Structural, Electrical, Magnetic, and Spectroscopic Properties of Ring-Oxidized Molecular Metals Produced by Iodination of Metal-Free and Nickel Tetrabenzporphyrins

Article abstract of DOI:10.1021/ic961490s

Detailed studies of the structure, conductivitity, magnetoresistance, optical spectra, and magnetic properties (susceptibility, EPR) for the new molecular metal tetrabenzporphyrin iodide (H₂(tbp)I) and the electrical, spectral, and magnetic properties of Ni(tbp)I are reported. Paramagnetic transition-ion impurities were carefully excluded during the synthesis of H₂(tbp)I and Ni(tbp)I, and both materials show much higher, metal-like conductivites than previously seen for less-pure Ni(tbp)I. Comparison of the specular reflectance data for Ni(tbp)I and H₂(tbp)I allows a distinction between purely ring π-transitions and metal-involved charge-transfer transitions, and the spectra fix the energy levels of the π orbitals involved in conduction. Transport, magnetic, and optical properties show that both H₂(tbp)I and Ni(tbp)I are ring-based conductors that have metal-like conductivities, varying as ～1/T. down to ca. 30-40 K. However, the remaining level of defects is higher in the tbp conductors than in H₂(pc)I, and whereas the latter is metallic down to the mK temperature range, the defects in the (tbp) compounds localize the conduction electrons at ～10 K (Ni(tbp)I) and ～30 K (H₂(tbp)I), leading to transport through one-dimensional variable-range hopping. EPR g-values for H₂(tbp)I and Ni(tbp)I are close to that for the free electron and are nearly temperature-independent. The line widths for both samples are extremely narrow and also are nearly temperature-independent. These results show that Ni(tbp)I does not display doubly-mixed valence, as thought earlier: Paramagnetic impurities significantly altered the EPR signals of the prior samples. H₂(tbp)I crystallizes in the space group P4/mcc with cell constants of a = 14.173(10) A? and c = 6.463(4) A?. Full-matrix least-squares refinement of 63 variables gave an R index of 0.061 on F₀².

Full text of DOI:10.1021/ic961490s

Inorg. Chem. 1997, 36, 3363-3369
3363
Structural, Electrical, Magnetic, and Spectroscopic Properties of Ring-Oxidized Molecular
Metals Produced by Iodination of Metal-Free and Nickel Tetrabenzporphyrins
Kazuhiko Murata, Kwangkyoung Liou, Julia A. Thompson, Ellen M. McGhee,
Dean E. Rende, Donald E. Ellis, Ronald L. Musselman,* Brian M. Hoffman,* and
James A. Ibers*
Departments of Chemistry and Physics and Materials Research Center, Northwestern University,
Evanston, Illinois 60208-3113, and Department of Chemistry, Franklin & Marshall College,
Lancaster, Pennsylvania 17604-3003
X
ReceiVed December 13, 1996
Detailed studies of the structure, conductivitity, magnetoresistance, optical spectra, and magnetic properties
(susceptibility, EPR) for the new molecular metal tetrabenzporphyrin iodide (H2(tbp)I) and the electrical, spectral,
and magnetic properties of Ni(tbp)I are reported. Paramagnetic transition-ion impurities were carefully excluded
during the synthesis of H2(tbp)I and Ni(tbp)I, and both materials show much higher, metal-like conductivites
than previously seen for less-pure Ni(tbp)I. Comparison of the specular reflectance data for Ni(tbp)I and H2(tbp)I
allows a distinction between purely ring π-transitions and metal-involved charge-transfer transitions, and the
spectra fix the energy levels of the π orbitals involved in conduction. Transport, magnetic, and optical properties
show that both H2(tbp)I and Ni(tbp)I are ring-based conductors that have metal-like conductivities, varying as
∼
1/T, down to ca. 30-40 K. However, the remaining level of defects is higher in the tbp conductors than in
H2(pc)I, and whereas the latter is metallic down to the mK temperature range, the defects in the (tbp) compounds
localize the conduction electrons at ∼10 K (Ni(tbp)I) and ∼30 K (H2(tbp)I), leading to transport through one-
dimensional variable-range hopping. EPR g-values for H2(tbp)I and Ni(tbp)I are close to that for the free electron
and are nearly temperature-independent. The line widths for both samples are extremely narrow and also are
nearly temperature-independent. These results show that Ni(tbp)I does not display doubly-mixed valence, as
thought earlier: Paramagnetic impurities significantly altered the EPR signals of the prior samples. H2(tbp)I
crystallizes in the space group P4/mcc with cell constants of a ) 14.173(10) Å and c ) 6.463(4) Å. Full-matrix
2
least-squares refinement of 63 variables gave an R index of 0.061 on Fo .
Introduction
spectral properties are significantly modified. For example,
negative thermoelectric power for Co(pc)I (pc ) phthalocyanine)
We have extensively studied quasi one-dimensional molecular
When the central
metals and ligands are changed, the electrical, magnetic, and
suggests that Co(pc)I is a metal-centered conductor where the
metals based on porphyrin molecules.1
-22
2
23
Co dz orbital is the path for conduction electrons. In contrast,
positive thermoelectric power for Co(tbp)I shows that oxidized
tbp rings are the pathway for conduction electrons.24 This is
because it is easier to ring-oxidize (tbp) than (pc). We have
also used single-crystal polarized reflectance spectroscopy to
determine the energy levels of the π-orbitals involved with
conduction in the macrocycle-based conductor. The lowest
energy ligand-to-metal and metal-to-ligand charge-transfer
transitions are dependent on the nature of the central metal and
*
To whom correspondence should be addressed: R.L.M., Franklin &
Marshall College; B.M.H. and J.A.I., Northwestern University.
X
Abstract published in AdVance ACS Abstracts, April 15, 1997.
(
1) Hanack, M.; Roth, S.; Schier, H., Ed. Proceedings of the International
Conference on the Science and Technology of Synthetic Metals;
Elsevier Sequoia: Lausanne, Switzerland, 1991.
(
(
2) Cowan, D. O.; Wiygul, F. M., Chem. Eng. News 1986, 29, 28-45.
3) Ferraro, J. R.; Williams, J. M. In Introduction to Synthetic Electrical
Conductors; Ed.; Academic Press: New York, 1987; pp 139-204.
4) Williams, J. M.; Schultz, A. J.; Underhill, A. E.; Carneiro, K. In
Extended Linear Chain Compounds; Miller, J. S., Ed.; Plenum Press:
New York, 1982; pp 73-118.
5) Miller, J. S.; Epstein, A. J. Prog. Inorg. Chem. 1976, 20, 1-151.
6) Hoffman, B. M.; Ibers, J. A. Acc. Chem. Res. 1983, 16, 15-21.
7) Marks, T. J. Angew. Chem., Int. Ed. Engl. 1990, 29, 857-879.
8) Ogawa, M. Y.; Martinsen, J.; Palmer, S. M.; Stanton, J. L.; Tanaka,
J.; Greene, R. L.; Hoffman, B. M.; Ibers, J. A. J. Am. Chem. Soc.
(
(15) Quirion, G.; Poirier, M.; Liou, K. K.; Hoffman, B. M. Phys. ReV. B:
Condens. Matter 1991, 43, 860-864.
(16) Godfrey, M. R.; Newcomb, T. P.; Hoffman, B. M.; Ibers, J. A. J. Am.
Chem. Soc. 1990, 112, 7260-7269.
(
(
(
(
(17) Newcomb, T. P.; Godfrey, M. R.; Hoffman, B. M.; Ibers, J. A. Inorg.
Chem. 1990, 29, 223-228.
(18) Newcomb, T. P.; Godfrey, M. R.; Hoffman, B. M.; Ibers, J. A. J. Am.
Chem. Soc. 1989, 111, 7078-7084.
1
987, 109, 1115-1121.
(19) Almeida, M.; Kanatzidis, M. G.; Tonge, L. M.; Marks, T. J.; Marcy,
H. O.; McCarthy, W. J.; Kannewurf, C. R. Solid State Commun. 1987,
63, 457-461.
(
9) Ogawa, M. Y.; Hoffman, B. M.; Lee, S.; Yudkowsky, M.; Halperin,
W. P. Phys. ReV. Lett. 1986, 57, 1177-1180.
10) Liou, K.; Ogawa, M. Y.; Newcomb, T. P.; Quirion, G.; Lee, M.;
(
(20) Yakushi, K.; Yamakado, H.; Ida, T.; Ugawa, A. Solid State Commun.
1991, 78, 919-923.
Poirier, M.; Halperin, W. P.; Hoffman, B. M.; Ibers, J. A. Inorg. Chem.
1
989, 28, 3889-3896.
(21) Yakushi, K.; Sakuda, M.; Hamada, I.; Kuroda, H.; Kawamoto, A.;
Tanaka, J.; Sugano, T.; Kinoshita, M. Synth. Met. 1987, 19, 769-
774.
(
(
(
(
11) Quirion, G.; Poirier, M.; Liou, K. K.; Ogawa, M. Y.; Hoffman, B.
M.; Ibers, J. A. Phys. ReV. B 1988, 37, 4272-4275.
12) Pace, L. J.; Martinsen, J.; Ulman, A.; Hoffman, B. M.; Ibers, J. A. J.
Am. Chem. Soc. 1983, 105, 2612-2620.
(22) Yakushi, K.; Ida, T.; Ugawa, A.; Yamakado, H.; Ishii, H.; Kuroda,
H. J. Phys. Chem. 1991, 95, 7636-7641.
(23) Martinsen, J.; Stanton, J. L.; Greene, R. L.; Tanaka, J.; Hoffman, B.
M.; Ibers, J. A. J. Am. Chem. Soc. 1985, 107, 6915-6920.
(24) Liou, K. Y.; Newcomb, T. P.; Heagy, M. D.; Thompson, J. A.; Heuer,
W. B.; Musselman, R. L.; Jacobsen, C. S.; Hoffman, B. M.; Ibers, J.
A. Inorg. Chem. 1992, 31, 4517-4523.
13) Liou, K. K.; Jacobsen, C. S.; Hoffman, B. M. J. Am. Chem. Soc. 1989,
111, 6616-6620.
14) Ogawa, M. Y.; Palmer, S. M.; Liou, K. K.; Quirion, G.; Thompson,
J. A.; Poirier, M.; Hoffman, B. M. Phys. ReV. B 1989, 39, 10682-
10692.
S0020-1669(96)01490-5 CCC: $14.00 © 1997 American Chemical Society

3364 Inorganic Chemistry, Vol. 36, No. 15, 1997
Murata et al.
the identity of the macrocycle.25,26 In particular for Co, the p
π
specular reflectance data for Ni(tbp)I and H2(tbp)I allows a
distinction between purely ring π-transitions and metal-involved
charge-transfer (CT) transitions, and the spectra fix the energy
levels of the π-orbitals involved in conduction. Transport,
magnetic, and optical properties show that both H2(tbp)I and
Ni(tbp)I are ring-based conductors, and they exhibit high, metal-
like conductivities down to ca. 30-40 K. However, the
remaining level of defects is higher than in H2(pc)I, and whereas
the latter is metallic down to the mK temperature range, the
defects in the (tbp) compounds localize the conduction electrons
at low temperatures, leading to transport through one-dimen-
2
4
2
orbitals are higher in energy for Co(tbp)I than for Co(pc)I
π
and the HOMO for the former is a p orbital not the metal dz
orbital. Thus Co(tbp)I is a ring conductor as are the Ni(pc)I
and Cu(pc)I, rather than a metal-spine conductor as is Co(pc)I.
Recent calculations that compared Ni(tbp)I with several M(pc)I
2
7
π
compounds predict that the p levels should be lower in
Ni(tbp)I than in Ni(pc)I, raising the possibility that they may
be below the metal dz level and that Ni(tbp)I will be a metal-
2
spine conductor.
We have recently found that electrical and magnetic properties
of porphyrinic molecular metals are far more sensitive to organic
46
sional variable-range hopping. Ni(tbp)I does not show doubly-
2
mixed valence, as thought earlier, where both Ni dz orbitals
impurities, structural defects, and paramagnetic metal centers
than we had appreciated.28 In order to remove as many of these
and tbp rings are oxidized owing to the hopping of the
2
44
conduction electrons between the dz orbitals and the rings:
factors as possible, we revisited an early system, the phthalo-
cyanines (pc’s), obtaining much higher purity H2(pc)I and
Ni(pc)I. The maximum conductivity for single crystals of these
Paramagnetic impurities had significantly altered the EPR
signals of the prior samples.
samples is about 5-6 times larger than for those reported
Experimental Section
previously.2
9-32
These “clean” background conductors enabled
us to study the effect of impurities on the electrical and magnetic
properties.33 Organic impurities and structural defects act as
sources of random potential and lead to a localized electronic
Zn(tbp). Zinc acetate dihydrate was prepared from ultrapure zinc
shot (Johnson Matthey, 99.9999%) and freshly distilled acetic acid.
Zn(tbp) was synthesized by the template cyclization of isoindolinone-
3-acetic acid with zinc acetate dihydrate by utilizing the method of
Edwards et al.47 but excluding the purification process. Zn(tbp) was
extracted from the reaction mixture with pyridine and recrystallized
from 1-chloronaphthalene until the impurity peak appearing at 460 nm
in the optical spectrum was removed.
3
4-36
state.
Electrical conductivity is reduced in the localized
37-39
state and is limited by variable-range hopping.
Paramag-
II
II
II
netic metal impurities, such as Cu , Co , and Fe , also reduce
the conductivity of porphyrinic molecular metals but by a
different mechanism known as the Kondo effect.4
0-43
2
H (tbp). Metal-free tetrabenzporphyrin was obtained by the de-
We report here structural, electrical, UV-vis polarized
specular reflectance, and magnetic properties of the new
molecular metal H2(tbp)I, prepared in a manner similar to the
purer pc’s in order to avoid the contamination of paramagnetic
impurities. We also report a restudy of electrical and magnetic
properties for a sample of Ni(tbp)I having a much lower level
of paramagnetic impurities than before.44,45 Comparison of the
metalation of Zn(tbp) in cold concentrated sulfuric acid saturated with
hydrogen chloride gas. The sulfuric acid solution of Zn(tbp) was poured
onto ice in order to avoid the sulfonation of the tbp rings. The resulting
precipitate was washed with cold water and dried before extraction
with 1-chloronaphthalene with the use of a quartz Linstead hot-
extraction apparatus48 and a high-purity glass thimble (Whatman). The
material obtained after cooling the solution had a blue-purple luster,
2
and its optical solution spectrum showed no trace of Zn(tbp). H (tbp)
-3
was further purified by sublimation under vacuum (less than 10 Torr)
at 450 °C. The overall yield was 7.9% from isoindolinone-3-acetic
acid.
(
25) Heagy, M. D.; Rende, D. E.; Shaffer, G. W.; Wolfe, B. M.; Liou, K.;
Hoffman, B. M.; Musselman, R. L. Inorg. Chem. 1989, 28, 283-
286.
Ni(tbp). H
of 99.9985% NiCl
2
(tbp) was metalated with twice the stoichiometric amount
‚6H O (Johnson Matthey) in a 10% quinoline/1-
(
26) Rende, D. E.; Heagy, M. D.; Heuer, W. M.; Liou, K.; Thompson, J.
2
2
A.; Hoffman, B. M.; Musselman, R. L. Inorg. Chem. 1992, 31, 352-
chloronaphthalene solvent. The solution was refluxed in a fused-silica
vessel for 72 h and then cooled in an ice bath. The dark-purple Ni(tbp)
crystals that precipitated were removed by filtration and washed with
hot water and acetone. The progress of the metalation was followed
by UV-vis solution spectroscopy, and the reaction was repeated until
358.
(
27) Liang, X.; Flores, S.; Ellis, D. E.; Hoffman, B. M.; Musselman, R. L.
J. Chem. Phys. 1991, 95, 403-417.
(28) Thompson, J. A.; Murata, K.; Miller, D. C.; Stanton, J. L.; Broderick,
W. E.; Hoffman, B. M.; Ibers, J. A. Inorg. Chem. 1993, 32, 3546-
3553.
the peak at 666 nm in the UV-vis spectrum associated with H
was removed. The resulting Ni(tbp) was sublimed twice in vacuum at
450 °C. The yield from H (tbp) was 91%.
(tbp)I and Ni(tbp)I. Single crystals of H
grown by oxidizing H (tbp) and Ni(tbp), respectively, with a 20% molar
excess of I in 1-chloronaphthalene at ca. 200 °C in shielded fused-
2
(tbp)
(
(
(
(
29) Schramm, C. J.; Scaringe, R. P.; Stojakovic, D. R.; Hoffman, B. M.;
Ibers, J. A.; Marks, T. J. J. Am. Chem. Soc. 1980, 102, 6702-6713.
30) Martinsen, J.; Greene, R. L.; Palmer, S. M.; Hoffman, B. M. J. Am.
Chem. Soc. 1983, 105, 677-678.
2
H
2
2
(tbp)I and Ni(tbp)I were
31) Martinsen, J.; Palmer, S. M.; Tanaka, J.; Greene, R. C.; Hoffman, B.
M. Phys. ReV. B 1984, 30, 6269-6276.
2
2
32) Inabe, T.; Marks, T. J.; Burton, R. L.; Lyding, J. W.; McCarthy, W.
J.; Kannewurf, C. R.; Reisner, G. M.; Herbstein, F. H. Solid State
Commun. 1985, 54, 501-504.
silica H-tubes. A higher temperature is needed to dissolve tbp as
opposed to pc. The silica tubes have an extra horizontal tube to equalize
the level of the solvent. If too much iodine is used, then tbp is
(33) Murata, K.; Ohashi, Y.; Murata, K.; Thompson, J. A.; Mori, M.;
Hoffman, B. M. Synth. Met. 1993, 56, 1777-1782.
overoxidized. Shiny dark-green crystals of H
collected by filtration. Anal. Calcd for C36
N, 8.78. Found: C, 69.0; H, 3.49; N, 8.68. Calcd for C36
2
(tbp)I and Ni(tbp)I were
: C, 67.8; H, 3.49;
Ni:
(
(
(
(
34) Lee, P. A. ReV. Mod. Phys. 1985, 57, 287-337.
35) Nagaoka, Y. Prog. Theor. Phys. Suppl. 1985, 84, 1-15.
36) Anderson, P. W. Science 1978, 201, 307-316.
H22IN
4
H
20IN
4
C, 62.28; H, 2.91; N, 8.07. Found: C, 62.07; H, 2.78; N, 8.13 (Nippon
Shokubai Central Research Laboratory, Osaka, Japan).
37) Mott, N. F. Metal-Insulator Transitions, 2nd ed.; Taylor & Francis:
London, 1990.
(
38) Mott, N. F.; Kaveh, M. AdV. Phys. 1985, 34, 329-401.
X-ray Diffraction Study of H
2
(tbp)I. From Weissenberg X-ray
(39) Ambegaoker, V.; Halperin, B.; Langer, J. S. Phys. ReV. 1971, 4, 2612-
photographs, we assigned the H (tbp)I crystals to Laue group 4/mmm
2
2620.
of the tetragonal system. Systematic absences are consistent with the
space groups P4/mcc and P4cc. Agreement among Friedel pairs (Rint
(
(
(
40) Kondo, J. Prog. Theor. Phys. 1964, 32, 37-49.
41) Gruner, G.; Zawadowski, A. Rep. Prog. Phys. 1974, 37, 1497-1583.
42) Kondo, J. In Solid State Physics; Seitz, F., Turnbull, D., Ehrenreich,
H., Ed.; Academic Press: New York, 1969; pp 183-281.
43) Brandt, N. B.; Moshchalkov, V. V. AdV. Phys. 1984, 33, 373-468.
44) Martinsen, J.; Pace, L. J.; Phillips, T. E.; Hoffman, B. M.; Ibers, J. A.
J. Am. Chem. Soc. 1982, 104, 83-91.
(46) Bloch, A. N.; Weisman, R. B.; Varma, C. M. Phys. ReV. Lett. 1972,
(
(
28, 753-756.
(47) Edwards, L.; Weisman, R. B.; Varma, C. M. Phys. ReV. Lett. 1972,
28, 7638-7641.
(
45) Euler, W. B.; Martinsen, J.; Pace, L. J.; Hoffman, B. M.; Ibers, J. A.
(48) Barrett, P. A.; Dent, C. E.; Linstead, R. P. J. Chem. Soc. 1936, 1719-
Mol. Cryst. Liq. Cryst. 1982, 81, 231-242.
1736.

Ring-Oxidized Molecular Metals
Inorganic Chemistry, Vol. 36, No. 15, 1997 3365
Table 1. Crystal Data and Experimental Details for H
2
(tbp)I
Single-Crystal Polarized Reflectance Spectroscopy. Polarized
specular reflectance spectra were obtained with an instrument described
in detail elsewhere.51 Spectra were recorded from elongated faces
compd
formula
fw
tetrabenzoporphyrin iodide
C H N I
36 22 4
637.51
2
((100) and (010)) of both H (tbp)I and Ni(tbp)I with the electric vector
aligned either parallel or perpendicular to the needle (c) axis of the
crystals. Since the macrocycle planes are aligned perpendicular to the
c axis (see below), parallel and perpendicular to the needle axis
correspond to out of plane and in plane, respectively. For each spectrum
400 data points were collected between 215 and 850 nm. Collection
at each data point for each individual spectrum proceeded until the
sample mean had a 99% probability of being within 1% of the
population mean. For each sample, three or four individual spectra
were averaged once the data were corrected for percent reflectivity
a, Å
c, Å
V, Å
14.173(10)
6.463(4)
1298
3
Z
d
2
calc, g cm-
3
1.631(3)
2
space group
µ, cm
transm coeff
D
4h -P4/mcc
-
1
12.5
0.826-0.851
0.061
0.104
0.039
2
R on F
o
2
R
w
on F
o
52,53
relative to a NIST standard mirror. For Kramers-Kronig analysis,
reflectivities beyond our experimental region were estimated in the
2
2)
R on F
o
with F
o
> 3σ(F
o
infrared region from values for Ni(pc)I and Co(pc)I2
3,31
and were
)
0.025 after correction for absorption) and successful refinement of
2
approximated in the vacuum ultraviolet and beyond so as to produce
baselines approaching zero absorbance in regions known to have no
absorbance in related systems in the vapor state. Deconvolution was
carried out with an interactive Gaussian and Lorentzian program
the structure support the choice of space group D4h -P4/mcc. Least-
squares refinement of 14 reflections centered on a Picker diffractometer
yielded cell constants of a ) 14.173(10) Å and c ) 6.463(4) Å (λ(Mo
KR
1
) ) 0.7093 Å). In addition to the normal Bragg scattering, diffuse
(
“FastGauss”54 ).
scattering was observed, similar to that found in the Ni(tbp)I and Ni(pc)I
2
9,44
Electron Paramagnetic Resonance Measurements. Measurements
at room temperature and 77 K were carried out with the use of a
modified Varian E-4 X-band ESR spectrometer. Data recorded between
systems.
The diffuse scattering, observable on oscillation photo-
*
graphs, was found in planes perpendicular to the c axis with a spacing
*
2
*
of c ′ ) / c . No diffuse layers with l′ ) 3n were observed.
3
1.8 and 4.2 K were taken on powdered samples with a modified Varian
The data for the Bragg scattering were collected at 123 K with the
θ-2θ scan technique. No systematic changes were seen in the six
standard reflections that were measured every 100 reflections. The
E-109 Q-band ESR spectrometer for H (tbp)I or a Bruker ER 200D-
2
SRC X-band ESR spectrometer for Ni(tbp)I. Experiments were
performed with a silvered TE102 sample cavity in a Janis Corp. liquid-
helium immersion Dewar flask. The temperature was regulated by
controlling the He pressure inside the Dewar. For the calibration of
the field, 2,2-phenyl-1-picrylhydrazyl (DPPH), recrystallized from
benzene (g ) 2.0036 at 300 K, g ) 2.0025 at 4.2 K), was used.
2
2
final refinement on F
observations and 63 variables. In this refinement the positions of
the hydrogen atoms were idealized (C-H ) 0.95 Å) and each hydrogen
atom was given an isotropic thermal parameter 1 Å larger than that of
the carbon atom to which it is bonded. Experimental details and crystal
data are found in Table 1. Other crystallographic details are in the
Supporting Information.
o
0
converged to R(F ) ) 0.061 for 999
4
9
2
Computer simulations of the powder patterns generated by the program
Qpow5
5-57
were used to determine g-values and line widths.
Magnetic Susceptibility Measurements. Static magnetic suscep-
tibility measurements were taken from 300 to 1.9 K with a Quantum
Design SQUID susceptometer. Gelatin capsules were used to hold the
samples, and the background was measured over the full temperature
range just prior to measuring the sample. Data were recorded at 5 kG
on 20-40 mg samples.
Intensity data for the diffuse scattering were collected at 123 K point
by point along a line perpendicular to the diffuse planes. Two data
sets were collected: (1) every 0.05° in 2θ with a counting time of 100
s per step; (2) every 0.10° in 2θ with a counting time of 200 s per
step. The data were collected out to l′ ) 11. The l′ ) 1 peak was not
discernible from the background. The two different data sets were
averaged and corrected for absorption and Lorentz-polarization effects.
Results and Discussion
As in Ni(tbp)I, the apparent thermal parameter of the iodine atom
2
was larger along the chain (U
|
) 0.0957(5) Å ). The diffuse lines
Structure of H (tbp)I. Not surprisingly, the structure of
2
may be ascribed to the disorder in the iodine chains that leads to this
anomalous thermal motion. A reasonable model that accounts for this
44
H2(tbp)I is very similar to that of Ni(tbp)I. The macrocycles
are stacked in chains that run parallel to the c axis, with chains
of iodine atoms in the channels between the stacks. The two
H2(tbp) molecules within the unit cell are staggered by 40°.
The porphyrin macrocycles are necessarily planar because they
lie upon sites of 4/m symmetry. Figure 1 is a perspective
2
9
-
3
disorder proposes that there are ordered chains of symmetric I
ions
with repeat distance c′ ) 3c/2 but that the origins of these chains are
out of register. This lack of register gives rise to the observed diffuse
scattering. From an analysis of that scattering, we can obtain the value
-
of d, the displacement parameter, and hence the I-I distance in the I
3
ion. A least-squares fit to the previously derived expression2
9,50
drawing of the H (tbp) molecule. The thermal parameter of
2
the iodine atom in the stacking direction is much larger than
2
|
F (l)| )
that perpendicular to the stack. This evidence supports the
d
-
29
1
2
2
disordered I3 model for H2(tbp)I.
(1/9)[(1 + 2cos[2πl(d - / )]) ][9 - (1 + 2 cos[2πl/3]) ] (1)
3
The molecular structure is similar to that of Ni(tbp)I, but there
are some notable differences (Supporting Information). Re-
placement of the Ni center with protons causes the coordinating
pyrrole nitrogen atoms to move away from the center of the
ring. Thus, the angles around the nitrogen atoms are larger in
H2(tbp)I than in Ni(tbp)I, and the central hole of the macrocycle
is expanded.
yields d ) 0.0213(1). The observed and calculated values of the diffuse
structure amplitudes (Supporting Information) lead to an R index for
these intensities of 0.088.
Single-Crystal Electrical Conductivity Studies. Single crystals
2
of both H (tbp)I and Ni(tbp)I were mounted on graphite fibers with a
locally prepared palladium paste. Four-probe ac (27 Hz) conductivity
measurements were performed along the stacking axis. A Stanford
Research SR-530 two-channel lock-in amplifier was used to measure
the voltage drop across the sample in the differential mode. A Quantum
Design SQUID susceptometer was used to apply a transverse field up
to 5 T and to control the temperature between 300 and 1.85 K. All
(
51) Desjardins, S. R.; Penfield, K. W.; Cohen, S. L.; Musselman, R. L.;
Solomon, E. I. J. Am. Chem. Soc. 1983, 105, 4590.
(52) Anex, B. G. Mol. Cryst. 1966, 1, 1-36.
(53) Kronig, R. J. Opt. Soc. Am. 1926, 12, 547.
-6
-3
(54) Haddon, H.; Musselman, R. L. FastGauss, Macintosh Application,
Franklin and Marshall College, 1992.
crystals showed ohmic properties for the current between 10 and 10
A.
(
55) Belford, R. L.; Nilges, M. J., Eds. In Computer Simulation of Powder
Spectra; EPR Symposium, 21st Rocky Mountain Conference, Denver,
CO, 1979.
2
2
2
2
(
(
49) R(Fo ) ) Σ|Fo - Fc |/ΣFo .
50) Guinier, A. X-Ray Diffraction in Crystals, Imperfect Crystals and
Amorphous Bodies; W. H. Freeman: San Francisco, CA, 1963.
(56) Nilges, M. J. Ph.D. Dissertation, University of Illinois, 1979.
(57) Maurice, A. M. Ph.D. Dissertation, University of Illinois, 1980.

3366 Inorganic Chemistry, Vol. 36, No. 15, 1997
Murata et al.
Figure 1. Perspective view of the tbp macrocycle, hydrogen atoms
omitted.
2
Figure 3. Absorbance spectra of H (tbp)I from Kramers-Kronig-
transformed reflectance data: (a, top) in-plane polarized; (b, bottom)
out of plane polarized; s, experimental data; ---, Gaussian deconvo-
lution; ‚‚‚, residual.
at 24 × 10 cm as a transition from I3- only in the solid
state. Peaks H and H′ are similar to two relatively low-intensity,
high-energy peaks in H2(pc)I; neither of these peaks is predicted
3
-1
27,58
in recent theoretical treatments,
and these remain unassigned.
The in-plane spectrum for H2(tbp)I, Figure 3a, shows a Soret
region, peak B, and a greatly reduced Q band relative to that in
Figure 2. Polarized specular reflectance spectra of a single crystal of
26
H
)
2
(tbp)I: s ) electric vector perpendicular to (out of plane) and ---
H2(pc)I, consistent with porphyrin Q bands being much lower
in intensity relative to the Soret band than the phthalocyanine
parallel to (in plane) molecular planes.
5
9
Q bands. This spectrum is remarkably similar to the in-plane
2
4
Single-Crystal Reflectance Spectroscopy. Specular reflec-
spectrum for Co(tbp)I, and we assign the transitions as we
3
-1
*
tance spectroscopy was used to obtain polarizations of electronic
transitions, since the high absorbances made transmission
spectroscopy impractical. Comparison of the spectra of Ni(tbp)I
and H2(tbp)I is helpful in distinguishing metal-involved CT
transitions from purely ring π-transitions. There is a complica-
did in that system: Q at 15.3 × 10 cm as 2a (π) f 7e (π )
1
u
*
g
3
-1
and R at 17.4 × 10 cm as 5a (π) f 7e (π ). These two
1
u
g
-
1
transitions are calculated to be at 15 136 and 20 7120 cm for
2
7
Ni(tbp)I.
The intense peak in Figure 3a at 26 300 cm- corresponds
1
-
tion from three out-of-plane I3 transitions, one of which falls
to the Soret band in solution spectra of porphyrins. We recently
2
26,27
in the region of the lowest energy MLCT band, the dz f π,pz
assigned this band in M(pc)I systems
and in Co(tbp)I to be
25,26
24
*
transition in M(pc)I
and Co(tbp)I; the low-energy out-
primarily 6e (π) f 4b (π ) with minor contributions from
several much weaker transitions. Our calculations on Ni(tbp)I
show several transitions around 27 000 cm , the strongest being
g
2u
27
of-plane-polarized transitions can be differentiated with the help
of spectra from H2(tbp)I.
-1
*
H2(tbp)I. We show the reflectance spectra of H2(tbp)I in
Figure 2 and the Kramers-Kronig-transformed absorbance
spectra in Figure 3a,b. The transformed spectrum with the
electric vector parallel to the c axis and out of the complex
plane is shown in Figure 3b. As with all peaks in this compound
and in Ni(tbp)I, Gaussian deconvolution gave a better fit for
well-defined peaks than did Lorentzian deconvolution and was
thus used. The lowest energy out-of-plane peaks, D-F in Figure
6e (π) f 4b (π ). Others include 2b f 7e and 4a f 7eg.
g
2u
1u
g
2u
*
We thus assign peak B as primarily 6e (π) f 4b (π ) with
g
2u
*
minor contributions from 2b (π) f 7e (π ) and 4a (π) f
7e (π ). Note that this assignment differs from that of the four-
where the Soret transition has been assigned
as 4a (π) f 6e (π ), equivalent to the 5a (π) f 7e (π ) that
1
u
g
2u
*
g
6
0,61
orbital model,
*
*
1
u
g
1u
g
we assign to peak R. The structure at the top of peak B we
attribute to instrument noise since it falls within the background
2
6
3
b, match very closely the corresponding peaks in H2(pc)I,
-
all of which were assigned as arising from I3 , the actual form
of the reduced I species in the crystal. We thus assign these
three peaks as we did their counterparts in H2(pc)I: peak D at
(58) Liang, X. L.; Ellis, D. E.; Gubanova, O. V.; Hoffman, B. M.;
Musselman, R. L. Int. J. Quantum Chem. 1994, 52, 657-671.
59) Schaffer, A. M.; Gouterman, M.; Davidson, E. R. Theoret. Chim. Acta
(
1
973, 30, 9.
3
-1
1
1
1
×
9.95 × 10 cm as the Sg f Su* transition, peak F at 30.4
(
60) Edwards, L.; Gouterman, M. J. Mol. Spectrosc. 1970, 33, 292.
3
-1
10 cm as a spin-orbit component of peak D, and peak E
(61) Basu, S. Ind. J. Phys. 1955, 28, 511.

Ring-Oxidized Molecular Metals
Inorganic Chemistry, Vol. 36, No. 15, 1997 3367
Table 2. I
3
--Based UV-Vis Peaks
wavenumber
at max abs
max abs
intensity,
q (Å )
-
3
-1
(10 M cm-1
3
-1
2
2
compd
peak
(10 cm
)
)
H
2
(pc)I
D
E
F
D
E
F
D
E
F
19.65
23.95
29.95
19.95
24
30.4
20
22.848
4.61
5.97
22.85
4.85
9.36
12
4.6
3.3
6.2
4.6
3.4
7
4.2
3.5
6.5
H
2
(tbp)I
Ni(tbp)I
24.35
30.05
3.34
4.38
-
peaks from I3 . Peaks D-F have been identified as arising
Figure 4. Polarized specular reflectance spectra of a single crystal of
Ni(tbp)I: s, out of plane; ---, in plane.
1
1
from a Σg f Σu* transition, a spin-orbit coupling component,
and a solid-state induced component, respectively. These peaks
also appear in Figure 5b, but the Gaussian deconvolution needs
-
to be described because of the close proximity of the I3 and
Ni(tbp)I CT peaks. Peaks D and F have appeared within 400
-1
-1
26
cm of 20 000 and 30 000 cm , respectively, in both H2(pc)I
and H2(tbp)I, and it is reasonable to assume that their positions
-
will not change in Ni(tbp)I since the environment of I3 anion
is similar in all three compounds. It is also reasonable to assume
that the relative integrated intensities will remain nearly constant
in all three compounds. The reliability of intensity data is the
principal limitation of our specular reflectance measurements.
Both inaccuracies in the raw intensities and inherent uncertain-
ties in the reflectance data beyond our experimental region affect
the intensities. The practice of estimating far UV reflectance
52
with an “effective reflectance” is essentially sound. Although
the energies of peaks are unaffected by uncertainties in
unobservable reflectances, the slope of the transformed ab-
sorbance baseline, necessary for the estimation of relative
intensities, is principally affected by changes in the shape of
the effective reflectance. Table 2 lists several characteristics
-
of the I3 transitions in H2(pc)I and H2(tbp)I, as well as the
results arrived at for Ni(tbp)I from these intensity and energy
2
considerations. The ratio of q for peaks D:F is 3.54 in H2(pc)I
and 2.44 in H2(tbp)I. If no allowance were made for peaks A
and G in Ni(tbp)I, this same ratio would have the value 1.31.
-
1
When we place peak F at 30 000 cm and target an intensity
ratio between 3.5 and 2.4 to obtain the adjusted peaks D and F
2
listed in Table 2, we obtain a q ratio of 2.66. The estimated
deconvoluted peaks A and G that result from this adjustment
are shown in Figure 5b.
Peak A at 18 550 cm is analogous to a1g(dz ) f a2u(pz,π )
Figure 5. Absorbance spectra of Ni(tbp)I from Kramers-Kronig-
transformed reflectance data: (a, top) in-plane polarized; (b, bottom)
out of plane polarized; s, experimental data; ---, Gaussian deconvo-
lution; ‚‚‚, residual.
-
1
2
*
25
26
24
transitions found in Co(pc)I, Ni(pc)I, Cu(pc)I, and Co(tbp)I.
2
*
The energies of the M(pc)I a1g(dz ) f a2u(pz,π ) transitions are
3
-1
signals in the deconvolution. The remaining peaks N, L, and
V fall in the region where at least five moderately intense (0.14
e f e 0.51) transitions are predicted for Ni(tbp)I: 6eg f 3b1u,
14.5, 18.4, and 18.8 × 10 cm , respectively, following the
expected energies of metal orbitals. The energy of the same
-1
-1
transition in Co(tbp)I is 18 000 cm , a blue shift of 3500 cm
2
4
6
eg f 6a2u, 5eg f 3b1u, 2a1u f 9eg, and 2b1u f 8eg.
from that in Co(pc)I. This was interpreted as resulting from
a larger central cavity for tbp, causing a drop in the d orbital
and a rise in the π orbitals in tbp owing to the less
electronegative methylene carbon atoms in the pyrole-bridging
positions. In the present case of Ni(tbp)I there is only a slight
blue shift of 150 cm . The end result is that the a (d ) f
a (p ,π ) transition in Ni(tbp)I is 550 cm higher than that in
Presumably these five transitions contribute to much of the
intensity in this region. The labels are chosen to be consistent
with earlier treatments of porphyrin spectra.2
2
z
*
6,60
Ni(tbp)I. The reflectance spectra of Ni(tbp)I are shown in
Figure 4, and the transformed absorbance spectra are shown in
Figure 5a,b. The in-plane spectrum (Figure 5a) is essentially
identical to that for H2(tbp)I (Figure 3a), and as noted above,
our calculations for the Ni(tbp)I spectra agree with the experi-
mental in-plane peaks. We will therefore assign peaks Q, R,
S, B, N, L, and V to the same transitions as in H2(tbp)I. Peak
T is not predicted in our calculations and likely arises from a
-
1
2
1g
z
*
-1
2
u
z
Co(tbp)I, consistent with parallel results in M(pc)I compounds.
Nevertheless, from the comparison between the a (d ) f
2
1g
z
*
a (p ,π ) transitions in Co(tbp)I and Ni(tbp)I we conclude that
2
u
z
2
the Ni 3dz orbital lies below that of Co and hence that partial
oxidation occurs on the macrocycle ring.
-
1
monochromator aberration at 22 000 cm .
Electrical Conductivity. The single-crystal electrical con-
ductivity along the stacking axis at room temperature is in the
The out-of-plane spectra of Figure 5b contain the first metal-
related information in this set, but we first need to identify the
-
1
range of 300-1200 S cm for both H2(tbp)I and Ni(tbp)I. The

3368 Inorganic Chemistry, Vol. 36, No. 15, 1997
Murata et al.
Figure 6. Single-crystal conductivity as a function of temperature for
(tbp)I (O) and Ni(tbp)I (0).
H
2
Figure 7. Magnetoresistance, (F
H
- F
0
)/F, as a function of the square
broad range arises from uncertainties in the size of the crystals.
The conductivity normalized to that at room temperature,
however, is reproducible; it is shown in Figure 6 as a function
of temperature for both H2(tbp)I and Ni(tbp)I. The conductivi-
ties show similar temperature dependences. With decreasing
temperature, the conductivities increase as do metals (dσ/dT <
2
2
of the magnetic field, H : H (tbp)I (b); Ni(tbp)I (0).
should not be attributed to a Schubnikov-deHaas oscillation.
As shown in Figure 7 the magnetoresistance is linear with H ,
being negative at low fields and becoming positive at high fields.
The magnetoresistance can be interpreted as the sum of a
negative term that tends to saturate with increasing field and a
positive term that does not saturate. The negative contribution
2
0), reach a local maximum, and then decrease as semiconductors
do (dσ/dT > 0). The maximum normalized conductivity (σmax/
σ300) is 4.1 at 35 K for H2(tbp)I and is 8.9 at 9 K for Ni(tbp)I.
The conductivity ratios at 1.85 K are 2.3 and 6.8 for H2(tbp)I
and Ni(tbp)I, respectively. The maximum conductivity of
Ni(tbp)I reported here is much higher than that reported
previously (σmax/σ300 ) 1.8 at 95 K), and it occurs at a lower
temperature (Tmax(K)). This major improvement in the con-
ductivity of Ni(tbp)I is brought about by the decreased level of
3
7-39
is explained by the Kondo effect.
Because of the trace
(
<50 ppm) of Cu(tbp)I impurity, the conduction electrons are
II
2
2
scattered at low temperatures by the local moments in Cu dx -y
orbitals. In a magnetic field the direction of the local spins is
frozen and s-d exchange interaction between conduction
electrons and local spins is prevented. As a result the resistivity
in the magnetic field decreases. Positive quadratic magneto-
resistance is observed in many atomic metals. Such quadratic
dependence is explained by the drift velocity model that
considers two types of carriers, for example electrons and
II
II
impurities, especially of paramagnetic metals, such as Cu , Co ,
II
and Fe . Only a trace (<50 ppm) of Cu(tbp)I was detected by
II
EPR measurements on H2(tbp)I and Ni(tbp)I. Note that Ni is
diamagnetic in square-planar coordination.
6
2
holes. Negative magnetoresistance, explained by a change
of Fermi energy owing to the Zeeman effect, is often observed
The increase in conductivity with decreasing temperature
results from the reduction of phonon scattering found in typical
metals. The temperature dependence of the conductivity below
Tmax shows variable-range hopping. For one-dimensional
conductors, the conductivity follows the equation, σ R exp(T0/
6
3
for localized electronic systems. But this would not lead to
the magnitude of the negative magnetoresistance being related
to the amount of paramagnetic impurities, as is found here. The
positive effect is also expected in localized systems with
electron-electron interactions. But this does not apply here
as the magnitude of the positive magnetoresistance is not related
to the magnitude of localization.
-
1/2
T) , where T0(K) is a parameter close to Tmax. The conduc-
tivities of both H2(tbp)I and Ni(tbp) below Tmax are limited by
such hopping. The combination of the two contributions,
phonon scattering that dominates at high temperatures and
variable-range hopping that dominates at low temperatures,
results in the conductivity maximum. Because of variable-range
hopping the conduction electrons are localized at low temper-
atures. We recently reported that the conductivity of Ni(pc)I
increases as the defects in pc rings are reduced; the defects are
the source of random potential that causes a localized electronic
6
4
Magnetic Susceptibility. The variable-temperature static
susceptibilities for H2(tbp)I and Ni(tbp)I (Figure S1 in Sup-
porting Information) and the susceptibilities of H2(tbp) (-3.25
-
4
-1
-4
-1
× 10 emu mol ) and Ni(tbp) (-3.82 × 10 emu mol )
were measured for diamagnetic core correction. The correction
1
-
-4
-1
for /3 I3 is -1.15 × 10 emu mol from Pascal’s constant.
2
8
state. Similar arguments could be made for H2(tbp)I and
Ni(tbp)I where the crystal defects in tbp rings localize the
conduction electrons at low temperatures. But owing to the
limitations in the respective synthetic routes (see above) it has
not been possible to reduce crystal defects in tbp as much as
they can be in pc. Although the conducting behavior for Ni(pc)I
is very similar to that for H2(pc)I, the conductivity for Ni(tbp)I
is higher than that for H2(tbp)I because Ni(tbp)I contains fewer
impurities.
The total diamagnetic correction for H2(tbp)I and Ni(tbp)I is
-4
-4
-1
-4.40 × 10 and -4.97 × 10 emu mol , respectively. The
susceptibility for both samples remains almost constant from
room temperature to 30 K, indicative of a metal-like, temper-
ature-independent Pauli paramagnetism. Upon further cooling
of the sample the susceptibility increases rapidly and can be fit
approximately to an equation composed of a temperature-
independent Pauli-paramagnetism term and a Curie-Weiss term,
Magnetoresistance (∆/F0 ) (FH - F0)/F0) at 1.85 K is plotted
(62) Kittel, C. In Quantum Theory of Solids; Wiley: New York, 1987; pp
237-248.
2
as a function H , the square of the magnetic field, for H2(tbp)I
(
63) Fukuyama, H.; Yoshida, K. J. Phys. Soc. Jpn. 1979, 46, 1522-
and Ni(tbp)I in Figure 7. Both H2(tbp)I and Ni(tbp)I show
similar behavior. The data for Ni(tbp)I are noisy because a
single crystal of Ni(tbp)I is extremely fragile; this behavior
1
528.
(
64) Kamimura, H.; Takemori, T.; Kurobe, A. Anderson Localization;
Springer-Verlag: Berlin, 1982.

Ring-Oxidized Molecular Metals
ø ) ø + C(T - Θ)
Inorganic Chemistry, Vol. 36, No. 15, 1997 3369
-1
Table 3. Computer-Simulated EPR g Values and Line Widths for
H (tbp)I and Ni(tbp)I
2
(2)
T
P
-
1
sample
T (K)
g
⊥
g
|
Γ
⊥
(MHz, G)
Γ
|
(MHz, G)
where T is temperature, C (emu K mol ) is the Curie constant,
65
and Θ (K) is the Weiss temperature. The resultant values for
H
2
(tbp)I 300
2.0026 2.0045
2.0031 2.0059
6, 2.1
8, 2.9
15, 5.4
20, 7.1
25, 8.9
25, 8.9
10, 3.6
13, 4.6
25, 8.9
25, 8.9
-
1
77
4
1
øP (emu mol ), C, and Θ are as follows: H2(tbp)I, 3.84 ×
.2 2.0012 2.0050
.9 2.0010 2.0044
2.0010 2.0070
10, 3.6
10, 3.6
10, 3.6
13, 4.6
15, 5.4
15, 5.4
-
4
-3
-4
1
0 , 4.28 × 10 , and -4.13; Ni(tbp)I, 3.75 × 10 , 1.01 ×
-3
1
0 , and -0.33. The Curie constant represents the localized
Ni(tbp)I 300
77
spins; these arise from localized conduction electrons and from
2.0020 2.0090
II
2
2
localized spins in Cu dx -y orbitals. But since the amount of
Cu(tbp)I impurity is small (<50 ppm), in the present instance
the Curie constant measures the strength of localization. The
larger Curie constant indicates that the conduction electrons are
more localized in H2(tbp)I than in Ni(tbp)I, probably owing to
the larger amount of crystal defects or organic impurities. The
result is consistent with the conductivity measurements that
show H2(tbp)I to have poorer conductivity than does Ni(tbp)I.
The Weiss temperature can be analyzed in a similar way. The
negative Weiss temperature indicates antiferromagnetic coupling
among localized conduction electrons. The larger negative
Weiss temperature for H2(tbp)I once again demonstrates a larger
number of localized electrons. The temperature-independent
Pauli paramagnetism is almost identical for H2(tbp)I and
Ni(tbp)I. According to the one-dimensional tight-binding
model, the transfer integral t (erg) is written as
4.2 2.0025 2.0140
.9 2.0025 2.0140
1
in the rings but also hopping into Ni orbitals at low temperatures.
-
Similar results were seen for Ni(pc)Y , Y ) SbF and
0
.5
6
-
70
AsF . Our new samples of Ni(tbp)I, from which paramag-
6
netic transition-ion impurities such as Cu(tbp) have been
carefully excluded, give totally different results. Values of the
g|, g⊥ , Γ|, and Γ⊥ , the parallel and perpendicular components
of g and of the line width, as determined by computer simulation
of the signals, are summarized in Table 3.
Both H2(tbp)I and Ni(tbp)I show similar behavior. The g
values are close to the free-electron value (g ) 2.0023), are
almost temperature-independent, and are similar to those values
in H2(pc)I and Ni(pc)I. The slightly larger value for g| than
for g⊥ results from back-electron-transfer of electron density
from iodine to the ring. The line widths for H2(tbp)I and
Ni(tbp)I are extremely narrow at room temperature and increase
only slightly as the temperature is lowered. The narrow line
widths are typical for one-dimensional conductors because the
direction of spin-lattice relaxation is restricted and slower. The
slight increase in the line widths probably results from the
presence of otherwise undetectably small amounts of paramag-
netic impurities. In these systems the g values and line widths
are described by the susceptibility-weighted average of those
for the exchange-coupled conduction electrons, with tempera-
ture-independent susceptibility, and the Curie-like unpaired d
electrons; thus the influence of the latter increase as the
temperature is decreased. The unexceptional behavior of the
EPR variables of the new materials and the absence of a
significant differences between H2(tbp)I, necessarily a ring-
centered conductor, and Ni(tbp)I show that Ni(tbp)I is not a
doubly mixed-valence conductor but rather is a simple ring-
2
-1
t ) Nâ [ø π sin(πF/2)]
(3)
P
23
-1
where N is Avogadro’s number (6.022 × 10 molecules mol ),
-
21
-1
â is the Bohr magneton (9.273 × 10
erg G ), and F is the
1
degree of oxidation. From the values of øP above and F ) /3,
we obtain bandwidths (4t) of 0.214 and 0.219 eV for H2(tbp)I
and Ni(tbp)I, respectively. These values are smaller than those
for pure H2(pc)I (0.296 eV) and Ni(pc)I (0.381 eV). There is
no clear relationship, though, between the bandwidth and
the strength of localization.
Electron Paramagnetic Resonance. A free radical signal
(
g ∼ 2) is observed for unoxidized H2(tbp) and Ni(tbp). The
origin of a free radical signal in diamagnetic porphyrins has
been the subject of several papers.6
6-68
We confirm that the
signal originates from trapped charge carriers at structural
defects or impurity sites.69 The much larger signal for the
precursors H2(tbp) and Ni(tbp) than for H2(pc) and Ni(pc)
suggests that tbp has more defects or impurities, and it is these
that make H2(tbp) and Ni(tbp)I less conducting than H2(pc)I
and Ni(pc)I at low temperatures.
II
oxidized conductor with a diamagnetic Ni center and that the
same is true for the M(pc)Yx.
7
0
Our earlier EPR study44,45 of what we now know was impure
Ni(tbp)I reported the following: (1) The powder EPR signal
for Ni(tbp)I at room temperature showed a roughly symmetric
line with the unusally large isotropic g value of gav ) 2.024.
Acknowledgment. We thank the National Science Founda-
tion for funding (Grants DMR-9523228 (B.M.H.), CHE-
8
911215 (R.L.M.), and DMR-9632472 (the MRL Program at
the Materials Research Center of Northwestern University
J.A.I., B.M.H.)). R.L.M. also thanks the the donors of the
(
2) gav increased from 2.024 at 295 K to 2.062 at 77 K. (3)
(
The peak-to-peak line width at room temperature, Γ ) 105 G,
was also unusually large. (4) The line width further increased
from 105 G at 295 K to 320 G at 77 K. From these results we
concluded that the [Ni(tbp)] cation exhibited an electronic
tautomerism with interconversion between [Ni (tbp)] and
Petroleum Research fund of the American Chemical Society
for partial funding (Grant 25966-B).
+
III+
Supporting Information Available: Tables of positional and
displacement parameters, calculated and observed diffuse amplitudes,
bond distances and angles, and a metrical comparison of H
Ni(tbp)I and Figure S1, showing the temperature dependence of the
magnetic susceptibility for H (tbp)I and Ni(tbp)I (5 pages). Ordering
II
+
[
Ni (tbp) ]. Thus, we believed Ni(tbp)I to be a doubly mixed-
valence conductor, where conduction electrons were not only
2
(tbp)I and
2
(
65) For the purposes of this paper a more detailed analysis is not warranted.
For example, the temperature-dependent part could be described by a
information is given on any current masthead page.
-
γ
term dT
.
(
(
66) Assour, J. M.; Harrison, S. E. J. Phys. Chem. 1964, 68, 872-876.
67) Raynor, J. B.; Robson, M.; Torrens-Burton, A. S. M. J. Chem. Soc.
IC961490S
1
977, 2360-2364.
68) Harbour, J. R.; Loufty, R. O. J. Phys. Chem. Solids 1982, 43, 513-
20.
69) Jerome, D.; Schultz, H. J. In Studies in Physics; 1982; pp 299-490.
(
(
(70) Yakushi, K.; Yamakado, H.; Yoshitake, M.; Kosugi, N.; Kuroda, H.;
Sugano, T.; Kinoshita, M.; Kawamoto, A.; Tanaka, J. Bull. Chem.
Soc. Jpn. 1989, 62, 687-696.
5