close

Вход

Забыли?

вход по аккаунту

?

Interfaces between Molecular and Polymeric УMetalsФ Electrically Conductive Structure-Enforced Assemblies of Metallomacrocycles.

код для вставкиСкачать
Interfaces between Molecular and Polymeric “Metals”: Electrically
Conductive, Structure-Enforced Assemblies of Metallomacrocycles
By Tobin J. Marks”
The design, synthesis, characterization, and understanding of new molecular and macromolecular substances with “metal-like” electrical properties represents an active research area
at the interface of chemistry, physics, and materials science. An important, long-range goal in
this field of “materials by design” is to construct supermolecular assemblies which exhibit
preordained collective phenomena by virtue of “engineered” interactions between molecular
building blocks. In this review, such a class of designed materials is discussed which, in
addition, bridges the gap between molecular and polymeric conductors: assemblies of electrically conductive metallomacrocycles. It is seen that efforts to rationally construct stacked
metal-like molecular arrays lead logically to structure-enforced macromolecular assemblies of
covalently linked molecular subunits. Typical building blocks are robust, chemically versatile
metallophthalocyanines. The electrical, optical, and magnetic properties of these metallomacrocyclic assemblies and the fragments thereof, provide fundamental information on the
connections between local atomic-scale architecture, electronic structure, and the macroscopic
collective properties of the bulk solid.
1. Introduction
Until only a few years ago, the idea that an organic or
metal-organic solid could exhibit the electrical, magnetic,
and optical characteristics normally associated with bulk
metals would have seemed far-fetched indeed. Among other
factors, simple band-theoretical concepts tell us that such
substances inherently lack the partially occupied, spatially
delocalized manifolds of electronic energy levels (bands)
characteristic of metals.f’] In molecular and polymeric
solids, such bands are built up of contributing, molecularlevel atomic and molecular orbitals, and, as in metals, only
the energetically close proximity of filled and empty levels
can give rise to electronic (charge) mobility and metallic
character. However, typical electron-occupancy patterns
found in organic and metal-organic solids lead to wide gaps
between filled and unoccupied bands. Such traditional materials, as exemplified by solid benzene, polyethylene, and
poly(tetrafluoroethylene), are electrical insulators. These
concepts are illustrated schematically in Figure 1.
The past several years have seen a dramatic reassessment
of the aforementioned historical picture. As a result of both
serendipity and rational chemical synthesis, whole new classes of molecular and polymeric materials have emerged with
macroscopic characteristics that are in many ways reminiscent of metals.[21These include molecular solids which become superconducting at low temperaturesf2‘. 31 and flexible
polymers with electrical conductivities approaching that of
copper metal.[41However, while many macroscopic properties may be similar to metals, the building blocks which compose these new substances and the synthetic chemistry necessary to prepare them are radically different. Indeed, the new
“synthetic metals”, “molecular metals”, and “organic
[*] Prof. T. J Marks
Department of Chemistry,
Department of Materials Science and Engineering, and
the Materials Research Center
Northwestern University
Evdnston. IL 60208 (USA)
Angnr. c‘l11~m.Int. Ed. Engl. 29
(1990) 857-879
0 VCH
Fig. 1. Schematic depiction of electron occupancy of allowed energy bands for
a classical metal, a semiconductor, an insulator, and a semimetal. The energy
of the highest occupied level is called the Fermi energy. The unoccupied energy
levels are white, the occupied energy levels gray.
metals” are infinitely richer in their architectural diversity
and tailorability. Furthermore, a host of fundamentally new
concepts such as “soliton conductor”, “bipolaron conductor”, and “Mott-Hubbard insulator” have emerged to challenge and enrich our understanding of the condensed state.
Equally important has been the close collaboration between
chemists, physicists, materials scientists, and engineers that
this new interdisciplinary area has spawned. Intellectual
bridges have been built which never before existed and which
allow condensed matter problems of unparalleled complexity to be attacked.
In addition to importance in fundamental solid-state
chemical and physical research, it is also apparent today that
conductive, processable, functionalizable, and light-weight
organic materials offer the possibility of new technologies.12“.4a1 Applications as diverse as energy storage
devices,[51sensors,[61switching devices,”. ‘1 chemoselective
electrodes,[’. 91 static charge-dissipating materials and electromagnetic shielding materials,[’ ‘1 nonlinear optical materials,“ ‘I transparent microphones,” ’] “smart” windows,[’’]
solar energy devices,[’] and electrochromic devicesrI3.l4]
have been proposed. Many of these concepts are in various
Veriagsgesellschaji mbH. 0-6940 Weinheim, 1990
0570-0833i90i0808-0857S 3 . 5 0 i .ZSjO
857
stages of development and, in some cases, commercialization.
Despite the major advances that have been achieved in the
areas of electrically conductive molecular and polymeric
substances, we must also concede that our current level of
synthetic chemical control over, ability to effectively process,
and physical understanding of such materials is at a very
rudimentary level. Chemically, it is apparent that while the
art of synthesizing individual molecules is at a very sophisticated level, the technology for constructing lattices of molecules with precise control over molecular packing and intermolecular interactions is embarrassingly primitive. This
problem of course transcends the subject of molecular metals
and is fundamental to many areas of condensed matter
molecular chemistry (e.g., nonlinear optics, molecular ferromagnets). Likewise, the methodology for turning the latest
generations of conductive organic substances into useful
films, fibers, monoliths, etc., is only now beginning to develop. For conductive polymers, it is apparent that presently
achievable mechanical and electrical properties are far below
the theoretical maxima that optimum processing may afford.
Finally, our fundamental understanding of how macroscopic physical properties connect with microscopic details of
molecular structure, crystal packing, counterion interactions, degree of band filling, disorder, interchain communication, defects, etc., is far from the level where the ab initio
design of new substances can be confidently pursued.
The purpose of this review is to summarize and to analyze
recent results in a research area that in many respects is at the
crossroads of molecular metals and conductive polymers :
electrically conductive assemblies of metallomacrocycles.
The long-range goal of this effort has been to develop rational and versatile synthetic approaches to low-dimensional
molecule-based conductors and, via closely connected physical and theoretical investigations, to better understand their
properties. The strategies to be discussed here have progressed logically from simple systems composed of aggregates of single molecules to more elaborate, structureenforced polymeric assemblies of linked molecular constituents. It will be seen that the design, synthesis, and properties of these assemblies impinge upon issues common to
both molecular and macromolecular conductors and
provide new insights of relevance to both. Due to space
limitations, it is only through these regions of common overlap that it will be possible to mention the excellent work of
numerous other groups now studying molecular and macromolecular conductors.
2. Design Considerations for Molecular Metals
In Figure 2 are shown some of the common building
blocks of molecular conductors. It is generally recognized
that two overriding features are essential for converting an
unorganized collection of organic or metal-organic molecules into an electrically conductive, multimolecular array.
x-x
E
xX
H
x X
7J
TCNQ
TTF(X=S)
TSF (X=Se)
m (X=S)
TSeT (X=Sel
CHI
q=&
;
NC
F
F
TCNQ - F4
TMITF (X=S)
TMTSF(XSe1
NMP
BEDT-TTF
Phen
DMTCNQ
P
aXxXD
x
x
HMTTF(X=S)
HMTSF (X=Se)
FA
DMDCNQl (R=CH,I
DMODCNQI (R=C€H,)
Fig. 2. Common building blocks for nonmetallomacrocylic molecular conductors. The two leftmost columns contain electron donors while the rightmost
contains electron acceptors. TTF = tetrathiafulvalene, TSF = tetraselenafulvalene, TMTTF = tetramethyltetrathiafulvalene, TMTSF = tetramethyltetraselenafulvalene, BEDT-TTF = bisethylenedithiotetrathiafulvalene, HMTTF
= hexamethylenetetrathiafulvalene,HMTSF = hexamethylenetetraselenafulvalene, TTT = tetrathiatetracene, TSeT = tetraselenatetracene. NMP = Nmethylphenazinium, Phen = phenazine, FA = fluoranthene, TCNQ = tetracyanoquinodimethane,
TCNQ-F, = tetrafluorotetracyanoquinodimethane,
DMTCNQ = dimethyltetracyanoquinodimethane, DMDCNQI = dimethylN,N-dicyanoquinodiimine, DMODCNQI = dimethoxydicyanoquinodiimine
First, the molecules must be located in close spatial proximity and in similar crystallographic and electronic/Coulombic
environments so that an energetically uniform (i.e., devoid of
Tobin J: Marks was born in Washington, D.C., in 1944. He received a B.S. degree from the
University of Maryland in 1966 and a Ph.D. from A4.I.T in 1970, working under I? A. Cotton.
He moved to Northwestern University as an assistant professor in 1970. Tobin Marks is currently
Charles E. and Emma H. Morrison Professor of Chemistry and Professor of Materials Science
and Engineering at Northwestern. He serves on a number of editorial boards and advisory panels,
has won several awards, including the American Chemical Society Award in Organometallic
Chemistry, and is currently Associare Editor of Organometallics. His present research interests
include polymer and solid-state chemistry, f-element organometallic chemistry, homogeneous and
heterogeneous catalysis, nonlinear optical materials, metal-organic chemical vapor deposition,
and carcinostatic metal complexes.
858
Angew. Chem. Inl. Ed. Engl. 29 (1990) 857-879
carrier-trapping hills and valleys), extended pathway exists
for electronic charge diffusion. This is equivalent to saying
that a band structure having sizeable bandwidths ( 20.5 eV)
must exist. Such a situation is commonly realized when flat,
conjugated molecules crystallize in various types of stacking
motifs which allow extensive overlap of the x-electron systems (e.g., 1 or 2). The way in which a band structure is built
o a
1
subunit (the occupancy of the HOMO), the quantities (2-e)/2
and lel/2 give the band occupancy (e.g., 2/3-filled, 1/3-filled)
for e > 0 and e < 0, respectively. Equally important is the
observation that materials with e = 1 (half-filled bands for
classical metals) are never found to be molecular metals and
are usually insulators.
The above complexities are due to a number of factors. As
was first noted for certain transition-metal oxides by Mott,
situations may occur in which the Coulombic repulsion (denoted by the parameter U) of having two charge carriers on
a single site [Eq. (l)] is comparable to or exceeds the band-
2
up from highest occupied molecular orbitals (HOMOs) as
increasing numbers of molecules are added to a cofacial
molecular stack is illustrated in Figure 3. In an simplistic
+
width. In the extreme case where U 4t, a material with a
half-filled band can actually be a magnetic insulator.[’. *I
That is, all carriers are confined to their individual molecules. A valence-bond picture of how mixed valency affects
charge mobility in molecular conductors is shown schematically in Figure 4. Currently, the degree to which U can be
00
-
00
+o
-
+o
+o
_
_
-
_
-
-
00
00
coo-;-
00
o+
-
Fig. 3. Schematic illustration of how a band structure is built up by arraying
increasing numbers of molecular subunit HOMOs in a molecular stack. The
parameter I is the tight-binding transfer integral. analogous to the Hiickel /j
integral. The tight-binding bandwidth is 41.
Hiickel-like tight-binding band-structure description,[’”. b1
the measure of how strongly neighboring HOMOs interact is
given by the transfer integral t , which is directly analogous to
the Huckel resonance integral, 8. In an extended stack consisting of a very large number of molecular constituents, the
bandwidth is given by 4t (48). This important condensed
state parameter is directly connectable to molecular properties and as such is an important quantity to accurately measure and understand.
The second key requirement for metallic characteristics in
a molecular stack is that the highest energy band be incompletely filled. This characteristic has already been discussed
in reference to Figure 1 for classical metals; for molecular
systems, however, several important caveats must be added.
In all of the molecular metals prepared to date, the substances are actually salts and the molecular subunits are in
formal ,fractional oxidation states (“partial oxidation” (or
reduction), “incomplete charge transfer”, “mixed valence”).
That is, the molecules have valence shells which, formally,
are fractionally occupied as in (TMTSFo.5@),XG,TTF0.59@TCNQ0.590,and [ H , ( P C ) ~ ~ ’ ~ @ ] (the
I ~symbol
) ~ , ~ ~e; is normally used to designate the subunit oxidation state. Since
each of the simple bands of the type shown in Figures 1 and
3 can accommodate two electrons per constituent molecular
Ang”s. (‘hem. In!. Ed. EngI. 29 (1990) XS7-879
00
00
c:+:
_+ o
+o
++
+o
c;-+:
00
_
00
_
00
_
-0 0
+o
-
00
_
_
_
00
_
+o
-
00
00
+o
_
+o
Fig. 4. Schematic depiction of how the degree of oxidation affects charge mobility in a typical molecular stack. Left: no oxidation. Middle: partial oxidation. Right: complete oxidation (one electron per subunit).
manipulated by alterations in molecular architecture, crystal
packing, neighboring counterions, and level of band filling
has been the subject of much speculation. Such relationships
are presently obscure, large U effects may be entangled with
other structural features, and accurate measurements of U
are not without ambiguity. It is generally agreed that enhanced mangnetic susceptibilities are, in most molecular
conductors, an indication of large U effects (“Coulomb enhancement”), as are certain types of band-filling-dependent
structural distortions (4k, instabilities, where k, refers to the
electronic wave vector at the Fermi level).[’b*51
In addition to Coulombic effects, the properties of a
molecular metal can be exceedingly sensitive to the coupling
of the electronic system to certain phonons (lattice vibrations). Such an instability is analogous to a Jahn-Teller distortion and is predicted theoretically for a low-dimensional
metal. The net consequence can be a periodic structural
distortion and charge buildup in the stack (a charge
density wave, Peierls distortion, or 2k, instability), the periodicity of which reflects the filling of the highest occupied
151 In the extreme case (especially for half-filled
bands or at low temperatures) the distortion “locks in” and
an opening of a gap in the conduction band can induce a
metal-to-insulator transition. Since other structural effects
such as interactions with counterions in the crystal lattice can
also induce periodic structural distortions in the stacking
’
859
geometry, identification of the origin of charge-density-wave
phenomena is not always unambiguous. Qalitatively, such
phenomena become less important as the molecular chain
deviates strongly from unidimensionality (i.e., interchain interactions become very large). Electron-phonon interactions
are also believed to play a crucial role in a variety of chargetransport mechanisms, including organic superconductivity[2e,3z161
and the cr T - 2 behavior frequently observed in
molecular metals.“ ’ 1
From a synthetic chemical perspective, the most obvious
way to clarify many of the above phenomena would be to
prepare and study molecular conductors in which there was
sequential variation of one crucial parameter (e.g., molecule-molecule interplanar spacing, off-axis coun terion, band
filling) while all other parameters were held fixed. Unfortunately, current routes to molecular metals are not subject to
such high degrees of control. Thus, conventional syntheses
usually rely on crystallization combined with a chemical or
electrochemical oxidation or reduction process as in Equation (2). The success of such a self-assembly process in form-
-
a ap+ *a ap+
a - ap+
A-
erties. It is therefore apparent that any rational synthetic
strategy for disentangling the numerous molecular-metal
architectural/electronic structural variables must possess a
means of rigorously controlling the subunit stacking rather
than relying upon capricious packing forces. The principal
focus of the present article, then, is upon the chemical and
physical consequences of effecting, via the introduction of
strong, covalent intrastack bonds, the connectivity modification shown in Equation (3).[’*J
a-0
(3)
It should also be appreciated that structure 6 is a macromolecule. The field of electrically conductive polymers has
developed in parallel to that of molecular metals, and representative structures are illustrated in Figure 5.r41 In common
/
Pol yacetylene
ing segregated stacks of closely arrayed radical cations (or
anions) and evenly distributed, complementary arrays of offaxis counterions is by no means guaranteed. The structural
outcome hinges upon a complex and largely uncontrollable
collection of Madelung, exchange, polarization, bandwidth,
van der Waals, core repulsion, crystal-growth kinetic, and
ionization potential/electron affinity factors, which determine, in poorly understood ways, the level of partial oxidation (e) and the crystallization architecture. Thus, charge
transfer may not occur, and if it does, numerous other packing modes are possible (e.g., 3-5).
Polypyrole
*Din,
N
H’
Polythiophene
Polyaniline
Polyphenylenevinyiene
Fig. 5. Precursors of common electrically conductive polymers.
3
4
5
Most embodiments of motifs 3 and 4 lack an obvious
conduction band structure and, to date, such materials have
been found to be insulators.[2. In addition, simple changes
in the molecular building block or off-axis counterion in
these syntheses typically induce major changes in stacking
architecture, hence in a myriad of correlated physical prop860
with stacked molecular systems, oxidation or reduction
(commonly called “doping” as for conventional inorganic
semiconductors) is necessary to mobilize charge carriers.
However, the bandwidths of most highly conductive polymers are greater, and charge-transport mechanisms may be
different. Furthermore, structural regularity is not as great
as in molecular conductors, and a number of classes of conductive polymers are poorly crystalline or amorphous. This
factor and the accompanying structural uncertainties severely complicate the interpretation of chemical and physical
data.
Angew. Chem. In!. Ed. Engl. 29 (1990) 857-879
3. Building Blocks and Model Materials
Reference to target structure 6 indicates that a suitable
molecular subunit must have chemical functionality perpendicular to the molecular plane. An ideal synthon for such a
strategy is a metallomacrocycle in which the central metal
atom has the possibility of bond formation perpendicular to
the square coordination plane. Early work in this laboratory
established that partial oxidation of glyoximates (7,8),[191
dibenzotetraazaannulenes(9),[”’ hemiporphyrazines (10),1211
and phthalocyanines (11)[”] with halogens afforded new
Fig. 6. The crystal structure of the macrocyclic molecular metal Hz(Pc)(IJo J 3
viewed parallel to the molecular stacking direction (crystallographic c axis).
10
11
families of low-dimensional molecular conductors. In all cases, spectroscopic analysis of the form of the halogen present
in the crystal structure showed unambiguously that the
metallomacrocyclic subunits were in formal fractional oxidation states.[’8b*231
Metallophthalocyanines, M(Pc), are robust, highly functionalizable commodity chemicals that are widely used as
pigments and oxidation catalysts.[241More importantly for
the present purposes, they form the basis for a class of
molecular metals in those cases where the tactic of Equation
(2) is successful. A particularly informative example is
H,(Pc)(I,),~,,[~~~
(11, where M = two H atoms), which exhibits many of the physical properties typical of a molecular
metal. As can be seen in Figure 6, the tetragonal crystal
structure consists of H,(Pc)O.~~@cations, arrayed at
3.251(2)-8, intervals and staggered by 40.0 with respect to
neighboring subunits. Chains of disordered I,” counterions
extend parallel to the stacking direction. This particular
structural pattern will be seen to be ubiquitous for phthalocyanine metals. The electrical conductivity of these exceedingly thin crystals (0.07 x 0.07 x 0.65 mm) in the stacking
direction (oil) is ca. 7 0 0 K 1 c m - ’ at room temperature.
When normalized for the large H,(Pc) cross-sectional area,
this room temperature conductivity is one of the highest yet
reported for a molecular metal. The high unidimensionality
of the charge transport is demonstrated by the result
o,,/o,
500. As the temperature of H , ( P C ) ( I ~ ) crystals
~,~~
is slowly lowered, the stacking-axis conductivity becomes
metal-like (Fig. 7, top, do/dT < 0), reaches a maximum at
ca. 15 K, and exceeds 3 5 0 0 K 1 c m - ’ even at 1.5 K. Also
shown in Figure 7, top, is the thermoelectric power (TEP) of
Angrw. Chem. Int. Ed. Engl. 29 (1990) 857-879
an H2(Pc)(13)o.33crystal measured in the stacking direction.
The positive sign of the Seebeck coefficient (S)indicates that
this material is a radical-cation conductor (the major conduction pathway is through the [ H , ( P C ) ~ . ~ ~chain),
@ ] , while
the relatively small magnitude and linear dependence on
temperature are classic signatures of metallic behavior.[26-281Note the change in the thermoelectric power toward semiconducting behavior at the temperature (ca. 15 K)
where a , ( T )passes through a maximum. In a one-dimensional tight-binding picture, the temperature dependence of
S can be related to the bandwidth (4t, Fig. 3) via Equation
(4).[26 2 8 , 291 Here, k , is the Boltzmann constant. Analysis
~
S=
23c2k~Tcos(n~/2)
3 e ( 4 t ) sinZ(xe/2)
(4)
of the S ( T ) data for Hz(Pc)(13)o.33yields a tight-binding
bandwidth of 1.0(1) eV. In agreement with this ligand radical-cation-based picture of charge transport, fully assigned
solid-state I3C CPMAS NMR data[30]evidence large, locally resolved Knight shifts (due to nuclear-conduction electron
interactions) consistent with a phthalocyanine radical cation. ESR spectra are characterized by nearly free-electron g
values and narrow line widths. These are signatures of a
radical-cation conductor with the narrow line width indicating a high degree of unidimensional character (in accord with
charge transport and optical data) and only minor interaction of the carrier spins with the I,” counterions (i.e., only
small contributions to electron-spin relaxation by phononmodulated spin-orbit interactions).[31.321
The optical and magnetic properties of Hz(Pc)(13)o.33
crystals also provide classic signatures of metal-like character. In
Figure 7, middle, is shown the reflectance spectrum of an
H2(Pc)(13)o.33
crystal polarized parallel to and perpendicular
to the macrocycle stacking direction. The small size of the
crystals necessitated the development of specialized microspectrophotometry techniques.[25]The most striking feature
of the spectrum is the parallel-Polarized edge[33,341 observed
at ca. 3100 cm-’(lg v =: 3.49). This feature is analogous to
861
480
4000
70
f
6o
s
50
[pV K-'I
H J P C ) ( I ~ ) is
~ ,only
~ ~ metallic (there is a significant bandwidth) in the stacking direction. Analysis of the optical data
yields the carrier relaxation time and mass, the plasma frequency (w,), and, via the tight-binding expression of Equation (5),[33,341the bandwidth. Here, N , is the carrier density. Analysis of the Hz(Pc)(13)o.33reflectivity data yields
40
41 =
111
20
500
10
/I
50
0
100
150
T IKI
"O
ZOO
-
250
300
I
0t
0
2 00
100
T [KI
-
300
Fig. 7. Physical properties of the macrocyclic molecular metal H2(Pc)(IJ0 J 3 .
Top. Variable-temperature electrical conductivity and thermoelectric power
measured parallel to the molecular stacking direction. Middle: Reflectance
spectrum of an HI(Pc)(I3)" single crystal; R., and R , denote reflectance parallel to and perpendicular to the molecular stacking direction. The perpendicularlypolarizedfeaturesat 14800(Igv = 4.17)and29000cm-' (Igv = 4.46)are
assigned to Pc-centered FIT* transitions. while the parallel-polarized feature at
18200cm-' (lg v = 4.26) is assigned to an I?-centered excitation. Bottom:
Variable-temperature magnetic susceptibility (1) data for a polycrystalline
H J P C ) ( I ~3)3~sample. The filled circles denote the Pauli-like component. the
open circles the paramagnetic Curie-like component.
the plasma edge observed in the reflectivity of metals["] and
can be associated with carrier excitation in the partiaIIy filled
conduction band. The marked polarization indicates that
862
e(hq,S
4 N , ez e2 sin (K p/2)
30
(5)
4t = 1.3(1) eV, in excellent agreement with the thermopower
data. Static magnetic susceptibility data for H2(Pc)(13)o.33
crystals are shown in Figure 7, bottom. When a small Curielike component of the type normally associated with defects
and impurities[35*361 is subtracted, the remaining susceptibility is only weakly paramagnetic and virtuaIly independent
of temperature. This behavior is reminiscent of the Pauli
susceptibility of bulk metals,['d*elwhich is a consequence of
the partially filled character of the conduction band. Assuming a one-dimensional tight-binding band and noninteracting electrons ( U = 0), a tight-binding bandwidth can be
derived from Equation (6).C3'1 The result for H2(Pc)(13)o.33is
that 4t = 0.37(3) eV, in clear disagreement with the afore-
mentioned TEP and optical data. The implication of this
finding is that the magnetic susceptibility is "enhanced",
that is, that U and therefore Coulombic repulsions are not
insignificant.[38*391 In summary, Hz(Pc)(13)o,33 exhibits
many of the classic features of a molecular metal, to which
we will make reference in later discussion: (1) metal-like
conductivity in the stacking direction; (2) thermoelectric
power which is small in magnitude and the absolute value of
which is linearly proportional to temperature; ( 3 ) a plasmalike edge in the reflectivity; (4) Pauli-like paramagnetism.
These results also clearly demonstrate that a metal-ion substituent is not required in the core of the macrocycle for
metal-like character. How reliable the formalisms are that
allow extraction of 4t from condensed-phase TEP and optical data is an important, unanswered question which will be
taken up in Section 5.
Although the molecular self-assembly tactic of Equation
(2) was a useful first-generation approach, subsequent studies have shown it to have limited scope. Several additional
salts have recently been prepared by high-temperature electrocrystallization procedures developed in this laboratory,
such as N~(PC)(BF,),[~*'and Ni(Pc)(C104),.[411 The crystal
structure of the latter material (y = 0.42) is shown in Figure
8 and is similar to that of H2(Pc)(13)o.33,as are those of
Ni( Pc)( I 3 ) o , [22c1 and Ni( Pc)( BF,),
.[401
Electrical, optical, and magnetic studies of these materials
indicate metallic character in the macrocycle stacking direction and further show that the presence of the Ni ion has a
reIativeIy minor influence on the charge-transport characteri s t i c ~ . [ ~ ~ Comparative
-~*]
charge-transport data are presented in Figure 9. Ni(Pc)(ClO,), also displays the interesting characteristic that y values for different crystals can vary
slightly (0.39 -0.47) as a function of electrocrystallization
conditions,14 'I presumably reflecting different occupancy
levels of the anion channels. Such modest variations in
Angen. Chem. Int. Ed. Engl. 29 11990) 8S7-879
I500
500
0
Fig. 8. The crystal structure of Ni(Pc)(ClO,),
viewed parallel to the macrocycle stacking direction. The ClOQ ions are drawn as open circles for clarity.
molecular-metal band-filling/stoichiometry level are unusual
but not completely unprecedented[431and, in this case, are
evident in TEP and optical measurements (cf. the e dependence in Eqs. (4) and (5)). Perhaps more interestingly, the
CIO? ions are rotationally disordered (and likely tumbling)
at room temperature in these crystals, but lock into ordered
positions at lower
The effects of this counterion ordering on the collective properties of Ni(Pc)(CIO,),
are rather small compared to the effects on (TMTSF),CIO,,
where cooling-rate-dependent anion-ordering phenomena
have a dramatic effect on superconducting behavior.[4.4 5 1
Beyond the above small anions, attempts via self-assembly
to probe phthalocyanine band-structureecounterion interactions are frustrated in that electrocrystallization experiments
with larger anions (e.g., A s F ~ SbF2)
,
lead to materials with
different packing arrangements and e values.'42f1Likewise,
the use of organic oxidants such as high-potential quinones
(e.g., TCNQ) fails to produce molecular metals, yielding instead insulators which are presumed to have integrated
stacking as in 3 or 4.14@Clearly, the forces retaining the
stacking architecture in these ligand-based conductors are
rather delicate and unreliable. Comparative physical data on
phthalocyanine molecular conductors are compiled in Tables
1 and 2.
i
300
4000t
0
50
100
200
150
T LKI
250
-
250
c
425
i'
4. Cofacial Condensation Assembly as an Approach
to Enforced Molecular Stacking
A particularly attractive approach to target structure
6 employs the solid-state condensation polymerization
of dihydroxy group 4 phthalocyaninato complexes
(Fig. 1O).[' *, 47 - 4 9 1 These robust, highly crystalline rigid-rod
polymers have been characterized by various end-group
analyses (radiochemical, FT-IR),r47a1
solid-state NMR spect r o ~ c o p y , [electron
~~]
microscopyJS0]and powder X-ray diffraction [47al (aided by computer simulation techniques and
single-crystal diffraction studies of model olig~mers[~']),
and solution diffraction techniques (for more soluble, ringfunctionalized derivative^).^'^] Together, these data provide
a detailed picture of [M(Pc)O], architecture and crystal
Angew Chrm. In:. Ed Engl. 29 (1990) 857-879
350
300
0
50
100
150
T [KI
-
ZOO
250
300
Fig. 9. Variable-temperature macrocycle stacking axis conductivity ( 0 ) and
thermoelectric power (S)data for representative crystals of Ni(Pc)(I& 3 3 (top),
Ni(Pc)(BF,), 3 3 (middle), and Ni(Pc)(CIO,),, y z 0.40 (bottom). Measurements on these exceedingly fragile crystals were performed with cooling rates as
low as 1 K h - I .
structure. For example, n is typically 50-200, depending
upon the polymerization conditions (n 2 100 is used for
most physical experiments), and Pc-Pc interplanar spacings
expand with M ionic radius in the following order: 3.32(2)&
863
Table 1. Structural, degree of charge transfer, transport, magnetic, and optlcal parameters for electrically conductive phthalocyanlne polymers and model compounds [a].
Compound
[a-' cm-'I
Tetragonal lattice [b] [A]
e
53001(
u = 13.97(5)
c = 6.60(4)
0.35
0.60
2.35(10)
4540
a = 13.70(7)
c = 6.58(4)
u = 13.96(7)
0.35
0.10
2.22(6)
4700
0.50
0.13
2.32(6)
5390
0.35
0.10
2.49(6)
4600
0.47
0.10
0.35
0.10
2.22(3)
4550
0.41
0.10
0.67
0.04
+ 26
3.13(6)
5000
0.19
0.09
i
49
1.93(14)
4870
0.35
0.10
2.70(10)
4210
0.33
ca. 700
8.0 [d]
ca. 500
7.5[d]
ca. 1000
6.0[d]
2.21(5)
6360
1.90(10)
5630
1.30(4)
6020
2.62(8)
7700-9000
S,,,, [pVK-']
~ [ l O ~ ~ e r n u m o l - ' ] w,[cm-']
~
c = 6.66(4)
u = 13.98(6)
c = 6.58(4)
u = 14.08(7)
c = 6.63(4)
u = 14.31(4)
c = 6.58(4)
u = 14.19(7)
c = 6.61(4)
u = 14.39(7)
c = 6.64(4)
u = 13.86(7)
c = 6.67(4)
u = 13.96(5)
c = 6.96(4)
a = 13.928(5)
c = 6.502(3)
u = 13.936(6)
c = 6.488(3)
u = 13.97(2)
c = 6.48(1)
u = 13.957(3)
c = 6.4672(9)
0.33
0.33
0.42
+0.3
i
62
+ 59
+ 50
+ 25
100-1500
[a] All data for polycrystalline samples except where indicated. [b] Pc-Pc interplanar spacing = 42.[c] Single-crystal specimen. Conductivity and thermopower
measured in the Pc stacking direction. [d] Comparative polycrystalline measurement.
Table 2. Representative bandwidth information for phthalocyanine conductors.
~~
~~
Compound
Interplanar
separation [A]
Bandwidth [a]
Optical [eV]
Bandwidth [b]
TEP [eV]
Bandwidth [c]
PES [eV]
Bandwidth [d]
Magnetic [eV]
Bandwidth [el
Theoretical [eV]
H2(PC)(I3)03 3
Ni(Pc)(I3f0J3
GW")Ol(L)o.35}n
&CW")Ol(BF& 35)n
USi(Pc)OI(BF&
3.251(2)
3.244(2)
3.30(2)
3.29(2)
3.33(2)
3.48(2)
1.2(1)
l.OO(9)
0.60(6)
0.64(6)
0.63(6)
0.48(5)
1.0(1)
1.0(1)
-
0.37(3)
0.43(3)
0.32(3)
0.37(3)
0.26(1)
0.28(2)
0.85
0.85
0.76
0.76
0.76
35}n
0.58(6)
0.58(6)
0.58(6)
0.38(6)
0.70(7)
0.50
[a] From optical reflectivity data analyzed via Eq. (5). [b] From thermoelectric power data analyzed via Eq. (4).[c] From gas-phase photoelectron spectrum of
ROM(Pc)OM(Pc)OR dimer. [d] From magnetic susceptibility analyzed via Eq. (6).[el From DV-Xlr calculation on a cofacial dimer.
M = Si; 3.53(2)& M = Ge; 3.82(2)& M = Sn. Combined
with the data base of model molecular conductors discussed
above, these stacking-enforced arrays of ligand-based con-
- H20
M
ductive building blocks offer a unique opportunity to probe
many of the poorly understood features of molecular metals
as well as conductive polymers. We begin with a discussion
of chemical and electrochemical redox phenomenology and
the physical properties of the resulting materials.
4.1. Chemical Doping of (M(Pc)O], Materials:
Interplanar Spacing and Counterion Effects on Collective
Properties
*
Partial oxidation of the [M(Pc)O], polymers with halogenrS3'and nitrosyl salts[541proceeds as shown in Equations
(7) and (8). X-ray diffraction studies reveal that the doping
= Si. Ge, Sn
[M(Pc)Ol,,
Fig. 10. Strategy for the cofacial assembly of structure-enforced group 4
metallophthalocyanine assemblies.
864
+
[M(Pc)O],
0.35n NO@Ze
{[M(Pc)O]Z,,,},
M = Si; Ze= BF?, PF:, SbFz
+
+ 0.351NO
(8)
Angew. Chem. Int. Ed. Engl. 29 (1990) 857-879
0-rh.
tetr.
tetr.
Fig. 11. Structural model of the changes accompanying chemical doping of as-polymerized [Si(Pc)O]. with I,. followed by thermal undoping. The view is parallel to
the phthalocyanine stacking dlrection. o-rh = orthorhombic, tetr. = tetragonal
processes are largely inhomogeneous. That is, incremental
addition of oxidant produces progressively larger amounts
of the fully doped phase, having a discrete stoichiometry
(fixed e) and a different crystal structure. This process is
illustrated schematically for [Si(Pc)O], in Figure 11. Until
full uptake of oxidant is achieved, the material is a heterogeneous mixture of starting material and product. This inhomogeneity has been verified by transmission electron microscopy.[5o1X-ray diffraction studies also indicate that the
tetragonal doped phases are essentially isostructural with the
molecular phthalocyanine analogues (cf. H2(Pc)(13),,33 in
Fig. 6 and Ni(Pc)(C104)o,42in Fig. 8), differing principally in
interplanar spacing. Relevant structural data are compiled in
Table 1. It can also be seen in Equations (7) and (8) that the
ultimate degree of partial oxidation (band depletion, e) is ca.
+ 0.35 for all oxidants. Born-Haber cycles using proper
AHsub,,
and ionization potential data for molecular phthalocyanines show that this is not an unexpected result.[541Drastic modifications in e and off-axis counterions are best
achieved using electrochemical techniques (see Section 4.2).
Interestingly. the products of Equation (7) can be “undoped” by heating under vacuum to yield a new [Si(Pc)O],
phase which is best modeled by a tetragonal crystal structure
(Fig. 1 1 ) .
The series Ni(PC)(13)0.33(1); &W“)Ol(13),.35}, (111, and
{[Ge(Pc)O](I,),,,,}, (111) provides an extremely informative
picture of how an expansion in interplanar spacing from 3.24
to 3.30 to 3.48 A modifies the properties of a molecular
metal. Comparative variable-temperature electrical conductivity data are shown in Figure 12. In viewing these data, it
should be noted that measurements on polycrystalline samples (high-quality single crystals of the doped polymers are
not available) are influenced by the isotropy of the macroscopic specimen and interparticle contact effects.[55.561
Hence, while full “metal-like’’ character is partially masked
(cf. the Ni(Pc)(13),,33data), an empirical rule of thumb[55,5 7 1
suggests that polycrystalline conductivity values can be multiplied by a factor of ca. 500 to estimate single-crystal conductivity in the stacking direction (all). This approximate
relationship has been verified for a number of molecular
conductors and, in the present case, allows meaningful, although not strictly quantitative, comparisons to be drawn.
Most importantly, the data evidence a dramatic diminution
in the efficiency of charge transport as overlap decreases
between the adjacent macrocycle x systems. In accord with
the inhomogeneous character of the doping processes, the
doping-level dependence of the conductivity is not monotonAngen. Chum. Int. Ed. Engl. 29 (1990) 857-879
100
t
~
lo-*
10-L
[R-’crn-’l
10-6
10-8
10-10
0
LO
80
120
1000 T-’ IK-’1
160
-
200230
270
T[KI
-
310
Fig. 12. Variable-temperature electrical conductivity data for compressed
(11). and
polycrystalline samples of Ni(Pc)(13)o.33(I), ([Si(Pc)O](I,),,,].
{[Ge(Pc)0](1,),,,}.
(111). Note that metal-like behavior observed in
Ni(Pc)(I,),33 single crystals (Fig. 9, top) is modified by the polycrystalline
nature of the samples.
ic, but exhibits an abrupt discontinuity which can be associated with a percolation threshold, as found in many composite materials.[581Such discontinuous behavior is observed
in mixtures of conductive and nonconductive particles when
the volume fraction of the former reaches a certain critical
threshold beyond which there is a continuous network of
contacting conductive particles. At this point-the percolation threshold-the material becomes highly conductive.
The temperature dependence of ([Si(Pc)O]X,), conductivity
can best be fit to a fluctuation-induced carrier tunneling
in which relatively large conductive regions are
separated by insulating barriers having a parabolic potential.
At low temperatures, transport is dominated by temperature-independent elastic tunneling, while it is thermally activated at higher temperatures. Importantly, the a( T ) data
cannot be convincingly fit to conventional carrier hopping
models which apply to many conductive polymers.[41
The sensitivity of band-structure-related properties to PcPc interplanar separation is also evident in comparative optical reflectivity data, where a progressive shift of the plasma
edge to lower energy is observed with increasing interplanar
spacing (Fig. 13). This behavior is in excellent agreement
with the simple tight-binding band-theoretical picture
[Eq. ( 5 ) ] . Finally, while the magnetic susceptibilities of 1-111
are all essentially Pauli-like (cf. Fig. 7), the susceptibility increases incrementally with increasing interplanar spacing,
again in accord with simple theoretical notions [Eq. (6)].
865
0 ,
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
3.6
3.8
4.0
lg v ~crn-’]
Fig. 13. Optical reflectance spectra of compressed polycrystalline samples of
N1(Pc)(L)o3 3 (I), { [ S ~ ( P C ) O I ( I ~ )(11).
~ . ~and
~ ) . {[Ge(Pc)Ol(~3)o.35).
(111). Spectra
are offset vertically by 0.2 reflectance units for ease of viewing.
-
igv ~ c m - ’ ~
Fig. 14. Optical reflectance spectra of compressed polycrystalline samples
of {[Si(Pc)O(SbFd0,),.
{[Si(Pc)Ol(PFJ0351n. {[Si(Pc)Ol(BFd)o3 S ) n and
~ [ S ~ ( P C ) O ] ( I , )Spectra
~ . ~ ~ ~ are
” . offset verticalIy by 0.2 reflectance units for ease
of viewing.
Relevant physical data and derived 4t values are set out in
Tables 1 and 2. It can be seen that reflectivity-derived (and
susceptibility-derived) 4t values fall with increasing interplanar spacing. All of the susceptibility-derived bandwidths
structure ( t , is small, in agreement with the other physical
evidence some degree of Coulombic enhancement. At this
measurements). The anion-dependent line width effects likestage, a completely independent measure of 4t, for example,
ly reflect minor interactions with the off-axis anions and
not tied to the many approximations inherent in the conresulting electronic relaxation via phonon-modulated spindensed-state derivation, would be highly desirable. This suborbit interaction^.[^ ’]
ject will be taken up in Section 5.
The present results stand in sharp contrast to those for
The intriguing question next arises as to the nature and
most organo-chalcogen-based molecular conductors and sumagnitude of band-structureecounterion interactions. If the
perconductors.[’. 3 , 451 In these cases, small changes in coun[S~(PC)~@)O],
stacking is essentially frozen in these materials,
terion effect significant changes in crystal packing, hence
to what degree are the charge carriers influenced by whether
major changes in charge-transport, optical, and magnetic
off-axis anions are large and highly polarizable (e.g., 1:) or
properties; differences in resulting properties are sometimes
small and hard (e.g., BF?)? This question was addressed via
as large as those expected for a superconductor rather than
magnetic, optical, charge-transport, and ESR measurean insulator. For the present [ S ~ ( P C ) ~ . ~ ~materials,
@ O ] , the
ments. Magnetic susceptibility data are one means to probe
insensitivity appears to arise from two major factors. The
the electronic structure and population of the highest occufirst and probably most important factor is the essentially
pied band, and the result is that all of the [S~(PC)~.~’@O], frozen character of the stacking architecture, so that bandmaterials that have been studied exhibit Pauli-like paramagstructure-counterion interactions must be Coulombic or innetism that is virtually identical in magnitude and temperavolve subtle static or dynamic modifications in Pc-Pc stagture dependence (Table 1). Any effects of the counterion on
gering angles. The second factor is likely connected with the
the Coulombic enhancement (“screening” effects of this type
unique M(Pc) x electronic structure and will be discussed in
have frequently been invoked for other molecular conducSection 5.
tors) are undetectable by this technique. Optical reflectivity
studies of these materials reveal essentially identical, plasma4.2. Electrochemical Doping of [M(Pc)O], Materials:
like edges (Fig. 14), with derived o,and 4t parameters virtuStructure, Counterion, and Band-Filling Effectsally independent of counterion (Tables 1 and 2). Likewise,
Tunable @
variable-temperature conductivity data for the partially oxidized [S~(PC)~.~’@O],
series evidence only slight differences
Electrochemical techniques have been employed extenof the type that can be associated with slightly different intersively to effect the “doping” of electrically conductive polyparticle resistances and different chain-packing densities in
m e r ~ . [In
~ ]general, those structural changes which accompathe respective crystal structures. Data are set out in Table 1 .
Studies of ESR data for polycrystalline ( [ S ~ ( P C ) ~ ~ ~ ~ @ O ] X ,ny
} , oxidation and reduction cycles are not well-understood.
For the [M(Pc)O], polymers, electrochemical studies offer an
reveal nearly isotropic free-electron g values, which are conopportunity to address these issues as well as to vary anions
sistent with ligand-based oxidation, and relatively narrow
and modify e to an extent not readily possible with chemical
line widths that increase in the progression BF,“ < PFZ <
reagents.
SbFF < I,” (0.36 to 2.9 G at 300 K). As noted in Section 3
Electrochemical studies of [Si(Pc)O], oxidative (reductive)
for the molecular phthalocyanine conductors, narrow line
doping have been carried out under rigorously anhydrous
widths can in general be associated with the absence of heavy
and anaerobic conditions in several experimental configuraatoms in extensive communication with the carriers as well
tions in which the polymer solid is in direct contact with a
as with the highly unidimensional crystal and electronic
866
Angew. Chem. Inr. Ed. Engl. 29 (1990) 857-879
o-rh
tetr.
tetr.
Fig. 15. Structural model of the changes accompanying initial electrochemical oxidation of as-polymerized [Si(Pc)O], . Initial “break-in” is accompanied by a
transformation to a more open structure which allows more facile redox cycling. The view is parallel to the phthalocyanine stacking direction.
liquid electrolyte containing the anion (cation) of interest.‘“. 6 1 1 Most studies involve an automated electrochemical potential spectroscopy (ECPS) experiment in which the
applied potential is changed in steps, and at each step Coulombs of charge passed in the doping/undoping process are
measured. The result is a map of applied potential versus
doping stoichiometry, with the precise shape of the curve
indicating where transformations between phases occur.
Preparative-scale experiments are used to provide macroscopic samples for elemental analysis and diffractometric,
magnetic, optical, and charge-transport characterization.
Perhaps the most interesting feature in the oxidative electrochemistry of “as-polymerized” [Si(Pc)O], (orthorhombic;
cf. Fig. 11) in common electrolytes such as nBu,N@BF?/
acetonitrile is the observation of a large overpotential/kinetic barrier upon initial oxidation.[601The oxidation product is
a tetragonal ([Si(Pc)O](BF,),}, material, structurally similar
to that prepared with NO@BFF [Eq. (S)] except that
y = 0.50 rather than 0.35. Importantly, subsequent electrochemical undoping and doping cycles for this material no
longer exhibit the large overpotential, and the shape of the
ECPS curve indicates that &) can be tuned smoothly between 0.00 and 0.50. The reason for the initially difficult
oxidation (commonly referred to as “break-in” behavior’,, 621 but poorly understood) followed by facile cyclability can be directly correlated with the structural information.
The break-in doping involves an orthorhombic to tetragonal
phase transformation (Fig. 15) in which considerable reorganization of the [Si(Pc)O], chains and expansion of the crystal
lattice is required. In contrast, subsequent cycling involves
interconversions between structurally similar doped and undoped tetragonal phases, requiring only anion migration
through “tunnels” parallel to the stacking direction and minimal reorganization of the [Si(Pc)O], chains (Fig. 15). The
undoped phase produced in the electrochemical cycling is
spectroscopically, electrochemically, and diffractometrically
indistinguishable from that afforded by thermally undoping
{ [Si(Pc)O](I 3)o. ) (Fig. 1 1). This electrochemical behavior
is unaffected by the degree to which the solid polymer sample
has been finely ground, the number of redox cycles performed, or the addition of electrochemical mediators such as
[0s(bpy)lze. Further confirmation that the doping level in
{ [Si(Pc)O]BF,),}, can be homogeneously tuned is provided
by the observation that the tetragonal a lattice parameter
varies monotonically with doping level (Fig. 16). This is the
first example where the band-filling level in a molecular
metal could be homogeneously tuned over a broad range
Angun Churn. Inr. Ed. Engl. 29 (t990) 857-879
0.3
4
0.2
-
0.1
-
0.0
-
Aa/c
[A1
-a-
I,;
I
-0.1
0
0.10
0.30
0.20
0.40
YFig. 16. Change in unit-cell parameters (I and c as a function of degree of partial
oxidation for {[Si(Pc)O](BF,),}..
without a major change in crystal structure. In the closest
previous example, e in the system of “alloys”
(NMP),(Phen), -,(TCNQ)
(NMP = N-methylphenazinium; Phen = phenazine) has been tuned from ca. - 0.50 to
641
- 0.66 in the TCNQ
Detailed studies of [Si(Pc)O], electrochemistry have been
carried out with a wide variety of anions of difrering size,
shape, polarizability, and charge.[601Representative ECPS
plots are shown in Figures 17 and 18. Excepting cases where
the electrochemical window of the electrolyte is exceeded, the
general pattern governing the maximum stoichiometry of
doping, hence the degree of band depletion that can be
achieved, appears to be largely determined by the anion size.
Simple packing calculations employing anion van der Waals
radii show that in most cases there is good agreement between maximum possible occupancy of the anion tunnels
and the doping level achieved. In the case of nBu,N@BFf,
no further oxidation of {[S~(PC)O](BF,),.,~},,
is observed at
potentials as high as 3.5 V vs. the saturated calomel electrode with NaCl instead of KCl (SSCE) in liquid SO, at
- 30”C.[651At this point, irreversible oxidation of the polymer begins. It can also be seen in Figure 17 that the present
electrochemical/synthetic results include the first case in
which the charge-compensating counterion in a molecular
conductor is a dianion (S0:O).
In the case of anions having very large dimensions, the
behavior upon doping is somewhat different. For nBu,N@
pyrenesulfonate in acetonitrile (Fig. 17), the threshold po-
+
867
0.50
0.6
O.’
Y
:I
lo
*
BF?
* TOSe
*
0.1
O-O
i
i
2.60
,”’,””
.
1.50
~
1-60
v
IVI
u
~
[
-
r
0.50
r
m
n
0.00
[
s
r
r
<
-0.50
Fig. 17. Electrochemical potential spectroscopy plots for electrochemical
oxidation of tetragonal [Si(Pc)O], slurries in electrolytes composed of nR,Na
salts (R = alkyl group) of various counterions in acetonitrile. Legend identifies
the counterions. TOS = p-toluenesulfonate, PYS = 1-pyrenesulfonate. Potential is given vs. SSCE.
0.6
0.4
-I
4
n = 3~
1 !
\
0.3
Y
0.2
-
0.1
-
l
is invariably found to be inhomogeneous, with the coexistence of undoped (or lightly doped) and doped (CHX,),,
y FZ 0.06) phases. Then, depending on the counterion, transformations are observed upon further doping to different
polymer-chain/counterion packing arrangements and/or
there is a modest tunability in band filling for a given crystal
structure. However, in no case is the overall limiting degree
of charge storage per monomer subunit as great as in the
phthalocyanine polymer, nor is comparable tunability in the
band filling possible. These differences no doubt reflect the
far larger dimensions of the delocalized phthalocyanine x
system and the attendant demonstrated capacity to store
positive or negative charge, as well as the spatially closer
polyacetylene-counterion interactions. The latter is understandable in terms of the higher concentration of charge per
polyacetylene subunit and a polymer structure which is more
favorable for close n-system-counterion interactions. The
magnitudes of these interactions doubtless favor different
polyacetylene packing arrangements for different stoichiometries and/or counterions, while we have shown that
direct R-system-counterion interactions are small for structure-enforced phthalocyanine conductors. The factors mentioned above also suggest that doped polyacetylene crystal
structures will collapse upon undoping and that well-ordered
broken-in crystal structures, which are more amenable to
facile doping/undoping cycles, will not be as accessible.
Presently available polyacetylene data appear to confirm
this.[661
How important is the structure-enforced character of the
[Si(Pc)O], chains in the aforementioned electrochemistry?
This intriguing question can be addressed via solid-state electrochemical investigations of Ni(Pc), which lacks covalent
Pc-Pc structural connectivity.r601As-sublimed Ni(Pc) has a
monoclinic, slipped-stack (“p” phase) packing arrangement
as shown in Figure 19. Not surprisingly, initial electrochem-
0.0 -
I
8
. . . . . . . . . . . . . . . . ....*
2.60
1.iO
1.60
v [VI
0.40
0.60
A
-0.50
7
............
Fig. 18. Electrochemical potential spectroscopy curves for electrochemical
oxidation of tetragonal [Si(Pc)O]. slurries in nR,NeCF,(CF2)nSO? ( n = 0, 3.
7)pacetonitrile (R = alkyl group). Potential is given vs. SSCE.
tential for oxidative doping of tetragonal [Si(Pc)O], is higher
than for other anions studied and the ultimate doping level
is rather low. In the case of [ ~ B ~ , N @ ] , [ M o , O , , ]in
~ ~acetonitrile, an even higher threshold is observed (ca. + 0.9 V
vs. SSCE) and an even lower ultimate doping level, y FZ 0.06,
is achieved!651 These results demonstrate the marked structure sensitivity of the [Si(Pc)O], electrochemical doping
process.
In addition to the present [Si(Pc)O], materials, electrochemical/structural relationships have been extensively investigated for trans-polyacetylene.‘6. ],’
Considerably
more information is presently available for reductive doping
(alkali-metal insertion) than for oxidative doping. In both
cases, however, the picture is quite different than for
[Si(Pc)O], oxidative doping. Initial doping of polyacetylene
‘,,
868
~
....
~...*
@(virgin)
Tetragonal
(doped)
y(undoped)
PWC
P4/mcc
PWC
Fig. 19. Crystal structure relationships in the electrochemical oxidation of
Ni(Pc) (left) to Ni(Pc)(BF,), where y = 0 00-0.48 (middle), followed by reduction (“undoping”) (right).
ical oxidation is accompanied by a substantial overpotential,
and the product is the tetragonal doped phase (Fig. 19). At
this point, however, the behavior diverges from that of
[Si(Pc)O], in that subsequent undoping also occurs with a
large overpotential, and the product is not an undoped tetragonal phase, but rather a second (known) monoclinic Ni(Pc)
(“y” phase). It is evident that an undoped tetragonal Ni(Pc)
phase must be unstable with respect to a slipped-stack structure and that further electrochemical cycling is necessarily
accompanied by D, F? slipped-stack structural reorganizaAngew. Chem. Int. Ed. Engl. 29 (1990)857-879
tions and accompanying overpotentials. Tunability of y in
the tetragonal Ni(Pc)(BF,), phase is rather modest.[60a1
Electrochemical experiments also answer the question of
whether reductive doping of the [Si(Pc)O], framework is possible. Partial occupancy of a band constructed from the lowest unoccupied molecular orbitals (LUMOs) of [Si(Pc)O],
might also yield a metal-like substance according to theoretical calculation^!^^^^^^ Figure 20 shows an electrochemical
-1
-7
# o-rh.
-8
8
'
1
1
'
~
I
~
1
.
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
-
Yo,
-0.0
1
1
Fig. 21. Electrical conductivity (lg u) at room temperature for polycrystalline
{[Si(Pc)O]X,}. samples as a function of the degree of oxidation per Si(Pc) unit
box).
They values for {[S~(PC)O](SO,),}~
have been multiplied by 2.0 to correct
for the dinegative charge of Sote.
Xed
-0.1
I " " I " " I " " I ' " '
cycling experiment with solid tetragonal [Si(Pc)O], in
Bu,N@BF?/THF. It can be seen that the polymer can be
reversibly reduced with a limiting reduction stoichiometry of
y = 0.09, that is, to form {(nBu,N),,,,,[Si(Pc)O]),.
Furthermore, the polymer can be cycled a number of times between
the oxidized and reduced states with minor, if any, decomposition. The difference in threshold potentials measured for
tetragonal [Si(Pc)O], oxidation and reduction is ca. 1.5 V,
which compares favorably with the HOMO-LUMO band
gap estimated from optical spectroscopy: 1.65 eV
(750 nm).[47a1Initial measurements indicate that the reductively doped [Si(Pc)O], polymers are electrically conducti~e.1~~'
As noted above, the electrochemically doped ([Si(Pc)O]X,}, macromolecules not only incorporate a great diversity of off-axis anions (and a dianion), but also represent
the first class of molecular metals in which band filling is
widely tunable. Detailed charge-transport, optical, and magnetic measurements have been carried out as a function of
anion and y.(60,I' Intriguing questions include whether the
additional Coulombic perturbation of an off-axis dianion
might cause "pinning" of the carriers and whether metal-like
characteristics are tunable over the entire range of mixed
valency, as suggested by simple band-theoretical considerations (Fig. I), or whether there is a certain threshold doping
level beyond which they "switch on".
In Figure 21 are shown room-temperature conductivity
data for a series of polycrystalline ([Si(Pc)O](BF,),),, {[Si(Pc)O](tosylate),),, and {[Si(Pc)O](SO,),), samples as a
function of y . It can be seen that the a(y) behavior i s rather
insensitive to the identity of the anion and that an abrupt
increase in conductivity at low doping levels is followed by a
leveling off at y = 0.20-0.30. Beyond this point, the conducAngew. C'hem. Int. Ed. Engl. 29 (1990) 857-879
tivity is virtually independent of y as far as y = 0.67. Variable-temperature studies indicate that, for y 5 0.20, the a(T)
data can be best fit to a model appropriate for a highly
disordered material or one in which there is carrier hopping
between localized states.[,'
Fory 2 0.20, the best fit is
to the previously mentioned fluctuation-induced tunneling
model.
Because the thermoelectric power measurement is a zerocurrent experiment and because temperature drops at grain
boundaries are expected to be less important than voltage
drops, this technique should be far less susceptible to polycrystalline sample effects than dc conductivity measurements.'26-281Indeed, we find that, in contrast to conductivity measurements (cf. Figs. 9, top, and 12), variable-temperature TEP data for single crystal and polycrystalline samples
of the same molecular metal are rather similar.[ss1Figure 22
shows room-temperature TEP data for polycrystalline
([Si(Pc)O](BF,),},, { [Si(Pc)O](tosylate),}n, and {[Si(Pc)O]samples as a function of y . With increasing y , the
Seebeck coefficient (9 drops steeply from large, positive
values characteristic of p-type semiconductors to small, positive values typical of p-type molecular metals (Fig. 22) and
then levels off at y 2 0.35. Variable-temperature studies as a
307
1
200
I
1
A
0
m
0
m
0
0.00
o.io
I
.
-
,
.
I
,
I
0.20 0.30 0.40 0.50 0.60 0.70
Yo,
Fig. 22. Room-temperature thermoelectric power (S) data as a function of
doping level for polycrystalline {[Si(Pc)O](BF,),},, {[Si(Pc)O](t~syIate)~)~,
and
{[Si(Pc)O](SO,),,,,}. conductive polymer samples. The y value shown for
{[Si(Pc)Ol(SO,), ),.
has been corrected for the dinegative charge of SO:".
869
1
function of y have also been carried out. Although the data
for the above three classes of doped polymers differ in small
details, all evidence a smooth, continuous transition from
behavior characteristic of a p-type semiconductor (large,
positive S with S l / T )to a radical-cation molecular metal
(small, positive S with S T). An example is shown in Figure 23. Importantly, the transition from a semiconductor to
-
-
150
IpV K-'I
#
y=0.4
y=O.B
y=O8
y=1.0
y=1.2
y=l.4
50
-50
y=1.8
25
50
100
150
250
200
T IK1-
300
350
Fig. 23. Variable-temperature thermoelectric power (S) data for polycrystalline { [Si(Pc)O](tosylate),} samples.
a metal occurs for all three series in the region y z 0.20. How
do these experimental S(y,T) results compare with the predictions of simple tight-binding band theory? Using Equation (4) and assuming U is negligible and 4t x 0.60 eV (cf.
Table 2), the curves in Figure 24 are generated. While the
agreement between experiment and theory for Sb) at fixed
T appears reasonable (Fig. 24, top), the theory predicts continuous metallic behavior ( S T ) for all doping levels
(Fig. 24, bottom). In contrast, the experimental TEP data
evidence a "switching on" of metallic character in the neighborhood of y % 0.20. That the thermopower of the present
{[Si(Pc)OJX,), materials with y = 0.50 does not exhibit temperature-independent IS1 = 60 pV K - ' behavior stands in
marked contrast to the great majority of low-dimensional
]el = 0.50 systems having very large U/4t ratios.[26.271
Detailed reflectivity, magnetic susceptibility, and ESR
studies have also been carried out for the {[Si(Pc)O](BF,),},,
{[Si(Pc)O](tosylate),$,, and {[Si(Pc)O](SO,),}, materials as a
function of y.1701
We have shown elsewhere that single-crystal and polycrystalline optical reflectivity data for low-dimensional conductors are readily interc~nvertible.[~~',
s 3 , 541
As exemplified by the optical data for the series of BF," salts
(Fig. 25), initial doping yields reflectance spectra characteristic of a molecular semiconductor. A well-defined plasma
edge is not evident until y 0.20, beyond which doping level
the energy of the edge shifts monotonically to higher energy.
Indeed, for y values within the metallic regime, the adherence
to Equation (S), derived for tight-binding, is quite good
(Fig. 26).
-
870
Fig. 24. Top: Band-filling dependence of thermoelectric power for a simple
low-dimensional tight-binding band structure with energy-independent scattering, negligible U , T = 300 K, and 41 = 0.60 eV calculated with Eq. (4). Bottom:
Temperature dependence of thermoelectric power as a function of degree of
oxidation e calculated using Eq. (4)for a simple tight-binding band structure
with energy-independent scattering, negligible U , and 41 = 0.60 eV.
Magnetic susceptibility studies of the {[Si(Pc)O](BF,),),
materials as a function of increasing y reveal a large Curie
component typical of localized radical-cation states for low
values of y (Fig. 27). This component of the susceptibility
peaks at y z 0.20 and then falls abruptly with increasing y .
Over this same y range, a Pauli component grows in and
remains essentially constant beyond y 2 0.25. As also shown
in Figure 27, this xb)behavior is distinctly different from
that found for the chemically (inhomogeneously) doped
{[S~(PC)O](I~)~}~
Because the chemical doping
process is largely inhomogeneous, the Pauli contribution is
linearly dependent upon the doping level (because the fraction of doped phase is linearly dependent upon the doping
level). The xk)dependence of the homogeneously doped
{[Si(Pc)O](BF4),), polymers can now be compared with the
predictions of simple tight-binding band theory [Eq. (6)] in
Figure 28. As for the thennopower and reflectivity data, the
agreement with theory is poorest at low doping levels where,
in the magnetic data, the BF? salt exhibits the large, y-dependent Curie component. However, some aspects of the
predicted flattening-out in xb) are observed at higher
doping levels. Magnetic Susceptibility measurements on
{[si(Pc)ol(tosYlate)o.~7)"and ~ ~ ~ ~ ~ ~ ~ > (in
~ or
l ~ ~ ~ 4
near the metallic regime) also reveal large Pauli-like compoAngen.. Chem. Int. Ed. Engl. 29 (1990) 857-879
0
'2' 0.20 n
v
v,
\
y = 0.41
I
.-tQ
0.10
-
v,
4
o j K +
7 = 0.38
--lo
0
I
I
.
l
0.10
0
~
l
0.20
Yo,
y = 0.27
y = 0.13
4
r
l
l
l
7
0.50
0.40
0.30
-
Fig. 27. Stoichiometry 0.)dependence of the Pduli-like (m) and Curie-like (A)
spin susceptibilities of ekCtrOChemiCdlly (homogeneously) doped {[Si(Pc)O](BF,),}, polymers. The dashed line indicates the Pad-like term ( 0 ) for the
chemically (inhomogeneously with 12) doped polymer.
0
y = 0.00
1 o3
1o2
loL
105
Fig. 2 5 . Optical reflectance spectra of compressed polycrystalline
{[Si(Pc)O](BF,),), samples as a function o f y . Spectra are offset vertically for
ease of viewing.
I
0
0.50
,
,
,
I
1.50
(
,
2.00
Fig. 28. Band filling dependence of the Pauli magnetic susceptibility (1,)calculated according to the one-dimensional tight-binding model of Eq. (6) for
4r = 0.60 eV and U = 0.
5000 45004000-
35004 ,
0.50
,
,
0.60
,
,
0.70
,
,
0.80
[sin(rrp/Z]'/Z
,
,
,
-
0.90
I
1.00
Fig. 26. Plot of {[Si(Pc)O](BF,),}, and {[Si(Pc)O](tosylate),). plasma frequencies wq(from reflectance data) as a function of degree of partial oxidation e(y)
according to the tight-binding band theoretical expression of Eq. (5).
nents and, in the case of the latter material, no evidence of
carrier localization by the dinegative counterion. Magnetic
data are compiled in Table 1 and magnetically derived bandwidths in Table 2. It is found that susceptibilities are again
enhanced, but also that the degree of enhancement depends
upon Q.[~'"I This is the first experimental verification of the
important theoretical prediction[38",d1
that U can b e e (band
filling) dependent.
ESR data for the above polycrystalline ([Si(Pc)O]X,), materials as a function ofy are in substantial agreement with the
aforementioned characterization results.[70.751 Incremental
doping is accompanied by an invariance in the essentially
free-electron g value (indicative of ligand-centered carriers)
and by a general decrease in line width. The magnitudes and
Angeu.. C'hrtn. I n ! .
.
P -
5500 -
-'I
I
1.00
Ed. EngI. 29 11990) 857-879
y dependence of the line widths are in general accord with
increasingly mobile (delocalized) conduction electrons upon
doping and with a relatively unidimensional band structure
(small interstack interactions). The only surprise in these
data is the observation that the ESR line widths of
{[Si(Pc)O](SO,),), are slightly larger than those of the BF,"
and tosylate salts (ca. 1.5 G vs. 0.5-1.0 G). It is not clear at
this point whether the enhanced broadening reflects decreased carrier mobility and/or greater two-dimensionality
for Coulombic reasons, or whether greater carrier density on
the SO:@ counterion induces more efficient electronic relaxation via phonon-modulated spin-orbit interactions as in the
130 and SbFF salts discussed in Section 4.1.
4.3. Origin of the (ISi(Pc)O]X,},
Insulator/Semiconductor Metal Transition
The dramatic changes in {[Si(Pc)O]X,}, properties that
occur in the vicinity o f y z 0.20 indicate that a major change
in electronic structure and carrier mobility occurs. The diffraction and electrochemical data give no indication of
equally drastic structural changes, but rather evidence a ftairly smooth evolution in structure as a function ofy. As will be
discussed below, the nature of they N 0.20 transition is from
a localized-carrier, insulating or semiconducting material to
871
,
a delocalized. metal-like material, the properties of which
continue to evolve in an instructive manner as y approaches
the maximum doping stoichiometry. The electronic structural model that best fits these >!-dependentchanges focuses on
the effects of disorder and defects on the properties of a
quasi-one-dimensional band structure. To d o so. we make
use of another band-structural concept which indicates the
number of energy leve!s per unit energy in the band: the
density of states.[’] In a classical description, many of the
properties of metal-like substances (e.g.. conductivity, magnetic susceptibility) are extremely sensitive to the density of
states near the highest occupied band level (the Fermi level).
Figure 29A illustrates a density of states versus energy diagram for a classical. one-dimensional tight-binding band.“]
This is the idealized model from which the thermoelectric
power, optical spectroscopic. and magnetic models of Equations (5) (Fig. 24), (6). and (7) (Fig. 28j, respectively. are
derived. The effects on this picture of disorder and defects
will be to alter the density of states as shown schematically
in Figure 29 B and to introduce a mobility edge (E,) between
b
c - It-----*
Ec
Er
Ec
Energy
Fig. 29. Schematic density of states model of the electronic structural consequences of incrementally oxidizing [Si(Pc)O].. A : Undoped. Classical quasione-dimensional tight-binding description. B: Part A in the presence ofdisorder
and defects. C: Lightly doped. D: Heavily doped. EF represents the Fermi
energy and Ec Ihe energy or a mobility edge.
localized states at the tails of the band and the more central.
delocalized (metal-like) states.[76.”I A transition of the Fermi level (E,) across such an edge as a function of composition or some other property is classically known as an Anderson transition.[76’In the commonly accepted description,
the localized Anderson state possesses a “Fermi glass” electronic structure having a relatively continuous density of
localized states.”6. ’I Whether such a situation has been
truly realized in the present materials is currently under further investigation.
At low oxidation levels of {[Si(Pc)O]X,),, it is reasonable
that the doped polymer is significantly disordered. pre872
sumably as a consequence of varying [Si(Pc)O], chain lengths
and/or crystallographic positions as well as possibly other
types of defects (including in anion locations). The effect of
these perturbations will be localization of conduction band
states at the tails of the band. In this regime, the carriers
(holes) will therefore have low mobilities. The charge-transport data are in good accord with this model: the dc conductivity is low. exhibits high apparent activation energes, and
can be lit to a transport model applicable to disordered systems or carrier hopping between localized states. At this
same doping level, the magnitude and temperature dependence of the thermoelectric power is also typical of a p-type
semiconductor (Figs. 22 and 23). Additional evidence for
carrier localization in this regime is noted in the optical reflectivity, which is in accord with a semiconducting or insulating electronic structure (Fig. 25), and the magnetic susceptibility. which evidences a large concentration of localized
Curie-like spins and little, if any. Pauli-like behavior
(Fig. 27).
In the region of y = 0.20, major changes are detected.
Both the magnitude and temperature dependence of the thermoelectric power evidence a transition to a metal-like substance (Figs. 22 and 23). Although interparticle contact effects are likely to obscure the sharpness of the transition, the
electrical conductivity begins to level off in the vicinity of
y =: 0.20, while the apparent activation energies (slopes in lg
(i vs. 1 / T )
For y 2 0.20, the temperature dependence
of {[Si(Pc)O]X,), conductivity can be better fit to a fluctuation-induced tunneling-transport model appropriate for a
more metal-like material. Also at y = 0.20, a metal-like plasma edge is first observed in the optical reflectivity, while
magnetic measurements indicate that the concentration of
Curie-like spins drops precipitously and the concentration of
Pauli-like spins rises sharply. ESR line widths exhibit a decrease to values characteristic of more mobile spins. In terms
of the model shown in Figure 29. they = 0.20 doping region
represents the region at which the Fermi level is crossing the
mobility edge. This picture stands in contrast to the classical
model (Fig. 29). which would predict a “turn-on’’ of metallic
characteristics at the lowest doping levels. To our knowledge,
such a doping level-dependent turn-on has not previously
been observed in a molecular metal.
Depleting the band structure beyond y z 0.25 evokes a
number of interesting responses. The room-temperature
electrical conductivity remains remarkably constant over a
very broad range. while the decrease in apparent activation
energy is smaller than in the y r 0.20 region. Beyond y =
0.35, the Pauli-like component of the magnetic susceptibility
remains nearly constant while the Curie-like component
declines to a low. constant level. Meanwhile. the ESR
line widths fall to levels typical of unidimensional phthalocyanine molecular metals not having heavy-atom counterions. For y increasing beyond ca. 0.25. the optical plasma
edge shifts continuously to higher energy. in good accord
with simple tight-binding band theory (Fig. 26). These observations generally agree with the picture of Figure 2911 in
which the Fermi level is in the metallic portion of the conduction band and in which the density of states varies only
modestly as a function ofy. The latter contention is supported by the weak sensitivity of the conductivity and Pauli-like
susceptibility to y.As noted elsewhere. agreement between
Angen. Chem. Inr. Ed. DigI. 29 ( 1 9 9 0 ) 857-879
experimental parameters, such as bandwidths derived from a
tight-binding analysis of optical reflectivity and thermopower data and those measured by other methods, is always poor
at low doping levels, but more satisfactory at higher doping
level^.[^'^^'^ Comparison of Figure 29A to Figures 29 B26D shows that one reason derives mainly from how the
density of states responds to disorder and defects. The deviation of the density of states from the idealized formalism
(Fig. 29A), upon which Equations (5)-(7) are based, is
greatest near the tails of the band, that is, at very low and
very high doping levels.
8
8
8
10-7
in4
1
L
0. 00
4.4. Tunable Approach to a Mott-Hubbard Insulator:
Quinone Doping of [Si(Pc)O],
0. 20
0. 60
0.40
0. 80
1. 00
YFig. 30. Room-temperature electrical conductivity (uRT)data of compressed
polycrystalline {[Si(Pc)O](DDQ),}, samples as a function of y.
The maximum ([S~(PC)O]X,}~
doping levels achieved via
electrochemical techniques are on the order of + 0.70, corresponding to a 65 %-filled classical band. With reference to
Figure 4 and earlier discussion, the fascinating question next
arises as to whether it is possible to incrementally tune y to
even higher levels, perhaps to the half-filled band (integral
oxidation state) level. Will the products be metals, semiconductors, or insulators? Doping of [Si(Pc)O], with the strongly oxidizing reagents 2,3-dichloro-5,6-dicyano-p-benzoquinone (DDQ) or 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (TCNQF,) proceeds as shown in Equation
(9).[42h,46a,
h, ”I Heating the doped polymers under vacuum
.
0. 0
0. 2
0. 4
0. 6
0. 8
I. 0
Y-
DDQ
TCNQ-F,
removes the quinones and regenerates undoped [Si(Pc)O], .
IR, Raman, and optical spectral data indicate that most, if
not all, of the quinone molecules are present as radical anions (Q’O). Hence, oxidation of the phthalocyanine subunits
has occurred. X-ray diffraction data indicate that the
quinone doping process is at least partially inhomogeneous
and that the doped polymer crystal structures are significantly disordered/poorly crystalline.
The electrical and optical properties of the ([Si(Pc)O]Q,},
materials are markedly different from those of quinonedoped molecular phthalocyanines. While the latter materials
were found to be insulators, in agreement with the proposed
integrated stacking crystal structures (3, 4), the quinonedoped stacking-enforced polymers exhibit electrical and optical properties reminiscent of the halogen- and nitrosyldoped [Si(Pc)O], materials discussed in Section 4.1 . [ 5 3 , 541
Thus, a steep rise in conductivity is observed upon initial
oxidation (Fig. 30) along with a drop in the thermoelectric
power from that characteristic of a semiconductor to that of
a x-radical-cation-based [Si(Pc)eo)O], molecular metal
(Fig. 31). The optical reflectivity at y N 0.40 exhibits a characteristic molecular-metal plasma edge virtually identical to
Angew Chrm. Int. Ed. Engt. 29 (1990) 857-879
Fig. 31. Room-temperature thermoelectric power (5‘) data for compressed
polycrystalline ([Si(Pc)O](DDQ),}, samples as a function of y.
that of the corresponding IF, BFF, and PFF salts
(Fig. 14).[53*54, 781 As the {[Si(Pc)O]DDQ,), doping level is
increased beyond y N 0.20, the room-temperature conductivity and thermoelectric power remain nearly constant, as
observed for other {[Si(Pc)O]X,}, materials (Figs. 21 and
22). In addition, the o(T) data can be fit to the aforementioned fluctuation-induced carrier tunneling model, while
the S ( T ) data are reminiscent of the XG = BFF and tosylate
analogues (Fig. 23).
As y is increased beyond ca. 0.50-0.60, dramatic changes
occur in the {[Si(Pc)O]DDQ,), physical properties. The electrical conductivity begins to fall and, by y z 0.90, the material has become an insulator with a ca. lo7 decrease in electrical conductivity versus that at y N 0.40 (Fig. 30). Indeed, the
conductivity at y N 0.90 is lower than for y N 0.00. In this
high-doping-level regime, a plasma edge is no longer discernible in the {[Si(Pc)O]DDQ,}, optical reflectivity, and the
strong carrier absorption commonly observed in the transmission infrared spectra of phthalocyanine molecular
metalsL531
is no longer apparent. As y increases beyond
ca. 0.50, the room-temperature thermoelectric power of
873
{[Si(Pc)O]DDQ,}, rises and becomes semiconducting by
y z 0.90 (Fig. 31). Interpretation of magnetic susceptibility
data is presently complicated by unknown and potentially
varying contributions of the paramagnetic QaQcounterions.
It is evident from the above discussion that {[Si(Pc)O]DDQ,), undergoes a dramatic metal to insulator transition
as y is tuned toward ca.
1.O (a half-filled classical band).
As noted above, the admittedly simplistic model of Figure 4
implies that, all other factors being equal, electron4ectron
Coulombic interactions will adversely affect carrier mobility
most at the integral oxidation levels of e = 0 and 1. The
latter situation would correspond to a molecular/macromolecular Mott-Hubbard insulator, and the result is effectively a gap at the center of the highest filled band. Several
isolated examples are known where such a regime is believed
to obtain;I2*38a, d *791 however, data interpretation is complicated by the crystal-structure and building-block alterations
needed to produce the e = 1 material. The present [Si(Pc)O]DDQ,}, system is the first example where the MottHubbard insulator state can be incrementally tuned into a
low-dimensional material. That the onset of insulating character is somewhat in advance of the @ = I .O oxidation level
supports the not unexpected presence of localized states at
the low-energy tail of the conduction band (cf. Fig. 29). If the
admittedly simplistic model of Figure 4 is correct, oxidation
beyond y z 1 .O should give rise to a second regime of metallike characteristics. Preliminary results with the more strongly oxidizing dopant TCNQF, support this
Thus,
while the sharpness of the effect is likely convoluted to some
degree with artifacts arising from doping inhomogeneity
and disorder, it is evident that an oxidative increase of y in
{[S~(PC)O](TCNQF,),}~
materials leads (Fig. 32) initially to a
+
-
'
-
.
11
1 0 . ~-
-0
10-7tl,, , ,
0
,
05
,
, , ,
I
10
Y-
,
,
, , I
15
,
,
,
,
I
20
Room-temperature conductivity data for compressed polycrystalline
{[~(Pc)O](TCNQF~),}.
samples as a function of I:
terms of Figure 4, charge mobility is again increased by the
creation of additional site vacancies.
5. Electronic Structure Considerations:
Connections between the Microscopic
and the Macroscopic
The dicussion to this point has focused upon how deliberate, chemically induced changes in metallomacrocycle stacking architecture, counterion environment, and band filling
influence band structure, hence macroscopic electrical, optical, and magnetic properties. We now consider these observations in light of what is known about the composition and
nature of the band structure. We also inquire as to the validity of the tight-binding formalism ['] employed here and elsewhere to extract band structure information from macroscopic, condensed-phase thermopower and optical reflectivity measurements. In particular, we ask whether small-molecule theoretical and experimental means exist to calibrate
this formalism.
As depicted in Figure 3, the [Si(Pc)O], band structure is
built up by suitable combinations of Si(Pc) molecular orbitals. The highest-energy filled band, the depletion of which
gives rise to metallic character, will be composed of Si(Pc)
HOMOS, while the lowest-energy unfilled band, which is
populated upon reductive doping, is composed of Si(Pc)
LUMOs. Electronic structure calculations have been
carried out at the first-principles discrete-variational local
exchange (DV-Xa) level [*'I
on the molecules Si(Pc)(OH), (12; M = Si, R = H) and the cofacial dimer
HOSi(Pc)OSi(Pc)OH (13; M = Si, R = H).f51.681
Species 13
is a prototype for known molecules of the type
Me,(tBu)SiOSi(Pc)OSi(Pc)OSi(tBu)Me,, which have been
characterized by single-crystal diffractometry E 5 'I (structural
results are in good agreement with those deduced for [Si(Pc)O],[~'"~).It will be seen that the availability of cofacial
dimers that model the simplest aspect of the macrocycle r-r
band-forming interaction is a particularly important dividend of the [Si(Pc)O], architectural connectivity
+
OR
6
OR
12
steep rise in conductivity followed by a leveling off by
y z 0.20 as seen for DDQ (Fig. 30). At this point, a plasma
edge is evident in the reflectance spectrum. As in the case of
{ [Si(Pc)O]DDQ,),, the conductivity begins to decline by
y z 0.50 and reaches a minimum near y z 1 .O. However, it
is possible to dope {[Si(Pc)0](TCNQF,),Jn beyond the integral oxidation state, and this chemistry effects a return in
conductivity to the levels observed in the metal-like,
0.20 5 0.70 stoichiometry region! This remarkable result
I .O state involves
suggests that oxidation beyond the y
electronic depletion of the lower Mott-Hubbard band. In
874
13
The DV-Xm calculations on 12 and 13 give results in excellent agreement with experimental optical and orbital ionization energy data.["] In Figure 33 are shown graphic representations of the HOMO and LUMO of Si(Pc)(OH),. The
most significant conclusion about the HOMO is that the
greatest orbital amplitudes are near the core of the macrocycle with a lesser contribution from the fused benzo ring
orbitals and almost negligible contribution (forbidden by
symmetry) from the orbitals of the (Si-0)" chain. These obAngen. Chem. In!. Ed. Engl. 29 (1990) 857-879
Xa formalism is known to be accurate in situations where
relatively small z-n orbital overlap obtains.'51.
From
the upper part of Figure 34, it can be seen that the bandwidth
of an infinite [Si(Pc)O], stack is sensitive to the ring-ring
eclipsing angle (with fixed interplanar spacing). It can also be
seen that the experimentally observed angles on the order of
40 only approach one of two possible overlap maxima at 0
and 45 '. The former may be disfavored due to nonbonded
repulsions. These calculations also reveal a zero bandwidth
minimum near 20". Clearly, all synthetic strategies for a
phthalocyanine molecular metal must avoid approaching
this eclipsing angle. The lower part of Figure 34 shows the
calculated effect on the bandwidth of varying the Pc-Pc interplanar spacing (with a fixed eclipsing angle). It can be seen
that the calculated 4t value falls off rapidly as the stacking
repeat distance increases. In Table 2, these results are compared with experimental thermopower- and reflectivityderived tight-binding bandwidth estimates at the three experimentally accessible interplanar spacings of 3.25, 3.30,
and 3.48 A. It can be seen that the agreement between the
condensed-phase measurements, interpreted within the
framework of a Hiickel-like formalism, and the DV-Xa theoretical calculations is quite good. However, an independent
experimental calibration of the solid-state measurement
would be highly desirable.
As can be appreciated from Figure 3, the electronic structure of dimers such as 13 should be directly extrapolable to
the bandwidth of the corresponding polymer. That is, the
HOMO-HOMO splitting of 2t should be one-half the tightbinding bandwidth. Molecules of structure 13 provide a
''9
Y
O
X
-
X
Fig. 33 Graphic representations of DV-Xa-derived molecular orbitals of
Si(Pc)(OH), viewed perpendicular to the macrocychc plane. Left: 7e,(LUMO).
Right: 2a,,(HOMO).
servations help to understand the largely ring-centered, zradical-cation character of the conduction process. Furthermore, the spatial remoteness of the band-structure-forming
orbitals from off-axis counterions (cf. also Figs. 6 and 8)
helps to explain the observed insensitivity of the collective
properties of ([Si(Pc)O]X,), to the identity of X. The DV-Xa
results for the Si(Pc)(OH), LUMO (Fig. 33) indicate that the
band structure populated upon reductive doping will be of
significantly different orbital composition. For example, it
will have a non-negligible contribution from Si orbitals.
Calculations on dimeric structure 13 and related dimers
have included an examination of how X--K interactions as
expressed by 4t (Fig. 3) vary with interplanar eclipsing/twist
angle and ring-ring interplanar spacing (Fig. 34). The DV-
O
4t
lev1
OR
I
+
:
:
:
6
4
:
:
!
:
:
:
*
12
10
8
:
:
!
:
16
14
i
l
18
20eV
a)
d
IAl
-
+
Fig. 34. Top: Calculated variation of the HOSi(Pc)OSi(Pc)OH conduction
(HOMO) bandwidth as a function of Pc-Pc eclipsing (twist) angle at an interplanar separation of 3.32 8,(0" = completely eclipsed, D., Si(Pc)oSi(Pc) geometry). Bottom: Pc-Pc interplanar separation with an eclipsing angle of 37
O.
Angen. Chem. Int. Ed. Engl. 29 (1990) 857-879
4
l
:
6
:
:
8
!
:
:
10
:
:
12
:
:
14
l
:
16
I
I
l
18
20.V
Fig. 35. Gas-phase He(1) photoelectron spectra of a) PcGe(OSiMe,tBu),, b)
(tBu)Me,SiOGe(Pc)OGe(Pc)OSIMe,(tBu), and c) expanded-scale presentation
of the aa' doublet in b.
875
unique ex situ measure of 4t since the HOMO-HOMO splitting energy can be measured in the gas phase by photoelectron spectroscopy.[”. mdl As exemplified by Figure 35, the
HOMO-HOMO interaction energy in the gaseous M = Ge
compound is clearly visible in the splitting of the lowest-energy Pc n ionization (aa’ band). The measured splitting in the
M = Ge dimer is 0.19(3) eV, while that in the M = Si dimer
is 0.29(3) eV. Assuming molecular geometies are similar and
monomer/dimer relaxation effects upon ionization are approximately constant, multiplication of these splittings by a
factor of two should provide an independent measure of the
tight-binding bandwidth. In Table 2 are compared 4t values
measured in the solid state (TEP, reflectivity, magnetic),
measured in the gas phase (PES), and calculated using the
DV-Xu formalism. With the exception of the bandwidths
derived from magnetic measurements, which are susceptible
to Coulomb enhancement (see above), the agreement is quite
good. Clearly, 4t values derived from properly analyzed condensed-phase measurements are reliable in these kinds of
materials. Moreover, the era is fast approaching when firstprinciples quantum chemistry will play a major predictive
role in guiding the design and synthesis of new electronic
materials.
6. Comparison of [M(Pc)O]X,), Materials to Other
Molecular and Polymeric Conductors
The initial goal at the outset of this research was to devise
a system of molecular conductors in which a number of key
parameters could be perturbed while retaining control of the
stacking architecture. No other system of molecular metals
has allowed wide uniaxial variation of interplanar spacing or
wide variation of off-axis counterions without major
changes in the stacking architecture. In regard to broad
excursions in the partial oxidation level, only
(NMP),(Phen), -,TCNQ[631 merits comparison and contrasting. The band filling in this low-dimensional molecular
conductor can be tuned from e = - 0.50 to - 0.66 in the
TCNQ stack by substitution of neutral Phen for NMPO.
Electron-phonon coupling is very large, giving rise to a
Peierls distortion in the stacking architecture and gap-related semiconducting charge-transport behavior even at room
temperature. As x is tuned from 1.0 to 0.5, the material
changes (at x = 0.65) from a two-chain conductor with a
relatively small U (due to interchain screening) to a large-U
single-chain (TCNQ) conductor. In the narrow region where
x N 0.50-0.56, excess charge is accommodated by maintaining a commensurate (dimerized) TCNQ stacking architecture and forming pairs of charged defects c bipolar on^).[^^]
This behavior is to be contrasted with the present phthalocyanine conductors, where the molecular M(Pc)X, systems
exhibit metallic characteristics over a broad temperature
range, where there is minimal to no evidence for a Peierls
instability down to low temperatures, and where the latticestiffening of the {[Si(Pc)O]X,], backbone and the lower
bandwidth are likely to further oppose electron-phonon
coupling-based structural distortions. In addition to these
contrasts, off-axis counterion screening effects appear to be
far larger in (NMP),(Phen), -x(TCNQ).[631
876
As macromolecules, the large family of { [M(Pc)O]X,},
conductors differs from the great majority of electrically
conductive polymers in that the conjugation pathway does
not coincide with the putative charge transport pathway and
electron-phonon coupling is much smaller. Thus, initial
doping of Peierls-distorted trans-polyacetylene results in a
decrease in the magnetic susceptibility and semiconducting
transport via spinless charged soliton (nfree radical) defect
states in the band gap.[4.661At the doping level of ca.
[(CH)X,,,], , there is an abrupt transition to a metallic state
with Pauli-like magnetic susceptibility. At this point, the material has been described as having a disorder-induced metallic density of states in the band gap,’’’] or as a polaron
Polyacetylene can be prepared in highly crystalline
microstructures, considerable crystal-structure-doping information is accumulating, and off-axis counterion effects
are presently being exp10red.I~.6 6 . 671
There is far less structural information on polypyrrole,
polyaniline, and polyphenylenevinylene (Fig. 5). The former
two materials have nondegenerate ground states and very
strong electron-phonon coupling. Incremental doping of
polypyrrole gives rise initially to polarons (radical-cation
sites bound strongly to lattice distortions), then predominantly to spinless bipolarons in the band gap.[841Thus, the
magnetic susceptibility initially increases, goes through a
broad maximum, and then drops to very low levels as partial
oxidation progresses (0.0 to 0.33 holes per monomer unit).
The magnitude and temperature dependence of the magnetic
susceptibility is never Pauli-like. The conductivity peaks near
the maximum in the susceptibility and then remains approximately constant. The temperature dependence of the conductivity at all doping levels is suggestive of carrier hopping
between localized states. To our knowledge, the homogeneity and crystal-structural aspects of polypyrrole doping remain obscure. Doping (via protonation) of polyaniline[s51
appears to be largely inhomogeneous using the type of xb)
arguments presented in Section 4.2 for {[Si(Pc)O]I,},
(Fig. 27). The fully doped material has been described both
as a granular polaronic metal above the percolation
and as a nonmetallic “Fermi glass”.[861A number of aspects of the undoped and doped polyaniline macromolecular structures are presently unclear.
Polyphenylenevinylene and its 1,4-dimethoxy derivative
undergo chemical and eIectrochemica1 oxidation to yield
conductive polymers.[87] Although doping/structural relationships are not yet fully resolved, it appears that two doped
phases are present in the electrochemistry of the dimethoxy
derivative. For this material, y can be brought as high as
+ 0.32. Charge transport has been discussed in terms of
polaron and bipolaron states in the band gap with a temperature dependence characteristic of hopping between relatively localized states.[”’
7. Conclusions
It is hoped that this review has conveyed some sense of the
rich and illuminating cornucopia of solid-state chemical and
physical phenomena exhibited by the ([M(Pc)O]X,), family
of electrically conductive, multimolecular assemblies. Such
substances clearly stand at the frontier between molecular
Angerv. Chem. 1m. Ed. Engl. 29 (1990) 857-879
conductors and conductive polymers. As such, they shed
new light upon each area and contribute to our better understanding of both classes of materials.
It is a sincere pleasure to acknowledge the outstanding contributions and stimulating collaboration of a host of talented
colleagues. Their names can be found in the references. The
author thanks the National Science Foundation through the
Northwestern Materials Research Center (Grants D M R
8520280 and D M R 8821571) and the Office of Naval Research f o r generous jkancial support of this research. He
thanks the Technical University of Berlin and the Free University of Berlin ,for a Graduiertenkolleg Lectureship, during
which much of this account was written.
Received: February 12, 1990 [A 774 IE]
German version: Angew. Chem 102 (1990) 886
[l] For excellent introductions to relevant condensed-matter science and
band-theoretical concepts, see: a) R. Hoffmann, Angew. Chem. 99 (1987)
871, Angew. Chem. Int. Ed. Engl. 26 (1987) 846, and references cited
therein; Solids and Surfaces. A Chemist’s View of Bonding in Extended
Structures, VCH, Weinheim/New York 1988; b) M.-H. Whangbo in J.
Rouxel (Ed.): Crystal Chemistry and Properties of Materials with QuasiOne-Dimensional Structures, Reidel, Dordrecht 1986, p. 27, and references
cited therein; c) J. S. Miller, A. J. Epstein, Prog. Inorg. Chem. 20 (1976) 1;
d) H. M. Rosenberg: The Solid State, 3rd ed., Clarendon Press, Oxford
1986; e) C. Kittel: Introduction to SolidState Physics, 6th ed., Wiley, New
York 1986.
121 a) Proc. Int. Con5 Sci. Technol. Synth. Metals, Synth. Met. 27-39(1989);
b) D. Jerome, L. G. Caron (Eds.): Low Dimensional Conductors andsuperconductors, Plenum, New York 1987; c) J. R. Ferraro, J. M. Williams:
Introduction to Syntheric Electrical Conductors, Academic Press, New
York 1987; d) J. Y. Becker, J. Bernstein, S . Bittner (Eds.): Isr. J. Chem. 27
(1986) No. 4, e) D. 0 . Cowan, F. M. Wiygul, Chem. Eng. News 64 (1986)
No. 29, p. 28; f ) M. L. Cohen, Science 234 (1986)549; g) R. L. Greene,
G. B. Street, ibid. 226 (1984) 651; h) J. S . Miller (Ed.): Extended Linear
Chain Compounds, Vol. 1-3, Plenum, New York 1982.
[3] a) H . Inokuchi, Angew. Chem. 100 (1988) 1817; Angew. Chem. Int. Ed.
Engl. 27(1988) 1747, and references cited therein; b) J. M. Williams, H. H.
Wang, T. J. Ernge, U. Geiser, M. A. Bens, P. C. W. Leung, K. D. Carlson,
R. J. Thorn, A. J. Schultz, M:H. Whangbo, Prog. Inorg. Chem. 35 (1987)
51; c) D. Schweitzer, E. Gogu, 1. Hennig, T. Klutz, H. J. Keller, Ber.
Bunsenges. Phys. Chem. 91 (1987) 890; d) H. Mayama, H. Yamochi, G.
Saito, K. Nozdua, T. Sugano, M. Kinoshita, S . Sato, K. Oshima, A.
Kawamoto, J. Tanaka, Chem. Lett. 55 (1985); e) J. M. Williams, Prog.
InorR. Chem. 33 (1985) 183; f) F. Wudl, Ace. Chem. Res. 17 (1984) 227.
[4] a) J. Heinze, Top. Curr. Chem. 152 (1990) 2; b) N. C. Billingham, P. D.
Calvert, Adv. Polym. Scr. 90 (1989) 3; c) S . Roth, H. Bleier, Adv. Phys. 36
(1987) 385; d) T. A. Skotheim (Ed.) Handbook of Conducting Polymers,
Vol. 1.2, Marcel Dekker, New York 1986; e) J. E. Frommer, R. R. Chance:
Encyclopedia of Polymer Science and Engineering 5 (1986) 462; f) K J.
Wynne, G. B. Street, Ind. Eng. Chem. Prod. Res. Dev. 21 (1982) 23; g)
R. H. Baughman. J. L. Bredas, R. R. Chance, R. L. Elsenbaumer, L. W
Shacklette, Chem. Rev. 82 (1982) 209; h) G. Wegner, Angew. Chem. 93
(1981) 352; Angew. Chem. Int. Ed. Engl. 20 (1981) 361.
[5] A. G. MacDiarmid, R. B. Kaner in [4d], Vol. 1, p. 689.
[6] a ) J. J. Miasik, A. Hooper, B. C. Tofield, J. Chem. Soc. Faraday Trans. I 8 2
(1985) 1117; b) F. Gamier, Angew. Chem. Adv. Muter. 101 (1989) 529;
Angew. Chem. Int. Ed. Engl. Adv. Muter. 28 (1989) 513; Adv. Muter. 1
(1989) 117.
[7] J. Kanicki in [4d], Vol. I. p. 543.
[8] M. J. Natan, M. S . Wrighton, Prog. Inorg. Chem. 37 (1989) 391.
(91 R. Naufi. J. Electrochem. SOC.130 (1983) 2126.
[lo] J. R. Ellis in [4d], Vol. 1, p. 489.
[ l l ] a) M. Sinclair, D. McBranch, D. Moses, A. J. Heeger in [2a], D 645; b) W.
Wu, S. Kivelson in [2a], D 575.
1121 S. Miydta, 3rd Symp. EIectroresponsive Molec. Polym. Systems, Brookhaven National Laboratory, Upton, NY 1989.
[13] M. Gazard in [4d], Vol. 1, p. 673.
[14] K. Kaneto, H. Agawa, K. Yoshino, J. Appl. Phys. 61 (1987) 1197.
[15] a) R. Moret, J. P. Pouget in [l b], p. 87; b) D. Jerome, H. J. Schulz, Adv.
Phys. 31 (1982) 299.
1161 S. Mazumdar, S. Ramasesha in [2a], p. A 105, and references cited therein.
[I71 a) P. D. Hale, M. A. Ratner, J Chem. Phys. 83 (1985) 5277, b) E. M.
Conwell. Ph.w Rev. 5. 22 (1980) 1761; c ) M. Weger, M. Kaveh, H. Gut
Angew. Chem. Int. Ed. Engl 29 (1990) 857-879
freund, SolidState Commun. 37 (1981) 421 ;d) H. Gutfreund, C. Hartstein,
M. Weger, ibid. 36 (1980) 647.
[18] T. J. Marks, Science 227 (1985) 881; b) T. J. Marks, D. W. Kalina in [2h],
Vol. 1, p. 197.
1191 a) A. Gleizes, T. J. Marks, J. A. Ibers, J. Am. Chem. Soc. 97 (1975) 3545;
M. A. Cowie, A . Gleizes, G. W. Grynkewich, D. W. Kalina, M. S. McClure, R. P. Scaringe, R. C. Teitelbaum, S. L. Ruby, J. A. Ibers, C. R.
Kannewurf, T. J. Marks, ibid. 101 (1979) 2921, c) L. D. Brown, D. W
Kalina, M. S. McClure, S. L. Ruby, S . Schultz, J. A. Ibers, C. R. Kannewurf, T. J. Marks, ibid. 101 (1979) 2937; d) D. W. Kalina, J. W Lyding.
M. T. Ratajack, C. R. Kannewurf, T. J. Marks, ibid. 102 (1980) 7854.
I201 a) L:S. Lin, M. S . McClure, J. W. Lyding, M. T. Ratajack, T- C Wang,
C. R. Kannewurf, T. J. Marks, J Chem. Soc. Chem. Commun. 1980,954;
b) L:S. Lin, M. S . McClure, J. W. Lyding, M. T. Ratajack, T.-C. Wang,
C. R. Kannewurf, T. J. Marks, unpublished.
[21] C. W. Dirk, T. J. Marks, Inorg. Chem. 23 (1984) 4325.
1221 a) T. J. Marks, J. Coat. Techno!. 48 (1976) 53; h) J. L. Petersen, C. S .
Schramm, D. R. Stojakovic, B. M. Hoffman, T. J. Marks, J. Am. Chem.
Soc. 99 (1977) 286; c) C. S . Schramm, R. P. Scaringe, D. R. Stojakovic.
B. M. Hoffman, J. A. Ibers, T. J. Marks, ibid. 102 (1980) 6702; d) D. R.
Stojakovic, Ph. D. Thesis, Northwestern University, Evanston. IL, USA
1978.
I
Chem. Soc. Chem.
[23] a) T. J. Marks, D. F. Webster, S . L. Ruby, S . Schultz, .
Commun. 1976, 444; b) T. J. Marks, Ann. N . Y. Acad. Sci. 313 (1978) 594.
[24] F. H. Moser, A. L. Thomas: The Phthalocyanines, Vol. I , 11, CRC Press,
Boca Raton, FL, USA 1983.
[25] T. Inabe, T. J. Marks, R. L. Burton, J. W. Lyding, W. J. MacCarthy, C. R.
Kannewurf, G. M. Reisner, F. H. Herbstein, Solid State Commun. 54
(1985) 501.
[26] a) J. F. Kwak, G. Beni, P. M. Chaikin, Phys. Rev. 513 (1976) 641; b) J. R.
Cooper, B. Alavi, L.-W. Zhou, W. P. Beyermann, G. Griiner, ibid. 35
(1987) 8794.
1271 a) K. Mortensen, E. M. Conwell, J. M. Fabre, Phys. Rev. 5 28 (1983) 5856,
and references cited therein; b) E. M. Conwell, ibid. 18 (1978) 1818.
[28] a) D. Schweitzer, I. Hennig, K. Bender, H. Endres, H. J. Keller, Mol. Cryst.
Liq. Cryst. 120 (1985) 213; b) K. Mortensen, C. S. Jacobsen. K. Bechgaard, K. Carneiro, J. M. Williams, ibld. 119 (1985) 401; c) A. Maaroufi.
C. Coulon, S. Flandrois, P. Delhaes, K. Mortensen, K. Bechgaard, Solid
State Commun. 48 (1983) 555, and references cited therein.
1291 This expression assumes a small value of U and pragmatically neglects the
energy dependence of the carrier scattering time: a) D. Ihle, T. Eifrig, Phys.
Status Solid. 5 91 (1979) 135; b) P. M. Chaikin, G. Beni, Phys. Rev. 5 13
(1976) 647.
1301 a ) P. J. Toscano, T. J. Marks, J. Am. Chem. Soc. 108 (1986) 437; b) S. M.
Tetrick, T. J. Marks, unpublished.
[31] a) L. Zuppiroli, in [2b], p. 307, and references cited therein; b) G. G.
Maresch, M. Mehring, J. U. von Schiitz, H. C. Wolf, Chem. Phys. 85
(1984) 333, and references cited therein.
[32] a) K. Carneiro, J. C. Scott, E. M. Engler, Solid State Commun. 50 (1984)
477, and referencescited therein; b) Z. Soos, S. R. Bondeson in 12 h], Vol. 3,
p. 229; c) D. Jerome in L. Alcacer (Ed.): The Physics and Chemisrrv of
Low-Dimensional Solids, D. Reidel, Dordrecht 1980, p. 123.
[33] a) C. S . Jacobsen, in [2b], p. 253; b) D. B. Tanner, in [2h], Vol. 2, p. 205;
c) R. L. Burton, W. I McCarthy, C. R. Kannewurf, T. Inabe, T. J. Marks,
N 5 S Spec. Publ. U S . 697 (1985) 54.
[34] a) M. R. Madison, L. B. Coleman, R. B. Somoano, Solid State Commun.
40 (1981) 979; b) B. A. Weinstein, M. L. Slade, A. J. Epstein, J. S . Miller.
ibid. 37 (1981) 643, and references cited therein; c) J. B. Torrance, B. A.
Scott, B. Welber, F.B. Kaufman, P. E. Seiden, Phys. Rev. 5.19(1979)730:
d) P. Delhaes, C. Coulon, J. Amiell, S . Flandrois, E. Toreiiles, J. M. Fabre,
L. Giral, Mol. Cryst. Liq. Cryst. 50 (1979) 43.
[35] M. Takahashi, T. Sugano, M. Kinoshita, BUN.Chem. SOC.Jpn. 57 (1984)
26; b) P. Delhaes, C. Coulon, S. Flandrois, B. Hilti, C. W. Mayer, G. Rihs,
J. Rivory, J. Chem. Phys. 73 (1980) 1452; c) L. C. Isett, Phys. Rev. 5. 18
(1978) 439; d) J. C. Scott, A. F. Garito, A. J. Heeger, ibid. 10 (1974) 3131.
[36] a) Z. G. Soos, S . R. Bondeson, Mol. Cryst. Liq. Cryst. 85 (1982) 19; b)
L. C. Tippie, W. G. Clark, Phys. Rev. 5 . 2 3 (1981) 5846; c) L. J. Azevedo.
W G. Clark, ibid. 16 (1977) 3252.
[37] a) H. Shiba, Phys. Rev. 5 6 (1972) 930; b) J. B. Torrance, Y Tomkiewicz.
B. D. Silverman, ibid. 15 (1977) 4738.
1381 a) S. Mazumdar, S . N. Dixit, Phys. Rev. 5 34 (1986) 3683, and references
cited therein; b) R. Laversanne, J. Amiell, C. Coulon, C. Garrigou-Lagrange, P. Delhaes, Mol. Cryst. Lry. Cryst. 119 (1985) 317; c) C. S. Jacobsen, K. Bechgaard, ibid. 120 (1985) 71; d) s. Maxumdar, A. N. Bloch,
Phys. Rev. Lett. SO (1983) 207; e) C. Coulon. P. Delhaes, S . Flandrois, R.
Lagnier, E. Bonjour, J. M. Fabre, J. Phys. (Paris) 43 (1982) 1059.
[39] Electron-phonon coupling can also lead to susceptibility enhancement: H
Gutfreund, 0 . Entin-Wohlman, M. Weger, Mol. Cryst. Liy. Cryst. 119
(1986) 457, and references cited therein.
[40] T. Inhabe, S. Nakumura, W.-B. Liang, T. J. Marks, R. L. Burton, C. R.
Kannewurf, K:I. Imaeda, J. Am. Chem. Soc. 107 (1985) 7224.
[41] a) M. Almeida, M. G. Kanatzidis, L. M. Tonge, T. J. Marks, H. 0.Marcy,
W. J. McCarthy, C . R. Kannewurf, SolidState Commun. 63 (1987) 457; b)
877
M. Almeida, M. G. Kanatzidis, L. M. Tonge, T. J. Marks, W. J. McCarthy,
H. 0. Marcy, C. R. Kannewurf, in [2a], p. F37.
[42] For examples of other work on phthalocyanine molecular metals, see: a)
T. Inabe, W.-B. Liang, J. F. Lomax, S. Nakamura, J. W. Lyding, W. J. McCarthy, S. H. Carr, C. R. Kannewurf, T. J. Marks, Synth. Met. 13 (1986)
219; b) J. G. Gaudiello, H. 0. Marcy, W. J. McCarthy, M. K. Moguel,
C. R. Kannewurf, T. J. Marks, ibid. 15 (1986) 115; c) M. Y Ogawa, J.
Martinsen, S. M. Palmer, J. L. Stanton, J. Tanaka, R. L. Greene, B. M.
Hoffman, J. A. Ibers, J. Am. Chem. SOC.109 (1987) 1115; d) K. Liou, C. S.
Jacobson, B. M. Hoffman, ibid. I 1 1 (1989) 6616; e) B. Guay, L. G. Caron,
C. Bourbonnais in [2a], p. F557; f) K. Yakushi, M. Sakuda, I. Hamada,
H. Kuroda, A. Kawamoto, J. Tanaka, T. Sugano, M. Kinoshita, Synth.
Met. 19 (1987) 769; g) S. Hino, K. Matsumoto, H. Yamakado, K. Yakushi,
H. Kuroda, ibid. 32 (1989) 301.
[43] B. A. Scott, S. J. LaPlaca, J. B. Torrance, 8 . D. Silverman, B. Welber, J.
Am. Chem. SOC.99 (1977) 6631.
[44] M. Almeida, M. G. Kanatzidis, L. M. Tonge, H. 0. Marcy, W. J. McCarthy, T. J. Marks, C. R. Kannewurf, unpublished.
[45] a) T. J. Emge, H. H. Wang, M. A. Beno, J. M. Williams, M.-H. Whangbo,
M. Evain, J. A m . Chem. Soc. 108 (1986) 8215, and references cited therein;
b) S. Tomic, D. Jerome, K. Bechgaard, J. Phys. 43 (1982) L 839; c) T.
Takahashi, D. Jerome, K. Bechgaard, ibid. 43 (1982) L 565.
[46] a) T. J. Marks, C. W. Dirk, K. F. Schoch, Jr., J. W. Lyding in F. L. Carter
(Ed.): Molecular Electronic Devices, Marcel Dekker, New York 1982,
p. 195; b) T. Inabe, J. W. Lyding, M K. Moguel, T. J. Marks, J. Phys.
(1983) C3; c) L. J. Pace, A. Ulman, J. A. Ibers, Znorg. Chem. 21 (1982) 199.
[47] a) T. Inabe, C. W. Dirk, K. F. Schoch, Jr., T. J. Marks, J. Am. Chem. SOC.
105 (1983) 1539; b) C. W. Dirk, K. F. Schoch, Jr., T. J. Marks in R. B.
Seymour (Ed.): Conductive Polymers (Polymer Science and Technology
Series, Vol. 1 5 ) , Plenum, New York 1981, p. 209; c) C. W. Dirk, E. A.
Mintz, K. F. Schoch, Jr , T. J. Marks, J. Macromol. Sci.-Chem. A 16 (1981)
275; d) T. J. Marks, K. F. Schoch, Jr., B. R. Kundalkar, Synth. Met. 1
(1980) 337; e) K F.Schoch, Jr., B. R. Kundalkar, T. J. Marks, J. A m
Chem. Soc. 101 (1979) 7071, f) R. D. Joyner, M. E. Kenny, ibid. 82 (1960)
5790.
[48] For cofacial polymers with other types of metallomacrocycles or bridging
groups, see: a) M. Hanack, S. Deger, A. Lange, Coord. Chem. Rev. 83
(1988) 115; b) M. Hanack, A. Datz, R. Fay, K. Fischer, U. Keppeler, J.
Koch, J. Metz, M. Mezger, 0. Schneider, J. H. Schulze, in [4d], Vol. 1,
p. 133; c) J. P. Collman, J. T. McDevitt, C. R. Leidner, G. T. Yee, J. B.
Torrance, W A. Little, J. A m . Chem. Soc. 109 (1987) 4606; d) K. J. Wynne,
Inorg. Chem. 24 (1985) 1339, and references cited therein; e) B. N. Diel, T.
Inabe, N. K. Jaggi, J. W. Lyding, 0. Schneider, M. Hanack, C. R.
Kannewurf, T. J. Marks, L. H. Schwartz, J. Am. Chem. SOC. 106 (1984)
3207; f) T. J. Moyer, L. A. Schechtman, M. E. Kenney, Polym. Prepr.
25(2) (1984) 234; g) M. Hanack, Chimia 37 (1983) 238; h) R. S. Nohr,
P. M. Kuznesof, K. J. Wynne, M. E. Kenney, P. G. Siebenman, J. Am.
Chem. Sor. 103 (1981) 4371.
[49] K. G. Beltsios, T. J. Marks, S. H. Carr, in [2a], p. F31.
[SO] a) X. Zhou, T. J. Marks, S. H. Carr, Mol. Cryst. Liq.Cryst. 118 (1985) 357;
b) X. Zhou, T. J. Marks, S. H. Carr, J. Polym. Sci. Phys. Ed. 23 (1985) 305.
[Sl] E. Ciliberto, K. A. Doris, W. J. Pietro, G. M. Reisner, D. E. Ellis, 1.
Fragala, F. H. Herbstein, M. A. Ratner, T. J. Marks, J. Am. Chem. Soc 106
(1984) 7748.
[52] W. Caseri, T. Sauer, G. Wegner, Makromol. Chem. Rapid Commun. 9
(1988) 651.
[53] B. N. Diel, T. Inabe, J. W. Lyding, K. F. Schoch, Jr., C. R. Kannewurf, T. J.
Marks, J. Am. Chem. SOC.f05 (1983) 1551
[54] T. Inabe, J. G. Gaudiello, M. K. Moguel, J. W. Lyding, R. L. Burton, W. J.
McCarthy, C. R. Kannewurf, T. J. Marks, J. Am. Chem. Soc. 108 (1986)
7595.
[55] T. Inabe, J. C. Butler, H. 0. Marcy, C. R. Kannewurf, T. J. Marks, unpublished.
[56] Single-crystal conductivity measurements are by no means immune to
experimental artifacts and can be severely affected by disorder, defects,
and impurities, as well as other factors related to the purity of the starting
materials and the crystallization methodology [40, 56a-d]. In addition,
even in cases of optimum electrical contact alignment [56e], measurementdependent phenomena such as contact strain [41, 551 and cooling rate
effects [56fl can have major and frequently unappreciated effects on the
observed u( Phthalocyanine conductors are particularly treacherous in
this regard. a) K. Khanna, W. W. Fuller, G. Griiner, P. M. Chaikin, Phys.
Rev. B. 24 (1981) 2958, and references cited therein; b) A. J. Heeger in J. T.
Devreese, R. P. Evrard, V. E. van Doren (Eds.): Highly Conducting OneDimensional Sohds, Plenum Press, New York 1979, p. 105; c) I. D. Begg,
K. J. Roberts, J. N. Sherwood, L. Groth-Anderson, C. S. Jacobsen, Chem.
Phys. L e t r . 79 (1981) 513; d) S. S. Parkin, F. Greuzet, D. Jerome, J. M.
Fabre, K . Bechgaard, J. Phys 44 (1983) 975; e) D. E. Schafer, F. Wudl,
G. A. Thomas, J. P. Ferraris, D. 0.Cowan, SolidState Commun. 14(1974)
347; f) S. Kagoshima, T. Yasunaga, T. Ishiguro, H. Anzai, G. Saito, ibid.
46 (1983) 867, and references cited therein.
(571 a) C. K. Chiang, M. A. Druy, S . C. Gau, A. J. Heeger, E. J. Louis, A. G.
MacDiarmid, Y W. Park, H. Shirakawa. J. Am. Chem. SOC.100 (1978)
1013; b) C. K. Chaing, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis,
n.
878
A. G. MacDiarmid, J. Chem. Phys. 69 (1978) 5098; c) J. P. Farges, A. Brau,
F. Gutmann, J. Phys. Chem. Solids 33 (1972) 1723.
I581 a) B. Lundberg, B. Sundqvist, J. Appl. Phys. 60 (1986) 1074, and references
cited therein; b) W. Y. Hsu, J. R. Barkley, P. Meakin, Macromolecules 13
(1980) 198; c) P. S. Clarke, J. M. Orton, A. J. Guest, Phys. Rev. B 18 (1978)
1813; d) C. H. Seager, G. E Pike, ibid. 10 (1974) 1435; e) S. Kirkpatrick.
Rev. Mod. Phys. 45 (1973) 574.
(591 a)P.Sheng,Phys. Rev.B21(1980)2180;b)Y.-W.Park,A. J.Heeger,M. A.
Druy, A. G. MacDiarmid, J. Chem. Phys. 73 (1980) 946.
[60] a) J. G. Gaudiello, G. E. Kellogg, S. M. Tetrick, T. J. Marks, J. Am. Chem.
Soc. 1I f (1989) 5259, b) G. E. Kellogg, J. G. Gaudiello, J. A. Schlueter,
S. M. Tetrick, T. J. Marks, H. 0. Marcy, W. J McCarthy, C. R.
Kannewurf, in p a l , p. F 15, c) M. Almeida, J. G. Gaudiello, J. C. Butler,
H. 0.Marcy, C. R. Kannewurf, T. J. Marks, Synth. Met. 21 (1987) 261 ; d)
J. G. Gaudiello, M. Almeida, T. J. Marks, W. J. McCarthy, J. C. Butler,
C. R. Kannewurf, J Phys. Chem. 90 (1986) 4917.
1611 For an alternative approach, see: E. A. Orthmann, V. Enkelmann, G.
Wegner, Makromol. Chem. Rapid Commun. 4 (1983) 687.
[62] a) P. G. Pickup, R. A. Osteryoung, J. Am. Chem. Soc. 106 (1984) 2294; b)
F. B. Kaufman, A. H. Schroeder, E. M. Engler, S. R. Kramer, J. Q. Chambers, ibid. 102 (1980) 483.
[63] a) J. S. Miller, A. J. Epstein, Angew. Chem. 99 (1987) 332; Angew. Chem.
Int. Ed. Engl. 26 (1987) 287; b) H. H. S. Javadi, J. S. Miller, A. J. Epstein,
Phys. Rev. Lett. 59 (1987) 1760, c) E. M. Connell, I. A. Howard, Phys.
Rev. B. 31 (1985) 7835
[64] Application of extreme pressures (up to 17 kbar) or cooling to liquid-helium temperatures effects more modest changes ( 512%) in the @ of TTFTCNQ: a) S. Kagoshima, T. Ishiguro, H. Anzai, J. Phys. SOC.Jpn. 41
(1976) 2061; b) G. K. Khanna, J. P. Pouget, R. Conis, A. F. Garito, A. J.
Heeger, Phys. Rev. 816 (1977) 1468; c) S. Megtert, R. Comes, C . Vettier,
R. Pynn, A. F. Garito, Solid State Commun. 37 (1981) 875.
[65] J. A. Schlueter, G. E. Kellogg, T. J. Marks, unpublished results.
[66] a) J. Chen, A. G Heeger, Synth. Met. 24(1988) 311; b) M. Winokur, Y. B.
Moon, A. J. Heeger, J. Barker, D. C. Bott, H. Shirakawa, Phys. Rev. Left.
58 (1987) 2329; c) Y B. Moon, M. Winokur, A. J. Heeger, J. Barker, D. C.
Bott, Macromolecules 20 (1987) 2457; d) D. C. Bott, C. S. Brown, J. N.
Winter, J. Barker, Polymer 28 (1987) 601; e) L. W. Shacklette, J. E. Toth.
Phys. Rev. B 32 (1985) 5892.
[67] a) G. Wieners, R. Weizenhofer, M. Monkenbusch, M. Stamm, G. Lieser,
V. Enkelmann, G. Wegner, Makromol. Chem. Rapid Commun. 6 (1985)
425; b) J. P. Pouget, J. C. Pouxvlel, P. Robin, R. Comes, D. Begin, D.
Billaud, A. Feldblum, H. W. Gibson, A. J. Epstein, Mol. Cryst. Liq.Cryst.
117 (1985) 75; c) H. W. Hasslin, C. Riekel, K. Menke, S. Roth, Makromol.
Chem. 185 (1984) 397; d) R. H. Baughman, N. S. Murthy, G. G. Miller, J
Chem. Phys. 79 (1983) 515; e) J C. W. Chien, F. E. Karasz, K. Shimamura,
Macromolecules 15 (1982) 1012.
[68] a) P. D. Hale, W. J. Pietro, M. A. Ratner, D. E. Ellis, T. J. Marks, J. Am.
Chem. Sor. 109 (1987) 5943; b) K . A. Doris, E. Ciliberto, I. Fragala, M. A.
Ratner, T. J. Marks in [2d], p. 337; c) W J. Pietro, T. J. Marks, M. A
Ratner, J. Am. Chem. SOC.107(1985) 5387; d) W. J. Pietro, D. E. Ellis, T. J.
Marks, M. A. Ratner, Mol. Cryst. Liq. Cryst. 105 (1984) 273.
[69] J. A. Schlueter, T. J. Marks, unpublished results.
[70] M. Almeida, J. G Gaudiello, G. E. Kellogg, S. M. Tetrick, H. 0. Marcy,
W. J. McCarthy, J. C. Butler, C. R. Kannewurf, T. J. Marks, J. A m . Chem.
SOC.i f 1 (1989) 5271.
(711 a) N. F. Mott, Festkorperprobleme 21 (1979) 333; b) N. F. Mott, E. A.
Davis: Electronic Processes in Non-Crystalline Materials, 2nd ed., Clarendon, Oxford 1979, Chapters 1, 2.
1721 a) A. J. Epstein, H. Rommelmann, R. Bigelow, H. W. Gibson, D. M. Hoffman, D. B. Tanner, J. Phys. C 3 (1983) 61; b) R. Colson, P. Nagles, J.
Non-Cryst. Solids 36 (1980) 129, and references cited therein.
[73] a) P. Sheng, B. Abeles, Y Arie, Phys. Rev. Lett. 31 (1973) 44; b) E. K.
Sichel, J. I. Gittleman, P. Sheng, Phys. Rev. B 18 (1978) 5712; c) E. K
Sichel, Appl. Phys. Commun. f (1981) 83; d) B. Abeles, P. Sheng, M. D.
Coutts, Y. Arie, Adv. Phys. 24 (1975) 407.
[74] a) Y. Tomkiewicz,N. S. Shiren, T. D. Schultz, H. Thomann, L R. Dalton,
A. Zettl, G. Griiner, T. C. Clarke, Mol. Crysf. Liq. Crysf. 83 (1982) 17; b)
K. Mortensen, M. L. W. Thewalt, Y. Tomkiewicz, T. C. Clarke, G. B.
Street, Phys. Rev. Lett. 45 (1980) 490.
[75] S. M. Tetrick, T. J. Marks, unpublished.
[76] a) N.F. Mott: Conduction in Non-CrystalIine Materials, Clarendon,
Oxford 1987, Chapters 2, 3; b) N. F. Mott: Metal-Insulator Transitions,
Taylor and Francis, London 1974, Chapters 1,6; c) V. L. Berezinskii, Sov.
Phys. JETP 38 (1974) 620; d) P. W. Anderson, Phys. Rev. 109 (1958)
1492.
[77] I n a rigorously one-dimensional chain, disorder should lead to localization
of all states, and conduction can only occur via phonon-assisted hopping
[76b, 77a, b]. To our knowledge, such a high degree of unidimensionality
has not yet been achieved, and a11 known molecular metals exhibit nonnegligible two- or three-dimensional character. a) A. N. Bloch, R. B.
Wilson, C. M. Varma, Phys. Rev. Lett. 28 (1972) 753; b) N. F. Mott, Adv.
Phys. 16 (1967) 49.
[78] M. K. Moguel, L. M. Tonge, H. 0. Marcy, W. J. McCarthy, C. R.
Kannewurf, T. J. Marks, unpublished.
Angew. Chem. Ini. Ed. Engl. 29 (1990) 857-879
1791 a) J. B. Torrance, Y. Tomkiewicz, R. Bozio, C. Pecile, C. R. Wolfe, K.
Bechgaard, Phys. Rev. B 26 (1982) 2267, and references cited therein; b)
J. B. Torrance, J. J. Mayerle, K. Bechgaard, B. D. Silvermann, Y.
Tomkiewicz, hid. 22 (1980) 4960, and references cited therein; c) A. N.
Bloch, S. Mazumdar, private communication.
[80] a) E. J. Baerends, D. E. Ellis, P. Ros, Chem. Phys. 2 (1973) 41; b) 2.
Berkovitch-Yellin, D. E. Ellis, M. A. Ratner, ibid. 62 (1981) 21; c) B. Delley, D. E. Ellis, J. Chem. Phys. 76 (1982) 1949.
[81] a) K. A. Doris, D. E. Ellis, M. A. Ratner, T. J. Marks, J. Am. Chem. SOC.
f06 (1984) 2491; b) K. A. Doris, B. Delley, M. A. Ratner, T. J. Marks,
D. E. Ellis, J Phys. Chem. 88 (1984) 3157.
[82] a) X. 2. Yang, D. B. Tanner, M. J. Rice, H. W. Gibson, A. Feldblum, A. J
Epstein, Solid Slate Commun. 61 (1987) 335; b) E. J. Mele, M. J. Rice,
Phys. Rev. B 23 (1981) 5397.
[83] J. Chen, A. J. Heeger, Synth. Met. 24 (1988) 311; b) A. J. Heeger, in [4c],
Chapter 21, and references cited therein; c) S. Kivelson, A. J. Heeger, Phys.
Rev. L e t f . 55 (1985) 308.
Angew. Chem. Inr. Ed. Engl. 29 (1990) 857-879
[84] a) G. B. Street in [4c], Chapter 8; b) P Pfluger, G . Weiser, J. C. Scott, G. B.
Street in [4c], Chapter 38; c) F. Genoud, M. Guglielmi, M. Nechtschein.
E. Genies, M. Salmon, Phys. Rev. L e f t . 55 (1985) 118.
[85] a) R. P. McCall, J. M. Ginder, M. G. Roe, G. E. Asturias, E. M. Scherr,
A. G. MacDiarmid, A. J. Epstein, Phys. Rev. B39 (1989) 10174; b) F. Zuo,
M. Angelopoulos, A. G. MacDiarmid, A. J. Epstein, ibid.36 (1987) 3475;
c) J. M. Ginder, A. F. Richter, A. G. MacDiannid, A. J. Epstein, Solid
Stale Commun. 63 (1987) 97.
[86] F.Wudl, R. 0. Angus, Jr., F. L. Lu, P. M. Allemand, D. J. Bachon, M.
Novak, 2.X. Liu, A. J. Heeger, J. Am. Chem. Soc. 109 (1987) 3677, and
references cited therein
[87] a) J. B. Schlenoff, J. Obrzut, F. E. Karasz, Phys. Rev. E 40 (1989) 11822; b)
I. M. Madsen, B. R. Johnson, X. L. Hua, R. B. Hallock, M. A. Masse,
F. E. Karasz, ibid. 40 (1989) 11751; c) M. J. Obrzut, F. E. Karasz, Macromolecules 22 (1989) 458; d) M. J. Obrzut, F. E. Karasz, J Chem. Phys. 87
(1987) 6178; e) T. Granier, E. L. Thomas, D. R. Gagnon, F. E. Karasz.
R. W. Lenz, J. Polym. Sci. Poiym. Phys. Ed. 24 (1986) 2793.
879
Документ
Категория
Без категории
Просмотров
0
Размер файла
2 475 Кб
Теги
electrical, structure, molecular, metallomacrocycles, enforced, уmetalsф, polymeric, interface, conducting, assemblies
1/--страниц
Пожаловаться на содержимое документа