Thermal dihydrogen elimination from Cp*2Yb(4,5- diazafluorene)

Article abstract of DOI:10.1021/om300876b

The reaction of 4,5-diazafluorene with Cp*₂Yb(OEt ₂), where Cp* is pentamethylcyclopentadienyl, affords the isolable adduct Cp*₂Yb(4,5-diazafluorene) (1), which slowly eliminates H₂ to form Cp*₂Yb(4,5-diazafluorenyl) (2); the net reaction is therefore 1 → 2 + H^?. The ytterbium atom in 1 is shown to be intermediate valent by variable-temperature L_III-edge X-ray absorption near-edge (XANES) spectroscopy, consistent with its low effective magnetic moment (μ_eff). The experimental studies are supported by complete active space self-consistent field (CASSCF) calculations, showing that two open-shell singlets lie below the triplet state. The two open-shell singlets are calculated to be multiconfigurational and closely spaced, in agreement with the observed temperature dependence of the XANES and χ data, which are fit to a Boltzmann distribution. A mechanism for dihydrogen formation is proposed on the basis of kinetic and labeling studies to involve the bimetallic complex (Cp*₂Yb)₂(4,5- diazafluorenyl)₂, in which the heterocyclic amine ligands are joined by a carbon-carbon bond at C(9)-C(9′).

Full text of DOI:10.1021/om300876b

Article
pubs.acs.org/Organometallics
Thermal Dihydrogen Elimination from Cp*2Yb(4,5-diazafluorene)
Gregory Nocton,†,‡,∥ Corwin H. Booth,‡ Laurent Maron,§ and Richard A. Andersen*,†,‡
́
†Chemical Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States
‡Department of Chemistry, University of California, Berkeley, California 94720, United States
§LPCNO, UMR 5215, Universite
́
de Toulouse-CNRS, INSA, UPS, Toulouse, France
S
* Supporting Information
ABSTRACT: The reaction of 4,5-diazafluorene with Cp*2Yb(OEt2), where Cp* is pentamethylcyclopentadienyl, affords the
isolable adduct Cp*2Yb(4,5-diazafluorene) (1), which slowly eliminates H2 to form Cp*2Yb(4,5-diazafluorenyl) (2); the net
reaction is therefore 1 → 2 + H•. The ytterbium atom in 1 is shown to be intermediate valent by variable-temperature LIII-edge
X-ray absorption near-edge (XANES) spectroscopy, consistent with its low effective magnetic moment (μeff). The experimental
studies are supported by complete active space self-consistent field (CASSCF) calculations, showing that two open-shell singlets
lie below the triplet state. The two open-shell singlets are calculated to be multiconfigurational and closely spaced, in agreement
with the observed temperature dependence of the XANES and χ data, which are fit to a Boltzmann distribution. A mechanism for
dihydrogen formation is proposed on the basis of kinetic and labeling studies to involve the bimetallic complex (Cp*2Yb)2(4,5-
diazafluorenyl)2, in which the heterocyclic amine ligands are joined by a carbon−carbon bond at C(9)−C(9′).
hole occupancy, such that when the contribution to the f14
INTRODUCTION
■
configuration is 0, nf = 1). The temperature dependences were
fit with the Boltzmann equation and shown to result from an
equilibrium between two open-shell singlet configurations. The
active space in the computational model was expanded to
include additional π* configurations on the substituted
bipyridine ligand, and a perturbative PT2 correction was
added in order to account for these experimental results.
The studies outlined above show that the magnetic and
spectroscopic properties can be described by a molecular model
based upon wave function methodology. A question that
naturally arises is whether the multiconfigurational ground state
plays a role in the chemistry, that is, breaking and making
chemical bonds, as indicated by ΔH° and ΔH⧧. In this article
we begin a systematic series of studies aimed at answering this
question. The initial studies begin by preparing the adduct
between Cp*2Yb and 4,5-diazafluorene (see Figure 1), which is
2,2′-bipyridine with a CH2 group annulated in the 3,3′-position.
The 2,2′-bipyridine adduct of decamethylytterbocene,
(C5Me5)2Yb(bipy), and related adducts with heterocyclic
amine ligands are classified as intermediate-valence com-
pounds.1 Thus, the valence of ytterbium in the Cp*2Yb
fragment is neither Yb(II) nor Yb(III); rather, its value is
between these extreme integral values.2−5 A molecular model
was developed from computational studies that accounts for the
following experimental data: (i) the low value of the effective
moment (μeff) at 300 K and its temperature dependence from 5
to 300 K and (ii) the ytterbium LIII-edge X-ray absorption near-
edge (XANES) spectra showing the presence of f13 and f14
features that are independent of temperature from 30 to 400 K,
with the f13 configuration dominant. A physical model,
consistent with all these observations derived from CASSCF
calculations, is that the ground state is an open-shell singlet
consisting of f13(π*)1 and f14(π*)0 configurations that are lower
in energy than the triplet configuration.
When the experimental studies were extended to methyl-
substituted bipyridine ligands coordinated to Cp*2Yb, the
adducts exhibited magnetic moments and nf values that
depended upon temperature, the number of methyl groups,
and their positions in the bipyridine ligands (nf accounts for f-
Special Issue: Recent Advances in Organo-f-Element Chemistry
Received: September 12, 2012
Published: November 29, 2012
© 2012 American Chemical Society
1150
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
9,9′-Bis-4,5-diaza-9H-fluorene. The ligand was prepared as
described in the literature.10−12 Stirring 4,5-diazafluorenone for 4 h
at 100 °C in the presence of hydrazine hydrate gave 9,9′-bis-4,5-diaza-
9H-fluorene in 45% yield. The two products (i.e., 4,5-diazafluorene
and 9,9′-bis-4,5-diaza-9H-fluorene) were separated on a silica gel
column (50% ethyl acetate, 50% hexane). 9,9′-Bis-4,5-diaza-9H-
fluorene did not sublime but was ground into a fine powder and
heated at 110 °C under reduced pressure for 1 week in order to
This ligand may be viewed as a cyclopentadiene that shares
edges with two pyridine ligands at the 1,2- and 3,4-positions.
Thus, the NCCN torsion angle is fixed to be close to 0° and
one of the two CH2 hydrogens is acidic; the pKa of CpH is 18,
and that of fluorene is 22.6.6 The experimental and computa-
tional studies show that the ground state is multiconfigura-
tional. The experimental studies also show that the adduct
eliminates dihydrogen stoichiometrically. Combining these
experimental and theoretical results allows for a molecular
level of understanding of the origin of the reactivity of
Cp*2Yb(4,5-diazafluorene).
1
eliminate the water of hydration. H NMR (CDCl3, 300 K): δ (ppm)
8.68 (d, 4H, J = 4.8 Hz), 7.31 (d, 4H, J = 8.0 Hz), 7.13 (dd, 4H, J = 7.8
Hz), 4.88 (s, 2H, −CH). The deuterated 9,9′-bis-4,5-diaza-9D-
fluorene was prepared in a similar way using ND2ND2·D2O (% H
1
∼5%). H NMR (CDCl3, 300 K): δ (ppm) 8.68 (d, 4H, J = 4.8 Hz),
7.31 (d, 4H, J = 8.0 Hz), 7.13 (dd, J = 7.8 Hz), 4.88 (s, 0.009H, −CH).
2D NMR (CDCl3, 300 K): δ 4.92 (s, CD2).
EXPERIMENTAL SECTION
■
Synthesis of Complexes. Cp*2Yb(4,5-diazafluorene) (1). A cold
toluene solution (10 mL) of Cp*2Yb(OEt2) (0.127 g, 0.245 mmol)
was added dropwise to a cold suspension of 4,5-diazafluorene (0.042 g,
0.245 mmol) in toluene (5 mL) at 0 °C. The brown-purple suspension
was stirred for 2 h and then filtered at 0 °C. The filtrate was cooled at
−20 °C, and a dark brown microcrystalline powder formed overnight
(16 h). The powder was collected by filtration (101 mg, 80%), washed
three times with cold toluene, and dried under reduced pressure (67
General Considerations. All reactions were performed using
standard Schlenk-line techniques or in a drybox (MBraun). All
glassware was dried at 150 °C for at least 12 h prior to use. Toluene
and pentane were dried over sodium and distilled, while CH2Cl2 was
purified by passage through a column of activated alumina. Toluene-d8,
pyridine-d5, and C6D6 were dried over sodium. All solvents were
degassed prior to use. Infrared samples were prepared as Nujol mulls
and taken between KBr plates and recorded on a Thermo Scientific
Nicolet IS10 spectrometer. Samples for ultraviolet, visible, and near-
infrared spectrometry were prepared in a Schlenk-adapted quartz
cuvette, and spectra were obtained using a Varian Cary 50 scanning
spectrophotometer. Melting points were determined in sealed
capillaries prepared under nitrogen and are uncorrected. Elemental
analyses and mass spectra (EI) were determined by the Micro-
analytical Laboratory of the College of Chemistry, University of
California, Berkeley, CA. X-ray structural determinations were
performed at CHEXRAY, University of California, Berkeley, CA.
Magnetic susceptibility measurements were made for all samples at 5
and 40 kOe in a 7 T Quantum Design Magnetic Properties
Measurement System, which utilized a superconducting quantum
interference device (SQUID). Sample containment and other
experimental details have been described previously.4 The samples
were prepared for X-ray absorption experiments as described
previously, and the same methods were used to protect the air-
sensitive compounds from oxygen and water.7 X-ray absorption
measurements were made at the Stanford Synchrotron Radiation
Lightsource on beamline 11-2. The samples were prepared and loaded
into a liquid helium-flow cryostat at the beamline as described
previously.7 Data were collected at temperatures ranging from 30 to
300 K, using a Si(220) double-crystal monochromator. Fit methods
were the same as those described previously.7
1
mg, 53%). H NMR (toluene-d8, 300 K): δ (ppm) 71.07 (2H), 30.50
1
(2H), 6.35 (2H), 4.65 (2H), 4.23 (30H, Cp*). H NMR (thf-d8, 297
K): δ (ppm) 45.59 (2H), 27.61 (2H), 9.51 (2H), 8.81 (2H), 3.05
1
(30H, Cp*). H NMR (py-d5, 297 K): δ (ppm) 35.80 (2H), 23.41
(2H), 9.35 (2H), 6.34 (2H), 2.94 (30H, Cp*). Mp: 273−275 °C.
Anal. Calcd for C31H38N2Yb: C, 60.87; H, 6.26; N, 4.58. Found: C,
60.63; H, 6.46; N, 4.45. Vis−near-IR (toluene; λ, nm (ε, cm−1 M−1)):
474 (2840), 501 (3410), 789 (1680), 909 (2450), 1005 (1390). IR
(cm−1): 2962 (w), 2903 (w), 2850 (m), 1595 (w), 1578 (w), 1563
(m), 1435 (m), 1406 (s), 1378 (m), 1296 (s), 1260 (m), 1232 (m),
1164 (m), 1089 (s), 1015 (s), 924 (w), 859 (m), 796 (m), 776 (m),
765 (s), 729 (s), 695 (w). MS: {Cp*2Yb(4,5-diazafluorene) − H} m/z
611.
Cp*2Yb(4,5-diazafluorene-d2) (1-d2). The synthesis was similar to
that used for the unlabeled compound using 145 mg (0.280 mmol) of
1
Cp*2Yb(OEt)2. H NMR (Toluene-d8, 300 K): δ (ppm) 71.07 (2H),
2
30.50 (2H), 6.35 (2H), 4.65 (0.16H, −CH2), 4.23 (30H, Cp*). H
NMR (C6D6, 298 K): δ 4.82. Mp: 273−275 °C. Anal. Calcd for
C31H36D2N2Yb: C, 60.67; H, 5.95; N, 4.59. Found: C, 59.56; H, 6.07;
N, 4.31. Vis−near-IR (toluene, λ, nm (ε, cm−1 M−1)): 474 (2840), 501
(3410), 789 (1680), 909 (2450), 1005 (1390). IR (cm−1): 2183 (m),
2117 (w), 1580 (m), 1538 (m), 1459 (m), 1396 (s), 1378 (m), 1292
(s), 1260 (m), 1246 (m), 1163 (s), 1095 (s), 1015 (s), 934 (w), 799
(m), 757 (m), 731 (s). MS: {Cp*2Yb(4,5-diazafluorene-d2) − D} m/z
612.
Kinetics Experiments and NMR Experiments. 1H NMR spectra
were recorded on Bruker AVB-400 MHz, DRX-500 MHz, AV-600
1
MHz, and Advance 300 MHz spectrometers. H chemical shifts are
Cp*2Yb(4,5-diazafluorenyl) (2). The complex Cp*2Yb(OEt2)
(0.127 g, 0.245 mmol) was mixed with 4,5-diazafluorene (0.042 g,
0.245 mmol), and toluene (10 mL) was added at room temperature.
The brown-purple suspension was stirred for 48 h at 70 °C, cooled to
room temperature, and filtered. The volume of the filtrate was reduced
to 2 mL, and the deep purple solution was cooled at −20 °C. The
small dark purple crystals that formed were collected by filtration and
given relative to δ 0 (TMS). Kinetics studies were performed using a
1
Bruker DRX-500 MHz H NMR at a given temperature ( 0.1 °C)
and by integrating data manually for each spectrum. All the data were
integrated relative to an internal reference (toluene, dihydroanthra-
cene, or the grease peak) to track mass loss during the progress of the
reaction. The measured loss of intensity did not exceed 5% over the
kinetic study.
1
dried under reduced pressure (77 mg, 52%). H NMR (toluene-d8,
Synthesis of Ligands. 4,5-Diazafluorene. The ligand 4,5-
diazafluorene was prepared according to a published procedure by
reducing 4,5-diazafluorenone by an excess of hydrated hydrazine
(NH2NH2·H2O; 16 h at 100 °C).8,9 After extraction in CH2Cl2, the
pure 4,5-diazafluorene was sublimed at 80 °C under reduced pressure.
1H NMR (CDCl3, 300 K): δ (ppm) 8.78 (d, 2H, J = 7.8 Hz, H3,6 or
H1,8), 7.94 (d, 2H, J = 7.6 Hz, H1,8 or H3,6), 7.35 (dd, J = 8.0 Hz, H2,7),
3.92 (s, 2H, −CH2). Mp: 172 °C dec (lit.8 mp 172 °C). Deuterated
4,5-diazafluorene-9,9-d2 was prepared in a similar way using
ND2ND2·D2O (% H ∼5%). 1H NMR (CDCl3, 300 K): δ (ppm)
8.79 (d, 2H, J = 7.6 Hz, H3,6 or H1,8), 7.92 (d, 2H, J = 8.0 Hz, H1,8 or
H3,6), 7.34 (2H, dd, J = 7.8 Hz, H2,7), 3.92 (s, 0.08H, −CHD), 3.90 (s,
0.04H, −CH2); 94% of the CH2 group is deuterated. 2D NMR
(CDCl3, 300 K): δ 3.92 (s, CD2).
300 K): δ (ppm) 122.19 (2H), 31.54 (2H), 16.78 (2H), 11.83 (1H),
3.69 (30H, Cp*) H NMR (thf-d8, 297 K): δ (ppm) 123.51 (2H),
1
31.94 (2H), 16.85 (2H), 11.57 (1H), 3.66 (30H, Cp*). 1H NMR (py-
d5, 297 K): δ (ppm) 124.38 (2H), 32.32 (2H), 17.16 (2H), 12.02
(1H), 3.98 (30H, Cp*). Mp: 275−278 °C dec. Anal. Calcd for
C31H37N2Yb: C, 60.97; H, 6.11; N, 4.59. Found: C, 61.04; H, 5.94; N,
4.48. Vis−near-IR (λ, nm (ε, cm−1 M−1)): 426 (1250), 502 (360), 538
(405), 581 (395), 634 (210), 963 (40), 908 (30), 1001 (55). IR
(cm−1): 2954 (w), 2924 (w), 2854 (m), 2725 (m), 1561 (w), 1528
(w), 1460 (s), 1408 (w), 1377 (s), 1327 (w), 1300 (m), 1261 (m),
1260 (m), 1164 (w), 1090 (s), 1020 (s), 947 (w), 888 (w), 798 (s),
781 (m), 726 (m), 684 (m).
Cp*2Yb(4,5-diazafluorenyl-d) (2-d). The synthesis was similar to
that used for the unlabeled complex from 51 mg (0.099 mmol) of
1151
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
Figure 1. Numbering scheme and names of the ligands used in this work.
Table 1. Solid-State Properties of Complexes 1−4
compd
color
mp (°C)
IR (cm−1
)
χT(300 K)
μeff(300 K) (μB)
Cp*2Yb(4,5-diazafluorene) (1)
Cp*2Yb(4,5-diazafluorenyl) (2)
Cp*2Yb(4,5-diazafluorene-d2) (1-d2)
Cp*2Yb(4,5-diazafluorenyl-d) (2-d)
(Cp*2Yb)2(9,9′-bis-4,5-diazafluorene) (3/3-d2)
dark green
dark red
273−275
275−278
273−275
275−276
2923, 2900
3044, 3009
2183, 2117
2360, 2342
0.15
2.55
0.17
2.54
0.31
1.1
4.52
1.2
dark green
dark red
4.51
1.6
a
dark brown
deep purple
b
(Cp*2Yb)2(9,9′-bis-4,5-diazafluorenyl) (4)
a
b
Obtained as a powder. Isolated product.
Cp*2Yb(OEt)2. Mp: 275−276 °C dec. IR (cm−1): 2923 (s), 2854 (s),
2360 (m), 2342 (m), 1529 (w), 1518 (w), 1463 (s), 1377 (m), 1328
(w), 1319 (w), 1309 (w), 1261 (m), 1204 (m), 1091 (m), 1018 (m),
799 (s), 780 (m), 737 (w), 724 (m), 668 (w). H NMR (toluene-d8,
297 K): δ (ppm) 11.98.
model corrects the absorption surface using a spherical harmonic series
based on differences between equivalent reflections. The structures
were solved by direct methods using SHELX. Non-hydrogen atoms
were refined anisotropically, and hydrogen atoms were placed in
calculated positions and not refined for 4, but their positions were
refined for 2.
2
In Situ Synthesis of (Cp*2Yb)2(9,9′-bis-4,5-diazafluorene-d2) (3/3-
d2). A 1:1 brown mixture of 9,9′-bis-4,5-diazafluorene-d2 and
Cp*2Yb(OEt2) in C6D6 was stirred vigorously for 30 s and then
transferred quickly into a NMR tube adapted with a J. Young valve and
quickly cooled in liquid N2 (77 K) once outside the drybox. The
sample was transported to the NMR spectrometer, rapidly warmed to
room temperature, and then inserted into the probe, precooled at a
given temperature. The kinetics data were collected from 290 to 315 K
RESULTS AND DISCUSSION
■
The results are presented in five separate sections: (i) synthesis
and general characterization; (ii) solution mechanistic studies
of the thermal elimination of dihydrogen as 1 (Cp*2Yb(4,5-
diazafluorene)) goes to 2 (Cp*2Yb(4,5-diazafluorenyl)); (iii)
electronic structural studies of 1; (iv) computational results; (v)
a general discussion that presents our molecular level of
understanding of the origin of the reactivity of 1.
1
over at least 3 half-times. H NMR (C6D6, 300 K): δ (ppm) 79.20
(4H), 25.08 (4H), −0.591 (4H), 5.40 (30H, Cp*), 3.84 (30H, Cp*).
The kinetics were also obtained at room temperature using an
analogous procedure but followed by visible spectroscopy for the
labeled and unlabeled complexes.
Synthesis and Characterization. The complex Cp*2Yb-
(4,5-diazafluorene) (1; see Figure 1 for the structural formulas
and numbering system for the ligands used in this article) is
prepared by mixing a cold (0 °C) toluene solution of
Cp*2Yb(OEt2) with a cold (0 °C) toluene suspension of 4,5-
diazafluorene. The resulting brown powder sublimes as dark
green crystals at 190 °C at 10−2 mm, gives a M − 1+ molecular
ion in the mass spectrum, and melts at 273−275 °C. Attempts
to grow crystals from solution were unsuccessful, and those
obtained from sublimation were not suitable for an X-ray
diffraction study. The temperature (0 °C) for the synthesis is
important, since reactions carried out at room temperature give
1H NMR spectra that contain resonances due to several
metallocene compounds. In the 0 °C temperature synthesis, the
Isolation of (Cp*2Yb)2(9,9′-bis-4,5-diazafluorenyl) (4). A cold
diethyl ether solution of Cp*2Yb(OEt2) was added to an ether
suspension of 9,9′-bis-4,5-diazafluorene at −77 °C. The brown
solution that formed was stirred at −77 °C for 5 h, warmed to −40
°C, and stored for 2 days at −40 °C and 5 days at −20 °C. The brown-
purple solution was then stirred at 0 °C for 12 h and at room
temperature for another 12 h. Solvent was removed under reduced
pressure, and the brown residue was analyzed by 1H NMR
spectroscopy. Two sets of signals in a 1:1 ratio were observed. The
two products were identified as complexes 2 and 4. Complex 4 was
identified only by X-ray crystallography, since only a small amount was
obtained. The purple X-ray-suitable crystals were obtained by cooling a
concentrated (30 mM) solution to 10 °C.
X-ray Crystallography. Single crystals of compounds 2 and 4
were coated in Paratone-N oil and mounted on a Kaptan loop. The
loop was transferred to either a Bruker SMART 1000 or SMART
APEX diffractometer equipped with a CCD area detector. Preliminary
orientation matrixes and cell constants were determined by collection
of 10 s frames, followed by spot integration and least-squares
refinement. Data were integrated by the program SAINT and
corrected for Lorentz and polarization effects. Data were analyzed
for agreement and possible absorption using XPREP. A semiempirical,
multiscan absorption correction was applied using SADABS. This
1
resulting H NMR spectrum is, however, largely free of these
additional resonances.
In order to examine the origin of the temperature effect on
the synthesis, the 1H NMR spectrum in C6D6 was monitored at
60 °C over time at millimolar concentration. The resonances
due to 1 convert into a major product with a 90% conversion
after 2 days. The product is shown to be the complex
Cp*2Yb(4,5-diazafluorenyl) (2) by heating 1 in toluene at 70
1152
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
°C for 2 days. Crystallization from toluene gave 2 as dark red
crystals in moderate yield. Some physical properties of 1 and 2
and their deuterio derivatives are given in Table 1. Compound
2 is derived from 1 by the loss of a hydrogen atom; when the
reaction is monitored by NMR spectroscopy, H2 is detected.
An ORTEP of 2 is shown in Figure 2; crystal data and
packing diagrams are available in the Supporting Information.
C(25)−C(26) angle is 107.7(2)°, the geometry at C(25) is
trigonal planar. (ii) The C(24)−C(25) and C(25)−C(26)
distances are 1.421(3) and 1.417(3) Å, respectively, in the
range found for C(sp2)−C(sp2) distances in nonfused hetero-
cycles of 1.41−1.43 Å.13 (iii) The average Yb−C(Cp) distance
of 2.609 0.009 Å is close to the YbIII−C(Cp) distance of 2.59
0.01 Å in [Cp*2Yb(bipy)]+[Cp*2YbCl2]− 14 and significantly
shorter than the YbII−C distance of 2.74
0.04 Å in
Cp*2Yb(py)2.15 (iv) The average Yb−N distance of 2.372
0.001 Å is identical with the YbIII−N distance in [Cp*2Yb-
14
(bipy)]+[Cp*2YbCl2]− of 2.372 0.005 Å and significantly
shorter than the YbII−N distance of 2.565
0.010 Å in
Cp*2Yb(py)4.15 These bond lengths and angles in 2 are in
agreement with a Cp*2Yb fragment based upon Yb(III) and a
diazafluorenyl fragment that carries a negative charge.
Accordingly, the Yb in 2 is f13, which implies that Yb is f14 in
1, since 1 and 2 differ by a single hydrogen atom.
1
Although 1 and 2 differ by one hydrogen, their H NMR
spectra show that both are paramagnetic compounds; the
chemical shifts at 25 °C are given in the Experimental Section.
The variable-temperature 1H NMR chemical shifts for 2 shown
as a δ vs 1/T plot (Supporting Information, Figure S1) are
linear in temperature and therefore follow Curie behavior. This
behavior is expected for an isolated Yb(III) paramagnet and
consistent with the formulation deduced from the X-ray
Figure 2. ORTEP for Cp*2Yb(4,5-diazafluorenyl) (2). Ellipsoids are
given at the 50% probability level, the non-hydrogen atoms are refined
anisotropically, and the hydrogen atoms are refined isotropically.
1
structure: viz., Cp*2YbIII(4,5-diazafluorenyl). Although the H
NMR chemical shifts of 1 are indicative of a paramagnetic
molecule, a plot of δ vs 1/T shows that all of the resonances
except those due to Cp* are decidedly nonlinear (Supporting
Information, Figure S2). The nonlinear behavior is reminiscent
of the resonances observed for Cp*2Yb(bipy) and its
substituted derivatives.1−4,7,14 The nonlinearity is the first clue
that the electronic structure of 1 is based upon intermediate-
valent ytterbium, rather than divalent ytterbium, a point
developed below.
Table 2. Selected Bond Distances (Å) and Angles (deg) for
Cp*2Yb(4,5-diazafluorenyl) (2)
1
Another notable difference between the H NMR spectra of
1 and 2 is the different solvent dependence of their chemical
shifts at 25 °C; the resonances of 1 depend strongly upon the
solvent, while those of 2 do not. Thus, the most deshielded
resonance, due to 2H's in 1 in toluene-d8, is a singlet at 71.1
ppm that moves upfield to 45.6 or 35.8 ppm in either thf-d8 or
pyridine-d5, respectively. The other resonances due to the
diazafluorene ligand shift slightly ( 3−4 ppm), and those due
to the Cp* resonances are shielded by about 1 ppm in either
thf-d8 or pyridine-d5; the values are given in the Experimental
Section. Although the resonance at δ 71.1 ppm in toluene-d8 is
not assigned with certainty, it is likely due to the CH that is α
to N, since these resonances are the most strongly deshielded
in the bipy adducts due to their proximity to the paramagnetic
center, giving rise to a large dipolar contribution to the total
chemical shift tensor.1,2 A possible interpretation of these
solvent effects is the equilibrium shown in eq 1. Synthetic
support for this postulate is obtained by heating 1 in pyridine-
d5 for several days at 60 °C in an NMR tube. The resulting
material has NMR resonances due to Cp*2Yb(py)2 and free
diazafluorene. Since the pyridine adduct is diamagnetic and 1 is
paramagnetic, the equilibrium illustrated in eq 1 implies a
reversible electron transfer. A similar equilibration between
Cp*2Yb(bipy) and 4,4′-Me2bipy and related adducts was noted
previously.4,14
bond
A
1.414(3)
1.354(1)
1.338(1)
1.403(1)
1.375(1)
1.398(1)
1.445(3)
1.419(1)
B, O, av
C, N, av
D, M, av
E, L, av
F, K, av
G, J, av
H, I, av
bond or angle
Yb−C (distance range)
Yb−C(ring), av
Yb-Cp(cent), av
2.599(2)−2.632(2)
2.609 0.009
2.31
Cp(cent)−Yb−Cp(cent)
Yb−N, av
torsion angle N−C−C−N
torsion angle C−C−C−C
140
2.372 0.001
0.5
0.7
Some selected bond lengths and angles are given in Table 2.
Four features of the solid-state structure are important relative
to the electronic structure. (i) Only one hydrogen atom is
attached to C(25), which is located and refined isotropically.
The angles C(24)−C(25)−H and C(26)−C(25)−H are
124(2) and 128(2)°, respectively, and since the C(24)−
1153
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
concentration dependence, the reaction is not a unimolecular
decomposition. A crossover experiment was studied by mixing
equimolar amounts of 1-h2 and 1-d2 in C6D6; following the
reaction at 60 °C shows that H2 and HD are both formed, but
the amount of H2 is greater than that of HD, as a result of the
primary isotope effect. The formation of HD is consistent with
crossover and a bimolecular process.
Cp*2Yb(4,5‐diazafluorene) + xS
⇌ Cp*2Yb(S)x + 4,5‐diazafluorene
(1)
Mechanism Studies. The mechanism of dihydrogen
elimination in the thermal rearrangement of 1 to 2 (Scheme
1) is not obvious but is of considerable interest.
At this point in the mechanistic study, the deductions are
that the rate of the reaction (i) depends upon concentration of
1, (ii) has a large primary isotope effect, (iii) free radicals are
not involved, (iv) neither H2 nor D2 influences the rate but HD
is formed under a D2 atmosphere and HD and H2 are formed
in a crossover experiment.
Scheme 1
The next kinetic study, however, is crucial in the evolution of
a proposed mechanism for the reaction. In a NMR tube, the
C−C-bonded dimer, 9,9′-bis-4,5-diaza-9H-fluorene (see Figure
1 for a drawing) was mixed with Cp*2Yb(OEt2) in C6D6 at
Since the reaction is relatively clean but slow at 20 °C,
1
room temperature. Examination of the H NMR spectrum
according to the NMR spectra, the time evolution was studied
1
within minutes showed that 2-h was formed cleanly (Scheme
2). A similar reaction was observed when the deuterated
analogue was used.
by observing the H NMR and vis−near-IR spectra. In C6D6,
the spectral intensity of 1 was estimated by the total numbers of
hydrogens in 1 and their decrease with time is illustrated in the
Supporting Information, together with a plot of ln(C/C0) as a
function of time, where C is the concentration of 1. The total
number of hydrogens is used, rather than just using the Cp*
resonances, because the Cp* resonances of 1 and 2 overlap at
some temperatures. These plots are nonlinear and cannot be
used to extract a rate law with confidence over the course of the
reaction; these are available (Figures S3 and S4 in the
Supporting Information). Hence, the values of t1/2 for the
reaction were estimated by determining the time at which the
concentration of 1-h2 equals that of 2-h and the values of t1/2
for 1-d2 yielding 2-d were determined similarly. Over the
temperature range of 20−87 °C, the values of t1/2(h)/t1/2(d)
vary from 6.4 to 6.2. These rather large values show that
breaking the C−H or C−D bond occurs before or during the
slow step or steps in the reaction. In addition, the t1/2 values are
used to determine the influence of added H2 or D2 on the
overall rate. Since the vis−near-IR spectrum of 1-h2 has a
feature at 909 nm that does not appear in 2-h, this absorption
was monitored as a qualitative indicator of the reaction rate.
Since the concentration of starting material is an order of
magnitude lower in the vis−near-IR spectrum, the rate is
correspondingly slower. As the reaction rate is much slower,
typically days (vis−near-IR) rather than minutes (NMR) at 20
°C, the reaction was studied over the first 10% of the reaction,
using the method of initial rates. The values of kobs(H) and
Scheme 2
Since this reaction is clean and rapid, the disappearance of
the two farthest downfield resonances was monitored in the
NMR spectrum from 17 to 42 °C over 3 half-lives. The ln(C/
C0) plots as a function of time are linear and yield a first-order
rate constant, kH, of 1.10 × 10−3 s−1 at 27 °C. Similarly, the
value of kD at 27 °C is 1.00 × 10−3 s−1. Plots of these data are
available in the Supporting Information. The kH/kD value is 1.1,
showing that the cleavage of the C−C bond does not involve
C−H bond cleavage before the slow step. The rate of the
reaction as a function of temperature gives ΔH⧧ = 23(1) kcal
mol−1 and ΔS⧧ = 3(1) cal mol−1 K−1. The Eyring plot is shown
in Figure 3. When 3 is allowed to rearrange to 2 under an
atmosphere of D2, no HD is observed in the 1H NMR
spectrum.
The concentration dependence of the rate of reaction of 3 to
2 was also followed by vis−near-IR spectroscopy, since the
concentration is an order of magnitude less than that in the
NMR experiment. The plot of ln(C/C0) against time was
k
obs(D) at 60 °C are available (Figure S5 in the Supporting
Information). The kobs(H)/kobs(D) value of 6.8 is in agreement
with values obtained from the NMR spectra. Since these rate
data are qualitative, additional studies were performed to gain
additional information about the mechanism.
The rate of rearrangement of 1-h2 to 2-h was followed under
an atmosphere of N2 and then compared to the rate under 1
atm of H2 or D2. In each case, the rates were unchanged when
followed by NMR spectroscopy. However, under D2, HD is
observed in the spectrum, but when isolated 2 is exposed to an
atmosphere of D2 in an NMR tube, no HD is observed and no
deuterium is incorporated into 2. The rearrangement was
monitored in the presence of dihydroanthracene or 1,4-
cyclohexadiene as potential radical traps. In each case no
resonances due to anthracene or benzene were observed, and
the rates were qualitatively similar to those observed in the
absence of the traps. As the rate of the reaction shows strong
Figure 3. Eyring plot for the complex [(Cp*2Yb)2(9,9′-bis-4,5-diaza-
9D-fluorene)] (3-d2) going to Cp*2Yb(4,5-diazafluorenyl-d) (2-d).
1154
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
obtained over 3 half-lives, and the kH value was 0.6 × 10−3 s−1
at 20 °C. The near equality of the rates determined in the vis−
near-IR and NMR experiments shows that the reaction does
follow a first-order rate law, and the low value of the entropy of
activation is consistent with the notion that the C−C bond
cleavage step occurs after the transition state. The fact that the
rate of reaction of 3 to 2 is much faster than the rate of reaction
of 1 to 2 is consistent with the intermediate formation of 3
from 1, implying the sequence of elementary reactions as 1 → 3
→ 2. This idea is outlined in more detail in the Discussion.
In the process of collecting the data for the Eyring plot for
the reaction of 3 to 2 (Figure 3), resonances of another
Scheme 3
1
compound were detected in the H NMR spectrum in the
reactions at temperatures below 27 °C. On a synthetic scale,
when 9,9′-bis-4,5-diaza-9H-fluorene is mixed with Cp*2Yb-
(OEt2) in Et2O at −78 °C and the mixture is warmed slowly to
0 °C over approximately 2 weeks, a brown powder identified as
2 by NMR spectroscopy was obtained that contained purple
crystals of a compound identified as 4 by X-ray crystallography.
The ORTEP (Figure 4) shows that 4 is 9,9′-bis-4,5-
ene) and Ru(PPh3)2(H)(N2)(4,5-diazafluorenyl) can be
interconverted by N2 and H2.18
Electronic Structure. The solid-state magnetism of 1, 1-d2,
2, and 2-d was studied over the temperature range 2−360 K,
and plots are shown in Figure 5 and in Figure S13 (Supporting
Figure 4. ORTEP drawing of complex 4. The hydrogens and the
cocrystallized solvent molecule are removed for clarity.
diazafluorenylidene coordinated to two Cp*2Yb fragments.
The Yb−C(Cp*) and Yb−N distances are in the range found
for Yb(III). The two carbene fragments are linked by a C−C
double bond of 1.463(9) Å with a torsion angle between the
two fragments of 39°. In the free ligand the C−C distance is
1.385(4) Å and the torsion angle is 38°.12 The C−C bond in 4
is therefore a stretched double bond and the complex is chiral
with C2 symmetry.
Although the resonances in the NMR spectra due to 3 and 4
are absent during the transformation of 1 to 2 at temperatures
greater than 27 °C, those due to 4 are observed at lower
temperature. Once isolated, 4 is stable at room temperature,
which implies that the rate of the transformation of 3 to 2 is
faster than that of 3 to 4 and 4 is not on the pathway from 3 to
2 and therefore is not the source of dihydrogen. The identity of
4 also is consistent with the postulate that 3 is the 2:1 adduct as
shown in Scheme 2.
The conversion of the diazafluorene adduct 3 to 4 is, in a
sense, related to the conversion of the isomers of (Me3C)2C5H4
to dihydrofulvalene and then to fulvalene (Scheme 3).16 The
analogy between the conversion of CpH or the 3,3′-CH2
annulated to bipyridine is invoked later when the reaction
mechanism is discussed.
Two studies involving formation of H2 from 4,5-diazafluor-
ene coordinated to d transition metals are relevant to the
studies outlined above. Mach and co-workers observed
dihydrogen evolution when 4,5-diazafluorene was added to
Cp′2Ti(Me3SiCCSiMe3) (Cp′ = C5H5, C5HMe4, C5Me5), and
they isolated the Cp′2TiIII(4,5-diazafluorenyl) adduct.17 Song
and co-workers reported that Ru(PPh3)2(H2)(4,5-diazafluor-
Figure 5. Plots of μeff and χ vs T (K) for the complex Cp*2Yb(4,5-
diazafluorene) (1) in the temperature range 2−300 K.
Information). Complex 1 exhibits a low magnetic moment, μeff
= 1.2 μB at 360 K, as observed for the decamethylytterbocene
adducts with various bipyridine ligands.1−4,7 As in these
adducts, when 4,5-diazafluorene is reduced by the Yb(II)
center, a hole is created in the Yb(III) center (f13) as the
electron enters the π* orbital of the ligand, and the calculated
effective moment for the complex is expected to be 4.8 μB at
2
300 K if the spins are uncorrelated (2F7/2 + S).
The low effective moment observed for complex 1 clearly
implies that the two spins are correlated. However, this
information is not sufficient to allow a quantitative description
of its electronic structure. Inspection of the χ vs T plot indicates
that the correlation is not straightforward, since a plateau is
observed in the temperature range 150−300 K (Figure 5).
Moreover, the very low magnetic moment observed in 1 makes
these data susceptible to so-called “Curie tails” from trace
amounts of magnetic impurities, such as J = 7/2 impurities. The
data in Figure 5 were obtained on a sample that was sublimed
and the Curie tail was minimized. Using the sublimed and
therefore highly pure material results in two discontinuities
occurring at 83 and 120 K. Thus, the temperature-dependent
magnetic data of 1 show a very low magnetic susceptibility and
temperature-independent paramagnetic (TIP) behavior in the
high-temperature regime as well as a low-temperature regime
with two distinct transitions between these regimes. In earlier
1155
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
studies, a molecular model was developed from CASSCF
calculational studies that traced the temperature-dependent
magnetic and XANES behavior to one or two open-shell singlet
states that lie below the triplet state. If the two open-shell
singlet states are close in energy, their population change as a
function of temperature can be modeled by a Boltzmann
distribution between these two singlet states, referred to as SS1
and SS2. The TIP behavior is thought to result from van Vleck
mixing of triplet character into the singlet states. The
experimental magnetic and spectroscopic data for Cp*2Yb-
(4,5-diazafluorene) (1) are treated in an analogous way as
outlined below.
The Yb LIII-edge XANES spectra were recorded at different
temperatures over a 30−300 K range (Figure 6). At room
Figure 7. Plot of nf vs T (K) for the complex Cp*2Yb(4,5-
diazafluorene) (1) and 1-d2. The best fits are obtained from a
Boltzmann distribution between two singlet states using the following
parameters: (1) nf1 = 0.81, nf2 = 0.38, ΔH = 5.1(3) kJ mol−1, ΔS =
43(5) J mol−1 K−1; (1-d2) nf1 = 0.81, nf2 = 0.38, ΔH = 5.1(3) kJ mol−1,
ΔS = 42(5) J mol−1 K−1.
Figure 6. Yb LIII-edge XANES data for the complex Cp*2Yb(4,5-
diazafluorene) (1) at various temperatures.
Figure 8. Plots of μeff and χ vs T (K) for the complex Cp*2Yb(4,5-
diazafluorene) (1). The best fit is obtained from a Boltzmann
distribution between two singlet states using the following parameters:
χSS1 = 0.001 74 emu mol−1, χSS2 = 0.000 38 emu mol−1, ΔH = 5.27(5)
kJ mol−1, ΔS = 47.5(2) J mol−1 K−1, Cimp/T = 0.000 55 emu mol−1
K−1.
temperature, two white line features are visible, in agreement
with the presence of both f14 and f13 character for complex 1 in
a 0.62:0.38 ratio. The f14 feature decreases while the f13 feature
increases with decreasing temperature. The magnetic suscept-
ibility data indicated the presence of TIP behavior at lower
temperature. It is possible to rationalize TIP behavior with the
presence of a hole on the Cp*2Yb fragment and an electron in
the ligand that results in the formation of an open-shell singlet
that is lower in energy than the triplet. The electron that is
transferred to the ligand has several accessible π* orbitals,
resulting in several possible open-shell singlets of different
configurations that are lower in energy than that of the triplet.
Each of the open-shell singlets has different magnetic properties
because of the relative amount of the π* contribution to the
configurations of the f13-π*1 states in comparison to the f14
configuration, which is diamagnetic (χ < 0). In the magnetic
data for 1 (Figure 5) two transitions are observed at 83 and 120
K. From these data, it is possible to rationalize the presence of
three possible open-shell singlet states (SS) below the triplet
that have comparable energies and are thermally populated as
the temperature increases, leading to the observation of the two
transitions SS1 ↔ SS2 and SS2 ↔ SS3. However, the
Boltzmann fits only used two singlet states below the triplet
because the valence change between SS1 and SS2 is likely too
subtle to be detected by the nf results, which are only precise to
a few percent. The enthalpy and entropy changes determined
from the Boltzmann fits to these XANES results are in
reasonable agreement with those obtained by similar fits to the
magnetic data and are given in the captions of Figures 7 and 8
and are discussed in the Discussion.
For complex 2, both the magnetism in the 2−300 K
temperature range and the Yb LIII XANES measurements at
room temperature as well as at 30 K are in agreement that this
complex is an isolated f13 paramagnet, Cp*2YbIII(fluorenyl), and
implies that the negative charge is delocalized on the
fluororenyl ligand as deduced above.
The vis−near-IR spectra were recorded for 1, 2 and 3 in
toluene in the 300−1100 nm wavelength range and are shown
in the Supporting Information. The spectra of 1 and 3 are very
similar and present a very broad and intense band in the vis−
near-IR region (700−1100 nm), which is attributed to π* → π*
transitions, since the spectral profile is very similar to those of
both Cp*2Yb(bipy) and Na(bipy).5,14 This band has a
maximum at 909 nm. On the other hand, the spectrum from
2 displays only a set of three low-intensity, yet relatively sharp,
bands in the range characteristic of f13 f → f transitions.5,14,19
The disappearance of the band with a maximum at 909 nm was
used to follow the time dependence and therefore the rate of
reaction because it is a signature of the presence of unpaired
spin density in the ligand in 1 and 3 and is situated in the range
in which 2 does not absorb.
Calculations. In order to gain insight into the electronic
structure of 1, computational investigations were carried out
1156
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
using the methodology described in previous studies.1,7
Geometry optimizations were carried out using f-in-core
RECP (adapted to the +II and +III oxidation states) at the
DFT level (B3PW91). The structure obtained with a +III
oxidation state RECP is in agreement with the X-ray structure
of 2, shown in Figure 2 (bond A is 1.413 Å and the Yb−N
distance is 2.396 Å). The structural parameters obtained for 2
were used in the CASSCF calculations for 1. As reported in the
bipyridine systems,1,7 a CAS definition of 4 active orbitals (2f
and the 2 lowest π*) and 4 active electrons leads to similar
results as a bigger CAS with 8 active orbitals (6f and 2π*) and
12 active electrons, so that the former was used. State-specific
calculations were carried out for the three lowest singlets and
the lowest triplet. The lowest electronic state is predicted to be
an open-shell singlet with a mixture of 84% f13(π*)1 and 16%
f14 in which π* is composed of 55% of π*1 and 45% of π*2. A
second open-shell singlet is found 0.24 eV higher in energy and
is also a mixture of 39% f13(π*)1 and 61% f14 in which π* is
composed of 74% of π*1 and 26% of π*2. The closed-shell
singlet is found 2.38 eV higher in energy. The lowest triplet is
predicted to be 0.55 eV above the lowest singlet and is based on
a pure f13π*1configuration. Despite our considerable effort, it is
not possible to locate a third singlet state lower than the triplet.
These computational results (Figure 9 and Table 3) are in
accord with the experimental observations for 1.
thermochemical quantities ΔH and ΔS are obtained for the
equilibrium SS1 ⇌ SS2 by fitting the Boltzmann distribution to
the values of χ and nf obtained as a function of temperature.
The values of ΔH and ΔS show that SS1 is favored by enthalpy
and SS2 is favored by entropy; at 300 K, SS2 is favored by
about −8 kJ/mol over SS1. A similar thermochemical result is
observed for the methyl-substituted bipyridine adducts. Since
CASSCF calculations indicate that the SS1 has an nf value of
0.84, Yb(III) dominates Yb(II), consistent with shorter Yb−
C(Cp*) and Yb−N bonds in the former adducts. In SS2, the
contribution is inverted and the longer Yb−C(Cp*) and Yb−N
bonds result in an increase in entropy. The computational
results provide a molecular level of understanding of the
thermochemistry. Although the 4,5-diazafluorene and bipyr-
idine adducts of Cp*2Yb have multiconfigurational ground
states, they differ in their reactivity; 1 eliminates H2, whereas
the bipy adducts are thermally stable.
A possible mechanism that is consistent with the
experimental facts for the transformation of 1 to 2 by way of
3 is outlined in Scheme 4. Adduct 1 has biradical character, and
Scheme 4
Figure 9. Representation (extended Huckel calculation) of the three
lowest π* orbitals of 4,5-diazafluorene, their relative energies and
corresponding labels in C2v symmetry.
̈
the 4,5-diazafluorene ligand has unpaired spin density
distributed in various pπ orbitals located on the nitrogen and
carbon atoms. Since the LUMO+1 of 2b1 symmetry has
unpaired spin density at C(9), the p orbital on C(9) is
presumably the “active orbital” that is the trigger for chemistry
at C(9). If this conjecture is correct, the unpaired spin density
at C(9) initiates cleavage of a C−H bond, resulting in
formation of H2 and a C−C bond at C(9) in 3. The mechanism
of this step is not clear, but the experimental studies show that a
bimolecular process is involved. A possible transition state is
shown in brackets, in which H2 is eliminated heterolytically as
the C−C bond forms. This step may account for the HD
formation in the crossover experiment.20 If this step (k2) is
slow, then the rate will show a concentration dependence. Since
3 can be generated and is shown to rearrange rapidly to 2, k3 is
faster than either k1 or k2. The step involving k3 is homolytic
cleavage of the C−C bond that results in formation of the 4,5-
diazafluorenyl adduct 2 or H2 elimination forming 4, depending
upon the temperature. The k2 step is reminiscent of the
dimerization of a cyclopentadiene to dihydrofulvalene and then
to fulvalene,21 as in Scheme 3.16 In the reaction shown in
Schemes 3 and 4, the important step in coupling two
cyclopentadienyl fragments together is generation of a
cyclopentadienyl radical or radicaloid. In Scheme 3, this is
accomplished by oxidation of the cyclopentadienyl anion by
Table 3. State Configuration Fractions for f orbitals and π*
Orbitals As Calculated by CASSCF
label
f13
f14
nf
SS1
SS2
0.84
0.39
0.16
0.61
0.84
0.39
configuration
π*1
π*2
SS1
SS2
0.55
0.74
0.45
0.26
DISCUSSION
■
The interpretation of the experimental studies outlined above is
that the ytterbium complex 1 has an intermediate valence
between +II and +III, while the valence in 2 is +III. Compound
2
2 is therefore an isolated paramagnet based upon a J7/2 state
but 1 is multiconfigurational, like the Cp*2Yb(bipyridine)
adducts reported earlier. The experimental results for 1 show
that at least two open-shell singlet states are lower in energy
than the triplet state and there is an equilibrium between the
two low-lying open-shell singlet states, SS1 and SS2. The
1157
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158

Organometallics
Article
iodine, and in Scheme 4, the coupling is achieved by
coordination of Cp*2Yb to 4,5-diazafluorene that results in
formation of the open-shell singlet ground electronic state of 1.
The set of elementary reactions outlined in Scheme 4 is
consistent with what is currently known about the mechanism.
The individual steps are set forth in the sense of a postulate as a
guide for further study rather than as an established set of facts.
The key postulate, however, is that the electronic structure of 1
stabilizes the biradical enough so that it can be isolated but not
so stable that it does not undergo chemistry.
REFERENCES
■
(1) Booth, C. H.; Kazhdan, D.; Werkema, E. L.; Walter, M. D.;
Lukens, W. W.; Bauer, E. D.; Hu, Y.-J.; Maron, L.; Eisenstein, O.;
Head-Gordon, M.; Andersen, R. A. J. Am. Chem. Soc. 2010, 132,
17537.
(2) Walter, M. D.; Berg, D. J.; Andersen, R. A. Organometallics 2006,
25, 3228.
(3) Walter, M. D.; Berg, D. J.; Andersen, R. A. Organometallics 2007,
26, 2296.
(4) Walter, M. D.; Schultz, M.; Andersen, R. A. New J. Chem. 2006,
30, 238.
(5) Da Re, R. E.; Kuehl, C. J.; Brown, M. G.; Rocha, R. C.; Bauer, E.
D.; John, K. D.; Morris, D. E.; Shreve, A. P.; Sarrao, J. L. Inorg. Chem.
2003, 42, 5551.
(6) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456.
(7) Booth, C. H.; Walter, M. D.; Kazhdan, D.; Hu, Y.-J.; Lukens, W.
W.; Bauer, E. D.; Maron, L.; Eisenstein, O.; Andersen, R. A. J. Am.
Chem. Soc. 2009, 131, 6480.
(8) Dickeson, J.; Summers, L. Aust. J. Chem. 1970, 23, 1023.
(9) Druey, J.; Schmidt, P. Helv. Chim. Acta 1950, 33, 1080.
(10) Kloc, K.; Mlochowski, J.; Szulc, Z. Heterocycles 1978, 9, 849.
(11) Mlochowski, J.; Szulc, Z. Pol. J. Chem. 1983, 57, 33.
(12) Riklin, M.; Von Zelewsky, A.; Bashal, A.; McPartlin, M.; Wallis,
J. D. Helv. Chim. Acta 1999, 82, 1666.
(13) Allen, F.; Kennard, O.; G., W. D.; L., B.; Orpen, A. G.; Taylor,
R. J. Chem. Soc., Perkin Trans. 2 1987, S1.
(14) Schultz, M.; Boncella, J. M.; Berg, D. J.; Tilley, T. D.; Andersen,
R. A. Organometallics 2002, 21, 460.
(15) Tilley, T. D.; Andersen, R. A.; Spencer, B.; Zalkin, A. Inorg.
Chem. 1982, 21, 2647.
CONCLUSION
■
Adduct 1 is an example of the growing class of decamethy-
lytterbocene adducts with 2,2′-bipyridine and its substituted
derivatives in which ytterbium has an intermediate valence: that
is, its valence is nonintegral and lies between +II and +III. The
4,5-diazafluorene ligand in the adduct is also a radical in which
the unpaired spin density is on the π orbitals of the carbon and
nitrogen atoms. The electronic structure of the adduct is a
multiconfigurational open-shell singlet state that lies 0.55 eV
below the triplet state according to CASSCF calculations. In
previously reported ytterbocene adducts with open-shell singlet
ground states, the biradical adducts are sufficiently stable,
relative to the hypothetical triplet biradical, and they do not
undergo chemistry. In contrast, the biradical 1 is an example in
this class of adducts that is sufficiently stable to be isolated and
characterized but is able to undergo chemistry; the heterolytic
cleavage of C−H bond generating the Yb(III) amide, 2, and
dihydrogen. Adduct 1 is therefore not just a molecule with
unusual magnetic properties; its biradical character also triggers
breaking and making chemical bonds.
(16) Brand, R.; Krimmer, H.-P.; Lindler, H.-J.; Sturm, V.; Hafner, K.
Tetrahedron Lett. 1982, 23, 5131.
(17) Witte, P. T.; Klein, R.; Kooijman, H.; Spek, A. L.; Polasek, M.;
Varga, V.; Mach, K. J. Organomet. Chem. 1996, 519, 195.
(18) Stepowska, E.; Huiling, J.; Datong, S. Chem. Commun. 2010, 46,
556.
(19) Carlson, C. N.; Kuehl, C. J.; Da Re, R. E.; Veauthier, J. M.;
Schelter, E. J.; Milligan, A. E.; Scott, B. L.; Bauer, E. D.; Thompson, J.
D.; Morris, D. E.; John, K. D. J. Am. Chem. Soc. 2006, 128, 7230.
(20) Boche, G.; Lohrenz, J. C. W. Chem. Rev. 2001, 101, 697.
(21) Doering, W. E. Theoretical Organic Chemistry, the Kekule
Symposium; Butterworths: London, 1959.
ASSOCIATED CONTENT
* Supporting Information
Figures, text, tables, and CIF files giving information
concerning magnetic data, vis−near-IR spectroscopy, kinetics,
and X-ray crystallographic data (CCDC 896946 (2) and
CCDC 896947 (4)). This material is available free of charge via
the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: raandersen@lbl.gov
■
Present Address
∥Laboratoire Het
́ ́ ́ ́
eroelements et Coordination, CNRS, Ecole
polytechnique, F-91128 Palaiseau Cedex, France.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Director, Office of Science,
Office of Basic Energy Sciences, and by the Division of
Chemical, Geosciences, and Biosciences of the Department of
Energy at LBNL under Contract No. DE-AC02-05CH11231.
X-ray absorption data were collected at the Stanford
Synchrotron Radiation Lightsource, a national user facility
operated by Stanford University on behalf of the Department of
Eneergy, Office of Basic Eneergy Sciences. L.M. is member of
the Institut Universitaire de France. Cines and CALMIP are
acknowledged for a generous grant of computing time. L.M.
would also like to thank the Humboldt Foundation for a
fellowship.
1158
dx.doi.org/10.1021/om300876b | Organometallics 2013, 32, 1150−1158