Neutral complexes of first row transition metals bearing unbound thiocyanates and their assembly on metallic surfaces

Article abstract of DOI:10.1021/ja055459d

A series of transition metal coordination complexes designed to assemble on gold surfaces was synthesized, their electronic structure and transitions analyzed, and their magnetic properties studied. By taking advantage of recently developed thiocyanate assembly protocols, these molecules were then assembled onto a gold surface, without the need for an inert atmosphere, to give a loosely packed monolayer. The assembled molecules exhibit properties similar to that of the bulk molecules, indicating little change in molecular structure outside of chemisorption.

Full text of DOI:10.1021/ja055459d

Published on Web 02/18/2006
Neutral Complexes of First Row Transition Metals Bearing
Unbound Thiocyanates and Their Assembly on Metallic
Surfaces
Jacob W. Ciszek,† Zachary K. Keane,‡ Long Cheng,† Michael P. Stewart,†
Lam H. Yu,‡ Douglas Natelson,*,‡ and James M. Tour*,†
Contribution from the Department of Chemistry and Smalley Institute for Nanoscale Science and
Technology, and Department of Physics and Astronomy, Rice UniVersity, 6100 Main Street,
Houston, Texas 77005
Received August 24, 2005; E-mail: natelson@rice.edu; tour@rice.edu
Abstract: A series of transition metal coordination complexes designed to assemble on gold surfaces was
synthesized, their electronic structure and transitions analyzed, and their magnetic properties studied. By
taking advantage of recently developed thiocyanate assembly protocols, these molecules were then
assembled onto a gold surface, without the need for an inert atmosphere, to give a loosely packed
monolayer. The assembled molecules exhibit properties similar to that of the bulk molecules, indicating
little change in molecular structure outside of chemisorption.
Introduction
metal-molecule-metal junctions. These analyses supported the
concept of molecular devices and also provided a bridge
connecting traditional synthetic and analytical chemistry with
molecular electronics.
Watershed events for the molecular electronics community
were the measurements of single molecule conductance by
several groups.1-7 Following these demonstrations, scientists
began to propose and measure electronic behavior attributed to
single or small ensembles of molecules. Recent significant
contributions to this field include papers on the Kondo effect8-10
and inelastic electron tunneling spectroscopy (IETS)11-14 in
individual molecules. These papers explored quantum
phenomena15-18 in individual molecules that were part of
We have extended this work by synthesizing a new series of
metal-containing molecules. The purpose of the metal core is
two-fold. First, one can discern which electronic effects in the
junction are due to the metal atom core by varying the metal.
Second, one can study the effect of different spin states on
molecular conductance. Here we describe the efficient synthesis
of these molecules and the simple techniques necessary for
assembling them along with detailed characterization data of
the assembled compounds on gold surfaces. From these data
we can determine with high confidence the nature of the
molecules after assembly.
The molecular system described here is of particular interest
as it has been examined by both the theoretical and experimental
branches of science. These compounds have been studied in
electromigration-fabricated single molecule transistors (SMT)
(Figure 1a). When examined in SMTs13,19,20 these molecules
have conductance features characteristic of the Kondo effect, a
coherent many-body state comprising an unpaired spin on the
molecule coupled by exchange to the conduction electrons of
the leads.13,20 Kondo temperatures in excess of 50 K were found,
comparable to those in a purely metallic system.20 In addition
to the Kondo resonance some of these SMTs exhibit inelastic
co-tunneling features that correspond energetically to vibrational
excitations of the molecule, as determined by Raman and
infrared spectroscopy.13 This is a form of IETS of single
† Department of Chemistry and Smalley Institute for Nanoscale Science
and Technology.
‡ Department of Physics and Astronomy.
(1) Reed, M. A.; Zhou, C.; Muller, C. J.; Burgin, T. P.; Tour, J. M. Science
1997, 278, 252-254.
(2) Bumm, L. A.; Arnold, J. J.; Cygan, M. T.; Dunbar, T. D.; Burgin, T. P.;
Jones, L, II; Allara, D. L.; Tour, J. M.; Weiss, P. S. Science 1996, 271,
1705-1707.
(3) Dorogi, M.; Gomez, J.; Osifchin, R.; Andres, R. P.; Reifenberger, R. Phys.
ReV. B: Condens. Matter 1995, 52, 9071-9077.
(4) Andres, R. P.; Bein, T.; Dorogi, M.; Feng, S.; Henderson, J. I.; Kubiak, C.
P.; Mahoney, W.; Osifchin, R. G.; Reifenberger, R. Science 1996, 272,
1323-1325.
(5) Joachim, C.; Gimzewski, J. K. Europhys. Lett. 1995, 30, 409-414.
(6) Joachim, C.; Gimzewski, J. K.; Schlittler, R. R.; Chavy, C. Phys. ReV.
Lett. 1995, 74, 2102-2105.
(7) Fischer, C. M.; Burghard, M.; Roth, S.; Klitzing, K. v. Appl. Phys. Lett.
1995, 66, 3331-3333.
(8) Park, J.; Pasupathy, A. N.; Goldsmith, J. I.; Chang, C.; Yaish, Y.; Petta, J.
R.; Rinkoski, M.; Sethna, J. P.; Abruna, H. D.; McEuen, P. L.; Ralph, D.
C. Nature 2002, 417, 722-725.
(9) Liang, W.; Shores, M. P.; Bockrath, M.; Long, J. R.; Park, H. Nature 2002,
417, 725-729.
(10) Yu, L. H.; Natelson, D. Nano Lett. 2004, 4, 79-83.
(11) Kushmerick, J. G.; Lazorcik, J.; Patterson, C. H.; Shashidhar, R.; Seferos,
D. S.; Bazan, G. C. Nano Lett. 2004, 4, 639-642.
(12) Wang, W.; Lee, T.; Kretzschmar, I.; Reed, M. A. Nano Lett. 2004, 4, 643-
646.
(13) Yu, L. H.; Keane, Z. K.; Ciszek, J. W.; Cheng, L.; Stewart, M. P.; Tour,
J. M.; Natelson, D. Phys. ReV. Lett. 2004, 93, 266802.
(14) Stipe, B. C.; Rezaei, M. A.; Ho, W. Science 1998, 280, 1732-1735.
(15) Kondo, J. Prog. Theor. Phys. 1964, 32, 37-49.
(16) Anderson, P. W. Phys. ReV. 1961, 124, 41-53.
(17) Jaklevic, R. C.; Lambe, J. Phys. ReV. Lett. 1966, 17, 1139-1140.
(18) Adkins, C. J.; Phillips, W. A. J. Phys. C: Solid State Phys. 1985, 18, 1313-
1346.
(19) Natelson, D.; Yu, L. H.; Ciszek, J. W.; Keane, Z. K.; Tour, J. M. Chem.
Phys. 2006. In press
(20) Yu, L. H.; Keane, Z. K.; Ciszek, J. W.; Cheng, L.; Tour, J. M.; Baruah,
T.; Pederson, M. R.; Natelson, D. Phys. ReV. Lett. 2005, 95, 256803.
9
10.1021/ja055459d CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 3179-3189
3179

A R T I C L E S
Ciszek et al.
plexation, ligand exchange, or complexation of trace species
such as water. Occupation of all coordination sites is also
essential as these molecules are designed to contain a moiety
(sulfur) for binding to metal surfaces in the presence of a metal
core. If there were free coordination sites they could lead to
complexation of the sulfur or cyanide with the metal core. The
two tridentate ligands therefore protect the molecule from self-
decomposition.
Experimental Section
Figure 1. (a) SEM image of a junction used in the testing of the molecular
devices.13,19,22 Current transport likely occurs through the small area at the
top portion of the junction where the gap distance is similar to the molecular
dimensions. When current transport occurs through a sufficiently small area,
it is dominated by a single molecule. (b) Targeted molecular systems
synthesized for this study.
Materials. The syntheses of all compounds can be found in the
Supporting Information, along with characterization including: 1H, 13C,
IR (powder and surface), magnetic susceptibility, EPR, CV, UV-vis,
DSC, TGA, MS, elemental analysis, and XPS (powder and surface).
Ethanol (100%) was obtained from Pharmco. THF was freshly distilled
(sodium, benzophenone) prior to use. Dichloromethane was distilled
from calcium hydride. All metals were of 99.99% or greater purity.
Silicon 100 wafers (p doped, test grade) were obtained from Silicon
Quest International. Gold coated mica was purchased from Molecular
Imaging.
EPR. All EPR samples were prepared by dissolving the metal
complexes in freshly distilled THF (ca. 10 µM, 2; 100 µM, 4). Solutions
were passed through a 0.2 µm nylon filter and then quickly frozen.
Spectra were taken in a Varian E6 X-band spectrometer with an Air
Products helium cryogenic system.
Magnetic Susceptibility. Magnetic susceptibility measurements were
taken in a Superconducting Quantum Interference Device (SQUID).
Analysis was carried out at temperatures from 300 to 1.7 K using a
Quantum Design MPMS-5S magnetometer.
Cyclic Voltammetry. Electrochemical characterization was carried
out with a Bioanalytical Systems (BAS CV-50W) analyzer. The
reference was a Ag|AgNO3 electrode. The counter electrode was a clean
Pt wire. Redox potentials were determined at the specified scan rates.
The solutions studied were 1 mM with 0.1 M tetrabutylammonium
tetrafluoroborate as the supporting electrolyte in THF under inert
atmosphere.
UV-Visible Absorbance. UV-vis spectra were acquired on a
Shimadzu UV-3101PC containing both a deuterium and halogen light
source. All spectra were collected in freshly distilled dichloromethane.
Preparation of Evaporated Metal Substrates. All metal substrates
were prepared by thermal evaporation with a base pressure of ∼1 ×
10-6 Torr on test grade 100 silicon wafers with a typical evaporation
rate of 1 Å/s. A chromium adhesion layer (5 nm) was applied to all
substrates prior to deposition of 150 nm of the metal. All substrates
were used within 20 min after removal from the vacuum chamber.
Assembly Procedure. Molecules were assembled onto freshly
prepared substrates in freshly distilled dichloromethane or THF at a
concentration of 0.5-1.0 mM. Assemblies typically proceeded for 24
h in the dark, after which they were vigorously rinsed with the same
solvent used for assembly and dried under a stream of nitrogen.
Ellipsometric Measurements. Measurements of surface optical
constants and molecular layer thicknesses were taken with a single
wavelength (632.8 nm laser) LSE Stokes Ellipsometer (Gaertner
Scientific). The surface thickness was modeled as a single absorbing
layer atop an infinitely thick substrate (fixed ns). The observed error
in repeated measurements of the same spot was typically 0.2 nm or
less. The index of refraction was set at 1.55.
Figure 2. Schematic of the assembly on gold.
molecules, with transistor geometry allowing in situ tuning of
the electronic states via a gate electrode. The molecules have
also been modeled using the Green function approach and have
been predicted to exhibit good hole conduction.21
The compounds 1-5 (Figure 1b) were synthesized with
several parameters in mind. First, these molecules were designed
to have a simple assembly protocol on metal where they are
stable to air yet active toward a metal surface without the use
of exogenous reagents. We have taken advantage of our recently
published thiocyanate assembly procedure, in which the thio-
cyanate is converted to a thiolate via the gold surface, and the
cyanide leaves as [Au(CN)2]-.22,23 Thus, the metal complexes
spontaneously assemble on a clean gold surface (Figure 2). Use
of the thiocyanates eliminates the need for the oxidatively
unstable aromatic free thiols or their thioacetate precursors.24-26
When designing these molecules we wanted to avoid parasitic
charges that may interfere with the solid-state electrical studies
of the molecules as SMTs. Thus the system was designed to be
overall neutral (no counterions). To facilitate molecular conduc-
tion, the molecules were designed to be small and contain a
continuous π-electron system that is interrupted only by the core
in order to force electrons through the metal atom. Finally, the
molecules were designed such that all coordination sites on the
metal atom were bound by multi-dentate ligands. This is
important since it minimizes ligand dissociation and retards
effects that may influence electrical studies, such as decom-
(21) Yan, L.; Seminario, J. M. J. Phys. Chem. A 2005, 109, 6628-6633.
(22) Ciszek, J. W.; Stewart, M. P.; Tour, J. M. J. Am. Chem. Soc. 2004, 126,
13172-13173.
Infrared Reflection Spectroscopy Measurements. A Nicolet Nexus
670 spectrometer with KBr beam splitter and MCT/A detector was
used in conjunction with a SMART/SAGA grazing angle accessory,
using a 16 mm diameter sampling area with p-polarized light fixed at
an 80° angle of incidence.
X-ray Photoelectron Spectroscopy (XPS) Measurements. A PHI
Quantera SXM XPS/ESCA system at 5 × 10-9 Torr was used to take
photoelectron spectra. A monochromatic Al X-ray source at 100 W
(23) Ciszek, J. W.; Tour, J. M. Chem. Mater. 2005, 17, 5684-5690.
(24) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.;
Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335.
(25) Tour, J. M.; Jones, L., II; Pearson, D. L.; Lamba, J. J. S.; Burgin, T. P.;
Whitesides, G. M.; Allara, D. L.; Parikh, A. N.; Atre, S. V. J. Am. Chem.
Soc. 1995, 117, 9529-9534.
(26) Stapleton, J. J.; Harder, P.; Daniel, T. A.; Reinard, M. D.; Yao, Y.; Price,
D. W.; Tour, J. M.; Allara, D. L. Langmuir 2003, 19, 8245-8255.
9
3180 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Synthesis and Assembly of Transition Metal Complexes
A R T I C L E S
Scheme 2. Formation of the Metal Complex Test Cases
Scheme 1. Synthesis of the Simple Ligand 8
was used with an analytical spot size of 0.15 mm × 1.4 mm and a 45°
takeoff angle, with a pass energy of 26.00 eV. High-resolution spectra
of the S 2p region used a 45° takeoff angle and 13.00 eV pass energy.
Unless noted, surface samples were referenced against the internal Au
4f7/2 line at 84.0 eV. Crystalline powder samples were referenced
externally to an Au 4f peak at 84.00 eV and internally to a C 1s binding
energy of 284.50 eV (NIST XPS database). Atomic concentration values
were calculated with PHI Multipak software using factory-calibrated
values for the sensitivity factors of the respective elements.
Computational Procedures. Theoretical calculations were per-
formed by using the Gaussian 9827 program. The calculations were
performed on the crystallographic structures. The energy calculation
was performed using the reduced Hartree-Fock method with the SDD
basis set.28,29 The output was visualized with gopenmol.30,31
Scheme 3. Synthesis of Thiocyanate-Containing Ligandsa
Synthesis
The basic design was to have a metal(II) atom surrounded
by two anionic ligands (Figure 1b). Hence, a ligand such as 8
was a natural choice (Scheme 1). Although this ligand had been
synthesized previously32 and even complexed on one occasion,33
we found all previous methods to be impractical for large-scale
synthesis. Thus, our preferred route to 8 begins with succinic
acid that was converted to the bis-acetimide 6 in 89% yield
using a slightly modified version of Schindlbauer’s method.34
This was then converted to the bis-pyridal system 7 with a
modification of conditions outlined by Owsley.35 Cyclization
using ammonium acetate32 gave 8. This ligand was deprotonated
and reacted with ZnCl2 and CoCl2 to produce 9 and 10,
respectively, as test cases for the complexation (Scheme 2).
With complexation successfully demonstrated, the possible
methods for attachment of the molecules to metal surfaces were
examined. The 4-positions of the pyridine rings and 3,4-positions
of the pyrrole ring both suggest themselves as viable sites for
functional groups that will attach to the surface.36 We focused
on substitution at the 3- and 4-positions of the pyrrole ring.
a Solvent conditions determine whether the mono- or dithiocyanate
species is preferentially formed.
Thiocyanates were chosen as surface-binding precursors since
they offer many of the advantages of the thiol on gold chemistry
(such as rearrangement on a surface,37 an important factor for
such a bulky molecule) without the oxidative instability found
in the free thiols.25 This system has the advantage of not
requiring any extraneous materials or inert conditions, and gives
the same structure on gold as an assembly from a free thiol
(Figure 2).
Accordingly, the mono- and dithiocyanate functionalized
ligands 11 and 12 were synthesized by treating the bare ligand
with thiocyanogen ((SCN)2, generated from Pb(SCN)2 and Br2).
In a mixture of 9:8 CH2Cl2/THF the monothiocyanate 11 was
formed preferentially, whereas in a more polar mixture (2:5 CH2-
Cl2/THF) the major product was the dithiocyanate 12 (Scheme
3). For complexation, 12 was deprotonated with NaH and then
combined with the metal halide to give the final metal
complexes 1-5 in varying yield. The iron compound 5 was
meta-stable, allowing only partial characterization (Scheme 4).
(27) Frisch, M. J.; et al. Gaussian 98, Revision A.11; Gaussian Inc.: Pittsburgh,
PA, 2001.
(28) Dunning, T. H. J.; Hay, P. J. Modern Theoretical Chemistry; Plenum
Press: New York, 1977; Vol. 3.
(29) Igel-Mann, G.; Stoll, H.; Preuss, H. Mol. Phys. 1988, 65, 1321-1328.
(30) Bergman, D. L.; Laaksonen, L.; Laaksonen, A. J. Mol. Graphics Model.
1997, 15, (301-306).
(31) Laaksonen, L. J. Mol. Graphics 1992, 10, 33-34.
(32) Jones, R. A.; Karatza, M.; Voro, T. N.; Civcir, P. U.; Franck, A.; Ozturk,
O.; Seaman, J. P.; Whitmore, A. P.; Williamson, D. J. Tetrahedron 1996,
52, 8707-8724.
(33) Hein, B.; Beierlein, U. Pharma. Zentralhalle 1957, 96, 401-421.
(34) Schindlbauer, H. Monatsh. Chem. 1968, 99, 1799-1807.
(35) Owsley, D. C.; Nelke, J. M.; Bloomfield, J. J. J. Org. Chem. 1973, 38,
901-903.
Metal Complex Characteristics
(36) The design is such that there is a continuous conducting π system throughout
the molecule except were the electrons are forced to travel through the
metal atom. The alligator clips must be in a position such that each metal
lead in a SMT attaches to only one ligand, and thus, electrons travel from
lead to lead through the metal center as opposed to from end to end of one
ligand.
The metal complexes are intended to be studied in a variety
of molecular electronic testbeds.38 The results of the tests will
(37) Ulman, A. Chem. ReV. 1996, 96, 1533-1554.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3181

A R T I C L E S
Ciszek et al.
Scheme 4. Complexation of the Free Ligand 12 with the Metal
Halide
Figure 4. Crystal field splitting and correlation diagram for 2.
tetragonal distortion (T, where T ) RS/RL)42 is not unusually
large (0.84-0.89) and supports treatment of the complexes as
compressed octahedra. Asymmetry that is not consistent with
compressed octahedra is observed in the bond angles. The
rigidity of the ligand results in N-M-N bond angles which
deviate from 90° (an average of 74-76°). This is only slightly
larger than the typical Jahn-Teller distortion commonly seen
for coordination compounds including Cu(dien)2(NO3)2 (79-
80°), one of the few well-studied compressed octahedral
compounds,42-44 which will serve as the basis for comparison
in our work. Though the true symmetry of the molecule is S4,
the symmetry elements are similar to D4h, and thus much of
the literature can be used interchangeably.
Table 1. Bond Distance and Angle Deviations from Oh Symmetry
av N−M−N angle
metal
RS (Å)
RL (Å)
(deg)
Zn (1)
Cu (2)
Ni (3)
Co (4)
Fe (5)
1.96
1.89
1.96
1.97
2.03
2.32
2.26
2.23
2.24
2.27
74
75
76
74
74
The crystal field splitting expected from S4/D4h symmetry is
illustrated in Figure 4. The splitting (and the associated
transitions in the UV-vis spectrum) is somewhat ambiguous.
2
The dz orbital is the highest in energy (by EPR spectroscopy
2
2
vide infra), lying above the dx -y orbital. It is generally
accepted39,44,45 that the dxz and dyz orbitals lie at a slightly higher
energy than the dxy orbital, although there has been some
question as to the latter assignment.43 Since we have no
experimental evidence to suggest otherwise, we use the standard
ordering (Figure 4).
UV-Vis 43000-6250 cm-1 (230-1600 nm). Building on the
preceding discussion of the relative energies of the d orbitals,
we examined the various d electron transitions in complexes
2-5. The assignments of the absorption features of 2 are
relatively straightforward (Table 2) and are consistent with the
expected transitions. Assignments for the other metals are
problematic. For nickel(II) (d8), the correlation diagram gets
more complex (Figure 5a) with 11 different terms for Oh
symmetry and 19 when S4 symmetry is taken into account.
Cobalt(II) (d7) has seven free ion terms, and 19 in the simple
case of Oh symmetry. Thus, the cobalt spectrum generally
contains many broad indistinguishable features that can only
be resolved at very low temperatures.46,47
Figure 3. Structural details of compounds 1-5. (a) The relevant bond
distances and angles of 1-5. (b) ORTEP of 4 from X-ray crystallography.
be correlated with spectroscopic data to show these individual
molecular effects in electronic devices.11-13,20 The physical
characterization data discussed here will focus primarily on
molecular properties that will contribute to the discussion of
the SMT data. An expansive version of the characterization can
be found in the Supporting Information.
Structure and Crystal Field Splitting. The target com-
pounds 1-5 share the same geometry. The complexes can be
called compressed octahedra, an uncommon species39,40 (Table
1, Figure 3). On average the M-Npyrrole bond distances (along
the Z axis) are 0.2-0.3 Å shorter than the M-Npyridine bond
distances. For a compressed octahedral species, it is common
to predict the crystal field splitting based from an Oh symmetry
with the appropriate distortions noted (taking into account only
the immediate environment of the metal41). The extent of
Nickel has been studied in greater detail, although low
temperature is again necessary for full detailed studies of this
system. For nickel we would expect three main sets of transitions
(Oh symmetry) corresponding to the spin allowed transitions
3
3
3
(3A2g to T2g, T1g, and T1g (P)) (Figure 5). Each of these
(42) Procter, I. M.; Hathaway, B. J.; Nicholls, P. J. Chem. Soc. A 1968, 1678-
1684.
(38) Tour, J. M. Molecular Electronics: Commercial Insights, Chemistry,
DeVices, Architecture and Programming; World Scientific: River Edge,
New Jersey, 2003.
(39) Halcrow, M. A. J. Chem. Soc., Dalton Trans. 2003, 4375-4384.
(40) Murphy, A.; Mullane, J.; Hathaway, B. Inorg. Nucl. Chem. Lett. 1980, 16,
129-134.
(41) Figgis, B. N. Introduction to Ligand Fields; John Wiley & Sons: New
York, 1966.
(43) Hathaway, B. J.; Bew, M. J.; Billing, D. E. J. Chem. Soc. A 1970, 1090-
1095.
(44) Stephens, F. S. J. Chem. Soc. A. 1969, 883-890.
(45) Pilbrow, J. R. Transition Ion Electron Paramagnetic Resonance; Clarendon
Press: Oxford, 1990.
(46) Ferguson, J.; Wood, D. L.; Knox, K. J. Chem. Phys. 1963, 39, 881-889.
(47) Lever, A. B. P.; Walker, I. M.; McCarthy, P. J. Inorg. Chim. Acta 1980,
39, 81-90.
9
3182 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Synthesis and Assembly of Transition Metal Complexes
A R T I C L E S
Table 2. Visible Transitions Observed in the Transition Metal Complexes
absorption featuresa
1
1
1
λ
, cm-
λ
, cm-
λ
, cm-
compound
(nm)
ꢀ
transition
(nm)
ꢀ
transition
(nm)
ꢀ
transition
b
b
2 (Cu)
14290
(700)
-
28
2B1gf 2B2g
12990
(770)
16420
(609)
46
17
2B1g f 2Eg
8670
189
2B1g f 2A1g
(1154)
10550
(948)
9550
(1047)
10830
(923)
3 (Ni)
-
-
“Oh T1g”
3B1gf 3Eg
+
31
31
19
3B1g f 3B2g
3B1g f 3Eg
(3T2g, Oh)
-
3B1g f 3A2g
4 (Co)c
15530
(644)
25
-
-
-
-
a The spectra that are not shown are included in the Supporting Information. b See previous section for discussion on these particular transitions. c Features
are very broad and possibly encompass multiple transitions.
Figure 5. (a) Correlation diagram for octahedral and compressed octahedral splitting of the d8 configuration. Energy is not to scale. For more details on
calculating energy transitions of nickel(II) in D4h fields see references 49-51. (b) Visible spectrum of 3.
transitions will be split into two bands when the S4 symmetry
is taken into consideration. The transitions to 3B2g and 3Eg (3T2g
in Oh) are expected to appear between 7500 and 12500 cm-1 47
.
Though the splitting between the two transitions is small (∼1000
3
cm-1) the spectrum can be deconvolved to show the B2g
transition at 9550 cm-1 and the 3Eg peak at 10550 cm-1 (Figure
3
5b). The two absorbances stemming from the T1g (Oh) state
(3B1g f Eg, A2g) contain excessive overlap from other terms
as well as charge transfer and π f π* transitions and hence
cannot be deconvolved. Thus, the peak at 16420 cm-1 is
assigned to the combination of those two signals. The remaining
spin-allowed and -forbidden transitions are obscured by the other
larger signals. For cobalt, six low-energy spin-allowed transitions
are expected. However, none of these are sufficiently resolved
in order to make definite assignments.
Although the conjugated system is advantageous for transfer-
ring electrons throughout the molecule, the π f π* transistions
of the conjugated system makes assignment of the charge-
transfer signals problematic. The ligand π f π* and charge-
transfer transitions are summarized in Figure 6 and Table 3.
EPR. The EPR spectrum of the copper compound allows for
some confirmation of the predicted crystal field splitting. The
measured g tensor can be predicted by the equation51
3
3
Figure 6. Charge transfer and π f π* transitions in complexes 1-4.
Table 3. UV Absorption Band for Compounds 1-4
cmpd
transitions cm-1 (nm)
12
-
-
30680 (326)
31250 (320)
31150 (321)
30770 (325)
30160 (322)
-
1 (Zn)
2 (Cu)
3 (Ni)
4 (Co)
40490 (247)
40160 (249)
37590 (266)
40490 (247)
-
28570 (350)
28650 (349)
28330 (353)
-
-
33110 (302)
34970 (286)
coupled orbital, L is the orbital angular momentum operator,
ge is the g-factor for a free electron, λ is the spin-orbit coupling
constant and E represents the respective energy terms. If the
n|LR|n′ n|Lâ|n′
2
unpaired electron lies in the dz orbital, the equation gives gxx
gRâ ) geδRâ + 2λ
∑
) gyy ) ge + 6λ/(Ez2 - Exz,yz) and gzz ) ge. Hence, an
experimental value of gzz near that of a free electron (2.0023)
and a larger gxx,yy is evidence for the predicted crystal field
En - En′
n′
where n is the orbital containing the unpaired spin, n′ is the
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3183

A R T I C L E S
Ciszek et al.
Figure 7. Magnetic properties of complexes 2 and 4. (a) EPR spectrum of the copper complex 2. Inlaid image is an enlarged portion of the spectrum. (b)
A histogram of the Kondo temperatures measured for the cobalt complex 4 and the copper complex 2.20 (c) EPR spectrum of the cobalt complex 4 at various
receiver gains.
Figure 8. Magnetization data for 2 and 3. (a) SQUID low-field data for copper compound 2. (b) High-field data for compound 2. (c) SQUID low-field data
for nickel compound 3.
splitting.52 The respective measured values for g| (gzz) and g
(gxx,gyy) are 2.020 and ∼2.22 (the small amount of uncertainty
in the latter being due to hyperfine splitting appearing near the
inflection point) (Figure 7a). This confirms the unpaired electron
lies in the dz orbital. The hyperfine splitting constant A is 200
G in the z direction. The hyperfine splitting constant is smaller
coupling between the transition metal and the conduction
electrons of the electrodes (through the ligands) onto which the
molecule is assembled. The measured Kondo temperatures are
comparable to those observed in scanning tunneling microscopy
measurements of transition metals directly bonded to gold
surfaces. This large extent of delocalization provides an
experimental explanation for such a high source-spin-drain
coupling.
For the cobalt species (Figure 7c) we observe large magnetic
anisotropy consistent with that seen in the literature for co-
balt(II) where S ) 3/2.57 The large g factor (4.3) becomes
anisotropic when the octahedral symmetry is distorted by
trigonal58 or tetragonal distortion.59 This results in additional
terms along the diagonal of the energy matrix which supplement
spin-orbit coupling. Further discussion of this phenomenon can
be found elsewhere.45,57,60 In our system g was observed to be
2.1 with g| equal to 8.27. This large anisotropy along the z axis
has also been observed in terpyridine complexes.61 For 4 the
hyperfine splitting constant A is 135 G in the z direction.
Magnetic Susceptibility. The magnetic susceptibilities of
compounds 1-5 were examined by SQUID. Analysis of
complexes 1-3 was straightforward. As expected, the zinc(II)
complex 1 has no unpaired spins and is diamagnetic with a
susceptibility of -4.68 × 10-4 emu/mol. The copper(II)
complex 2 with a 3d9 configuration behaves like a well-defined
2
than the line width for gxx,yy
.
By combining these results with the observed UV transitions,
one can calculate λ′ the observed spin-orbit coupling con-
stant.53,54 The ratio of λ′/λ provides a measure of the extent of
covalency, or an indication as to the delocalization of the spin
onto the ligand.51,55 With dxy as the lowest lying d orbital (dxy
2
f dz ) 700 nm, see previous section), λ′/λ is equal to 0.57.
This is quite low when compared to other simple copper(II)
complexes.56 It also provides experimental evidence that this
compound contains a large amount of spin delocalization onto
the ligands and has very good transport through the metal center.
Such experimental evidence is welcome as we have previ-
ously shown unusually high Kondo temperatures in SMTs
(Figure 7b).20 Kondo transport15,16 (which can simplistically be
thought of as electron transfer through unpaired spins) has a
characteristic temperature that is exponentially sensitive to the
(48) Schreiner, A. F.; Hamm, D. J. Inorg. Chem. 1973, 12, 2037-2048.
(49) Maki, G. J. Chem. Phys. 1958, 28, 651-662.
(50) Rowley, D. A.; Drago, R. S. Inorg. Chem. 1967, 6, 1092-1096.
(51) Atherton, N. M. Principles of Electron Spin Resonance; Horwood, PTR
Prentice Hall: New York, 1993.
(57) Abragam, A.; Pryce, M. H. L. Proc. R. Soc. London, Ser A. 1951, 206,
173-191.
(58) Jesson, J. P. J. Chem. Phys. 1966, 45, 1049-1056.
(59) Bleaney, B.; Ingram, D. J. E. Proc. R. Soc. London, A. 1951, 208, 143-
158.
(60) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition
Ions; Clarendon Press: Oxford, 1970.
(61) Kremer, S.; Henke, W.; Reinen, D. Inorg. Chem. 1982, 21, 3013-3022.
(52) Palmer, G. Biochem. Soc. Trans. 1985, 13, 548-560.
(53) Dunn, T. M. J. Chem. Soc. 1959, 623-627.
(54) Owen, J. Proc. R. Soc. London, Ser A. 1955, 227, 183-200.
(55) Cotton, F. A.; Goodgame, D. M. L.; Goodgame, M. J. Am. Chem. Soc.
1961, 83, 4690-4699.
(56) Hathaway, B. J.; Billing, D. E.; Nicholls, P.; Procter, I. M. J. Chem. Soc.
A 1969, 319-325.
9
3184 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Synthesis and Assembly of Transition Metal Complexes
A R T I C L E S
Figure 9. Magnetization data of for the iron complex 5. (a) SQUID low-field data for 5. (b) SQUID high-field data for 5.
Table 4. Reversible Reductions/Oxidationa of the Metal Centers
1
spin /2 paramagnetic compound (Figure 8a,b). Low-field data
cmpd
metal reduction
potential (V)
metal oxidation
potential (V)
gives a magnetic susceptibility of 1.76 µB, similar to the 1.73
µB expected for a spin /2 compound with no orbital contribu-
tions. The high field data can be fit with a Brillouin function
with j ) 0.5 and g ) 2.02. The nickel(II) species 3 is similarly
straightforward at high temperatures with an effective magnetic
moment of 4.3 µB.
The behavior of the cobalt(II) complex 4 was much more
complex. We discovered from measurements on powders that
this material’s susceptibility is highly anisotropic; as a result
we were forced to collect data from a particularly large single
crystal (2.6 mg). The single crystal data indicates that the cobalt
complex is strongly paramagnetic along one axis and weakly
paramagnetic along the other two. According to low-field dc
susceptibility measurements, the weakly paramagnetic axes have
an effective magnetic moment of 2.6 µB, and the strongly
paramagnetic axis gives an effective magnetic moment of 9.2
µB. The crystallographic axes are coincident with the molecular
axes.
1
2
3
4
10
5
CuII f CuI
NiIII f NiII
CoIII f CoII
CoIII f CoII
FeIII f FeII
-0.80
0.77
-0.012
-0.42
0.20
CuI f CuII
NiII f NiIII
CoII f CoIII
CoII f CoIII
FeII f FeIII
-0.66
0.87
0.14
-0.30
0.39
a The oxidations/reductions were performed on the metal complexes 1-5,
10 (1 mM) in THF with 0.1 M TBABF4 as the supporting electrolyte. The
scan rate for all of the redox potentials was 20 mV/s with the exception of
5 which was at 600 mV/s. The potential is vs Ag/Ag+.
then back to -1.0 V(Figure 11a). We surmise that this
phenomenon is responsible for the reduction current peak (Ipc)
always being smaller than the oxidation current peak (Ipa).
Because the ratio of these two peaks approaches unity as the
scan rate is increased (Figure 11b), it is natural to suggest that
this is a rearrangement phenomenon (be it ligand association
or dissociation), which is consistent with the lack of stability
for 5 as mentioned earlier.
It was unsurprising that the redox potentials of the complexes
3-5 occur at a potential that is negative with respect to that of
the analogous 2,6-bis(2-pyridyl)pyridine (terpy) complexes (E1/2
for the II f III oxidation was 1.65 V for nickel, 0.31 V for
cobalt, and 1.13 V for iron, vs SSCE).66,67 The terpy ligand is
neutral while our complexes use 12 as an anionic ligand, and
thus a substantial difference in redox potential was expected.
Following the same logic, the redox behavior of complexes 3-5
should be closer to the analogous bis(2-pyridylcarbonyl)amine
(bpca) complexes. The E1/2 reported for the II f III redox
process was -0.28 for cobalt and 0.35 V for iron.68 These values
suggest that the nature of 12 lies somewhere inbetween the bpca
and terpy ligands, possibly due to the electron-withdrawing
effect of the SCN moieties.
The magnetization data on the iron(II) complex 5 also yielded
interesting information. While the data above ∼6 K is highly
suggestive of a system with four unpaired spins (with an
effective magnetic moment of 5.9 µB), it is clear that the
magnetization data plotted vs H/T below 6 K fall on different
curves (Figure 9). This suggests that zero field splitting plays a
significant role in the magnetic properties of the iron com-
plex.62,63 Alternatively, it is possible that the low-temperature
data hint at the onset of a transition to a low-spin state.64,65
CV. From cyclic voltammetry we can see that all molecules
are overall neutral with a metal(II) center (Table 4, Figure 10).
As the potential is increased, nickel(II) 3, cobalt(II) 4, and
iron(II) 5 complexes undergo a reversible oxidation to nickel-
(III), cobalt(III), and iron(III), respectively. This reversible
oxidation and reduction occurs at 0.87 and 0.77
V
(vs Ag|AgNO3, 20 mV/s) for nickel, -0.01 and 0.14 V for
cobalt (20 mV/s), and 0.39 to 0.20 for iron (at 600 mV/s). It is
worth noting that there is an extra feature at -0.69 V for iron.
None of the other compounds (1-4) demonstrated this feature.
When we further investigated this phenomenon we saw that it
was not present when scanning from -0.1 to -1.0 V but was
present when the scan was cycled from -1.0 to 0.75 V and
To examine this phenomenon, the redox potential of 4 was
compared with that of the SCN-free cobalt compound 10, which
(62) Behere, D. V.; Mitra, S. Inorg. Chem. 1980, 19, 992-995.
(63) Chen, W.-Z.; Cotton, F. A.; Dalal, N., S.; Murillo, C. A.; Ramsey, C. M.;
Ren, T.; Wang, X. J. Am. Chem. Soc. 2005, 127, 12691-12696.
(64) Gutlich, P.; Garcia, Y.; Goodwin, H. A. Chem. Soc. ReV. 2000, 29, 419-
427.
(66) Rao, J. M.; Hughes, M. C.; Macero, D. J. Inorg. Chim. Acta 1976, 16,
231-236.
(67) Arana, C.; Yan, S.; Keshavarz-K, M.; Potts, K. T.; Abruna, H. D. Inorg.
Chem. 1992, 31, 3680-3682.
(68) Kajiwara, T.; Sensui, R.; Noguchi, T.; Kamiyama, A.; Ito, T. Inorg. Chim.
Acta 2002, 337, 299-307.
(65) Gutlich, P.; Goodwin, H. A. Top. Curr. Chem. 2004, 233, 1-47.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3185

A R T I C L E S
Ciszek et al.
Figure 10. Representative plots of the cyclic voltammetry data. The reversible reduction/oxidation of (a) copper complex 2 and (b) the iron complex 5. The
linear relationship between the oxidation peak currents and square root of scan rates, as shown in the insets, indicates a diffusion-controlled redox process.
Figure 11. (a) Cyclic voltammagram of 5. (b) Ipa/Ipc at various scan rates.
Figure 12. Potential map of 9 (a), 1 (b), and 12 (c). Images are overlaid visualizations of electron density and potential maps. The calculation was performed
using the Gaussian 98 software suite27 with SDD28,29 basis set on the X-ray crystal structure data. The output was visualized with gopenmol.30,31 Red denotes
the most electron-rich areas, blue, the most electron-deficient.
has its oxidation/reduction at -0.42 and -0.30 V, respectively.
From such comparison we see that the thiocyanate moiety has
the effect of decreasing the reduction potential of the metal
(-0.38 V for 10, -0.01 V for 4), apparently by reducing the
electron density available in the pyrrole ring, and in turn the
metal center. This is further supported by theoretical modeling
of the compounds which shows a decrease in the electron density
of the pyrrole ring (illustrated in Figure 12) when the thiocyanate
moieties are on the pyrrole ring.
This reduction and oxidation occurs respectively at -0.80 and
-0.66 V. The zinc compound 1, as expected, shows no
oxidation or reduction within a window from +1.0 to -1.6 V.
The zinc compound 1 does, however, show the oxidation and
reduction of the thiocyanate group (Figure 13). In agreement
with recently published data,69,70 we clearly see the thiocyanate
reduction to thiolate at -2.0 to -2.2 V for 1 (zinc) and 12 and
3 (nickel), and 4 (cobalt) (although in the latter two cases the
Neither the zinc compound 1 or the copper compound 2 are
readily oxidized to the metal(III) oxidation state. With the copper
compound 2, the lowest-energy transition is to the metal(I) state.
(69) Houmam, A.; Hamed, E. M.; Hapiot, P.; Motto, J. M.; Schwan, A. L. J.
Am. Chem. Soc. 2003, 125, 12676-12677.
(70) Houmam, A.; Hamed, E. M.; Still, I. W. J. J. Am. Chem. Soc. 2003, 125,
7258-7265.
9
3186 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Synthesis and Assembly of Transition Metal Complexes
A R T I C L E S
Figure 13. Oxidation and reduction of the SCN moiety. (a) The oxidation and reduction of the free ligand 12. (b) The oxidation and reduction of 1.
Table 5. Thiocyanate Infrared Stretches for the Metal Complexes
precise location cannot be determined due to possible metal
redox features appearing at nearby potentials). The correspond-
1-5 and the Free Ligand 12
cmpd
12
4
3
5
2
1
ing oxidation features (ca 0.3 V) are attributed to the formation
of disulfide bonds. These data show the sensitivity of the
reduction to the electron density in the adjacent ring, as has
been demonstrated previously.69,70 As a general trend, when the
ring is electron deficient, the reduction occurs at a more positive
voltage. In our experiments the electron-poor ligand 12 shows
the reduction at -1.96 V. If these data are then compared to
those of the zinc, nickel, and cobalt complexes (-2.2, ∼-2.1,
and ∼-2.1 V, respectively), we observe this same trend.
Infrared 4000-400 cm-1 (2500-25000 nm). Several IR
assignments can be made unambiguously and can be useful in
corroborating the relative electron densities of various parts of
the ligand. The most obvious stretch is the ν(CN) which occurs
at ∼2150 cm-1. The signal at ∼680-690 cm-1 can also be
attributed to the thiocyanate functionality. The stretches at
∼1590 are ν(CdC) and ν(CdN). A stretch at 1583 cm-1 is
generally attributed to pyridine CdC and CdN stretches; it is
well-known that this stretch shifts to higher wavenumbers for
pyridine complexation.71 In our case the shift is not as large
due to the longer than usual pyridine-metal bond length. Most
of the other peaks vary too much in location or intensity to
make an assignment. The stretches falling between 1520 and
1430 cm-1 belong to the ring breathing (ν(CdC) and
ν(CdN)), with the two stretches at 1460-1430 likely belonging
to the pyridine.
metal
None
Co
Ni
Fe
-
Cu
-
Zn
-2.2
SCN reduction -1.96 ∼-2.1 -2.1
ν(CN), cm-1
2152.0 2150.1 (a)
2149.1 2148.9 2148.7
ν(C-S), cm-1 733.9
ν(C-S), cm-1 686.6
743.8
696.5
744.1 743.0
699.3 698.3
743.7
698.5
743.3
699.0
(a) Two separate peaks appear for the nickel compound 3 in the
thiocyanate region as a result of a slight difference in bond lengths of one
of the thiocyanates.
Figure 14. A model of the cobalt-containing molecule 4 assembled on a
gold (111) surface.
These IR assignments can be useful in the discussion of the
electronic effects of the metal centers on the SCN reductions.
Should the arguments regarding the SCN reduction potentials
and their relationships with the electron density of the pyrrole
ring (made in the previous section) be valid, we would also
expect to see corresponding changes in the SCN frequencies.
The trend is obvious in the case of the electron-dependent CN
stretch (Table 5). As changes in the electron density of the
pyrrole ring for 1-5 are primarily the result of covalent
interactions with the metal both of these frequencies should also
provide some insight into the amount of delocalization of the
ligand electrons onto the metal.
prevent disulfide polymerization,22 and the rigidness and
orthogonality of the ligands prevent both ligands from binding
to gold at the same time (Figure 14), confirmation was needed.
We also examined the molecules for ligand loss, metal oxidation/
reduction, or ligand decomposition.
X-ray Photoelectron Spectroscopy (XPS). XPS is a power-
ful tool available for discerning the state of the molecule on
the surface. By comparing the spectral data of the powder
sample to that of the assembled molecule, we can observe the
bonding of the sulfurs, and any change in the oxidation state of
the metal center, as well as other changes that may occur upon
assembly.
Initially, we sought confirmation that the molecules were
covalently bound to the surface. An advantage of the thiocyanate
assemblies is that thiolates (Au-S) generated during the
assembly have substantially more electron density than the
thiocyanates (-SCN) and can be distinguished easily by XPS
Surface Characterization. Surface characterization of the
molecular assemblies focuses on analytically determining which,
if any, aspects of the molecule may have changed during the
assembly procedure. Although thiocyanates are known to
(71) Gupta, R. R. Physical Methods in Heterocyclic Chemistry; Wiley-
Interscience: New York, 1984.
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3187

A R T I C L E S
Ciszek et al.
Figure 15. Deconvolution of the sulfur 2p region of the XPS spectrum. The left pair of peaks in each figure is due to the thiocyanates (-SCN), the right
pair from the thiolates (Au-S). (a) Deconvolution of the S 2p region in the surface-assembled sample of 4. (b) Deconvolution of the S 2p region in the
surface-assembled sample of 2.
Table 6. Comparison of the XPS Binding Energies of the
Powdered and Assembled Forms of 4
no change in the S-Au to S-CN ratio from assemblies done
on gold-silicon surfaces. Another possibility, disproportion-
spectral
surface spectrum
powder spectrum
ation, is ruled out (vide infra). X-ray meditated decomposition
of the thiocyanate can be ruled out as multiple scans in the same
region show no difference in intensity. We are left to conclude
that the difference is due to a combination of a slight difference
in the efficiency of the photoelecton process and/or a charging
effect. Charging can retard the formation of photoelectrons. A
difference in the effectiveness of the neutralizer on the gold
thiolates vs the thiocyanates will in turn cause a difference in
the measured intensity by XPS. Since the difference appears
consistently from sample to sample, such a mechanism seems
likely.
component
binding energy (eV)
binding energy (eV)
S 2p3/2
S 2p1/2
Co 2p3/2
Co 2p1/2
N 1s
161.9 (Au-S)
163.2 (Au-S)
165.0 (-SCN)
166.4 (-SCN)
781.6
796.5
399.3
164.8
166.2
780.4
796.3
398.6
(Figure 15a,b).22,23 Thus, these two peaks are an indication that
one set of sulfurs is bound to gold, the other to CN. All
compounds showed similar sulfur regions. As a general trend,
all thiolate regions were of slightly larger intensity than the
thiocyanate region. A possible reason for this discrepancy is
that, at several areas on the substrate, surface roughness (which
is closer to curvature on the atomic scale) may allow three or
four sulfurs to reach the surface. When this is averaged out over
all the surfaces, we would expect to see a ratio of thiolate to
thiocyanate that is larger than 1:1. However, when assemblies
were performed on gold on the flatter mica surface, there was
XPS allows the examination of the surface assembly for other
signs of change in the molecule. Large changes in binding
energy of the elements would be indicative of a change in
oxidation state or bonding motif. Table 6 and Figure 16
summarize the binding energies of each element of the complex
showing both the powder spectrum and surface assembly. No
substantial changes in binding energies were seen, indicating
Figure 16. Comparison of powder and surface XPS spectra of 4. Results are representative of the data for 1-4.
9
3188 J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006

Synthesis and Assembly of Transition Metal Complexes
A R T I C L E S
Table 8. Stretching Frequencies of the CN Triple Bond
Table 7. Film Thickness ((10%) for Assembliesa of 1, 3, 4
1
1
measured
cmpd
ν
(CN) (powder) cm-
ν
(CN) (surface) cm-
molecule
thickness (Å)
3
4
2155.6, 2146.8
2150.1
2156.5, 2147.1
2151.7
Zn (1)
Ni (3)
Co (4)
13.5
10.5
12.8
thiocyanate was clearly visibly by IR and had a stretch nearly
identical to that of the powder. The surface stretch had shifted
∼1 cm-1 higher but appeared at values similar to those of the
powder spectrum (Table 8).
a The calculated height for surface bound-metal complex is 14 Å. The
calculated height is the distance from the gold surface to the outermost
atom of the complex while taking into account a 20° sulfur tilt angle.
Conclusions
the assembled molecule was similar to the powder form except
for the fact that two of the thiocyanates had been converted to
thiolates. The small shift in both nitrogen and cobalt energies
can be attributed to the difference in electron density of the
ligand when the electron-withdrawing thiocyanates are converted
to electron-rich thiolates. According to atomic concentration
calculations performed using the PHI Multipak software, the
ratio (corrected for the atomic sensitivity factors of each
respective element) of cobalt to S in the crystalline powder
sample of 4 matched the ratio found in the surface-assembled
sample. This indicates that the metal complex has not dispro-
portionated or decayed and is most likely intact on the surface.
Ellipsometry. Ellipsometry was carried out on the self-
assembled monolayers (SAMs). Due to the bulky nature of the
molecules we did not expect the packing to approach the
theoretical height of the molecule. For a layer of bulky molecules
on the surface we would expect a thickness of about 11-12
Å,72 while for a well-packed monolayer the thickness would
be 14 Å. Ellipsometric values lie between these two, but do
not give a definitive indication due to the error in the
measurement. Thiocyanate assembly has been previously shown
to prevent multilayers,22 and ellipsometry confirms this for these
complexes (Table 7).
We have shown a simple and efficient method for the
synthesis of the transition metal complexes 1-5. Characteriza-
tion of these compounds via UV-vis, cyclic voltammetry,
magnetic susceptibility, and EPR allows us to predict and
explain many of the electronic properties of these molecules in
SMTs. The complexes have a high degree of covalency between
the metal core and ligands, explaining the origin of high Kondo
temperatures in SMTs. Assembly of the molecules on gold gives
a thickness consistent with what is expected for a loosely packed
layer of bulky molecules. Other surface characterizations (XPS,
surface IR) show species pre- and post- assembly that are
virtually identical except for the two thiocyanates that have been
converted to thiolates during assembly.
Acknowledgment. J.M.T. acknowledges support from DAR-
PA, via ONR and the AFOSR. The NSF, CHEM 0075728,
provided partial funds for the 400 MHz NMR. D.N. acknowl-
edges financial support from the Research Corporation, the
Robert A. Welch Foundation, the David and Lucille Packard
Foundation, an Alfred P. Sloan Foundation Fellowship, and NSF
CAREER award DMR-0347253. We thank Professors Marian
Fabian, Graham Palmer, Kenton Whitmire, and Lon Wilson as
well as Dr. Douglas Ogrin for their assistance.
Infrared Reflection Spectroscopy Measurements. IR re-
flection spectroscopy was used for a secondary confirmation
of the presence of the thiocyanate after assembly. Both the nickel
and cobalt complexes were examined. In both cases the
Supporting Information Available: Synthetic details, char-
acterization data, and the complete ref 28. This material is
available free of charge via the Internet at http://pubs.acs.org.
(72) Wu, C.-G.; Chiang, S.-C.; Wu, C.-H. Langmuir 2002, 18, 7473-7481.
JA055459D
9
J. AM. CHEM. SOC. VOL. 128, NO. 10, 2006 3189