Mol. Cryst. Liq., 1980, Vol. 62, pp. 181-200 
0026-8941/80/6204-0181 $06.50/0 
© 19XO Gordon and Breach Science Publishers, Inc. 
Printed in the U.S.A. 
Electrochemically Generated 
P eryl e n i u myl- H exafl u oro p hos ph ate 
and Hexafluoroarsenate: New 
One- Dimensional Metals 
H. J. KELLER, D. NOTHE, H. PRITZKOW, D. WEHE, and M. WERNER 
Anorganisch-Chemisches Institut der Universitat, 
1m Neuenheimer Feld 270, 6900 Heidelberg, West Germany. 
and 
P. KOCH and D. SCHWEITZER 
Max-Planck-Institut filr Medizinische Forschung, 
Abteilung filr Molekulare Physik, Jahnstr. 29, 6900 Heidelberg, West Germany. 
(Received May 28, 1980) 
Compounds of stoichiometry (pe),(PF,) " , x 0.8CH 2CI 2(I)(pe = perylene),(peh(AsF6 )" x 
0.7 CH 2CI 2 (2), (peh(PF6 )1.4 x 0:6 THF (3), (peh(AsF6)1.5 x 0.5 THF (4) and (peh(SbF6h x 
0.75 CH 2C12 (5) have been obtained as crystalline samples by electrochemical deposition from 
CH 2Cl2 [(I), (2) and (5)J or from THF [(3) and (4)J solutions of peryJene, containing the appro-
priate counterion. The three compounds (1)-(3) crystallize in isomorphous orthorhombic 
lattices. (I) forms black needles: space group Pnmn with a = 4.285 A, b = 12.915 A and c = 
14.033 A, z = I. (2) gives black needles, orthorhombic space group Pnmn with a = 4.294 A, 
b = 13.077 A, and c = 14.132 A, z = L The structures of (I) and (2) were solved by direct 
methods and refined by least squares to final R = 0.148 and R = 0.088 based on 476 and 322 
observed reflections. The perylene forms segregated stacks in direction of the a-axis with 
interplanar distances of 3.40 A and an angle of 37.7° between the he-plane and the perylene. The 
channels between the segregated stacks are filled by anions and solvent molecules. The d.c. 
conductivities (four probe measurements) of (1)-(4) fall in the range of 70-12000- 1 ·em -1 at 
room temperature. The conductivities show a metallic regime down to about 2000K and drop 
off below that temperature. 
181 
182 H. J. KELLER et al. 
1 INTRODUCTION 
Electrochemically induced redox reactions in organic solvents are a common 
tool to generate and investigate open shell organic ions. I - 3 In several cases 
this procedure led to the isolation of solid products containing radical ions 
with remarkable solid state properties.4- 8 The recent interest in crystallized 
organic radical ions is connected mainly with their high electrical conduc-
tivities.9- 11 Some of the very promising materials have recently been ob-
tained electrochemically.8 The electrochemical method is especially well 
suited in the search for highly conducting (metallic) radical salts. This is 
because electrochemical deposition of solid organic radical salts from solu-
tions "inevitably" selects in favor of the highly conducting crystals. From all 
thermodynamically favorable compounds which could crystallize from 
solution the metallic ones grow fastest. After nucleation at the electrode 
surface those crystals with the lowest resistivity will enable the most intense 
electron flow. Only those compounds which allow electron transport for 
the redox reaction over a distance of several millimeters and in a high density 
flux can be deposited as reasonable crystals. The advantages of electro-
crystallization have been demonstrated earlier by different groups growing 
single crystals of one-dimensional metals of the mixed valence tetracyano-
platinate (II, IV) type electrochemically.12.13 
Electrochemical oxidation reactions on several aromatic hydrocarbons 
have been carried out in this laboratory, following the above mentioned 
lines. By varying experimental conditions, using different hydrocarbons 
(like coronene, anthracene, pyrene, triphenylene, etc.) and supporting 
electrolytes like (C4H9)4N+PF6", (C4H9)4N+ AsF6, etc., it was found as a 
preliminary result that perylene is especially well suited to form highly con-
ducting solid radical species. Structural and physical investigations on some 
of these solids obtained by electrocrystallization procedures are presented 
here. 
2 EXPERIMENTAL 
2.1 Preparation of compounds 
2.1.1 Apparatus for electrochemical syntheses 
The electrolyses were carried out in a 100 ml V-tube with a glass filter or in a 
75 ml electrochemical cell of 3 compartments separated by 2 glass filters from 
each other. In some cases the crystal size of the reaction products depend 
on the used electrochemical cell. Platinum wire was used as electrode 
material. Ni, Ag and Cu can be used as cathode material, too. The potential 
was selected to give a current of 20-50 /-lA. Depending on the cell, solvent 
PERYLENE WITH AsFi AND PF6 183 
and concentration of the supporting electrolyte, potentials between 2.5 and 
6 V have been used. This results in a current density of 2-5 J.lA/cm 2• This 
current density is conclusive for the velocity of the crystal growth. 
2.1.2 Chemicals 
Methylenechloride (CH 2CI2) was dried over CaClz and a molecular sieve 4 A 
and saturated with nitrogen. Tetrahydrofurane (THF) 99.9 % (Gold-
Marke) saturated with nitrogen was purchased from EGA Chemie. Tetra-
butylammoniumhexafluorophosphate (BU4PF 6) was obtained by mixing 
equimolar amounts of BU4N J and KPF 6 in aqueous solution. 14 Bu4NAsF 6 
was obtained in a similar fashion by using potassium hexafluoroarsenate 
instead of K PF 6' Bu4NSbF 6 was prepared in a similar manner as 
(CH3)4NSbF 6. 15 To a solution of 13.15 g Potassiumhexahydroxoantimonate 
(V), K[Sb(OH)6J in 100 m140% HF a 40% aqueous solution of Tetrabutyl-
ammoniumhydroxide was added. The mixture was evaporated to half of its 
original volume, the precipitate was filtered and washed with a few milliliters 
of water. Recrystallization from isopropanol gave white needles, Fp.: 220°C, 
Yield: 11 g ,;; 46 % of the theoretical yield. 
Analyses: (C4H9)4NSbF6 Molemass: 478.22 
Calc.: C 40.19% H 7.59% N 2.93 % 
Found: C40.19% H7.79% N2.80% 
All these supporting electrolytes were dried under high vacuum. Perylene 
99 + % Gold label was purchased from Aldrich Chemical Company and 
used without further purification. 
2.1.3 Preparation of pe-P F 6 compounds 
2.1.3.1 (peMPF6 )1.1 x 0.8 (CHzCiz). 0.1 g perylene and 0.5 g Bu4NPF6 
were dissolved in 100 ml absolute CHzClz, filtrated into the electrochemical 
cell and electrolized by applying about 3 Vand 30 J.lA for about 2 days under 
exclusion of light. Crystal growth starts after a few minutes and after two days 
black needles (which appear green by looking through the crystals), of 
several cm in length, were obtained. A typical experiment is shown in 
Figure 1. The crystals decompose above 180°C, giving up neutral perylene. 
Analyses: (peh(PF6)1.1 x 0.8 CH2CI3 : 
C4o.sH25.6CI1.6F 6.6P 1.1 
Molemass: 732.06 
Calc.: C 66.94 H 3.52 C17.75 F 17.13 P 4.65 
Found: C 66.71 H 3.66 C18.05 F 17.00 P 4.50 
C 66.57 H 3.72 P 4.53 
184 H. J. KELLER el al. 
Closer inspection of the prepared samples showed, that principally two dif-
ferent types of crystals grow simultaneously at the surface of the anode: 
thinner ones with a rectangular or trapezoid cross-section (" type one") 
while the thicker ones have rectangular, quadratic or preferably hexagonal 
cross-sections ("type two "), (for typical dimensions see Section 2.3. "Con-
ductivity measurements"). The latter crystals are the main product during 
electrocrystallization. 
Analytical data suggest that type one crystals contain much less CH2Cl2 
but more PF6" ions than the main product. This results in a higher formal 
oxidation number for the peryleniumyl stacks in "type one" crystals com-
pared to the "type two" modification. The analytical data of the mixture 
which are given above are very close to the values of the main product 
(type two). 
2.1.3.2 (pe)z(PF6 )1.4 x 0.6 (THF). By using THF instead of CH 2Cl2 as 
solvent the same condition with a current of 20 JlA leads after two days to 
long but very thin black needles. No "type two" crystals have been observed. 
The crystals decompose above 170°C. Above this temperature perylene 
crystals grow at the surface of the black needles. 
Analyses: (pe)z(PF6)1.4 x 0.6 THF 
C42.4H2B.BOO.6P 1.4F 8.2 
Molemass: 751.46 
Calc.: C 67.77 H 3.86 F 21.24 ° 1.28 P 5.77 
Found: C 67.48 H 3.96 F 20.62 0- P 5.77 
2.1.4 Preparation of pe-AsF 6 compounds 
2.1.4.1 (peh(AsF6kl x 0.7 (CH 2Cl2 ). The electrocrystallization was 
carried out with the same apparatus and under the same conditions just 
using BU4NAsF 6 instead of Bu4NPF 6. Long black needles are obtained 
(green in absorption) which decompose above 260°C giving up neutral 
perylene. 
Analyses: (pe)z(AsF6) 1. I x 0.7 CH2Cl 2 
(C40. 7H 25.4As 1.lCI 1.4F 6.6) 
Molemass: 771.90 
Calc. C 63.34 H 3.32 As 10.62 Cl 6.42 F 16.25 
Found: C 63.30 H 3.44 As 10.8 Cl 6.20 F 17.0 
Again two different types of crystals (thinner and thicker ones) have been 
obtained simultaneously at the anode. The results are very similar to those 
described for the PF6" -derivatives (see Sections 2.1.3.1. and 2.3). 
PERYLENE WITH AsF6 AND PF6 185 
II 
(a) (b) 
(c) (d) 
2cm 
FIGURE I Eleclrocrystallization 01"(1) at a platinum anode: (a) after 30 min. (b) after 2 h, 
(c) after 12 h, (d) after 20 h. (Only the anode half cell is shown). 
186 H. 1. KELLER el al. 
2.1.4.2 (pe}z(AsF6)1.5 x 0.5 (THF). Similar procedures as described 
above, using THF as a solvent, give black long but thin needles which de-
compose above 260°C. Again no second phase has been found. 
Analyses: (peh(AsF6)1.5 x 0.5 THF 
C42H2sAS1.5F 9.200.5 
Molemass: 824.08 
Calc.: C 61.22 H 3.42 As 13.67 F 21.21 00.97 
Found: C 60.92 H 3.32 As 14.1 F 20.38 0-
2.1.5 Preparation of (peh(SbFh compounds 
Electrocrystallization from CH2Cl2 and from THF gives small black 
platelets and prisms which cover the electrode but which do not grow into 
the solution. Only C and H analyses have been carried out because the 
crystals are not usable for physical investigations. 
Analyses: (peh(SbF6)2 x 0.75CH2C12 
C60.7sH37.sCI1.5F 12Sb2 
Mo1emass: 1292.16 
Calc.: C 56.67 H 2.93 
Found: C 56.74 H 3.14 
from 
THF: (peh(SbF6h x THF 
C64G44F 12Sb2 
Molemass: 1300.57 
Calc.: C 59.10 H 3.41 
Found: C 58.93 H 3.62 
2.2 X-Ray investigations 
For the collection of the X-ray intensity data only the type two crystals 
of (1) and (2) could be used. Nevertheless we investigated thin crystals of 
(3) and showed them to be isomorphous with the type tw.o phases of (1) 
and (2) (see Table I). There seems to be no doubt that all phases of (1)-(4) 
are isostructural. 
Lattice parameters were determined from Weissenberg films and refined 
with 22 for (1) respectively 19 for (2) reflections measured on a Syntex R3 
diffractometer. Intensity data were collected on the diffractometer (MoKa 
radiation, A. = 0.7107 A, graphite monochromator). Reflections were mea-
sured by fJ/2fJ scan with a scan range of 1.9°. Crystal data and details of 
PERYLENE WITH AsF;; AND PF;; 
TABLE J 
Compound 
Crystal System 
Space group 
Cell parameters a (A) 
b (A) 
c (A) 
Cell volume (A) 
z 
Density calc. (g/cm 3 ) 
Linear absorption factor (cm ~ 1) 
Size of crystal (mm) 
28 max. 
Number of measured refl. 
Unobserved refl. 
(I) 
orthorhombic 
Pnmn 
4.285 
12.915 
14.033 
776.56 
I 
1.57 
13.0 
0.5 x 0.2 x 0.1 
67.6 
781 
305 
(2) 
orthorhombic 
Pnmn 
4.294 
13.077 
14.132 
793.55 
I 
1.65 
12.1 
0.3 x 0.2 x 0.1 
69.6 
1226 
904 
187 
(3) 
orthorhombic 
Pnmn 
4.26" 
12.79" 
13.89' 
756.80' 
I 
1.61 
0.3 x 0.2 x 0.05 
, Cell parameters from Weissenberg films were not refined by diffractometer data. 
measurements are given in Table I). LP and absorption corrections were 
applied to the data. Reflections with 1 < 3 a{l) were considered unobserved 
and not included in the refinement. The calculations were carried out on an 
IBM 370/168-computer (Universitiits-Rechenzentrum, Heidelberg). 
2.3 Conductivity measurements 
The dc-conductivity was measured by the standard four-probe method l6 
along the needle axis (length 2-4 mm) parallel to the perylene stacks. The 
samples were mounted on gold wires in a Teflon block as described by 
Coleman. 1 7 The current and voltage contacts were made with silver-
(Dupont 4929) or gold-paint (Doduco Auromal GM7). The current contacts 
covered the ends of the crystal to assure uniform current distribution. 
Currents from 0.5-10 {lA (HP 4329A) were used for the measurements. 
The voltage contacts were carefully painted completely around the crystal. 
The voltage was measured with a Digital Nanovoltmeter (Keithley Model 
180). The samples were placed in a continuous flow cryostat (Oxford In-
struments CF 204 with Digital Temperature Controller DTC-2) for lowering 
the temperature. 
The temperature dependence of the conductivity was measured on about 
twenty (pe)iPF 6)1.1 x 0.8 CH2Cl2 crystals. The room temperature con-
ductivity (1"(300) for these crystals varied between 160 (n cm)-l and 1.100 
(n cm) - 1. A typical value for the majority of the samples was about 600 
(n cm)- I. Very similar results were obtained for the (pe)2(AsF 6)1.1 x 0.7 
CH2Cl 2 crystals. The room temperature conductivity (1"(300) varied between 
210 (n cm)-l and 1200 (n cm)-I, a typical value was about 700 (n cm)-I. 
For (1) as well as for (2) two types of crystals were measured. The first 
type of crystals was very thin C-O.OOS-O.01 mm) and at least 10 to 50 
188 H. 1. KELLER et al. 
times as broad, while the second type of crystals was thicker (~0.05 mm) 
and only 1 to 3 times as broad. The room temperature conductivities of 
both types of crystals are comparable but the temperature dependence is 
somewhat different. Figure 2 shows the typical temperature dependence of 
the conductivity normalized to the room temperature value 0"(300) for the 
type one crystals of (1) and (2). Figure 3 shows the same temperature de-
pendence for the second type of crystals. The main difference of both con-
ductivity characteristics is that in the second case the conductivity makes a 
sudden jump at a temperature of about 110-120 K. In the moment it is not 
clear if this change in conductivity is due to a phase transition. 
In the case of(3) and (4) only one type of crystal is found [similar to the type 
one crystals of ( 1) and (2)]. The room temperature conductivity 0"(300) of these 
crystals of (3) and (4) varied between 70 and 700 (n cm)-I and is therefore 
somewhat lower than the r.t. conductivity of both forms of (1) and (2). 
The temperature dependence of the normalized conductivity of (3) and 
(4) is shown in Figure 4 and is similar to the one of type one crystals of 
(1) and (2) in Figure 2. Nevertheless, in the semiconducting region below 
180 K the observed activation energy for (3) and (4) is only about half as 
large as for type one of (1) and (2). 
2.4 EPR measurements 
X-Band EPR spectra were taken on a Bruker B-ER 420 spectrometer 
equipped with a Bruker B-NM 12 NMR Gaussmeter for field calibration. 
Absorption intensities were obtained by double integration of the usual 
first derivative lines by means of a Bruker ASPECT 2000 computer. 
EPR spectra of three thick (type two) single crystals of (1) show one 
strong, exchange narrowed line of Lorentzian shape. The peak to peak line-
width of only 0.6 Gauss remains constant between 300 K and 190 K. Below 
that temperature the linewidth slightly increases with decreasing temperature, 
being 0.8 Gauss at 100 K. The EPR absorption intensity is constant within 
± 2 % from 300 K down to 190 K. At lower temperatures one observes an 
activated paramagnetism. Because of the small temperature interval in 
which that paramagnetism could be observed in our experiment the activa-
tion energy can be given only very roughly. We estimate the value as 0.025 eV. 
3 STRUCTURE DETERMINATION 
The structure determination of (2) is described in detail. The structure was 
solved in the space group Pnmn by direct methods using the SHELX 76 
program. The highest peaks in the E-map are assigned to the carbon atoms of 
a(T) 1 
U(300K) 
10-1 
x (Perylene~PF6 U<300K):!900Q-1cm-1 
o (Perylene~Asfs a<300K)=12ooQ-1cm-1 
~"Oo \~ 
x 0, 
X 0 
\x '\0 
\x 
\\ 
X 
\ 
X 
\ 
x 
O.045eV \ o O.035eV 
o , 
0.2 0.4 os 0.8 t2 1.8 22 2A 
2 
FIGURE 2 Temperature dependence of the normalized d.c. conductivity of H thin" crystals 
(first type: see text) of (1) [H (Perylene)2PF 6 "] and (2) [H(PerylenehAsF 6 "]. 
189 
an) t 
a<300K) 
2 
10-2 
10 
10-4 
o 0.2 Q.4 0.6 OB 
x 
x (Perylene)2 PF6 
n-1 -1 
a<300K)=880.l." em 
o (Perylene)2 As F6 
n-1 -1 
a<300K)=760.l." em 
0' \ x 
'''\ ~o
x \. 
x 
0.038eV\ 
1.2 1.4 1.6 1.8 
0.027eV 
o 
2.2 2.4 2.6 2.8-
100fT IK-'I 
FIGU RE 3 Temperature dependence of the normalized d.c. conductivity of" thick" (second 
type: see text) of (1) and (2). 
190 
--
10-1 -
PERYLENE WITH AsF6" AND PF6" 
.... nK) 
3002?O 100 I 
70 
I 
x (Perylene)2 PF6 U<300Kl=700-'cm-
1 
o (Perylene)2AsF6 U<30OKl=580 O-'cm-' 
(THF) 
191 
0.2 0." 0.6 0.8 I 1.2 1.4 1.6 1.8 I 2.2 2." 2.6 2.8'-" 
1 2 100/T (K-') 
FIGURE 4 Temperature dependence of the normalized d.c. conductivity of (3) 
["(Perylene)2PF6(THF)"] and (4) ["(PerylenehAsF6(THF)"]. 
the perylene, which lies in the special position 2b (0, t, 0). The peryJene mole-
cules form isolated stacks in the direction of the a-axis with a translational 
period of 4.29 A. The channels between these stacks are filled by anions 
and solvent molecules. The localization of these molecules is more difficult. 
Another strong peak in the special position 2a (0, 0, 0) gives the position of 
the As atom, but the peak height is too small and no F atoms can be found. 
As the chemical analysis gives a perylene: AsF 6 ratio of 2: 1.1, the special 
position 2a can only partially be occupied by anions. Considering the cell 
and the form of the anion this is sensible. In the direction of the a-axis the 
distance between anions in the same channel would come to 4.29 A. less 
than the shortest extension of the anion. Therefore, only every second cell 
in a channel is occupied by an anion, the holes between are filled by solvent 
molecules. This distribution gives a perylene: AsF 6 ratio of 2: 1. With an 
192 H. J. KELLER e/ al. 
analysis ratio of 2: 1.1, some anions have to occupy adjacent cells in the same 
channel. This is only possible, if they are shifted from one another in the 
direction of the a-axis. The very big anisotropic temperature factor in this 
direction for As (see Table III) may be produced by this displacement. 
In the case of an ordered arrangement the cell parameter has to be doubled 
and in the reciprocal space intermediate layers must appear, to which only 
the anions and solvent molecules would contribute. On rotation photo-
graphs of (2) diffuse streaks appear in the middle between the layers per-
pendicular to a. We conclude from this that the distribution in the channels 
has a translational period of 2a, but without correlation between different 
channels in the bc plane. A refinement with all carbon atoms and the As 
atom gives R = 0.27. The difference Fourier map shows several peaks 
around the As atom. Three of these form with the As atom an octrahedron 
with distances of about 1.7 A and angles of 90°. The mirror plane of the space 
group goes through the AsF 6-octahedron, but doesn't coincide with the 
TABLE II 
Coordinates and temperature factors for (pe)zPF6 (CH 2CI 2) 
x y Z U1l 
C(I) 0.0936 0.5000 0.8314 0.0463 
(0.0050) (0.0014) (0.0052) 
C(2) 0.2999 0.5000 0.9189 0.0387 
(0.0047) (0.0013) (0.0046) 
C(3) 0.3985 0.5955 0.9592 0.0392 
(0.0029) (0.0010) (0.0008) (0.0028) 
C(4) 0.2901 0.6871 0.9204 0.0507 
(0.0036) (0.0011) (0.0010) (0.0039) 
C(5) 0.0800 0.6853 0.8394 0.0577 
(0.0040) (0.0012) (0.0011) (0.0043) 
C(6) -0.0094 0.5938 0.7999 0.0559 
(0.0034) (0.0012) (0.0009) (0.0037) 
H(I) 0.2450 0.7685 0.9025 0.05 
H(2) 0.3143 0.7009 0.8086 0.05 
H(3) -0.1626 0.6169 0.7403 0.05 
P 0.0000 0.0000 1.0000 (1) 
F(I) 0.2577 0.0816 0.9676 0.0438 
(0.0145) (0.0045) (0.0040) (0.0133) 
F(2) 0.1330 -0.0776 0.4263 0.0238 
(0.0143) (0.0053) (0.0046) (0.0146) 
F(3) 0.1858 -0.0412 0.5623 0.1675 
(0.0155) (0.0047) (0.0047) (0.0264) 
Cl(I) -0.0361 0.0397 0.5979 0.0724 
(0.0099) (0.0026) (0.0026) (0.0097) 
(1) Temperature factors for P: 
UII Un U33 U23 U13 U12 
0.1375 0.0419 0.0637 0.0 -0.0488 0.0 
(0.0209) (0.0084) (0.0103) (0.0130) 
PERYLENE WITH AsFi AND PFi 193 
mirror plane of the octahedron, thus producing two orientations for AsF 6' 
Two other peaks are assigned to the chlorine atoms of CH2Cl2 . A refinement 
with all non-hydrogen atoms (As anisotropic) lowers the R-values to R = 
0.102 and Rw = 0.077, adding the H-atoms in calculated positions gives the 
final values of R = 0.088 and Rw = 0.061. 
For (1) similar results were obtained. But the assignment of the peaks 
around the P gave no reasonable results. Thus we assume that the smaller 
PF 6 octahedron is not as well fixed in the available space between the cation 
stacks and thus no special orientation for PF 6 could be found. 
Rotation- and Weissenberg-photographs of (3) agreed exactly with that 
of (1). 
4 DESCRIPTION OF THE STRUCTURE 
Tables II and III show the atom coordinates and temperature factors for 
(1) and (2) respectively. The coordinates for the carbon atoms show the 
similarity of the structures of both compounds (1) and (2). The perylene 
molecule occupies the special position 2b (0, t. 0). The angle between the 
perylene and the be-plane is 37.7° and the intermolecular distance is 3.40 A. 
Atomic distances and angles of the perylene ion in (2) with the standard 
deviations are shown in Figure 5. Figure 6 shows a projection oftwo adjacent 
perylenium cations perpendicular to the perylene plane. A projection of the 
H'(3) 
FIGURE 5 Atom distances and angles of the peryleniumyl ion in (2). 
194 H. 1. KELLER et af. 
TABLE III 
Coordinates and temperature factors for (pehAsF 6(CH 2CJ 2) 
x y Z VI! 
C(I) 0.0960 0.5000 0.8386 0.0427 
(0.0025) (0.0) (0.0008) (0.0028) 
C(2) 0.2999 0.5000 0.9194 0.0367 
(0.0025) (0.0) (0.0008) (0.0028) 
C(3) 0.3967 0.5946 0.9592 0.0407 
(0.0016) (0.0006) (0.0005) (0.0020) 
C(4) 0.2863 0.6865 0.9177 0.0462 
(0.0019) (0.0006) (0.0006) (0.0023) 
C(5) 0.0895 0.6838 0.8392 0.0563 
(0.0020) (0.0006) (0.0006) (0.0024) 
C(6) -0.0074 0.5937 0.8002 0.0520 
(0.0021) (0.0006) (0.0005) (0.0021) 
H(l) 0.3639 0.7617 0.9468 0.05 
H(2) 0.0053 0.7575 0.8082 0.05 
H(3) -0.1686 0.5942 0.7397 0.05 
As 0.0000 0.0000 1.0000 (1) 
F(I) 0.2641 0.0549 0.9259 0.1668 
(0.0037) (0.0019) (0.0041) (0.0042) 
F(2) 0.6913 0.0837 0.9601 0.1207 
(0.0032) (0.0016) (0.0013) (0.0041) 
F(3) 0.0906 0.0723 1.0902 0.2853 
(0.0032) (0.0014) (0.0012) (0.0037) 
CI(l) -0.0553 -0.0958 0.9478 0.1386 
(0.0042) (0.0031) (0.0028) (0.0043) 
Cl(2) -0.0248 -0.0217 1.1071 0.0735 
(0.0040) (0.0023) (0.0017) (0.0041) 
H(4)a 
-0.2311 0.0280 0.9846 0.05 
H(5)" -0.1699 0.0565 0.9791 0.05 
(I> Temperature factors for As: 
VII Vn V33 V 23 V,3 V,2 
0.1668 0.0341 0.0507 0.0 0.0249 0.0 
(0.0037) (0.0023) (0.0026) (0.0033) 
a H-atoms for CH2CJ 2 in the AsF6 compound. 
unit cell on the be-plane shows Figure 7, the anions, which lie between the 
perylene cations are omitted. A cation stack shows Figure 8. This corres-
ponds to a projection onto the ac-plane. The distances and angles of the 
perylene ion are in agreement with the values found for perylene. 1B Because 
of the agreement of the structures of perylene [Pt(mnthJ 19 and the present 
compound it must be assumed that the perylene determines the structure 
by making segregated stacks, whereas the remaining place is filled by anions. 
In this case the place is filled by AsF 6 and CHzCl z , where the AsF 6 octa-
hedron has an orientation between the cation stacks, which gives a suitable 
FIGURE 6 Projection of two adjacent perylenium ions perpendicular to the perylene plane. 
FIGURE 7 Projection of the unit cell on the be-plane (anions are omitted). 
195 
196 H. 1. KELLER et al. 
FIGURE 8 Projection of a cationic stack onto the ac plane. 
filling of the space in the channels. As the anions are disordered, the found 
distances and angles (AsF: 1.63, 1.70, 1.81 and F-As-F: 92.3, 94.4,93.9) are 
not very reliable. 
5 DISCUSSION 
It is well known that perylene can easily be oxidized by halogens-especially 
bromine20- 24 and iodine2o,23,25-31_by different metal salts27 ,32-34 like 
antimonyhexachloride,32-35 silverperchlorate34 and electrochemically.4,6 
Though it has been pointed out very early that some of these oxidized solids 
are quite good electrical conductors (at least compared to other typical 
organic compounds, which are insulators),22 detailed physical and structural 
data on single crystals, so far have only been obtained for (pe}z Pt(mnt)2 
(mnt = anion of dithiomaleodinitrile}.19 The latter solid was obtained by 
oxidizing perylene with iodine in the presence of tetra-n-butylammonium-
Pt(mnth,36 as well as electrochemically.6 The principle arrangement of 
peryleniumyl cations in the hexafluoroarsenate (V) and hexafluorophosphate 
(V) is very similar to the columns in (pe)2 [Pt(mnthJ.1 9 
The d.c.-conductivities parallel to the a-axes of all crystals of (1), (2), (3) 
and (4) in the temperature range between 300 and 190 K show a metallic-
like behaviour. Below 190 K the conductivities of the crystals decrease with 
decreasing temperatures. In this semiconducting region the conductivity 
of the crystals of (1) and (2) depend on the "shape" of the crystals. In the 
case of the thicker crystals (type two) of (1) and (2) a sharp drop of the con-
ductivity at a temperature of about 110-120 K is observed. The thinner 
crystals (type one) of (1), and (2) show a smooth decrease in the conductivity 
with decreasing temperature below 190 K. The observed activation energies 
in the semiconducting region are smaller for (3) and (4) than for (1) and (2) 
(see Figure 2-4). Nevertheless, this activation energies are very similar to 
PERYLENE WITH AsF'; AND PF,; 197 
the one estimated from the ESR-measurements. Temperature cycling ex-
periments between 300 and 50 K have shown a degradation of the room 
temperature conductivity which was particularly sensitive for those crystals 
of (1), (2), (3), and (4) which primarily showed the highest conductivities. 
Similar observations were made earlier for other one-dimensional organic 
metals. 37 
The relatively high conductivities of these compounds are possibly con-
nected with the nonintegral formal oxidation numbers on the organic part 
of the structure. The .. substitution" of some of the counter anions (like 
PF6 or AsF6) in the channels by neutral solvent molecules (CH2 C12 or THF) 
leads to a mixed valence state for the cationic stacks. This mixed valence 
state which is recognized as a prerequisite for high conductivities,9-11 is 
normally brought about by partial charge transfer but in the cases discussed 
here determined by stoichiometry. In this respect it is not surprising that 
crystals with different CH2CI2/PF6 or different CH2CI2/AsF6 ratios show 
different d.c. conductivity behaviour (type one and type two crystals of (1) 
and (2», (see Sections 2.1.3.1; 2.1.4.1 and 2.3). 
In by far the most highly conducting organic charge transfer complexes 
one observes very sharp epr lines, only very few exceptions are known. 38 •39 
A linewidth of 0.6 Gauss as in our case thus clearly hints to a rather strong 
exchange interaction between neighbouring molecules. This interaction 
apparently leads to a metallic-like behaviour between 300 K and 190 K 
as can be seen from temperature independent paramagnetism and con-
ductivity data. Such a temperature independent paramagnetism has been 
likewise observed in NMP-tCNQ,40 a well-known organic metal. The 
similarity of the magnetic properties of NMP-TCNQ and (1) is also ap-
parent from the temperature dependence of the linewidth. We are thus quite 
sure that (1) can be considered as an organic metal at higher temperatures 
whereas it behaves as a paramagnetic semiconductor at temperatures below 
190 K. 
The hexafluoroantimonate (V) salt (5) which differs in composition from 
(1)-(4) shows a d.c. conductivity which is about two orders of magnitude 
lower than those of (1)-(4). Consistent with this observation is the result of 
electrocrystalJization experiments using SbF 6" as counterion: only small 
black prisms covering the electrode have been obtained (see Section 2.1.5). 
6 CONCLUSION 
Electrocrystallization is a valuable tool in the search for new highly con-
ducting solids. This method selects crystals with the highest electrical con-
ductivities and not those with the lowest free energies under the experimental 
conditions as found by using chemicals as oxidation agents. 
198 H. J. KELLER el al. 
Acknowledgements 
We would like to thank Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, and 
Stiftung Volkswagenwerk for financial support. The help of P. Hagg and R. H. Harms in 
preparing Figure I is gratefully acknowledged. 
References 
I. F. Beck, Elektroorganische Chemie, Verlag Chemie, 1974. 
2. Free Radicals, ed.l. K. Kochi, lohn Wiley, N.Y. 1973. 
3. Methods in Free Radical Chemistry, Vol. 3, ed. E. S. Huyser, Marcel Dekker, N.Y. 1972. 
4. T. C. Chiang, A. H. Reddoch, and D. F. Williams, J. Chem. Phys., 54, 2051 (1971). 
5. C. V. Ristagno and H.l. Shine, J. Org. Chem., 36, 4050 (1971). 
6. L. Alcacer and A. H. Maki, J. Phys. Chem., 78, 215 (1974). 
7. H. P. Fritz, H. Gebauer, P. Friedrich, P. Ecker, R. Artes, and U. Schubert, Z. Naturforsch., 
33b, 498 (\ 978). 
8. a) D. Jerome, A. Mazaud, M. Ribault, and K. Bechgaard, J. Physique Lel/re.l. 41, L-95 
(1980). b) K. Bechgaard, unpublished results. 
9. Chemistry and Physics of One-dimensiunal Metals, ed. H. J. Keller, NATO-ASl Series, 
Vol. 258, Plenum Press, N.Y. 1977. 
10. Synthesis and Properties of Low-dimensional Materials, ed. J. S. Miller and A. 1. Epstein, 
Ann. N. Y. A cad. Sci., 313 (1978). 
11. Molecular Metals, ed. W. A. Hatfield, NATO-Conference Series VI: Malerials Science, 
Vol. I, Plenum Press, N.Y. 1979. 
12. J. S. Miller, Science, 194, 189 (1976). 
13. J. M. Williams, D. P. Gerrity, and A. J. Schultz, J. Am. Chern. Soc., 99,1668 (1977). 
14. J. E. Lind,jr., H. A. A. Abdel-Rehim, J. Phys. Chern., 70,3610 (1966). 
15. W. Lange and K. Askitopoulos, Z. Anorg. Allg. Chern., 223,369 (1935). 
16. L. B. Coleman, J. A. Cohen, A. F. Garito, and A.l. Heeger, Phys. Rev., 87, 2122 (1973). 
17. L. B. Coleman, Rev. Sci. Instrument, 46,1125 (1975). 
18. A. Camermann and J. Trotter, Proc. Roy. Soc., A279, 129 (1964). 
19. L. Alcacer, D. Chasseau, and J. Gaultier, in preparation. 
20. K. Brass and E. Clar, Ber., 65, 1660 (1932), Ber., 69,1977 (1936). 
21. A. Zinke and A. Pongratz, Ber., 69,1591 (1936) and Ber., 70, 214 (1937). 
22. H. Akamatu, H. Inokuchi, and Y. Matsunaga, Nature, 173, 168 (1954). 
23. H. Akamatu, H. Inokuchi, and Y. Matsunaga, Bull. Chem. Soc. Japan, 29, 213 (1956). 
24. T. N. Anderson, D. W. Wood, R. C. Livingston, and H. Eyring, J. Chem. Phys., 44, 1259 
(1966). 
25. T. Uchida and H. Akamatu, Bull. Chem. Soc. Japan, 34, 1015 (1961) and earlier references 
cited therein. 
26. J. Kommandeur and F. R. Hall, Bull. Am. Phys. Soc., 4, 421 (1959) and J. Chem. Phys., 34, 
129 (1961). 
27. M. S. Frant and R. Eiss, J. Eleclrochrn. Soc., BO, 769 (1963). 
28. S. Aronson, 1. S. Mittelman, and f. B. Bramwell, J. lnorg. Nucl. Chern., 39,1095 (1977). 
29. H.-C.l. Kao, M. Jones, and M. M. Labes, J.C.S. Chem. Comm., 1979, 329 (1979). 
30. R. C. Teitelbaum, S. L. Ruby, and T. 1. Marks, J. Arner. Chem. Soc., 101, 7568 (\979) and 
references cited therein. 
31. P. Coppens in Extended Linear Chain Compounds, ed. J. S. Miller, Plenum Press, N.Y., in 
prepaRItion. 
32. K. Brass and E. Tengler, Chern. Ber., 64,1650 (\931). 
33. G. E. Blomgren and J. Kommandeur, J. Chem. Phys., 35,1636 (1961). 
34. Y. Sato, M. Kinoshita, M. Sano, and H. Akamatu, Bull. Chem. Soc. Japan, 42,3051 (1969). 
35. S. Hilpert and L. Wolf, Ber. deut. chern. Ges., 46,2215 (1913). 
36. L. Alcacer, H. Novais, and F. Pedroso, in Molecular Metals, ed. W. A. Hatfield, pp. 415. 
PERYLENE WITH AsF6 AND PF6 199 
37. M. J. Cohen, L. B. Coleman, A. F. Gar-ito, and A. J. Heeger, Phys. Rev., lOB, 1298 (1974). 
38. Y. Tomkiewicz and E. Engler, Bull. Am. Phys. Soc., 20, 497 (1975). 
39. R. B. Somoano, A. Gupta, and V. Hadek, J. Chern. Phys., 63, 4970 (1975). 
40. A. J. Epstein, S. Etemad, A. F. Garito, and A. J. Heeger, Phys. Rev., B5, 952 (1972).