are dictated by the thickness and the skew angle of the organic ligands. T h b distance is much too large for electron transport along the metal ion stack. Therefore one dimeneional conductivity occure through the.DCNQ1 stacks only and not through the metal ions.

402

Y ii 10’

lo0

lo-’

lo-’

s. HUNIG

Fig. 5 : a,b-projection;

[2,5-Mel-DCNQI] &u, space group I4 1 a

Fig. 6: perspective view of one stack of copper ions surrounded by four stacks of DCNQI’s

As can be Been from Fig. 5, however, due to the bifunctionality and the special geometry of the DCNQI units, each ligand column is connected to two cation columns, thereby creating a network of parallel stacks of radical anions and cations. The very special but still general arrangement may even allow conductivity perpendicular to the axis of stacking, if the metal ions can easily change their oxidation levels a t the appropriate potentials. This rather unique feature has indeed been observed, but with copper salts only. The radical salts [2-R1,5-R3- DCNQIJ2M are therefore discussed in two separate groups.

1. DCNQI-Non-Copper Radical Anion Salts All non-cepper salts investigated 80 far [38,39,37] are “metal like semiconductors” (401 as can be judged from the temperature dependence of their electrical conductivity. This is exemplified by the behaviour of differently substituted silver salts shown in Fig. 7. The rather high conductivities a t room temperature drop slowly but steadily until they are reduced by ca. lo6 Scm-I a t ca. 50 K.

td - 10’ -

1 c I

v) 5 - - lo-‘ e -

lo-’ - lo-’

1 1 1 I

0 50 100 156 200 256 300 T [Kl

I 0 50 100 I 150 200 256 1 500 r T [Kl

Fig. 7: (2-X,5-MoDCNQI]2M Temperature depen- Fig. 8: [ ~ - X , ~ Y - D C N Q I ] ~ C U Temperature depen- dence of conductivity dence of conductivity

Aromatic/quinoid systems: principles and applications 403

2. DCNQI-Copper Radical Anion Salis All copper salts so far investigated clearly show metallic conductivity. As it is well known from other metallic organic conductors, on cooling at a certain temperature, the conductivity sharply drops, because of a phase transition (Peierls distortion) which produces a semiconductor. This behaviour is demonstrated in Fig. 8 and has also been proven by temperature dependent X-ray analysis [36]. But some of the copper salts exhibit extraordinary and unprecedented features: In salts with certain substituents metal conductivity is prevailed down to < 1 K without any phase transitions: [2,5Me,-DCNQI)&u u - 500,000 Scm-’ a t T 5 10 K (Fig. 8, [41]), [2-1,5Me-DCNQI]&u u - 2,500 Scm-’ a t T 5 1.2 K (Fig. 8, [42]), [2,5(MeO)rDCNQI]&u (u - 100,000 Scm-’ a t T < 10 K (431. With 500,000 Scm-’ the conductivity of copper is already approached if one counts the number of movable electrons per volume unit.

The metallic behaviour of DCNQI copper salts is connected to pseudo threedimensional conductivity, which has been measured to be 100 Scm-’ perpendicular to the crystal axis of [2,5Me2 DCNQI]&u [44). The special role of copper has been substantiated by other methods [45,46,47], although the interpretation of the results seems still to be controversial. Short bond distances, (Cu-N 199 pm) cannot be the only reason for the specific behaviour of the DCNQI-copper salts, because the short Ag-N distance of 230 pm also indicates strong coordination [32] similiar to [M(MeCN),]CIO,,d(M-N):Cu:195-202 pm [48], Ag: 218-233 pm [49]. Single crystals of DCNQI salts can be grown by electrocrystallisation [41], a well developed technique, which has already been used with TCNQ salts. For DCNQI copper salts, an extremely simple and useful method was found, which is also applicable to silver salts [50]. Instead of slowly reducing the DCNQI’s electrochemically, the metal itself serves as the reducing agent:

mDCNQl+ nMo + (0CNQl)mMn

Indeed, on reaction of acetonitrile solutions of substituted DCNQI’s with wires of copper or silver, the sparingly soluble anion radical salts deposit as crystals of 1-30 -mm length in many cases. These salts have been shown to be identical with thoee grown by electroreduction.

9 ALLOYS FROM DCNQI RADICAL A N I O N SALTS

Since the space group 141/a is preferred for the DCNQI-metal salts discussed so far, despite of a broad variety of 2,bsubstituents, one may anticipate that alloys of DCNQI salts should exist. Indeed, numerous alloys from two different DCNQI’s with copper as cation [51] can be synthesized either by electrocrystallisation or by the metal wire approach.

NC

X

‘C N

+

CuBrz + e or

c uo - McCN

NC

X

- m

The ratio of the two components may vary from 1:l to 13, very often being independent of the stoichiomtry of the two DCNQI’s employed. This ratio certainly does not depend directly on the difference of the reduction potentials; e. g. for 2,5-Mez-DCNQI and %Me,S-Br-DCNQI A E amounts to 0.21 V = 4.8 kcal/mol. If only thermodynamic factors govern the compoeition of the alloy, a ratio of 1:3000 would be expected and not 1:3 88

observed. It is not yet clear if the two different DCNQI’s are statistically ordered or occur in clusters within the stacks. Alloying of two (or more) DCNQI’s to one radical salt allows the manipulation of the temperature dependence of the electric conductivities together with the phase transitions M demonstrated by the two following extreme examples. As pictured in Fig. 9, the allays can behave as a mixture of the two components. The temperature dependence of conductivity is unchanged and the conductivity lies between thoee of the com- ponents [52]. By contrast, according to Fig. 10, the sharp phase transition with drastic loes of conductivity of [2,5-Br,Me-DCNQI]zCu at 150 K is completely eliminated by alloying with [2,51,Me-DCNQI]&u (1.2 : 0.8). In this allay the conductivity remains nearly constant down to 3 K (521. In addition, alloys from three and even four DCNQI’s have also b a n obtained.

404 s. HUNIG

1 1 I I I 1 I I 1 I 1 I I

b

0 ( r - D . , M O o c N Q ~ . 0 p.l.6uODcNq~c. 10-a

+ (+Dr.6WDCrrq~~%l,&U-DCNqq,&.

0 501001w200wosoo 0 50 loo 150 200 250 300 T [Kl T [Kl

Temperature dependent conductivities of alloys and their components.

Fig. 9: [~,~B~,M~-DCNQI]I[~,~CI,M~-DCNQI]ICU. Fig. 10: [~,~-B~,M~-DCNQI]I.~[~,~-I-M~-DCNQ~]O &u.

10 CONCLUSIONS

The scope of the general structural principle for two step redox systems outlined here is remarkable broad. It covers i. a. aromatic/quinoid Wurster and Weitz type systems. From the examples shown it is evident that redox potentials and molecular geometry, including substituents with ligand propertiea, can be varied within very wide limits. Within the high range of thermodynamic stability of the intermediate radical ion stage (SEM) systems can be taylored to meet specific needs. In this way, in aadition to the discussed examples, new applications may be found in the areas of chemical and biochemical electron transfer catalysis as well as in the field of organic conducting materials.

Acknowledgements I am very grateful to a great number of enthusiastic coworkers who developed this chemistry over the last 25 years, including the electrochemical inveatigations and the growing of numerous single crystals. Physical and structural data for the conducting organic materials were provided through excellent cooperation of physicists and crystallographers. All their names are given in the cited literature. Financial support is gratefully acknowledged by Deutsche Forschungsgemeinschaft, Stiftung Volkswagenwerk, Fonds der Chemischen Industrie and BASF AG, Ludwigshafen/Rhein.

REFERENCES

[l] E. Weitz, Angew. Chem., 86,658-677 (1954).

[2] S. Hiinig, Liebiga Ann. Chem., m, 32-35 (1964). [3] K. Deuchert and S. Hiinig, Angew. Chem., M, 927-938 (1978); Angew. Chem. Int. Ed. Engl., n, 875-886

(1978).

[4] S. Hiinig and H. Berneth, Top. Curr. Chem., 92 1-44 (1980). [5] L. Micheelm and E. S. Hill, w, fi, 1481-1493 (1933). [6] S. Hiinig and I. Wehner, Liebigs Ann. Chem., m, 299-301. [7] B. A. Hess Jr. and L. J. Schaad, J . Pure Appl. Chem., a, 1471-1494 (1980). [S] S. Hiinig and B. Hagenbruch, Liebiga Ann. Chem., m, 340-353. [9] C. F. Bernasconi, R. G. Bergstrom and W. J. Boyk Jr., w, 96, 4643-4650 (1974).

B. C. Bennion, J. J . Auborn and E. M. Eyrig, J. Phys. Chem., a, 701-703 (1972). C. F. Bernssconi and H. Wang, -, 29, 2214-2221 (1977).

1101 Z. Goren and I. Willner, -, ulli, 7764-7765 (1983).

Aromatic/quinoid systems: principles and applications 405

[ll] H. Simon, J . Bader, A. Gunther, S. Neumann and I. Than-, Angew. Chem., U, 541-542 (1985); Angew. Chem. Int. Ed. Engl., 29, 539-540 (1985). H. Skopan and H. Simon, b e w . Chem., a, 139-141 (1987); Angew. Chem. Int. Ed. Engl., 26, 139-141 (1987).

[12] Cf. International Conference of Synthetic Metals, Santa FC 1988, Synth. Metals, ypL2z, Bl-B556 (1988).

[13] R. C. Wheland and J. L. Gilleon, -, a, 3916-3925 (1976).

(141 S. Hunig, G. KieBlich, H. Quast and D. Scheutzow, Liebigs Ann. Chem., m, 310-323.

[15] D. 0. Cowan, A. Kini, L.-Y. Chiang, K. Lerstrup, D. R. Talham, T. 0. Poehler and A. N. Bloch,

[16] 1. B. Torrance, h S h a u & & 12, 79-86 (1979).

[17] R. C. Wheland, J A m A u d k , 28, 3926-3930 (1976).

[la] G. Saito, J . P. Ferraris, Bull. Chem. SOC. Jpn., U, 2141-2145 (1980).

[l9] J . M. Williams, H . H. Wang, T. J. Emge, U. Geieer, M. A. Beno, P. C. W. Leung, K. D. Carison, R. T.

Mol. Cryet. Liq. Cryet., 88, 1-26 (1982).

Thorn, A. J . Schultz and M.-H. Whangbo, Prog. Inorg. Chem., a, 51-218 (1987).

and 1. Tanaka, Chem., m, 55-58. [20] H. Urayama, H. Yamochi, G. Saito, K. Nozawa, T. Sugano, M. Kinoshita, S. Sato, K. Oshima, A. Kawamoto

[21] T. Mori, A. Kobayaehi, Y. Sasaki, H. Kobayashi, G. Saito and H. Inokuchi, Bull. Chem. SOC. Jpn., U, 627-63 1 ( 1984).

[22] K. D. Carleon, U. Geiser, A. M. Kini, 8. H. Wang, L. K. Montgomery, W. K. Kwok, M. A. Beno, J . M. William, C. S. Cariss, G. W. Crabtree, M.-H. Whangbo and M. Evain, Inorg. Chem., 21, 965-967 (1988).

[23] L. Birckenbach and K. Huttner, 2. Anorg. Allg. Chem., &Q, 1-26 (1930).

[24] H . Kohler, B. Eichler and R. Saleweki, Z. Anorg. Allg. Chem., m, 183-192 (1970).

[25] Our first approach used a conventional route via N,N'-dicyan-pphenylene diaminee [27] WBB independantly also tried by G. D. Andreetti, S. Bradamante, P. C. Bizzarri and G.A. Pagani, Mol. Cryat. Liq. Cryst., XU, 309-314 (1985) (private communication September 1984). Only a few species are accmeable by this route and purification is difficult.

[26] A. Aumuller and S. Hunig, Angew. Chem., M, 437-438 (1984); Angew. Chem. Int. Ed. Engl., 23, 447-448 (1984).

(271 A. Aumuller and S. Hunig, Liebigs Ann. Chem., m, 142-164.

[28] A. Aumuller, P. Erk, S. Hunig, H . Meixner, J.-U. von Schutz and H.-P. Werner, Liebigs. Ann. Chem., 1982,

[29] U . Schubert, S. Hunig and A. Aumuller, Liebigs Ann. Chem., m, 1216-1220.

[30] H. Gunther and W. Leber, unpublished results.

[31] A. Aumuller and S. Hunig, Liebigs Ann. Chem., m, 165-176.

997-1006.

[32] A. Aumuller, P. Erk, S. Hunig, J. U. v m Schutz, H.-P. Werner and H. C. Wolf, Mol. Cryet. Liq. Cryet. Inc. Nonlin. Opt., m, 215-221 (1988).

[33] A. Aumuller, E. Hiidicke, S. Hunig, A. Schatzle and J . U. von Schutr, Angew. Chem., a, 439-440 (1984);

[34] S. Hunig, A. Aumuller, P. Erk, 8. Meixner, J . U . von Schutz, H . J . GroB, U. Langohr, H.-P. Werner, H . C.

(351 A. Aumuller, * , Univ. Wiirzburg, 1985.

(361 A. Kobsysehi, T. Mori, H. Inokuchi, R. Kato and H. Kobayashi, Synth. Metals, 21, B275-B280 (1988).

[37] R. Kato, H. Kobayaehi, A. Kobayaehi, T. Mori and H. Inokuchi, Synth. Metals, 21, B263-B268 (1988).

1381 H.-P. Werner, 3. U. von Schuta, H. C. Wolf, R. K. Kremer, M. Gehrke, A. Aumuller, P. Erk and S. Hunig,

Angew. Chem. Int. Ed. Engl., 2 , 4 4 9 4 5 0 (1984).

Wolf, C. Bumchka, G. Klebe, K. Peters and H.-G. von Schnering, Synth. Metals, 22, B181-Bl88 (1988).

-, fi, 809-813 (1988).

406 s. HUNIG

[39] J . U. von Schutz, M. Bair, H.-J. Gro6, U. Langohr, H.-P. Werner, H. C. Wolf, D. Schmeieeer, K. Graf, W. Gopel, P. Erk, H. Meixner and S. Hiinig, Synth. Metals, 22, B249-B256 (1988).

[40] J , S. Miller and A. J . Epstein, Angew. Chem., a, 332-339 (1987); Angew. Chem. Int. Ed. Engl., 26,287-294 (1987).

[41] A. Aumiiller, P. Erk, G. Klebe, S. Hunig, J . U. von Schiitz and H.-P. Werner, Angew. Chem., u, 759-761

(42) P. Erk, S. Hiinig, J . U. von Schutz, H.-P. Werner and H. C. Wolf, Angew. Chern., m, 286-287 (1988);

(1986); Angew. Chem. Int. Ed. Engl., a, 740-741 (1986).

Angew. Chem. Int. Ed. Engl., 2, 267-268 (1988).

[43] H . Kobayashi, R. Kato, A. Kobayashi, T. Mori and H. Inokuchi, , 65, 1351-1354 (1988).

[44] T. Mori, K. Imaeda, R. Kato, A. Kobayashi, H. Kobayashi and H . Inokuchi, J . Phys. Soe. Jpn., a, 3429- 3431 (1987).

D. Schmeieeer, K. Graf, W. Gopel, J . U. von Schiitz, P. Erk and S. Hiinig, Chem. Phye. Lett., 148, 423-428 (1988).

D. Kongeter, F. Hentsch, €I. Seidel, M. Mehring, J . U . von Schutz, H . C. Wolf, P. Erk and S. Hunig, -, &5,453-456 (1988).

H. Kobayashi, R. Kato, A. Kobayashi, T. Mori, H . Inokuchi, Y. Nishio, K. Kajita and W. Sasaki, Synth. Metals, 22, A289-A297 (1988).

I . Csoregh, P. Kierkegaard and R. Norrestam, Acta Cryetallogr. See. B, m, 314-317 (1975).

K . Nilwon and A. O s k a m n , -, a , 7 9 - 8 5 (1984).

P. Erk, S. Hunig, H. Meixner, J. U. von Schiitz and H.-P. Werner, Liebigs Ann. Chem., m, 157-159.

P. Erk, H.-J. Groea, S. Hunig, U. Langohr, H. Meixner, H.-P. Werner, J. U. von Schutz and H. C. Wolf, Angew. Chem., U, (1989); Angew. Chem. Int. Ed. Engl., 28, (1989) in print.

H.-J. Groea, U. Langohr, J. U. von Schiitz, H.-P. Werner, H. C. Wolf, S. Tomic, D. JCrorne, P. Erk, H. Meixner and S. Hiinig, J . Physique, a, (1989) in print.