Electrochemical synthesis, characterization and magnetic studies of Ni/TCNQ salts

Article abstract of DOI:10.1016/S0020-1693(00)00260-7

Two new materials, Ni(TCNQ) and Ni₃(TCNQ)₂, have been obtained by an electroplating technique and characterized by SEM, XPS, vibrational spectroscopy, X-ray diffraction and magnetic techniques. The compound Ni(TCNQ), obtained at lowe

Full text of DOI:10.1016/S0020-1693(00)00260-7

www.elsevier.nl/locate/ica
Inorganica Chimica Acta 313 (2001) 1–14
Electrochemical synthesis, characterization and magnetic studies
of Ni/TCNQ salts
Gregory Long a, Roger D. Willett b,*
a Department of Chemistry, Minnesota State Uni6ersity, Mankato, MN 56002, USA
b Department of Chemistry, Washington State Uni6ersity, Pullman, WA 99164, USA
Received 24 March 2000; accepted 28 June 2000
Abstract
Two new materials, Ni(TCNQ) and Ni3(TCNQ)2, have been obtained by an electroplating technique and characterized by SEM,
XPS, vibrational spectroscopy, X-ray diffraction and magnetic techniques. The compound Ni(TCNQ), obtained at lower applied
voltages, contains paramagnetic Ni(II) cations and diamagnetic TCNQ−2 anions. In contrast, the material obtained at higher
applied voltages, Ni3(TCNQ)2, contains diamagnetic Ni(II) cations and S=1/2TCNQ−3 anions. Thus, the higher oxidation states
of the organic ligand have been stabilized by control of the applied electroplating voltage. Magnetic studies of polycrystalline
samples of both compounds show the presence of dominant ferromagnetic interactions at higher temperatures, but antiferromag-
netic interactions dominate at low temperatures. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Electrochemistry; Vibrational spectroscopy; X-ray diffraction; Magnetic studies; Ni/TCNQ solutions
sition metal ion and an organic constituent such as a
protonated amine, can be classified according to the
orbitals in which the spins responsible for the magnetic
behavior reside [9]. The organic moiety may be an
active magnetic component of the material with its spin
sites contributing to both the non-zero magnetic mo-
ment as well as any associated spin coupling. Alterna-
tively, the organic moiety may be a passive component
by only providing a means of positioning the spins,
which reside solely on metal ions, so as to facilitate
their coupling. As such, molecular based magnets are
qualitatively distinct from conventional inorganic based
magnets such as CrO2.
During the past few decades, chemists and material
scientists have been actively employing an interdisci-
plinary molecular engineering strategy, by coupling in-
organic and organic techniques and synthetic methods,
so as to achieve novel magnetic systems that undergo a
spontaneous magnetic phase transition at a finite tem-
perature [10–12].
1. Introduction
The development of present day experimental and
theoretical condensed matter chemistry and physics is
deeply indebted to magnetism and the magnetic behav-
ior of many advanced materials [1].
Initially, compounds based on molecular and/or
polymeric materials that display significant magnetic
behavior were thought unlikely due to their relatively
low spin densities. However, the discovery in the mid
1980s [2–4] of the transition metal–phthalocyanine
magnets, all of which display significant magnetic tran-
sition temperatures in the range of liquid helium mea-
surements, and the reports [5–8] in 1991 of transition
metal–TCNE based polymeric magnets and of a Prus-
sian Blue analog with transition temperatures above
room temperature, have accelerated research efforts
directed at the development and understanding of
molecular/polymeric magnets.
In general, molecular-based magnets, which are usu-
ally composed of an inorganic component such a tran-
In recent years there has been renewed interest in
molecular magnets, due to the report of two systems
which exhibited permanent moments at room tempera-
ture. Verdaguer et al. [8] used a systematic theoretical
* Corresponding author. Fax: +1-509-3358867.
E-mail address: willett@mail.wsu.edu (R.D. Willett).
0020-1693/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(00)00260-7

2
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
approach to design a series of Prussian blue analogs
that have high values of Tc. This has spurred further
studies of cyanide based systems [13–15]. In addition,
electron transfer salts have become an aggressively
studied class of low dimensional magnetic materials
due, in part, to the recent discovery of the material
[V(TCNE)x]·y(CH2Cl2) which exhibits spontaneous
magnetization at room temperature [16–18]. Although
this material is pyrophoric and is difficult to character-
ize, it nonetheless represents a significant advance in the
development of molecular magnets as a potential device
for industrial applications.
In our efforts with new synthetic strategies for molec-
ular magnets, we have decided to focus initially on the
electron transfer-salts of the transition metal ions and
the electron acceptor tetracyanocarbon ligand 7,7,8,8-
tetracyanoquinodimethanide (TCNQ) [19]. Potember
and coworkers [20–22] discovered that the material can
switch from a high impedance to a low impedance
hydrolysis product of TCNQ, the anion of a,a-dicyano-
p-toluoylcyanide (DCTC−1), in particular, is a com-
mon impurity in solution derived materials that can
have pronounced affects on the bulk magnetic proper-
ties of metal–TCNQ salts [38] due to its S=0 diamag-
netic ground state.
In addition to the difficulty of preparing impurity-
free materials, the solution metathesis methods have
also been relatively unsuccessful in their ability to con-
trol morphologies and stoichiometries of metal–TCNQ
thin-films or in yielding stable materials containing
higher oxidation states of the TCNQ anion [39]. The
lowest oxidation state, TCNQ−3
, is particularly
difficult to stabilize in solution prepared materials.
In this work, a unique electrochemical method of
preparation similar to that outlined by Shields [40] has
been employed to synthesize Ni(II) salts of TCNQ and
to characterize their magnetic properties. As will be
shown, this method of preparation offers several dis-
tinct advantages over traditional solution methods,
such as the minimization of impurities, improved con-
trol over product stoichiometry, as well as the ability to
produce stable compounds of the higher TCNQ anionic
oxidation states.
After this project was underway, it was learned the
Professor Kim Dunbar was investigating the transition
metal–TCNQ species obtained by solution methods.
For M=Ni(II), a Ni(TCNQ)2 salt containing the
TCNQ−1 anion is obtained [41].
The next section describes the synthetic method as
well as the battery of investigative methods used to
characterize the materials produced via the electro-
chemical technique, while the subsequent sections docu-
ment the results and conclusions. In particular, we
show that the electrochemical technique can be used to
control the oxidation state of the TCNQ species.
material by applying an electric field across
a
MꢀTCNQꢀM% (where M=Cu or Ag) device. In addi-
tion, this impedance change can also be induced by
laser radiation [23–25]. Although other inorganic mate-
rials such as the chalcogenide semiconductors exhibit
this behavior, the MꢀTCNQꢀM% systems were the first
hybrid organic–inorganic compound to do so. Al-
though no device has been produced to date, this
optical switching capability of Cu(TCNQ) suggests that
the material may be hold potential applications as an
optical data storage device [26–28].
Due to its demonstrated versatility as an electron
acceptor molecule and its ability to form several stable
radical species [29], and the ability of these reduced
species to coordinate to a metal center as a s-donor via
one or more of its nitrile groups [30], TCNQ is a useful
organic acceptor molecule that has yielded a wide vari-
ety of interesting molecular magnets [31].
The most common synthetic method for TCNQ
based materials involves a metathesis reaction between
an alkali metal salt of the TCNQ anion and a suitable
salt, typically a halide, of a transition metal ion in
acetonitrile [32–34].
2. The synthetic method and characterization techniques
In this work, an electrochemical method of prepara-
tion similar to that outlined by Shields [40] has been
employed to produce both bulk and thin film metal–
TCNQ compounds. As will be shown, this method of
preparation appears to offer several distinct advantages
over traditional solution methods [42], including mini-
mization of impurities and enhanced control over both
product stoichiometry and thin-film deposition
morphologies.
nLi(TCNQ)+MA [ M(TCNQ)n+LinA
In most cases, however, such preparative methods
lead to materials that do not exhibit significant temper-
ature dependent magnetic susceptibilities [35]. This is
due, at least in part, to the tendency of the S=1/2
TCNQ−1 anion to dimerize into one-dimensional anti-
ferromagnetic chains in the solid state [36] such that
coupling between paramagnetic metal ions is
minimized.
2.1. Experimental
The questionable degree of purity that can be at-
tributed to the very fine microcrystalline powders de-
posited from solutions can also have a pronounced
effect on their electrical and magnetic properties [37]. A
The electrochemical cell used for the synthesis of all
materials presented in this work consists of a two
electrode system as shown schematically in Fig. 1. The
cell was powered by a McKee–Pherson Electronics Inc.

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
3
d.c.-stabilized potentiostat and, in each case, the intrin-
2.3. Method
sic impedance (MV) of the cell assured that the current
wasB1 mA. The distance between the electrodes was
kept constant at 1 cm. The effects of varying voltage
will be discussed in a subsequent section.
In all cases, 50 ml of the HPLC grade acetonitrile
was transferred to the reaction cell and purged with dry
helium gas via the purge needles illustrated in Fig. 1 for
15 min prior to the introduction of the TCNQ so as to
minimize the formation of the hydrolysis product
DCTC. Re-crystallized TCNQ (0.255 g) was then added
to the cell so as to form a 0.05 M solution (solution is
nearly saturated). The solution was then purged again
with dry helium for 10 min. Upon introduction of the
source and substrate electrode assembly, the cell was
again purged with helium and then biased to the de-
sired d.c. voltage. Within 15–120 min, depending upon
source material and cell bias, the desired product mate-
rial began precipitating out of solution at the surface of
the cathodic electrode. The cell was allowed to run (30
min to 3 days depending on source and bias) until the
surface coverage of the cathodic platinum foil was
deemed sufficient by eye so as to provide adequate yield
of product. Upon completion of the product deposi-
tion, the substrate electrode was removed from the
assembly and the product was washed with several
volumes of degassed acetonitrile and allowed to dry in
a dessicator at room temperature prior to subsequent
analysis.
2.2. Reagents
The solvent used was Aldrich HPLC spectroscopic
grade degassed acetonitrile. TCNQ was also from
Aldrich and was sublimed twice under vacuum at
180°C and subsequently re-crystallized twice from spec-
troscopic grade degassed acetonitrile prior to use. The
same batch of purified TCNQ, which was stored in a
dessicator, was used for all of the materials synthesized
in this work.
The substrate electrodes consisted of platinum foil
purchased from Aldrich at 99.999% purity. The foil was
cut to the desired size and cleaned with a 1:1 HNO3
acid:H2O solution, dried at 150°C overnight and cooled
in a dessicator prior to use.
The nickel source electrodes were also purchased
from Aldrich as foils at 99.999% purity. As before, each
was cut to the desired size and cleaned with a 1:1 HNO3
acid:H2O solution, dried at 150°C overnight and cooled
in a dessicator prior to use.
Fig. 1. A schematic illustration of the electrochemical cell used to synthesize metal–TCNQ compounds from 0.05 M acetonitrile solution. The cell
atmosphere consisted of dry helium gas during the synthesis.

4
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
Table 1
2.6. X-ray photoelectron spectroscopy
Analytical data for electrochemically synthesized Ni–TCNQ com-
plexes
XPS spectra for all of the materials reported in this
work were determined using a Kratos Analytical Axis
165 electron spectrometer using Mg–Ka X-rays. All of
the samples, which had been isolated from the platinum
cathode using a Kel-F spatula and were washed and
dried as specified earlier, were analyzed as powders
pressed into clean indium foil. No compensation for
sample charging effects was necessary. The metal 2p3/2
binding energies, nitrogen 1(s) binding energies, Auger
line energies and Auger parameter energies are the
average of at least five replicate measurements with the
confidence limits taken as the standard deviations of
the sample set.
Compound Anal. Found (Calc.) (%)
C
H
N
Ni
Ni(TCNQ) 54.7 (54.8) 1.5 (1.5)
(1)
Ni3(TCNQ)2 49.3 (49.3) 1.4 (1.4)
21.3 (21.3) 22.5 (22.4)
19.2 (19.0) 30.1 (30.3)
(2)
The low voltage product Ni(TCNQ) (1) and the high
voltage product Ni3(TCNQ)2 (2), were synthesized at a
constant 5 and 10 d.c. V applied cell bias, respectively,
for a period of 72 h. In both cases the size of the
platinum cathode was 3.0×0.30 cm. Analytical data
for these materials are given in Table 1.
2.7. Powder X-ray diffraction
No samples of the electrochemically synthesized tran-
sition metal complexes of TCNQ suitable for single
crystal X-ray diffraction analysis have been produced.
However, in order to explore relative differences in
their structures, particularly as a function of applied
cell voltage, the materials described in this work have
been analyzed via powder X-ray diffraction.
The X-ray powder patterns were collected with a
Siemans D501 diffractometer using copper Ka radia-
tion. The samples were removed from the platinum
electrode, washed and dried as discussed earlier, and
homogenized with an agate mortar and pestle prior to
analysis. Indexing of the patterns was completed using
the Appleman [45] software in all cases.
2.4. Scanning electron microscopy and energy dispersi6e
X-ray analysis
The SEM and EDX [43] analyses for all materials
produced in this work were conducted using a JEOL
JSM 6400 scanning electron microscope (SEM) fitted
with a LINK Analytical Si(Li) EDX detection system.
The EDX spectra for all of the washed and dried
M(TCN)Q products were acquired prior to removal
from the platinum substrates at an SEM accelerating
voltage of 20 kV and a working stage distance of 8 cm.
The platinum electrode was in all cases anchored to the
brass SEM stage with both conductive graphite tape
and a metallic grounding screw. The SEM micrographs
of the materials were acquired at magnifications and
accelerating voltages as specified on an individual basis.
Due to the absence of significant charging effects [44],
none of the materials analyzed were sputter coated with
an Au film prior to analysis.
2.8. Magnetic susceptibility
The magnetic susceptibility for powder samples all of
the M(TCNQ) compounds reported in this work were
determined from 4.4 K using a Lakeshore 7000 a.c.
susceptometer. All of the data were background cor-
rected between 4.4 and 300 K using a delrin sample
container. Corrections for diamagnetic behavior were
estimated from Pascal’s constants [46].
2.5. Infrared spectroscopy
The washed and dried product materials were re-
moved from the surface of the platinum cathode using
a Kel-F spatula. A small sample of the materials were
then homogenized, using an agate mortar and pestle,
with KCl which had been dried at 200°C overnight and
then cooled in a dessicator prior to use. The diffuse
reflectance IR spectra of the materials were acquired
with an IBM Instruments IR-98 Fourier transform
infrared spectrometer with a TGS detector. All of the
spectra, which were background corrected and are re-
ported in units of absorbance, were acquired at 4 cm−1
resolution (2 cm−1 per data point) with 512 scans
coadded.
3. Results and discussion
The EDX spectrum exhibits the characteristic Ka, Kb
and average Lab X-rays of nickel at 7.5, 8.3 and 0.85
eV, respectively. No other impurities to which EDX is
sensitive, such as metals or non-metallic elements with
Z\10, are present at a detectable level. The EDX
spectrum of compound 2 is identical to that of com-
pound
1
and again shows no evidence of
contamination.

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
5
3.1. Structural studies
and 3, respectively. Both materials are polycrystalline in
nature and exhibit different powder patterns thus sug-
gesting the electrochemical method of preparation may
give rise to structural variations in the nickel products
Powder X-ray diffraction data for compounds 1 and
2, presented in Section 2, are also presented in Figs. 2
Fig. 2. X-ray diffraction powder pattern of 1.
Fig. 3. X-ray diffraction powder pattern of 2.

6
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
Table 2
Vibrational peak positions (cm−1) for l(CꢀH) IR bands for TCNQ, TCNQ anions and DCTC salts
TCNQ
859
TCNQ−1
TCNQ−2
822
TCNQ−3
Cu(DCTC)
834
Ag(DCTC)
835
Assignment
802
823
812
819
822
b3u
b3u
b3u
b2g
1046
as a function of applied cell voltage. Indexing of the
major reflections present in the powder pattern of com-
pound 1 suggest a cubic lattice with cell dimensions of
cm−1 and is consistent with other nickel salts of
TCNQ−2 prepared by solution methods [56]. The char-
acteristic b3g band of DCTC−1 at 835 cm−1 is not
evident in the l(CꢀH) spectra of compound 1, suggest-
ing that no significant contamination from the hydroly-
sis product is present. Furthermore, the lack of
absorption maxima in the l(CꢀH) region 850–880
cm−1 suggest that no neutral TCNQ0 is present in the
material.
,
10.78(2) A. When additional reflections, which are also
shown in Fig. 2, are included in the indexing routine,
no reliable solution is possible. These additional reflec-
tions, which do not correspond to known reflections of
crystalline TCNQ [47], are, although relatively weak,
reproducible and are above the estimated background
noise present in the diffractometer. They are therefore
considered to be real and suggest that the true structure
has a lower symmetry structure.
The IR spectrum of compound 2 in the l(CꢀH)
region is shown in Fig. 6. For the high voltage product,
two strong bands are present at 809 and 828 cm−1
,
The major reflections present in the powder pattern
of compound 2, as shown in Fig. 3, can be indexed on
a tetragonal cell with dimensions a=10.898(12) and
which are consistent with stretches observed for the
TCNQ−3 anion in the solid state [48,57]. As with
compound 1, no evidence of either the hydrolysis
product DCTC nor neutral TCNQ0 is detectable in the
spectrum.
The w(CꢀN) bands for compounds 1 and 2 are shown
in Figs. 7 and 8, respectively. This region of the IR
spectra can yield information regarding the coordinate
status of the TCNQ anion within the solid state mate-
rial. In uncomplexed TCNQ−2 and TCNQ−3 only two
,
c=3.008(29) A. However, as in the case of the low
voltage product, when additional reflections, which
again are reproducible and above background, are in-
cluded in the indexing routine no reliable solution is
found. These additional reflections suggest a distortion
of the lattice to one of symmetry lower than tetragonal.
3.2. Spectroscopic studies
The number and frequencies of the IR bands of
TCNQ are highly indicative of its oxidation state and
coordinate status [48–53]. In addition, solution synthe-
sized polycrystalline samples of copper and silver salts
of DCTC have been studied via IR spectroscopy [54].
As is shown in Table 2, the IR spectra of the copper
and silver salts of DCTC, which are relatively indepen-
dent of ion environment, are distinctively different from
those of TCNQ in the l(CꢀH) IR region and can thus
be used as a probe for the presence of this hydrolysis
product. Further studies of solution prepared copper
salts of TCNQ deposited on the surface of copper foil
[55] suggest that an impurity level of as little as 3% of
DCTC by mass is detectable by infrared spectroscopy
for metal–TCNQ salts in the solid state.
The diffuse reflectance infrared spectra in the l(CꢀH)
region for compound 1 are given in Figs. 4 and 5. As is
shown, compound 1 gives a single band in the l(CꢀH)
regions at 815 and 1050 cm−1 which is indicative of the
TCNQ−2 anion [48,51]. The slight shift in the frequen-
cies of these two bands, relative to observed stretches in
alkali metal salts of TCNQ, is on the order of ca. 4
Fig. 4. l(CꢀH) infrared bands in the 800–880 cm−1 region for 1.

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
7
Fig. 5. l(CꢀH) infrared bands in the 1030–1060 cm−1 region for 1.
Fig. 7. w(CꢀN) infrared bands for 1.
w(CꢀN) stretches in the 2000–2300 cm−1 region are
expected and found [58]. The occurrence of multiple
w(CꢀN) bands, as observed in both of these com-
pounds, has been attributed to a reduction in symmetry
of the TCNQ anion caused by an interaction between
the highly polarizing positively charged metal center
and the cyano groups of the TCNQ anions [59]. Such
interactions have been observed in the compound
Fig. 6. l(CꢀH) infrared bands in the 800–880 cm−1 region for 2.

8
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
Studies of transition metal derivatives of TCNQ by
Azcando et al. [63] suggest that the splitting of the
w(CꢀN) bands of TCNQ anions into three or more
bands in the solid state is a result of direct alpha
coordination of the anions to the metal center through
the negative charge density of the cyano groups, as
opposed to coordination of the metal ion to the TCNQ
anions through the p-electron density of the ring. In
addition, studies of metal–TCNQ derivatives by Cor-
nelisson suggest that the splitting of these bands is often
accompanied by a frequency shift of the w(CꢀN) bands
to slightly higher energies relative to bands observed in
uncoordinated ligands [64]. Comparison of the w(CꢀN)
band splitting observed in compounds 1 and 2 with that
observed by Azcando et al. suggest that the TCNQ
anions in both electrochemically synthesized materials
may be coordinated directly to the metal center through
one or more of the terminating cyano groups.
X-ray photoelectron spectroscopic (XPS) measure-
ments for the nickel and nitrogen atoms of compounds
1 and 2 are shown in Figs. 9 and 10, and were obtained
with a magnesium Ka X-ray source. In the past several
decades, a plethora of studies of transition metal com-
pounds have demonstrated that experimentally mea-
sured binding energies of metal ion core electrons often
differ significantly between materials where the metal
ions are either exposed to different chemical environ-
ments in the solid state or are present in different
Fig. 8. w(CꢀN) infrared bands for 2.
(C5H2)2VBrꢀTCNE [60] and in the silver cyanocarbon
derivatives AgC(CN)2NO [61] and AgC(CN)2NO2 [62].
Fig. 9. Nickel and nitrogen XPS spectra for compound 1.

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
9
Fig. 10. Nickel and nitrogen XPS spectra for compound 2.
oxidation states [65–69]. Measured differences in core
electron binding energies for a variety of different tran-
sition metal materials have been tabulated [70] and are
often used as a probe of the ion environments and
charge states.
state for the nickel ion and also suggests that the metal
ion is coordinated to a nitrogen containing ligand such
as an organo–cyanide [75]. The binding energy of the
measured nitrogen 1s XPS transition at 398.3 eV is in
further agreement with literature values of reduced
organic amines, cyanide in particular, coordinated to
metal atoms in the solid state [76]. No evidence of other
nickel transitions appear present in the spectrum, which
suggests that the material does not contain oxidation
states of the metal atom other than the +2 state.
The nickel 2p3/2 binding energy for compound 2, at
855.8 eV, is strongly indicative of a Ni(II)ion coordi-
nated directly to one or more cyanide ligands such as in
K2Ni(CN)4 [77]. The slight reduction of the 2p3/2 bind-
ing energy of this line in the XPS spectra of compound
2, relative to compound 1, is predictable due to the
decrease in electronegativity of the TCNQ anion with
increasing charge. The Auger parameter calculated for
this material, based on the measured kinetic energy of
the LMM2 Auger transition of the nickel atom is 1695.9
eV, and is further consistent with values observed for
nickel atoms in the +2 oxidation state coordinated to
cyanide ligands [77]. As seen in compound 1, the mea-
sured binding energy of the nitrogen 1s XPS line for
compound 2, at 398.2 eV, is again in excellent agree-
ment with values observed for Ni(II) compounds in
which direct coordination of the cyanide ligand to the
metal center is present [61]. Again, no evidence of
The most reliable, and most common method of
determining metal ion oxidation states using electron
spectroscopy is by the determination of the Auger
parameter [71], which is defined by Wagner [72] as the
sum of the kinetic energy of the most intense Auger line
and the binding energy of the most intense photoelec-
tron line. Similar to the XPS core electron binding
energies, the Auger parameters of transition metal com-
pounds are also extensively tabulated [73]. The Auger
parameters and binding energies of the XPS 2p3/2 lines
differ significantly as the chemical environment and
oxidation state of the copper ion changes. In this work,
the oxidation state of the metal ion in the M(TCNQ)
products were estimated by comparison of tabulated
Auger parameters and XPS 2p3/2 line energies of similar
materials with experimentally determined values.
The measured nickel 2p3/2 binding energy for com-
pound 1, 856.3 eV is consistent with a 3d8 Ni(II) ion
which may be coordinated to a nucleophilic moiety
such as a cyanide ligand or an organic amine [74]. The
kinetic energy of the LMM2 Auger transition, measured
at 840.9 eV, yields a calculated Auger parameter of
1697.2 eV. This is further indicative of a +2 oxidation

10
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
oxidation states other than +2 for the metal is present
in the spectrum.
MT to increase with decreasing temperatures and a
weaker antiferromagnetic exchange that becomes im-
portant at lower temperatures which tends to align the
resultant moments antiparallel. Compound 1 and 2
show maxima in MT at Tc=28.3 and 8.6 K,
respectively.
3.3. Magnetic susceptibility and EPR studies
The magnetic susceptibility of compounds 1 and 2
were measured from 4.4 up to ca. 200 K. Both materi-
als were studied at a frequency of 375 Hz and an a.c.
excitation field of 0.4 Oe. EPR measurements on pow-
dered samples of both materials were taken with a
Bruker EP300 X-band spectrometer at a microwave
frequency of 9.3 GHz at room temperature. The aver-
age g values obtained from these measurements were
used to fit the susceptibility data. Diamagnetic correc-
tions for both materials were estimated from Pascal’s
constants.
The deviation from Curie behavior of the data at low
−1
M
temperatures present in both 
plots is characteristic
of systems which are not pure Curie magnets and
suggests the presence of magnetic coupling between
paramagnetic moieties, significant zero-field splittings
or a combination of the two, as is often the case with
nickel compounds [78]. At higher temperatures, where
the data are expected to follow Curie behavior, linear
least-squares regression of the data for compound 1
suggests a Curie constant of 1.19 emu K mol−1 and a
Curie–Weiss constant of 22.6 K. These values are
consistent with the reported stoichiometry of
Ni(TCNQ) which contains one paramagnetic S=1 3d8
Ni(II) ion and one S=0 diamagnetic TCNQ species
The magnetic susceptibility data of compound 1 and
2, shown in Figs. 11 and 12, respectively, show evidence
of two types of magnetic interaction: a dominant ferro-
magnetic exchange at high temperature which causes
Fig. 11. Plot of MT versus T for 1. (Experimental data, ꢀ; best fit, —).

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
11
Fig. 12. Plot of MT versus T for 2. (Experimental data, ꢀ; best fit, —).
per mole of compound. The positive value of the
Curie–Weiss constant suggests the presence of predom-
inant ferromagnetic behavior at high temperatures [79].
The value of the Curie constant obtained from the
regression, 1.19 emu K mol−1, is also consistent with a
high temperature extrapolation of the MT data in Fig.
11.
The EPR spectrum of compound 1 at room tempera-
ture consisted of a single symmetric broad line centered
at g=2.20 with peak-to-peak line width of ca. 600 Oe
as expected for an S=1 Ni(II) ion. The diamagnetic
TCNQ−2 is not expected to contribute to the EPR
spectrum of the material, and no evidence of a sharp
g=2.00 signal due to a TCNQ−1 or TCNQ−3 species
is present.
the fitting as 47.2 K, is positive as expected for ferro-
magnetic interactions.
The value of the Curie constant obtained from these
calculations, which is consistent with the high tempera-
ture extrapolations of the data in Fig. 12, is interesting
in that it is approximately an order of magnitude lower
than expected if the contributions of three paramag-
netic S=1 Ni(II) ions per mole of material are consid-
ered. The diamagnetic behavior of the metal ions is
expected, however, if the Ni(II) ions have a square-pla-
nar D4h geometry as is known to occur with strong field
ligands such as the cyanides [80]. Preliminary ab initio
calculations conducted by Professor Ron Poshusta of
Washington State University of TCNQ in the −1, −2
and −3 oxidation states suggest an increase of electron
density on the terminal cyano groups of TCNQ accom-
pany a decrease in oxidation state. These calculations
yield approximate an electron density value of −0.5
for each of the terminal nitrogen atoms in the
TCNQ−3 anion which may be sufficient to allow the
Linear regression of the high temperature Curie data
for compound 2 yields a Curie constant of 0.8 emu K
mol−1 which is consistent with paramagnetic contribu-
tions only from two S=1/2 TCNQ−3 species per mole
of material. The Curie–Weiss constant, calculated from

12
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
TCNQ anion to behave as a strong field ligand and
thus force the nickel ions into a diamagnetic square
planar geometry.
0.9999 over the specified temperature range, suggest an
intrachain coupling constant of J/k=39.190.8 K and
an interchain coupling constant of J%/k= −0.029
0.004 K. The best agreement between the experimental
and calculated molar susceptibilities was given by fixing
the value of the number of equivalent nearest neighbor
chains parameter to 2. The anisotropic g value, as
measured from the EPR spectrum, was fixed at a value
of g=2.000 throughout the fitting routine.
The qualitatively susceptibility data are indicative of
low dimensional ferromagnetic coupling with weaker
intersystem antiferromagnetic coupling. Thus, the sus-
ceptibility data was modeled with an appropriate low
dimensional model. The susceptibility data above 30 K
for compound 1 was fit to a high-temperature series
expansion S=1 linear chain Heisenberg model [81]
based on the Hamiltonian H= −2JijSi·Sj that incorpo-
rates zero-field splitting terms for the nickel ions and
mean-field correction terms which account for inter-
chain interactions. Non-linear curve fitting of the data
to this model, which gave an estimated goodness-of-fit
parameter of 0.9998 [82] over the specified temperature
range, suggests an intrachain coupling constant of J/
k=14.690.02 K, an interchain coupling constant of
J%/k= −1.8890.04 K, and a zero-field splitting con-
stant of D/k= −16.990.07 K. The anisotropic g
value, as measured from the EPR spectrum, was fixed
at a value of g=2.20 throughout the fitting routine.
The agreement between the calculated and observed
susceptibility data and the values for the exchange
constants derived from the fitting suggest that within
this temperature range, the system exhibits predomi-
nant ferromagnetic interactions and that the material
consists of a low dimensional magnetic lattice. The
magnitudes of the intrachain coupling and zero-field
splitting constants are in fairly good agreement with
literature values reported for other linear chain Ni(II)
compounds [83–86]. The sign of the interchain coupling
constant J%/k= −1.8890.04 K suggests that this in-
teraction is antiferromagnetic in nature, which is ex-
pected from the low temperature behavior of the MT
data. The magnitude of this exchange constant is also
within the range observed for other Ni(II) linear chain
materials [83–86] and may suggest that the crystal
lattice for the material is fairly small, as mentioned
earlier, so as to allow for significant interchain contacts
between paramagnetic moieties.
The magnitude and signs of the exchange constants
derived from the fitting suggest that, as expected, ferro-
magnetic interactions predominate over this tempera-
ture range. The magnitude of the intrachain coupling
constant, which stems from the paramagnetism of the
TCNQ anions, is consistent with other Ni(II) deriva-
tives which contain S=1/2 TCNQ−1 anions [88]. The
sign and magnitude of the interchain coupling constant
suggest that overall this interaction is antiferromagnetic
in nature and is quite weak. However, due to the sharp
increase in MT at ca. 30 K may suggest the onset of
long range order due to the presence of a second
ferromagnetic pathway. Thus the nearly zero value
derived for the interchain coupling may well be the
consequence of nearly offsetting ferromagnetic and an-
tiferromagnetic contributions. In the absence of single
crystal structural data, the significance of the small
negative value for the nearest neighbor chain parameter
is difficult to assess. The aforementioned increase in
MT at 30 K and the onset of significant antiferromag-
netic interactions at low temperatures suggest that the
system may not be a true one-dimensional system at
any temperature. Attempts to fit the susceptibility data,
however, with several magnetic models that incorporate
two-dimensional interactions, proved unsuccessful.
4. Conclusions
An electrochemical method has been used to success-
fully synthesize two new nickel complexes of TCNQ.
Unlike solution methods, in which the oxidation state
of the organic anion is difficult to control, the electro-
chemical method appears to readily yield stable com-
pounds of various anion oxidation states as a function
of applied cell bias. The elemental, EDX, IR and XPS
data suggest that the stoichiometry of the products are
Ni(TCNQ) and Ni3(TCNQ)2 when the reaction is car-
ried out at 5 and 10 d.c V applied bias, respectively,
with Ni(II) in both cases and TCNQ in the −2 and
−3 oxidation states, respectively. This is in contrast to
the Ni(TCNQ)2 material obtained by Dunbar et al.
from solution preparations, which contains the
TCNQ−1 anion. In addition, minimization of the hy-
drolysis product DCTC is possible when appropriate
precautions are taken during the synthetic procedure.
The EPR spectrum of compound 2 at room tempera-
ture shows a single sharp line centered at g=2.000 with
a peak-to-peak line width of ca. 50 Oe as expected for
the S=1/2 TCNQ−3 anions. The diamagnetic nickel
ions are not expected to contribute to the EPR spec-
trum of the material, and no evidence of their contribu-
tion is present at detectable levels.
The susceptibility data above 35 K for compound 2
was fit to a S=1/2 one-dimensional ferromagnetic
chain Heisenberg model [87] based on the Hamiltonian
H= −2JijSi·Sj which incorporates a term for the num-
ber of equivalent nearest neighbor chains and a mean-
field correction term for the interchain interactions.
Non-linear curve fitting of the data to this model,
which gave an estimated goodness-of-fit parameter of

G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
13
infrared or elemental analysis. Clearly, single crystalline
samples are necessary to completely characterize this
material in the solid state.
The redox reaction of the TCNQ with the source
metal that occurs in the cell is thought to take place via
the following general reaction scheme:
Recent studies [41] by Dunbar et al. at Michigan
State University suggest that a similar magnetic behav-
ior to that of compound 2 is seen in solution prepared
nickel salts of TCNQ−1. Although no single crystal
structural data or magnetic modeling data are currently
available for these materials, the similarities of the
magnetic data suggest that the dramatic increase in the
ferromagnetic interactions at 30 K is due to the onset of
exchange via the S=1/2 TCNQ anions in the lattice.
Although TCNQ anions in the −1 and −3 oxidation
state are known to exhibit both ferro and antiferromag-
netic behavior via the ring p-electron density in the
solid state [94-98] usually through the formation of
dimeric units, the magnitude of the coupling is usually
on the order of a few Kelvins and is thus unlikely to be
solely responsible for the behavior seen in compound 2.
Ni(s)“Ni+2+2e−
TCNQ+ye−“TCNQ−y (y=2, 3)
Ni+2+TCNQ−2 or TCNQ−3“Ni(TCNQ)(s) or
Ni3(TCNQ)2(s)
where the reduction of the TCNQ takes place at the
surface of the substrate electrode. The metal ions and
the TCNQ anions then interact to form the insoluble
complex which deposit at the surface of the substrate
electrode.
Temperature dependent magnetic susceptibility mea-
surements suggest that both materials exhibit low di-
mensional behavior with dominant ferromagnetic
behavior at high temperatures with antiferromagnetic
exchange interactions that dominate at low tempera-
ture. The magnetism in 1 appears to be due to spin-spin
interaction between the paramagnetic Ni(II) ions medi-
ated by diamagnetic TCNQ−2 bridging ligands. For 2,
the TCNQ−3 ions act as strong field ligands, forcing
the Ni(II) ions to form diamagnetic square planar
complexes. Thus, the magnetism is due to interactions
between the spin 1/2 ground states of the TCNQ−3
anions. The high voltage product Ni3(TCNQ)2 exhibits
a sharp increase in MT at ca. 30 K that may suggest
the presence of additional antiferromagnetic interac-
tion. Due to the observed splitting of the w(CꢀN) bands
in the IR spectra of the materials, coordination of the
TCNQ anions to the metal center is possible, but in the
absence of single crystal structural data, it is difficult to
determine the geometries of such exchange pathways.
References
[1] National Research Council, Materials Science and Engineering
for the 1990s, National Academy Press, Washington DC, 1989,
p. 94.
[2] J.S. Miller, P.J. Krusic, A.J. Epstein, W.M. Reiff, J.H. Zhang,
Mol. Cryst. Liq. Cryst. 120 (1985) 27.
[3] J.S. Miller, J.C. Calabrese, H. Rommelmann, S.R. Chittipeddi,
J.H. Zhang, W.H. Reiff, A.J. Epstein, J. Am. Chem. Soc. 109
(1987) 769.
[4] S. Chittipeddi, K.R. Cromack, J.S. Miller, A.J. Epstein, Phys.
Rev. Lett. 58 (1987) 2695.
[5] J.S. Miller, A.J. Epstein, The interconnection of chemical and
electronic structure, in: Proceedings of the Nobel Symposium
NS-81 on Conjugated Polymers and Related Materials, Oxford
University Press, London, 1991, p. 461.
[6] J.M. Manriquez, G.T. Yee, R.S. McLean, A.J. Epstein, J.S.
Miller, Science 252 (1991) 1415.
,
Assuming a tetragonal cell with a short axis of ca. 3 A,
[7] J.S. Miller, G.T Yee, J.M. Manriquez, A.J. Epstein, The inter-
connection of chemical and electronic structure, in: Proceedings
of the Nobel Symposium NS-81 on Conjugated Polymers and
Related Materials, Oxford University Press, London, 1991, p.
461.
it is possible that the material may consist of NiꢀTCNQ
chains that run parallel to the long axis of the unit cell.
Stacking of these chains parallel to the short axis may
then allow for short ligand–ligand contacts which are
responsible for the observed magnetic behavior below
30 K. It is difficult, however, to envision the precise
connectivity of the metal–TCNQ−3 species in this or
any other unit cell so as to yield a square planar
geometry for the nickel ions given the predicted stoi-
chiometry. In order to provide such a geometry for
each of the three nickel ions suggested by the molecular
formula Ni3(TCNQ)2, a total of twelve coordinating
species are required. Even assuming that the nickel ions
are coordinated to the potentially bidentate di-
cyanomethylene groups of the TCNQ anions, the pre-
dicted stoichiometry only allows for eight such
interactions. The presence of other coordinated species
however, such as acetonitrile, is not supported by the
[8] T. Mallah, S. Thiebaut, M. Veraguer, P. Weillet, Science 262
(1993) 1554.
[9] J.S. Miller, A.J. Epstein, in: P. Coronado, D. Delhaes, J.S.
Gatteschi, D Miller (Eds.), Molecular Magnetism: From Molec-
ular Assemblies to the Devices (NATO ASI Series E), Kluwer,
Dordrecht, 1995.
[10] J.S. Miller, A.J. Epstein, New Aspects of Organic Chemistry,
VCH, New York, 1989, p. 237.
[11] A. Caneschi, D. Gatteschi, R. Sessoli, P. Rey, Acc. Chem. Res.
22 (1989) 392.
[12] O. Kahn, Y. Journaux, Inorganic Materials, Wiley, New York,
1993, p. 59.
[13] A.J. Epstein, in: J.S. Miller (Ed.), Synthesis and Properties of
Low Dimensional Materials, New York Academy of Sciences,
New York, 1978, p. 313.
[14] W.R. Entley, G.S. Girolami, Science 268 (1995) 397.
[15] S. Ferlay, T. Mallah, R. Ouahe`s, P. Veillet, M. Verdaguer,
Nature 378 (1995) 701.

14
G. Long, R.D. Willett / Inorganica Chimica Acta 313 (2001) 1–14
[16] J.S. Miller, D.A. Dixon, J.C. Calabrese, C. Vazquez, P.J. Krusic,
M.D. Ward, E. Wasserman, R.L Harlow, J. Am. Chem. Soc. 112
(1990) 381.
[17] P. Zhou, B.G. Morin, J.S. Miller, A.J. Epstein, Phys. Rev. Sect.
B 48 (1993) 1325.
[18] J.S. Miller, A.J. Epstein, Chem. Eng. News 73 (1995) 30.
[19] L.R. Melby, R.J. Harder, W.R. Hertler, J. Am. Chem. Soc. 84
(1962) 3374.
[51] J. Stanley, D. Smith, B. Latimer, J.P. Devlin, J. Phys. Chem. 70
(1966) 2011.
[52] C. Pecile, R. Bozio, A. Girlando, J. Chem. Soc. Faraday Trans.
71 (1975) 1237.
[53] S.E. Bell, J.S. Field, R.J. Haines, M. Moscherosch, W. Matheis,
W. Kaim, Inorg. Chem. 31 (1992) 3269.
[54] M. Harris, J.J. Hoagland, U. Mazur, K.W. Hipps, unpublished
results.
[20] R.S. Potember, T.O. Poehler, D.O. Cowan, P. Brant, Chem. Scr.
17 (1980) 219.
[21] R.S. Potember, T.O. Poehler, D.O. Cowan, A. Rappa, A.N.
Bloch, Synth. Met. 4 (1982) 371.
[22] R.S. Potember, T.O. Poehler, R.C. Benson, R.C. Hoffman, E.
Bourkoff, Appl. Phys. Lett. 42 (1983) 855.
[23] W.M. Risen, E.I. Kamitsos, Solid State Commun. 45 (1983) 165.
[24] W.M. Risen, E.I. Kamitsos, J. Chem. Phys. 79 (1983) 5808.
[25] E.J. Kamistos, C.H. Tzinis, W.M. Risen, Solid State Commun.
42 (1982) 561.
[26] R.C. Wheland, J.L. Gillson, J. Am. Chem. Soc. 98 (1976) 3916.
[27] E.I. Kamistos, W.M. Risen, Mol. Cryst. Liq. Cryst. 134 (1986)
31.
[28] S. Yamaguchi, C.A. Viands, R.S. Potember, J. Vac. Sci. Tech-
nol. B9 (1991) 1129.
[29] S.E. Bell, J.S. Field, R.J. Haines, Inorg. Chem. 31 (1992) 3269.
[30] B. Lanelli, C. Pecille, J. Chem. Phys. 52 (1970) 2375.
[31] J.S. Miller, A.H. Reis, E. Gebert, J.J. Ritsko, W.R. Salaneck,
T.W. Kovnat, R.P. Cape, J. Am. Chem. Soc. 101 (1979) 7111.
[32] A.R. Siedle, G.A. Candela, T.F. Finnegan, Inorg. Chim. Acta 35
(1979) 125.
[55] J.J Hoagland, PhD Thesis, Washington State University, 1992.
[56] A.R. Siedle, G.A. Candela, Inorg. Chem Acta 35 (1979) 125.
[57] M.S. Khatkale, J.P. Devlin, J. Chem. Phys. 70 (1979) 1851.
[58] R.G. Kepler, J. Chem. Phys. 39 (1963) 3528.
[59] A. Bieber, J. Andre, J. Chem. Phys. 5 (1974) 166.
[60] R.P. Shibeva, L.O. Atovmyan, L.P. Rozenberg, Chem. Com-
mun. (1969) 649.
[61] Y.M. Chow, D. Britton, Acta Crystallogr., Sect. B 30 (1974)
1117.
[62] Y.M. Chow, D. Britton, Acta Crystallogr., Sect. B 30 (1974) 147.
[63] M.T. Azcando, B. Loreto, L. Calderon, A. Gutierrez, M.F.
Perpinan, Polyhedron 14 (1995) 2339.
[64] J.P. Cornelissen, J.H. van Diemen, L.R. Groeneveld, J.G. Haas-
noot, A.L. Spek, J. Reedijk, Inorg. Chem. 31 (1992) 198.
[65] G.J. Leigh, J. Bremser, J. Chem. Soc. Dalton Trans. (1972) 1216.
[66] W.M. Riggs, Anal. Chem. 44 (1972) 830.
[67] S.O. Grim, L.J. Matienzo, J. Am. Chem. Soc. 94 (1972) 5116.
[68] P. Finn, Jolly, Inorg. Chem. 11 (1972) 1434.
[69] L.J. Matienzo, W.E. Swartz, S.O. Grim, Inorg. Nucl. Chem.
Lett. 8 (1972) 1085.
[70] D. Briggs, M.P. Seah, Practical Surface Analysis, Wiley, New
York, 1983.
[71] C.D. Wagner, Faraday Discuss. Chem. Soc. 60 (1975) 291.
[72] C.D. Wagner, H.L. Gale, R.H. Raymond, Anal. Chem. 51
(1979) 466.
[73] D. Briggs, M.P. Seah, Practical Surface Analysis, Wiley, New
York, 1983 Appendix 4.
[74] S.O. Grim, L.J. Matienzo, W.E. Swartz, J. Am. Chem. Soc. 94
(1972) 5116.
[75] L.O Pont, A.R. Siedle, Inorg. Chem. 13 (1974) 483.
[76] J.F. Weiker, L.R. Melby, R.E. Benson, J. Am. Chem. Soc. 86
(1964) 4329.
[77] C.D. Wagner, L.H. Gale, R.H. Raymond, Anal. Chem. 51
(1979) 466.
[78] R.L. Carlin, Magnetochemistry, Springer-Verlag, New York,
1986.
[79] C. O’Conner, Inorg. Chem. 29 (1982) 203.
[80] A.F. Cotton, Advanced Inorganic Chemistry, Wiley, New York,
1988.
[81] M.E. Fisher, Am. J. Phys. 32 (1964) 343.
[82] J. Kuo, Transforms and Curve Fitting, Wiley, New York, 1982.
[83] N. Achiwa, J. Phys. Soc. Jpn. 27 (1969) 561.
[84] C. Dupas, J. Renard, J. Chem. Phys. 61 (1974) 3871.
[85] J.F. Acherman, G.M. Cole, S.L. Holt, Inorg. Chem. Acta 8
(1974) 323.
[33] M.T. Azcondo, L. Ballester, L. Calderon, A. Gutierrez, F.
Perpinan, Polyhedron 14 (1995) 2339.
[34] J.S. Miller, A.J. Epstein, W.M. Reiff, Chem. Rev. 88 (1988) 201.
[35] J.S. Miller, A.J. Epstein, Prog. Inorg. Chem. 20 (1974) 1.
[36] D.S. Acker, W.R. Hertler, J. Am. Chem. Soc. 84 (1962) 3370.
[37] K. Nakatani, T. Sakata, Tsubomura, Bull. Chem. Soc. Jpn. 48
(1975) 2205.
[38] J. Gong, Y. Osada, Appl. Phys. Lett. 61 (1992) 2787.
[39] (a) J.J. Hoagland, X.D. Wang, K.W. Hipps, Chem. Mater. 5
(1993) 54. (b) L. Ballester, M.C. Barral, R. Gutierrez, J.M.
Jimenez-Aparicio, J.M. Martinez-Muyo, M.F. Perpinan, A.
Monge, C. Ruiz-Valero, J. Chem. Soc. Chem. Commun. (1991)
1396.
[40] L. Shields, J. Chem. Soc. Faraday Trans. II 81 (1985) 1.
[41] K. Dunbar, private communication.
[42] A.R. Siedle, J. Am. Chem. Soc. 97 (1975) 5931.
[43] J.C. Russ, Fundamentals of Energy Dispersive X-ray Analysis,
Butterworths, London, 1984.
[44] C.E. Fiori, A Procedure for Resolving Spectral Interferences in
Energy Dispersive X-ray Spectrometry, Special Technical Publi-
cation 604, National Bureau of Standards, US Department of
Commerce, 1981, p. 238.
[45] D. Appleman, USGS ACA Meeting, Cambridge, 1963.
[46] O. Kahn, Molecular Magnetism, VCH, New York, 1993.
[47] G.S. Long, unpublished results, Washington State University,
1995.
[86] A. Stebler, H.U. Gudel, A. Furrer, J.K. Kjems, Inorg. Chem. 21
(1980) 380.
[87] J.S. Miller, A.J. Epstein, Angew. Chem., Int. Ed. Engl. 33 (1994)
385.
[48] B. Lunelli, C. Pecille, J. Chem. Phys. 52 (1970) 2375.
[49] S.E. Bell, J.S. Field, R.J. Haines, M. Moscherosch, W. Matheis,
W. Kaim, Inorg. Chem. 31 (1992) 3269.
[50] M.S. Khatkale, J.P. Devlin, J. Chem. Phys. 70 (1979) 1851.
[88] H. Wang, Y. Chen, J. Li, Inorg. Chem. Acta 148 (1988) 261.
.